Rationality in Question: On Eastern and Western Views of Rationality

RATIONALITY IN QUESTION On Eastern and Western Views of Rationality EDITED BY

SHLOMO BIDERMAN AND

BEN-AMI SCHARFSTEIN

E.J. BRILL LEIDEN • NEW YORK • K0BfcNHAVN • KQLN 1989

046004 ISSN ISBN

0169-8834 90 04 09212 9

© Copyright 1989 by E J. Brill, Leiden, The Netherlands All rights reserved No part of this book may be reproduced or translated in any form, by print, photoprint, microfilm, microfiche or any other means without written permission from the publisher PRINTED IN THE NETHERLANDS

Contents

PREFACE INTRODUCTION: For a Reasonable Rationalism

vii ix

PART I: ON THE NATURE OF RATIONALITY Rationality and Logic, by J. Kekes

3

Rationality in Decision Theory and in Ethics, by H. Putnam

19

Modern Myths Concerning Rationality: The Imperatives of Consistency and (Constrained) Maximality, by R Sylvan

29

The Rationality of Asceticism, by R. W. Perrett

57

On The Rationality of Context, or, How Rational Attention to Context Leads to Total, Irrational Relativity, by B. -A. Scharfstein

77

PART II: COMPARATIVE RATIONALITY The Rational and Irrational in Medieval Jewish and Islamic Philosophy, by L. E. Goodman

93

Confucius, Mencius, Hume and Kant on Reason and Choice, by J. J. Kupperman

119

Rationalism and Anti-Rationalism in Pre-Buddhist China, by A. C. Graham

141

On the Limits of Empirical Knowledge in Chjnese and Western Science, byN. Sivin

165

Dharma and Rationality, by B. K. Matilal

191

Indian Philosophical Tradition: The Theory of Pramana, by J. N. Mohanty 217 Thinking on Empty: Madhyamaka Anti-Realism and Canons of Rationality, by M. Siderits

231

CONTRIBUTORS

251

INDEX OF NAMES

255

Preface

The present book is the first in a series dealing with comparative philosophy and religion, East and West. The editors of this series are members of the Department of Philosophy of Tel-Aviv University, which has for years maintained a program of comparative philosophy and religion. The two subjects are taught together because their separation would be false to the historical experience we are interested in understanding. The editors are now planning a second volume, the tentative name of which is Myths and Fictions: Their Place in Philosophy and Religion, East and West. We invite any reader who would like to contribute to this volume to communicate with us. We are grateful to The Isaac Manasseh Meyer Fund, created by Sally Meyer, for the means to publish the present book and to continue with the planned series. We are also grateful to Tel-Aviv University, in particular to its Research Authority, for help in the administration of the Fund. Finally, we thank Norma Schneider for her careful editing of the manuscript, for other help in its technical preparation and, in general, for the devotion with which she has carried out this work.

Introduction: For a Reasonable Rationality S. Biderman and B. -A. Scharfsteln

The dictionary tells us, quite plausibly, that a rational person is one endowed with reason, that is, one who not only has the ability to re_ason_but.wliQ_also makes proper use of this ability. By definition, such a person cannot easily be foolish, absurd or extravagant, either in thought or in action. It is evident that philosophers, whose usual goal is rationality, pride themselves on their ability to reason especially well; what they think of their actions is less sure. The effort to reason especially well has led some of them, such as Plato and Plotinus, to the ideal of reasoning that leads beyond reason itself, up to the ineffable experience of what they take to be reality as such. Other philosophers, such as the Neo-Positivists of the twentieth century, whose devotion to reason does not allow them to believe in the possibility of surpassing it, take the ideal to be the kind of clarity and intersubjective agreement found in mathematics and physics, a clarity and agreement with the help of which they hope to improve our usual human failure to be consistently rational. Still other philosophers, in our generation often those inspired by Wittgenstein, allot reason a negative or therapeutic function and hope to use it to resolve the muddles in our thought and so, perhaps, in our action. The attitude of modern Western writers, such as Kafka, resembles that of Buddhist thinkers in often displaying a quite skeptical attitude toward rationality. Consider, as a lighthearted example, the conversation in Wonderland between Alice and the Cat: "But I don't want to go among mad people," Alice remarked. "Oh, you can't help that," said the Cat: "we're all mad here. I'm mad. You're mad." Now, when the Cat says, "We're all mad here," he includes himself. But surely, if the cat is mad, then his claim that everyone else is mad should be taken seriously only in a really mad

S. Biderman and B. - A. Scharfstein world. However, if he is sane enough to realize his own madness and to say so explicitly, maybe he is not mad after all. For it is surely sane to recognize the mad impulses that a sane but merely human creature can never wholly escape. Whether made by a Carrolian Cat or a philosopher, skeptical self-references are challenging. The Indian skeptic (and Buddhist believer) Nagarjuna says "All statements are empty" — his own statement included. But surely, if all statements are really and wholly empty, then it is not reasonable to take any statement seriously — including the one that states that all statements are empty. However, if Nagarjuna is self-conscious enough to admit the emptiness of his own words, maybe there is a hidden grain of something like truth in them, by way of a hint so subtle that it seems to deny itself. When one tries to draw a first, bold map of rationality, its clarity and coherence are obscured if one takes account, as logic and conscience dictate, of paradoxes, riddles and dilemmas. Put metaphorically, one can see these paradoxes, riddles and dilemmas as "the enemy within," the cunning instruments by which reason confronts and confounds itself. The obstacles in the way of anyone who wants to construct a system that will explicate the notion of rationality are massive. The first obstacle is the inevitable ambiguity of the notion itself. What do we mean by rationality? Serious answers to this question must twist and turn elusively. As Max Weber has succinctly put it, "A thing is never irrational by itself, but only from a particular point of view."1 Beyond this first obstacle, there are the difficulties involved in trying to produce an adequate model of rational human action. One can develop Weber's analysis and offer all sorts of distinctions between formal and substantive rationality, objectively and subjectively rational actions, and so on. Or, one can advance further and elaborate on the intricate distinctions involved in any analysis of the notion of rationality, such as that made between means and results, or between factual procedures or tactics and desirable ends — all in the attempt to construct a consistent, plausible model of rational human action. However that may be, even if such an effort at explication were successful, there would still remain the immense gap between the description of rational human action and its application to human action in fact. As the paradoxes of the Cat and of Nagarjuna

For a Reasonable Rationality demonstrate, it may well be the case that life as we really live it does not always presuppose or require rationality in any direct sense. Or, as Jon Elster has nicely put it, "Contradictory and paradoxical intentions may be intelligible in terms of a project even if that project is not rational."2 Until quite recently, few Western philosophers have been attracted to this 'irrationalistic' extreme. Since Descartes at least, the debates between most philosophers have been on the measure of success of their use of reason, some of the philosophers claiming or appearing to claim a total rationality, and so arousing the suspicion that their passion to be reasonable, like any other unchecked passion, has led them to be foolish, absurd or extravagant. If we share this suspicion, we might say of them that they were rational in that they were possessed by the ideal of rationality and wielded its instruments with a seemingly impartial yet unreasonable ruthlessness, because their claim was excessive and therefore unwise. We should then be making a common enough distinction between, on the one hand, rationality in the sense of faith in the ideal of rationality and in one's own ability to fulfil it, and, on the other, reasonableness in the sense of one's awareness of the fallibility both of the ideal and of all particular versions of it — including, of course, one's own. By this distinction, rationality might be attributed to seventeenth-century Rationalists such as Spinoza and Leibniz, while reasonableness might be attributed to Hume, who stressed how swayed he was by animal faith, or to Montaigne, who took such pleasure in describing his own and others' inconstancies and inconsistencies. Surely, a wise philosopher (we assume no necessary contradiction between the adjective and the noun) should have learned enough from the history of thought, if not from personal experience, to restrain his claims. Seen in this light, the philosopher's problem is to reconcile the passion he may have for rationality with the doubts held by any experienced and reasonable person. If the reconciliation is successful, we may be tempted to think the philosopher not only reasonable but also wise; what else can we call the passion to understand that has been tempered but not subdued by experience? It is now perhaps a generation since philosophers whose basic inclination is positivistic have recognized the unattainability of their ideal of verification in science and philosophy, and, in philosophy, of discourse purged of everything irrational. Logic itself

S. Biderman and B - A. Scharfstein has come to appear more mutable than before, more humbled by its need to confront paradoxes, more adjustable to different needs. Physics has long since exceeded our ability to visualize things; it has learned to achieve great predictive success with theories known to contain mathematical inconsistencies; and in quantum mechanics it speaks of effects that may reasonably be considered to be without causes. Sometimes it seems as if physics attempts to reconcile the equivalent of Plato's Ideas with the probabilism of his unlikely heirs, the Academic skeptics, not to speak of the free swerve, ungoverned by any law, introduced by the ancient atomists. In even greater contradiction to what has been considered rational, current cosmological speculation may take the universe to begin in a random fluctuation of the vacuum and, without violating even the law of the conservation of energy, to appear by 'accident' out of 'nothing' — 'nothing,' that is, except (as Plato would approve) for the pre-existing laws of nature, with whatever they incorporate of mathematics and logic.3 Now we are better aware of the importance of randomness, which Charles Peirce thought intrinsic to nature, and of the importance, stressed by Henri Poincare, of small initial differences that lead to great final ones. We have become aware that determinism itself often makes prediction impossible because almost all dynamical systems are subject to deterministic 'chaos,' the fundamental order of which is of a sort that precludes knowing in advance just what will happen. 4 Yet it is not surprising that philosophers, like others, still have difficulty in adjusting their thought to a world that cannot be easily characterized as either deterministic or indeterministic. Whether we tend to believe in full determinism or not, what are we to make of, and how can we name, a world in which no experiment can be reproduced really exactly, if only because its initial conditions cannot be reproduced with absolute accuracy; in which observations of the solar system or any other astronomical system cannot be fully accurate because full accuracy would have to take (impossible) account of every body, force and physical relationship in the cosmos; in which the subsystems of a deterministic astronomical or other system may appear undetermined although they are not really so; in which events may be both determined and unpredictable; and in which macroscopic phenomena may exhibit essentially the same kind of indeterminism as microscopic ones? 5 What complicates our understanding of physics makes our

For a Reasonable Rationality physics, of course, more adequate to our experience, so that the gap between experience and explanation grows smaller. There is an interesting analogy between the growing complexity and adequacy of physics and what we are learning of the behavior of living cells, for example, the extraordinary adaptability of immune systems to invaders of all sorts with a sufficient likeness to previous invaders. The even more extraordinary adaptability of the nervous system cannot be explained by any kind of hardwired computation we understand, for the 'wires' are overlapping dendritic trees that grow or wither in competition with one another, as is decided by experience and the complex chemistry that mediates the system. A neuron can change its structure as associations are learned 6 ; the early structure of every person's brain is influenced by that person's particular experience7; and the nervous system learns, changes structurally and functionally, and, in ways we can only dimly perceive, translates learning into changes in behavior.8 Everywhere, clear old theories give way to more complicated, more flexible new ones, or at least show their inadequacy to deal with the world we in fact perceive. The once apparently simple unit of inheritance we call a 'gene' is now a player in a game the intricacies of which we are — to our astonishment — just beginning to learn. Though we are in the process of mapping the human genome, when it has been mapped there will still be an apparently endless system of genetic relationships to study. What we learn about genes must change our grasp of evolution. Even if we have begun with a notion of the 'struggle for existence,' we do not know how to translate the pure selfishness this notion appears to teach into the many examples of real or apparent altruism we encounter, among insects no less than among human beings. We do not know why even ants and bees sacrifice themselves for others of their kind, that is, for their colony- or hivemates. Perhaps colonies and hives, given their close genetic makeup, are to be regarded as superorganisms, any part of which is genetically conditioned to sacrifice itself for the good of the whole. If this is true, does this bear on the altruism of genetically more diversified animal societies, including those of human beings? Beyond immunology, neurology and ethology, the human sciences are undergoing a process that fits them to give more adequate explanations. The philosophical study of language is being

S. Biderman and B. -A. Scharfstein broadened, made more accommodating and potentially linked with adjoining disciplines. In anthropology, the study of the symbolism and the cognitive systems of peoples distant from us bares more of the nature of these peoples' rationality and, often, reasonableness, and is beginning to teach us that rationality has many possible expressions different from those we ourselves most often cling to. Although the Western tradition alone developed Euclidean geometry with its strict modes of proof, alone developed an elaborate symbolic logic and alone developed the theoretical structures of modern science, we are learning to honor the rationality, reasonableness and wisdom of cultures other than our own. We are learning, that is, to suspend the natural conviction of our superiority — natural in the sense that its counterpart is held by the members of every culture — and to begin trying to explore other cultures and modes of thought without prejudice. Everything we as editors are saying we say for ourselves; but although we chose the contributors to this book on the grounds of their competence and their interest in the problem of rationality, we have found that they too reflect what we take to be a contemporary, more complex, more open, more reasonable view of what rationality is or, rather, should be perceived as being. Thus, John Kekes insists, reasonably enough, that conformity to logic is not a necessary condition of rationality and that there is no canonical decision-procedure for rationality, but that this conclusion — which is so drastic in relation to our traditional ideals — does not doom rationality, which he describes as the endeavor to proceed from genuine problems by the use of suitable methods to the achievement of valuable goals. It should be added that Kekes has written elsewhere that a satisfactory conception of rationality should balance what is theoretical and truth-directed with what is practical and action-directed, but that he does not see how to remedy the inability of the traditional conception to reach such a balance. 9 Hilary Putnam finds a flaw in rational preference theory; and then he brings up a related, theoretically serious point, the solution which shows that ethics and science have a basic likeness. Putnam is concerned with the person who makes a moral decision he wants to be rational, and who therefore makes it only after considering what its outcome will probably be. However, the decision may be of a kind that will never recur — maybe once

For a Reasonable Rationality made, it cannot be rescinded, or maybe it is made toward the end of one's life, when few moral decisions of any kind remain. If the decision is unique in an individual's life, it cannot be thought to be one of a series of similar ones, and probability (on a frequency interpretation) has nothing to say of it. A solution considered by Putnam is for the decider to regard himself as the member of a moral community. Then his decision is only one in a series of similar ones made by other members of the community, so that thinking in terms of probability becomes theoretically reasonable. This solution by means of identification with others is what Charles Peirce suggests for scientists: A scientist takes himself to be one of a community of investigators, so that he can regard a single experiment he undertakes as only one of a series, even if he himself never repeats it. On this analogy, to be reasonable, the moral vision by which one acts must be fundamentally social. But though Putnam respects Peirce's perceptiveness, he regards his view, at least as applied to ethics, as incredible, for he has come to the conclusion that there is no exclusively and exhaustively correct basis for ethics. Instead, insists Putnam, for critical decisions each individual falls back on moral convictions for which he or she can give no final rational justification. Richard Sylvan points out that, in practical work as in logic, the usual requirements for rationality have prevented the acceptance of more adequate practices and theories. In broad agreement with Kekes, Sylvan holds that consistency is not a necessary condition for true rationality, for there can be adequate grounds for holding explicitly contradictory beliefs. As if meaning to strengthen the position of Kekes and Sylvan, Roy Perrett gives the particular example of asceticism. His view is that asceticism — which he characterizes as the doctrine that one ought to deny or minimize a whole class of one's desires in favor of some moral or spiritual good — can be rational. To show this, he says, is to shed light on the nature of rationality itself. BenAmi Scharfstein, who deals with the idea of context, argues that a full acceptance of the need to establish context leads to an extreme that in effect nullifies itself. That is, when pushed to a rational extreme, consistency in the pursuit of contextual understanding proves to be self-defeating and therefore less than reasonable. At this point, the focus of the book changes from the nature of rationality in general to rationality as displayed in other

S. Biderman and B. -A. Scharfstein cultures. Thus, Lenn Goodman writes of a number of Islamic and Jewish thinkers who, he believes, show us that we cannot simply identify the irrational with the emotional, the religious or even the mystical. As in the case of Plato and Aristotle, says Goodman, reason may and should warrant, inform and guide the content of our religious experiences. "We are seeking boundary-conditions of rationality-seeking, that is, not so much to apply as to define a concept, to define it in use." The spirit that pervades his essay recalls Einstein's remark that the most incomprehensible thing about the universe is its comprehensibility, its marvelous yielding to the human effort to understand, as if, that is, its very substance were rationality. Here, in the medieval thinkers whose models of rationality Goodman explains, we experience an ideal that we may or may not share, but that has its pride and use, if not as a metaphysical certainty then at least as a Kantian Ideal — as something we cannot avoid assuming for our own welfare even though we cannot prove it. Even Hume, despite his famous statement that reason is and must remain a servant to the passions, believed that reason is important in discovering that which will fulfil the passions and enable us to constitute, recognize and modify our moral beliefs. Finally, for Hume, as no doubt for Einstein, reason itself "is nothing but a wonderful and unintelligible instinct in our souls." In such words, Hume transfers the unintelligibility of God, almost unquestionably accepted by the Islamic and Jewish thinkers, to the nature and presence in us of reason, which he sees as the unintelligible source of intelligibility. Speaking of ancient Chinese philosophy, Joel Kupperman points out that we must not expect Confucius and Western philosophers to be answering exactly the same questions. The method he exemplifies is that of a modern Westerner submitting Confucius and Mencius to an interrogation unlike that quite natural to them or their students. In the course of his judicious comparison and interrogation, Kupperman distinguishes carefully between the persons involved, in order not to oversimplify any of them. He takes care not to require an impossibly logical clarity and consistency from them, but also not to create any artificial barriers between them and ourselves. His abstract and untrivial conclusion is that all perhaps of the Western and Eastern thinkers he deals with share the insight that there should be a

For a Reasonable Rationality special impersonality when one speaks or thinks of what is right or on the whole best. A.C. Graham, who finds the plainest example of rationalism in the Chinese tradition among the Mohists of about 300 B.C., points out that the same contrast between philosophy and protoscience that developed in Greece developed in China as well, except that in Greece the contrast was sharpened by the use of syllogistic logic. He agrees with the view that the language and thought of China were correlative, but holds that correlative thought was also characteristic of Europe. Graham points out that most Chinese thinkers do not share the Mohist and Western hope of a perfect system of definitions that would establish an independent realism of reason, but says that this hope does not mark the difference between rationalism and anti-rationalism, for many modern philosophers have found it necessary to abandon this hope except in logic and mathematics. We should not forget, he concludes, that such 'irrationalistic' poetry as that of Lao-tzu breaks down dichotomic thinking and helps Chinese thought to preserve the latitude that allows creativity and originality to flourish. Nathan Sivin shows another aspect of Chinese thought, that of the limits of scientific exactness, especially as evidenced in the arguments of Chinese astronomers over the indeterminacy they discovered in their data. Though he is careful to point out contextual differences and to argue that the appeal to the inherent irregularity of nature was a defense against new, more precise theories, the quotations he cites show a curious similarity to the positions of Peirce and Boutroux, as well as to contemporary doctrines — whether of the inherent randomness of particles, the deterministic chaos of dynamic systems or the increasing theoretical precision that may show that apparent exceptions in fact conform to a hitherto unknown rule. Sivin demonstrates both the likeness and unlikeness of scientific rationality as perceived in the West and in China. He also proposes that the professional astronomer Shen Kuo (Shen Gua, 1031-1095) was the first person in history to make a clear distinction between our unconnected experiences and the unitary causal world we postulate to explain them. (It could be argued that a view like Shen Kuo's was inherent in Heraclitus, Plato and even Democritus.) When we turn to India, we find B. K. Matilal's affectionate

S. Biderman and B -A. Scharfstein assessment of dharma, which he defines as the standard of moral virtue implicit in traditional Indian culture. Referring to the old literature of the epics for stories that pose moral issues and even dilemmas, Matilal points out that in ancient India the Buddhists, Jains and Hindus all refused to believe in a God who had established an authoritative moral code, and that this refusal left the way open to reasoned human judgment. Even the Hindus, who obeyed scriptures considered authoritative in their own right, made reasoned examinations of the scriptural injunctions. Matilal's final judgment is that, while the ancient Indian writers were often primitive in their judgments by our standards, they often judged wisely, even when they decided that a decision was too difficult. And for this, he says, they deserve our praise. The project of J.N. Mohanty is somewhat similar to that of Matilal, for he tries to look at the Indian tradition from within, so that he can continue the tradition by advancing it creatively. Mohanty explains the Indian view of rational proof by means of reliable 'instruments of perception,' and takes this view to be rather different from anything found in Western epistemology. By defining the Indian tradition in terms of its concerns rather than its explicit beliefs, he frees himself, he says, to learn from it and use it. Taking on the subtle, troubling Indian philosophy of 'emptiness,' Mark Siderits interprets its chief exponent, Nagarjuna, as meaning to say that an acceptable canon of rationality must reflect human needs, interests and institutions, and cannot, therefore, be absolute. He then draws an interesting parallel between this view and contemporary anti-realism, which also holds that philosophical rationality is doomed to the extent that it seeks a truth free of all taint of human needs and interests. And he then suggests that the tactics used in India and in the contemporary West are not, perhaps, very different. Like a contemporary anti-realist, the ancient Indian counterpart systematically displays the paradoxes that arise when worldly categories are subjected to analysis in accord with the prevailing standards of philosophical rationality. True, Nagarjuna had his notion of paradoxes, as we have ours, but the upshot, according to Siderits, can be similar. We have not tried to summarize the various essays this book contains. Fortunately, they are rich in the sort of substance to which a summary is inadequate. We have, however, tried to show

For a Reasonable Rationality that there is a general similarity of spirit in the essays, all the more impressive for its having been neither planned nor expected. This spirit is in accord with what we take to be the increasing sophistication and adequacy of scientific and philosophical thought, of a kind that allows us to hope — is it against hope? — that in philosophy too there can be a kind of general progress of rationality in the sense of reasonableness.

Notes 1. 2.

3. 4. 5. 6. 7. 8. 9.

M. Weber, The Protestant Ethic and the Spirit of Capitalism, trans. T. Parsons (New York: Scribner, 1958), p. 194. J. Elster, Ulysses and the Sirens: Studies in Rationality and Irrationality, revised edition (Cambridge: Cambridge University Press, 1984), p. 153. J.D. Barrow, The World within the World (Oxford: Oxford University Press, 1988), p. 231. P. Davles, The Cosmic Blueprint (New York: Simon and Schuster, 1988), p. 53. I. Ekelund, Mathematics and the Unexpected. (Chicago: University of Chicago Press, 1988). Scientific American, Jan. 1989, pp. 16-17. A. Aoki and P. Siekevitz, "Plasticity in Brain Development," Scientific American, Dec. 1988. G.M. Edelman, Neural Darwinism (New York: Basic Books, 1987). J. Kekes, "Some Requirements of a Theory of Rationality," The Monist, 71(1988), 3.

Part I On the Nature of Rationality

Rationality and Logic J. Kekes

I It is a basic assumption of the Western intellectual and moral tradition that rationality is a central value. I think that this assumption is correct, but many thoughtful and decent critics disagree. My purpose here is to argue that one type of criticism of rationality is misdirected due to a misconception about the relationship between rationality and logic. The misconception I have in mind is shared by many defenders and critics of rationality, so in arguing against it I shall also be arguing against a particular conception of rationality. According to this conception, or rather misconception, conformity to logic is a necessary condition of rationality. If we can understand why this is not so, we shall have come much closer to understanding just why rationality should be highly valued. Part of my thesis is that 'rationality' is a controversial concept. I cannot, therefore, begin with a definition of it. But we must begin somewhere, and I shall do so by sketching out an intuitive conception of rationality, not as a definition, but merely as a starting point. It is obvious, I think, that we all want to have as much truth and as little error as possible. We have to live in the world, and we have to respond to it in order to satisfy whatever needs or wants we have. Success in this endeavor requires having a correct understanding of the aspect of the world to which we are responding. This correct understanding gives us the truth we all want to have and saves us from the error we all want to avoid. But understanding comes in degrees, and, correspondingly, we have various degrees of confidence in its deliverances. These degrees are indicated by three related concepts: knowledge, justification and rationality. Knowledge is the possession of truth and the avoidance of er-

4

J. Kekes

ror. When we know that something is so, we have the truth and we are not mistaken about what we know. Justification is both weaker and more inclusive than knowledge: it is having sufficient grounds for believing that something is so. If we are justified in believing that something is so, we have reasons that warrant the belief that it is so. A justified belief may amount to knowledge, but it may also fall so far short of it as to be actually false. The explanation of this is that there are situations in which it is reasonable to believe what subsequently turns out to be false. So justification includes knowledge, and it may also include error, as well as the many intermediate states between knowledge and error. Rationality is even weaker and more inclusive than justification. It indicates engagement in the enterprise of trying to find the truth and avoid error. We are rational if we are concerned with trying to justify and criticize our beliefs so that we increase our chances of achieving truth and avoiding error. Rationality does not guarantee that we shall achieve these goals, it merely indicates the resolve to pursue them in a particular way. The question I shall be considering is the relationship between this initial understanding of rationality and logic. I take logic to comprise predicate logic (including Aristotelian logic and its scholastic refinements) and mathematical logic (incorporating propositional logic, set theory, and modal and intensional logic). One of the aims of logic is the codification of arguments in a formal system, in a calculus. The calculus can then be applied to test the validity of arguments. Conformity to logic, the observance of the rules of valid inference, thus provides one necessary condition of the rationality of arguments. According to this view, logic is like an initial screening of the candidates for truth; it eliminates many, which, due to their violations of logical rules, have no chance of being true. I shall begin to question this view by asking about the source of the authority that logical rules have. Clearly, such rules as identity, non-contradiction and the excluded middle are binding. We disregard them at our peril. But, what gives such logical rules their obvious authority? II 1 Let us begin with the suggestion that logical rules are necessarily true. The trouble with this is that, unless it is explained why they

Rationality and Logic

5

are necessarily true, the suggestion is no more than a pious injunction to abide by logic. Attempted explanations, of course, are available. The difficulty with them is that there are serious doubts about the necessity of logical rules due to the supposed incoherence of the notion of analyticity, to the availability of deviant and alternative logics and to the extension of fallibilism to logic.2 Thus, before we accept the claim that logical rules are necessarily true rules of correct reasoning, and that is why we should follow them, we should know what makes them necessary, and doubts about them being so should be removed. The second answer to our question attempts to provide an explanation: the logical-illogical distinction coincides with the meaningful-meaningless distinction. We should be logical because, if we are not, we fail to make sense. Thus, Ryle tells us that Underlying the familiar distinction between truth and falsehood, there is a more radical distinction between significance and meaninglessness. True and false statements are both significant, but some forms of words, with the vocabulary and construction of statements, are neither true nor false, but nonsensical — and nonsensical not for reasons of wording or grammar, but for logical reasons. ... [L]ogic is from the start concerned with ... what can or cannot be significantly said. ... All logic and philosophy are enquiries into what 3makes it significant or nonsensical to say certain things. If "meaningful" is used in the standard dictionary sense, meaning the opposite of incoherent babble, gibberish, then this answer fails. For many illogical utterances are meaningful. Indeed, their identification as illogical requires that they be understood and hence be meaningful. If "meaningful" is used in some technical sense, then what that sense is must be made clear. The history of recent philosophy does not warrant optimism on this score. The verifiability principle, for instance, has not fared well in distinguishing between meaningful and meaningless utterances. At any rate, the difficulty with the identification of some technical sense of meaningfulness and logicality is that the sense must be made clear and coherent and the identification must be justified. I know of no candidate that succeeds on both counts. The third answer is that the logical-illogical distinction corresponds to the distinction between what can and cannot be

6

J. Kekes

thought. The historical origin of this answer is the notion that logic is concerned with discovering the laws of thought. This is the controversial doctrine of psychologism.4 The traditional criticism of psychologism is that fallacies demonstrate that the logical does not correspond to what can be thought. The illogical not only can be, but frequently is, thought. The attempt to connect logic with human mental processes, psychologism, is thought to be fallacious because our mental processes are fickle. Some people think better than others, and one and the same person thinks differently at different times. Part of the value of logic is supposed to be that it provides objective standards for appraising how people think. If psychologism were correct, then logic would have to be revised if everybody took a drug that permanently impaired logical thinking. Psychologism, however, has found some recent defenders. 5 The occurrence of fallacies is compatible with logic being a description of the laws of thought if, following Kant, 6 fallacies are attributed to interference with the proper functioning of our faculties. And the question of how even a correct description of how we think can amount to a justification of how we think is answered by denying that there is a sharp distinction between questions of fact and questions of justification. But these claims are unconvincing if viewed as answers to the question of why logical arguments are authoritative. If logic describes how we think, logic cannot justify the claim that we ought to think that way. Logical arguments may be good, but they are not good because they are logical. If they are good, it follows from the present suggestion, they are so for whatever reason it is good to think the way we do. It may be suggested that the concern of logic is not with how we think, but with how we ought to think. No one will deny that we could think better than we actually do. However, the question of why it is logic that would bring about this felicitous improvement is the very one we are considering, so it should not be begged. The fourth answer is that the logical-illogical distinction defines the limits of correct usage: conformity to logic is a necessary condition of correct usage, and failure to conform is a sufficient condition of incorrect usage. According to this view, logical rules express the limits within which language must operate. 7 This fails to answer the question, for it merely redescribes the situation without explaining it. It says nothing about why we are

Rationality and Logic

7

bound by particular logical rules. Clearly, if we are to speak, then some utterances must count as correct and others as incorrect. But why is it that many of the rules that define what is correctly sayable in English are shared by many other languages? Logical rules seem to have a wider scope than the regimentation of correct usage in a particular language, although, to be sure, they do have that role. But if they have only that role, different languages would have different logics, j u s t as they have different vocabularies and phonemes. One way of meeting this difficulty is to identify logic with grammar. The identification rests on distinguishing between the surface and the depth grammar of a language. Depth grammar is then identified with logic. If one is willing to make the additional assumption that depth grammar is connected with innate human mechanisms (physiological or otherwise), then the objection I raised above would be met. For the invariance of logical rules, conformity to depth grammar and, consequently, the rules of correct usage would be seen as depending on universally human mechanisms. 8 But I do not think that this great and fruitful idea helps answer the question I am considering. For even if the extremely speculative hypothesis about innate grammar turned out to be true, the question of why logical arguments are good would still be left unanswered. The identification of logic with depth grammar is a powerful impetus for psychologisiri (or physiologism) in logic. But, as we have seen, it does not follow from the discovery of why we think the way we do that it is good to think that way. If this linguistically inspired approach to logic were correct, then we would know what makes an argument logical. However, we would still not know what I want to find out, namely, why a logical argument deserves acceptance. And this question needs to be answered, even if it turns out that we are so constructed that, if our thinking proceeds unimpeded, then we cannot help thinking logically. After all, there are many things human beings do, and perhaps cannot help doing, but it does not follow that they ought to do them. Would it follow from the discovery of innate mechanisms, say for aggression, hostility or cruelty, that these deplorable traits are now to be valued? The fifth answer will seem tautological, but it is not. The distinction between the logical and the illogical is supposed by many people to be the distinction between valid and invalid arguments.

8

J. Kekes

A logical argument is valid and an illogical argument is invalid. What makes validity valuable is that it is truth-preserving. A valid argument transmits the truth from the premises to the conclusion, and anybody concerned with rationality must recognize that as a good thing. The answer, however, fails, because, while it is true that valid arguments are logical, there is no good reason to believe that invalid arguments are illogical.9 An argument can be said to be valid if it conforms to an established rule of inference. The difficult question is: what makes an argument invalid? The answer that first comes to mind is that an argument is invalid if it fails to conform to an established rule of inference. But this is unsatisfactory because it rests on the unprovable assumption that all rules of inference have been established. There is no reason to suppose that the number of such rules is finite. If they are infinite, then it is impossible to establish all rules of inference. And even if it turns out to be finite, how could one show that the number of established rules coincides with the number of possible rules? Consequently, apparently invalid arguments may in fact be valid, because they conform to not yet established rules of inference. If we change to the claim that an argument is invalid if it fails to conform to a rule of inference, regardless of whether or not it is established, then the difficulties just noted are avoided, but only to be replaced by new ones. The logical merits of arguments do not depend on what rules of inference happened to have been established, but on what rules there are. One consequence is that we cannot, then, distinguish between valid and invalid arguments. For there is no guarantee that the rules of inference we have established do indeed coincide with the rules of inference as they really are. Furthermore, there is a fundamental difficulty even if we do grant that the established rules of inference accurately reflect some of the existing rules. The difficulty is that, whenever an argument fails to conform to an established rule of inference, there are two possibilities: one is that the argument is invalid, the other is that it is valid, because it conforms to a not yet established rule of inference. As long as not all rules of inference are established, these two possibilities will persist. The result is that no argument can be shown to be invalid. Let us now try another suggestion. Why not say that an argu-

Rationality and Logic

9

ment is invalid if it violates an established rule of inference? This differs from the previous two suggestions in that it involves a conflict between an established rule of inference and an argument, whereas the previous two merely noted lack of conformity. But what happens if there is an argument that we find prelogically valid or vouchsafed for by scientific investigation, and it turns out to be invalid because it conflicts with established rules of inference? Let us consider an example of each. The first is made possible by the standard practice of making use of ordinary terms after changing their meaning. Thus, logicians refine the ordinary uses of terms like "and," "if," "or" by codifying acceptable senses and then constructing the truth table on the basis of their codified meaning. Occasionally, this results in there being perfectly valid ordinary inferences that turn out to be invalid given the logically permitted inference. It is a valid ordinary inference to conclude from the statement "I ate with abandon and had to diet" that first I indulged and later I paid for it. But the truth table would not license this inference, since it uses "and" in a nontemporal sense. The ordinary inference is valid because the ordinary meaning of "and" may include a temporal dimension; but the same inference is invalid if the logical sense of "and" is adopted, since it is restricted to a nontemporal sense of the word. Should we then say that ordinary inference is invalid? Surely not. The second example in which valid arguments may contradict established rules of inference comes from the possibility of conflict between logic and science. In quantum mechanics, the statement that a particle with a certain momentum either is or is not at a particular position at a particular time is not taken to be true. Does this mean that the relevant inferences in quantum mechanics are illogical? Not at all! What this statement does indicate is that we need a new logic which would be better suited to quantum mechanics. These two examples show that the third suggestion for determining the invalidity of arguments also fails. For, it does not follow from the fact that there is a conflict between an argument and an established rule of inference that the argument is invalid. It may be that the conflict is a sign that the established rule of inference needs to be revised. If it is true that "[l]ogic is in principle no less open to revision than quantum mechanics or the theory

10 J. Kekes of relativity," 10 then the conflict of an argument with an established rule of inference cannot be sufficient to show the invalidity of the argument. I do not think that this establishes that the invalidity of an argument cannot be determined. But it does produce a dilemma that must be faced, and one I do not know how to resolve. If an argument seems to be invalid, there are two possibilities. Either it is genuinely invalid, or it is a case calling for the revision of established rules of inference. Unless the second alternative is eliminated, the first cannot be justifiably asserted. The consequence is that, as things are now, invalidity need not coincide with illogicality, and so the logical-illogical distinction does not coincide with the valid-invalid distinction. Therefore, while it is reasonable to hold that arguments may be justified logically, it is not reasonable to criticize them logically. The last answer I shall consider is that logical rules are the most general laws of nature. Their generality consists in universal applicability. Thus, it can be truly said of each and every thing that it is identical with itself, or that it cannot both have and not have one and the same characteristic at any given time. According to this suggestion, the authority of logical rules is grounded upon the world being what it is. This answer, however, also fails, for the dissimilarities between rules of logic and laws of nature make their identification implausible. It is far from clear what the laws of nature are. There will be general agreement, however, that, whatever else they may be, they are empirical regularities. Rules of logic and laws of nature differ, because the first do not and the second do admit counterexamples. Thus, we cannot conceive of a thing failing to be identical with itself, but we can conceive of a person flying unaided. The general point is that laws of nature are subject to empirical refutation, while logical rules are not, at least not in the same sense. Any theory which does not allow for the radical difference between the status of physical truisms (such as "All rocks are heavy") and logical truisms (such as "All rocks are rocks") ... must be unsatisfactory. We feel that such a logically true proposition is true not because it describes the behavior of all possible facts but simply because it does not take the risk of being falsified by any fact; it does not exclude any possible fact, and it therefore does not assert anything whatsoever of any fact at all. 11

Rationality and Logic

11

The difference between laws of nature and logical rules is not that the former are subject to revision, while the latter are not. Both may equally be revised in the light of experience. The difference between the one and the other kind of rule is due to the different ways — whether counterexample or degree of serviceability — in which experience may force their revision. [Scientific hypotheses can meet with counterexamples, the propositions of logic and mathematics are not invalidated by experience, but at worst found to be unserviceable. We do not say that Euclidean geometry has been discovered to be false, but only that for certain purposes another geometry serves us better. If we were to give up the law of excluded middle ... it would not be that the proposition "p or not-p", as it is now understood, would be invalid, but only that we had found it preferable to operate with different logical constants. We might retain the same signs but we should give them a different meaning. 12 The objection to the identification of logical rules and natural laws is not that one is analytic and the other is empirical. The objection is that there is an important disanalogy between them. So even if they are both empirical, they are so in different ways. Thus, to explain the authority of logical rules by analogy with the authority of laws of nature fails.

Ill But if the authority of logical rules does not derive from being necessarily true, setting the limits of meaningfulness, establishing what can be thought, defining correct usage, excluding invalid arguments, or being the most general laws of nature, then what gives them the authority they so obviously have? The beginning of an answer is suggested in Locke's trenchant observation that God has not been so sparing to men to make them barely two-legged creatures, and left it to Aristotle to make them rational. ... God has been more bountiful to mankind than so. He has given them a mind that can reason, without being instructed in the methods of syllogizing: the understanding is not taught to reason by those rules; it has a native faculty to perceive the coherence or incoherence of its ideas. 13

12 J. Kekes The authority of logical rules does not come from logicians. Although logicians do indeed formulate the rules, they do not invent and then impose them: logical rules are already there awaiting to be discovered. Logic, therefore, is not a legislative enterprise. This, however, still does not answer the question about the authority of logical rules. The answer is that The validity of deduction depends not upon conformity to any purely arbitrary rules we may contrive, but upon conformity to valid rule. ... But how is the validity of rules to be determined? ... Principles of deductive inference are justified by their conformity with accepted deductive practice. Their validity depends upon accordance with particular deductive inferences we actually make and sanction. If a rule yields inacceptable [sic] inferences, we drop it as invalid. Justification of general rules thus derives from judgments14rejecting or accepting particular deductive inferences. Thus, the authority of logical rules rests upon the successful practice from which they have been extracted. A logical rule is implicit in practices we wish to perpetuate. There were reasoners before Aristotle and strategists before Clausewitz. The application of rules of reasoning and strategy does not have to await the work of their codifiers. Aristotle and Clausewitz were, in fact, only able to extract these rules because they were already being applied. The crystallization of performance-rules in rule-formulae is ... not the condition of their being applied but a product of studies in the methodology of the practice in which they have already been applied. People who construe logicians' rule-formulae as descriptions of the spine and ribs of the world are committing only a more ambitious form of the same error as that committed by those who construe these rule-formulae as premisses requiring to be intellectually acknowledged before intelligent performance can begin. 15 Logical rules are crystallizations of methodological principles that have proved successful" in the past. Logical rules, therefore, presuppose successful practice, and, at least initially, the success of a practice cannot be determined by reference to them. This conclusion needs to be qualified, because not all logical

Rationality and Logic

13

rules are derived from successful practice. Some are derived from other logical rules. Logical rules have consequences that are often very difficult to discern. Discerning them is part of what is meant by research in logic. So I do not want to give the mistaken impression that the task of logic is to scrutinize practice for the purpose of extracting rules from them. The rules have already been extracted. Logic is concerned with making explicit what is only implicit in them. Logical rules have their authority because they supply, in a pure form, the methods employed in successful practices. Of course, logical rules do not and are not meant to characterize the methods of all successful practices. Their role is only to exhibit the method of reasoning from premises to conclusions. Since much reasoning is of this kind, logical rules have an obvious connection with rationality. If we wish to pass successfully from premises to conclusions, we are well advised to employ and abide by logical rules. Thus, part of the relation between rationality and logic is that logic provides an important part of reasoning. But this is only the beginning of the story. Logical rules are extracted from successful practices. But what practices count as successful? Generally speaking, a practice is successful if it accomplishes its goals. This involves solving whatever problems prompt the practice in accordance with whatever goals the agents accept. If the agents are mistaken about having the problems they believe themselves to have, or if the goals in service of which they aim to solve the problems are irrational, then no matter how well-suited the practice is to the solution of the misidentified problem and to the pursuit of the irrational goal, it cannot be said to be successful in any straightforward sense. There is no need to be dogmatic about this. We can distinguish between weak and strong senses of success. A practice may be weakly successful if it leads to the solution of a misidentified problem in accordance with an irrational goal; while the strong sense of success implies a genuine problem and a rational goal. The point I want to make concerns the strong sense of success. In that sense, the correct identification of the problems and the rationality of the goals are necessary conditions of a practice being successful. Since logical rules are extracted from successful practices, the authority of logical rules depends on the correct identification of problems and on the rationality of the goals in-

14 J. Kekes volved in the practice. And this means that we cannot always appeal to the authority of logic in identifying problems and establishing the rationality of goals. There is, therefore, a rational task that precedes the kind of reasoning that can be logically justified or criticized. This argument should not be taken to imply that the identification of problems and establishment of the rationality of goals proceed independently of logical rules. Of course logical rules are involved, for such rules are extracted from successful practice and, at least sometimes, the identification of problems and the establishment of the rationality of goals are successful. What the argument implies is that logical rules cannot be appealed to in determining the rationality of the practice from which they are extracted. The precedence of rationality over logic is not in their temporal order, but in the order of justification. It follows from the fact that logical rules are extracted from successful practices that there must be practices whose rationality cannot be established by appealing to logic. Practice precedes logic. Practical reason is prior to theoretical reason. In the words of one commentator: Kant would wish to put it in this way: in the "practical use" of reason we find grounds for accepting those important propositions which are postulated, but which speculative reason is demonstrably powerless to establish. This seems to say that practical reason yields, so to speak, a bonus or dividend not procurable by any other means. 16 I hasten to emphasize that what the primacy of practice implies is not that some postulates cannot be justified and must be accepted on faith, but that their rationality cannot involve appeal to logic.17

IV It is an unavoidable feature of human life that we must cope with the problems we encounter. The penalty for failure is not to get whatever it is we want. And since all of us want a large variety of things, it is in our interest to cope with the problems that prevent us from getting them. But we do not merely want to cope, we want

Rationality and Logic

15

to cope in order to achieve the goals we value. These goals may or may not ultimately be valuable. We could value a goal and be wrong, because its achievement would not yield the benefit we anticipate, or because it would incapacitate us from pursuing more important goals, or because its pursuit would change us in ways we would find undesirable. If the goals are indeed worthy of pursuit and thus represent desirable solutions to our problems, there is still the further question of how we should get from problems to goals. This question is answered by following or inventing a method that is an appropriate bridge connecting the particular problems we have to the particular goals we value. Rationality is rightly regarded as a central value, because it is the endeavor, indispensable to human welfare, to proceed from genuine problems to the achievement of valuable goals by the use of suitable methods. If we care about human welfare, we must, therefore, care about rationality. Yet, commitment to rationality, thus understood, does not mean that we can find or that we should search for formulae that would provide a uniform decision-procedure for the identification of genuine problems, the formulation of suitable methods and the evaluation of goals. The search for a canonical decisionprocedure is misguided. I have concentrated herein on showing that logic cannot be the object of this search, although logic is the most promising candidate. But it is not the only one. People also look to the physical sciences, to biology, to history, to various sacred texts, to economic forces or to psychology to provide a framework both universal and necessary for connecting problems, methods and goals. One consequence of recognizing that there is no such framework to be found shapes what I take to be the correct response both to a certain defense and to a certain criticism of rationality. Critics of the identification of rationality with a canonical decision-procedure are correct in saying that no such decision-procedure can be found and that the attempt to force life into various Procrustean beds is damaging. But they are incorrect in supposing that this shows that rationality is doomed. Defenders of the identification of rationality with a canonical decision-procedure are correct in valuing rationality as highly as they do and in regarding attacks upon it as undermining human welfare. But they go astray in supposing that the defense of rationality must stand or fall with the finding of such a decision-procedure. The alter-

16 J. Kekes native missed by both these critics and defenders of rationality is that the decision-procedure may be reliable and nevertheless vary with problems, methods and goals.18

Notes 1. In sections II—III. I use material first published as "Logicism," in Idealistic Studies, 12(1982), 1-13. 2. See, e.g., S. Haack's Deviant Logic and Philosophy of Logics (Cambridge: Cambridge University Press, 1974 and 1978). 3. G. Ryle, Collected Papers (London: Hutchinson, 1971), vol. I, p. 252. Ryle is describing Wittgenstein's views in the Tractatus. 4. "Logic, then, is the science of the operations of the understanding which are subservient to the estimation of evidence: both the process itself of advancing known truths to unknown, and all the other intellectual operations in so far as auxiliary to this." J.S. Mill, A System of Logic (New York: Longmans, 1898), Introduction. 5. See, e.g., W.V. Quine, "Epistemology Naturalized," in Ontological Relativity (New York: Columbia University Press, 1969). J. Piaget, Psychology and Epistemology (Harmondsworth: Penguin, 1972). S. Haack, "The Relevance of Psychology to Epistemology," Metaphilosophy, 6(1975), 161-176. 6. "The origin of all error must be sought solely in the unobserved influence of the sensibility on the understanding. ... It is owing to this influence that in our judgments we mistake merely subjective reasons for objective, and consequently mere semblance of truth with truth itself." I. Kant, Introduction to Logic, trans. T. Abbott (New York: Philosophical Library, 1963), p. 44. 7. This view is defended by M. Black, "Necessary Statements and Rules," in Models and Metaphors (Ithaca: Cornell University Press, 1962). 8. See N. Chomsky, Aspects of a Theory of Syntax (Cambridge: M.I.T. Press, 1965), and Language and Mind (New York: Harcourt, Brace and World, 1970). 9. The argument I here present is G.J. Massey's; see his "Are There Any Good Arguments that Bad Arguments Are Bad?" Philosophy in Context, 4(1975), 61-77. 10. W.V. Quine, Philosophy of Logic (Englewood Cliffs: Prentice-Hall, 1970), p. 100. 11. K.R. Popper, "Why Are the Calculi of Logic and Arithmetic Applicable to Reality?" in Conjectures and Refutations (New York: Harper and Row, 1968), 207-208. 12. A.J. Ayer, The Central Questions of Philosophy (London: Weidenfeld & Nicholson, 1973), p. 203. 13. J. Locke, An Essay Concerning Human Understanding, ed. A.C. Fraser (New York: Dover Press, 1959), vol. II, chap. XVII, p. 391.

Rationality and Logic

17

14. N. Goodman, Fact, Fiction and Forecast (Indianapolis: BobbsMerrill, 1953), pp. 63-64. 15. G. Ryle, Collected Papers, vol. II, p. 233. 16. G.J. Wamock, "The Primacy of Practical Reason," in Studies in the Philosophy of Thought and Action, ed. P.F. Strawson (Oxford: Oxford University Press, 1968), p. 224. 17. My view is close to N. Reseller's in Methodological Pragmatism (Oxford: Basil Blackwell, 1977), chaps. XHI-XTV". 18. I have not, of course, said anything in this paper about how the rationality of problems, methods and goals could be determined. I do so in A Justification of Rationality (Albany: State University of New York Press, 1976), and The Nature of Philosophy (Oxford: Basil Blackwell, 1980).

Rationality in Decision Theory and in Ethics" H. Putnam

We are haunted by a certain culturally accepted distinction between "science" and "ethics," but we are also haunted by another culturally accepted distinction, that between "absolute" and "instrumental" values, in effect the distinction between valuing and engineering. Kant himself was in the grip of this dichotomy when he insisted that all "imperatives" must be either "hypothetical" or "categorical." The assumption has always been that hypothetical imperatives, or statements about what one ought to do if one wants to attain a particular end, are unproblematic in exactly the way that scientific statements are thought to be unproblematic. My purpose herein will be to show that this is wrong; that if we are in a position that seems troubling in "ethics," we are in exactly the same position in "engineering"; that the hypothetical imperative is in the same situation as the categorical one; that rationality is as difficult a thing to 'explain' in both cases. I shall begin by discussing the formal assumptions behind what is called "decision theory," because I think those assumptions themselves are more dubious than is usually recognized. Then, in the latter part of the paper, I shall examine a case proposed by Peirce. The Axioms of Rational Preference Theory When I first encountered modern decision theory and rational preference theory in the classic work of von Neumann, which I read many years ago, I was awed. Who wasn't? Like many other people, I felt that von Neumann had succeeded in recovering This paper first appeared in Critica, vol. VIII, no. 54, Dec. 1986. It appears here with minor emendations. 19

20 H. Putnam everything that was sound in classical utility theory without assuming the dubious psychology upon which that theory based itself. But, very soon, I began to have doubts — doubts I would like to share with you. I shall not spend much time on the ingenious use of the notion of an ideally rational gambler which underlies von Neumann's way of getting a utility scale. Von Neumann imagines a gambler who can answer such questions as "Do you prefer a gamble that gives you a chance r of getting X and a chance l-r of getting Y, or a gamble that gives you a chance s of getting X* and a chance Is of getting V, where X,Y,X',Y are themselves "commodity bundles," or combinations of things as different as a concert, a place to live, a friend, etc., and where one's choices are required to be rational as defined by the axioms for "rational preference." And he proves a beautiful theorem showing that, if your preferences are perfectly rational and defined on such "gambles," then it is possible to assign utilities to the individual commodities in a way that rationalizes all the bets you are willing to make. What I want to discuss in particular is the notion of "rational preference" itself. (In the second part, I shall talk about the other great notion of decision theory — that of "subjective probability.") Of course, in order to derive his result, von Neumann needs to assume certain axioms for "rational preference." These axioms imply that all choices — including what are intuitively choices of "incomparable" alternatives — can be rank ordered: any two "commodities" are either (a) unequal (one is preferred to the other), or (b) equal (I am "indifferent" as to which I shall get). Many students feel uncomfortable when they encounter these axioms. The student often wants to say that there is a tertium quid, that he may, sometimes in his life, have to choose between "incomparable" alternatives, and that choosing between "incomparable" alternatives is not the same thing as choosing between alternatives that are "indifferent" from his point of view. But the economist, or whoever, challenges the student to give the distinction between indifference and incomparability any content. And here the student normally finds himself stuck. I want to defend the intuitive sense that a real distinction has been wiped out. First, however, let us identify the relevant axiom. I don't think anyone objects to the assumption that — for an ideally rational agent — preference is transitive. Thus, using "xPy" to symbolize "x is preferred to y," we have

Rationality in Decision Theory and in Ethics

21

{xPy & yPz) -*xPz. What the rational preference theorist is assuming is something much stronger, however. He is assuming that both preference and its complement are transitive. That is, he is assuming both the validity of the previous axiom and that (~xPy & ~ yPz) -+~xPz. (In words: "If x is not preferred to y and y is not preferred to z, then x is not preferred to z.) Together these two axioms justify the claim that "indifference is an equivalence relation" {"xly" — the chooser is indifferent as between x and y — is defined by xly = df~ xPy & ~ yPx). The properties of an equivalence relation — reflexivity, symmetry and transitivity — follow from the above axioms together with the assumption of the irreflexivity of preference. The "work," however, as measured by the strain upon intuition, is done by the second axiom: the transitivity of the complement of the preference relation. Why is it so difficult to reject this axiom? The difficulty that one appears to get into if one rejects this axiom can be easily sketched. Suppose I prefer xto y and I claim that a third "commodity" z is indifferent1 to both x and y, in the technical sense that zlx & zly. Then a decision theorist can, it seems, convict me of being "irrational" in my expressed preferences by the following argument (analogous to a "Dutch book" argument 2 in probability theory): "Suppose," he says, "I were to offer you a choice between x and y. Since you prefer x to y, you would choose x. But, suppose instead that you are confronted with the alternatives of x and z. Since you are indifferent, you cannot complain if, instead of offering you a choice, I just give you z rather than x. If you complain, that would show that, after all, you did prefer x to z, contrary to your expressed statement that ~ xPz & ~ zPx. Isn't that right?" Having gotten you to agree, he goes on: "But now, having gotten you to agree that it's all right if I give you z, I can say 'Since you don't care whether you get z or y, and it's turned out to be inconvenient for me to give you z after all, I will give you y instead'. If you complain at this stage, that will show you did prefer y to z,

22 H. Putnam contrary to your expressed statement that ~ yPz & ~ zPy. But, if you don't complain, then in two steps I will have 'moved' you — with your consent at each step — from receiving x to receiving y, that is, from a preferred to a less preferred alternative." What I have to do, in order to make good on my program of defending the intuitive objection to the assumption that the relation I is transitive, is defuse this argument. But this is not easy to do, not because the argument is invulnerable, but because it rests on nothing less than a whole way of thinking, and it is necessary to expose that way of thinking, and not just to think of a "counterexample" to an axiom. A counterexample is, nonetheless, needed. And here is a simple one: Suppose I am torn, as Pascal imagined me to be in his famous "wager," between an ascetic-religious way of life and a hedonistic-sensual one. I may be quite sure that if I choose the hedonistic-sensual way of life, I would prefer to have a beautiful and responsive lover to a plain and unresponsive one. Call these choices x and y, and let z be the ascetic-religious life. If I regard the two ways of life as "incomparable," then I might insist that, prior to my making my existential choice, ~ xPz & ~ zPx, and also ~ yPz & ~ zPy (i.e., xlz and ylz). This is certainly the kind of case the student has in mind when he calls for a distinction between "incomparability" of alternatives and mere indifference. Why is the student not convicted of irrationality by the argument I just described, the argument that is so analogous to a Dutch book argument? The problem with the Dutch book type argument just described is very simple: it ignores the one value that cannot itself be represented as just one more "commodity" to be combined with the various "bundles," among which or between which I am to choose — the value Kant called autonomy, or the value of making the choice myself as opposed to having it made for me. It is part of calling the choices between x and z and between y and z choices between ways of life, that I view them as choices to be made by me in the process of deciding who I am to be. Regarding all choices as choices between external goods, goods that someone else may alot to me provided he respects my subject value assignments, is precisely the heart of the bureaucraticmanagerial outlook that underlies the whole subject of decision theory. (Of course, reading xly as "indifference" helps to conceal what is at stake.). If someone "decides" to give me x rather than z

Rationality in Decision Theory and in Ethics

23

on the grounds that I do not (yet) have a formed preference between these two ways of life, he deprives me of precisely what is most important to me, namely, that the decision, whichever it is, shall be my own. Could one defend rational preference theory by assuming that an ideally rational agent will already have made all his existential choices? To have made all possible existential choices is precisely to have stopped growing, to have become utterly rigid as a human being. I cannot believe that anyone would really want to pack this attribute of some human personalities into "rationality!" Peirce's Example3 The example I wish to discuss is one that Peirce uses to draw a certain connection between scientific problems and ethical problems, though not the one I would draw. In my opinion, Peirce's great contribution lies in his perception of the depth of individual problems, even if he did not succeed in building a unified system out of all those wonderful perceptions. One of these great flashes of genius occurs when Peirce discusses the question: Why should a person do what is most likely to work? Suppose I am in a situation in which I have to do X or Y and the probability of success is very high if I do X and very low if I do Y. We can put Peirce's question this way: Why should I do X? Why is the fact that X will probably succeed a reason to do it? The Importance of Peirce's Puzzle Many philosophers would say that the reason one should be guided by the probabilities is that the frequency of successes one will enjoy will be higher if one does so. Observe that the case is not one in which the probabilities themselves are at all uncertain; we are supposed to know the probabilities, and so the problem of induction, that is, the problem of ascertaining the probabilities, is not the issue here. The issue is that we know the probability of success is high if one does X, low if one does Y, and the question is why should we do X? Observe also, that the given knowledge is of precisely the type that is supposed to "justify" the hypothetical imperative "Do X if you want success." It is at this point in the argument that Peirce's genius shows

24 H. Putnam itself. Suppose that I am an old man, or that for some other reason I don't believe I have many years of life ahead of me. What do beliefs about what my success-frequency would be if I were to live a long time and be involved in a great many of these situations have to do with what I should do in this one situation? In fact, Peirce considers a situation in which the choice is between "eternal felicity" and "everlasting woe." By the very nature of this situation, there isn't going to be any further "gambling situation" which the rational agent will have to deal with. Specifically, Peirce's thought example is this 4 : One has to choose between two arrangements. Each arrangement is probabilistic; under each arrangement, one will select a card from a well-shuffled pack with twenty-five cards in it, one of which is specially designated. The outcome depends in both cases on whether or not one draws the specially designated card. Under arrangement A, one gets everlasting woe if one draws the designated card and eternal felicity if one draws any other card, so that one's chances of eternal felicity are twenty-four to one; while under the second arrangement it is the other way around: one gets eternal felicity if one draws the designated card and everlasting woe if one draws any other card, so that one's chances of everlasting woe are twenty-four to one. (Those for whom the notion of immortality is troubling can substitute "an easy death" and "a hard death" for eternal felicity and everlasting woe, respectively.) We all believe that a rational person would choose arrangement A. Peirce's question is: Why should he? Reichenbach holds that probability statements about the single case are simply a fictitious transfer of relative frequencies in the long run, 5 or of knowledge of relative frequencies in the long run. Notice that this is yet another example of the use of the notion of a projection; Reichenbach was saying that the very statement that Jones will have only one chance in twenty-five of eternal felicity this one time under arrangement A is a "projection." There is no fact about the single unrepeatable situation, which is the fact that choice A gives Jones twenty-four chances out of twenty-five of eternal felicity. (Recently, Stephen Leeds has written a stimulating paper 6 arguing that the whole notion of probability is a projection.) Peirce's problem comes out very clearly if we take the view that probability just is relative frequency in the long run. The

person in the situation knows a fact which is utterly irrelevant to what he should do. He knows that, if there were a series of situations like this one, then he would have eternal felicity twenty-four times out of every twenty-five, if he were to choose arrangement A each time. But a person can have eternal felicity or everlasting woe only once! His problem is not how to achieve eternal felicity twenty-four times out of every twenty-five; his problem is to obtain eternal felicity this time. Why should he pick arrangement A? The only answer we can give is that it is more probable that he will have eternal felicity under arrangement A. But, remember, the question was: Why should one expect what is probable? If you say that you should expect what is probable because it is likely to happen this time, you're not answering the question; you're just, as it were, repeating the advice: Expect what is probable. If you say, "Well, it's reasonable to expect what is probable," well — in this situation — isn't "reasonable" j u s t a synonym for "probable," in the Keynes-Carnap sense of "logical probability?" Isn't "It's reasonable to expect what is probable to happen" just another way of saying "It's probable (in the logical sense of probability) that what will probably happen (in the frequency sense of probability) will happen in any individual case (unless we know of some respect in which the individual case is atypical)?" We are forced back, then, to the view that a reasonable person adjusts his expectations to the logical probability; and this time, any beliefs we may have about how this will lead us to fare in the long run are seen to be irrelevant to the problem. That there is such a thing as the "logical probability," that it corresponds to the frequency in a long series (if there were a long series), and that a reasonable person adjusts his beliefs to it become just Ultimate Logical (read: metaphysical) Facts. According to Peirce, one can only be rational if one identifies himself psychologically with a whole ongoing — in fact, potentially infinite — community of investigators. It is only because I care about what might happen to people in similar situations that I do what has the best chance in my own situation. My belief that I, in this one unrepeatable situation, am somehow more likely to die easily than by torture is fundamentally, then, just what Reichenbach said it was, a fictitious transfer, on Peirce's view. What is true, and not fiction or projection, however, is that my fellows, the members of the community with which I identify,

26 H. Putnam will have eternal felicity twenty-four times out of twenty-five if they follow this strategy; or, more generally, even if this one particular situation is never repeated, if in all the various uncorrelated cases of this kind or any other kind that they find themselves in they always follow the probabilities, then in the long run they will experience more successes and fewer losses. But can it really be that the reason I would choose arrangement A is that I am altruistic? Maybe I am, but isn't it obvious that I would choose arrangement A first and foremost because it would avoid everlasting woe in my own case? Peirce's argument is that I ought to choose arrangement A for what one might describe as "Rule Utilitarian" reasons: in choosing this arrangement I am supporting, and helping to perpetuate, a rule that will benefit mankind (or the community of rational investigators) in the long run. Is this really what is in my mind when what I am facing is torture ("everlasting woe")? Frankly, it isn't. I cannot give a reason for doing what I would do in this case if the only reasons allowed are in terms of "What will happen in the long run if?" And this shows that, even in the means-end kind of problem, I must fall back on intuitions that I am powerless to explain. Today, many people7 think that the only reason for being reasonable at all is that one will arrive at truth in theory and success in action more often if one is reasonable. Some people8 have even proposed replacing the notion of a "reasonable" method by the notion of a reliable method: one that, as a matter of fact, leads to successful outcomes with a high relative frequency. Notice that (if you agree with me in finding Peirce's own solution incredible) these approaches are helpless in the face of Peirce's problem. If my only reason for believing that I should be reasonable were my beliefs about what will happen in the long run if I act or believe reasonably, then I would have absolutely no reason (apart from the implausible reason of altruism) to think it better to be reasonable in an unrepeatable single case like the one described. In fact, as I came close to the end of my life, and found myself unable to make many more "bets," then my reasons for doing what is reasonable or expecting what is reasonable should diminish very sharply, on this view. The fact is that we have an underived, a primitive, obligation of some kind to be reasonable; not a "moral obligation" or an "ethical obligation," to be sure, but nevertheless a very real obligation to be reasonable, which — contrary to Peirce — is not

Rationality in Decision Theory and in Ethics

27

reducible to my expectations about the long run and my interest in the welfare of others or in my own welfare at other times. I also believe that it will work better in the long run for people to be reasonable, certainly; but, when the question is Why do you expect that, in this unrepeatable case, what is extremely likely to happen will happen?, here I have to say with Wittgenstein, "This is where my spade is turned. This is what I do, this is what I say." My reason for discussing this, when the more usual question today is what to do about the "bottomless pit" phenomenon in ethics, the lack of a Foundation in ethics, is that, in the case just described — a case which has to do with reasonableness about "means and ends," rather than with ethics — my epistemic situation is exactly the same. I do think, and I think it warranted to think, that "acting on the probabilities" is the only rational thing to do, and that one ought to do the rational thing even in unrepeatable situations. In the ethical case, I do think, and I think it warranted to think, that a person who has a sense of human brotherhood is better than a person who lacks a sense of human brotherhood. A person who is capable of thinking for himself about how to live is better than a person who has lost or never developed the capacity to think for himself about how to live; but, whether the question be about single case probability or about ethics, I don't know how I know these things. These are cases in which I find that I have to say, "I have reached bedrock and this is where my spade is turned."9 Recognizing that there are certain places where one's spade is turned; recognizing, with Wittgenstein, that there are places where our explanations run out, isn't saying that any particular place is permanently fated to be "bedrock," or that any particular belief is forever immune from criticism. This is where my spade is turned now. This is where my justifications and explanations stop now. To recognize that a loyal human being is better than a disloyal human being; that a person capable of philia is better than a person incapable of philia; that a person capable of a sense of community, of citizenship in a polis, is better than a person who is incapable of a sense of community or of citizenship in a polis — and so forth — is not to say that any one of these values, or any one of the moral pictures that may lie behind and organize these values, is final in the sense of being exclusively or exhaustively correct. Our moral images are in a process of development and reform. But it is to say that, at each stage in that development

28 H. Putnam and reform, there will be places, many places, at which we have to say, "This is where my spade is turned." None of this goes against the idea that rational criticism of a moral vision is possible. A moral vision may contradict, for example, what we know or think it rational to believe on other grounds, be they logical, metaphysical or empirical. But we cannot any longer hope that these kinds of criticism will leave just one moral vision intact. Ultimately, there is still a point at which one has to say, "This is where my spade is turned." Notes 1. 2.

3. 4. 5. 6. 7.

8. 9.

I say "x is indifferent to y" as short for "the chooser is indifferent as between x and y." A "Dutch book" is a system of bets that cannot result in a favorable outcome for the bettor, no matter how the gambles turn out. The axioms of subjective probability theory are frequently justified by proving that violation of them always leads to the possiblity of being required to accept a "Dutch book." This section of the present paper coincides with the conclusion of my Cams Lectures, The Many Faces of Realism (LaSalle, 111.: Open Court, 1987). Peirce discusses this example in "The Doctrine of Chances," p. 69; reprinted in Chance, Love and Logic, ed. Morris R. Cohen (New York: Harcourt Brace, 1923). Reichenbach discusses the single case in The Theory of Probability (Berkeley: University of California Press, 1949), pp. 372 ff. "Chance, Realism, Quantum Mechanics," Journal of Philosophy, 81(1984), 97-107. For a discussion of this kind of "epistemic utilitarianism," see R. Firth's presidential address, "Epistemic Merit, Intrinsic and Instrumental," in Proceedings and Addresses of the American Philosophical Society, 55(1981), 5-23. For a sophisticated version of this view, see A.I. Goldman, "What is Justified Belief," in George Pappas, ed., Justification and Knowledge (Boston-London-Dordrecht: Reidel, 1979). Philosophical Investigations, see 217. That Wittgenstein here uses the first person — where my spade is turned — is very important; yet many interpreters try to see his philosophy as one of simple deference to some "form of life" determined by a community. On this, see also S. Cavell's fine discussion in The Claim of Reason (New York-Oxford: Oxford University Press, 1979), esp. part one, chapter V: 'The Natural and The Conventional."

Modern Myths Concerning Rationality: The Imperatives of Consistency and (Constrained) Maximality* R. Sylvan

It is easy to become ensnared in issues concerning rationality. I became involved, not through any theory of action or from social theory — or because of my incumbency in a locally prominent school devoted to the social sciences — but primarily as a result of practice, of engagement with environmental issues. There, as in logic, prevailing rationality requirements stand as serious obstacles to the attainment of more satisfactory alternative practices or theories, and as major elements reinforcing prevailing practices and entrenched theories. To illustrate, let us consider only four significant areas of substantial blockage to alternative paths: 1. In forestry, and likewise in fisheries and other primary production areas, rational conservation practices are almost axiomatically taken to have maximum sustained yield as an objective. But in practice that means trying to drive forests, fisheries and farms to the hilt in perpetuity. More moderate, doingwith-enough conservation practices emerge as irrational. 2. The mistreatment and exploitation of animals was, and in some places still is, defended, on the grounds that animals are not (really) rational. A rather similar approach used to be adopted toward indigenous peoples, and is often taken with respect to those accounted mentally ill (e.g., schizophrenics, chronic nonachievers). Rationality has long been, and still is, a basis for discrimination. The format and contents of this paper have been much improved through critical responses to earlier presentations of it in 19851986, at the universities of Queensland, Waikato and Auckland, and through comments of the editors.

30 R. Sylvan 3. In nuclear and associated political debates, deterrence and other current superpower practices are defended as rational, on the basis of maximization principles or positions that incorporate them, e.g., utilitarianism or Bayesianism. 1 4. New logics that seem to get things right where entrenched logics has gotten them badly wrong are criticized as irrational. In particular, paraconsistent logics, 2 which investigate nontrivial inconsistent theories, have been dismissed as violating rationality canons, especially the requirement of universal consistency. The above examples offer sufficient cause for involvement; but there are many other issue-areas that call for engagement on the issue, for instance, what is accounted rational in consumer behavior or in current business transactions. There are two main sorts of options open to reformers and radicals who hang in: first, giving the notion of rationality to the status quo, or giving it away; and second, criticizing and amending entrenched notions. To adopt the first approach, taken by tougher environmentalists, would be to abandon an extremely important and useful analytic tool, and there is little longer-term historical basis or good reason for that. But the second approach, which I pursue herein, leads far beyond minor tampering; it means tearing the fabric of the dominant social paradigm. Separating the Target Myths, of Maximality and Consistency, from their Matrix There are two pervasive and interconnected myths concerning rationality that deserve singling out for concentrated attack. The first of them, which is decidedly modern, is that rationality involves maximality, and the second is that rationality requires consistency. Certainly there are other subsidiary and localized myths connected with and derived from these, for example, the supremacy of classical deductive reasoning (reflecting maximal consistency) and the irrationality of nondeductive methods such as induction. But these are, in large measure, derivative and separable myths. There is also a somewhat different group of ancient myths that connect rationality with (proper) humans and with the divide between humans and other animals. For almost 2000 years Aristotle's dictum that man is a rational animal, together with

Modern Myths Concerning Rationality

31

the ubiquitous Greek assumption that rationality is what separates humans from the brutes, held sway. These assumptions were very damaging to animals, over-generous to humans and detrimental to the development of the social sciences. But these assumptions have already been heavily (in my view, devastatingly) criticized in contemporary environmental philosophy. It is the first group of myths, which underpin much contemporary social science and theoretical biology, that form the present target. To be sure, singling out maximization and consistency requirements for separate criticism is a characteristic analytical activity, already threatening the fabric of interwoven myth. But since deductive principles like Simplication do hold — A & B does entail each of A and B — this is legitimate separation and targeting. Nonetheless, it is important to notice how the two groups of myths, ancient and modern, tie together as a whole, and with what familiar additional assumptions they cohere. Among the characteristically associated assumptions we shall shortly arrive at are, unsurprisingly, those of self-interested individualism. Political control through voting provides a convenient parable of the whole framework in action. There is an ideal of rational control at work here, and the rational individuals in a given region are those who exercise this control. There is a time constraint limiting individuals to presently living ones, and there is the assumption of the state which, through its territory, provides a spatial constraint. Political control is optimized — maximized subject to admittedly demanding constraints, given especially the imperfections of the material involved — in the democratic framework (so a major American myth would have a client world believe, though votes in the super-state are not distributed across clients). Most adult humans in the state are rational (those that are conspicuously irrational are excluded), and it is assumed that rationality is confined to them (and that what is excluded is supposed to be represented by their preferences); this supplying of a base class (local rational individuals) is the contribution of ancient myth. Then, by maximizing on the consistent preferences of the base class, as recorded in voting, consistently aggregated, rational political will can be determined; this is the modern myth at work. Thus such rational political control optimizes the relevant choices of rational

32 R. Sylvan individuals; and this is what the modern democratic framework optimally provides. Any critique of modern rationality and political ideals gets hotly contested (as but little experience reveals); the modern myths now have many salaried apologists. There has been more than 2000 years of intellectual work put into defending consistency, above all as an absolute precondition of rationality (ultimately for political reasons, so Lukasiewicz, who began the process of demythologization, contends 3 ). Moreover, there are now very substantial and elegant theories built on the consistency assumption, for instance, all those that depend essentially on classical logic. Accordingly a great deal is at stake intellectually, including many assumptions that will not easily be relinquished or displaced from their absolute positions. The idea that consistency is not an imperative, that is does not furnish an immediate, decisive test of correctness, is thus generally dismissed as absurd. Those apparent exceptions to the ideal, which do not rest on mistakes, are superficial and disappear under proper analysis (e.g., removal of equivocation, or fragmentation of inconsistent sets into consistent subsets). But, contrary to wellentrenched orthodoxy, there need be no mistakes, and inconsistency, rationally maintained inconsistency, does not always conveniently disappear under analysis. 4 The orthodox defense of the modern imperative of maximization tends to be different, because it is usually admitted that nonmaximizing, but of course irrational, behavior occurs, at least among humans. Instead the orthodox response to objections takes the form that requisite constraints have been neglected; and it almost invariably takes that form after expanding the mix of factors maximized upon (the 'objective function') has failed.5 What is more, as soon as anti-maximization arguments are outlined, or even before, it will be said that nothing novel is involved, nothing that challenges maximization principles. It will also be patronizingly said that the requisite elaboration, effectively fitting constrained maximization (e.g. in administrative practices) into mainstream theory, has essentially all been done and duly put abroad by Simon, and duly recognized, as he won a Nobel Prize largely on the strength of this work.6 As part of my later argument is designed to show, it hasn't all been done by Simon; indeed it hasn't been done at all by Simon. For instance, Simon did not address the biological issues. 7 Even worse, his no-

Modern Myths Concerning Rationality

33

tion of satisficing fits easily within the optimizing paradigm, with satisficing a (time-) constrained maximization. The notion of rationality (e.g., of economic theory) does not change a jot under Simon's adaption of constrained maximization to organization and business theory. By contrast, under the alternative notion to be advanced, that of satisizing as it is now called, much has been changed and much theory is up for change; perhaps most strikingly, the standard axioms of rational choice theory fall.8 The target myths are logically independent. There can be paraconsistent maximizers,9 and there can be consistent satisizers. 10 Nonetheless the myths are commonly coupled, for instance in the logically fundamental notion of the maximally consistent class, 11 and, in more complex fashion, in the pervasive position called Bayesianism, which involves maximization in mainstream economic fashion. But the subjective probabilities, 12 or beliefs, which yield the expectation values, are essentially restricted by consistency requirements. So, rational belief, it is assumed, has to be duly consistent. 13 Consistency requirements only enter in other less conspicuous ways. For instance, the class of exclusive and exhaustive alternatives considered in the selection of an optimal alternative are consistency delimited. However, there is more to the linkage of the target myths than mere coupling; as will emerge, they help support one another. Why bother with these myths? After all, humans tend to cocoon themselves in them. Admittedly, myths are ubiquitous In. human social life and affairs, and many of them are relatively harmless. But the targeted myths are both fundamental in social practices and the social sciences and, as already glimpsed, far from benign, leading to damaging and destructive policies and practices. 14

The Importance of Rationality in the Social Sciences The notion of rationality has substantial roles to play in social anthropology, where, for instance, it is supposed to guarantee common ground between otherwise radically different cultures; in social psychology, where it figures significantly in the determination of normality; and in ethics, where practical rationality is often supposed to underpin much of the theoretical enterprise. But it is in economics and (positive) political science that the no-

34 R. Sylvan tion becomes especially prominent. Indeed, it is even claimed that many economists would "separate economics from sociology upon the basis of rational or irrational behavior, where these terms are defined in the penumbra of utility theory."15 But, while rationality does play a largeTTole in economics than sociology, the notion is also fundamental in sociology, as elsewhere in the social sciences. In particular, the notion of rational individual self-interest or enlightened self-interest - has a long history in the development of social theory. It is the basis of the entire body of economic theory, as well as of efforts towards the development of deductive political science — originating with the social contract theory of Hobbes and Locke, which attempts to 'explain' the construction of the state as a 'contract' based on rational calculations made by individual citizens. 16 The same doctrines extend into contemporary social-choice theory and positive political theory. There are three key assumptions in virtually all this theorizing: individual reductionism; self-interest (i.e., "the assumption that the primary objective of the individual is the maximization of his own welfare"),17 commonly strengthened to possessive self-interest; and rationality, the assumption that the individual is an appropriate maximizer. All these assumptions are in serious doubt, all need heavy qualification, if not outright rejection; but our present concern is not with the first two, those of self-interested or possessive individualism, 18 but rather with the rationality assumption. The matter of rationality is not, however, really separable from the other assumptions. For one thing, individuals are taken to be the basic ontological objects, onto which all else devolves, certainly intensional features such as rationality. For another, and more directly, in attempts to remove anomalies from the mainstream conception of rationality, the notion is typically broken down into more tractable subforms: individual and collective. 19 Individual rationality is explained, as before, in terms of consistency and maximality, but reductionistic schemes are entertained for collective rationality, which is supposed to be arrived at as some sort of synthesis of individual rationalities. But a series of paradoxes (e.g., Newcomb's paradox, which we shall

Modern Myths Concerning Rationality 35 encounter below) have put spanners in these reductionistic works. While such fragmentation and the emergence of paradoxes is both typical and indicative of conceptual breakdown, such features do not establish it. Nor do they much help the credentials of a rival theory of rationality (especially given the range of possibilities). To support such an alternative theory and to underpin the critique of the mainstream conception, some account of and conditions on the notion of rationality are required. Such a return to basics is important for other reasons as well, for instance, to pull the issues out of the hot and charged atmosphere where they tend to be debated and down to earth. There is a connected reason: to sneak around the intentionally forbidding obstacle to fresh investigation provided by the mountain of mainstream literature concerning rationality. Fortunately, as careful groundlevel exploration will show, we can bypass almost all this entire body of intimidating literature, as it incorporates the target myths and is mostly far above the ground.

A Preliminary Account of Rationality, as a Type of (Coherent) Adequacy Notion The common or core meaning of the term "rational," so dictionary definitions reveal, has to do with (the exercise of or capacity for) reason, "based on [adequate] reason," conforming to reason. The dictionary senses of "rational" do not themselves lend any support to maximization (or consistency) themes. On the contrary, they convey a clear picture of moderation: "rational" is also defined as "sensible, not foolish, not extravagant." This picture of moderation, of rationality through mere sufficiency, can be reinforced by appeal to etymological features: the Latin ratio is also the source of the term "ration," meaning "a fixed allowance" or "portion," and, less directly (by way of the notion of thinking or judging), of the terms "rates" and "ratio," which indicate comparative or due proportion. 20 If maximization enters analytically, it must then be in an indirect and unlikely way, through the notion of reason or adequate reason. And, certainly, what counts as reason, or a reason, is eventually critical for any account of rationality. But what a

36 R. Sylvan reason is has been, and remains, a matter of much debate, namely, because of the evaluative loadings of key terms and the different connected value systems embedded in different underlying world views. In "C17 social, political and intellectual arguments, ... a reason associated with faith, precedent and established law was challenged both by the new reasoning and concepts of the reasonable and ... by an appeal beyond (mere human) reason."21 Reason supplied from faith or legal precedent, though it still works wonders for insiders, may carry little weight with someone outside the framework involved, and may indeed be rejected as inadequate or even worthless. Evaluations, and the characteristic framework-dependence of evaluation, thus enter and complicate the picture. This also helps explain why there can be, and are, different and competing accounts of rationality (as there are not of such technical notions as massless particles and the ideal gas). What is rational and rationality have, then, to do with reason and reasoning: that much has already become evident, and is, in any case, what dictionaries and etymology tell us. But fools and psychotics (as well as bishops and judges) reason, sometimes with uncommon cunning. It is not reasoning, any old reasoning, that is required, but reasoning sufficiently well. That is where evaluation enters, strikingly, with one of the most general words of appraisal in the language getting into the act. Terms in the "rational" nexus are evaluative, terms of (logical) appraisal. As James observed more than half a century ago, "the word 'rationality' ... is one of those eulogistic words that both sides claim — for almost no one is willing to advertise his philosophy as a system of irrationality." 22 James was mistaken, however, about the effect of the stigma, and recent history has seen waves of irrationalism — reactions against entrenched (and false) standards of rationality. The evaluative aspect of rationality enables a simple account to be given of the seemingly inconsistent claims made concerning both the stability of the notion and its variability, and as regards the significance of the notion and its apparent increasing emptiness. There is a stable sense to "rational," in terms of having or exercising reason up to acceptable standards; but there is considerable variation in what counts as such a 'satisfactory reason,' depending on the background framework or, more comprehensively, on the underlying paradigm.

Modern Myths Concerning Rationality 37 A uniform account of rationality, which reflects the evaluative element and delivers the dominant requirements of maximality and consistency as limiting cases, has two components: a local analytic component concerned with (sufficient) acceptable reasons and an environmental component concerned with fitting such reasons into a wider framework. The first component has already emerged from the common (dictionary) account. But the second, holistic, component is also essential. Psychotics can have reasons as adequate under given circumstances as those of more normal people for what is known or generally taken for granted. 23 What has happened under the dominant social paradigm is that the local component has been upgraded from sufficient to maximal reasons (the high redefinition step), while the environmental component has been flattened to consistency with what is known, i.e., roughly, coherence is explicated through consistency. Both adjustments have appeal and are now academically pervasive (in part as a corollary of the ideological domination of classical logical paradigm, and in part because academics are usually maximizers, at least in theoretical commitment. That rationality is an adequacy notion rather than a maximizing one comes not only from more ordinary usage, as reflected in dictionaries (sources often despised by philosophers), 24 and not only from etymology; it also comes from reflective accounts offered by philosophers suitably distanced from positive social theory. 25 Thus, for example, Oakeshott arrives at a conclusion to the effect that rationality is "knowledge of how to behave well."26 And, as Black implies, the older expansive rational ideals of 'right thinking' and 'admirable living' were likewise adequacy and not maximizing ideals. 27 The same applies to the common ground, in accounts found by Black, that to be rational "is to use and respect reasons." Although rationality is an adequacy notion rather than a maximizing one, there are undoubtedly maximizing analogues, where adequacy thresholds are pushed to the maximum. But: 1. These maximizing analogues are not as well determined as has been supposed, and, for that matter, not uniquely determined (as various paradoxes reveal). 2. They by no means exhaust rationality notions, and are strictly not rationality notions; and many rational creatures do not conform with them in their practices. They constitute only a

38 R. Sylvan quite proper sub-class of rational procedures, of creatures (those we can distinguish as maximizers), etc. A corollary is that these analogues are only of limited value in social theory; they are valid only for maximizers. This accords with empirical investigations, wherein maximizing notions have not fared well. Thus Eells, an exponent of maximization, reports: "Today, the subjective expected utility maximization theory is generally regarded by psychologists as a poor descriptive theory." 28 That is further motivation for a different theory. If descriptive and prescriptive accounts can be better integrated, a more empirically adequate theory is likely to result. Similarly, although rationality is not defined by consistency, it is constrained by requirements adjoining and often confused with consistency, such as coherence.29 Thus too there are consistency analogues, where sufficient coherence is upgraded to unexceptional consistency; and there is an associated subpopulation of consistencizers, which includes all those strictly committed to requisite features of mainstream logic (i.e., classical or intuitionistic logical theories). The analogues are only analogues however; mere adequacy which suffices for rationality requires neither. Thus rationality requires neither consistency nor maximality, as we shall now try to explain, in greater detail in the case of maximality, since the case against consistency has been made elsewhere (e.g., OP, Priest).

Removing Consistency as a Minimal Requirement of Rationality The imposition of consistency as a minimal requirement is certainly ubiquitous; but for some reason the message is especially strident in and strongly propagated from Scandinavia 30 and Finland, 31 where consistency is logical bedrock. 32 That the assumption is so widespread is one reason why those who cite it rarely bother to adduce any arguments for it. This is perhaps just as well for the assumption, since the main arguments for it rest on rotten foundations, such as the paradoxes of implication, that inconsistency implies triviality. A typical argument for consistency — which applies the received, but thoroughly rotten, theory of entailment as some sort of strict implication 33 — goes as follows: If per impossible a

Modern Myths Concerning Rationality

39

person rationally believed (desired, or whatever) something inconsistent then that person would be rationally committed to everything (which is absurd). 34 The objection, that if this were so people would be unable to hold inconsistent beliefs (which they do), is avoided by arguing, further, that they can, because — whereas rationality requires closure under entailment — the consequences of inconsistent beliefs are not followed through. A further assumption typically appears when this argument is filled out; namely, that no one, not merely no rational creature, can have explicitly contradictory beliefs.35 But both the further assumption and its rationality-qualified repairs are mistaken, as numerous counterexamples from daily intellectual practice, from the history of dialectics and from elsewhere serve to show.36 Indeed this whole classical approach, with these assumptions, begins to fall apart once an improved theory of entailment and intensionality, which can accommodate inconsistency without breakdown, is glimpsed. And it comes entirely apart in the dialethic 37 reaches of paraconsistent theory, where the following counter-themes are asserted and argued: 1. Creatures sometimes have not only inconsistent beliefs (a commonplace matter), but adhere to explicitly contradictory beliefs. 2. It is sometimes rational to adhere to such beliefs. Thus, further, 3. Consistency is not a necessary condition for rationality. The positive argument for Theme 1 is initially by way of cases and then builds on these; the negative argument deals with a range of allegations designed to remove the cases, especially of equivocation, fragmentation and lack of bona fides in the cases. The further argument for Theme 2 appeals to features of rationality, as already discerned, and shows that these are satisfied in some of the cases involved. A related approach to Theme 2 is by way of modeling; a semantical modeling38 shows how all the requisite conditions can be nontrivially met. It is worth sketching certain direct lines of elaboration of Theme 2 (Theme 1 being well-covered in previous work, such as LB). A first argument takes this form: As already concluded, a sufficient condition for rational adherence to a belief is that there are adequate reasons for it. But there are adequate reasons — as adequate as those for a variety of everyday and scientific claims — for inconsistent beliefs39; whence Theme 2. The details

40 R. Sylvan of the argument, filling out the schematic form, are those offered in defense of paraconsistent theories of various sorts (see especially OP). Indeed the reasons for accepting inconsistent theories in foundational areas, such as set theory and semantics, are more solid than those for competing consistent theories. One bolder, more explicit argument along these lines runs as follows: Contradictions, such as those delivered by logical and semantical paradoxes, are believed (by Priest and others) to be fatal to rationality. But there are appealing and acceptable arguments, certainly measuring up to everyday and scientific reasoning standards, for accepting these beliefs. Hence such (paradoxical) beliefs are rational. Whatever kind of argument it takes to make something rationally acceptable, an inconsistent structure can have it (to adopt Priest's way of summarizing a key point). What is more, this sort of argument can be strengthened. For, at least as seen dialethically,40 the contradictions involved are true and the arguments for them deductively sound. But it is eminently rational to believe that what can be so demonstrated is true. 4 1 Thus what is inconsistent can sometimes be rationally believed. For consistency is not required for rationality. What is, in general, required is enough consistency, not total consistency. 42 The Fundamental Argument against Maximization Maximization is a different - and now even tougher - business, infiltrated as it is in basic theories in both the 'natural' sciences (physics and biology especially) and the social sciences (conspicuously, economics). While at first many are prepared to concede that some choosers are nonmaximizers, as soon as it becomes clear that these choosers are not covert constrained maximizers, their rationality gets questioned. On mainstream theories, satisizing rationality has to be repudiated even though it appears to be regularly exemplified in a variety of behavior. 43 It is therefore fortunate for theoretical progress that some of the crucial divisive issues as to maximization can be boiled down to claims concerning the following elementary diagram: Ranking of otherwise undifferentiated objects (of the same sort) in some given setting.

Modern Myths Concerning Rationality

41

A • better

•B adequacy threshold

The claim is that B is a rational choice. The argument for this claim is this: Since B exceeds the adequacy level, there is adequate reason for choosing B; but adequate reason is essentially all that rationality requires here. For such choices can often enough be fitted into a coherent pattern of such choice making. Accordingly, the conclusion is: Maximization is refuted. For A is definitely maximal, but B is a rational choice. That is, non-maximal choice is rational. Observe that the picture I have drawn has been simplified as far as feasible, with a view to focusing issues. But it can be elaborated and complicated in a variety of ways (for instance, to please philosophers, who often prefer complexity, richness of cases and even murkiness). For example, the betterness or quality ranking can be relativized — to contexts, goals, persons, etc. — or it can be seen as some sort of aggregate of relativized or personalized rankings, and so on. Though reductive analyses are mistaken, analysis of betterness is not excluded; but, for the most part, it does not matter how such an analysis goes; it can be of betterness as correct preference (Brentano's illuminating account), of betterness as preferences aggregated over time, as preferences all things considered or all relativized rankings taken into account, and so on. Typically it will not matter. But however the picture is presented, it is subject to much criticism and counterargument; and this is most easily met in terms of the simple picture. The first of many counterarguments is simply this: A is better than B, so it is surely a better choice. But, to respond at once, that is not in dispute. The claim is that, nonetheless, B is a rational choice. Surely (a sure indicator of argumentational deficiency) there must be some further factor relied upon that accounts for the suboptimal choice. For example, the chooser is in a hurry and B is the first adequate object she saw, or she is leaving A for someone

42 R. Sylvan else to take, or A is too good for her, or doesn't fit her lifestyle, or she is trying to devise a simple counterexample to optimization theory, or whatever. But, then, in each case the choice is again optimizing, maximizing under constraints of time, or theory design, or whatever. The reply is that, in some imaginable cases, there may — but need not — be such further differentiating factors. There are also imaginable cases where the chooser is in no hurry, knows little or nothing about optimization theory, and so on, and still chooses B. To insist that there must be a further factor is to presuppose the theory in dispute (as to insist that there must always be a further factor revealing equivocation where someone maintains something contradictory, is to assume a consistency theory). The choice of B would then be suboptimal; but to make suboptimal choices when a better one is freely available is irrational. That is simply to restate the optimizing theme. What is claimed, what the case is designed to show, is that there are rational suboptimal choices. But the chooser has reason to choose A rather than B, namely that A is better. That can be granted; she may also have a reason, of the sort indicated before, not to go on to choose A — or she may not. That does not tell against the rationality of choice of B. But there is an asymmetry in the choice situation that has so far been neglected; namely, the agent has reason to go on to A because A is better. The asymmetry alleged is illusory, since there is a matching reason for persisting with B, namely, because B is (quite) good enough. 44 This can serve, moreover, as a reason for choosing B rather than A. If the agent persists with choice of B, then that shows that her revealed preference is for B rather than A and hence that she does not rank A as better than B. Thus the counterexample is no counterexample and the diagram is simply misleading. Granted, one can work wonders with the 'revealed' terminology, which has proved a great boon in helping economists slide dubious propositions past rightly suspicious audiences. But here it does too much — ordinarily revealed magic relies upon a switching back, before it is too late, to a non-revealed terminology. For it would follow that no one — rational or not — could choose the worse. Suppose an agent chooses E rather than D; then E is revealed as not worse than D. But agents can act irrationally and make irrational

Modern Myths Concerning Rationality

43

choices, for instance, choosing C, which falls below the adequacy threshold. In any event, the divergence of individual-preference rankings from quality (betterness) rankings is hard enough data — straight from ethical theory — though problematic for certain subjectivist and utilitarian positions and, by extension, for mainstream economics. Yet much of the resistance to rational suboptimal choices, and also of the appeal of the so-called axioms of rational choice, appears to come from just such a conflation of preference and quality rankings. A pair of diagrams, showing possible cases, will help to illustrate the point.

Case 1: Preference coincides with quality and diverges from choice. A

•B

better

agent's preference

agent's choice/ revealed preference

Case 2: Preference matches choice and differs from quality assessment. A • better B

agent's preference

agent's choice/ revealed preference

Both cases are possible and both occur; in both cases the agent can be entirely rational; but in both cases some further explanation seems to be called for. In Case 1 an agent's choices cannot be at odds too often with her given preferences, without satisfactory explanation, or doubt builds up about the genuineness of those preferences. However, on any occasion (as distinct from on every

44 R. Sylvan or too many occasions) an agent may choose from adequate objects in preference-reversing ways without irrationality and without needing to offer a reason or further explanation. Of course, we can impute motives; but they needn't imply constrained maximizing, e.g., "he chose to take the Jaguar today though he prefers the Rolls" (normally, so for a change?); "she chose to sleep on her left side, though she prefers the right." Case 2 is more straightforward, because an agent's preference can regularly (and consistently, in this sense) run counter to quality rankings and, indeed, to the agent's own relativized quality assessments as to what is better for her. Countering Rationality "Axioms," and Separating Off Satisjlcing First let us consider axioms for rational choice, which play a fundamental role in microeconomic theory. What has been acclaimed as "perhaps the most basic criterion of rationality in the literature" is the following axiom of rational choice: a rational agent chooses an option A from a set of alternatives S, only if A is at least as good as any other option in S, on the agent's ranking of the set. 45 The fundamental argument, with Set S the pair {A,B}, counters this axiom. The same sorts of arguments also counterexample maximum expected utility criteria for rationality. For, let quality be estimated by way of utility (or replace the quality ranking by a utility ranking); and observe that in a situation of certainty, such as the elementary choice situation affords, expectation values drop out. Then, as under the maximality criteria, A is the rational choice under this utility criterion; B cannot be — what it is — a rational choice. This type of argument also refutes Bayesianism. For, as Suppes claims, the Bayesian position leads "to the overall single principle of rationality, the principle of maximizing expected utility: one action or decision should be chosen over another if the expected utility of the first is at least as great as that of the second ,..."46 (These arguments also refute connected principles, such as that of sure thing, appealed to regularly by Harsanyi 47 in derivations of maximization criteria, and that of ratifiability or "anticipatory Bayesianism," appealed to by Jeffrey48 in a sustained but failed attempt to turn a flood tide of problems threatening to overwhelm Bayesianism.)

Modern Myths Concerning Rationality

45

Satisizing — doing-with-enough — thus counters the orthodox rationality axioms; but satisficing — doing-as-well-as-can-bedone-in-the-circumstances — does not. For satisficing, as is usually explained, amounts to constrained maximization, typically time-constrained maximization. The satisficer, the organization man in Simon's picture, always chooses the maximum from among alternatives already enumerated; thus, in the elementary diagram, he chooses A, not B (as satisizers may). There is already a literature at pains to point out that satisficing does not interfere with the rationality axioms, although that it does not was clear enough from the limited (or bounded) rationality picture Simon built integrally into satisficing. 49 Thus, for example, Riker and Ordeshook 50 try to turn criticisms of the principle of maximizing expected utility on satisficing grounds, by arguing (incorrectly) that, in all practical cases, maximizing and satisficing realize identical results. But their case boils down to a fallacious argument from authority; in the crunch, they appeal twice to Simon's authority: While some enthusiasts have misinterpreted his argument as anti-rational, in fact Professor Simon does not suppose that, when better or worse alternatives are clearly available to the chooser, he may reject the better for worse, just because the worse is satisfactory. ... Certainly Professor Simon is not asking for this, because even in his terms, it is irrational to reject better for worse. As an 51 irrationality satisficing cannot then be expected to occur. We need not join the enthusiasts; but we are supposing what Simon did not: 52 unlike Simon's satisficer, a satisizer may well choose to select not the better, but the less good, just because it is adequate. There may appear to be serious obstacles in the way of theory such as that of satisizing, which repudiates received rationalchoice axioms. For it is now well-enough known that it is possible to reinstate rational-choice axioms — as it is possible to guarantee (a definitional) egoism, which should give pause — by suitable redefinitions of 'interests,' 'preference' and 'utility,' given only consistency assumptions. "If you are consistent, then no matter whether you are a single-minded egoist ... [or not], you will appear to be maximizing your own utility in this enchanted world of definitions."53 The strategy is to replace preference by

46 R. Sylvan revealed preference, and then to define personal revealed utility as "a numerical representation of this 'preference,' assigning a higher [revealed] utility to a 'preferred' alternative. With this set of definitions you can hardly escape maximizing your own utility, except through inconsistency."54 But, since inconsistency is deemed irrational, the rational creature is revealed as a maximizer. The usual response to this sort of hocus-pocus is that such redefinitions are illicit, for a variety of reasons, and should be avoided. For example, Sen has rightly argued against such economic redefinitions of 'preference' and 'welfare' on the dual grounds that "there are non-choice sources of information on preference and welfare, as these terms are usually understood," and that "choice may reflect a compromise over a variety of considerations of which personal welfare may be just one." 55 The present response, however, goes further, in that it also claims that the underlying requirements of consistency not only "have surprising cutting power," as Sen puts it 56 but are excessively strong as well. For inconsistent choices and preferences may well not be irrational. The maximizing result requires, in particular, a consistent pattern of choices over time; but the rational satisizer may not exhibit this. Consider an experiment involving iterated choices of an elementary type (as illustrated in the elementary diagram), where a creature is asked to choose one of A and B, both of which are adequate. Now consider, in this same experimental situation, a satisizer with the following coherent procedure: choose the first adequate object encountered; and suppose the satisizer encounters A in the first trial, B in the second, A in the third, and so on, in an oscillating pattern. From a maximizing angle, now written into requirements of consistency over time (i.e., of regularity, strictly speaking), this is an impermissible procedure. But it is perfectly rational. The Eliminative Impact on Supposed Paradoxes Concerning Rational Behavior Various paradoxes in rational-choice theory now appear in a new light; and in some cases paradox even largely disappears. For the effect or impression of paradox derives either from competing maximization criteria or from analogues of pragmatic or seman-

Modern Myths Concerning Rationality 47 tic paradoxes (e.g., the assertion is true but the asserter does not believe it, a statement that says of itself that it is not rationally acceptable). Thus there are two (again interconnected) groups of paradoxes of rationality: those that are products of consistency imperatives, and those resulting from maximization imperatives. Both types have been said to issue in 'inevitable irrationality'; but what irrationality is involved is no more inevitable than the imperatives. A good example of maximization at work is provided by Newcomb's paradox, which is supposed to arise in the following situation: There are two boxes before you: one transparent and one opaque. You can see that there is $1,000 in the transparent box, and you know that there is either $1,000,000 or nothing in the opaque box. You must choose between the following two acts: taking the contents only of the opaque box or taking the contents of both boxes. Furthermore, there is a being in whose predictive powers you have enormous confidence, and you know that he has already determined the contents of the opaque box according to the following rules: If he predicted that you would take the contents only of the opaque box, he put the $1,000,000 in the opaque box, and if he predicted that you would take the contents of both boxes,57he put nothing in the opaque box. What would you do? The 'paradox' arises because two well-regarded maximizing principles — utility and dominance — which are normally compatible, prescribe different rational choices. The criterion of MEU (maximizing expected utility; here dollar-value) prescribes taking the contents of only the opaque box as rational. The argument is this: if you take the contents only of the opaque box, then, since the predicter is so accurate, he probably predicted you would do that, in which case he would have put the $1,000,000 in the opaque box and you would walk away with $1,000,000. If you take the contents of both boxes, then, since the predicter is so accurate, he probably predicted that is what you would do, in which case he would have left the opaque box empty and you would walk away with only $1,000. 58 However, the principle of dominance prescribes a different choice

48 R. Sylvan as rational, namely, choosing both boxes. According to this (maximizing) principle, if, no matter what circumstances obtain, an agent is better off with A than B, then it is rational to choose A. Now, whether the $1,000,000 is in the opaque box or not, you get $1,000 more by taking the contents of both boxes than you would get by taking only the contents of the opaque box. In recent American literature, elaborate and sometimes ingenious variations on Newcomb's paradox, relying on dominance, have been designed with a view to refuting MEU and Bayesianism. However, maximization in some guise is always assumed. The argument against maximization, insofar as it succeeds, rejects dominance along with MEU; so the framework for Newcomb's paradox and its variations is thereby removed. What does the rational satisizer choose? There is no problem of choice, because all a satisizer can reasonably expect to obtain is $1,000. And if her adequacy threshold is reasonably set at some positive sum not exceeding $1,000, then she will choose both boxes, as under dominance. (Given a zero threshold, where there are no material costs to choice, any choice would suffice.) But suppose she needs large funds for a charitable organization, or is simply greedy (Marx's term for a maximizer), and her threshold accordingly far exceeds $1,000. Then, given the probabilities, she will choose only the opaque box. In no case is there any paradox. 59 What holds for Newcomb's paradox holds near enough for near-equivalent puzzles such as Prisoners' Dilemmas and Commons' Tragedies, which also typically involve conflict of maximization principles. Often they are similarly dissolved; they present serious fixes for maximizers only. But aren't maximizers rational? Not always: the image maximizers have of themselves as preeminently rational is illusory.

Reversing the Charges: The Limited Rationality of Maximization The main thrust of the later argument has been that maximization does not have a monopoly on rationality, that, in enough significant cases, satisizing is also a rational procedure. However, the argument in favor of satisizing and satisficing procedures and practices really shows more than this: it challenges the rationality of maximization. There are now well-known ranges of cases where, even by its own standards, maximizing be-

Modern Myths Concerning Rationality 49 havior is sometimes counterproductive, excessive, undesirable, or the like. 60 In some of these cases maximization is irrational, for instance because (but not only because) it is self-defeating; in many such cases satisizing is, so it can also be argued, more rational than maximizing.

Appendix Towards an Integrated View of Rationality Against any uniform account of rationality such as that proposed and argued from, it is likely to be objected that there can be no such account, because the term "rational" applies to so many different sorts of things: actions, beliefs, enterprises, people, preference... . And it also means different things and plays different roles in different applications. This reflects degenerating fragmentation (recently carried to extremes by Elster). To the protest that dictionaries do not offer any such bundle of different senses, it will be responded that one can hardly rely upon dictionaries. Dictionaries don't even mention consistency, which, according to Elster and others, provides the thin notion of individual rationality, as applied, for example, to action. And how can dictionary accounts be sensibly applied over the range of things involved? How does 'having sufficient reason' strictly apply to individuals, when we speak of rational individuals, or, for that matter, to beliefs or desires, main elements in modern explications of intentional action. To meet these quite legitimate points, something more does have to be done; but fragmentation is neither the right nor the required response. Initial integration can be accomplished by way of the theory of procedures (or processes). The core notion turns out to be that of rational procedure, of which rational action is one special case and rational argument another. Rational states are those characterized in terms of appropriate dispositions to follow rational procedures. A rational belief is one that is founded on rational evidential procedures (including, especially, those of argument). A rational creature is one that characteristically follows and would follow rational procedures in its day-to-day practices. A rational enterprise is one that involves rational procedures and fits into a wider setting of rational procedures. And

50 R. Sylvan so on. So there are reasonable initial grounds for supposing that, if a suitable account of rational procedure can be obtained, then all other straightforward elaborations can also be accounted for along the lines sketchily indicated. What then is a 'rational' procedure? It is a procedure, an agentinfluenced process, that involves the control of reason in some appropriate way by the agent. This account places the matter squarely within cybernetics. It remains, before setting out a typical block diagram, to expand a bit on the control. The 'control' involved can be quite weak, since the (due) exercise of reason is usually considered sufficient. Williams remarks that, in contrast to the term 'reasonable' and most of its cognates, "rational in its predominant sense, has remained relatively constant. It still means having or evidently exercising the faculty of reason, and its negative irrational, quite strictly corresponds to this." 61 He also gives the primary sense of 'rational' (and of 'reasonable') as that of "being endowed with reason or being characterized by reason, as an act or argument." Upon introducing the uniform account already arrived at, the initial, still primitive, cybernetic diagram looks like this: procedure structure

structure

rational control (programming) component

control by agent

agent has reason

procedure62 component

style of components

analytic holistic

The control may be exercised through feedback or in other ways. The agent need not be able to supply a reason; indeed the "agent" may not be the sort of thing that can do such supplying. 63 But whatever passes for the reason must pass an appropriate adequacy threshold. For it is in this, especially, that the rationality of the procedure lies.

Modern Myths Concerning Rationality

51

Notes 1. This and other technical terms will be explained in the notes after their first use. Thus, according to R.C. Jeffrey, The Logic of Decision, 2nd ed. (Chicago: University of Chicago Press), p. 1, in which a full and copiously illustrated exposition is given, Bayesianism supplies a general framework for deliberation, decision and action, on consequentialist lines, incorporating the following features: an "agent's notions of the probabilities of the relevant circumstances and the desirabilities of the possible consequences are represented by numbers that collectively determine an estimate of the desirability for each of the acts [alternatives] under consideration. The Bayesian principle, then, is to choose an act of maximum estimated desirability [i.e., which maximizes expected desirability]. ... The numerical probabilities and desirabilities are meant to be subjective in the sense that they reflect the agent's actual beliefs and preferences, irrespective of factual or moral justifications." Bayesians assume that the Bayesian procedure is the rational course. For such a defense of super power practices as rational, see D. Gautier, "Deterrence, Maximization, and Rationality," Ethics, 84(1984), 474-995. 2. A logical system is paraconsistent if it can serve for the formal representation of an inconsistent theory (i.e., one where both a statement A and its negation not-A hold for some statement A) without trivializing it (i.e., admitting the derivation of an arbitrary statement B, that is, of every statement). In general, the mark of a paraconsistent logic is the unavailability of the spread rule — from A and not-A to derive B — which admits the inference of an arbitrary statement from an explicitly inconsistent pair. 3. See G. Priest and R. Routley, On Paraconsistency (Canberra: Research School of Social Sciences, Australian National University, 1983), pp. 29-31; hereinafter OP. 4. See below. See also R. Routley and R.K. Meyer, "Relevant Logics and their Semantics Remain Viable and Undamaged by Lewis's Equivocation Charge," Topou 2(1983), 205-215; OP. 5. R. Routley, "Maximizing, Satisficing, Satisizing: The Difference in Real and Rational Behaviour under Rival Paradigms," Discussion Papers in Environmental Philosophy, no. 10, (Canberra: Research School of Social Sciences, Australian National University, 1984); hereinafter MSS. 6. H.A. Simon, Administrative Behavior, 3rd ed. (New York: The Free Press, 1976). 7. MSS, p. 16 ff. 8. There are three, not just two, contrasting practices or policies involved in administrative behavior and elsewhere: maximizing (taking the lot, on a simple picture of the matter), satisficing (taking as much as time permits) and satisizing (taking enough). They are discussed and contrasted in the background paper, MSS from which this critical exercise grew.

52 R. Sylvan 9. For example, G. Priest, "Contradiction, Belief and Rationality," Proceedings of the Aristotelian Society, n.s. 86(1986), 99-116. 10. Perhaps M. Slote, "Satisficing Consequentialism," Proceedings of the Aristotelian Society, Supplementary Vol. LXVIII(1984), 139164. 11. A maximally consistent class is a class of statements (or open formulae) that is consistent and also maximal, that is, no statement consistent with it is not a member of it. 12. Subjective or personal probability "represents a relation between a statement and a body of evidence... . It is a quasi-logical relation and the numerical value attached to it [which] is not uniquely determined, represents a degree of belief ... [i.e., coherent belief. For while] ... any degree of belief in any statement is permissible ... there are [coherence] restrictions placed on the distribution of degrees of belief among related statements." See H.E. Kyburg, Jr., and H.E. Smokier, eds., Studies in Subjective Probability (New York: John Wiley, 1964), pp. 5, 7. 13. Although these requirements are often called coherence requirements, they amount to consistency requirements. They represent a direct extension of classical consistency requirements to probability and 'rational belief as arguments justifying them show. A more comprehensive notion of coherence admits inconsistent beliefs and theories; see R. Sylvan, "On Making a Coherence Theory of Truth True," Research Series in Unfashionable Philosophy, no. 3 (Canberra: Research School of Social Sciences, Australian National University, 1988). 14. See further, Routley, ibid, (note 5). 15. P.A. Samuelson, The Foundations of Economics (Cambridge: Harvard University Press, 1985), p. 90. 16. T.R. Cotton, Public Man: A Model of Rational Cooperation, Ph.D. Thesis, Stanford University 1978 (Ann Arbor: University Microfilms, 1979). 17. Cotton, ibid, (note 16). 18. According to C.B. Macpherson, The Political Theory of Possessive Individualism (Oxford: Clarendon Press, 1969), p. 3, the possessive quality of modem individualism consists "in its conception of the individual as essentially the proprietor of his own person or capacities owing nothing to society for them. The individual [is] seen neither as a moral whole, nor as part of a larger social whole, but as an owner of himself. The relation of ownership ... was read back into the nature of the individual." With a few standard liberal assumptions, possessive individualism quickly leads in the direction of 'economic rationalism': "Society becomes a lot of free equal individuals related to each other as proprietors of their own capacities and of what they have acquired by their exercise. Society consists of relations of exchange between proprietors. Political society becomes a calculated device for the protection of this property and for the maintenance of an orderly relation of exchange."

Modern Myths Concerning Rationality

53

19. J. Elster, Sour Grapes (Cambridge: Cambridge University Press, 1983). 20. Roger Lamb suggested some of these etymological points. 21. R. Williams, Keywords: A Vocabulary of Culture and Society (Glasgow: Fontana/Croom Helm, 1976). 22. W. James, A Pluralistic Universe (London: Longmans Green, 1909), p. 319. 23. See, for instance, the examples presented and discussed following the claim that "anyone who has ever mistakenly tried to argue a paranoic individual out of his delusions can vouch for the persistence of a stubbornly dogmatic, but strangely logical vein of thought"; P.C. Wason and P.N. Johnson-Laird, Psychology of Reasoning (London: Batford, 1972), p. 237. What controlled testing has been done appears to show "no significant differences between the normal and schizophrenic groups on the task of logical reasoning"; E.B. Williams, "Deductive Reasoning in Schizophrenia," Journal of Abnormal and Social Psychology, 69(1964), 47-61. 24. E.g., M. Black, "Why Should I Be Rational?," Dialectica, 36(1982), 148-168. 25. The most detailed contemporary account of this sufficiency type I have encountered appears in Gibson's text, where rationality is characterized in terms of having good reasons, and a purely sufficiency account is given of having good reasons. But the ideological 1 significance of this account passes unrecognized (and is now repudiated by the author). 26. M. Oakeshott, "Rational conduct," The Cambridge Journal, 4(1950), p. 26; emphasis added. 27. Black, ibid, (note 24), p. 150. 28. E. Eells, Rational Decision and Causality (Cambridge: Cambridge University Press, 1982), p. 36. 29. Coherence is a matter of fitting together, through requisite relations, into a suitable system. A coherent system may be inconsistent; dialectical set theory is one example, and the totality of truths may be another. Conversely, consistent sets may well be incoherent, because they are all over the place, i.e., because they are not closed under appropriate relations. For more on coherence, see Sylvan, ibid, (note 13). 30. E.g., Elster, ibid, (note 19); D. Follesdal, 'The Status of Rationality Assumptions in Interpretation and Explanation of Action," Dialectica, 36(1982); 301-316; K. Tranoy, "Norms of Inquiry: Rationality, Consistency Requirements and Normative Conflict," in R. Hilpinin, ed., Rationality in Science (Dordrecht: Reidel. 1978). 31. E.g., Hilpinin, Hintikka and von Wright. See Hilpinin, ibid, (note 30), pp. 191-202; G.H. von Wright, "Norms, Truth and Logic," Deontic Logic, Computational Linguistics and Legal Information Systems, ed. A. Martino (Amsterdam: North-Holland, 1982), pp. 320; and for Hintikka, A.N. Prior, Objects of Thought, eds. P. Geach and A. Kenny (Oxford: Clarendon Press, 1971).

54 R. Sylvan 32. Some of these authors present logic only as the logic of consistency or, more explicitly, generate it as reflecting the requirements of consistency, requirements in turn supported as demands of rationality. Thus, for instance: The avoidance of ... inconsistency is listed by Hintikka as one of the marks of 'rationality' in belief. But as Makinson points out [in discussing the paradox of the preface] it may be very rational indeed (possibly a reasonable induction from what we have found with books we have written previously) to launch quite deliberately into the inconsistency just mentioned [effectively that of acknowledging error in a book presented as correct]. ...Hintikka's equation of rationality with consistency is at this point implausible. (Prior, ibid., note 31, pp. 84 f.) 33. According to strict implication, a modal explication of entailment, i.e., of logical implication, one statement A strictly implies another B if and only if A conjoined with the negation of B (i.e., A & not-B) is logically impossible or, equivalently, if it is logically necessary that either not-A or B. 34. An argument like this was recently dusted off and wheeled out again in R.G. Stalnaker, Inquiry (London: Bradford Books, 1984), p. 82 ff. Fuhrmann, who carefully takes to pieces several recent attempts to rule out, or explain away, inconsistent beliefs, calls such an argument, the worst argument. He also neatly dissects arguments from functionalism, from knowledge and using fragmentation 'theory' (such as it is); see A. Fuhrmann, "Relevant Logics, Modal Logics and Theory Change," unpublished Ph.D. thesis, Australian National University, 1988. 35. The assumption enters because the holding of inconsistent beliefs leads — by closure principles and their implicational analogues, notably Bp & Bq -• B(p & q) — to the holding of explicitly contradictory beliefs. Note that belief is only representative of much more widespread features. Though the argument above and in the text is carried through in terms of beliefs, similar considerations apply to other intentional phenomena, e.g., to wishes, desires, preferences, recommendations, etc. 36. See, e.g., R. and V. Routley, "The Role of Inconsistent and Incomplete Theories in the Logic of Belief," Communication and Cognition, 8(1975), 185- 235; hereinafter LB; OP; Priest, ibid (note 9) — and sources cited therein. 37. A dialethic system is a paraconsistent one in which contradictory statements in fact hold, e.g., contradictions are explicitly maintained or are derivable. 38. In a semantical model, a truth predicate, or analogous semantical evaluation function, is defined over a relational structure, or frame, which commonly comprises a system of worlds, their domains of objects and certain of their interrelations. 39. One group of grounds for this claim is outlined in OP, p. 204 ff.,

Modern Myths Concerning Rationality

40.

41.

42.

43.

44.

45.

46.

55

where it is argued that rational criticism and rational assessment are commonly not based upon, but are independent of, assumptions of consistency. A view is effectively criticized, for instance, if it can be shown to lead to something that is not rationally acceptable — be it a contradiction or not. Dialethism is the position that some statements are such that both they and their negations are true. A dialethic logic is a paraconsistent logic that explicitly asserts some such contradictory statements. Of course, here as almost everywhere else, there is a counter-dialectic to be faced. Not uncommonly it starts from the queemess of contradictions' being true, or even being taken to be true (though nothing stops their assumption, even in classical natural deduction procedures). Preferably, enough consistency at the right places: for example, locally, in straightforward situations. For explanation and some detail, see R. Routley, "Relevantism, Material Detachment and the Disjunctive Syllogism Argument," Canadian Journal of Philosophy. 14(1984), 185 ff. See also MSS. Important examples include not only satisizing reproduction behavior, which is widely enough practiced, but also satisizing economic behavior, for example, where firms operate on reasonable returns and do not aim to maximize profits. There are documented examples of such firms (e.g., certain importers of satisfactorily produced third world goods) having become well established, and not having been driven out of business by profitmaximizing competitors. This response, to a point developed by Andre Gallois, also stops Pettit's first argument that sub-maximization is profoundly irrational. Pettit's second argument is based on a common confusion between betterness and preference. While to prefer C to D is, other things being equal, to be disposed to choose C rather than D, an analogous principle does not hold for ranking better in place of preferring — which is what Pettit assumes. And it is accordingly false that sub-maximization "deprives the notion of evaluation [betterness] therefore of its usual content." P. Pettit, "Satisficing Consequentialism," Proceedings of the Aristotelian Society, Supplementary vol. LXVTII(1984), 165-176. See, e.g., H.A. John Green, The Logic of Social Enquiry (London: Routledge and Kegan Paul, 1960), chap. 2. There is some ambivalence in the formulation of these axioms, as between more plausible preference formulations and less plausible quality formulations. For many — though by no means all — economic claims, the more plausible formulation can be adopted. P. Suppes, "Probabilistic Empiricism and Rationality," in Hilpinen, ibid, (note 28), p. 178. Suppes is also no doubt right in his contention that Bayesian theory is an extension of classical logic (pp. 179-180). However, it is not the only extension, but an important one of a competing set.

56 R. Sylvan 47. J.C. Harsanyi, "Advances in Understanding Rational Behaviour," Logic, Foundations of Mathematics, and Computability Theory, eds. R. Butts and J. Hintikka (Dordrecht: Reidel), pp. 315-343. 48. Jeffrey, ibid, (note 1), pp. 16 ff. 49. This point is discussed in MSS, p. 10. 50. G. Riker and P.C. Ordeshook, An Introduction to Positive Political Theory (Englewood Cliffs: Prentice-Hall, 1973). 51. Riker and Ordeshook, ibid., pp. 21-22, 23. 52. Thus, for example, according to Simon, "when a firm has alternatives open to it that are at or above its aspiration level, it will choose the best of those known to be available"; see H.A. Simon, "Theories of Decision-Making in Economic and Behavioral Science," American Economic Review, XLIX(1959), 253- 283. In some more recent work, he may have shifted ground. Simon's satisficing also differs from satisizing in other important respects. For example, according to him, all choices "ranked above" what is chosen are indifferent; this is not so with satisizing. The distinction is not endangered because some of those who declare themselves as marching in Simon's satisfying band sometimes lapse from satisficing to a less demanding and more appealing satisizing; many a clear-enough distinction could be demolished in that way. 53. A. Sen, "Rational Fools: A Critique of the Behavioural Foundations of Economic Theory," in Scientific Models and Man, ed. H. Harris (Oxford: Oxford University Press, 1976), p. 6. 54. Sen, ibid., p. 5. 55. See Sen, ibid., p. 6, where two references to the detailed case are indicated. The case against definitional egoism is elaborated in several philosophical sources, e.g., R. and V. Routley, "Against the Inevitability of Human Chauvinism," Ethics and the Problems of the 21st Century, eds. K.E. Goodpaster and K.M. Sayre (South Bend, Indiana: Notre Dame University Press, 1978). 56. Sen, ibid., p. 7. 57. Eells, ibid, (note 28), p. 1. 58. Ibid., p. 2. 59. Though some more careful estimations may be needed where the threshold is not much above $1,000. 60. Cf. Elster, ibid, (note 19); Pettit, ibid, (note 44), pp. 172-173. 61. Williams, ibid, (note 21), p. 212. 62. Reasoning is, of course, a procedure; but it is not, by any means, the only procedure that may be assessed rational. Indeed, the use of the term "rational" tends to be vacuous or curious when applied to reasoning. 63. Though there is a difference between 'having reason' and 'having a reason', commonly the having of reason would be elaborated, and defended, where it is or can be, by appeal to a reason. However a creature's having reason may be experiential, and not articulatable, or such that something is lost in trying to set it down.

The Rationality R. W. Perrett

As agents we frequently experience a tension between what we want and what we can get. Faced with such dissonance between our desires and the conditions requisite for their fulfillment we have two strategies open to us: we can attempt to modify our desires, or we can attempt to modify the constraints upon their fulfillment. Both strategies are prima facie rational and, at least on some accounts, need not be mutually exclusive. One variant of the first strategy is asceticism: i.e., the doctrine that the agent ought to deny a whole class of his or her desires, or at least seek to minimize them. In this paper I want to discuss two sorts of objection to the rationality of asceticism. Consideration of these arguments is of interest not only for the purpose of determining whether asceticism is indeed irrational, but also for the light thereby shed on the nature of rationality itself.

I In the history of Western thought, ascetic ideals were prominent in Cynicism, Stoicism and medieval Christianity. 2 Asceticism appears to have entered Greek philosophy from the mystery religions which influenced Pythagoreanism in the sixth century B.C. Plato's arguments in the Phaedo, for the repression of bodily desires in order to free the soul in its search for knowledge, continue this tradition. Some of the Cynics rejected worldly desires so as to pursue virtue in independence. The early Stoics denounced emotion as irrational desire and aspired to the ideal of self-sufficiency (autarkeia), a state wherein the sage desires nothing outside the immediate control of his will and reason. Plotinus stressed the ascetic strand in Platonic thought in his condemnation of matter as the source of evil.

58 R.W. Perrett These elements in Greek philosophy subsequently had their influence on the development of Christianity. The ascetic strain was not dominant in early Christianity, but the period from the third to the fifth century saw an extraordinary upsurgence of asceticism. 3 The intellectual foundations for this were provided by theologians like St. Athanasius, St. Gregory of Nyssa, St. Ambrose and even St. Augustine. They argued for the repression of fleshly desires in order to achieve virtue and to further divine contemplation. Figures like St. Simeon Stylites were held up as models of ascetic virtue. Of this anchorite it is said that he bound a rope tightly around himself until it was embedded in his flesh, which then became infested with worms. As the worms dropped from his putrefying sores he replaced them, saying, "Eat what God has given you." Such doctrines and exemplars underpinned the growth of the monastic institutions that were so potent an influence on medieval religious life. Modern Western philosophy, however, has tended to downgrade the value of asceticism, Schopenhauer's advocacy of the importance of annihilating the will being the only significant exception to this trend. In a well known passage in the second Enquiry, Hume sets the tone of dismissal with his sneer at Celibacy, fasting, penance, mortification, self-denial, humility, silence, solitude, and the whole train of monkish virtues; for what reason are they everywhere rejected by men of sense, but because they serve no manner of purpose; neither advance a man's fortune in the world, nor render him a more valuable member of society: neither qualify him for the entertainment of company, nor increase his power of self-enjoyment? We observe, on the contrary, they cross all these desirable ends; stupify the understanding and harden the heart, obscure the fancy and sour the temper. We justly, therefore, transfer them to the opposite column, and place them in the catalogue of vices; nor has any superstition force sufficient among men of the world, to pervert entirely these natural sentiments. A gloomy, hairbrained enthusiast, after his death, may have a place in the calendar; but will scarcely ever be admitted, when alive, into intimacy and society, except by those who are as delirious and dismal as himself.4 And philosophers as different as Bentham and Nietzsche concur: in the Introduction to the Principles of Morals and Legislation

The Rationality of Asceticism

59

Bentham treats the "principle of asceticism" as just the confused inverse of the principle of utility; in On the Genealogy of Morals Nietzsche tells a Just So story about how those incapable of living well disguise their impotence and promote ascetic ideals to excuse their moral sickness and maintain their power over the herd. Other cultural traditions, however, have not conformed to this Western pattern. In India, for example, ascetic strands are still very much part of the religio-philosophical traditions of Hinduism, Jainism and Buddhism. 5 To take but one case: we have, as recently as 1955, the widely admired ritual self-starvation of the Jaina ascetic Santisagara in Maharashtra. 6 And thousands of other Indian religious are daily involved in ascetic practices of varying degrees of austerity. This cultural difference probably requires a complex explanation, but for many modern Westerners such asceticism seems deeply irrational. This widespread cultural confidence is generally unargued for. However, at least two sorts of general objection to the rationality of asceticism can be reconstructed: (1) the claim that the ascetic strategy is 'strongly' irrational in that, for pragmatic or even logical reasons, it is unactualizable for a rational agent; and (2) the claim that the ascetic strategy is 'weakly' irrational in that, although it may have some beneficial consequences, it is not the optimal strategy for maximizing expected value. Before I discuss these arguments, I shall have to say a little more about the nature of asceticism.

II I characterize asceticism as, roughly, the doctrine that one ought (on principle) to deny a whole class of one's desires, or at least seek to minimize them, in the pursuit of some moral or spiritual goal. This characterization may seem rather permissive, but I believe it picks out the relevant phenomena. Certainly I believe it specifies a necessary condition for asceticism. Moreover, it highlights two points about asceticism. First, the disjunctive form implies that asceticism may be more or less severe, depending on how restricted its scope is. Thus a very rigorous asceticism may aim at complete desirelessness; a moderate asceticism merely at the elimination of certain sorts of desires. And this feature is surely a desideratum, for not all varieties of asceticism exhibit-

60 R.W. Perrett the extreme severity of a St. Simeon Stylites. Stoicism, for instance, is usually counted as an ascetic philosophy, but it only requires that the sage desire nothing outside the immediate control of his will and reason. In other words, the Stoic ideal of autarkeia does not entail denial of every desire. And, although Buddhism (particularly in its monastic forms) is markedly ascetic, the "middle way" it preaches does not seek to suppress every desire; instead it teaches that we should satisfy our desires, but desire little. 7 On the other hand, not every form of self-denial counts as asceticism: ascetic practices are directly concerned with bringing about the achievement of moral or spiritual goals. This latter feature ties in with my second point about the nature of asceticism. Ascetic practices are performed for the achievement of some moral or spiritual end, not for their own sakes. Although such practices are claimed by the ascetic to be valuable, they are not held to be valuable in themselves: they have only instrumental value. The justification of ascetic practices is that they lead to intrinsically valuable ends. As to what these ends are, ascetics have offered various characterizations: release from suffering, virtue, closeness to God, happiness, etc. For our purposes we need a general term for the states that are supposed to be the intrinsically valuable ends of ascetic practice: call such states alpha-states. With these considerations in place it is obvious that there are at least three distinct ways to deny the rationality of asceticism. First, we can deny the value of the alpha-states. This, however, is implausible. The sorts of states typically nominated by ascetics as the ends of their practices do indeed seem states that would be intrinsically valuable, if they could be actualized. A second way to deny the rationality of asceticism is to admit the intrinsic value of the alpha-states but deny that ascetic practices produce such states. This denial can take a strong form or a weak form. The strong claim is that ascetic practices cannot bring about alpha-states; the weak claim is that ascetic practices in fact do not bring about alpha-states. The weak claim is an empirical one, and it seems to me that such evidence as is available fails to establish the truth of the thesis. (Of course, some critics will here want to dismiss asceticism as irrational because of the supposed falsity of the ascetic's world view. But any assessment of the rationality of asceticism must begin from the ascetic's own epistemic situation and background beliefs.) The strong claim

The Rationality of Asceticism 61 about the inefficacy of ascetic practices looks more like an a priori one and is much more interesting philosophically. For this reason I shall return to concentrate on that version of the argument. A third way is to admit the intrinsic value of the alpha-states and also that ascetic practices bring about such states, but deny that asceticism maximizes value. This line of objection builds upon the point that ascetic practices are only instrumentally valuable. It might be argued here, for instance, that asceticism is not the optimific strategy for maximizing the production of the intrinsically valuable alpha-states. Alternatively, it might be claimed that alpha-states, though intrinsically valuable, are not maximally valuable and that the actualization of them is incompatible with the actualization of more valuable states. Of course, any criticism of the rationality of asceticism presupposes some account of the nature of rationality, the concept of which has long been the subject of extensive discussion by philosophers and social scientists. 8 Fortunately for our purposes here, we do not require a fully articulated theory of rationality. Suffice it to isolate two conditions for rationality widely assumed in the literature: consistency and maximality. 9 The second line of objection to the rationality of asceticism outlined above claims that asceticism violates the consistency requirement on rational action. The third line of objection outlined above claims that asceticism violates the maximization requirement on rational action. These two lines of objection correspond, respectively, to the arguments in Section I that the ascetic strategy is 'strongly' irrational and that it is 'weakly' irrational. Are these arguments sound?

Ill The first argument against the rationality of asceticism, then, is that it is strongly irrational because it violates the consistency requirement. The ascetic is guilty of some kind of inconsistency. For the rational justification of ascetic practices requires that such practices are productive of alpha-states; but such practices cannot produce alpha-states and hence asceticism is irrational. The precise nature of the objection here depends upon the sense of "cannot" intended. One claim is that asceticism involves a logi-

62 R. W. Perrett cal paradox, i.e., that "cannot" signifies logical modality. A different claim is that asceticism involves a 'pragmatic paradox.' Let us take each claim in turn. One preliminary point first however. My suggestion that inconsistency implies 'strong' irrationality may well be objected to as tendentious, particularly by paraconsistentists. For paraconsistent theory admits non-trivial inconsistent theories. Dialetheism goes further and admits some contradictions as true, or that some things are both true and false. Both reject the claim that consistency is a minimal requirement for rationality. 10 I do not intend here to try to defend the popular support for the consistency condition on rationality. Rather, I shall assume for the moment that this common view is correct and follow the arguments through on that assumption. Perhaps it is worth remarking, though, that advocates of dialetheism do not claim that all contradictions are true; indeed they admit it to be overwhelmingly the case that most contradictions are just false. On both dialetheism and paraconsistent theory, then, consistency may not be a necessary condition of rationality; but it could still be a general desideratum. For instance, if most contradictions are simply false then it is presumably desirable to avoid contradiction if trying to increase the probability of holding only true beliefs. And insofar as it is rational to adopt a strategy designed to increase the probability of holding only true beliefs, then it is also rational to seek to avoid contradictions where possible. Anyway, what about the charge that asceticism is inconsistent in that it involves a logical paradox? Some critics have suggested that Buddhist asceticism is paradoxical in this sense. 11 The ultimate Buddhist goal is the elimination of suffering [duhkha). Suffering (like all phenomena) is caused. The elimination of this cause will bring about the elimination of suffering. Desire is held to be the principal cause of suffering and hence Buddhism is presumably committed to the elimination of desire. But to eliminate desire one must desire to do so, and the cultivation of this latter desire is itself incompatible with the Buddhist prescription to eliminate desire. Hence there is a paradox inherent in the Buddhist ascetic's attempt to desire a way to desirelessness and nirvana. In fact exegetically it is clear that Buddhism is not open to the charge leveled against it here. For the paradox to arise the ascetic has to be committed to a strong ascetic goal of complete desire-

The Rationality of Asceticism

63

lessness. However, Buddhism is not so committed. It is not desire per se that has to be eliminated for Buddhists. (Indeed all agencyinvolves desire for some end or other on most Indian accounts of action, and the Buddhas and bodhisattvas are said to be tirelessly active on behalf of sentient beings.) Rather it is attachment to one's desires that has to be eliminated if one is to be free from suffering, for it is that sort of attached desire or craving (tr^na) that is. the principal cause of suffering. (An attached desire for something typically involves both a first-order desire for that thing and a second-order desire for that first-order desire.) But there is no paradox in desiring freedom from such attached desire, unless it is impossible to do this in a non-attached way (which there is no reason to suppose is so). The logical paradox of desiring desirelessness, then, is not a problem for Buddhism. What is needed for a target here is a very strong form of asceticism that aims at complete desirelessness (perhaps Jainism begins to approach such an extreme). But even if we can find some form of asceticism committed to this extreme goal of complete desirelessness, it is still unclear that a paradox would really be involved. Suppose it is true that in order to eliminate desire the ascetic has to desire to do so. Hence to get rid of one's desires one must begin by adding to them. Now it might seem that, since desire causes suffering, any attempt to eliminate suffering actually adds to one's suffering. But is this so? Surely not. Often the achievement of an end requires the employment of an indirect strategy. In order to obey an injunction it may be necessary to begin by acting in a way that is apparently in opposition to that injunction. Consider an example from the theory of differential games, where there is a rather nasty little pursuit game called "The Homicidal Chaffeur Game." 12 We are to suppose we are in command of an automobile in an infinite unobstructed parking lot and have been instructed to run down a pedestrian. As pursuer we have higher speed than our target, but poorer mobility due to the larger radius of curvature of the car. Now it can be proven that, if the pursuer is initially very close to the pursued, the optimal strategy for the pursuer is to drive away from the target and, then, when some distance away, turn around to initiate a direct pursuit. One step backwards really is required for two steps forward on some occasions. Just as there is nothing paradoxical in obeying the injunction to run the pedestrian down by initially driving away from her, there may well be nothing

64 R. W. Perrett paradoxical in the indirect strategy that requires the ascetic initially to add to his or her desires in order to obey the injunction to eliminate desire.

IV Perhaps, however, the real paradox of asceticism is a pragmatic paradox. Asceticism is not so much guilty of logical inconsistency; rather it is to be criticized as a self-defeating practice. One way to pursue this line of criticism is to invoke the notion of what Jon Elster has called "states that are essentially by-products." 13 These are mental and social states that can be produced by action but cannot be produced as the intended effect of the action. The attempt to bring about such states intelligently or intentionally precludes the state the agent is trying to bring about. Hence such states are attainable only by those who do not seek them. Failure to appreciate the special character of such states leads to two sorts of fallacy. The 'moral fallacy of by-products' occurs when one tries to bring about a state that is essentially a by-product (for some of these states are useful or desirable). The 'intellectual fallacy of by-products' occurs when one tries to explain the occurrence of such a state by reference to an agent's intention to bring it about. Elster offers a number of examples of such states and the selfdefeating nature of attempts at 'willing what cannot be willed.' These include the conscious attempt to 'become natural' that obsessed Stendhal; the unhappy dilemma of the insomniac chasing the elusive state of sleep; and the doomed struggle to will forgetfulness, where desiring the absence of some thought ensures the presence of the object that one wants to be absent. Such activities are irrational in that they are inherently self-defeating. The states that are the desired ends of such activities cannot be brought about by such agency, for these states are essentially byproducts. Now, perhaps the ascetic's alpha-states are similarly states that are essentially by-products. But if this is so, ascetic practices performed to effect such states must be self-defeatingly irrational. Ascetic practices are only rationally justifiable in terms of the valuable alpha-states they are claimed to bring about. If, however, alpha-states are essentially by-products, then they

The Rationality of Asceticism

65

cannot be brought about as the intended effect of action, and hence the practice of asceticism is self-defeating. This is the force of the charge that asceticism is irrational because it involves a pragmatic paradox. To assess this charge we need to distinguish three separate questions. First, are there any states that are essentially by-products? Second, supposing there are, are alpha-states such by-products? Third, supposing alpha-states are essentially by-products, does this make asceticism irrational? Although the first of these questions is in a sense logically prior to the other two, it will be convenient for the moment to set it aside in order to concentrate on the second. Are alpha-states essentially by-products? Certainly some ascetic traditions seem to suggest so. Elster offers the example of the Zen doctrine of "No-Mind."14 Zen masters repeatedly warn against taking this 'goal-less goal' of Zen to be attainable as the result of striving. And the Indian ascetic traditions also offer examples. In Jainism, for instance, there are certain forms of ascetic action traditionally held to be conducive to the attainment of the enlightened status of a GrthaAkara. However, no Jaina should pursue these activities with the aim of becoming a tirthankara; this would be a forbidden desire for worldly gain in return for one's deeds that would preclude the attainment of the desired state. One becomes a tirthankara unawares, as it were. 15 Again, a traditional class of Indian philosophers [qjativadins) specifically denies that one can attain liberation as the effect of one's actions. Although the possibility of liberation is not itself denied, these philosophers (both Buddhist and Hindu) all maintain that there is no causal series of actions and effects that provides a path to freedom [mok?a}.16 Nor is the idea of alpha-states as essential by-products alien to Western thought. Happiness and even pleasure have certainly been claimed by some ascetics as alpha-states, and the suggestion that these states are essentially by-products is very much part of the Utilitarian tradition. Consider, for instance, this passage from John Stuart Mill's Autobiography: I never, indeed, wavered in the conviction that happiness is the test of all rules of conduct, and the end of life. But I now thought that this end was only to be attained by not making it the direct end. Those only are happy (I thought) who have their minds fixed on some object

66 R.W. Perrett other than their own happiness... . Aiming thus at something else, they find happiness by the way. The enjoyments of life... are sufficient to make it a pleasant thing, when they are taken en passant, without being made a principal object. Once make them so, and they are immediately felt to be insufficient. They will not bear a scrutinizing examination. Ask yourself whether you are happy, and you cease to be so. The only chance is to treat, not happiness, but some end external to it, as the purpose of life.17 This passage must have been familiar to Sidgwick, though he does not specifically refer to it in his Methods of Ethics when he develops what he calls "the fundamental paradox of Hedonism": ...that the impulse towards pleasure, if too predominant, defeats its own aim... . It is not only that the exercise of our faculties is insufficiently stimulated by the mere desire of the pleasure attending it, and requires the presence of other more objective, "extra-regarding", impulses, in order to be fully developed: we may go further and say that these other impulses must be temporarily predominant and absorbing, if the exercise and its attendant gratification are to attain their full scope. Many middle-aged Englishmen would maintain the view that business is more agreeable than amusement; but they would hardly find it so if they transacted the business with a perpetual conscious aim at the attendant pleasure. Similarly, the pleasures of thought and study can only be enjoyed in the highest degree by those who have an ardour of curiosity which carries the mind temporarily away from self and its sensations. In all kinds of Art, again, the exercise of the creative faculty is attended by intense and exquisite pleasures: but it would seem that in order to get them, one must forget them: the genuine artist at work seems to have a predominant and temporarily absorbing desire for the realisation of his ideal of beauty. 18 Of course, the mere fact that some traditions have regarded alpha-states as essentially by-products does not in itself establish that they are so, though some may find that various of the examples cited increase the plausibility of the thesis. Let us look more closely, then, at the first question distinguished above: Are there really any states which are essentially by-products? It is important to understand here that the thesis in question is not the

The Rationality of Asceticism

67

(implausible) claim that some products of action logically exclude intentional achievement. Rather the thesis that some states are essentially by-products amounts to the empirical claim that (in our world) some products of action cannot nomically be other than by-products. 19 But, even so understood, the thesis is apparently open to objection. Are there not indirect strategies available whereby the agent can intentionally effect those states that resist a direct approach? Thus Sidgwick, after presenting the "fundamental paradox of egoistic hedonism," goes on to say: ...but though it presents itself as a paradox, there does not seem to be any difficulty in its practical realisation, when once the danger indicated is clearly seen. For it is an experience only too common among men, in whatever pursuit they may be engaged, that they let the original object and goal of their efforts pass out of view, and come to regard the means to this end as ends in themselves: so that they at last even sacrifice the original end to the attainment of what is only secondarily and derivatively desirable. And if it be thus easy and common to forget the end in the means overmuch, there seems no reason why it should be difficult to do it to the extent that Rational Egoism prescribes: and, in fact, it seems to be continually done by ordinary persons in the case of amusements and pastimes of all kinds. 20 Again, consider the argument of the last part of Pascal's Wager.21 The skeptic in this dialogue claims that he cannot accept the wager because he is so constituted that he cannot believe in God. In reply Pascal seems willing to concede that belief in God may not be attainable as a direct product of intentional action; that, as the Humean tradition has it, belief is not a matter of the will. Nevertheless, he wants to insist that the causes of belief may be within the scope of one's agency: by intentionally acting as if one believes in God one may come to believe in God. In this sense belief can be the product of intentional agency. The availability of these sorts of counterexamples requires of us a more careful characterization of the states that are essentially by-products. In the face of the objection that purported by-product states are just states that may be intentionally achievable only by unobvious and indirect means, Elster makes a number of concessions. First, he admits that some by-products can be produced intelligently, but only because they are foreseen

68 R. W. Perrett and not through being intended. This point rests upon the viability of a distinction between doing something knowingly and doing something intentionally, so that an agent can be truthfully said to do something knowingly yet not intentionally. This distinction is familiar from Catholic moral theology, where it is often claimed there is an important moral difference between, for instance, a doctor intentionally killing a healthy human fetus and knowingly causing the fetus to die. The moral claim here is much disputed. Many have seen the distinction as an indefensible piece of moral casuistry epitomized by the figure of the nineteenth-century theologian, St. Alphonsus Liguori, who thought it wrong to commit suicide, but permissible to jump from a high window in the knowledge that it would cause certain death, provided the building was on fire and one's thoughts were firmly fixed on escaping the blaze. 22 However, the controversy over the moral significance of the distinction between acting knowingly and acting intentionally must not be allowed to blur the fact that there is an important distinction for the philosophy of action here. It may be that one cannot escape responsibility for an action by pleading that one did it knowingly but not intentionally; nevertheless the act-types can still be distinct. Second, Elster is willing to allow that some by-products may be produced through being intended, but only if the relevant causal chain is deviant. He offers the example of his eight-yearold son ordering him to laugh and the absurdity of such an instruction in fact causing him to laugh. 23 Again, Zen masters frequently employ teaching techniques that take advantage of the novice's naive attempts to attain by striving that which is essentially a by-product. These two concessions permit a more precise characterization of states that are essentially by-products: "they are states that cannot be brought about intelligently and intentionally." 24 But the objection that they are nevertheless accessible through indirect strategies still remains. To this three further concessions are made by Elster. First, even if a certain state can be achieved by indirect means, it does not follow that it can be achieved at will. Second, some indirect strategies are not rational to employ because they are too costly. Third, many indirect strategies are unviable. Some, for instance, are vitiated by "The Hammock Problem": "Gently rocking myself to sleep in a hammock, I found that just when sleep was coming, my body became so relaxed that

The Rationality of Asceticism

69

I could no longer sustain the rhythmic motion that led me to sleep, and so I woke up and had to start all over again."25 Others require faking; but a person with the skill to fake convincingly may be constitutionally unable to exercise that skill or, alternatively, the would-be faker may overdo the job and produce the genuine thing. The upshot of all this is a more modest, relativized thesis about states that are essentially by-products: "for all practical purposes we may at least talk of states that are essentially byproducts for a given person."26 And this at least seems plausible. Some indirect techniques that are nomically possible for one agent may not be possible for another. (As John Bishop puts it: "Not all of us, for example, can obey the kindly command, 'Just let yourself relax and it won't hurt a bit!'")27 What does the argument establish so far? Well, it seems at least that (1) there are perhaps some states that are essentially byproducts (in the weaker, relativized sense); (2) some alpha-states may indeed be essential by-products in this sense; and (3) such alpha-states cannot rationally be intended as the product of ascetic practices by agents for whom those states are essential byproducts. Does it follow from this, however, that the practice of such asceticism is irrational for those agents? Surely not. Even if the alpha-states that are the products of the ascetic way of life cannot be rationally intended, it may still be rational to engage in that way of life only because one knows it will produce the valuable alpha-states that are the desired essential by-products. (I take this point to be implicit in the resolution of the 'paradox of hedonism' quoted from Sidgwick earlier. The hedonist in pursuit of the pleasures of business or artistic creation would act self-defeatingly if he or she were to make that pleasure the direct object of his or her actions. However, such a person's motivation for engaging in those activities is precisely that pleasure, and there is no irrationality in pursuing business or art only in order that these pleasures may be enjoyed as a by-product of commercial or artistic activities.) But if this is right, then asceticism does not involve a pragmatic paradox. And, since we have found no reason to suppose that it involves a logical paradox either, the argument that asceticism is strongly irrational fails.

70 R. W. Perrett V

,

What, however, of the charge that asceticism is weakly irrational in that it fails to maximize value? There are two distinct questions at issue here: (1) are ascetic practices suboptimal; and (2) is non-maximizing irrational? It is very difficult to see how to determine a general answer to the first question. Of course, the ascetic is often thought to renounce something valuable, but he or she will argue that the alpha-states gained are more valuable still. The issue is further complicated by unclarity as to what is meant by reference to the expected value of an alternative. Is this the objective expected value of an outcome or the subjective expected value? And if it is the former, then how reliable a guide to this is the latter? 28 Usually, however, the interest in providing an answer to the first question, about the optimality of ascetic practices, is largely a function of the assumption that the second question, about the irrationality of non-maximizing, has to be answered affirmatively. This latter assumption is highly disputable. For a start, if maximality is a condition of rationality, then all rational choice requires a complete ordering of options. Does choice in the absence of such completeness really imply, however, a failure of rationality? It is hard to see that it does. There are plausible models of rationality that do not require completeness of ordering. Amartya Sen, for example, allows that rational choice may be based on an incomplete ordering, provided that a not inferior alternative is chosen. 2 9 On this model the real irrationality of Buridan's ass was not its inability to rank the two haystacks, but its refusal to choose either of the two haystacks without being certain that one was better than, or at least as good as, the other. Rational choice based on incomplete ordering requires only that a not inferior alternative be chosen: Buridan's ass may rationally opt for either haystack, but not neither, which is clearly an inferior alternative. An alternative satisficing (or satisizing) model of rationality goes further still. Once again, rational choice can be based on incomplete orderings, provided that an adequate alternative is chosen. This understanding of rationality as an adequacy notion, however, also undercuts the whole maximization assumption that is a presupposition of the argument for the weak irrationality of asceticism. For on this model of rational choice,

The Rationality of Asceticism

71

picking a suboptimal alternative is not irrational, provided that the alternative chosen satisfies an adequacy threshold. This model owes much to the work of the economist Herbert Simon. 30 The philosophical development of the idea, however, is primarily due to Michael Slote and Richard Sylvan (ne Routley). 31 In his original work on administrative behavior, Simon argued that, in reality, administrators frequently do not seek to maximize (i.e., select the best from alternatives available to them), but instead satisfice (i.e., select a course of action that is satisfactory or 'good enough"). Moreover, any model of rational choice that incorporates the actual properties of human beings must do justice to this fact. Exegetically it is unclear whether Simon understood satisficing as a genuine alternative to the classical maximizing model of full rationality, or merely as an instance of 'limited' or 'bounded' rationality that reduces to (constrained) maximization. Certainly some authors have claimed that satisficing (understood as limited rationality) is reducible to a species of maximizing or total rationality. 32 The idea here is that the use of rough principles or rules of thumb can be optimal because it permits economizing on the information costs that the search for an abstractly optimal choice would involve. The economist Sidney Winter, however, has argued that any such attempt to reduce satisficing to maximizing fails because it leads to an infinite regress. For the "choice of a profit-maximizing information structure itself requires information, and it is not apparent how the aspiring profit maximizer acquires this information, or what guarantees that he does not pay an excessive price for it."33 But then it seems that, not only is there a point in every decision process where calculation stops and an unsupported choice has to be made, but that this point may as well be as close to the action as possible. Satisficing, then, is not reducible to (constrained) maximization. (Hence, as I construe it, satisficing is identical with what Sylvan prefers to call "satisizing.") Furthermore, this irreducible satisficing model of rationality rejects the maximization assumption behind the argument for the weak irrationality of asceticism. Whereas maximization entails that only the best is good enough, satisficing allows for the rational choice of suboptimal alternatives, provided that an adequacy threshold is satisfied. And this is surely right: choosing less than the best available is not obviously irrational. Take Slote's example of our

72 R.W. Perrett reaction to fairy tales in which the hero or heroine is offered a single wish and asks for something good, but less than what he or she could have asked for.34 We are not usually convinced that such a wish is irrational. Perhaps we might have wished otherwise for ourselves in such a situation, but this need not imply that we believe the other's moderation to amount to a failure of rationality. Indeed we might go further: not only is there nothing irrational in choosing a suboptimal (adequate) alternative, but neither is there any moral obligation to do otherwise. Thus (following Slote) it may be possible to extend the point about the rationality of satisficing and have a form of moral satisficing which rejects the usual consequentialist demand for maximization in order to make room for supererogation within a consequentialist framework. The relevance of all this to the argument for the weak irrationality of asceticism is clear enough. On a satisficing model of rationality, even if asceticism does not maximize value, it can still be rational provided that it satisfies an adequacy threshold. Thus the general argument for the weak irrationality of asceticism fails. Of course, it can now be disputed whether asceticism meets the adequacy condition. But it is hard to find a plausible general argument why it does not. After all, the attractiveness of the argument for the weak irrationality of asceticism was in part due to the fact that it conceded both that asceticism does bring about the production of alpha-states and that such states are intrinsically valuable. The claim that asceticism nevertheless does not maximize value has some plausibility; the claim that the intrinsically valuable alpha-states are insufficiently valuable even to be adequate options is much more implausible. This point is quite compatible with our intuitions that some very rigorous varieties of asceticism fail to meet the adequacy requirement on rational choice. For what I am concerned with here is the possibility of a general argument for the irrationality of asceticism, rather than an argument for the irrationality of some specific form of ascetic practice. Since it is surely implausible that even a moderate asceticism fails to satisfy the adequacy condition, there can be no plausible general argument for the irrationality of asceticism from the claim that it fails to maximize value.

The Rationality of Asceticism

73

VI The lessons of this paper are of two kinds: one about the rationality of asceticism and one about the nature of rationality. In the first case I have argued that asceticism is neither strongly nor weakly irrational. Moreover I know of no other satisfactory general argument for the irrationality of asceticism. This conclusion does not preclude the exclusion of particular forms of ascetic practice as irrational. But this exclusion will require careful case by case evaluation of individual options. There are no grounds for a widespread contemporary Western cultural confidence that asceticism is not a rational option. The arguments I have presented for this conclusion have also highlighted certain features of the nature of rationality itself. In particular: (1) that some states which are the products of action cannot themselves be rationally intended and hence there are limits on rational agency; and (2) that maximization is not a minimal condition of rationality, for there is a more plausible satisficing model which allows for the rationality of suboptimal choices provided that they satisfy an adequacy condition.

Notes 1. My thanks to Richard Sylvan for access to some of his unpublished work on rationality and for his critical comments. Thanks also to Graham Oddie and Valerie Perrett for helpful suggestions. 2. Useful historical surveys of asceticism can be found in T.C. Hall, "Asceticism," in J. Hastings, ed., Encyclopedia of Religion and Ethics (Edinburgh: T. & T. Clark, 1909); O. Hardman, The Ideals of Asceticism: An Essay in the Comparative Study of Religion (London: Society for Promoting Christian Knowledge, 1924), chap. 2; C. Wellman, "Asceticism," in Paul Edwards, ed.. Encyclopedia of Philosophy (New York: Macmillan, 1967); and W.O. Kaelber, "Asceticism," in Mircea Eliade et al., eds., Encyclopedia of Religion (New York: Macmillan, 1986). 3. On this phenomenon see W.E. Hartpole Lecky, History of European Morals from Augustus to Charlemagne, vol. 2 (London: Longmans, Green, 1869), pp. 101-146, 154-194. 4. Hume's Ethical Writings, ed. Alasdair Macintyre (New York: Collier, 1965), p. 111. Compare the retort in Edward Gibbon, The Decline and Fall of the Roman Empire, vol. 1 (New York: Modern Library, n.d.), p. 413: "... it was not in this world that the primitive Christians were desirous of making themselves either agreeable or useful." (I owe this latter reference to Charles Pigden.)

74 R.W. Perrett 5. J.M. Masson, "The Psychology of the Ascetic," Journal of Asian Studies, 35(1976), 611-625, provides a useful introduction to the Indian ascetic tradition. Characteristically, however, the author is hostile to the phenomenon and regards the ascetic as fundamentally a psychopath. 6. See P.S. Jaini, The Jaina Path of Purification (Berkeley: University of California Press, 1979), p. 1. 7. Tantra is an interesting case here, for it encourages the non-suppression of desires and yet is regarded by the Indian tradition as a kind of ascetic discipline. The resolution of this tension is clear from the remarks of a contemporary Hindu practitioner: The true tantrik is always in a state of non-suppression and enjoyment. The purpose of every moment of life is to experience ananda. Ananda is active enjoyment of everything that comes your way. ... Ideally, a tantrik is in such a state of attunement with his environment, with what is possible, that his desire awakens just at the moment when the environment is willing to grant it. ... A tantrik has only those desires which the environment is ready, willing and in a position to satisfy. This is not because he denies any of his wishes or rationalizes them later, but because he has developed his capacity for attention and is intensely aware of where he is and what he is doing at every single moment of time. For example, in my own case ... I can tell you that I have the appropriate desire for whatever food is put before me. If it is fish, I discover with pleasure that fish was exactly what I wanted; if it is meat, I cannot imagine it could have been anything else. So every day I eat exactly the meal I want to. (S. Kakar, Shamans, Mystics and Doctors (New York: Alfred A. Knopf, ,1982), pp. 166-167.) On my characterization of asceticism tantra is ascetic in that it requires practitioners to refrain from conserving any desires which circumstances do not allow them to satisfy but to conserve those desires which are satisfiable. The tantric task is to minimize one's desires to that set which are satisfiable. Tantric practices of awareness-development are designed to act directly on one's desires and preferences to effect this result. 8. For a useful review of some of the recent literature see J. Elster, "Rationality," in Guttorm Floistad, ed., Contemporary Philosophy: A New Survey, vol. 2 (The Hague: Martinus Nijhoff, 1982). 9. For an attack on both of these conditions as modern myths about rationality, see Sylvan, this volume. 10. Sylvan, ibid., rejects the consistency requirement on these sorts of grounds. See also G. Priest, "Contradiction, Belief and Rationality," Proceedings of the Aristotelian Society, 86(1985-86), 99-116. On paraconsistency, more generally, see G. Priest and R.

The Rationality of Asceticism 75

11.

12. 13. 14. 15.

Routley, On Paraconsistency (Canberra: Philosophy Department, Research School of Social Sciences, Australian National University, 1983). On dialetheism, see Priest, In Contradiction: A Study of the Transconsistent (Dordrecht: Martinus Nijhoff, 1987). On this question, see the exchange represented by the following articles in Philosophy East and West J. Visvader, "The Use of Paradox in Uroboric Philosophies," 28(1978), 455-467; A.L. Herman, "A Solution to the Paradox of Desire in Buddhism," 29(1979). 91-94; W. Alt, "There Is No Paradox of Desire in Buddhism," 30(1980), 521-528; Herman, "Ah, But There Is a Paradox of Desire in Buddhism: A Reply to Wayne Alt," 30(1980), 529-532; Visvader, "Reply to Wayne Alt's There Is No Paradox of Desire in Buddhism'," 30(1980), 533-534. Cf. Elster, Ulysses and the Sirens: Studies in Rationality and Irrationality, rev. ed. (Cambridge: Cambridge University Press, 1984), p. 13. Idem, Sour Grapes: Studies in the Subversion of Rationality (Cambridge: Cambridge University Press, 1983), chap. 2. Ibid., pp. 48-49. On this Zen notion, see D.T. Suzuki, The Zen Doctrine of No-Mind (London: Rider, 1969). Jaini, ibid, (note 6), p. 266.

16. See K.H. Potter, Presuppositions of India's Philosophies (Englewood Cliffs: Prentice-Hall. 1963), chap. 11. 17. Essential Works of John Stuart Mill, ed. Max Lerner (New York: Bantam Books, 1961), p. 88. 18. H. Sidgwick, The Methods of Ethics, 7th ed. (London: Macmillan, 1907), pp. 48^9. 19. For this reading of Elster, see J. Bishop, "Review of Jon Elster's Sour Grapes," Australasian Journal of Philosophy, 63(1985), 246247. 20. Sidgwick, ibid, (note 18), pp. 136-137. 21. Pascal's Pensees, trans. M. Turnell (New York: Harper and Row, 1962), pp. 135-136. 22. Cf. J. Rachels, The End of Life (Oxford: Oxford University Press, 1986), p. 105. 23. Elster, ibid, (note 13), pp. 53-54. 24. Ibid., p. 56. 25. Ibid., p. 57. 26. Ibid., p. 58. 27. Bishop, ibid, (note 19), p. 247. 28. On these matters see F. Jackson, "A Probabilistic Approach to Moral Responsibility," in Ruth Barcan Marcus et al., eds., Logic, Methodology and Philosophy of Science VII: Proceedings of the Seventh International Congress of Logic, Methodology, and Philosophy of Science, Salzburg, 1983 (Amsterdam: NorthHolland, 1986); G. Oddie, "Culpable Error in Action and Thought," unpublished. 29. A. Sen and B. Williams, eds., Beyond Utilitarianism (Cambridge: Cambridge University Press, 1982), p. 17.

76 R.W. Perrett 30. H.A. Simon, Models of Man (New York: John Wiley, 1957), chap. 14; Administrative Behavior, 3rd. ed. (New York: The Free Press, 1976). 31. M. Slote, Common-Sense Morality and Consequentialism (London: Routledge and Kegan Paul, 1985), chap. 3; R. Routley, "Maximizing, Satisficing, Satisizing: The Difference between Real and Rational Behaviour under Rival Paradigms," Discussion Papers in Environmental Philosophy No. 10 (Canberra: Department of Philosophy, Research School of Social Sciences, Australian National University, 1984); Sylvan, ibid, (note 9). 32. E.g., W. Riker and P.C. Ordeshook, An Introduction to Positive Political Theory (Englewood Cliffs: Prentice-Hall, 1973), chap. 2. 33. S.G. Winter, "Economic 'Natural Selection' and the Theory of the Firm," Yale Economic Essays, 4(1964), p. 262. See also Winter, "Optimization and Evolution in the Theory of the Firm," in R.H. Day and T. Groves, eds., Adaptive Economic Models (New York: Academic Press, 1975); Elster, ibid, (note 12), p. 59. 34. Slote, ibid, (note 31), p. 43.

On The Rationality of Context, or, HowRational Attention to Context Leads to Total, Irrational Relativity B.-A. Scharfstein

Ordinarily, we accept the idea of context without any question; but this is only because we do not pursue it very far. We do not see that dependence on context is a limited kind of relativism and that relativism, looked at philosophically, is hard to limit. We seem to feel that the idea of context is clear, stable and sufficient and does not need either qualifications or a context of its own. Because it is most often used in literary study, history and social science, it is not analyzed as closely as the standard philosophical questions, such as that of relativism itself. The price paid is intellectual slackness, because to neglect to see where the idea of context leads is as intellectually unjustified as to neglect to use it. While it appears to me that both forms of neglect have some justification, it will be more useful here to pursue the idea of context consistently and see to what it leads. The very attempt to be consistent makes the idea of context hard to manage, because, if we do not limit it either intuitively or arbitrarily, it is unrestrained by any natural limit of its own. This lack of natural limit is the abstract explanation of many of the problems we run into. What happens when we try to apply the idea with scrupulous consistency? We discover that context extends so far, from the most encompassing framework of experience to its most minute particles (or waves), that we can never finish with it. We discover that it always remains possible and in some perspective true that our understanding has been limited because we have not paid enough attention to context. It should not be supposed that we get rid of this endless demand if we confine ourselves to a homogeneous cultural environment

78 B. -A. Scharfstein or, within it, to standard problems. Serious investigation has a way of making itself interminable: there are always more details to be gathered, more observations to be made, more side-issues to be investigated. When the painter Robert Rauschenberg wanted to force spectators to look really closely, he painted a surface all black. If anyone was curious to see what was there, he had to concentrate on the texture of the torn, glued-together pieces of newspaper. The near-uniformity of the painting stimulated the spectator to pay attention and see that, looked at closely, it was not uniform at all. In the same way, nothing human we think about is uniform through-and-through. There is always something left over to see and understand. Every culture, philosophy, philosophical problem, book, article, sentence, phrase and word, can be declared insufficiently intelligible if we demand that it be grasped in all its distinctiveness, if, that is, we insist that the context that distinguishes it cannot be detached from it, is in effect immanent to it, and separates It from anything that might be supposed the same. By attaching things more firmly to their contexts, which all vary at least in detail, we increase our recognition of their individuality and make them more nearly unique. At the extreme limit of distinction by means of context, they should become really unique, which is to say absolutely incomparable, and, in consequence, subject to a paradox that was developed by skeptics: To recognize something different or new we have nothing to go by but our experience with what we have known up till now; but what we have known up till now resembles what we now want to know only in those aspects that tell us nothing of its uniqueness; but we cannot recognize the uniqueness itself because we have never met it before; and if we recognize only that we are coming into contact with something new or different, we do not have the experience or ideas with which to grasp it. A traditional way out of the paradox depends on atomism, according to which unique things are made so by the rearrangement of their identical particles; but this way seems not to be available to the extreme partisan of context, who must, I imagine, be a holist of some sort. The person who insists that context creates uniqueness drives away the threat of sameness by pointing out small differences of context that interrelate and constitute a whole that has no double anywhere. Viewing, thinking, proving can all be declared in some degree misleading

On the Rationality of Context 79 if not viewed in a special enough light, through the prism of relationships and meanings that make everything human unique. When a context that is relatively uniform provokes attention to relatively small differences, and these to progressively smaller ones, which are the differences of the differences, the possibility of deciding that things are the same becomes progressively weaker. Of course, the key is the decision to take small differences into account. The process reminds me of the measuring of a coastline and the associated paradox (which now has a mathematical solution). The paradox is this: Measuring is by means of conventions that experience has dictated; but if, in defiance of the conventions, one tries, in thought or practice, to measure a coastline very exactly, and therefore measures not only the smaller protrusions and indentions, but their protrusions and indentions, and the protrusions and indentions of the protrusions and indentions, the coastline grows in length until it threatens to become immeasurably long. The further we regress into conditions and conditions of conditions, the more pedantic, more slow-moving we become, and the greater the threat that all intellectual motion will cease. Suppose we spell out the difficulty in more concrete detail. Suppose that we begin with no more than an interest in intellectual life and the commitment to investigate its context systematically and thoroughly. What would this require? How many disciplines would we have to turn to, and in what detail would we have to apply them? In giving a possible answer, I will avoid extravagance and not interpret the word "systematic" as demanding recourse to every discipline that might be relevant to the context of intellectual life. There are a number of disciplines, however, that research has long associated with this context. Among them are history, geography, economics and psychology. To this reasonable minimum, we may feel obliged to add philology, paleography and other disciplines invented to help in the understanding of texts. We decide, then, to concentrate on a manageable number of disciplines; but of those I have named, only philology and paleography can be said to have a single, unambiguous outlook or method to apply. The rest do not, and are complicated by their division into subspecialties and rival outlooks or methods. For ex-

80 B. -A. Scharfstein ample, the psychology we choose may be behavioristic, psychoanalytic, cognitive, affective, or holistic, or, as is often the case, eclectic, meaning pieced together inconsistently but conveniently out of fragments taken from everywhere. Or, one might turn, as seems reasonable in comparative studies, to ethnopsychology or ethnopsychiatry, which have gathered a body of special evidence, but which may be practiced in the spirit of any of the earlier named outlooks or methods — behavioristic, psychoanalytic, and so on.1 Suppose, however, that we simplify things as much as possible, and suppose, too, that we overcome the difficulty in choosing outlooks and methods. Even so, we come to recognize that the idea of context must be broken up or qualified in order to bear systematic examination. One way of breaking it up is to establish a sequence of levels. Without this, how can we hope for a complete and systematic view? The levels can, of course, be few or many and qualitative or quantitative, that is, arranged on a numerical scale. A simple, natural sequence of levels might be made of these five: the microcontext, the correlative context, the macrocontext, the metacontext and the universal or meta-metacontext. Any particular subject-matter would suggest its own variations on this set. I begin with the correlative context because it is the one to which we turn immediately and instinctively. In the case of a philosophical text, it would include the book in which it occurred, the text or texts on which it drew or to which it was responding, the other writings of its author, and so on. Philosophers, who rarely understand crucial texts as well as they would like, spend a good deal of effort in such contextualizing. They usually discover, I am sure, that the effort helps, but rarely to their complete satisfaction. The exact meanings prove to be extraordinarily elusive. It seems to me no exaggeration to say that no one, philosopher or layman, ever understands any other person, philosopher or layman, to the other person's complete and lasting satisfaction; and if the test of the accuracy of our interpretation of a text were to be the agreement of its author, I assume, though without proof, that we should usually fail, whether or not we had taken pains to establish the context. And if the test were to be the agreement of others who had taken pains to establish the context, then we should also usually fail, I assume.

On the Rationality of Context 81 By 'microcontext' I refer to a more minute scrutiny, sentence by sentence and concept by concept; but also to the personal context, which gives the words their personal resonance; and to the style, which can be analyzed by many techniques; and to things said as contrasted with those that are merely implied. It is evident that, as I use it, the 'microscopic' quality of the microcontext is vague and could be made more exact by further distinctions. How fine should the distinctions be, and how far down the scale of size should we go? Barring the discovery of more illuminating relationships between them and texts, the subatomic, atomic, and cellular levels of human existence should be left out, even though the peculiarities of the individual cells of the nervous system and their interconnections and electrochemical environments are presumably reflected in what and how one thinks. The genetic level is intriguing, but still too distant to help us with the understanding of texts. The sensitivity of the eye, the ear and some parts, perhaps, of the central nervous system, is so great that they can be affected by the chance fluctuations with which quantum mechanics deals. But, although this may be the ultimate level for us in physics, it would be comical to assume that it could give us real help at present in understanding a text. How small can something get and still stay relevant? Scrutiny minute enough to earn the name 'microtextual' could take the form of a detailed commentary with philological and other excursuses. If taken from a well-worked field, such as Greek philosophy, the text might be immersed in a sea of commentary, something on the order of a brief phrase to a page or more of commentary. As an example of an inquiry in principle microtextual one might take the essay, "How to Render Zweckmassigkeit [purposiveness] in Kant's Third Critique,"2 or the question how to render "purposiveness" in the Critique's first preface, on which a book-length commentary has been published. 3 We know that every thinker has not only his own intellectual preoccupations, but his own vocabulary, syntax, word relationships or semantic fields, and verbal eccentricities. Since in reading prose we are mainly attentive to the information conveyed, a microtextual investigation is most helpful in instances of ambiguity, ambivalence, or apparent or real inconsistency. However, interpreters have different sensitivities. Bergson went so far as to say that one could not understand a philosopher — he instanced Descartes — unless one read him aloud with the proper emphasis

82 B. -A. Scharfstein and intonation, and Wittgenstein made similar remarks about the precise emphasis or melody of what people say. Locke's style has been reasonably linked to his philosophical attitudes and desire to be and appear plausible.4 It is true that certain aspects of personality are much more evident in speech than in writing, but there are many things in writing apart from the explicit message that exert influence and repay analysis. There are metaphor, simile, syntax, relation between style of reasoning and style of expression, directness or evasiveness, the disregard that can be interpreted as conscious or unconscious silence about something of concern, the fit or lack of fit between what is said and how it is said — for example, purportedly rational refutation with the help of strongly emotional words; kindness preached in a harsh, militant vocabulary; calm recommended in an anxious way; and freedom or spontaneity recommended in a naggingly obsessive way. In brief, a fully serious attempt to understand a text would have to descend to its microcontextual qualities. While the microcontext is likely to be personal or subpersonal, the macrocontext is impersonal and deals with such larger matters as the disputes between schools to which the text is relevant, the cultural conditions it reflects, and so on. Such would be a study of Greek philosophy in Athens in the context of ancient Athenian life, or of Chinese philosophy in the context of the Chou dynasty, or of Renaissance thought in its Florentine context. The inclusiveness of macrocontexts creates a constantly overlapping effect. To study the ideals of Greek, Indian, Chinese or European education one needs to make reference to the appropriate philosophies; but one can also study philosophy in any culture in the context of its education, and in this context make clear just what was taught, by whom, in what way and to what purpose or effect.5 There is a similar overlapping between the history of philosophy and that of political thought, law and literature. A study of modes of reasoning in India concentrates on philosophy but also recalls dialectical reasoning in literature and law, which formed part of the context of intellectual life in that civilization.6 A somewhat similar attempt, confined to early Greek culture, finds it necessary to begin with pre-philosophical tendencies, with metaphor in cosmological theory, with argument by means of comparison or analogy, and so on; and a continuation of this

On the Rationality of Context 83 study deals with the critique of magic, with dialectic and demonstrations and with empirical research and science.7 The overlapping may be of another kind. For example, the study may be of the interplay of the rational and the irrational in Greek thought, or, more generally, of mental life in ancient Greece from the standpoint of psychological thought then and now.8 Or the macrocontext can be geographical, in the sense, say, of the book called Oxford in the Age of John Locke, which studies the 'inner' and 'outer' worlds of Oxford and the relations of Locke himself to the university. 9 Another kind of geographical microcontext is exemplified by a book on the holy city of Banaras, which gives Indian theology a sense of place, a physical and spiritual habitat. 10 So much for macrocontexts and their possible varieties and overlappings. The metacontext, as implied by the name, deals with the text from above. Looking from above, one asks why the kind of questions the text deals with are raised at all, or why such arguments are used, and so on. If reasons for using particular kinds of arguments fall into the scope of the metacontext then the studies of reasoning I have cited, one for India and two for Greece, belong here rather than in the category of macrocontext. The study of the relation of Greek or Chinese reasoning to the nature of Greek or Chinese as compared with other languages might also be put in either category, but if the comparisons are far-reaching, they seem preferable here, in the category of metacontext. 11 Studies that attempt to clarify whole cultures by philological means probably belong here as well, as do studies of the whole development of the language in which the text was written. 12 The category of metacontext also fits the study called Dreams, Illusion, and Other Realities, which deals with the Indian sense of reality and illusion and compares them with Chinese and European equivalents. 13 Here, too, belong the other wide-ranging comparative studies of all kinds. The universal or meta-metacontext is established by joining all the other, partial contexts and setting them in their relationships to one another so as to make visible the full intellectual universe of the text. The universal contexts that we do in fact construct are no more than hopes or sketches. For the breadth and audacity of their work, sociologists such as Emile Durkheim and Max Weber fit the universal category. For his universalizing ambition, so does the anthropologist Alfred Kroeber.14 Another

84 B. -A. Scharfstein candidate is Oswald Spengler, and a more recent one, Benjamin Nelson.15 On a more abstract plane, there are, to be sure, the encompassing metaphysical systems, some of which, notably Hegel's, construct a much fuller universal context than others. Though beginning modestly, the demand for context has grown beyond all measure. I say this because of the number of disciplines and levels that can reasonably be taken as relevant to context. Logically, however, this is only the beginning. A moment's thought shows that each discipline can also provide the context for the study of itself on a higher level as a metadiscipline. For instance, the study of history inspires the study of the history of history; that of geography, the study of the geographic differences in the study of geography; that of economics, the study of the economic characteristics of the profession of economics; and that of psychology, the study of the psychology of psychology or of psychologists. That is not the end. Each discipline can be regarded as providing the context for each of the others, and, more comprehensively, for all of the others together. Each has its history, geography, economics and psychology; and because they are all, in common, learned professions, they have a common history, economics, and so on. Yet the end is still not in sight. Each level of context can reasonably be regarded in the light of each of the others. One can discover or formulate a macrocontext for the use of microcontexts, and, likewise, a microcontext to explain the development and nature of macrocontexts, to which I add, not for the first time, "and so on." To recapitulate, we began by asking what it would mean to investigate the context of intellectual life thoroughly. Then, to be practical, we narrowed the context of intellectual life to that of some philosophical text. As a result, we obligated ourselves to reconstruct an indefinitely great part of intellectual and social life, as seen from a great number of angles and angles of angles. The natural comment is that such a procedure is absurdly unwieldy. It does not take much philosophy to see that this progressive increase in demands for context leads to the conclusion that everything in the universe of the text we have been thinking of provides a context for everything else in it. A step further, and we reach the conclusion that everything in the universe of the text's universe, meaning the universe without any qualifications, provides a context for everything else. Still another step, and we

On the Rationality of Context 85 learn, in the light of the need for context, that we cannot know anything unless we know everything. But since we cannot know everything, we cannot know anything — unless, that is, we accept the possibility that knowing any one thing fully is equivalent to knowing everything. In this last possibility we recognize the ideals of such philosophers as Leibniz and Hegel; but I am not sure that we have helped ourselves, because there is no reason to suppose that knowing any one thing quite fully is easier than knowing everything — unless one is God, the Monad of Monads, or the Hegelian Spirit. Certainly, if everything is the context for everything else, everything is in a sense everything else, constitutes everything else, or pervades everything else. If this idea seems vague, an example or two will clarify it. A house that has been moved from one place to another remains the same house, we usually say. But if a person argues that to him the house was what it was in virtue of the scenery around it, its immediate geographical context, we should understand him, because architects design houses for certain localities and certain localities fit the character of certain houses better than others. The person might therefore argue that in his mind the house was inseparable from the scenery and the scenery from the house. House and scenery would in that case affect one another, qualify one another and belong to one another, in direct experience, in memory and in conception. They could then be reasonably considered only different aspects of the same experiential or conceptual unit and their relations with 'one another' would be, in philosophers' terminology, internal or mutually pervasive. If this argument holds for simply material things such as houses, it holds all the more for human beings. Parents are the genetic context of their children's lives and are in this sense present in them and inseparable from them. And if this holds biologically, it holds psychologically, under normal circumstances, and not only of parents but of everyone who exerts a psychological influence: We are ourselves but are also made from and inseparable from the ideas and emotions of others. If all this is true, we arrive at the old view, now suggested in physics by "bootstrap" theory, which says, as Anaxagoras did long ago, "In everything there is a portion of everything."16 There is a further interesting consequence of the demand to see everything in context. It is this: If we assume that nothing can be

86 B. -A. Scharfstein understood outside of its particular context, the same must be true of the doctrine itself that nothing can be understood outside of its context; and the context of the doctrine must have its context or contexts; and so on. So, too, our personal act of setting something, anything, including what is now being said about context, must be set into its context, or, to be accurate, into as many contexts as are relevant to such an act, and, in addition, to the contexts of its no doubt many contexts. We have run into an old aporia, an encounter that teaches us that the process of reasoning we have begun is not subject to any consistent intellectual limit, and that we have demanded of ourselves not only more than is possible, but indefinitely more — if I were not now afraid of the word, I would say infinitely more. Granted this situation, it is a comfort to realize that others were here before us, and that not everyone was afraid. The fearless ones embraced the idea that everything is the context of everything else, relative to everything else, contained within everything else, immanent to everything else. Some of them embraced the idea with a seductively poetic exuberance. Talking, as they did, of the noumenon, the hidden reality, and the phenomenon, its visible effects, they said that the noumenon and phenomenon are as separate and yet as identical as the ocean and its waves, so that in principle one could know all the ocean from a single wave — rather as the astronomer learns the constitution of a whole distant star from the spectrum in his telescope. Using the metaphor of ocean and wave, they said that though the ocean is concentrated in the wave, the ocean does not shrink, and though the wave includes all the ocean, the wave does not expand. And "though the ocean simultaneously extends itself to all waves, it does not by this fact diversify itself; and though all waves simultaneously include the great ocean, they are not one ... .While one wave includes the great ocean, all other waves also include the ocean in its entirety. There is no obstruction whatsoever between them." 17 In such a philosophy, the universal and particular and the particular and every other particular were said to interpenetrate, yet without any loss of the difference between them. "Every actual entity," it was said, "is present in every other actual entity," and "Every spatio-temporal standpoint mirrors the world," for an atom "stretches in all ... directions, yet it does not move from its

On the Rationality of Context 87 local position. So it is far and also near, stretching and also remaining ... ,"18 In the last two paragraphs, I have run passages from Alfred North Whitehead together with others taken from a Chinese source. The Chinese source is the school of Buddhism called Flower Garland or Hua-yen, or, in Japan, Kegon. If we insist on literal precision and faithfulness to the immediate context of philosophizing, everything in Whitehead and these Buddhists should be construed to be different, and even the translated pair of words, "noumenon" and "phenomenon," should be rejected as suggesting Kantian, not Buddhist philosophy. Between Tu Shun (also Fa Shun), the first patriarch and most creative thinker of the school, and Whitehead there are the geographical and cultural distance between China and Europe, the temporal distance of some 1300 years, and the cultural distance between a Buddhist hermit who was a philosopher-saint and miracle-worker, and a professor of applied mathematics, partner in the composition of Principia Mathematica, who later turned metaphysician. The comparison I have made is a sin against context, but only on the assumption that the context chosen is the most conventional one — for China is not Europe, a Chinese Buddhist is not a British mathematician, and so on. I will argue below that while context-choosing is essential, what people mean by context differs widely. However, at the moment, I argue only that if we choose a high enough level of abstraction, it makes sense to say that Tu Shun and Whitehead were faced by the same problem, which the Buddhist put as the principle of mutual interpenetration. He meant, in his own words, that "one includes all and enters all, all includes one and enters one, one includes one and enters one, and all includes all and enters all."19 In my paraphrase, it is true that everything is the context for everything, is relative to everything, includes everything; but that makes no difference because just as things seem, so they are, just as independent as they are relative. It appears that the problem of total relativity, which is that of context made completely general, has been faced and answered in different times and places in not dissimilar ways, sometimes, as in Sextus Empiricus and Montaigne, with merely skeptical conclusions, and sometimes, as in Chuang-tzu and Nagarjuna, with mystically skeptical ones. Sometimes, too, the problem has been faced as in the Flower Garland School, with what I must call,

88 B. -A. Scharfstein without explaining, positively interpreted negation or reality identified with illusion, as if, to make an analogy, the selfsame coin had a reality-face and an illusion-face that only the enlightened person could both distinguish and recognize to be the same. Sometimes it has been faced as in Whitehead, with his principle of universal relativity or interdependence, influenced, I assume, by his interest in Einstein's theory of relativity and aimed at characterizing a creative, emergent universe. These answers, for all their metaphysical boldness, their brave show of consistency and sometimes humor, their ardor and implicit tolerance, are intractably vague when taken out of their local context. When I say this, I remind myself of the sin I denied I had necessarily committed. We get an insight that, taken very generally, is arguably the same in its very different advocates. But the attempt to capture the truth of relativity, when taken, not so generally but in its local detail, shows itself to be as subject to time, place and circumstance as any other philosophical idea. If the problem of context or of relativity troubles us, we can begin to think about it in the spirit of these past thinkers; but their help will not absolve us from doing our own intellectual work, with the assumptions and distinctions that appear plausible to us now, in the light of our knowledge of science and the history of thought. Our path has led us from the ordinary, reasonable idea of context to the extraordinary idea of total relativity. Total relativity is a difficult position to hold because, in the end, it must accommodate its own refutation. Relativity, as we have seen, is fortified by the consistency of its reasoning; but if it attempts to complete its train of reasoning, it must relate itself as a whole to its opposite, which is the absolute to which relativity as a whole can itself be relative. In accommodating its own refutation, total relativity shows itself to be irrational. This is not surprising because our survival depends in practice on the stubborn constancies by which nature has taught us to live. However, what I have suggested in this concluding paragraph must be argued in another place. 20

On the Rationality of Context 89 Notes 1. For example: B. Berlin and P. Kay, Basic Color Terms: Their Universality and Evolution ( Berkeley: University of California Press), 1969; G. Devereaux, Ethnopsychoanalysis (Berkeley: University of California Press) 1978; J. W. D. Dougherty ed. Directions in Cognitive Anthropology (Urbana: University of Illinois Press), 1981; A. I. Hallowell, Culture and Experience (Philadelphia: University of Pennsylvania Press, 1955); C. R. Hallpike, Foundations of Primitive Thought (Oxford: Oxford University Press, 1979); A. Kleinman and S. Good, eds., Culture and Depression (Berkeley: University of California Press, 1985); R. A. Shweder and R. A. LeVine, eds.. Culture Theory (Cambridge: Cambridge University Press, 1984); D. G. Spindler, ed., The Making of Psychological Anthropology (Berkeley: University of California Press, 1978); G. M. White and J. Kirkpatrick, eds., Person, Self, and Experience (Berkeley: University of California Press, 1985). 2. W. Pluhar, "How to Render Zweckmassigkeit in Kant's Third Critique," in M. S. Gram, ed., Interpreting Kant (Iowa City: University of Iowa Press, 1982). 3. H. Mertens, Kommentar zur Ersten Einteilung in Kants Kritik der Urteilskraft (Munich: Johannes Berchmans Verlag, 1973). 4. R. Colie, "The Essayist and His Essay," in J. W. Yolton, ed., John Locke (Cambridge: Cambridge University Press, 1969). 5. H. I. Marrou, A History of Education in Antiquity (London: Sheed and Ward, 1956); S. A. Gait, A History of Chinese Educational Institutions, vol. 1 (London: Probsthain, 1951); R. J. Mookerji, Ancient Indian Education (London Macmillan, 1947). 6. E. A. Solomon, Indian Dialectics, 2 vols. (Ahmedabad: B. J. Institute of Learning and Research, 1976, 1978). 7. G. E. R. Lloyd, Polarity and Analogy (Cambridge: Cambridge

University Press, 1966); Magic, Reason and Experience (Cambridge: Cambridge University Press, 1979). 8. E. R. Dodds, The Greeks and the Irrational (Berkeley: University of California Press, 1957); B. Simon, Mind and Madness in Ancient Greece (Ithaca: Cornell University Press, 1978). 9. W. N. Hargreaves-Mawdsley, Oxford in the Age of John Locke (Norman: University of Oklahoma Press, 1973). 10. D. J. Eck, Banares (New York: Knopf, 1982). 11. C. Kahn, The Verb "Be' in Ancient Greek (Dordrecht: Reidel, 1973); D. Hansen, "A Tao of. Tao in Chuang-tzu," in V. H. Mair, ed., Experimental Essays on Chuang-tzu (Hawaii: University of Hawaii Press, 1983); A. C. Graham, " 'Being' in Classical Chinese," in J. W. M. Verhaer, ed., The Verb 'Be' and Its Synonyms: Classical Chinese, Arthapaskan, Mundan (Dordrecht: Reidel, 1967). 12. S. Gonda, Change and Continuity in Indian Religion (The Hague: Mouton, 1965); L. Renou, Histoire de la langue Sanskrite (Lyon: Editions IAC, 1956).

90 B. -A. Scharfstein 13. W. D. O'Flaherty, Dreams, Illusion, and Other Realities {Chicago: University of Chicago Press, 1984). 14. A. Kroeber, Configurations of Culture Growth (Chicago: University of Chicago Press, 1944); The Nature of Culture (Chicago: University of Chicago Press, 1952). 15. E. E. Walter et al., Civilizations East and West: A Memorial Volume for Benjamin Nelson (Atlantic Highlands: Humanities Press, 1985). 16. R. Capra, "Holonomy and Bootstrap," in K. Wilbur, ed., The Holographic Paradigm (Boulder: Shambala, 1982); M. Schofield, An Essay on Anaxagoras (Cambridge: Cambridge University Press, 1980). 17. G. Chang, The Buddhist Teaching of Totality (University Park: University of Pennsylvania Press, 1971), p. 147. 18. A. N. Whitehead, Process and Reality, corrected ed. (New York: Macmillan, 1978 [1929]), p. 20; Chang, ibid, (note 17), p. 220; A. N. Whitehead, Adventures of Ideas (New York, Macmillan), p. 91. 19. Chang, ibid, (note 17), p. 222. 20. See my book, The Dilemma of Context, to be published by New York University Press in 1989, from which this essay is largely taken.

Part II Comparative Rationality

The Rational and Irrational in Medieval Jewish and Islamic Philosophy L. E. Goodman

Some forty years ago, when E. R. Dodds opened his Sather Lectures at Berkeley,1 he recalled the need to vindicate, in the face of Roger Fry and others, the thesis that the Greeks were aware of and sensitive to dimensions of experience not readily assimilable to "what was commonly recognized as rationality. The sense of mystery, of energy, of paradox, or of mere fun had been so filtered from the scholarly accounts that it came as a shock to many to discover that the 'austere' statues of Greek sculptors were originally painted in polychrome and often clothed. Students who had pored over their Greek Iliads could hardly fail to perceive what they were doing as the opposite of adventure, and were as unlikely to discover irony in Plato, energy in Pindar, terror in Aeschylus or hesitancy in Aristotle as were the Protestant Divines (who often were their teachers) to uncover the ironies, parodies, mocking singsong voices and banter of the Hebrew Bible. The Greeks were robbed of their sense of mystery (despite Aristotle's saying that philosophy begins in wonder) by the same exigency that deprived the Hebrew writers of their laughter: not for any failing of their own but for the sake of the function they performed and the image they held in the eyes of their later exponents and exploiters. Dodds and others have cleared the matter up — so thoroughly, in fact, that one can scarcely find a production of a Greek drama that is not over-starkly lighted and over-expressionistically played. Art historians now seek the same elemental energies in Greek terra cotta vase paintings that Picasso found in Benin bronzes. And Plato, Aristotle and even Socrates are as often celebrated (or damned) for their prejudices as assayed for their arguments. In the present paper I seek responses to the idea of the irra-

94 L. E. Goodman tional among a few of the major philosophic thinkers who wrote in Arabic in the Middle Ages. These men were in several ways heirs to the Greek traditions of philosophy, natural science and mathematics. It seems unwise here to attempt to define the irrational at the outset. Perhaps the reference of the term is not to a single univocal category. And it seems best in any case to let the sources speak for themselves. Part of our story is the clarification of what a given author will identify as irrational, and part involves appreciation of his or her grounds for doing so. But we do need a few definitional caveats: I do not think that here — where all of the greatest thinkers we confront were ethicists, most were natural theologians, and many were rational mystics — we can simply identify the irrational with the emotional, the religious or even the mystical. To assume such an identification is not only to disparage whole branches of discourse without warrant but also to narrow, a priori, the scope of what we shall deem rational. Reason may and should warrant, inform, guide and direct our emotions, as was a major theme of Plato and of Aristotle. Reason may warrant our religious convictions, and can — perhaps must — accompany, guide and even inform the content of our religious experiences, at least if we are to make anything much of them, as was a major theme of Ibn Tufayl, Plotinus and, again, Plato. Yet, clearly, not every emotion can be rational; not every belief is rationally held; not every experience is amenable to the direction of the rational ego. In seeking responses to the irrational we are also seeking boundary conditions of rationality — seeking, that is, not so much to apply as to define a concept, to define it in use. This can be a slippery job. Freud is commonly celebrated for paying new heed to the irrational. But, more coherently and more accurately, we can say that Freud discovered dimensions of rationality in what had often been dismissed as irrational, incomprehensible experiences. Dreams, for example, were found not only to have a language but even to do work. The death wish itself was made intelligible, not as a pursuit of annihilation per se, but as an evasive maneuver of self-protection — protection against exposure and the possibility of failure and disgrace — or as the introjection of others' hostility in the form of feelings of guilt. By such dialectics the irrational was not encountered as such but overcome, made rational within the context of a newly broadened conception of rationality, much in the pattern that

Medieval Jewish and Islamic Philosophers

95

Augustine discovered that his youthful destructiveness with stolen pears was not the pure pursuit of evil for its own sake but a misguided groping at the good, seen under the guise of pure freedom and abandon, power and fellowship. Or, in much the way that Socrates — that consummate rationalist, patron saint and martyr of rationalism — denied that we ever knowingly choose the less good. If confronting and accommodating the irrational is a desideratum, then, in a way, Milton confronts the irrational more squarely than Freud, when Milton has Satan utter the world-wrenching paradox: "Evil, be thou my good." Three models emerge for addressing the irrational: a contrastive model, in which the rational and the irrational are played off against one another; a reductive model, in which the irrational is in essence eliminated by the discovery of its inner rationality; and an assimilative model, in which what we take to be rational and irrational are accommodated or related to one another under some higher concept, neither being dissolved into the other, but both being subsumed in a larger unity where their opposition is revealed to be neither insignificant nor absolute. Among the Muslim and Jewish philosophic thinkers who wrote in Arabic and who responded to the translated philosophic and scientific writings of the Greek tradition, three of the earliest, alKindi, al-Razi and Saadiah, represent the contrastive approach. The classic Aristotelian neoplatonists — the falasifa, as they proudly styled themselves, using the Greek term — al-Farabi, Ibn Sina and Ibn Rushd, represent the reductive approach. And al-Ghazali and Maimonides, the great synthetic thinkers who sought to answer and rebut the falasifa while accommodating what was sound in their findings and absorbing what was rigorous in their method, were exponents of what we call the assimilative approach.

The Rational versus the Irrational The first significant philosopher to write in Arabic was al-Kindi (d. 867), the scion of an Arab princely family, a courtier, physician, theorist of music and astronomy, and a patron of the translators whose work in rendering Greek writings into Arabic made Arabic philosophy possible, and in some sense necessary. Influenced by the sixth-century Christian cosmologist John

96 L. E. Goodman Philoponus, Kindi defended the monotheistic thesis of creation. Against the eternalism of the Aristotelian tradition he argues that if the cosmos had an infinite past, it would never have reached the present. 2 Clearly, Kindi did not know or did not fully share that sense of rigor that was to reach other Arabic-writing philosophers in the arguments of Aristotle to the effect that, if the world began, then there was a beginning to time, change and the causal process — an irrational disruption of the seamless continuity of becoming, in which each event always has a prior cause, and none is left a widow or an orphan. Kindi saw no irrationality in the idea of God's creating something out of nothing {'ayson 'an laysin).3 This was not an instance of the irrational, unaccountable and therefore unacceptable derivation of what is from what is not. God's power was the prior something which precluded creation from being an irrational violation of the universal rule ex nihilo nihilfit. Clearly, irrationality is unacceptable; but Kindi has his own standards of rationality. Accordingly, Kindi follows a Mu'tazilite mode of argument in glossing a well known Qur'anic passage (55:5) about the prostration of the heavenly bodies before God: The verse portends divine sovereignty over the cosmos. 4 Monotheism and rationalism go hand in hand, since only monotheism provides a comprehensive accounting for the cosmos. The theme is taken up in the Hebrew liturgy in the hymn still sung on Sabbaths: "El Adon 'al kol hama'asim ... shevah notnim lo kol tseva' marom ..." — "God is Lord over all of His works ... All the supernal host pay homage to Him ..." All, indeed stand in dread of Him; they shed their influences by a delegated power; none are to be feared or held in awe in their own right. Monotheism presents a clear as well as a comprehensible universe. There are no disparate powers exercising ungoverned or arbitrary passions upon nature. What is Divine is wise and good as well as powerful. Indeed, as Saadiah argues in a Platonic argument elicited from scripture: it is because God is just that God is king.5 Kindi's rationalism is not confined, of course, to the heavens. Writing for his princely patron, who requested a kind of philosophic anodyne against grief and anxiety, Kindi exemplifies the best rationalist tradition in urging that, "For any kind of malady, where the cause is not known, no cure is to be found." The cause of grief is loss, and that of anxiety is the fear of loss. 6 Here philosophic analysis plays the role of medical diagnosis and

Medieval Jewish and Islamic Philosophers

97

links itself immediately to the prescription of a remedy: we must shift our assets, our desires and affections, to the intellectual realm, whose goods are not subject to loss like those of the transitory, sensory realm, and indeed cannot elude the mind that seeks them. Detachment from the evanescent — and attachment to the permanent and unchanging — is the condition of our happiness. We must regard as already lost all that is capable of loss. Acceptance and resignation, made habits of the mind, will liberate us from struggle, anxiety and grief.7 Kindi detects no irony in the Stoic notion that freedom is the acceptance of necessity. His attraction to the Platonism he promotes conspires with the demands of his prince's assignment, pressing him to conceal the sophistry of a consolation that equates loss with liberation and purports to value all that we love — persons as well as things — only insofar as it never can be lost. Kindi retells the tale of Alexander charging his mother to summon all those who had never suffered a loss to a great feast upon his death. When the feast was announced and no guest appeared, Olympias knew that her wise son had meant to comfort her: she was not the first mother to lose a son. 8 But if losses are real as well as universal, what is the meaning of detachment? We cannot banish pain by dismissing it or cease to cherish those we love by recognizing that our ties to them are temporary and our claims on them only partial. If giving hostages to fortune is inevitable in the human condition, and if effort and persuasion are required to feel as though these bonds were never ours, then the notion is illusory that those who have suffered loss will be the same through Kindi's tactics as those who have never lost. We acknowledge the reality and indeed legitimacy of our attachments in the very tenderness and insistence we employ in seeking to break free of them. Yet Kindi is capable of turning the tables on such logic with the same deftness he employs in turning the tables on Aristotle's eternalism. 9 If giving hostages to fortune is inseparable from the human condition, that fact is not to be bewailed: "If we want never to suffer a loss, then what we really want is never to have been at all."10 Recognizing that attachments to the ephemeral will be sources of grief, we should minimize, not maximize them: Grief is a burden we impose upon ourselves; to load ourselves down with its occasions is senseless and irrational. Even animals, with their instinctual sense of order and proportion,

98 L. E. Goodman are more reasonable: "The irrational man observes no order or sense of proportion in what he does. His mind is in disorder, and his actions betray it. We should be ashamed to live the life of such a wretch, object of the laughter of the foolish and the pity of the wise." 11 Saadiah makes a similar point in his commentary on the Book of Job. Glossing the verse which he translates, "A man with seeming heart of wisdom, that man is a wild ass at birth" (Job 11:12), Saadiah explains that "a wild beast eats what it can find and mounts any mate it encounters. The wicked do the same ... The analogy is all the more pointed because he composes them not with the domestic but with the wild ass, and not when it is mature but at birth." 12 The life of appetites and passions is the height of unreason and the antithesis of the Divine life, where there are neither appetites nor passions. Just as we know that God rules because He is just, so God knows that His angels will not rebel and that His righteous servants will pass the tests they must confront from birth to death. For character and reason go hand in hand: A being that has anger or spite will also have appetites, concupiscence and the fear of death. Only one who is ignorant, venal or vulnerable can be unjust. 13 Kindi illustrates his conception of the rational economy of our attachments with an allegory of the human condition drawn from the traditional repertoire of popular philosophy: our passage through this world is like a journey home by ship. When the vessel puts in for provisions on an isolated promontory, some of the passengers go ashore and are lost among the woods and fields, soon forgetting the vessel and the purpose of their journey. Others enjoy "the meadows filled with flowers of all colors, smell their fragrant blossoms, wander in the fields and the lovely woods full of strange new fruits, hear the enchanting calls of unseen birds, and see the varieties of brilliantly colored rocks in the soil of that country, that are so fine to look at, and the delightful seashells with their wonderful shapes." They then return to their old seats, "since the better ones are taken." Still others forget to gather supplies but come back laden with rocks and shells, flowers and fruit, which they cramp into the little space remaining to them. But their booty soon grows foul and noisome and must be jettisoned, since it brings their owners — now their servants — only worry and trouble. Such are all the transitory things men scrabble for: sources of worry and grief.14 An ascetic, intellectual

Medieval Jewish and Islamic Philosophers

99

course is far the wisest, because it teaches renunciation of what is not truly worth the having. "To be sorry because we have nothing to be sorry over is the mark of the irrational."15 The key to rationality in Kindi's view, as in the view of the Stoics, is found in discriminating what we can control from what we cannot control. 16 And the key to that discrimination is found in the Platonic claim that intellectual goods are ours for the seeking and permanent once grasped, and that sensory goods are perishable and vexatious to acquire and retain. The real value of any good is coextensive with its attainability and durability. Kindi's Philoponean creationism underscores his commitment to a Platonizing dualism. For, in an argument that Kindi took most seriously, 17 Plato had argued that all temporal things are transitory, that whatever is subject to change must come to be and pass away, that only the changeless is eternal. It was on this basis that Plato had argued for the world's origination, and it was on the same basis that both Philoponus and Kindi found, in the transitoriness of all natural things, an almost triumphal affirmation of God's sovereignty over nature as its Creator. 18 The destruction and the finitude of all things natural is the hallmark and impress of their eternal Counterpart. Kindi pictures the nutriment in our livers, ignorant of the world beyond and in trepidation at the higher realization awaiting it on being transformed into semen. Just as semen might fear the womb, the fetus would fear life, and man himself fears death. The changeable world is an anteroom to the higher and better life of Platonic permanence, where no suffering or loss are possible. Rationality is cleaving to that higher world, and death is no more to be feared than life. Irrationality is cleaving to the world of specious familiarities and the objects of the appetites and passions, which by their very natures render us subject to suffering and loss — in exact equilibrium with all that we have gained.19 Like Kindi, Saadiah affirms creation and believes this world to be a prelude. He does not wholly condemn those who linger a bit to admire the colorful rocks and smell the flowers. Pure asceticism, he argues, is just another monomania, like the obsessive quest for vengeance, or for sex, or romantic passion, urban or agrarian development or any other partial good whose exclusive pursuit does not enhance but rather precludes the good life. Treating abstention as the sole overriding good yields a melancholy and misanthropic life, bitter, angry and lacking in the

100 L. K. Goodman warmth, appreciation and sense of grace and graciousness that characterize true piety.20 Like Kindi, Saadiah retells his own version of the allegory of the fearful fetus and explains that only ignorance accounts for our regarding death with dread. 21 If we knew that the trials of this world presage the enduring bliss of the hereafter, where the righteous will be nourished by the same light that, in the form of fire, wastes the wicked, then we would know how little we have to fear in death. But such knowledge comes to us only through reason, mediated by our conceptual knowledge that God is just. If God were to tell us that the sufferings to which our embodiment makes us liable are mere trials of our mettle and to be requited, the trials would be trivialized. Life would lose its meaning. For its worth lies in its authenticity. The value of our endurance rests in our intellectual fortitude.22 This world is a means to an end, for Saadiah, the theater and vehicle of God's justice. But it is not, for that reason, of no intrinsic value. Rationality lies in recognition of the disproportion between eternal and temporal goods and ills; and irrationality, in taking any finite good as an absolute. Yet transcendent goods are achieved and transcendent evils perpetrated in this world: hence the justice of irrevocable rewards and punishments. And it is only through this world and our doubting and halting choices in it that infinite requitals are justly assigned. Thus life in the world acquires an intrinsic value, not from its contrast with but from its perfusion by the eternal. Our existence is genuinely better than non-existence, and our choices do genuinely matter. For, only by these means do we merit eternity and overcome the disproportion — the incommensurability — between the finite and the Infinite.23 Al-Razi, too, is a creationist, although an advocate of Jormatio mundi rather than of the absolute creation of a Kindi or a Saadiah. And Razi, too, subscribes to the Platonizing two-tiered world that Kindi and Saadiah employ in interperting the nexus between the temporal and the eternal. But Razi's dualism is more Gnostic in flavor than that of Kindi or Saadiah. In his philosophic iconography, the image of the frightened fetus, peering out, as it were, at a forbidding yet bounteous world from the confines of the womb, is displaced by the image of the cosmic soul, yearning toward the still unfounded world, in which she longs passionately to be embodied, not knowing, in her eagerness, of the pains and sufferings that await her there. 24

Medieval Jewish and Islamic Philosophers

101

Pains and sufferings are, for Razi, very much of the stuff that life in the world is made from. For Razi follows the Epicurean mode of re-reducing the old Platonic calculative model of rationality in the world to a calculus of pleasures and pains. 2 5 Embodiment remains the heel of our vulnerability. And through the exposure that embodiment represents, Razi can be certain (by an argument Saadiah accepts from him) that pleasures will always be outweighed by pains in this world; for every pleasure, he argues, in true Epicurean style, always presupposes a corresponding pain or loss, but the reverse cannot be claimed within this world. 26 It is perhaps because of the pessimistic, Gnostic flavor of Razi's cosmology that the irrational acquires a somewhat more positive aspect in his philosophy than it has in those of Kindi and Saadiah. Seizing hold of Aristotle's introduction of chance or spontaneity as a nominal third alternative to nature and art in the generation of things, 27 Razi argues that the world originated neither by a natural nor by a constrained motion, neither by God's all-seeing wisdom nor by His arbitrary will, but by the irrational and ill-informed, spontaneous but uncontrollable desire of Soul. Soul, the supreme principle of motion and discursive thought, can learn only from experience. It would have done her no good for God to inform her in advance of the dangers and pains in store for her in the world she was to generate. An understanding father knows that his headstrong son will not learn the dangers of the lovely garden he longs to explore until the boy has entered the garden, seen its beautiful flowers, felt its thorns and the sting of its insects for himself. He lets the boy explore, not because he is powerless to stop him but because he knows that the boy will know no peace until he has tasted for himself. In the same way, and for the same reason, God yields to the impulsive desire of the Soul, whose fall into nature was a spontaneous and unreasoning event, comparable, the physician argued, to a dignitary's farting in public. 28 When the Soul fell she set the entire preexistent realm of matter whirling in a chaotic and disordered motion. But God, in His mercy and wisdom, imparted reason to her, enabling her to form and order nature, discover her alienness within it and rediscover her true home. Without reason nature would be an irrecoverable chaos, and man would be irremediably lost. But without the ignorant and ungoverned passion of the unreasoning soul, nature —

102 L. E. Goodman with all its pains and all its beauties — would never have come to be. Razi retains the Platonizing dualism of Kindi and of his younger contemporary Saadiah. But by heightening the contrast between the vulnerability of this world and the impassivity of the other, Razi stretches the notion of the insufficiency of this life into a claim as to its inadequacy. He reverts to the Gnostic idea that creation is a mistake, or at least the product of unreason. The world, in Gnostic terms, is a trap to be escaped, a sepulcher. We are, as it were, buried alive. At the same time, however, the world is a garden for Razi, its pleasures not to be denied or rejected simply because they are outweighed by pains. On the contrary, Razi counsels a life that maximizes pleasures by moderating desires. The Epicurean sense of an intimate relation, perhaps equation, between voluptas and voluntas remains. Reason warns us of dangers, but reason is not the principle of life or motion that reaches out to combine things in new ways. Reason urges us beyond life, and only unreason urges us to linger. Rationally speaking we know that death is better than life, even if there is no eternal aftermath. 29 As a result, paradoxically, and to the extent that life holds any joy or good, irrationality becomes a beneficial and indeed creative principle for Razi.

Reducing the Irrational to Rationality Razi's treatment of the Soul's irrationality calls to mind the saying of R. Samuel ben Nahman: "Without the evil inclination no man would build a house, take a wife, beget a family and engage in work." 30 Here the principle of irrationality is subordinated to a higher rationality through its identification as a principle of creativity, the energizing drive of all that is productive. The difference, of course, between Razi's dialectic and that of Rabbi Samuel lies in Razi's greater ambivalence about the value of this life: Razi believes the world contains less good than evil because he reckons it contains more pain than pleasure. On this basis the inferred irrationality of the world's generative principle will never be fully reducible to rationality. But such a reduction is the objective of the Jalasifa, whose model of creation is not the polar tension between God and nature that Razi accentuates into an opposition, but rather the emanative pervasion of nature by the

Medieval Jewish and Islamic Philosophers

103

being and light of the Divine. In such a schematism, irrationality is naturally absorbed into the pure rationality that pervades the whole. The natural world, al-Farabi argues, is in effect a precipitate of the intellectual activity of the celestial intelligences. 3 1 Accordingly, it bears their imprint, not by its mere complementarity to the Divine realm — being contingent and temporal where the eternal intelligences are necessary and eternal — but because it expresses their character in its own. For the natural world as a whole is eternal in Farabi; and even its particulars vie for eternity and, in a sense, attain it. Given the manner in which natural things acquire their existence, Farabi writes, "once they exist they tend to endure and persist." 32 Composite particulars can, of course, be destroyed by an opposing form to their own. But each deserves, in a sense, to retain the form it has. At the same time, in virtue of its matter, each particular has a "desert or entitlement" to the opposing form, to which its matter is no less susceptible. "Since these two entitlements cannot be satisfied simultaneously, particulars must alternate, fulfilling the one for a time and then the other. Thus a thing comes into existence and endures for a time, its existence protected, and then it is destroyed and its opposite comes into existence — and so,unceasingly. For neither of the two deserves existence or persistence more than the other; but each gets its own share in existence and survival."33 Matter must alternate forms.34 As it does, counterclaims, as it were, arise, by whichever of two opposites is excluded for the moment: "The justice that operates in this is in that matter is taken away from this and given to that or taken away from that and given to this, in alternation. It is because justice must be satisfied vis-a-vis these existents that no one thing can last forever in terms of its arithmetic identity. A thing's eternal duration must be assigned entirely to its identity as a species."35 Species have a special unity and thus a special claim on being, as is clear from the necessity and universality of the propositions of the sciences: Only of species can we say that they are what they are in the full and proper sense, not partially, transiently or equivocally. Where the unity of a particular clearly does not imply self-sameness of a sort that would require perpetuity, the unity of a species does: each species exhausts the potentialities of a type and so does not need to yield its ground to another. All

104 L. E. Goodman species are eternal and exemplify, through their eternal duration, the timeless eternity of their archetypes. Individuals are not eternal. Yet the laws of succession, alternation of forms and equilibrium of types bespeak a higher harmony and unity. Farabi's point elaborates an idea of Anaxagoras', brought to fuller articulation by Aristotle when he took Empedocles to task for treating "love" and "strife" as opposites, and not seeing that generation and destruction are one and the same process viewed from opposite (and thus still partisan) perspectives. 36 Farabi equates the natural justice he finds here with the Divine maintenance of order and perfection (ihkam, itqari) in nature, that is, with God's providential concern ('inayd) and wisdom. It is on the basis of his finding these qualities in nature that Farabi denies Divine neglect, deficiency and injustice toward the world — the cardinal elements of the Epicurean problem of evil.37 The justice Farabi finds or vindicates in the natural changes and interactions of the cosmos, then, are emblematic of the higher justice which the idea of God's governance traditionally represents: the denial of the Epicurean claim that God lacks the power, wisdom or concern to govern the world for the best. Where Razi denies that God's justice extends to this world, except insofar as it affords us a way out of it, and where even Saadiah agrees that God's justice requires His provision of an alternative to this life, lest its many unresolved tragedies be taken as evidence of God's unconcern, Farabi takes the pattern of nature itself to argue for God's justice and thereby for the supremacy of God's wisdom. The cosmos is a projection of that wisdom, and the process by which it precipitates is itself intellectual in nature. Emanation reaches particularity not because the inexhaustible source of being peters out, but simply because the outflow arrives at the stage of instantiation, •where each product can no longer exhaust the reality it is capable of, and each must pay deference seriatim to the demands of the matter and form which give it being as a particular. Only by giving up all claim to perpetuity as a particular can any particular reach concreteness as a particular. Cosmically, there is no irrationality. For the Muslim rationalist philosophers, as for their radical monotheist predecessors in the Mu'tazilite schools of kalam, God has no real attributes distinct from His essence. Indeed, as Avicenna is careful to point out, the idea of God's uniqueness as a necessary being — the only absolutely necessary being — is cap-

Medieval Jewish and Islamic Philosophers

105

tured in the recognition that in God and in God alone, essence is identical with existence. If God had a character or nature distinguishable from the Divine Identity [dhat), God would be a composite being and thus contingent, that is, God would, in effect, have matter and would be temporal like any physical object, subject to change, corruptible and in fact created.38 Yet, while affirming the absolute simplicity and t h u s Plotinian impassivity of the Divine39 in the Mu'tazilite manner — the manner that the Mu'tazilites boasted of when they called themselves proponents of God's unity and justice — the falasifa, as exponents of the arguments of Aristotle and Plato, fused an Aristotelian method of reasoning to their kalam thesis, in much the way that Saadiah found a Platonic argument in scripture to deploy on behalf of the thesis of God's justice: God's simplicity could be defended on the basis of God's intellectuality. For God was pure Intelligence, as Aristotle had reasoned — inasmuch as mind is the most actual and self-sufficing of things — and thought, the noblest and least externally dependent of activities. It follows that God is conscious and aware, knowing all things in the forms which are the principles of all things, and causing and governing through that knowledge. But it also follows — despite all the diversity God sends in train — that God is self-simple and undifferentiated. For, as Aristotle had argued, a mind is identical with what it knows and with its act of consciousness. It is no different thing. 40 If all of God's traditional attributes reduce to intelligence, God's life and power, compassion and concern, justice and mercy, do not generate a plurality in His nature, and God remains selfsufficing and self-explaining, in need of no external cause to create or sustain Him. Only so, the Philosophers reasoned, is God capable of being God. What the Avicennan/Aristotelian program called for, in effect, as the price of eliminating unreason from the principles of the universe, was the reduction and collapsing of God's will into reason. By the time of Averroes this collapsing had become a theological cause celebre. Clearly alarmed by the theistic subjectivism and apparent anti-naturalism of the Ash'arite doctrine of divine voluntarism, Ibn Rushd rejects the notion that God's authorship of the world is an arbitrary act: We have experience of two kinds of causes. The first produces only a single effect in accordance with its identity, as what is hot produces heat and what is cold produces

106 L. E. Goodman cold. These are what philosophers call natural causes. The second sort are things that can produce one effect at one time and the opposite at another. These we call voluntary or choosing. They act on the basis of knowledge or opinion. But the First Cause, praised be He, transcends both of these descriptions in any sense that would be applied by philosophers to transitory causes. For one who chooses and wills lacks what he wills, and God, exalted be He, lacks nothing that He might will. One who chooses prefers one good to another for himself, but God needs no improvement. Further, when one wills, generally one's willing ceases when one achieves what was desired, for the will is an affect and the outcome of a change. But God, exalted be He, transcends all change and passivity. Still more does He transcend the natural mode of acting. For the action of a natural thing is a matter of necessity in its very substance, not a necessity of its will but of its entelechy. And natural causation does not work by knowledge; but that God's acts, glorified be He, issue from knowledge is a matter of proof.41 Thus, Ibn Rushd discovers a profound asymmetry between the traditional attributes of will and wisdom: God acts out of knowledge, as a rational being, but God's acts are not choices in the human sense, not expressions of preference. God's will is His wisdom, and so it must be, lest God be subjected to the vagaries of time, change and imperfection. God's act is not automatic — neither spontaneous, as in Razi, nor discretionary, like a human act that might have taken one direction or another. For, if God's act is arbitrary, God is placed under the sway of subjectivity, the world's determinations become matters of preference (rather than absolute requirements of pure wisdom) and God becomes not merely a model of tyranny and caprice but an author of evil. The issue of voluntarism versus intellectualism in God is bound up with the issue of creation versus eternity and thus with the Muslim philosophers' commitment to the idea of a full universe. 42 The Philosophers' rejection of voluntarism for God — or, rather, their reduction of Divine volition to rational wisdom — is a rejection of the ascription of irrationality to God. For Farabi, Ibn Sina and Ibn Rushd all follow the ancient arguments of Proclus and Simplicius, running back to Aristotle and Parmenides, regarding an origin for the world as a disruption of the continuity of nature and the rationality of God.43 If reality came to be, Parmenides had asked, how did this come

Medieval Jewish and Islamic Philosophers

107

about, out of what, and for what reason did it occur now rather than earlier or later? 44 Similarly Aristotle finds incoherence in the quest for a becoming of becoming. 45 What would be the substrate of such a change? And Proclus makes explicit the theological dimensions of the argument: What was there in one still non-existent instant that made it preferable to another as the moment of creation? Or, are we to believe that the Divine gave preference to one moment over another for no reason whatsoever, or that God chose to act after eons of inactivity, as though suddenly changing His mind about the relative merits of creation versus solitude and sterility? On the contrary, Proclus argued, if the Divine is a Creator in essence, then the act of creating and its product must be eternal. Creation is no mere whim. 46 And, for the Muslim neoplatonists, as for their Greek predecessors, the means of creation was thought, the one act that contains no flaw or compromise with multiplicity.47 For the Philosophers, intellectuality was no more a threat to God's justice than to His unity. Reducing God to intellect, that is, to the Divine attribute of wisdom, allowed them to regard nature as wisely fashioned in all respects, a consummate expression of God's perfect wisdom. Retaining an unreduced attribute of pure will in God seemed tantamount to the admission of irrationality in God and the retention of unreduced and irreducible surds in nature. Such concessions to positivity seemed defeating to the rationalist project of universal explanation — the reduction of the cosmos to intelligibility — and frustrating to the larger cosmological project of Neoplatonism, the derivation of the cosmos as a whole and in detail from the intellectuality of God, that is, from the Platonic forms. The same concessions, moreover, seemed to bear noxious implications for the polity — for the state, and for religion, the "elder brother" of the state, as the Ikhwan alSafa' called it — in the form of legal positivism, the admission of incomprehensible, purely positive and arbitrary obligations. With these thoughts in mind, Averroes insists that God's role as Author of nature be assimilated to no natural or voluntary model: He is above the entire system of such causes, the Cause of all causes, "bringing all into being out of nothingness and preserving all in a higher sense than is applicable to mere empirical causes" — not temporally, of course, but as the eternal Source and Goal of emanation. 48 God has a will, Averroes insists, and God does make choices, but not arbitrarily. The notion of will applied

108 L. E. Goodman to God is not in any sense applicable to human beings or apprehensible in nature. God's will must be commensurate with His being and power. For, unless God's will were congruent with His wisdom, in no way exceeding or overpowered by it, God's 'faculties' or powers would be on a par with ours. We human beings have a will that can outrun our wisdom and a knowledge that can outreach our power. But, in God, all is harmony and perfection, and all powers resolve to one — to God's unity itself.49 To be sure, Averroes argues, the world may appear to us, in our ignorance (and in proportion to our ignorance), to contain many arbitrary determinations. To the ignorant it appears that the highest heavens might have revolved from West to East, or that the cosmos as a whole might have had a different size than it has. Many ascribe such determinations to the arbitrary choice of God. But the wise, who study nature through the sciences and thereby recognize necessity in nature and discern the consummate wisdom of the Divine plan in nature's workings, understand that there is no arbitrariness in nature, just as there are no pettiness or passions in God. To the ignorant a crab might appear a silly creature, whose sidewise gait and ungainly limbs bespeak the playfulness or foolishness of its maker. But scientists, who understand the anatomy and environment of the crab, would no more expect a crab to walk like a man than a man to walk like a crab. 50 In a similar vein, the Ikhwan al-Safa' argued that, when scripture reports in God's name, "We formed man at the fairest height" [Qur'an 95:4), the reference is applicable to all animals and addresses the fitness of each species to its environment.51 Those who understand the wisdom in the fabric of the cosmos find no irrationality there: only from the sorts of elements that our world contains could any possible world be constructed. 52 Only with the movements that these elements undergo could nature persist: "Were a single one of the motions of these elements to cease, the entire order of the cosmos would be destroyed."53 The size of the cosmos and the directions of its motions are as natural and proper to it, as necessary, in fact, as are the anatomy and natural history of any natural organism; for, in a way, of course, the cosmos is a natural organism: The course determined for the body of the cosmos by its nature is the best possible course for it to move in, because it is the best of all bodies.... One who studies an artifact will not appreciate its wisdom unless he

Medieval Jewish and Islamic Philosophers

109

recognizes the wisdom of its purpose, the object it is intended to achieve. If one does not in the least understand the wisdom of a thing, one might readily suppose that the same object could have been made with any chance form or dimensions, any conformation or arrangement of parts whatsoever. This, precisely, is the error the mutakallimun fell into regarding the celestial bodies — all from thinking unreflectively. But, just as one who falls into such notions about man-made things has no understanding of their design and is ignorant of their maker but possesses only untrue suppositions, so it is also with created things. 54 For Averroes, God plays the role in nature that the young Socrates 55 expected Nous to play in the world of Anaxagoras: excluding all irrationality and seeing that all things are governed for the best. The world is consummately designed, and all is as it must be. God's will is no way different from His wisdom, and all natural events are expressions of that wisdom, whether or not we understand it or, indeed, miss its general principles, as the mutakallimun tended to do. To reject the conclusion of the Philosophers as to the perfect rationality of nature, argues Ibn Rushd, is not merely to reject the Aristotelian unity and harmony of the cosmos, but, in effect, to reject the unity of God as well: The old, pre-monotheistic theory was that two principles were needed to account for nature, "one for the good, one for the evil."56 In radical monotheism the idea of the absolute simplicity of God stands guard over the synthetic and integrative unity Of the world: When the ancients pondered reality more closely, they saw that all things that exist pursue a single goal, which is the order found in the cosmos, like the order in an army before its commander or in a city before its ruler. They concluded that the world as a whole must be organized in just this way. That is the idea expressed in His words: "Were there any other god but God in either world, both would be destroyed" [Qur'an 21:25]. They reached the conviction — from the place that good occupies in all being — that evil occurs only incidentally, as punishments are established in cities by good rulers. For these are evils instituted for the sake of the good, not for their own sake. 57 All finitizing conditions of existence — man's animality, for

110 L. E. Goodman example, Averroes argues — are ordained not for the inadequacies they bring in train, but because such conditions are concomitant to the manifestation of some intrinsic good. In this way, or in more sophisticated equivalents of this way, all things can be shown to be governed wisely and well, ultimately but consummately for the best. 58 Toward a Synthesis of Rationality with its Alternative Because the Mu'tazilites and their philosophic successors in Judaism and Islam were rationalists and optimists about nature, intellectualists about God, voluntarists about man and objectivists about morals, it is easy to fall into the trap of accepting their self-advertisement as "the proponents of monotheism and theodicy," taking them indeed to be the children of light, while casting their Ash'arite adversaries as children of darkness, defenders of unreason. Mu'tazilites were once commonly labeled the liberals' of Islam. Leibniz even paid them the compliment of calling them veritable Christians. 59 And sophisticated specialists like George Hourani, rightly impressed by the vigor of Mu'tazilite reasoning, sought analogies with recent schools of ethical thought by intentionally ignoring the theological preoccupations which are the nerve of Mu'tazilite reasoning. 60 There are, of course, no liberals in the strict sense among the classic Islamic schools of theology; and, in fact, the Mu'tazilites were notorious among the orthodox for their mihna, or inquisition, an attempt to ferret out the true beliefs of other Muslims whose faith they deemed heretical. The Mu'tazilites had never fully purged themselves of the Kharijite heritage of their school. Their mihna could not fail to appear to its victims and their sympathizers as a survival of the most odious tenet of Kharijite rationalism, the doctrine that mortals could adequately gauge the faith of their fellows from their acts (or, indeed, from their associations). In Kharijite thinking, this notion warranted — indeed mandated — the persecution and slaying as unbelievers of all those whose professions of faith were judged by such overt signs to be inadequate or insincere. The Ash'arite behaviorism and denial of enduring dispositions or capacities for action were entrenched as much by the dialectic of opposition to the Kharijite and Mu'tazilite conceptions of sincerity as they were by deference

Medieval Jewish and Islamic Philosophers

111

to the omnipotent sovereignty of God, in whom all power must be vested, and to whom, accordingly, all acts are due. 61 While Aristotelian philosophers and later scholars who gave them ear could see apologetics for tyranny in the Ash'arite doctrine that God acts as He pleases, Ash'arite mutakallimun had their own ideas of tyranny and did not readily forget that it was Mu'tazilite theologians and their princely supporters who had been the cause of their indignities and suffering.62 Politics aside, there is a counterstory to the charge of the Philosophers that Ash'arism made God a despot, the world unpredictable, the sciences impossible, a shiver or a twinge undistinguishable from a voluntary act and daily life undistinguishable from dreaming. The monovalent capacities of alAsh'ari do allow a kind of proto-naturalism. For events do not occur where no substrate is prepared for them. Ash'ari does not subscribe to the radical occasionalism he describes in figures like Salih Qubba and Salihi63; and, in fact, he does allow for science. For God in his grace follows a customary pattern in His dealings with nature and creates in our minds a habitual or accustomed expectation, through which we may know what to anticipate. The essence of a miracle is in God's making known to His chosen prophet when the familiarly expected outcome is not to be anticipated. 6 4 The epistemology is no more arbitrary or capricious than that of Hume. For Ash'ari, like Hume, is no mere skeptic. He relies on a Divine custom or habit ['adah) in much the way that Hume relies on the habit or custom of human nature to generate our familiar expectations. Al-Ghazali presses further in the same direction, evidently having abandoned the atomism of the occasionalist kalam, whose refutation he rehearses with such gusto in The Intentions of the Philosophers.65 He clearly espouses the naturalist thesis that events do not take place without their natural prerequisites, including the material substrates and dispositional properties which are their basis. Miracles await a change in the nature of a substrate (rather like transsubstantiation), when God's will so chooses; but creation does not need a substrate, because creation is not a natural event.66 Here we have the basis of Maimonides' accommodation of kalam voluntarism and creationism with Aristotelian naturalism: nature follows its ordained course within the cosmic system. But the necessities of nature are not absolute. The laws of nature, just as Ghazali had insisted, have

112 L. E. Goodman not the force of logic. Nor can they be derived deductively from knowledge of the wisdom and perfectness of God. For there is more than one course that nature might have taken. 67 Like Ghazali, Maimonides argues that God does distinguish — and, indeed, choose — between like alternatives, for example, in choosing the moment of creation. In so choosing, Ghazali argues, God does not find a difference among rival and virtual moments that might be made the first; rather, the nature of will (as distinct from wisdom) is to make a difference. Many such determinations must be made if a finite world is to exist, and they cannot all be prejudged by logic or even by the preference for the better or more rational, which philosophers assimilate to logic.68 Maimonides follows Ghazali in questioning the pertinence of ascribing the world to God's authorship if the world is to be God's act eternally — and thus without a genuine alternative. Authorship, Maimonides argues, is more properly conceived as bound up with choice among alternative possibilities. Only so can we argue that the world is as it is because God made it so. If there was never a time when the world did not exist with the precise nature we now observe, and if logic itself affords no room for alternative possible natures, then the notion of God's authorship is attenuated to mere concomitance. The world itself becomes necessary — ultimately frozen and changeless. And God's reality is collapsed into the world's eternity.69 For Ghazali the falasifa are atheists malgre lui. Maimonides balks at making that charge. He believes that an eternalist can be as sincere a theist as a creationist. But he thinks that creationism — rather than being the mere myth that Farabi and Ibn Sina had thought it to be, pictorializing emanation — harbors philosophic truths which the falasifa themselves lost sight of: God's transcendance of the world rather than "coexistence" with it as a kind of supernumerary correlative is perhaps the first of these truths. 70 But there are other values as well. Aided by Ghazali and by the seminal if unregimented insights of the kalam, and even by Ibn Sina, the Maimonidean philosophy identifies contingency in nature, empiricism in science, volition in God, historicity in creation and freedom in human choices as among the values that are bruised, slighted or overlooked by the eternalist logicism and intellectualism of the falasifa. Today we might add the concepts of emergence and the open future as items lying in the same slighted dimensions. 71

Maimonides does not believe that the ascription of a will to God can survive the demand, which is his demand as well, for absolute simplicity in a self-sufficient being. But neither does he accept the bias of the Aristotelian falasifa, who hold that wisdom can be identified with God's reality (dhat) in a way that will cannot. Both will and wisdom, as applied to God, are projections of our human notions; and, as such, both express our limitations. 72 Neither is privileged vis-a-vis the other. What we understand in nature we ascribe to wisdom, just as we call what seems conducive to our interests good.73 What we fail to understand we ascribe to chance or matter or satan or evil. But all of these terms name no real being; they merely refer to what has not yet been brought under the coverage of some concept. The Biblical idiom was to ascribe all events and acts to God. We recapture the force of that idiom when we recognize that what we cannot refer to God's wisdom we can still assign to God's will. In God, will and wisdom are in reality (of course) the same. But, since we are finite, it is inevitable that, from our perspective, there will be acts that we regard under the rubric of wisdom, and others (like the discrimination among equivalents that is necessary if finite actuality is to emerge from indeterminate virtuality) that we must assign to will. Even here, we can see that God's grace and goodness are at work. But why God's wisdom should achieve its ends by these particular means will inevitably elude our finitude. 74 The Divine will is not, of course, irrational. But neither can its expressions — the means the Law employs in pursuit of its rational ends, the images embedded in the pictorial imagination of the prophets, the seemingly patternless events that result from a superfluity of natural causes, the embodiment of man and the vulnerabilities it engenders, 75 or the primary differentiation of emanation, by which creation is disengaged from the One — be subsumed under the rubric of rationality in our comprehension. If finite beings are not self-determining and cannot in their own natures account for their own determinations, then, clearly, there are determinations in nature which can only be ascribed to the arbitrary determination of God, following not what God ascertains to be best but rather what He determines shall be or shall be done. Such positivity is amenable to reason, within measure, by reference to the larger purposes of God; for example, the selection of

114 L. E. Goodman the moment of creation, by reference to the larger preference for existence over non-existence, or the selection of the modalities of Israelite sacrificial worship, by reference to the larger Divine purpose of weaning humanity away from pagan worship and all its trappings, symbols and associated images. 76 But these large purposes underdetermine the event. They alone do not bring specificity to the brink of instantiation. They alone — the rational principles of grace, goodness and wisdom — do not explain why one individual imagination, language or genealogy should be chosen to bear the symbols by which common humanity are apprised of the Divine ideas accessible to all philosophers through reason and the Active Intellect. And they alone do not explain why another individual, imagination or genealogy should require, in history, to receive those images and symbols through the mediation of the chosen first expositors of an ideal — the poets, prophets and lawgivers whose words and images, symbols and institutions, most adequately and faithfully convey the ideal to the level of concrete action and expression. Such determinations, like the determination that the highest heaven will revolve from east to west, are arbitrary with reference to the purposes they serve, although necessary with reference to the attainment of those purposes in the weak sense that the rational goal could not be served without the arbitrary determination being made. Their recognition represents a qualification to the program of rationalism, the ceding of a place to the irrational. While such determinations are not construed as irrational, and still less as violent, they are conceived as representing a universe of values beyond our comprehension, not amenable to conceptualization by reference to the service of our ends. This qualification represents a recognition on the part of Maimonides, and of Ghazali before him, of the soundness of a criticism voiced by Plotinus against Aristotle: Plotinus agreed with Aristotle that intelligence is pure actuality. Accordingly, it must be perfect — indeed, it was to be equated with Platonic Being, the realm of forms. Yet for the same reason, intelligence cannot be regarded as the highest god; for, Plotinus argued, intelligence is not the highest of things. 77

Medieval Jewish and Islamic Philosophers

115

Notes 1. The Greeks and the Irrational (Berkeley: University of California Press, 1964; first printed in 1951 from the Sather Lectures of 1949). 2. See Fi al-Falsafa al-'Ula, trans, in A. Ivry, Al-Klndi's Metaphysics (Albany: SUNY Press, 1974), p. 74; Rasa'il al-Kindi al-Falsqfiyya, ed. M.A. Abu Rida (Cairo, 1950-1953), 1, p. 121. 3. Ivry, 72; Abu Rida, 1, p. 118. 4. See R.fi Ibana 'an Sujud al-Jirm al-'Aqsa in M.A. Abu Rida, 1, pp. 238-261. See also R. Walzer, Greek into Arabic (Oxford: Cassirer, 1963), pp. 196-199. 5. See Saadiah's Book of Theodicy, trans. L.E. Goodman (New Haven: Yale University Press, 1988), pp. 358-360. 6. R. Ji Dafi T'Akhzan (On How to Banish Sorrows), eds. H. Ritter and R. Walzer (Rome: Academia dei Lincei, 1938), pp. 3-64, i. 7. Sorrows, ii. 8. Sorrows, vi. 9. Following Philoponus, Kindi turns the Aristotelian arguments for the finitude of the cosmos to use in behalf of the finitude of time; see R. Ji ma'iyya ma la yumkinu 'an yakun la-nihaya lahu..., in Abu Rida, l,pp. 194-198. 10. Sorrows, vi. 11. Sorrows, vi. Kindi's expression for irrational is 'adm al-'aql, literally: mindlessness. 12. Book of Theodicy, p. 240. 13. For the life of appetites and passions, see K. al-Mukhtar fi 'l'Amanat wa-'l-'Itiqidat (cited below as ED), X, and my discussion in JAOS, 100(1978). For the absence of passions and appetites in the angels and God's knowledge of their faithfulness, see Book of Theodicy, pp. 155-157. For God's knowledge of his human servants' trustworthiness, 125; for the three roots of injustice, pp. 127, 359; cf. Epicurus Kyriai Doxai 1. 14. The parable was not invented by Kindi; even the sea shells are found in Epictetus, Enchiridion 7. 15. Sorrows, xi. Here Kindi's word for irrationality is jahl, literally: benightedness. 16. Cf. Enchiridion, 1. 17. See Timaeus, 27D and First Philosophy, Abu Rida, pp. 113-114, Ivry, pp. 67-68. 18. See Timaeus, 27C, 28A; First Philosophy, Abu Rida, pp. 120-122, Ivry, pp. 73-75. For Philoponus, see C. Wildberg, trans., Against Aristotle on the Eternity of the World (Ithaca: Cornell University Press, 1987); R. Sorabji. ed., Philoponus and the Rejection of Aristotelian Science (Ithaca: Cornell University Press, 1987); Seymour Feldman, "Philoponus on the Metaphysics of Creation," in R. Link-Salinger, ed., A Straight Path, Studies...in Honor of

116 L. E. Goodman Arthur Hyman (Washington: Catholic University of America Press, 1988), pp. 74-85. See also H. Davidson, Proofs for Eternity, Creation and the Existence of God in Medieval Islamic and Jewish Philosophy (New York: Oxford University Press, 1987). 19. Sorrows, xii. 20. See ED, X, and JAOS 100. 21. Book of Theodicy, pp. 126-127. 22. For trials as harbingers of bliss, see ED, V, 3; Book of Theodicy, pp. 125-126. For the light sown for the righteous and its dire effect on the wicked, see ED, IX, 5; Book of Theodicy, 353-354. For the need to rely on reason and God's unwillingness to vitiate our tests by informing us directly of their meaning, see ibid., pp. 382-383. 23. See my discussion in The Book of Theodicy, pp. 303, 362. For the incommensurability, see pp. 123-124. 24. See my "Razi's Myth of the Fall of the Soul," in G. Hourani, ed., Essays on Islamic Philosophy and Science (Albany: SUNY Press, 1975), pp. 25-10. 25. See my "The Epicurean Ethic of M. b. Zakariya' al-Razi," Studia Islamica, 34(1971), 5-26. Cf. Plato, Philebus. 26. See my "Razi's Psychology," Philosophical Forum, 4(1972), 26-18, and "Saadiah Gaon on the Human Condition," Jewish Quarterly Review, n.s. 67(1976), 23-29. 27. Metaphysics Zeta 7, 1032al2. 28. "Munazarat," in Paul Kraus, ed., Razi's Opera Philosophica (Cairo, 1939), pp. 289-316. 29. See Tibb al-Ruhani, trans. A.J. Arberry (London: Murray, 1950), pp. 103-107. 30. Genesis Rabbah, IX 7; cf. Ecclesiastes Rabbah, III 11.3. 31. See Mabadi' Ara' 'Ahl al-Madinatu 'l-Fadila, ed. and trans. R. Walzer (Oxford: Clarendon Press, 1985), pp. 89-135. 32. Ara', p. 145. 33. Ara', pp. 144-147. 34. Cf. Maimonides, Guide, I, Introduction, and III, 8, ed. Munk (Paris, 1856) 1, 8a, glossing Proverbs 7:6-21 and 6:26. 35. Ara', pp. 148-149. 36. Metaphysics, Alpha 4, 985a 21-28; cf. 985a 3-10. See also the fragment of Anaximander: "the source of coming to be for existing things is that into which destruction too happens 'according to necessity; for they pay penalty and retribution to each other for their injustice according to the assessment of Time'," G.S. Kirk and J.E. Raven, The Presocratic Philosophers (Cambridge: Cambridge University Press, 1985), p. 118. In view of the powerful impact of this line of thought on Aristotle, it seems strange for Walzer to say in his brief comments on this chapter that he has found "only one really good ancient parallel" to Farabi's application of the idea of deserts and entitlements here, and that in Simplicus' pedagogical introduction to his commentary on Aristotle's Physics. But Walzer does note that the issue of cosmic justice is crucial for Farabi. For the argument from the sciences, see Daniel Frank,

Medieval Jewish and Islamic Philosophers

117

The Arguments from the Sciences in Aristotle's Peri Ideon (New York: Peter Lang, 1984). 37. Ara\ pp. 276 11. 15-16. Cf. Book of Theodicy, pp. 192-193, 331-333; Maimonides, Guide, II, 20, III, 17-18. 38. See the texts translated in G. Hourani, "Ibn Sina on Necessary and Possible Existence," Philosophical Forum, 4(1973), 74-86; P. Morewedge, trans., Danish Namai Ala'i [named for Avicenna's patron 'Ala' al-Dawla], trans, in The Metaphysica of Avicenna (New York: Columbia University Press, 1973); A.J. Arberry, Avicenna on Theology (London: Murray, 1951). 39. Cf. the Qur'anic epithet of God as al-samad (112:2) and Plotinus, Enneads, III 6, "The Impassivity of the Unembodied." 40. De Anima, III 7, 431bl7; III 4, 430a4; 429al6; Aristotle, Metaphysics, Lambda 9. 41. Tahafut al-Tahqfut (Incoherence of the Incoherence; cited below as TT), ed. M. Bouyges (Beirut: Catholic Press, 1930), pp. 148-149. 42. See my "Ghazali's Argument from Creation," International Journal of Middle East Studies, 2(1971), 67-85, 168-88; "Did Ghazali Deny Causality?" Studia Islamica, 47(1978), 83-120. 43. See, for example, A.R. Badawi, ed., Prodi De Aeternitate Mundi in Neoplatonici apud Arabes {Islamica 19: Cairo, 1955); M. Mahdi, "Alfarabi Against Philoponus," Journal of Near Eastern Studies, 26(1967), 233-260; Shahrastani K. al-Milal wa-'l-Nihcd, ed. Cureton (London, 1846), p. 338. 44. Frag. 8, ap. Simplicius Phys, 145, 1, in Kirk and Raven, p. 249. 45. Metaphysics Lambda 3; cf. Physics VIII 1-2. 46. See Thomas Taylor, trans., The Fragments that Remain of the Lost Writings qfProclus (London, 1825) and note 43 above. 47. See, for example, al-Farabi, Ara' chaps. 1-9, and Aristotle, Metaphysics, Lambda 9; Nicomachaean Ethics, X 7. 48. TT 151 11. 5-6. Ibn Rushd appropriates the concept of ex nihilo creation in holding that there is nothing other than God on which all things depend; but he rejects the idea in its more familiar temporal sense, commonly expressed in Arabic by saying "God created the world after it was not." 49. TT, p. 152. 50. SeeTT, pp. 50-51. 51. The Case of the Animals vs. Man, trans. L.E. Goodman (Boston: Twayne, 1978), pp. 57-58. Cf. Ibn al-Munayyir, cited in E. Ormsby, Theodicy in Islamic Thought The Dispute Over al-Ghazali's Best of All Possible Worlds (Princeton: Princeton University Press, 1984), p. 164. 52. TT, p. 46. 53. TT, p. 47. Cf. the Hebrew liturgy: "Were a single one of these ducts to be closed..." 54. TT, pp. 50-51; cf. Ikhwan al-Safa', loc. cit.: "Do you not realize that a slight to the work is a slight to its maker?" 55. Phaedo,97C. 56. TT, p. 176,1. 10; De Caelo I 8.

118 L. E. Goodman 57. TT, pp. 176-177. Cf. Metaphysics, Lambda 8. 1074a 32-37. 58. TT, p. 177. 59. Leibniz "Observations on...Maimonides..." trans. L.E. Goodman, in "Maimonides and Leibniz," Journal of Jewish Studies, 31(1980), 234. 60. G. Hourani, Islamic Rationalism: The Ethics of 'Abd al-Jabbar (Oxford: Clarendon Press, 1970), pp. 15-16. 61. Al-Ash'ari K. al-Luma, ed. and trans. R.J. McCarthy (Beirut. Catholic Press, 1953), 53-103. 62. Cf. W. Montgomery Watt, The Formative Period of Islamic Thought (Edinburgh: Edinburgh University Press, 1973), pp. 25, 34, 85, 120, 123, 209, 213, 229-230. 63. Maqalat al-Islamiyyin, ed. H. Ritter, 2nd ed. (Wiesbaden: Steiner, 1963), pp. 309-311, 406-407. 64. Ibn Hazm, K. al-Fisal 5, 15, 11., 21-23; Tusi, K. Al-Muhassal 29; Ghazali, Tahafut al-Falasifa, ed. Bouyges, 2nd ed. (Beirut Catholic Press, 1962, cited below as TF), p. 199. Cf. Harry Wolfson, The Philosophy of the Kalam (Cambridge: Harvard University Press, 1976), pp. 546-555. 65. Maqasid al-Falasifa, ed. S. Dunya (Cairo: Dar al-Maarif, n.d.), pp. 147-154. 66. TF, pp. 200-205. 67. Guide, II, 17. 68. TF, pp. 59-62. 69. Guide, II, 21-22; cf. TF, pp. 70-73. 70. See note 42, above. 71. See my essay, "Three Meanings of the Idea of Creation," in D. Burrell and B. McGinn, eds., proceedings of the 1987 Notre Dame conference on God and Creation. 72. Guide, I, 52-60. 73. Guide, I, 2. 74. See my article, "Matter and Form as Attributes of God in Maimonides' Philosophy," in the Hyman Festschrift cited in note 18 above. 75. See Book of Theodicy, pp. 178-180, and Maimonides, Guide, III, 12, citing Galen. 76. Guide, III, 32. 77. Enneads, VI, 9.2.

Confucius, Mencius, Hume and Kant on Reason and Choice J. J. Kupperman

Herbert Fingarette has suggested that Confucius is interestingly and dramatically different from Western thinkers in that he omits the "whole complex of notions centering around 'choice' and 'responsibility'."1 This is one of a number of provocative and valuable points that Fingarette makes about Confucius; but, like the Analects itself, this statement is subject to interpretation. One might distinguish between the strong interpretation — Confucius does not think that the people he talks about make what we would term choices — and a more moderate interpretation — of course Confucius thinks that people make what we would term choices, but they are not or should not be (in some sense, to be explicated) "real" choices. In what follows I will discuss the concept of choice in order to explicate the difference between the strong interpretation (in which the claim about Confucius would be false) and the moderate one (in which it is true). I will then broaden the discussion to include a comparison of how Mencius, and then Hume and Kant treat choice, as opposed to Confucius' treatment. As part of this comparison, I will contrast the roles that reason seems to be assigned in choice by these four thinkers. Before beginning, I should like to make some remarks about how this investigation fits into the growing discipline of comparative philosophy. First, the student of comparative philosophy can expect to find some interesting similarities between Asian and Western philosophies, but must be cautious about them. For there are differences not only in substantive claims and emphasis, but also in the ways in which key terms must be interpreted. For example, it is often said that Confucian ethical philosophy contains no concept closely parallel to the Western

120 J. J. Kupperman concept of "morality" (as a special realm of emphatic normative judgments, in some accounts closely linked to generalizations), although there certainly are subtle and enlightening discussions in the Confucian canon of problems that we would term moral. In much the same way, we cannot count on philosophers such as Confucius and Mencius to have concepts that correspond closely to the traditional Western concepts of reason and choice. Thus step one toward a more sensitive discussion of Confucian and Western treatments of reason and choice is the understanding that we must not expect Confucian and Western philosophers to be answering exactly the same questions. Step two is to refrain from thinking of Eastern and Western philosophies as two tightly unified philosophical traditions. For the West, itself, does not subscribe to a single conceptual map. Thus the very differences between Confucian and Western ethics with regard to the concept of "morality" that have already been pointed out can also be argued to obtain between, say, Aristotle and Kant. And the differences between the conceptualizations of reason and choice in Hume and Kant may be as great as those between Hume and the Confucians. Nor does the East subscribe to a single conceptual map. We need not even assume that, because Mencius is a Confucian and in many ways identifies his position with that of the master, his conceptualization of what in ordinary English would be termed "choice" is identical with that of Confucius. I Here is a provisional definition of "choice." X has made a choice in situation S if and only if more than one course of action was possible for X and X decided among these possibilities. This is provisional in part, because it leaves open the question of the sense in which we can speak of a course of action as "possible" for an agent. If what is meant is "causally possible, given all antecedent conditions," and if determinism is correct, then it follows from the definition that there are no choices. We can postpone discussion of this counter-intuitive view for a while, because there is nothing to suggest that either Confucius or Mencius was a determinist; also, I will suggest an alternative reading of "possible." Another disturbing loose end is the word "decided." Must decision involve an articulation of alternatives along with some

Confucius, Mencius, Hume and Kant

121

process of weighing relevant factors or reasons for action? Or does it count as decision if an agent merely sees two possible paths to follow and immediately follows one? Those of us who are not wedded to an extremely intellectualized image of moral life will surely want to answer "Yes" to this latter question. The matter becomes less easy if we suppose that one of the two possible paths does not even occur to the agent, who immediately goes along the other, especially if there is no hesitation or break in his or her motion. But, then, has the agent in any sense decided or chosen anything? Those who are heavily influenced by the vision of human life found in such works as Sartre's Being and Nothingness may still want to answer "Yes"; the alternative is to concede that in many cases vital parts of people's lives are not decided upon or chosen by them, and it may be held to follow from this concession that people are not responsible for those parts of their lives. If it is true that important alternatives often do not register in people's thinking, we are left with a number of unattractive accounts from which to choose. Either we deny that people are responsible for that about which they have been totally thoughtless; or we say that they are responsible even though they have not decided that for which they are responsible; or we say that they have made decisions that they were (in any normal sense) unaware of having made. We can now return to the word "possible." Let me suggest that the sense appropriate to what we call "choice" is as follows: Y is possible for X in situation S if and only if a reasonable judge, knowing X's character, Xs perception of S, and all other knowable antecedent conditions, would not find it astounding if X did Y. This definition of "possible" is deliberately loose ("astounding" is hardly a precise term), and it fudges on determinism. But it does capture the sense in which it is not possible that various prominent scholars of Asian philosophy will run naked through a session of the next American Philosophical Association meeting, and in which their appearing clothed would not be taken to be the result of choice; one takes the disreputable alternative not to be "possible," whether or not one accepts determinism, and even if one mutters that, in the abstract, "anything is possible." It also captures the sense in which, even if determinism is true, it is possible either that Z will wear a blue tie at the A.P.A., and also possible that he will wear a red one instead.

122 J. J. Kupperman The point here is related to a phrase Fingarette uses more than once: "genuine option/alternative." (In William James's pioneering discussion, the phrases are "live hypothesis'* / "living option": a living option is one between two hypotheses, both of which are live.2) I shall use the phrase "live option" for anything that is a possibility in the sense just defined. As James's discussion suggests, even though a person might in some sense have any one of an indefinite number of religious beliefs, in practice it is usually the case that the live options are restricted to two: the belief the person was brought up with, and disbelief. Along similar lines, we might say that people's live options in moral matters are often rather restricted. How restricted they are is an interesting empirical question. One of the goals of the elementary moral education to which people are exposed at a very young age is to render certain kinds of behavior impossible. In this, moral education may build on natural sentiments. Whatever the respective contributions of natural sentiment and education may have been, it probably is impossible for any of us to torture a child to death, whatever the circumstances; one can say this without having to take a stand on the issues of determinism vs. free will. Beyond this very minimal decency, a great many other things may be impossible for a wellbrought-up person. The question may naturally arise: is anything possible, apart from the life the person actually leads or one roughly like it? It should be emphasized again that this is not the same as the metaphysical question of determinism, and that an anti-determinist can hold that, in matters of consequence, our live options are drastically restricted, while a determinist can hold that we often do (in the sense defined) have more than one live option. There is one further point that will be explored in what follows: It may be that the range of live options varies significantly from person to person, depending on the tightness of integration of different people's lives. It may be, that is, that some people have many live options apart from what they actually do or are, and that other people have essentially none. With this thought, let us turn to the Analects. As I have argued elsewhere, it is possible to view Confucius as an advocate of a higher naturalness, in which a person's own mind becomes a controlled work of art, and art becomes nature. 3 The life of such a person would be very highly integrated indeed. The scope of the integration is suggested by a passage in the Confucian Doctrine of

Confucius, Mencius, Hume and Kant

123

the Mean, which claims that "the superior man is watchful over himself, when he is alone."4 The degree of integration aimed for is suggested by the passage from the Songs, endorsed by the Master, As thing cut, as thing filed, As thing chiselled, as thing polished, and by the autobiographical report that, only at the age of seventy, could Confucius "follow the dictates of my own heart" without worrying about overstepping the boundaries of right. 5 Plainly, the process of integration should continue throughout a lifetime. The rewards in terms of inner satisfaction are considerable, and to a great extent render one invulnerable to luck or chance. Thus "A true gentleman is calm and at ease; the Small Man is fretful and ill at ease."6 He "that is really Good can never be unhappy." 7 'The Gentleman has nothing to fear."8 Thus the picture of the sage fits Fingarette's remarks very well. Such a man would seem to have no live options other than that of the Way (that is, of a sage-like existence), and perhaps, occasionally, when old desires flicker, that of abandoning the Way. The ideal of human life might therefore seem to be, at the extreme, a person who has no genuine choices. To present this as Confucius' picture of what human life is like would be to overstate the case in two respects. First, the person who has no genuine choices represents an ideal; furthermore, there are many people, who are not especially sage-like and who are far from this ideal. The Analects lay very heavy emphasis on the presentation of the ideal, as well as on descriptions of Confucius himself, who certainly came very close. But virtually no one has reached the ideal, for a number of reasons, one of them suggested by Confucius' remark that "I have never yet seen anyone whose desire to build up his moral power was as strong as sexual desire." 9 And, in any case, Confucius not only does not deny that there are other forms of human life besides that which is close to the ideal, but also talks about them. I will return to this point shortly. Second, even someone who is committed to following the Way, and whose psyche is chiselled and polished, can have moments when she or he does not know what is the best thing to do. It is in the light of this that we can understand the frequent references in the Analects to Confucius' desire to learn, and to his sense of his

124 J. J. Kupperman own fallibility. One of his merits is that he is "never obstinate, never egotistic."10 "If you have made a mistake," he says, " ... do not be afraid of admitting the fact and amending your ways."11 In reading these passages, one must bear in mind Confucius' repeated acknowledgement of the factor of skill in politics and morality. The advice to first enrich and then instruct the people is presumably meant for someone who, even with all the good will in the world, might not be clear about the priorities. 12 Similarly, such matters as "using the labour of the peasantry only at the proper times of year" are important in government, and it is imaginable that a decent but poorly informed person could go wrong about this. 13 At the very least, then, someone who is committed to following the Way might have to choose between doing X or Y because he or she does not know at this point which of them represents the Way. This, however, is a somewhat shallow rendering of what decisions of the kind under discussion amount to. Someone committed to following the Way, who must decide between X and Y without knowing which truly represents the Way, may well to some degree also be choosing her or his character; moreover, one of the alternatives may not be entirely virtuous. As Confucius saw it, part of the problem is that virtues can degenerate into vices; indeed, he probably would have agreed with La Rochefoucauld's contention that "Our virtues contain an element of vice, as medicaments are in part compounded of poisons. Prudence selects and tempers these ingredients. ... "14 A Confucian version of this Maxim would, however, substitute the word "learning" for "prudence." The problems that we have been discussing occur at a fairly high level, one that is not enormously far from the ideal. Many people who have some virtue are not entirely committed to following the Way, or may claim that they lack the necessary strength. J a n Ch'iu, to whom Confucius directed the advice "enrich, then instruct," was one such person. At one point, he said to Confucius, "It is not that your Way does not commend itself to me, but that it demands powers I do not possess." Confucius' reply was, "He whose strength gives out collapses during the course of the journey (the Way); but you deliberately draw the line."15 Jan Chi'iu's ultimate moral failure occurred when the head of the Chi family entrusted him with collecting revenues; rather than showing concern for the people, Jan Ch'iu extorted

Confucius, Mencius, Hume and Kant

125

more money. It was on this occasion that Confucius said, "He is no follower of mine ... . You may beat the drum and set upon him ... ,"16 My first comment is that Confucius' language here sounds like the language of responsibility: there is at least a suggestion that Jan Ch'iu made a choice, a bad one, and should be held responsible. Furthermore, there is nothing in the story of this unworthy disciple to suggest that he had only two options: the Way or depravity. As Confucius' reply to his excuse about lack of strength makes clear, even if J a n Ch'iu ultimately was not going to follow the Way, he could have tried harder and come closer. By and large the world of the Analects is polarized. At one extreme are Confucius, the ancient sage-kings and most of the disciples: either they follow the Way or come close by trying hard. At the other extreme are various contemporary miscreants, along with the great mass of the common people: they do not know the Way but can be helped toward a life that is in many respects good. Viewing these two groups, one may easily see the alternatives as following the Way or not following the Way. J a n Ch'iu, however, offers us the case of a man in the middle, whose alternatives have to be seen as considerably more complicated. Let me gather together the threads of the discussion thus far. We have been exploring a conception of choice in which someone has a choice if and only if she or he has more than one live option. Mental cultivation can diminish one's live options by rendering certain forms of conduct highly repugnant or unthinkable. What Fingarette has to say about Confucius is correct in this sense: Confucius thinks that people should not have any live option other than following the Way. This represents an ideal to which Confucius himself only approximates; but, beyond its biographical function, the Analects is dedicated to developing this ideal. When we move from Confucius to Mencius, the focus changes. Mencius also has something to say about the ideal, and much of this (as in his remarks about ch'i as a source of energy) goes well beyond what is in the Analects. But the major difference in relation to our present discussion is this: The Analects is largely devoid of what might be termed "existential moments," or points at which a choice is made that will determine not only matters of substance but also a person's character for the rest of his or her life. The closest to this that we come in the Analects is in relation

126 J. J. Kupperman to such behavior as Jan Ch'iu's in extorting money; and even here, it is suggested that what takes place is not a change of character but rather a continuation or working out of past unfortunate tendencies. In any case, we come on the scene not at the moment of choice but afterwards, in time to see a response. Mencius, in contrast, does give us existential moments. He suggests that a character can change drastically, that the potential for change is already present and that, with the right prodding from a great philosopher, someone can become a different person. The best-known existential moment in Mencius concerns the king who spared the sacrificial ox, 17 which shows the king's innate benevolence. Mencius suggests that this benevolence could be extended to the people: the king might also ensure that the people have enough to eat, and that education is provided in village schools. This famous passage really centers on two existential moments. First, there was the moment at which the king spared the ox, without fully understanding why; this points toward a second existential moment, which may or may not be realized, in which the king, having been aided in bringing his benevolence to clear self-consciousness, decides to extend this benevolence to his people systematically. He has the choice. None of this denies Confucius' view that the perfect sage has no choices. Nor does it represent the simple optimism about unalloyed human goodness that Mencius is sometimes, simplistically, represented as avowing. Instead we are led to realize that ordinary men and women who are not sages have within them a mixture of forces, including benevolence and selfishness, which give them the potential for a variety of types of action, both virtuous and blameworthy. This image of divided selves, who can be led to create their own unity by means of an appropriate focus, anticipates Pascal and the modern existentialists. The king can focus on his own innate benevolence and thereby become a different, and more successful, person. Mencius' argumentative strategy is repeatedly to get people to become aware of the benevolent element within themselves. Sometimes the manifestations of benevolence are slight and easily missed: for example, someone who is fond of music, whether sophisticated or popular, will want to share the experience with others. 18 One may have to set up an extreme and abstractly presented case as a thought experiment, for example the case of the

Confucius, Mencius, Hume and Kant

127

child about to fall into the well, in order to make clear that there is some degree of benevolence in everyone.19 It is as important to avoid a romanticist reading of Mencius as it is to avoid reading him as a simple-minded optimist. Mencius' view is not merely that, if we become clear about our benevolence and exercise it constantly, we will thereby behave well on all occasions. Like Confucius, he believes that proper awareness of the details of a situation is important in deciding what is appropriate to it; thus when he praises Confucius' flexibility, he is also proclaiming his own.20 Also like Confucius, he believes that benevolence needs to be complemented by technique; "Goodness alone is not sufficient for government ... ."21 What is essential, in the end, is acting out of benevolence; you "go into what you have learned in detail so that in the end you can return to the essential," but the details do matter. 22 Because the details matter, even someone who is entirely dominated by benevolence will have genuine choices to make. Again like Confucius, Mencius leaves room for two kinds of situation in which one might have more than one live option. One is the case in which an entirely benevolent person has to work out what the benevolent course of action is when there are two or more plausible alternatives. The second kind is the case in which someone who has not been entirely benevolent has a chance to behave benevolently. Mencius differs from Confucius in stressing the importance of this latter kind of case, and in arguing for the ongoing live possibility of benevolent behavior for any of us. As any commentator must be wary of possible misinterpretation or over-interpretation, a note of caution should be sounded here. Mencius never says in so many words that someone who has not been thoroughly good can change his or her character drastically. On the other hand, he urges people who fit that description (such as the king who spared the ox) to begin to behave in a thoroughly benevolent way. And this certainly does suggest that Mencius believed a drastic change of character is possible. Even so, however, we might wonder how quickly Mencius believed someone could change his or her character. Could a drastic change be sudden, or must it occur in stages? Hume suggests a very cautious response to this last question. The dominant view, in the Treatise of Human Nature and elsewhere, is that character does not (normally) change drastically,

128 J. J. Kupperman and that whatever changes are possible must occur gradually. Hume's view on these matters is so far removed from what people who have read only a little of his work tend to ascribe to him, that the background of his account of character change must be outlined. First, as Annette Baier has pointed out, Hume develops a much more positive account of self in Books 2 and 3 of the Treatise than in Book 1,23 At the root of this difference may be Hume's distinction "betwixt" personal identity, as regards our thought and imagination, and as it regards our passions or the concern we take in ourselves." 24 "The true idea of the human mind," Hume goes on to say, "is to consider it as a system of different perceptions or different existences, which are linked together by the relation of cause and effect... ." Memory does not so much produce as discover personal identity, and it does so by showing us the relation of cause and effect among our different perceptions. 25 The self, in the account that Hume begins to develop, is a "connected succession of perceptions," which is the object of both pride and humility. 26 The nature of a person's self is stable enough so that we can speak of her or his character, and this is the object of moral judgment; indeed, Hume says that people are not to be blamed for acts not caused by their characters or dispositions. 27 "If any action be either virtuous or vicious," according to Hume, 'tis only as a sign of some quality or character. It must depend upon durable principles of the mind, which extend over the whole conduct, and enter into the personal character. Actions themselves, not proceeding from any constant principle ... are never consider'd in morality.28 Normally when we blame someone, it is for something that we consider to be under his or her control. Thus, if Hume is right, if judgments of blame ultimately focus on character, it would seem to follow that character can be controlled, that someone can change his or her character if he or she wants to. But Hume denies this implication by denying a close connection between praise and blame on the one hand, and possible control of what is praised or blamed on the other. First, as Hume sees it, there is no sharp distinction between natural abilities and moral virtues; we praise and disparage the former as well as the latter. If it is said that natural abilities are involuntary and therefore have no merit, Hume's response is that many of the moral virtues "are

Confucius, Mencius, Hume and Kant

129

equally involuntary and necessary ... it being almost impossible for the mind to change its character in any considerable article ... ."29 The word "almost" in the last passage admits qualifications. These emerge in Hume's essay, "The Sceptic." This remarkable work centers on two topics: the objectivity of various kinds of knowledge, and also what nowadays would be termed "philosophy of life." In discussion Hume pays his debts to Cicero and the stoics, while developing an account of the neo-stoic virtue of gaiety. Hume suggests that one cannot change one's character by means of edicts addressed to oneself: "Mankind are almost entirely guided by constitution and temper, and ... general maxims have little influence."30 Indirect methods such as the formation of ideals can, however, have an effect. The prodigious effects of education may convince us, that the mind is not altogether stubborn and inflexible, but will admit of many alterations from its original make and structure. Let a man propose to himself the model of a character, which he approves: Let him be well acquainted with those particulars, in which his own character deviates from this model ... and I doubt not but, in time, he will find, in his temper, an alteration for the better. 31

'

Habit is, Hume suggests, "another powerful means of ... implanting ... good dispositions and inclinations." 32 For if one can impose appropriate habits upon oneself, changes of character will gradually follow. This adds up to a view in which one may have genuine choices in the long run that one does not have in the short run. Existential moments are as much missing from the moral universe that Hume presents as they are from the world of the Analects. Instead we are presented with the idea that a well-chosen regimen can gradually lead to a change of character. The concept of responsibility, though, remains in Hume's world, because in this world the connection with genuine choice that it is often assigned has been severed. The answers that Confucius, Mencius and Hume provide to the question, "How much choice does a person genuinely have?" (or that can be inferred from what they say) are all univocal but carefully qualified, and, because of the subtlety of the qualifications, difficult to grasp. In contrast, Kant at first appears to be

130 J. J. Kupperman bold and straightforward. My view is that this seeming straightforwardness is deceptive, and that there is more ambiguity and confusion in Kant's view than in those of the philosophers thus far considered. We can see this if we begin with a summary of the standard account of Kant's view of choice, and then go on to raise some questions about this account. What every properly trained philosophy graduate student knows about Kant's view of freedom is that he gives two answers to our question: (1) that a person never has any genuine choice, and (2) that, at least in moral matters, a person always has a genuine choice. Answer (1) holds if our purview is restricted to the phenomenal world, but consideration of the noumenal world leads us to answer (2). The reason for answer (1) is that one of the categories we impose on experience is that of cause and effect; and, indeed, if the world were a blooming, buzzing confusion of phenomena not governed by causal laws, experience and human intelligence would be impossible. Hence it is a synthetic a priori truth that every event has a cause. Thus determinism is true of every event in the phenomenal world, even of actions which would seem to involve moral choices. "If we could exhaustively investigate all the appearances of men's wills, there would not be found a single human action which we could not predict with certainty ... ,"33 Why, one might wonder, should "X causes Y" be taken to imply "Given X, Y is predictable with certainty"? Kant never gives a satisfactory explanation of this, but plunges ahead with determinism. Determinism, however, cannot be the last word about actions that we take to involve moral choice; the very existence of morality implies that there are at least some moral choices that are genuine ("Ought" implies "can"), and hence there are genuine alternatives to some of the things that people do. Thus we must look outside of the phenomenal world for a source of human freedom. That source is noumenal reason, which we are entitled to posit as the best and only explanation of the freedom which the morality of which we are aware presupposes. Hence we can tell two stories about actions that seem to involve moral choice: (1) that the determinism story is good psychology, and is true; but (2) that the freedom-grounded-in-noumenal-reason story is also true and represents a deeper (more ultimate) truth. This is a received interpretation of Kant, and it is also widely accepted that there are great difficulties in Kant's introduction of

Confucius, Mencius, Hume and Kant the noumenal world to account for the freedom that he thinks he must posit. First, we should mention a resolution of the problem of freedom that Kant entertains but does not accept: that a cause in the field of appearance of someone's behavior is not "so determining that it excludes a causality of our will — a causality which, independently of those natural causes, and even contrary to their force and influence," can produce empirically determinable behavior. 34 To resolve the problem thus would involve reason's "venturing beyond the limits of possible experience" and coming into conflict with itself,35 which would work only if appearances were things in themselves. 36 The correct resolution requires that we see freedom and determinism not as in conflict, but rather as true in different ways. Freedom applies to the intelligible character of a person, and determinism to the empirical character. Kant says this clearly, but he also says some things that sound different, remarking that how much of the moral character of a person's actions "is ascribable to the pure effect of freedom, how much to mere nature ... can never be determined ... "; on his own account this is to ask how two very different modes of explanation can be combined and weighed together.37 The intelligible character is outside of time: as intelligible, one is "free from all influence of sensibility and from all determination through appearances." But, inasmuch as this intelligible character is noumenon, "nothing happens in it; there can be no change ... ."38 Where in the realm of freedom, one might wonder, is the possibility of moral improvement, or the sheer fact of moral decision? The absence of any developed account of moral education from Kant's work is hardly an accident. Freedom, in Kant's view, requires actions that are in some sense determined by reason. Now reason has a phenomenal aspect: there are, after all, recognizable phenomena of people thinking and behaving rationally, and a person's rational activity can play as causal a role in her or his actions as any sensuous impulse. Our freedom, though, points beyond phenomenal rationality to what Kant calls reason "in its intelligible character." "The causality" of pure reason, he tells us, is outside of time. 3 9 Here a possible serious difficulty arises: many commentators have pointed out that cause and effect is a category and hence cannot be applicable to the noumenal world. Indeed, how can there be causality as we know it if there is no temporal order that can separate cause and effect? An alternative version

132 J. J. Kupperman of this line of criticism is that the word "cause" is ambiguous in the Kantian context: whatever it means in relation to phenomena it cannot mean in relation to noumena. Kant was sensitive to this, and returned to the point in the second Critique, arguing that, even though he had deduced the objective reality of the category of causality "only with regard to objects of possible experience," the seat of the concept was fixed in the pure understanding, "and it can be used even of noumena." 40 Kant, however, must go on to concede that this is so, "without our being able in the least to define the concept theoretically so as to produce knowledge."41 Thus he remains vulnerable to the charge that he is, in his own terms, in no position to know what he means when he speaks of the causality of intelligible reason. One other passage in the Critique of Practical Reason deserves mention here. Kant occasionally descends to discussion of how matters appear to common sense. He gives an example of someone who claims to be governed by uncontrollable lust: we ask him, if a gallows were erected and he were to be hanged immediately after the gratification of his lust, could he not indeed control it? The answer, Kant assures us, will surely be "Yes." We can also ask the same person If his sovereign ordered him, on pain of the same immediate execution, to bear false witness against an honourable man, whom the prince might wish to destroy under a plausible pretext, would he consider it possible in that case to overcome his love of life, however great it may be. He would perhaps not venture to affirm whether he would do so or not, but42he must unhesitatingly admit that it is possible to do so. I would like to suggest that this example rings true, although in my view it is for very different reasons from those Kant was prepared to supply. Much as the religious live options of most people include the religion in which they were brought up, along with unbelief, so it is arguable that, for virtually everyone, even people who might be considered depraved, a virtuous choice in a case like the one Kant describes will always be a live option. (The smallness of this claim should be noted: a single virtuous choice might be a live option for someone for whom a sustained virtuous way of life might not be possible.) If this is true, then it may be that one of Kant's starting points is the apprehension of a downto-earth fact about the possibilities of human choice, which fact

Confucius, Mencius, Hume and Kant

133

he then gilds and decorates with the heights of philosophical sophistication and confusion. What distinguishes Kant, indeed, from Confucius and Hume is that Kant's starting point lies not in a thoughtful awareness of how a person's possibilities are almost always more limited than they might seem to be, but rather in an ideal of scientific rigor that anticipates Laplace's determinism. I have already suggested that the presuppositions of coherent experience do not, as Kant supposes, require determinism: if every event has a cause in the weak sense of there being a set of antecedent conditions which, when grasped, would make it seem to a reasonable person much more likely and less surprising that the event would occur, this is more than enough to give us the sense of a minimally orderly world. To realize this is to see determinism as science fiction. Therefore there is no reason to prefer a determinist model of what is possible or not possible to the common-sense model Kant invokes in reporting the case of the man threatened with the gallows. Nor is there any reason to assume that the question, "What is possible for a person in the direction of her or his life?", must take some highly general form that is rooted in metaphysical doctrine. Underlying the subtle differences among what Confucius, Mencius and Hume have to offer in relation to this issue is the shared assumption that individual character matters to the answer, that the range of what is possible depends on what a woman or man has become. II To turn from the topic of choice to reason is to look at much the same textual data in a new way. We look again for situations in which a philosopher seems to think that people have a genuine choice to make. This time, though, we go on to look for recommended ways of arriving at a choice, or at what are held to be relevant considerations. This may seem to some to be a naive approach to our four philosophers. If one thinks of reason along the lines of what is operative in mathematics or in formal logic, it may seem that Confucius and Mencius have nothing to say about it, that Hume is in some respects dismissive of it, and that only Kant upholds the honor of rationality. Many philosophers nowadays would see things in this way, but this is surely a superficial view.

134 J. J. Kupperman For one thing, the account that Kant builds up of the role of reason in moral decision-making concerns something that cannot readily be assimilated to deductive or inductive inference. Unfortunately, Hume's view also turns out to be more complicated than many people would like to believe. It is true that he says "Tis not contrary to reason to prefer the destruction of the whole world to the scratching of my finger."43; but he also says what is "vulgarly call'd ... reason" is the calmer passions. 44 "By reason," Hume says, we mean affections that are like passions, but such as operate more calmly, and cause no disorder in the temper ... "; and it is clear that there is more than one important role in his ethics for something that he would be willing to call reason. 45 Once we abandon a rigid notion of what is to count as reason, we can see also that there is a role for reason in the ethical philosophies of Confucius and Mencius. I will suggest one similarity between Hume's and Mencius' views of rational sympathy. In my judgment, the comparisons that follow are useful, and help to shed light on Confucius and Mencius. But it should be emphasized now, especially, that these comparisons represent something like forced interrogation, in that the questions are by and large ones that Confucius and Mencius do not themselves ask. If one gets to know a person fairly well, one can imagine what she or he would be likely to say or do in some unfamiliar situation; and in much the same way, we can use things that Confucius and Mencius did say to reconstruct plausibly what they would have said, if our questions about reason were somehow translated not too inadequately for them. But it is important to remember that the thrust of inquiry was different for them from what it was for, say, Kant. For that matter, it would be wrong to see Hume and Kant as asking exactly the same questions. Among the many differences between them, there is one that stands out: Kant's approach to moral judgment is generally prospective, so that he wants to know how we can decide what we ought to do, while Hume's approach to the same question is generally retrospective, so that his questions center on how people have arrived at a shared vision of what they ought to do. In Kant's view, it is the form rather than the matter of maxims that is morally crucial: "It is the legislative form, then contained in the maxim, which can alone constitute a principle of determination of the [free] will."46 The bare form of the law can only be conceived by

Confucius, Mencius, Hume and Kant

135

reason ... ."47 What is grasped here by reason is not a pattern of deductive or inductive inference. It is rather the form that is required by a certain kind of thought. Just as aesthetic judgment, as Kant argues in the third Critique, requires disinterestedness, so also moral judgment carries with it certain special requirements, such as a kind of universalizability, respect for persons and the suitability to an ideal world of what is willed. Someone whose thought fails to meet any of these requirements (a failure which Kant seems to suggest entails a failure to meet all of them) has failed to think morally. In short, the, rationality that Kant sees as essential to moral reflection is the ability to enter into specifically moral forms of thought. This connects with other elements in Kant's philosophy: the first Critique, after all, centers on an examination of the forms of thought that rationality in general demands of us. Nor is Kant's view of morality entirely unprecedented. Robert Paul Wolff has pointed out that the synthetic a priori truths at the core of how we must think are prefigured in Hume as propensities of human thought as it moves from idea to idea. 48 Similarly, the a priori truths generated by entry into a specifically moral mode of thought are prefigured in Hume by the concept of the moral point of view. To use moral language, according to Hume, is to pass from merely personal appeals to others' sympathy or support to more general appeals, grounded in the way human beings ought to respond (can be expected to respond) to the case at hand. 49 One might be reminded also of Confucius' remark that "A gentleman can see a question from all sides without bias." 50 Of course, neither Hume nor Confucius is talking about a source of a priori truths; nor is either attempting to design a universal test of purportedly moral injunctions: this remained to be patented by Kant. But what Confucius and Hume (and, arguably, Mencius as well) share with Kant is the insight that there is a special impersonality that is, or should be, involved when one speaks or thinks of what is right or on the whole best. Hume is perhaps most sharply different from Kant in that he does not directly suggest that some feature of what is involved in entering this distinctive mode of thought and speech can be used in arriving at a judgment of what is right or best in a case at hand, although such a suggestion does not run counter to anything Hume says and I will argue that it is, in some sense, implicit in his view. Confucius' remark about the (presumably superior) gentleman's

136 J. J. Kupperman point of view does seem to me to make such a suggestion, and there is no reason to think that Mencius would not have endorsed it. Confucius' unbiased perception is an arrow that points in the same direction as that in which the machinery of Kant's categorical imperative moves; but this is a similarity in the midst of vast differences. What Confucius provides is in no sense moral machinery, dependent as it is on the character and temperament of the person who takes up Confucius' suggestion; and there is no hint of synthetic a priori truths. Nevertheless, there is something like "reason" operative in Confucius' ethics, if we are willing to construe the word "reason" so broadly that it is implicated in any positive answer to the question "Is there a good way to decide such matters as whether to increase taxes?" Certainly Confucius does not believe that an acceptable policy about taxation will come to a man if he merely closes his eyes and waits for an answer to pop into his head; nor does he believe that it will come in a surge of emotion. Everything in the Analects points to something like this: one should gather and weigh facts about the effects of taxation, and all of this should take place in the general context of a sympathy for the needs of the people. Someone influenced by the ethical philosophy of, say, Charles Stevenson might want to analyze this into (1) a basic attitude (sympathy for the needs of the people), and (2) the formation of beliefs about such things as the effects of taxation. But it is far from clear that the text of the Analects can sustain this separation, in that its attitude toward the suffering of the people is shaped and given substance by awareness of the facts, and in that a properly sympathetic attitude counts as what we would call a "reasonable" one. The process of cognitive transformation of an attitude is outlined with especial clarity by Mencius: Each of us has a basic attitude of sympathy, which is extended occasionally (to the ox about to be slaughtered, the child about to fall into the well). The process of becoming a good person does not require that we have a different attitude; rather it requires that the basic attitude be applied in a systematic and intelligent way. In many cases we fail to behave sympathetically, either because the basic attitude is blocked or overridden (say, by selfishness, or because we are preoccupied with special goals), or because we fail to realize a need for sympathy (perhaps because of insensitivity). Part of the process of becoming thoroughly virtuous is to generalize and free the

Confucius, Mencius, Hume and Kant

137

sympathy that is already in us. But we must also be intelligent in the way we express this sympathy: we should be sure that helpful actions really do help those they are intended to benefit, and we should not be over-eager, like the man who pulled out his plants to help them grow.51 Something like what is sometimes called "reason" is involved both in the ability to generalize sympathy and in the ability to apply it intelligently. As many commentators have noted, Hume does assign reason a role in his ethics. First of all, there is a role for the milder emotions, which (in a passage already quoted) he equates with what is commonly called "reason." Without these milder emotions we would be like the savage Scythians, whose ethical views can be dismissed because they have lost the sentiments of humanity. 52 Second, according to Hume, reason (in the sense of capacities involved in gathering, establishing relations among and assessing facts) does all of the preliminary work of ethical judgment. Even if we are inclined to approve of what is generally useful, a careful survey of the case may be required for us to know exactly what this is. 5 3 If the nature of moral discourse requires us to frame an appeal that will be valid from everyone's point of view, then it would seem to follow that everyone's well-being should be given equal weight. Thus, although Hume does not suggest a heuristic role for impartiality in moral thought, it is implicit in the conditions he outlines for acceptable moral judgments. A simplified modern view of reason and choice is as follows: Each of us encounters many situations in which there are two or more alternative things which we perfectly well could do. Reason — more or less the same intellectual processes that we use in inference or calculation — tells us which alternative is best. But none of our four thinkers fits this model. Kant comes closest, in that he provides an elaborate argument for saying that we often have genuine moral choices, although he also provides an argument for the opposite, along with enough ambiguity and obscurity to leave the issue permanently unclear. In any case, Kant does not have a view of reason in morality that matches that of the model. Neither Hume nor Confucius, nor Mencius, would accept this view of reason's perception of choice, and I have tried to indicate the degrees of their difference from the model and the reasons for their difference. We can attribute to Confucius, Mencius and Hume views of what might be translated as "reason" and "choice"; but Hume is careful to point out the

138 J. J. Kupperman ambiguity of the word "reason," and in the case of these three thinkers the translation or interpretation is inexact and the differences interesting.

Notes 1. Herbert Flngarette, Confucius — The Secular as Sacred (New York: Harper & Row, 1972), p. 18. 2. William James, "The Will to Believe," in Selected Papers on Philosophy (London: J.M. Dent, 1917). 3. J. Kupperman, "Confucius and the Problem of Naturalness," Philosophy East and West, 18(1968). 4. The Doctrine of the Mean, trans. James Legge, in The Chinese Classics, vol. 1 (New York: Hurst & Co., 1870), chap. I, 3, p. 124. 5. The Analects of Confucius, trans. Arthur Waley (New York: Vintage Books, n.d.), Book I, 15, p. 87; Book II, 4, p. 88. 6. Ibid., Book VII, 36, p. 131. 7. Ibid., Book DC, 28, p. 144. 8. Ibid., Book XII, 4, p. 163. 9. Ibid., BookK, 17, p. 142. 10. Ibid., Book IX, 4, p. 138. 11. Ibid., Book IX. 24, p. 143. 12. Ibid., BookXni, 9, p. 173. 13. Ibid., Book I, 5, p. 84. 14. Ibid., Book XVII, 8, pp. 211-212; The Maxims of the Due de la Rochefoucauld, trans. Constantine Fitz Gibbon (London: Allan Wtogate, 1957), Maxim 182, pp. 66-67. 15. The Analects, Book VI, 10, p. 118. 16. Ibid., Book XI, 16, pp. 156-157. 17. Mencius, trans. D.C. Lau (Harmondsworth: Penguin Books, 1976), Book I, Part A, 7, pp. 54-59. 18. Ibid., Book I, Part B, 1, p. 60. 19. Ibid., Book II, Part A, 6, pp. 82-83. 20. Ibid., Book II, Part A, 2, p. 79. 21. Ibid., Book IV, Part A, 1, p. 117. 22. Ibid., Book IV, Part B, 15, p. 130. 23. Annette Baier, "Hume on Heaps and Bundles," American Philosophical Quarterly, 16(1979). 24. A Treatise of Human Nature, ed. L.A. Selby-Bigge, 2nd ed. rev. P.H. Nidditch (Oxford: The Clarendon Press, 1978), Book I, Part IV, Section VI, p. 253. 25. Ibid., pp. 261-262. 26. Ibid., Book I, Part IV, Section VII, p. 277. 27. Ibid., Book II, Part III, Section II, p. 411. 28. Ibid., Book III, Part I, Section I, p. 575 (Hume's emphasis). 29. Ibid., Book III, Part III, Section IV, p. 608.

Confucius, Mencius, Hume and Kant

139

30. "The Sceptic," In Essays, ed. Eugene F. Miller (Indianapolis: Liberty Classics, 1985), p. 169. 31. Ibid., p. 170. 32. Ibid., pp. 171-172. 33. Critique of Pure Reason, trans. Norman Kemp Smith (New York: St. Martin's Press, 1965), A 550, B 578, p. 474. 34. Ibid., A 534, B 562, p. 465 (emphasis added). 35. Ibid., A 535, B 563, p. 465. 36. Ibid., A 535, B 563, p. 466. 37. Ibid., A 551, B 579, p. 475n. 38. Ibid., A 541, B 569, p. 469 (Kant's emphasis). 39. Ibid., A 551, B 579, p. 475 (emphasis added). 40. Critique of Practical Reason, in Kant's Critique of Practical Reason and Other Works on the Theory of Ethics, trans. Thomas Kingsmill Abbott (London: Longmans, Green and Co., 1898), 171172, p. 144 (Kant's emphasis). 41. Ibid. 42. Ibid., 141, pp. 118-119. 43. Treatise, Book II, Part III, Section III, p. 416. 44. Ibid., Book II, Part III, Section IV, p. 419. 45. Ibid., Book II, Part III, Section VIII, p. 437. 46. Critique of Practical Reason, 140, p. 117. 47. Ibid., 139, p. 116. 48. Robert Paul Wolff, "Hume's Theory of Mental Activity," in V.C. Chappell, ed., Hume (Garden City: Doubleday Anchor Books, 1966). 49. An Enquiry Concerning the Principles of Morals, ed. J.B. Schneewind (Indianapolis: Hackett, 1983), Section V, p. 49; Section K, p. 76n. 50. Analects, Book II, Chapter 14, p. 91. 51. Mencius, Book II, Part A, 2, p. 78. 52. Enquiry, Section VIII, pp. 64-65. 53. Treatise, Book III, Part III, Section I, pp. 689-5^0.

Rationalism and Anti-Rationalism in Pre-Buddhist China A. C. Graham

The most familiar texts of Chinese philosophy, the Analects of Confucius, Lao-tzu and the Book of Changes, contain nothing that a Westerner would recognize as rational demonstration; they may tempt us to see China as the Shangri-La of irrationalism. With a closer look at the literature of the classical period of Chinese philosophy (500-200 B.C.), one finds that the situation is more complicated. Confucius (551-479 B.C.) has no need of rational demonstrations because what he is offering in his wise sayings is an interpretation of traditional values that are selfevident to him and not yet open to question. But his first rival, Mo-tzu (late fifth century B.C.) — with his ten new doctrines, such as universal love, promotion of worth, rejection of fatalism, reward and punishment by the spirits — already has to defend his doctrines by giving reasons. Mo-tzu introduces the term pien "disputation" (literally: distinguishing, arguing to distinguish the right alternative from the wrong), and lays down three tests for judging a doctrine: the authority of the wisest in the past, observation by the eyes and ears of the multitude, and whether consequences are beneficial or harmful.1 In the core chapters of the corpus of his school [Mo-tzu, chaps. 8-37) all ten Mohist doctrines are defended at length in organized essays, appealing to the authority of the sages, to practical consequences and, in the case of factual issues (the existence of spirits and the non-existence of Destiny), to the evidence of common observation. With the appearance of new schools debate intensifies, with neater arguments and more precise definitions; this progress continues down to Hsiin-tzu and Han Fei in the third century B.C.2 In the philosophical mainstream analysis tends to be of questioned similarities and differences, as in the debates of Mencius and Kao-tzu,3 suggesting that thought is conceived primarily as a

142

A C. Graham

synthesizing process that goes wrong when it fails to assimilate and differentiate correctly; we shall return to this point in discussing 'correlative thinking'. 4 Toward the end of the fourth century B.C., however, the Sophists (Pien-che those who argue out alternatives) are offering paradoxical theses supported by logical demonstration. About 300 B.C. the Later Mohists undertake the enterprise of grounding the whole Mohist ethic in the analysis of moral concepts. This surely is rationalism as we find it in Greece, the plainest example in the Chinese tradition. But the Sophists have already provoked the reaction of the Taoist Chuang-tzu (c. 320 B.C.), who will have a much more lasting influence in Chinese thought. To use a terminology which we shall explain later, 5 Chuang-tzu's position is "anti-rationalism" (denial that reason is the right means to see things as they are) rather than "irrationalism" (which allows you to see things as you like). After 200 B.C. Chinese thinking channels in the orthodox Confucian direction (ethical, practical, conventional) and the unorthodox Taoist (spontaneous, mystical, disreputable). The former is often "rational," in that it checks its synthesizing by analysis, but not "rationalistic" in the sense of Later Mohist or Greek thought, which tries to detach rational demonstration wholly from common-sense synthesizing; the latter remains anti-rationalist as philosophical Taoism, and its continuation as Ch'an or Zen in Chinese Buddhism. As for true irrationalism, it seems to have no place in Chinese thought; it is never doubted that in the vision of the sage things show up as they are, as clearly as the detail of beard and eyebrows reflected in clear water. Let us look first at the brief episode of rationalism, which, since it failed to take root in China, has left only sparse and mutilated literary remains. Of the Sophists we have only paradoxes reported without their explanations by hostile witnesses, and the book Kung-sun Lung tzu ascribed to Kung-sun Lung (early third century B.C.), which has turned out to be a forgery from between A.D. 300 and 600. This forgery does, however, preserve early stories of Kung-sun Lung (chap. 1) and three genuine essays: the "White Horse" (chap. 2), "Pointings and Things" (chap. 3) and probably the "Left and Right" dialogue that introduces chap. 4, the rest of the book being banality or nonsense. 6 The most accessible essay is the "White Horse," which defends the thesis that "A white horse is not a horse." On the class/member analysis natural to a Western reader, this seems to start with a non sequitur

Rationalism and Anti-Rationalism in China

143

("By 'horse' we name the shape, by 'white' we name the color. To name the color is not to name the shape. Therefore I say a white horse is not a horse"), and to continue with a systematic confusion, of identity with class membership, which is ceasing to seem credible as our understanding of Chinese philosophy advances. It has only recently been noticed that Kung-sun Lung seems rather to be thinking in terms of whole/part, with "white horse" as the combination of "white" and "horse," and that the argument becomes fully intelligible as a demonstration that, since the whole is not one of its parts, the white horse is not the horse which is part of it.7 In any case it is clear that Kung-sun Lung's demonstrations, however interpreted, are purely analytic, excluding the analogical arguments so common throughout mainstream philosophical literature. For the Later Mohists we have a much more substantial literature, preserved in Mo-tzu chaps. 40-43 and 44, 45; it raises complicated textual problems which long deterred students of Chinese thought, but they have by now been sufficiently resolved for it to be usable as a source.8 1. The Canons (chaps. 40^11) and their Explanations (chaps. 42-43), divided between a series of 75 definitions and 12 analyses of ambiguous terms (A 1-87), and a series of propositions (A 88 B82). 2. The Big Pick (chap. 44) and the Little Pick (chap. 45), collections of fragments from two documents, including their titles: Expounding the Canons (12 more canons and explanations), and Names and Objects (a consecutive treatise). There must also have been another document defining the words in the formulations of the ten Mohist doctrines, since these crucially important terms are conspicuously absent from the 75 defined in the Canons.9 The organization of this collection has a considerable bearing on the nature of Later Mohist rationalism. The Canons are grouped in five sequences with common topics, not marked by titles, but running parallel throughout the two divisions, the definitions and the propositions. Both divisions have a central sequence on the problem of knowledge and change (A 40-51, B 13-16), followed by a sequence on the sciences — geometry, optics, mechanics and economics — (A 52-69, B 17-31) and preceded by a sequence on ethics (A 7- 39) and a gap between B 12 and B 13 per-

144

A. C. Graham

fectly fillable by Expounding the Canons. At first sight, the first and last sequences of both divisions seem to be about logic or proto-logic. The interesting point is that the arrangement of the five sequences implies that these are conceived as dealing with disciplines as unlike as either is to ethics or science. In the first sequence [Canons A 1-6, A 88-B 12), about the testing of similarities and differences, we are at the same level of rational discourse as in the philosophical mainstream. The independent treatise Names and Objects — on the same discipline — examines the effects of idiomatic variations on such formally parallel sentences as: 1. "White horses are horses, riding white horses is riding horses": 2. "Huo's parent is a person, but Huo serving her parent is not 'serving a person' [shihjen: 'serving a husband')." The Mohist criticizes a thesis that wrongly assumes parallelism with the first sentence: "Robbers are people, killing robbers [sha too: 'executing robbers') is killing people [shajen 'murder')." This implies that executing robbers is murder. The Mohist's answer is that the true parallelism is with the second sentence (and with a series of others with similar idiomatic shifts), so that one is entitled to say: "Robbers are people, but killing robbers is not killing people." Since this argument is about idiomatic sentences and not about propositions, it clearly does not belong to logic. Not that it is unsound: If someone objected, "Well, I prefer to class with the first sentence and say 'Killing robbers is killing people'," one would answer, "Certainly, but then it loses its pejorative force by shifting in meaning from 'executing robbers is murder' to the neutral 'killing robbers is killing persons'." The last sequence {Canons A 70-75, B 32-82) is about logical demonstration as we find it in Kung-sun Lung. There is no search for logical forms such as the syllogism, but the demonstrations — which multiply in the Explanations toward the end of the propo-

Rationalism and Anti-Rationalism in China

145

sition series — are strictly logical. Here, for example, is the refutation'of an objection to the Mohist doctrine of universal love Canon B 73: Their being limitless is not inconsistent with doing something to the total of them." Explanation (Objection): "The south if limited is exhaustible, if limitless is inexhaustible. If whether it is limited or limitless is not yet knowable, then whether it is exhaustible or not, whether men fill it or not, and whether men are exhaustible or not, are likewise not yet knowable, and it is fallacious to treat it as necessary that men can be exhaustively loved." (Answer): "If men do not fill the limitless, men are limited, and there is no difficulty about filling the limited. If they do fill the limitless, the limitless has been exhausted, and there is no difficulty about exhausting the limitless." This quite different discipline seems to be pien 'disputation' itself, a term defined in this section. The definition is in terms of a corrupt word which we might expect to mean "contradictory," but if rightly identified as/an, 1 0 a word found only in this document, it seems from other contexts to mean "converse." If so, the Mohist thinks of calling some objects "ox" as the converse of calling all the rest "non-ox." Canon A 74: "Disputation is contending over fan. To win in disputation is to fit." Explanation: "One calling it 'ox,' the other 'non-ox,' is contending over fan. Such being the case, they do not both fit; and if they do not both fit, necessarily one of them does not fit. (Not like fitting 'dog'.)" True pien must be over an 'ox or non-ox' issue, not — a point further established elsewhere in the Canons11 — over such an issue as "Is it a whelp or a dog?", of which both or neither might fit the object. In true pien one alternative fits 'necessarily' {pi). Inferences are regularly pronounced necessary in this discipline, never in the discipline treated in the first sequence. In an Explanation classifying types of connection we read: "If necessarily there is not one without the other, it is

146 A. C. Graham 'necessary.' What is from the sages, employ but do not treat as necessary. The 'necessary' admit and do not doubt."12 A central preoccupation of the Canons is to arrive at judgments unaffected by change. The middle sequence of definitions, concerned with the relation of knowledge and change (A 40-51), concludes with the definition of the "necessary" as the invulnerable to change. Canon A 51: "The necessary is the unending." Explanation: "It is said of cases where the complements are fully formed. Such cases as 'elder brother or younger brother' and 'something so in only one respect or something not so in only one respect' are the necessary and the unnecessary. 13 Being this or not this is necessary." Pi is used regularly of logical and also causal necessity, in disputation and in the sciences, the two disciplines that follow the bridging sequence on knowledge and change. The Later Mohists have fully detached the logical demonstrations of the Sophists from the criticism of assimilations and differentiations more characteristic of Chinese philosophy; the Sophists themselves, from whom so little remains, may well have done so earlier. Certainly the Later Mohists are 'rational,' but are they 'rationalists'? One no longer hesitates to accord them this title when one perceives that this school, which like all others in China is primarily concerned with philosophy of life, is trying to establish its basic ethical concepts as logically necessary. In this regard, a very striking feature of the Canons is a use of hsien 'beforehand,' which approaches our a priori. Thus a passing reference to the circle as 'known beforehand' 14 is explained when one notices scattered definitions interlocking in a system which, all measurements from the center being alike, builds up the definition of the circle from the term jo 'like,' which does not need to be defined because it is the basis of the whole theory of naming (i.e., that the same name is given to objects that are alike).15 We also meet a reference to "desiring beforehand": "Anything which the sage desires or dislikes beforehand on behalf of men, men necessarily learn about from him

Rationalism and Anti-Rationalism in China

147

by means of its ch'ing [close to the Aristotelian essence, but the essential not to being X but to being named 'X1.16] Anything from which desire and dislike are born in the conditions which they encounter, men do not necessarily learn about from him [the sage] by means of its ch'ing."17 This too becomes comprehensible when we notice another system of interlocking definitions beginning with the pair "desire" and "dislike." These do not have to be defined since the purpose is to show that benevolence and the right are what the wisest man desires 'beforehand,' but their different senses are distinguished in a section on ambiguous terms, including an irrelevant sense "about to" of the yil generally translatable as 'desire'. Canon A 84: "Desire. Immediate; having weighed up. Be about to. Dislike. Immediate; having weighed up." From these derive "benefit" and "harm" (but since the former, "desire," is not retrospective, it is replaced by "be pleased with"). Canon A 26: "Benefit is what one is pleased to have got." Canon A 27: "Harm is what one dislikes having got." The "on behalf of" of desire "on behalf of men" is defined in terms of the weighing of desires: Canon A 75: "To do on behalf of is to give the most weight in relation to the desires, in the light of all that one knows." The ethical definitions depend also on definitions of the words "love" and "total," in the combination "love of the total" (universal love), lost to us with the document defining words in the formulations of the ten Mohist doctrines. Judging by Mohist usage, the definition of "love" would have been something like "desiring benefit and disliking harm to, on his own behalf." For the pair "unit/total" [t'i/chien), used both of "individual/class" and "part/whole," we do have the definition of "unit":

148 A. C. Graham Canon A 2: "A unit is a part in a total." The central moral concepts, for Mohists a s for Coitfueispis, are benevolence (/en) and the right (t/i). Canon A 7: "Benevolence is love of units" — love of individuals, in contrast with 'love of the total,' the universal love which is not a moral virtue but a principle behind the virtues. Canon A 8: "The right is the beneficial." The system, which has further ramifications, is evidently designed to show that benevolence and the right are necessarily, by definition, what the sage desires 'beforehand' on behalf of men. We are even told that "If there were no men at all in the world, what our master Mo-tzu said would still stand." 18 The account of "desiring beforehand" recognized, however, that men can have no necessary knowledge of how to act in "the conditions they encounter," and continues: "The loving which involves benefiting is born of thinking. Yesterday's thinking is not today's thinking, yesterday's love of man is not today's love of man. The love of man involved in love of Huo is born from thinking about Huo's benefit, not from thinking about Tsang's benefit; but the love of man involved in loving Tsang is the love of man involved in loving Huo."19 For ethical choice in particular situations, Expounding the Canons therefore develops a procedure for weighing up benefits and harms, for which — as for the criticism of similarities and differences — no logical necessity is claimed. Since benefit to the whole is greater than benefit to the part, the principle is to prefer the 'total' to the 'unit,' sacrificing a finger to save the arm, the arm to save the individual, the individual to save mankind. A difference from Western Utilitarianism, which the system so strongly resembles, is that it starts not from the actual desires of all men, but from the desires of those who know the most ("Anything which the sage desires ..."). By the time of the Later Mohists, the logical acrobatics of the Sophists, always alien to Confucian and Legalist common sense,

Rationalism and Anti-Rationalism in China

149

had provoked a principled anti-rationalist reaction from the Taoist Chuang-tzu. We can trace a continuing debate over the value of reason, with Chuang-tzu. attacking the fourth-century Sophist Hui Shih and in turn becoming the target of the Later Mohists. 20 For Hui Shih, as for all the Sophists except Kung-sun Lung, we have little but sophisms — mockingly reported without their explanations — notably a list in the last chapter of the Chuang-tzu itself.21 They appear to be mostly spatio-temporal paradoxes, such as: "The south has no limit yet does have a limit," and "Simultaneously with being at noon the sun declines; simultaneously with being alive a thing dies." The list ends with the dictum "Love the myriad things indiscriminately, heaven and earth are one unit." Perhaps Hui Shih's purpose, like Zeno's, was to show that, since all division leads to contradiction, everything is one, and therefore we should love others as much as ourselves. 22 Although we lack Hui Shih's proofs, some of his conversations with Chuang-tzu show his style of argument, which is that of Kung-sun Lung and the Later Mohists. "Chuang-tzu and Hui Shih were strolling on the bridge above the Hao river. 'Out swim the minnows, so free and easy,' said Chuangtzu. That's how fish are happy.' 'You are not a fish. Whence do you know that the fish are happy?' 'You aren't me, whence do you know that I don't know the fish are happy?' 'We'll grant that not being you I don't know about you. You'll grant that you are not a fish, and that completes the case that you don't know the fish are happy.' *Let's go back to where we started. When you said 'Whence do you know that the fish are happy?', you asked me the question already knowing that I knew. I knew it from up above the Hao'."23 Hui Shih plays the logical game according to the rules; Chuang-tzu kicks over the board. But in his final stroke of wit he is not merely taking advantage of the accident that Hui Shih said "Whence (an) do you know..,?" instead of, for example, "By what means {ho-yi)...?n For Chuang-tzu, all knowing is from a standpoint, which is the whole concrete situation in which one stands.

150 A. C. Graham We learn from a story of Chuang-tzu passing Hui Shih's grave 24 that, much as he enjoyed making fun of the Sophist, he remembered Hui Shih as his only worthy opponent in debate. Indeed, we may even see him as drawing the full consequences of Hui Shih's presumed case, that one cannot make spatio-temporal distinctions without contradiction. It would take only one more step to recognize that all pien, as the arguing out of alternatives, starts from the drawing of distinctions, and so to abandon the proof that "heaven and earth are one unit" for immersion in the undifferentiating experience of things. Chuang-tzu. once takes up Hui Shih's observation that at the moment of death a thing is simultaneously alive, and draws the conclusion that both "It is alive" and "It is dead" may be simultaneously admissible. He is criticizing the distinguishing of fixed alternatives: 'this' for what fits one's naming, and 'that' for everything outside its scope. "Hence it is said 'That comes out from this, this too depends on that', the opinion that that and this come to life simultaneously. However, 'Simultaneously with being alive something dies', and simultaneously with dying it is alive: the admissible is simultaneously inadmissible, the inadmissible simultaneously admissible." 25 There is much subtle, elliptical argument showing through the poetry of Chuang-tzu, which is becoming more intelligible as we learn more of the technical terminology of disputation; Chuangtzu's complex sensibility includes a genuine taste for logic that is missing in later Taoists, including the later writers in the book Chuang-tzu itself.26 But the purpose is always to discredit pien, to show that by distinguishing alternatives, fixing them by names, arguing over which name fits and which course of action is right or wrong or beneficial or harmful, we become imprisoned in the viewpoint from which what fits our arbitrary choice of a name is "this," and thereby obscure our vision of the whole. For example, as soon as I lay down a distinction between waking and dreaming, it is impossible for me to know whether I am awake or dreaming. 27 Every statement that is admissible is also inadmissible, even "The myriad things and I are one" (Chuang-tzu's variation on Hui Shih's "Heaven and earth are one unit"), since as soon as I say it there are two: the one and my words about it. 28

Rationalism and Anti-Rationalism in China

151

The kind of language of which Chuang-tzu approves is that which spontaneously shifts viewpoints, makes only fluid distinctions, does not tie words to fixed meanings, guides in a direction without committing to any one statement — discourse which he compares to a type of vessel designed to tip over if filled to the brim 29 — that language, poetic rather than logical, in which he himself writes. Confucians and Mohists formulate in words a supposedly right way to behave and to order the Empire: the "Way {tad} of the ancients," or the "Way of the sage kings." For Chuang-tzu, on the contrary, the Way is the direction in which I find myself moving when submerged in the whole, when I recognize the fluidity of all distinctions (even between the direction and the whole), and simply mirror — from moment to moment, from my own viewpoint • and surrender to spontaneity: "The utmost man uses the heart like a mirror; he does not escort things as they go or welcome them as they come; he responds and does not store."30 The model for the sage is the craftsman, as in Chuang-tzu's story of a cook carving an ox, 31 and the many stories of his school about swimmers, anglers, painters and cicada-catchers. The craftsman does not analyze; he stills thought and emotion, concentrates attention on the object and responds, trusting to his knack. The crucial point is the total concentration of attention, as a hunchback catching cicadas on a sticky rod explains to Confucius: "I settle my body like a rooted tree-stump, I hold my arm like the branch of a withered tree; out of all the vastness of heaven and earth, the multitude of the myriad things, it is only the wings of a cicada that I know. I don't let my gaze wander or waver, I would not take all the myriad things in exchange for the wings of a cicada. How could I help but succeed?"32 Chuang-tzu's metaphor of the mirror is further developed io the writings of his school: "When the sage is still, it is not that he is still because he

152

A. C. Graham says "It is good to be still"; he is still because none of the myriad things is sufficient to disturb his heart. If water is still, its clarity lights up the hairs of beard and eyebrows, its evenness is plumb with the carpenter's level; the greatest of craftsmen take their standard from it. If mere water clarifies when it is still, how much more the stillness of the quintessential-and-daimonic, the heart of the sage! It is the reflector of heaven and earth, the mirror of the myriad things. Emptiness and stillness, calm and indifference, quiescence, doing nothing, are the even level of heaven and earth, the utmost reach of the Way and the Power; therefore emperor, King or sage finds rest in them. At rest he empties, emptying he is filled, and what fills him sorts itself out. Emptying he is still, in stillness he is moved, and when he moves he succeeds."33

A Westerner is tempted to assume that Taoists share his own 'reality/appearance* dichotomy, and credit the sage with an intuitive knowledge of a supra-sensible reality behind the perceived. Chuang-tzu does raise the question of how we know whether we are awake or dreaming, and once speaks of an "ultimate awakening" from the dream of life.34 For him, however, we are deluded not by the senses but by the dichotomies drawn by the intellect: life/death, waking/dream, something/nothing. Pre-Buddhist Chinese epistemology in general is "naive realism," and there is nothing to suggest that, even for Chuang-tzu, the world at the ultimate awakening ceases to be as it looks to the eye. The sage, like the craftsman, concentrates attention on the object, but he has no special source of information. Like everyone else, he thinks about things, makes momentary and relative differentiations, but does not trap himself in fixed distinctions and try to prove himself right. It is thinking about how to deal with things, instead of trusting to one's spontaneous reaction in full awareness of them, which Chuang-tzu forbids. Taoism — which recommends the stilling of the passions that disturb or distract attention, and the undeliberated reaction in perfect clarity of vision from one's momentary viewpoint — is quite unlike Western Romanticism, which has a similar preference for spontaneity but exalts the subjective vision in heightened emotion. It therefore seems inadequate to class Taoism as

Rationalism and Anti-Rationalism in China

153

'irrationalism.' It is different in kind from what we commonly recognize as irrationalism, the principled refusal to take account of facts that conflict with one's values or desires, as with the Nazi who does not merely ignore facts incompatible with the genuineness of the Protocols of the Elders ofZion, but scorns one's appeal to them as barren intellectualism. Even a rationalist will concede to Chuang-tzu. that spontaneity, as well as reason, has a function in adjusting intelligently to external conditions; that, for example, a tightrope walker cannot afford to pause and analyze his movements as he makes them (there was a New Yorker cartoon of a ski jumper seeing just ahead of him one of those office signs saying "Think!"). What the rationalist would deny is that the tightrope walker can be our model in dealing with the fundamental issues of life. I therefore prefer to class as 'anti-rationalism' doctrines such as Chuang-tzu's, which belittle the value of reason in adjusting to external conditions, and as 'irrationalism' those which entitle us to ignore those conditions. 35 Admittedly, in his shifting usage for "know," Chuang-tzu sometimes derides the knowledge of one verbally formulated alternative, and exalts ignorance; but he always has other words, such as ming, 'be clear about,' for the sort of awareness he prefers. Nothing could be more alien to him, or to the Chinese tradition in general, than Nietzsche's truly irrationalist question: "Granted we want truth, why not rather untruth? And uncertainty? Even ignorance?" 36 If rationalism is no more than a brief episode in the Chinese tradition, and anti-rationalism is limited to philosophical Taoism and its descendant Ch'an Buddhism, how shall we categorize mainstream thinking? So far, we have been getting along on such approximations as 'synthesizing' and 'common sense.' We may approach the question obliquely by considering the YinYang schematizing of Chinese cosmology, which we follow Needham in calling "correlative thinking." 37 This has often impressed Westerners as peculiarly Chinese, and the book by Granet which remains the best introduction to it is, in fact, called La pensee chlnolse.38 It organizes the cosmos by a scheme of correlated pairs, starting from Yang and Yin (heaven/earth, light/darkness, male/female, ruler/subject...), and branching into correlated sets of fours or fives, starting from the Wu hsing, or "Five Phases," 39 (Four Directions, Four Seasons, Five Colors, Five Notes...), and is the basis of court calendars and of pro to-

154

A C. Graham

sciences such as medicine, divination, musicology, alchemy and geomancy. In the chain of pairs A and B — and, correspondingly, in the larger sets — Al compares with Bl as A2 with B2, and Al connects with A2 as Bl with B2. Thus in the course of expounding its cosmogony Huai-nan-tzu (c. 140 B.C.) explains the ignition of tinder by the concave mirror (the "Yang sui") by contrasting it with a parallel phenomenon, the dew collecting at night on the "square chu," said to be another kind of mirror: "Therefore, when the Yang sui sees the sun, it ignites and makes fire, and when the square chu sees the moon, it moistens and makes water." *° Examining the whole scheme, we see that Yang compares with Yin as sun with moon, round with square and fire with water; therefore the connections between Yang sui, sun and fire compare similarly with those between square chu, moon and water. To a modern such thinking may be more alien than anything in Laotzu or the "White Horse." It is not merely that the explanation is obscure or fallacious; for us it is not an explanation at all. But far from being peculiarly Chinese this mode of thinking is common to proto-science everywhere, including Europe (with its Four Elements, Four Humors and Pythagorean numerology) right down to the Renaissance. Moreover, the philosophers from Confucius to Han Fei (d. 233 B.C.) never engage in this schematizing; except for Hsiin-tzu in the third century B.C. they do not mention even the Book of Changes, which even Hsiin-tzu does not yet include in the Confucian Classics.41 The entry of correlative schematizing into the philosophical literature is at the final eclectic stage of the classical period, in the syncretistic philosophical encyclopedia Lil-shih ch'unch'iu.42 Previously we meet it only in historical and other nonphilosophical sources, as the lore of physicians, music masters and diviners. The philosophical schools conspicuously neglect the sciences, with the single exception of the Later Mohists, who ignore the Yin and Yang, mention the Five Phases only to say "The Five Phases have no constant ascendancies," 43 and offer only the purely causal explanations for which they claim pi "necessity." We may contrast the Huai-nan-tzu explanation of the burning mirror with an account in the Canons of another peculiarity of the concave mirror: the inversion of the image if you

Rationalism and Anti-Rationalism in China

155

stand outside the center of curvature (which the Mohist fails to distinguish from the conjugal focus): Canon B 23: "When the mirror is concave the shadow is at one time smaller and inverted, at another time larger and upright. Explained by: outside or inside the center." Explanation: "Inside the centre. If the man looking at himself is near the centre, everything mirrored is larger and the shadow too is larger; if he is far from the centre, everything mirrored is smaller and the shadow too is smaller; and it is necessarily upright. This is because the light opens out from the centre, skirts the upright object and prolongs its straight course. Outside the centre. If the man looking at himself is near the centre, everything mirrored is larger and the shadow too is larger; if he is far from the centre everything mirrored is smaller and the shadow too is smaller; but it is necessarily inverted. This is because the light converges at the centre, ... and prolongs its straight course." In the West we find the same type of contrast between philosophy and proto-science, but much sharper. On the one hand, syllogistic logic (absent in China except for a brief flowering in the seventh century A.D. of the Buddhist logic imported from India)44 goes back to Aristotle; on the other, nearly two thousand years later, Kepler, discoverer of the first three modern laws of nature, is still trying to organize his cosmos by correlating the distances between the planets with the five regular solids, and sun, stars and planets with the three persons of the Trinity. But although correlative cosmos-building, whether Chinese or Western, belongs to the proto-sciences, it would seem that the correlative thinking of which it is an exotic offshoot is basic to all thinking and to the operation of language itself. This already follows from Roman Jakobson's description of words as related "paradigmatically" (on the dimension "similarity/contrast") as members of sets from which we select them, "syntagmatically" (on the dimension "contiguity/remoteness") as elements in the phrases and sentences in which we combine them. 45 In English such compound words as "daylight," and such formulae as "black as night," "the darkness of ignorance / of evil," imply that, before forming sentences, we are already linking paradigms syntagmat-

156

A. C. Graham

ically in chains of oppositions which at their simplest are binary. < Paradigm A

1

1

1. Day 2. Light 3. White 4. Knowledge 5. Good

B

Night Darkness Black Ignorance Evil

(The numbering is of course only for convenience) Then before the analytic thinking that starts from the sentence, we are already organizing the world around us by comparing and connecting words as in the Yin-Yang scheme: Day : night :: light : darkness (Day compares with night as light with darkness; day connects with light as night with darkness.) We begin to analyze when doubt arises as to a correlation, such as "White man : black man :: good : evil"; we demand a causal connection and fail to find one between moral qualities and color of skin. On this account, it becomes easier to understand why, in both China and the West, the correlating that constructs our first skeleton of a cosmos is extended throughout cosmology and the proto-sciences. In both civilizations there has always been plenty of causal thinking, and whole episodes, such as the Later Mohist, when its superiority to correlative schematizing is clearly seen; but piecemeal causal explanations do not add up to a cosmology. Until the Scientific Revolution in seventeenth-century Europe discovered an alternative — the mathematization of laws of nature tested by controlled experiment — the choice was between a cosmos organized by correlation and no cosmos at all. That much, or most, of the thinking of ancient China is correlation-guided and tested by analysis — as is our own as long as it remains on the common-sense level — is clearly exhibited by the strong tendency to parallelism in Chinese style, apparent even in

Rationalism and Anti-Rationalism in China

157

the Mohist account of the concave mirror. We noticed that the Later Mohists distinguish from logical demonstration a whole discipline devoted to analyzing similarities and differences, in particular to the testing of formal parallelism. As characteristic of ordinary Chinese argument, we might take the following example from the Legalist Book of Lord Shang (c. 240 B.C.): "T'ang and Wu rose to kingship without following antiquity, Yin and Hsia fell to ruin without change of conventions. Consequently, rejection of antiquity is not necessarily to be condemned, and conformity to convention does not deserve to be made much of." ^ Here the parallelism reflects a chain of oppositions in which the ratio "Rise to kingship : rejection of antiquity :: fall to ruin : conformity to convention" transfers approval and disapproval from one contrasted pair to the other. But the argument has none of the artificiality for us of the Huai-nan-tzu explanation of the burning mirror, in the first place because the connections between the pairs are acceptable as causal, as they tend to be in the political and other practical situations that discipline correlation by quickly fulfilling or disappointing expectations. Moreover, the inference has been tightened by analysis; as only two cases of rise to kingship following rejection of antiquity have been given, the conclusion is not that the rejection is good, but that it "is not necessarily to be condemned." Most Chinese thinkers, however much they may reason, do not share the Later Mohist and Western hope of establishing an independent realm of reason by a perfect system of definitions of the concepts from which it starts, leaving argument from analogy as loose thinking outside its borders. This, however, has nothing to do with being rationalists or anti-rationalists, for much modern philosophy in effect abandons this hope except in logic and mathematics — Ryle's undermining of arguments by exposing category mistakes, Kuhn's insistence that all scientific hypotheses are paradigm-dependent, Derrida's attempt to deconstruct the chain of oppositions underlying the logocentric tradition of the West. Thus Ryle sees the mind/body problem as arising from a "para-political myth" (in terms of the correlative ratios we have been using, "Mind : head, hands, feet :: ruler : subjects"), transformed by the ascendancy of modern science into a "para-

158

A. C. Graham

mechanical myth" ("Mind : head, hands, feet :: governor engine : the other engines"). He claims to dissolve the problem by proposing more adequate correlations: "Mind : head, hands, feet:: university : colleges, libraries, playing fields," "Mind : head, hands, feet :: the British constitutions : Parliament, the judiciary, the Church of England." 47 The implication is that any philosophical argument, however impeccably developed, is vulnerable to the exposure of a correlation at its foundations, with the further implication that, although analysis can criticize correlation, it always remains dependent upon it. In a sense which we need no longer treat as pejorative, the correlative undercurrent of thought is 'pre-logical,' although it does require logic to test its adequacy for problem-solving. An old correlation is fixed by habit, a new one appears in an unanticipated flash of insight. Kuhn describes the paradigm-shift as a "relatively sudden and unstructured event like the gestalt switch," after which scientists "often speak of 'the scales falling from the eyes' or of the 'lightning flash' that 'inundates' a previously obscure problem, enabling its components to be seen in a new way that for the first time permits its solution"; there are even times when "the relevant illumination comes in sleep."48 That it is acceptable to speak of the creative thinking even of scientists in this way may help us to understand why a Chinese thought in which we can trace no logical steps does not — as our own "appearance/reality" dichotomy tempts us to suppose — assume some mystical access to reality beyond the reach of reason. What it does imply is that there are opposite ways of correcting correlative thinking, by rational criticism and by unseating conventional oppositions to restore its fluidity, both of which are employed in the Chinese tradition. Granted that fixed chains of oppositions imprison thought, as is very obvious in the case of Yin-Yang schematism, how does one loosen them? We see from Derrida, as well as from Chuang-tzu and Lao-tzu, that language that deconstructs oppositions has to take the direction of poetry. The rationalist, who recognizes his own kind only in the Sophists and Later Mohists, is not bound to exclude even Lao-tzu and the Book of Changes — texts most alien to him — from the domain of philosophy. Lao-tzu is a poem that sets out to break the habit of thinking in dichotomies — (something/nothing, knowledge/ignorance, male/female, above/below, before/behind). These differ from the oppositions

Rationalism and Anti-Rationalism in China

159

(reality/appearance, nature/culture, life/death, good/evil) that Derrida finds at the back of Western logocentrism, in that the latter tradition strives to abolish B in favor of A; as it has long been a commonplace, China tends to complementary Yin-Yang polarities and the West to conflicting opposites. 49 However, China does regularly assume the relative superiority of A to B. The reversals by which Lao-tzu presents B as superior to A are strikingly similar to the move by which Derrida reverses, for example, our habitual preference for speech over writing.50 It is not that Lao-tzu wishes us to prefer nothing to something or female to male; his move, like Derrida's, is a step toward undermining the oppositions. The divination system of the Book of Changes, for a rationalist surely the most repellent book in early Chinese literature, may likewise be seen as a technique for recovering the fluidity of correlative thinking. The 64 hexagrams, with multitudinous correlations for each line, trigram and hexagram as a whole, seem at first sight to present a cosmology much more complicated, artificial and confused than the neat standard system with Five Phases corresponding to Four Seasons, Five Colors and so forth. Since the two systems share little except the binary distinction of Yin and Yang, from which they start, one may wonder how the Chinese succeed in reconciling them. Even Needham, charitable as he is to Chinese proto-science, treats the Book of Changes as a disaster in its history. 51 Insofar as the system of the Changes was allowed to intrude into the explanation of the more coherent system, one may agree with him. But the Five Phase system is designed for the proto-sciences and for action in regularly recurring situations, such as the changes of ritual throughout the Four Seasons, while the Book of Changes is intended for divination in particular situations (those in which artificial classifications are most restrictive and free correlation is indispensable), with the 64 hexagrams representing every mathematically possible sequence of six successive choices between two alternatives. Western enthusiasts for the Wilhelm-Baynes / Ching or Book of Changes generally assume that the divination could not be successful unless there really is an a-causal principle in nature (Jung's "synchronicity"), by which the fall of the yarrow sticks that selects the Yin or Yang lines responds to the pattern of the diviner's situation. 52 But it is possible to argue that these enthusiasts may not be misguided even though the fall of the sticks, or

160

A. C. Graham

of the coins that replace them, is the result of pure chance. If the hexagrams thrown up by chance carried unambiguous instructions they would of course, except by a lucky accident, be grossly misleading; but since they offer only enigmatic auspices and an almost unlimited latitude to correlate with a variety of images, the diviner's interpretation in the light of his personal situation becomes a meditation on his circumstances opened up to new perspectives. The effect will be to break down preconceptions and give binary thinking a fresh start. To relate Chinese divination to creative thinking in the sciences may seem fanciful. But in fact the Great Appendix of the Book of Changes includes the invention of tools beside divination among the purposes of the system 53 and also presents a legendary history of the major inventions, identifying the hexagrams which inspired them: "They hollowed out wood to make boats, shaved wood to make oars. The advantage of boat and oars was to cross to the inaccessible and deliver over distances to the advantage of the whole world. Evidently they took it from Huan "54

So the idea of the boat first came to an ancient sage contemplating the trigram correlatable, among many other things, with wood ( ), on top of the trigram correlatable with water ( • ), with its attached auspice "Advantageous for crossing a great river." If one reads this not as fabulous history but as an illustration of how an original thought comes into one's head, it does not look silly at all. It is remarkable that no account of the standard cosmological correlations ever achieved canonical status in China, while the Changes were eventually admitted to the Confucian classics and Lao-tzu became the most influential of all the unorthodox books. It is as though Chinese civilization has been careful to preserve a certain latitude in the organization of its cosmos, in order that throughout its long history originality and creativity should never die out. Perhaps we have here an answer to a question that Needham's great history of Chinese science helps us to put more

Rationalism and Anti-Rationalism in China

161

clearly: How is it that China, which never came within sight of our seventeenth-century Scientific Revolution, has been so extraordinarily fertile in basic inventions, until late in the Middle Ages perhaps the most fertile of all? As for the negative question in which most people seem more interested, of why the Scientific Revolution never happened in China, this unique and complex event depended on a variety of social and other conditions centered on the combining of Greek logic and geometry with Indian numerals and algebra; and since, for primarily geographical reasons, these converged in the Arab world, from which they passed to Latin Christendom, it is surely pointless to go into possible reasons why the 'discovery of how to discover' did not happen somewhere else. We generally ask why an event did happen, not why the same complex set of conditions did not come together at some other time or place.55 We sum up with some highly debatable generalizations: 1. The mainstream of Chinese thinking is correlative, on the one hand guided and tested by analysis as in our own common-sense thinking, on the other restored to fluidity by the poetry of Lao-tzu and diagrams of the Book of Changes. 2. There is a brief episode of rationalism, in which the Sophists and Later Mohists exalt analytic reason and detach it from correlation, and also a minority tradition of anti-rationalism, as found in philosophical Taoism and Ch'an Buddhism, which prefers spontaneity to reason in the conduct of life. 'Irrationalism,' as the principled refusal to acknowledge objective fact, does not seem to be represented in the tradition. 3. As in the pre-modern West, correlative cosmology is an extension of correlative thinking to the proto-sciences, in the absence of the mathematized and experimentally testable laws which, in the West, were finally to provide an alternative means of organizing knowledge of nature.

Notes L The three tests appear in various forms in the introductions of the three anti-fatalist chapters, Mo-tzu chaps. 35-37, trans. Y.P. Mei, The Ethical and Political Works of Motse (London: Probsthain, 1929), pp. 183, 189, 194 : chap. 35, trans. B. Watson, Mo Tzu: Basic Writings (New York: Columbia University Press, 1963), p. 118.

162

A C . Graham

2. See, in general, A.C. Graham, Disputers of the Tao: Philosophical Argument in Ancient China (Peru, Illinois: Open Court, 1989). 3. Mencius 6A:l-5. The analogical argument in Mencius is analyzed in D.C. Lau, Mencius (Harmondsworth: Penguin, 1979), pp. 235-263. 4. Cf. pp. 17 ff., below. 5. Cf. pp. 16 f., below. 6. A.C. Graham, Studies in Chinese Philosophy and Philosophical Literature (Singapore: Institute of East Asian Philosophies, National University of Singapore, 1986), pp. 126-166. 7. Chad Hansen, Language and Logic in Ancient China (Ann Arbor: University of Michigan Press, 1983), pp. 140-171. Hansen's detailed application of his proposal to the "White Horse" is very complicated. I have suggested a much simpler analysis of the argument starting from Hansen's premises in ibid, (note 6), pp. 196-210. 8. A.C. Graham, Later Mohist Logic, Ethics and Science (London-Hong Kong: SOAS-Chinese University Press, 1978), pp. 73-238. The numbering of the Canons here follows this work, in which Mo-tzu, chaps. 40-45 are translated in full. 9. Ibid., pp. 235 f. 10. Ibid., pp. 184 f. 11. CanonB 35. 12. Canon A 83. 13. I formerly thought this understanding of the sentence syntactically unnatural, and emended the text in Graham, ibid, (note 8), p. 299; but I have since noticed the same syntax in Expounding the Canons 8, as edited in ibid., p. 252. 14. Canon A 93. 15. Graham, ibid, (note 8), pp. 56-58. 16. For ch'ing, cf. ibid., pp. 179-182. 17. Expounding the Canons 1. 18. Loc. cit. 19. Loc. cit. 20. For dialogues between Hui Shih and Chuang-tzu, cf. A.C. Graham, Chuang-tzu: The Inner Chapters, and other writings from the Book "Chuang-tzu" (London: Allen and Unwin, 1981), pp. 46 f., 82 f., 100102, 122-124. Chuang-tzu's anti-rationalist theses are criticized in Canons B 35, 48, 68, 71-72, 79, 82. 21. Chuang-tzu, chap. 33, trans, in Graham, ibid, (note 20), pp. 283- 285; trans. B. Watson, The Complete Works of Chuang Tzu (New York: Columbia University Press, 1968), pp. 374-377. 22. The argument of Reding that the theses of Hui Shih are merely remarks about government and other common-sense themes, quoted out of context, is a reminder of how little we have to go on in interpreting them. I think, however, that there is enough evidence (as in the Mohist reference to the south as limited or limitless, and Chuang-tzti's to being both alive and dead at the moment of death, quoted on p. 6, above, and p. 13, below) to assure us of the

Rationalism and Anti-Rationalism in China

163

orthodox reading of the theses as spatio-temporal paradoxes. See J.P. Reding, Les fondements phUosophiques de la rhetorique chez les sophistes chinois (Berne: Peter Lang, 1985). 23. Chuang-tzu, chap. 17, Graham, p. 123; Watson, pp. 188 f. 24. Chuang-tzu, chap. 24, Graham, p. 124; Watson, p. 269. 25. Chuang-tzu, chap. 2, Graham, p. 52; Watson, p. 39 f. 26. For the composition of Chuang-tzii, in which only the seven Inner Chapters and some passages from the Mixed Chapters are likely to be the writing of Chuang-tzu himself, cf. Graham, ibid, (note 20), pp. 27-29, and ibid, (note 6), pp. 283-321. 27. Chuang-tzii, chap. 2, Graham, p. 61; Watson, p. 49. 28. Chuang-tzu, chap. 2, Graham, pp. 56 f.; Watson, p. 43. 29. Chuang-tzu, chap. 27, Graham, pp. 106 f.; Watson, pp. 303-305. 30. Chuang-tzu, chap. 7, Graham, p. 98; Watson, p. 97. 31. Chuang-tzu, chap. 3, Graham, pp. 63 f.; Watson, pp. 50 f. 32. Chuang-tzii, chap. 19, Graham, p. 138; Watson, p. 200. 33. Chuang-tzu, chap. 13, Graham, p. 259; Watson, p. 42. 34. Chuang-tzu, chap. 2, Graham, pp. 59 f.; Watson, pp. 47 f. 35. 1 have developed this distinction between anti-rationalism and irrationalism in Reason and Spontaneity (London: Curzon Press, 1985), pp. 156-227, taking Sade, Nietzsche (in one of his aspects) and Hitler as exemplars of irrationalism, and Chuang-tzu, three modernisms (Futurism, Dada and Surrealism) and Bataille as exemplars of anti-rationalism. 36. Friedrich Nietzsche, Beyond Good and Euil, trans. R.J. Hollingsworth (Harmondsworth: Penguin, 1983), p. 15. 37. J. Needham, Science and Civilization in China, vol. 2 (Cambridge: Cambridge University Press, 1954), pp. 279-363. 38. Marcel Granet, La pensee chinotse (Paris: Albin Michel, 1934). 39. "Five Phases" is coming to replace, as equivalent for Wu Hsing, the "Five Elements" still used by Needham; see J. Major, "A Note on the Translation of Two Technical Terms in Chinese Science," Early China, 2(1976) 1-3. For pre-Han usages I prefer "Five Processes"; see A.C. Graham, Yin-Yang and the Nature of Correlative Thinking (Singapore: Institute of East Asian Philosophies, 1986), pp. 74-76. 40. Huai-nan-tzu, chap. 3, trans, in ibid., pp. 30-32. (There is not yet a complete translation of Huai-nan-tzu.) 41. Ibid., p. 9. The Book of Changes is mentioned in the standard text of the Analects, 7:17, but there is a variant reading preferred, for example, by Lau; see D.C. Lau, Confucius, the Analects (Harmondsworth: Penguin, 1979), p. 88. 42. The correlations appear in the introductory calendrical sections of Lu-sfuh cfi'un-cfi'fu, chaps. 1-12, and in the account of dynasties succeeding each other by the conquest cycle of the Five Phases in chap. 13:2. The full text is available in the German translation of Richard Wilhelm, Frilhling und Herbst des Lil Bu We (Jena: Diederichs, 1928).

164

AC. Graham

43. Canon B 43. 44. C. Harbsmeier, Language and Logic in Ancient China in Needham, ibid, (note 37), vol. 7(1). 45. Roman Jakobson, Selected Writings, vol. 2 (The Hague-Paris: Mouton, 1971), especially "Two Aspects of Language," pp. 239-259. 46. Shang-tzu, chap. 1, trans, by J.J.L. Duyvendak, The Book of Lord Shang (London: Probsthaln, 1928), p. 173. 47. G. Ryle, The Concept of Mind (London: Hutchinson, 1949), pp. 11-24. 48. T.S. Kuhn, The Structure of Scientific Revolutions (Chicago: University of Chicago Press, 1970), pp. 122 f. 49. For a recent essay on using Chinese polarities to deconstruct the Western tradition, see D.L. Hall and R.T. Ames, Thinking through Confucius (Albany: SUNY Press, 1987), pp. 11-25 and passim. 50. J. Derrida, Of Grammatology, trans. G.C. Spivak (Baltimore: The Johns Hopkins University Press, 1976), pp. 1-94. 51. Needham, Ibid, (note 37), pp. 305-345. 52. C. Jung, preface to Wilhelm-Baynes, The I Ching or Book of Changes, the Richard Wilhelm trans, rendered into English by C.F. Baynes (London: Routledge & Kegan Paul, 1951), p. xxiv. 53. Great Appendix, A 9, ibid., p. 314. 54. Diagram of one of the changes; Great Appendix, B 2, ibid., p. 332. 55. For a criticism of this deliberately provocative claim (which I developed in "China, Europe and the Origins of Modern Science," in S. Nakayama and N. Sivin, eds., Chinese Science (Cambridge: Harvard University Press, 1973), pp. 45-70; see W.Y. Qian, The Great Inertia: Scientific Stagnation in Traditional China (London: Croom Helm, 1985), pp. 92-94 and passim.

Oil the Limits of Empirical Knowledge in Chinese and Western Science* N. SMn

My theme is the limits of scientific inquiry, that is, ancient Chinese concerns about whether nature can be comprehended fully by rational, empirical investigation.1 We find the limitations of observational knowledge taken up regularly in writings on astronomy, the most exact of the ancient sciences, but not in astronomy alone. Because this theme mainly appears in technical discussions rather than in writings of a general kind, we can avoid the dangerous assumption that the opinions of philosophers determined what scientists thought in China. What does emerge from the writings of fifteen hundred years is an abiding interest in the idea that the scale of the cosmos is too large, and the texture of nature is too fine, too subtle, too closely intermeshed (wei, miao and so forth) for phenomena to be fully predictable. This proposition denies that the physical world can be fully penetrated by study or fully described in words or numbers. The cognitive strategy behind it evolved, and its history can be traced. This is not the indeterminacy of contemporary theoretical physics, a point to which I will return in the conclusion, but a range of qualitative convictions drawn from mundane experience. A philosopher might divide this gamut into ontological and epistemological indeterminacy. Epistemological indeterminacy denies that it is possible to comprehend the order and regularity of the universe through study. Ontological indeterminacy asserts An earlier version of this article has been published in Time, Science, and Society in China and the West (The Study of Time V), J.T. Fraser et al., eds. (Amherst: University of Massachusetts Press, 1986.) It appears herein courtesy of the Publisher and the Editors.

166 N. Sivin that, beyond a certain point, the universe lacks the order and regularity that empirical study strives to find. The Chinese did not make such a distinction, which has analytic uses that will become clear when in my conclusion I summarize the evolution of thought about the limits of knowledge. Before looking at this idea historically, let me introduce two short but typical statements about astronomy's inherent limitation as a science. Here is an early assertion of this idea, which the polymath Cai Yong wrote in A.D. 175: The astronomical regularities are demanding in their subtlety, and we are far removed from the Sages [who founded this art]. Success and failure take their turns, and no technique can be correct forever. ... The motions of the sun, moon, and planets vary in speed and in divergence from the mean; they cannot be treated as uniform. When the technical experts trace them through computation, they can do no more than accord with [the observations of] their own time. Thus there come to be [differences between] the techniques of various periods. 2 Cai's lack of confidence should not be dismissed as a simple reflection of the crude techniques available in the second century, as we will see when we return to that period. First, let us take a passing look at a much later time, when Western astronomy was widely known. Perhaps the last such statement on the part of a scholar well qualified in astronomy comes from Dai Zhen (1724-1777), the leading philologist and in many respects the most influential intellectual of his time, in his essay on solar theory: In all prediction of celestial phenomena, as time passes there are bound to be errors that are due neither to inaccuracies in positional data nor to the need for periodic revision of computational methods. The sphere of the sky is so enormous that number and measure cannot get to the end of it, just as when we measure something an inch or a grain at a time, there is bound to be discrepancy by the time we have counted up to a foot or an ounce. Because this is so, we define units of time and observe phenomena so as to make the most of our techniques. Our best course is to continue using a technique so long as its inaccuracies remain imperceptible, and to correct it once they have been noticed. This is a matter of 3indeterminacy, as error accumulates over a long period.

Chinese and Western Science

167

Now let us look at some of the early philosophical ideas that may have formed the backdrop to statements about cosmic indeterminacy. Then we can consider the historical development of such statements themselves, in astronomy and in other departments of knowledge. Finally, we can ask what light this theme casts on the character and history of prediction as a goal of Chinese science. It is important to look at each statement, not as a great idea that must be taken at face value, but as a reflection of the viewpoint of someone with certain interests in certain historical circumstances. Since this is only a sketch of work in progress, I will consider only a few sources, and summarily indicate their circumstances. In the pre-Han classics, it is remarkable how seldom words such as wei, miao and xuan, which later imply subtlety and indeterminacy, refer to the possibility of knowledge. One pertinent treatise is the Great Commentary to the Book of Changes, the major source of orthodox cosmology from the Han on. There the word wei refers to the gentleman's sensitivity to the ethical implications of a situation as soon as they begin to evolve, long before they become obvious. Its statements are clearly not about factual or theoretical knowledge of the natural world. In the Laozi we find several other pertinent ideas. The Way itself in its constant and unchanging aspect, we are told, is shadowy and indistinct, and cannot be described. Wei and similar words are never clearly applied to the empirical world or to theoretical knowledge, but wei, miao and xuan appear together in one line that describes the exemplary gentleman: Of old those adept in the Way Their mastery recondite, subtle, and mysterious, Were too profound to be known ... One might guess from familiarity with the Laozi as a whole that the Sage becomes indeterminate as he models himself on the indeterminate Dao that he contemplates, but the text does not go quite that far.4 To sum up, by 300 B.C. certain aspects of the Dao were described as indeterminate, but these aspects are not identified with the phenomenal world, which can be described. Words implying indeterminacy rather than ineffability are used to describe the character of the ideal person rather than that of the cosmos. Not

168 N. SMn surprisingly, the key words above, which later appear in astronomical discussions, do not occur in any germane sense in the Zhuangzi. That book consistently rejects the humanistic orientation that we have found in the Laozi and the Great Commentary, and finds the logical description of experience useless. 5 The indeterminacy of the cosmos finally appears in less ambiguous form in the Chunqiufan In (135 B.C.), Dong Zhongshu's attempt to construct for the Han state a new intellectual orthodoxy that used the cosmic order to undergird the political order: The Ancients had a saying that if you do not know the future you can see it in the past. Now, in the study of the Spring and Autumn Annals, statements about the past are used to clarify the future. But because its words embody the subtlety of the 6natural order [tian zhi we], they are hard to comprehend. Dong is using the subtlety of nature, its resistance to being understood, as a metaphor for the arcane language of the orthodox classic. Indeterminacy in Astronomy Now let us pass on to the earliest statements about the limits of astronomical prediction. Most such assertions appear in the Standard Histories that chronicle the affairs of each dynasty. Computing the ephemeris and interpreting ominous phenomena were matters of concern to the state, which attempted to center this activity in its Astronomical Bureau and Imperial Observatory. Once astronomy was thus tied to the imperial charisma, a succession of computational systems for predicting the positions and chief phenomena of the sun, moon and planets was officially adopted; there were nearly fifty such systems between the beginning of the first century B.C. and the middle of the seventeenth century A.D. Improvement in technique did not lead to revision of the system in use, but to its complete replacement by a new one. This, at least, was the principle; in practice we find occasional traces of piecemeal revision, and several replacements that were no improvement at all. New systems were sometimes ordered up to signal a new dynasty or to announce a "new deal." On such occasions innovation was too much to expect.

Chinese and Western Science

169

Nearly all the judgments about astronomical systems in the Standard Histories were set down when a new system was presented for adoption, or when an old system was regularly failing to give accurate predictions. Today a sensible person who plans to buy an automobile, and who wants to find out about the limitations of a certain design, would not ask a salesman who sells that model, but would consult someone who has been driving one for some time. In astronomy, as well, familiarity breeds frankness. We usually find doubts about the extension of knowledge expressed, not with respect to a system newly presented for adoption, but when the shortcomings of an established system have become apparent; and this is all the more true when, because a competitor is in the offing, the tenure of the established system seems limited. Probably the most serious period of crisis and reassessment in early mathematical astronomy began shortly before the end of the first century A.D., when the Grand Inception system (Taichu lil, adopted in 104 B.C. and greatly developed as the Triple Concordance system (Santong 10 a century later, was about to be replaced. 7 In A.D. 92, Jia Kui presented the throne with the first major document of this crisis, his "Discussion of Calendrical Astronomy" (Lun 10. He writes of what we would call the imprecision of constants. Even the constants instituted by the legendary Sages who founded astronomy in the Golden Age, unable to endure unchanged [lit., "run through"] for thousands and myriads of years, must be altered and replaced. We [in later times] first determine angular measures and numerical quantities from the observations made over long intervals, and select those that accord with the positions of the sun, moon and planets. ... [Our] methods will thus differ from one period to another. The Grand Inception system [of two centuries earlier] cannot give accurate predictions for the present day; nor can the new system provide correct computations back to the beginning of the Han period. The computational methods of a single school can only be applicable within an interval of three hundred years. ... When the Han first attained power, it would have been appropriate to adopt the Grand Inception system [because time had come for renewal]; but there was no such reform until 104 B.C., 102 years later. Thus, early in the dynasty, there were lunar conjunctions the day before the last day of the month [i.e., two days before mean lunation], but by the

170 N. Sivin time of Emperors Cheng and Ai (32 B.C. to A.D. 1), the second day of the month was being taken as the day of lunation [i.e., the civil month was routinely set back one day], so that most conjunctions would occur on the last day of the month [which was allowable in the early Han]. This is clear proof [that calendar reform is periodically necessary]. 8 Why cannot even the Sages discover constants precise enough to be used forever? That Jia explained earlier in his report: "The Celestial Way being irregular, lacking uniformity, there are bound to be remainders. These remainders will have their own disparities, which cannot be made uniform."9 Imprecision is not a characteristic of the constants, that is, but of the universe. By the end of the second century, as my earlier quotation from Cai Yong indicates, the implications of indeterminacy had become much broader than in Jia Kui's time, and were affecting prediction in ways that did not depend only on the precision of constants. Although the sun moves along the ecliptic, and the moon is never more than six degrees from it, Han astronomers measured their positions along the equator. The ability to convert mathematically from positions on the ecliptic and the lunar orbit to equatorial right ascension and vice versa was beyond the simple linear techniques then in use. This made major improvement in eclipse theory — the central problem in traditional astronomy — seem hopeless. A report of lunar-eclipse prediction of the late second century outlines this difficulty at some length, and then draws an eloquent conclusion: In view of this [limited feasibility of mathematical solution], there is no point in rejecting any method that does not conflict with observation, nor in adopting any method whose utility has not been demonstrated in practice. The Celestial Way is so subtle, precise measurement so difficult, computational methods so varying in approach and chronological schemas so lacking in unanimity, that we can never be sure that a technique is correct until it has been confirmed in practice, nor that it is inadequate until discrepancies have shown up. Once a method is known to be inadequate, we change it; once it is known to be correct, we adopt it: this is called "sincerely holding to the mean." 10

Chinese and Western Science

171

The anonymous author is expressing resignation in the face of the crisis I have referred to. Imprecise constants could always be revised, but it was now clear — puzzlingly clear — that Han assumptions about the character of the celestial phenomena were beginning to break down. It was beginning to be apparent that certain phenomena, especially eclipses, could not be described by simple cyclic or linear methods. Finally, when this difficulty could not be resolved over several centuries, astronomers stopped trying. Cosmological hypotheses no longer ordered their computational techniques. They bought the power to predict in the simplest possible way, at the cost of the power to explain. This is a cost that greatly limits the power to predict in the long run, a lesson that did not become apparent until the seventeenth century, when the best astronomers of the time enthusiastically recognized the explanatory power of the geometric models introduced by the Jesuits. 11 Not everyone had reason to accept the idea of astronomical indeterminacy. We might expect people defending new astronomical systems rather than criticizing old ones to argue against it; indeed, examples are not hard to find. There is the spirited and rather exasperated rejoinder of Zu Chongzhi, one of China's greatest mathematical astronomers, against the attack on his new system by Dai Faxing. In or near 463, Dai developed an extensive argument along the lines of those I have quoted. Zu's defense, pragmatic rather than theoretical, is too long to cite completely: The writings of the Xia, Shang and earlier dynasties have been lost; but the historical chronicles of the Spring and Autumn period and the Han period record eclipses and lunations with care for detail; they constitute clear evidence. Testing my astronomical system by their use, I find the data entirely in accord [with my computations]. There is truly nothing speculative in [my system]. It takes precision as far as possible, so that over a span of a thousand years there is no discrepancy; whatever it be, far away though it be, it can be known. Now, I have studied all the ancient methods, and there I find much that is inexact. Computations are off by as much as three days, and the beginnings of qi periods by as much as seven hours. I know of no [ancient system] that can accurately predict the phenomena of the present time. 12

172 N. Sivin Zu's claim that there would be no discrepancy in predictions over a thousand years was excessive. It did not accord with informed opinion, could not be proven in practice, and he did not make it persuasive in principle. For reasons as much political as technical, he did not carry the day. Despite its excellence, his system was not officially adopted for fifty years. In 729, Ixing discussed the technique for predicting lunations in his new Great Expansion system (Dayan li). He politely suggested that, even if the course of the cosmos were inherently too irregular to be fully comprehended — and he did not minimize its irregularity — that would be irrelevant to the work of prediction: If the anomalies in the celestial positions [of the moon] actually fluctuated with time, providing rebukes [to the ruler] that the regularity of astronomical constants cannot encompass, and substituting for regularity a mutability [that derives] from the inaccessible [fine structure of the cosmos], this would be a matter beyond even [the ability of] Sages to assess. It can hardly lie within the scope of mathematical astronomy. This is a more meaningful statement than its brevity and skeptical tone make it look. In the first place, it reminds us that Ixing was anticipating exactly the sort of argument that Zu Chongzhi had had to fight off. The idea that astronomical knowledge was inherently limited was now being used even against the best new systems rather than just to explain the failure of old ones. Second, for Ixing the conceptual crisis I have mentioned was long over, and a disinterest in cosmology was the norm among astronomers. They rarely took up questions of the actual spatial relations or physical realities underlying the phenomena. It is curious, considering Ixing's lack of interest in these questions, that he remains the last great astronomer to give cosmology an important role in his computational system. His cosmology was not physical, however, but drew in a curiously antiquarian way on the numerology of the Great Commentary to the Book of Changes. 13 Still, his curiosity in such matters was much narrower than that of his Han predecessors. The astronomical systems that followed Ixing's were narrower still from the viewpoint of cosmology. Decreasing interest in the metaphysical and physical patterns that underlie such phenomena is not necessarily associated with more "scientific" trends in astronomy, either in China or in the

Chinese and Western Science

173

West, because in scientific work (as distinguished from certain ideal schemes of philosophers and historians of science) analysis of data and thought about their ultimate significance interact. It was the demand for a coherent and intelligible cosmic order that motivated Copernicus, Galileo and Kepler to innovate in directions that became decisive for modern science.

New Issues in the Song By the Northern Song period, discussions of the sort I have summarized were either too rare or too familiar to record. Many of the difficulties that had originally suggested inherent limitations to knowledge were no longer difficulties; for instance, the time of lunar eclipses could be predicted with some confidence. The issue for the working astronomer, as I have said, had become not knowability but technical progress. Whether some day his science might reach those ultimate limits, or would always fall short, was not an urgent problem. In the Song period, the idea of indeterminacy suggested ultimate questions of a new kind. These questions came from astronomers better prepared than their predecessors to explore methodological and epistemological aspects of their science, and from philosophers whose main interest was those aspects. My examples will be Shen Kuo (or Shen Gua, 1031-1095), who counted professional astronomy as one of his enormous range of accomplishments, and Zhu Xi (1130-1200) and Cai Yuanding (11351198), to represent the philosophers. In his Brush Talks from Dream Brook [Mengqi bitan), Shen summarizes the lost preface to his Oblatory Epoch system [Fengyuan li), an innovative document: Those who discourse on numbers [by which he means all regularities that make prediction possible], it seems, [can only] deal with their crude after-traces [/i]. There is a very subtle [wei] aspect to numbers that those who rely on mathematical astronomy are unable to know; [what they can know of] this aspect is, all the more, only aftertraces. As for the ability [of the sagely mind as exemplified in the Book of Changes] "when stimulated to encompass every situation in the realm," after-traces can play no role in that [wisdom]. That is why "the spirituality that makes foreknowledge [possible]" cannot readily be

174 N. Sivin sought through after-traces, especially when one has access only to the crudest ones. As for the very subtle traces I have mentioned, those who in our time discuss the celestial bodies depend on mathematical astronomy to know them, but astronomy is no more than the product of speculation [yH. Shen proceeds to develop an epistemological point that comes up several times in his writing, namely that in order to know, we break the continuity of nature into blocks of time that we treat as though each were uniform. As he puts it, The uninitiated say that, mathematical knowledge of the heavenly bodies being difficult to be sure of, only correlations between the Five Phases and time periods are reliable, but this is also untrue. The uninitiated who discuss the cyclic alternations [xiaozhang] of the Five Phases consider only the year. Thus [they know that], after the winter solstice the sun's motion is in the phase of Expansion [i.e., the equation of center is negative] and thus yin, and at the equinoxes corresponds to the mean. They do not realize that in the course of a month there is also an alternation. Before opposition, the moon's motion is in the phase of Expansion and thus yang; after opposition, it is in the phase of Contraction and thus yin; and at the quadratures, it corresponds to the mean. As for the associations of Spring with Wood, Summer with Fire, Autumn with Metal and Winter with Water, these are also true of the month, and not only of the month but of the day. The "Basic Questions" of the Inner Canon of the Yellow Lord [Huangdi neijing su wen] says "when the disorder is in the hepatic system, the onset [of an attack] is between 3 and 7 A.M., and the most serious time is between 3 and 7 P.M. When it is in the cardiac system, the onset is between 9 A.M. and 1 P.M., and the most serious time is between 9 P.M. and 1 A.M." Thus a single day has four seasons of its own. How do we know that there are not four seasons in each hour, or in each mark, 14 each minute, each instant? And how do we know that there are not a greater four seasons in each decade, century, Era cycle, Coincidence cycle and Epoch cycle? As for the association of Spring with Wood, within a period of ninety days there must be one [completed] cycle of alternation within another. It is impossible that the last hour of the 30th of the third month should belong to Wood, and the first hour of the next day

Chinese and Western Science

175

abruptly belong to Fire. Matters of this sort are not to be settled by the methods abroad in the world. In this second part of his short essay, Shen is writing about techniques of foreknowledge that depend, not upon observations of celestial events, but on associating in rotation the Five Phases with periods of time (for instance, the year, month, day and hour of birth), in order to yield interpretations of the latter. This simple, repetitive approach may have begun with astrology, but has been completely abstracted from what happens in the sky. Such methods cannot be reliable, Shen argues, because they imply regular and abrupt transitions from one block of time with its corresponding phase to the next. But such "quantum" transitions belie the continuous variation in motion of the celestial bodies from which the validity of the methods ultimately derives. This underlying continuity of variation, ignore it though we may, pervades time at every level from the fleetest instant to the long cycles of calendrical reckoning (the Epoch Cycle of the Han was 4,560 years). Anyone familiar with the philosopher of physics Alfred North Whitehead (1861-1947) will find this line of reasoning familiar. Shen, like Whitehead nine centuries later, was saying that a central problem for science is the gap that seems to separate our unconnected experiences from the unitary causal world that lies veiled in back of them. 15 Scientific mensuration is necessarily an act of abstraction. Near the beginning of his proposal of 1074 for a new armillary sphere, Shen argues this point with great clarity — for the first time in history, I believe: Degrees [on the equator and ecliptic] are invisible; what is visible are stars. [The paths] followed by the sun, moon and planets are occupied by stars. Twenty-eight stars are located [exactly] on a degree division; they are called "mansions" [she]. It is mansions that make it possible to measure degrees, and degrees that make it possible to create numerical regularities. Degrees are things that exist in the sky. When we make an armillary sphere [to measure intervals between real bodies], the degrees exist in the instrument. Once the degrees are in the instrument, then the sun, moon and planets can be isolated [tuan] in the instrument, and the sky no longer is involved. It is because the sky is no longer16 involved that what is in the sky is not difficult to know.

176 N. Sivin Shen implies that one can know either about the organismic universe as a whole ("the sky") or about particular phenomena in it. Observational, empirical science can yield only knowledge of the second kind (a point about which Whitehead would disagree). In doing so it rules out perceptions of the first kind. They can be reached only by other kinds of knowledge — intuition, illumination and so on — in which Shen is equally interested. Cai Yuanding, another polymath, did away with one of the basic confusions of the Han astronomers. Cai wrote at least one book on mathematical astronomy. This detailed study of bang's astronomical system is lost, but certain important arguments are preserved in the conversations of Cai's mentor and friend, Zhu Xi, with Zhu's disciples. It is clear from Zhu's paraphrases that Cai believed inaccuracies of prediction do not imply indeterminacy. Beneath an irregularity may lie a more complicated regularity waiting to be discovered: When an astronomical system is first being designed, the discrepant measures of the celestial rotations are combined and included in the computations. So, many years later, there will be discrepancies of so many fractions of a degree, and after so many additional years, of so many degrees. If, from these discrepant quantities, correct quantities are computed, and this process is repeated to the limit [i.e., until the magnitude of the correction becomes negligible, jintou], the astronomical system can be made essentially correct and free of discrepancies. People today, never having reached a comprehensive and correct understanding [da tong zheng], simply claim that the discrepancies are inherent in the celestial rotations. They make systems of computation seeking accord with the celestial phenomena, but their ephemerides become increasingly discrepant. The point is this: they do not understand that, if the sky is able to manifest a certain discrepancy, it is precisely because the celestial rotation must be of that kind. A discrepancy does not, as Cai's contemporaries think, reflect an anomaly in nature, but rather a more complicated regularity than originally assumed. Ad hoc technical adjustments simply obscure the discrepancies. In doing so they also obscure the underlying complex pattern that will keep generating discrepancies until it is understood. This is an important perception about method. When we

Chinese and Western Science

177

remember the gradual discovery in Europe of the various inequalities that complicate the moon's motion, we are reminded that the failure of Hipparchus' (fl. ca. 130-150 B.C.) first inequality to give perfect predictions suggested to Ptolemy (ca. A.D. 100 ca. 165) the evection, the second inequality. The discrepancies for which the evection could not account suggested to Tycho Brahe (1546-1601) the third and fourth inequalities. Similar processes of discovery can be traced in the history of Chinese astronomy, but Cai was the first (at least the first reflected in the surviving record) to make the point explicit. His own attitude toward the determinate character of the phenomena was decidedly nuanced, even though he did not accept his contemporaries' reasoning about what implies indeterminacy. Zhu quotes him elsewhere to the effect that "there is no constancy in the celestial rotations; the sun, moon, and planets are accumulations of qi; they are all moving things [dong wu\. Their angular motions may be faster or slower, beyond the mean or short of it; they are not naturally uniform." We have already seen that Cai considers these motions predictable. As Zhu remarks, "Cai was not saying that there is nothing determinate in the rotation of the sky, but that the angular motions of the luminaries are as they are." 17 The last quotation from Cai is best understood in the light of similar beliefs held by such Occidental luminaries as Plato and Ptolemy, for whom the planets are divine and self-propelled. This view was an alternative to the idea that the planets are passively driven in their rounds by some common source of motion that determines the speed of each. A philosopher who finds no evidence for mechanical linkages powering the celestial luminaries is likely to find the idea that each planet is the source of its own motion more plausible. Its velocity is thus internally determined and arbitrary with respect to those of other planets. If constant, it is arbitrarily constant. The 'erratic' retrogradations of the planets are thus accounted for; they could not be explained by those who considered the planets passive. In Greece and Hellenistic Egypt, a source that determined its own motion would be divine; in China, it was an animal-like "moving thing." Neither implies that its motion must necessarily be irregular or unknowable. The astronomer simply attempts to impose order upon whatever irregularities his observations reveal. What matters about Cai's attitude is that he faced the issue

178 N. Sivin of indeterminacy instead of making assumptions that render it all the more problematic. He could thus imply that, even, when taken seriously, it need not impede astronomy. Zhu Xi, like Cai Yuanding, did not believe that there were inherent limits to the astronomer's power to predict. His attitude emerges in several discussions of a chapter from the Mencius {Mengzl, 4B:26), which, in explicating the innate moral nature of man (xing), refers to the work of the astronomer. As Zhu explicates the relevant passage, it would mean: "Consider the sky so high, and its markpoints so distant; if we seek the traces of actual events [gu], without leaving our seats we can bring before us the solstices of a thousand years." There is no basis for reading into Mencius' casual statement a pronouncement on the limits of empirical knowledge in astronomy. But Zhu Xi, in one of several conversations about the chapter, relates this passage to that question: Mathematical astronomers computing backward from the present day are able to proceed without error even to the moment when the physical cosmos was formed. This is possible only because they follow traces of actual events [i ran zhiji]. There are sometimes irregularities in the true motions of the sky and of the sun, moon and planets, but as time passes these recur spontaneously to the norm. Here Zhu Xi understands gu as traces of what exists or has existed; in other conversations, referring to other occurrences in the same chapter, he explains the word more subtly as "why something is so" [suoyi ran) and "what something does" (suo wei). He is using Mencius' undefined gu, relating it to Shen Kuo's undefined "traces," to refer to phenomenological patterns. In a society never touched by Plato's opposition of phenomenon and reality, this is a more original step than it might appear. 18 In the discussion of astronomical prediction that provided the long paraphrase from Cai Yuanding quoted above, Zhu Xi begins more or less at the point where Cai stopped, but moves off in a significant new direction: Someone asked why calendrical systems are repeatedly inaccurate. "How can it be that in ancient and modern times no one has studied this matter thoroughly?" Zhu Xi replied: "It is precisely because no one has studied this

Chinese and Western Science

179

matter thoroughly enough to rule out further change that there are repeated discrepancies. If it were studied with enough precision to yield a definitive method of computation, there would be no further discrepancies. ... The astronomical techniques of the Ancients were imprecise [shukuo, lit., loose], but there were few discrepancies. The more precise [mi, lit., tight] the systems of today are, the more discrepancies appearl" At this point, he measured off one side of his desk with his hands, saying, "For instance, if we divide this breadth into four sections, each is limited in width by its borders with the others. If a discrepancy [between the widths] appears, it will be restricted to one of the sections. Large though [the discrepancy] may be, even so extreme that it involved a second or third section, it would still be restricted to the four sections. So it would be easily computed, and any discrepancy could easily be seen. The astronomical systems of today [in effect] divide these four sections into eight, and the eight into sixteen. As the limits [of the sections] become more precise, the frequency of discrepancies becomes greater. Why is this? Because, as the limits become more precise, they are increasingly overstepped. The discrepancy may be identical, but19the precision of ancient and modern systems differs." Zhu Xi is saying, if I understand him correctly, that increases of precision have led to greater expectations of accuracy, and that it is against these expectations that recent systems were failing; early systems satisfied lower standards. But that is not my point. This is the first clear explanation in Chinese, I believe, of the difference between accuracy and precision. This is not a small contribution to the methodology of the exact sciences. This was certainly not Zhu's aim, but it is hardly a by-product. These concerns with method and with theory of knowledge, although ignored by modern historians, were carried on by the leading scholars of the Qing period. In a recent book, John Henderson traces the growing importance in Ming and Qing philosophy of arguments from mathematical astronomy. Henderson shows that these concepts of quantitative origin largely replaced earlier conceptions such as yin-yang and the Five Phases. For example, prominent humanists between the mid-seventeenth and late eighteenth century became aware, through Western astronomical writings which they studied eagerly, of such secular changes as the slow decrease in the obliquity of the ecliptic. They

180 N. Sivin came to believe, as did many Europeans in the later part of the same period, that these were not entirely predictable, and their magnitudes could only be known through observation: A number of Ch'ing [i.e., Qing] scholars of varying scholastic affiliations thus identified several of the astronomical anomalies and deviations known to them as basically indeterminate, frequently drawing the conclusion that the patterns of the cosmos in general shifted in an irregular and even capricious fashion. They even regarded anomalies not so much as departures from a predictable order as themselves constitutive of the fundamental order, or disorder, of the cosmos. 20 Limits of Inquiry in the Qualitative Sciences It is not surprising that the issue of indeterminacy should have arisen mainly in astronomy, the one science that was both quantitative and concerned with prediction. The idea that empirical knowledge and understanding may be inherently limited also turns up in areas of inquiry that are not computational. Sometimes it is brought up by polymaths who are aware of the issue within astronomy. Shen Kuo, for instance, discusses a case in which lightning, striking a house, left its wooden structure unharmed but melted metal objects inside it. People insist that fire will burn things of vegetable origin before it melts things made from metals or minerals, but in this instance the latter all fused while not one of the former was destroyed by fire. This is not a matter that human capacities can fathom. A Buddhist treatise says "Water makes the Naga fire blaze up, but puts out the human fire." How true that is! People only know about matters in the realm of mankind. Outside that realm what limit can matters have? We may aspire, by our insignificant worldly wisdom and common sense, to get to the 21 bottom of ultimate truths, but that is hardly possible. And I have already quoted one of Shen's references to medical theory. We also find Fang Yizhi, in his Little Notes on the Principles of Things (completed 1643/1650), using what he had learned from Jesuit missionaries about optics to argue that the tendency of

Chinese and Western Science

181

light rays to diverge and of shadows and images to converge renders certain optical phenomena unexplainable. Fang describes an experiment in which a piece of paper with four or five small holes yields multiple images of the sun. But, as the paper is moved upward, away from the surface on which they are projected, the multiple images blend — in a way that puzzles him — to form a single image of the sun. "Sound and light," he argues, "are always more subtle than the 'number' of things." By "number" Fang means amenability to exact quantitative description, as "by acute angles and straight lines," i.e., geometric constructions. 22 In medicine, the idea that one can hope to understand only so much about the vital processes of the human body is natural enough. Ever since the sixth century, medicine has been strongly influenced by Buddhist ethics. Since its beginnings, physicians and medical scholars have drawn on numerology, yin-yang, and Five Phases cosmology, and even on astronomy, in order to investigate the links between the internal order of the body and the order of nature that surrounds it. 23 An obvious example is Zhang Jiebin's statement, in his Collected Treatises Jingyue quanshu (preface dated 1593), about what is needed to comprehend vital processes: "Anyone who does not possess transcendent wisdom is not prepared to master their subtleties [weimiao]; anyone who does not possess clarity of moral judgment is not prepared to make fine distinctions concerning what is correct." Empirical observation, in other words, must be supplemented by self-cultivation. In some such statements, Buddhist influence is plain to see. Typical is Yin Zhiyi's preface to his father, Yin Zhongchun's, little handbook of diagnosis and therapy, entitled Mental Dharmas of Eruptive Disorders (probably shortly before 1621): "Medicine" [yil means "meaning" [yi\. [The inner meanings of medicine, the patterns of vital processes] may be apprehended by the mind, but cannot be transmitted in words. Because these inherent patterns attain such arcane subtlety [weiao], even though the mind may achieve great constancy [in contemplating them], in [therapeutic] doctrine there can be no fixed rules. The interaction of hot and moist as governed by yin and yang, the relations of mutual production and overcoming among the Five Phases, change from one moment to the next ...

182 N. Sivin Yin begins with a familiar punning definition of medicine, and moves immediately to the Chan Buddhist notion of wordless teaching. Yin's word for therapeutic doctrine or method [fa] is the same as the term for dharma in the title of his father's book (in which "mental dharmas" means at one level "doctrines or truths to be grasped by the mind"). Such instances from therapeutic manuals could readily be multiplied. In astronomy, as we have seen, the limit of observation was a live and evolving issue. In medicine, however, what we find is reiteration of a familiar theme — a formula, more or less — that is seldom examined critically. My preliminary conclusion, pending deeper study and reflection, is that indeterminacy in medical writings is less significant for the history of medical thought than for epistemology in general. 24

Conclusion The sources on which I have drawn indicate that, in Han astronomy, themes of both epistemological and ontological limits on knowledge appear, but that, as experience led to confidence, the limit came to be seen consistently as one of imprecision rather than of inherent disorder. This situation continued until the two postulates were again combined in the attack on yin-yang and the Five Phases, as the basis of Confucian orthodoxy that John Henderson has documented.25 I have suggested that ideas of astronomical indeterminacy first arose to explain what would now be considered the failure of crude predictive techniques, and that this idea gradually became established to account for what historians would now explain as the failure of crude assumptions about the character of the celestial motions. As these assumptions were given up, and the crisis subsided, it seems that ideas of indeterminacy were for a while used more as a weapon to beat back innovation than as a means to reexamine past failures. These ideas began playing a productive role once more from the Song on, when they were used for diverse purposes, among them to direct critical attention to issues of what we would now call epistemology and method. Some who used them, including Cai Yuanding and Zhu Xi, did not accept the idea that empirical knowledge was necessarily limited. Why should a notion that looks so obscurantist, so opposed to

Chinese and Western Science

183

the idea of progress, have played such an enduring collection of roles in the history of science? To take first things first, since the idea of progress entered Western scientific thought in the eighteenth century, and that of China much later, how progressive a given early idea seems to moderns is beside the point. The point is rather what it meant and how it was used in its time. I suggest that the idea of indeterminacy was the one proposition that consistently challenged astronomers to come to grips with the distinction between two issues: First, what is involved in predicting future observational data from past observational data? And second, what is involved in making intelligible the nature from which we draw observational data? This is the difference between astronomy as a collection of data and techniques, and astronomy as a science. Despite the crisis in astronomy that began in the Later Han, the urge to make astronomy a science again never entirely subsided. It became a strong motivation from the eleventh century on, as impulses from philosophy stimulated astronomers, and vice versa. I think that, ultimately, it will be possible to show that discussions of the limitations of inquiry in the Ming and early Qing encouraged both the concern for causes and explanations in the seventeenth century and the prompt and positive response of leading Chinese astronomers to Western astronomy. 26 The issue I have outlined does not seem to have been important in Western science after Heraclitus and Parmenides. The concept of quantum indeterminacy in modern physics is concerned with a quite different, purely mathematical, issue. It states that there is a small, constant limit on the combined precision with which we can simultaneously measure two parameters, for instance position and momentum, of a particle event. The better we know one, the worse we know the other, up to that limit. In translating the equations into ordinary language, some popularizers have read into this very abstract scenario portentous implications about the subjectivity of the observer, and have even concluded that, at the limit, theoretical physics collapses (or rises, as the case may be) into mysticism. Sympathetic though one may be toward attempts to combat a mechanistic arrogance for which there is also no warrant in the equations, these reinterpretations are not a triumphant revival of an old theme, but rather bad philosophy of science. They are irrelevant to the present topic. For Plato, observation of phenomena alone cannot lead to

184 N. Sivin knowledge of reality. It can yield only a third-hand reflection of the abstract Ideas that real knowledge is about. The study of mathematics helps us toward them, but direct apprehension of the ideas is a contemplative, not an empirical, process. Aristotle believed that the reality of things was within them, and could be deduced directly from them, but "the advances made by the arts and sciences in each civilization were the fulfillment of the potentialities of their natural form beyond which they could not go." 27 The Skeptics denied that one could know with certainty; but their discussions of this point served to suspend judgment on all matters. This was not a doctrine that could greatly influence natural science. The Stoics were the school closest in intellectual temper to Chinese cosmology, and they were influential in science, especially medicine. The empiricists among them opposed Skepticism and, thus, indeterminacy. They were much concerned with the possibilities of knowledge, and considered all truth built up from what the senses deliver, judged by what Stoics called "right reason." In the European Middle Ages, the analogous issue — a weak analogue — was the relationship of faith and reason. In the midthirteenth century, we find St. Thomas Aquinas quoting with approval the words that St. Hilary of Poiters had set down nine hundred years earlier: For he who devoutly follows in pursuit of the infinite, though he never come up with it, will always advance by setting forth. Yet pry not into that secret, and meddle not in the mystery of the birth of the infinite, nor presume to grasp that which is the summit of understanding: but understand that there are things thou canst not grasp. 28 The faith in unlimited knowledge, in untrammeled understanding, is not a characteristically Western faith; it is a modern faith. It is as welcome today in Beijing as in New York, perhaps more so as skepticism gains ground in the overdeveloped nations. In seeking valid Western analogies to the role of indeterminacy in Chinese intellectual history, I would take an entirely different direction. I would prefer to ask whether we can find ideas that appear irrelevant or "unscientific" from a vulgar positivist point of view, but that nevertheless played enduring roles in encouraging discussions of scientific issues. It is not hard, in fact, to

Chinese and Western Science

185

think of examples. One is Zeno's paradoxes. It is well known that every important discussion of the continuity of points on a line, from the Greeks to the end of the nineteenth century (Georg Cantor, 1845-1918), focused on those paradoxes. 29 One more important contrast between East and West remains to be drawn. Whether one begins with Parmenides or Plato, classical European philosophers who wished to find a way past mere speculation about Nature insisted on asking, and arguing about, what they saw as the most fundamental question: How can knowledge be certain? What we know with varying degrees of likelihood, what is merely probable, has no place in science. This axiom dominated Western science until the advent of statistical thermodynamics made it meaningless. The modern vision of the world that we experience as merely the summation of innumerable random atomic, nuclear and subnuclear phenomena, left this drive toward certainty a quaint relic of a dead faith. It had no analogue in Chinese thought. Empirical knowledge is neither certain nor probable, merely given. The pattern one discerns may or may not be objectively there, but that is no more than to say that one may be empirically right or wrong. For certainty, one looks to illumination, introspection and other alternatives to purely cognitive processes. Certainty is, in the last analysis, a spiritual and moral stance. In conclusion, let me return to the beginning of this essay. J.T. Fraser once wrote that limits of inquiry divide the world into those phenomena that are predictable and those that are not. The dividing line between these phenomena constitutes a statement of belief, one of those irrationalities on which rationality must always rest. That demarcation usually amounts to a claim about the nature of time. 30 As we consider the many Chinese statements that we have reviewed through history, the issue is indeed in one sense the domain of prediction. Even more fundamentally, it is what the activity of prediction rules out. What it rules out is an uninterrupted response, at once intuitive and rational, to the concreteness and endless variety of phenomena in nature. That response is what theory in the traditional qualitative sciences, such as medicine, alchemy and siting, seems always to have striven for. Theory is necessarily abstract, based on rigorously defined concepts. Chinese scientists looked for a balance of concept and phenomenon, for accounts of nature that did justice to its rich-

186 N. Sivin ness. That, not Occam's razor, nor the necessity of geometric demonstration, was their aesthetic criterion. No one familiar with the sources would argue that their stance is irrational. I would hesitate to say that it is less rational than that of the European positivists of half a century ago, whose starry-eyed faith led them to draw and defend an ultimately indefensible line between positive knowledge and the outcomes of all other mental activity. The Chinese thinkers I have cited were very much concerned with the theoretical ordering of phenomena, and with prediction, in astronomy and medicine. What we see them saying, more and more explicitly, is that prediction is a reductive act and thus an inherently limiting one.

Notes 1. The Study of Time V (Amherst: The University of Massachusetts Press, 1986), pp. 151-169 For Chinese characters, see the 1986 version. 2. Hou Han shu, "Lull zhi," 2:1492. Astronomical and astrological treatises of the Standard Histories are cited from the series Lidai tianwen Mi deng zhi huibian, 10 vols. (Beijing, 1975-1976). 3. "Ying ri tui ce ji," pp. 113-118, in Dai Zhenji (Shanghai: Shanghai Guji Chubanshe, 1980); esp. p. 115. Let me reiterate that my use of the philosophical term "indeterminacy" here and below merely translates Chinese assertions that there are inherent limitations to observational knowledge. 4. Zhou yi, "Xici da zhuan," B:4-5; Wilhelm, The I Chtng, pp. 367, 370, where wei is translated as "that which is hidden" and "first imperceptible beginning." Laozi, 21, 1 and 15; cf. the translation of line 15 with that of D.C. Lau, Chinese Classics. Tao Te Chtng (Hong Kong: The Chinese University Press, 1982), p. 21. 5. The famous anecdote about Ding the Cook, in Chapter 3 of the Zhuangzi, is about a distantly related conviction, namely, that manual skill is not a matter of technical rules but rather of being "in touch through the daemonic in me," as A.C. Graham's translation puts it. A number of similar "knack passages" occur in chapters of the Zhuangzi outside the original corpus. In the story of Bian the Wheelwright in Chapter 13, for instance, the intuitive still that conies from long practice is adduced to argue against learning from books, not just from orthodox classics, but from all books that purport to transmit human experience. For exceptionally perceptive translations, see Graham, Chuang-tzu: The Seven Inner Chapters and Other Writings from the Book Chuang-tzu (London: George Allen and Unwin, 1981), pp. 63-64, 135-141. 6. Chunqiufan lu (Si bu bei yao ed.), 3: 10a.

Chinese and Western Science

187

7. The discussion of Han astronomy that follows is documented in Sivin, Cosmos and Computation in Early Chinese Mathematical Astronomy (Leiden: E.J. Brill, 1969). See also the insights of Yabuuti Kiyosi (Yabuuchi Kiyoshi), in "The Calendar Reforms in the Han Dynasties and Ideas in their Background," Archives internationales d'histoire des sciences, 2(1974), 51-65, and the monograph by Yabuuchi and Noda Churyo, Kansho ritsurekishi no kenkyu (Kyoto: Zenkoku Shobo, 1947). 8. Hou Han shu, "Liili zhi," 2: 1482. 9. Ibid. 10. Ibid., p. 1496. For a fuller translation, see Sivin, Cosmos and Computation, pp. 61-62. The last line contains an allusion to the Confucian Analects, 20:1. 11. Sivin, "Copernicus in China," Studia Copernicana (Warsaw), 6(1973), 63- 122, esp. pp. 72-73. 12. Sung shu, 13:1768-1769 et passim. The beginning date of each qi period, of which there were twenty-four in a tropical year, was a basic element of the Chinese ephemeris. 13. Xin Tang shu, 27A:2177. Ixing's computational system was in fact attacked in 733, four years after it was adopted and six years after the astronomer died, being was not charged with technical inadequacy — perhaps because he had anticipated criticisms on that count — but with plagiarism from Indian sources. The best account is in Christopher Cullen, "An Eighth Century Chinese Table of Tangents," Chinese Science, 5(1982), 1-33, esp. 30-32. On the numerological cosmology of the Great Expansion system, see the section of Ixing's treatise on the "Rationale of the Basis of the Ephemeris," in Xin Tang shu, 27A:2169-2173, trans, in Ang Tian Se, "I-hsing (A.D. 683-727): His Life and Scientific Work," unpublished Ph. D. dissertation, University of Malaya, 1979, pp. 419^145. 14. A mark (ke) is 0.01 day, approximately fifteen minutes. 15. Mengqi bitan. Item 123 {Mengqi bitan jiaozheng; rev.ed., Beijing, 1960; 1:292). Of the two quotations in the first part, one quotes the Great Commentary, A. 9, and the other alludes to it (see Wilhelm, trans., 1:339). The Huang ti nei ching su wen citation is to 7(22): 125 in the Shanghai, 1954, edition. On the Era and other long cycles, see Sivin, Cosmos and Computation, pp. 12-21. For an extremely satisfactory discussion of Whitehead, see Dictionary of Scientific Biography, s.v. 16. The Hun yi yi is preserved in Song shi, "Tianwen zhi," 48: 800-808; I cite the critical text in Mengqi bitan. Item 127, note 6 (I: 297), and accept the emendation of zhou to hua. Shen is using tuan in a technical sense that draws on several of its early meanings: to shape into a ball with the hands, to gather, to tie in a bundle, exclusive. For further reflections on the complementarity of scientific and other modes of knowledge in Shen's thought, see Sivin, "Shen Kua," Dictionary of Scientific Biography, s.v. 17. Zhuzi yu lei (Zhuzi yu lei da quan, Kyoto, 1668, reprint by Zhongwen Chubanshe, Kyoto, 1973), 86: 12a: 2:14b-15a. There is an interesting

188 N. Sivin

18.

19.

20.

21.

22.

23.

discussion of these and other passages in Zhu Xi's writings in Yamada Keiji, Shushi no shizengaku (Zhu Xi's studies of nature; Tokyo: Iwanami Shoten, 1978), pp. 279-301. Cai Yuanding's astronomical monograph was entitled Dayan xiang shuo. See his biography by Rulan Chao Pian, pp. 1037-1039, in Herbert Franke, ed., Sung Biographies, 3 vols. (Wiesbaden: Munchener ostasiatische Studien, 16, 1976). These conversations are recorded in Zhuzi yu lei, 57: 14a-17a. James Legge, who in his 1861 translation of the Mencius often relied on Zhu Xi rather than on more philologically rigorous later commentators, translates gu as phenomena (Hong Kong: The Chinese Classics, 1861), II: 206-207. For an especially penetrating discussion of the passage from Mencius, see Patrick E. Moran, "Key Psychological and Cosmological Terms in Chinese Philosophy: Their History from the Beginning to Chu Hsi (11301200)," unpublished Ph.D. dissertation, University of Pennsylvania, 1983, pp. 68-71. Zhuzi yu lei, 86: lib-12a. "Definitive method of computation" is a tentative translation for ding shu, which may mean nothing more elaborate than "definitive constants," or conceivably (but in my opinion less likely), "measures corresponding to true rather than mean motions." I translate cha as discrepancies rather than anomalies, since the example concerns error in measurement rather than inequality of motion. A pertinent essay is Hashimoto Keizo, "Seido no shiso to dento Chugoku no temmongaku" (Ideas of precision and traditional Chinese astronomy), Kansai Daigaku ShakaigakubuKiyo, 1979, 11. 1:93-114. The Development and Decline of Chinese Cosmology (New York: Columbia University Press, 1984), p. 249. I am grateful to Prof. Henderson for his comments on an early draft of this essay. His doctoral dissertation, from which the book is extensively revised, was in part responsible for inspiring me to take up this topic, and provided useful references. I do not agree with some of Henderson's interpretations, but his work takes up heretofore neglected questions and demonstrates the importance of astronomical writing in research in Chinese intellectual history. id=hen Mengqi bitan, Item 347. Mark Elvin has remarked on the implications of this passage for "the probable limitations of human understanding" in The Pattern of the Chinese Past (Stanford: Stanford University Press, 1973), p. 233, note. Wulixiao zhi (1st ed. of 1664), I: 34a-b. The whole passage is translated (with some misunderstandings) by Willard J. Peterson in "Fang I-chih: Western Learning and the 'Investigation of Things,'" pp. 369-409, in Wm. Theodore De Bary et al., The Unfolding ofNeoConfucianism (New York: Columbia University Press, 1975), esp. p. 391. See, for instance, Sivin, Traditional Medicine in Contemporary China (Science, Medicine, and Technology in East Asia, 2; Ann

Chinese and Western Science

24.

25. 26.

27.

28.

29.

30.

189

Arbor: Center for Chinese Studies, University of Michigan, 1987 [published 1988]), pp. 43-94. Zhang, Jingyue quanshu (photolithographic reprint of 1624 ed., Taipei, 1972), 3: 75b; mistranslated in Paul U. Unschuld, Medical Ethics in Imperial China: A Study of Historical Anthropology (Berkeley: University of California Press, 1979), p. 82; Yin, Zhenzi xinfa or Shazhen xtnfa, printed with Yin Dachun's Yizang shumu (Shanghai, 1955), pp. 107-108. The title of the latter work, meaning "Bibliography of the Medical Triptaka," also draws on Buddhist imagery. For a number of additional quotations, see Henderson, ibid, (note 19), esp. pp. 246-253. For useful data in this connection, see Henderson, and Benjamin Elman, From Philosophy to Philology. Intellectual and Social Aspects of Change in Late Imperial China (Harvard East Asian Monographs, 110; Cambridge: Harvard University Press, 1984), esp. pp. 29-32. A.C. Crombie, "Some Attitudes to Scientific Progress: Ancient, Medieval and Early Modern," History of Science, 13(1975), 213-230. See also on this topic, which is best distinguished from the one discussed in the present essay, E.R. Dodds, The Ancient Concept of Progress and Other Essays (Oxford: Oxford University Press, 1973). St. Hilary, De trinitatise 2: 10, 11, cited in The Summa Contra Gentiles of Saint Thomas Aquinas Literally Translated by the English Dominican Fathers from the Latest Leonine Edition (London: Oates, 1924), 1.8 (I, 16). No adequately detailed monograph on the role of Zeno's paradoxes has been published, but see G.E.L. Owen, "Zeno and the Mathematicians," Proceedings of the Aristotelian Society, n.s., 58(1957-1958), 199-222. Personal communication, 14 October 1982.

Dharma and Rationality B. K. Matllal

Introduction: Primitive Form of Rationality and Free Riders I wish to confess to my readers at the outset that some of the stories I heard in my childhood still hold a certain fascination for me. Although this may be tantamount to bias, I nevertheless wish to start with such a story, which is often heard in the rural part of Bengal. For I believe that the morals of this story and other similar tales often contain rudiments of some primitive theory of rationality. One day, a king wanted to create a lake of while milk because he was tired of looking at pools and lakes full of muddy water. He imagined what it would be like to watch the beauty of the milky waves glittering in the sunshine. Besides, people would be able to have milk to drink whenever they wanted. So the order was given, and thousands of diggers got busy digging in order to create the lake. The digging was successfully completed, but the final ingredient was still missing: the milk. Where could one locate a quantity of milk huge enough to fill the lake? (Remember, the dairy industry was at a very primitive stage in those days, and there was not even enough milk to feed the babies!) But, since the king wanted it, an order was given that, on a certain day, every citizen had to pour a bucketful of milk into the lake so that the lake that had been dug would be filled with milk. This meant that about one hundred thousand bucketfuls of milk was expected to come from the citizens. But, what happened the next day? Each citizen thought to himself, "Since milk is so dear and so scarce, I will pour a bucketful of water into the lake in the darkness of the night, and when my little bit of water is mixed with all the milk the others will pour in, nobody will be any the wiser." Next morning, when the king woke up he went to take a look at his new lake of milk, but what he saw instead was a huge lake full of

192 B. K. Matilal muddy water. Why? Because not a single drop of milk had come from any of the citizens, since everybody had thought that the others were dealing with the milk and that he could therefore silently slip into their group without paying the price. In other words, everybody wanted to be a 'free-rider'! But everybody acted with rational self-interest in mind. Let us assume that everybody wanted the lake of milk to be created because it would have been 'good' for the society in general — since it would create a common resource from which 'free milk' would be distributed to everybody — and not just because it would fulfil the king's whim. Then, one could say, at least from the point of view of this primitive theory of rational behavior, that, by pouring in water instead of milk, every citizen acted not only immorally but also irrationally. For, based upon this socalled rational self-interest, the argument of each citizen was in fact irrational, since, among other things, everybody knew that everyone else could take the same line of action — based on the same argument — and that, if everybody did this, nobody's selfinterest would be served because nobody would receive the intended benefit. From the moral point of view, we cannot say anything very different. If it is assumed that each citizen was reasonably capable of making the sacrifice — i.e., contributing a bucketful of milk — and that in making that contribution none would have been subjected to unbearable hardship or a great deal of suffering, then, according to any moral theory, it would be considered immoral to be a 'free-rider.' Indeed, in accordance with one of the most well-entrenched and widely held views of morality today — Utilitarianism — 'free-riders' such as those described in the story are clearly immoral because this action is irrational. In spite of many objections that have been raised against classical Utilitarianism, it unquestionably enjoys popularity and the support of many today because of the prime importance it gives to reason or rationality. In the past, Utilitarianism fought against the 'intuitionists' in political and moral fields, by emphasizing an impartial rational test for judging the existing moral principles and traditional social institutions. In many modern, modified versions of Utilitarianism, there is one common thread, viz., adherence to the principle that moral issues must be decided by rational tests. In fact, it has been even claimed that any theory of

Dharma and Rationality

193

morality should itself be part of a general theory of rational behavior. > By the term dharma in the title of this essay, I understand nothing short of moral virtue, or rather, a theory of moral behavior, as it is found implicit in India's traditional wisdom. Thus, I prefer to call the attempt to be a 'free-rider' an a-dharma, or a violation of dharma. For, in the wider tradition of India, dharma stands for neither "religion" nor the narrower caste-oriented duties. The best evidence for my claim are the frequent discourses on dharma, found throughout the two great Indian epics, the Ramayana and the Mahabharata — in all their different versions, composed at different periods of history and in different regional languages — as well as in many folktales, stories and fables told at different times.

Dharma and its Rational Critique Dharma is the popular subject of inquiry often found in all this narrative literature. The nature of dharma is often hotly debated and argued for; no other principle has been regarded as sacred. This need not be very surprising, for neither in Buddhism nor in Jainism, or even in Hinduism, was God cited as the authority on dharma. Hence search for a rational basis of dharma is often compatible with these religious traditions. There were, of course, the Hindu Scriptures. But these scriptures proved to be flexible, sometimes to the point that they seemed to have meant whatever their interpreters chose to make them mean. Furthermore, even when the literal text of the scriptures was taken seriously, the interpreters of the Mimamsa school undertook to make a rational examination ("mimamsa" means rational examination) of the meaning of the Vedic (scriptural) statements. It is true, of course, that, in those days, search for a rational basis for moral or social behavior was not free from unconscious bias and inherited prejudices; very few are, even in our own days. The supposed rational argument often turned into a form of 'rationalization' or 'apologia.' The point, however, is that the tradition did not have to wait until something like the Age of Enlightenment came in order to question the basis of moral and religious beliefs. Even some actions of Krishna and Rama — who

194 B. K. Matilal were regarded as incarnations of God on earth — were subjected to rational criticism over the ages. There is enough textual evidence to prove it. 1 My claim so far has been that the dharma tradition developed through an attempt of rational criticism of itself. Another piece of evidence in support of this claim can now be briefly cited. Consider the caste-hierarchy that is almost as old as the Vedic Hinduism. The Sramana tradition provided a rational critique from outside. But, even within the domain of Vedic Hinduism, there occurred a search for a rational basis. What resulted was an interpretation of the karma doctrine that was intended to provide a 'rational' basis for the apparently irrational practice of caste-hierarchy or social inequality. It was Max Weber who, in 1920, characterized the castedharma as "anti-rational," because it denied the "'natural' equality of man." 2 But, Weber then went on to pay glowing tribute to the karma doctrine of Hinduism, for he saw how the latter provided the rational basis of the caste-hierarchy. For him, the karma doctrine of Hinduism was "the most consistent theodicy ever produced in history." 3 The caste-hierarchy was, it may be assumed, historically prior to the development of the full-fledged karma doctrine with its rebirth hypothesis. There is some justification for this conjecture, for the Puru?a-Sukta in the Rg-Veda clearly refers to the caste-hierarchy. But, although the rudiments of the notion of karma or 'just deserts' may be found scattered in various places of the Vedic hymns, it was explicitly referred to as a doctrine (although still deemed an esoteric or secret doctrine) in the Brhadaranyaka Upani§ad. It was Yajnavalkya, the wellknown philosopher of the day, who mentioned the karma doctrine in the court of King Janaka, in reply to a question from Artabhaga. Yajnavalkya did not, of course, expound the doctrine publicly. This is how it is supposed to have happened. All the wellknown brahmin-scholars (both men and women) and priests assembled in the court of Janaka, and the one who could prove himself to be the best among the scholars gathered would win a prize of one thousand cows and ten thousand coins. 4 When Yajnavalkya claimed supremacy, a very tense session of debate, a form of question-and-answer session among the brahminscholars, took place, all the others asking questions that tested the depth of Yajnavalkya's knowledge, and the latter endeav-

Dharma and nationality

iyo

oring to satisfy them. When Artabhaga rose, 5 he started asking questions on matters concerning various rituals. But then he asked a very important question: "Assuming that the person survives his bodily death, what substratum would the person have when his material body dies and dissolves into all its [material] ingredients?" Yajnavalkya replied by saying: "Hold my hand and let us go to a secluded place. Only two of us will consider [and know] the answer, not the others present here." Then the two retired to a secluded place for a while. What they discussed has not been reported in the text. Only a brief account of the meeting has been given, in four lines. It is said that they talked about karma — a doctrine that should be admired by everybody, "For good deeds earn merit for the person while bad deeds earn demerit" (or, the person becomes good in his next life through good deeds, and he becomes evil through evil deeds). 6 This was the outline of the primitive karma theory which was later developed in the tradition. And this is how the tradition itself rationalized the ever persistent caste-hierarchy in Hinduism: the birth of each individual is predetermined by his or her karma in previous births. Such rationalizations, however, are not often seen as very rational from our modern point of view (they often amounted to apologias). But, the point that I wish to insist on is that, search for a rational basis was considered, at least implicitly, necessary for supporting an existing moral principle, and the caste- dharma unquestionably had a moral dimension since it legitimized social inequality. It should be mentioned here that, while people like Louis Dumont 'valorized' the Hindu caste-system and considered it to be a "value,"7 Max Weber hailed the linkage between caste and karma as a pure product of "rational ethical thought," even quoting from The Communist Manifesto, for he was struck by the obvious similarity in logic: The pious Hindus of low caste were in the same situation as the proletariat; they had nothing to lose, and it was open to them to climb up the ladder of caste-hierarchy gradually, and even to become a god in future life, through good deeds in this one. Although I hesitate to share Weber's enthusiasm regarding the implication of the karma doctrine, it cannot be denied that the doctrine did have that sort of significance. I also disagree with Dumont when he considers the Hindu castehierarchy as a "value" and also a "rational" practice. 8 Another illustration of the linkage between dharma-ethics

196 B. K. Matilal and the search for a rational basis may be in order here. I shall refer to another ancient text, the Chandogya Upani$ad, for another story. Satyakama grew up with his mother Jabala, and he wanted to have an education. He approached the well-known teacher of his time, Sage Gautama, for initiation and to join the group of his pupils. For the initiation ritual, Gautama wanted to know the name of Satyakama's father or his gotra, i.e., his family name. Unfortunately, Satyakama did not know who his father was. He said that he would come back after asking his mother. But Jabala had been a maid who had had to sell her body in order to survive. Thus, she did not know the name of Satyakama's father. Jabala told this to Satyakama, who went back to teacher Gautama and told the truth about his birth in the presence of all the other young pupils. There was a ripple of suppressed laughter out of contempt from the assembled pupils. But Gautama had to make a decision — a moral decision. He got up, embraced the boy, and announced: "Now I have no doubt that you belong to the highest caste, that you are a brahmin, for such courage, firmness and truthfulness can only be the constitutive qualities of a brahmin. I would accept your mother's name as your family name. You will henceforth be called Satyakama Jabala. Come, I will initiate you."9 Jabala Satyakama became a famous Upanishadic sage. Here, again, a moral decision was made on the basis of a rational argument. Caste-dharmct does not always depend upon birth; possession of moral virtues should also be a criterion for one's claims to supremacy. The above is reminiscent of an episode in the epic, the Mahabharata, where a similar problem of dharma was posed, and the decision or rational preference was made on similar grounds. I shall refer to it only briefly, as I have already discussed it elsewhere. 10 In the Vanaparva of the epic, a discourse on dharma took place between Yudhisthira and King Nahusa who, under curse, took the form of a huge python. (As I have already noted, such discussions on dharma are frequently found in the epic literature, which only showed the arguability of any dharma-preference.) When asked to define brahminhood, Yudhisthira said emphatically that what constituted brahminhood was not birth but a collection of moral virtues such as truthfulness, generosity, forgiveness, goodness, kindness. For, each birth was accidental due to the copulation of a man and a woman out of lust [raga), over which the person born had little

Dharma ana Kationaiuy

i»/

control. 11 This, too, shows the ambivalence of the later tradition to accept as entirely rational the prevalent resolution of castehierarchy in terms of the past karma of the individual. In the Bhagavad-GIta,12 Lord Krishna says, "I have created the four Varnas (castes) in reliance upon the division of qualities and actions." This can be read more as a critique of the birth-based division of caste rather than as an endorsement of it. 13

Dharma, Its Non-Theistic Basis and Moral Conscience I have said above that, in the Indian dharma theory of ethics, authority was seldom ascribed to God. Let us now turn briefly to the Dharmas'astras to examine their version of what the authorities had to say about dharma. I shall use the most well-known Mdnava-dharmaSdstra or Manu-samhita14 for reference. Verse 1 of chapter 2 says: Listen [my pupils], I shall describe dharma — that is always honored by the honest and the wise [or the learned]; it is followed by those who are above attachment [greed] and aversion [hatred], and it is approved by their heart. ' Different commentators have given varying interpretations of some of the epithets, but all have agreed that they constitute a general definition of dharma (Kulluka, e.g., dharma-samanyalaksana). In those days, any learned' person would have to be, among other things, an expert in Vedic studies. Hence, the commentators, being eager to connect the basis of dharma with the Vedas in some way or other, interpreted vidvan, or "the learned," as vedavit, or "versed in the Vedas." However, the later (commentorial) tradition was aware already that everything enjoined in the Vedas was not a moral duty or a dharma. Commenting upon this verse, Kulluka, for example, quoted from the Mimamsa-sutra 1.1.2 of Jaimini, where dharma was defined as "Codand-laksano'rtho dharmah."15 But he noted immediately that, although codana means Vedic injunctive statements by which the dharma duties would be signified [laksanah laksyate anena), there is the word "arthah," which means Sreyah-sadhanam, or "that which is conducive to good." Hence Kulluka 16 explains:

198 B. K. Matilal Both are meant by "injunction in the Vedas" [codana]: good acts that bring about good, e.g., the jyoti$toma ritual, and evil acts [an-artha] that bring about evil or moral lapses [pratyavaya], e.g., Syena, or ritual. Hence the meaning of the [Jaimini's] sutra is this: only such Vedic injunctions signify dharma as is conducive to good, such as the jyoti$toma ritual. It should be noted that, although they were enjoined in the Vedas, since rituals such as Syena or abhicara were meant for harming others, it was immoral to perform them. This is another piece of evidence on how a rational critique developed within the tradition itself to separate the morally unjustifiable (scriptural) injunctions or action-guides from those that were morally justifiable or even morally neutral, e.g., the quest for personal good such as perfect bliss, without harming others. This last type of action-guide may, therefore, be put among those that were rationally justifiable. The other qualifications in the verse of the Manusamhita quoted above are somewhat self-explanatory. For our point of view, it is significant to note that dharma-ethics, or the dharmaprescriptions cannot be based upon personal greed or hatred. Impartiality is an essential ingredient in the constitution of dharma, as it should be in any viable theory of rational behavior. Dharma-morality can hardly be assessed from a self-centered or emotionally biased, partisan point of view. We need the viewpoint of an impartial observer, just as equity and justice is commonly symbolized by a lady blindfolded and holding a scale. The last qualification in the verse quoted is also important. It enlightens us about the human side of dharma-morality. Whatever action is prescribed by dharma must also meet the approval of the heart of the honest and the wise; hypocrisy can never be a part of the dharma behavior. Traditional commentators, however, suggest several explanations. Kulluka thinks that the prescribed action must lead to something good (Sreyas), for then it will be the natural inclination of the mind to perform such acts. Govindaraja interprets it as an absence of any doubt in the mind of the learned with regard to a particular action. Kulluka gives a taunting reply in rejecting this interpretation. He says that, according to this interpretation, if a Vedic Scholar decided to travel to the countryside and had no doubt in his mind for reaching such a decision, then that would also be called a

Dharma and Rationality

199

dharma — an absurd consequence. Medhatithi first agreed that it meant "approval of the mind," but then suggested a queer alternative: he interpreted "heart" (hrdaya) as the Vedas (for Vedas had to be learned by heart), and then — in an awkward manner — claimed that the phrase meant that dharma was approved by the Vedas. This was somewhat embarrassing to the commentators, for the verse, while giving an almost impeccable definition of dharma, did not make any explicit reference to the scriptures or the Vedas. Hence various exegetical devices were used to establish the connection between the Vedas and dharma. (It may be that the definition was intended to be general enough to be acceptable to the Vedic and the non-Vedic people.) Medhatithi's alternative explanation, however, was too far-fetched to be true. As I have already explained, I believe that this last epithet was intended to establish the vital connection between dharma and the role of moral emotions, as well as the moral inclination of the person in determining the dharma. In this way, it supplements the impartiality criterion (see above) of rationality with the requirement of sympathetic understanding of the situation under which a moral decision is taken. We may pursue this point a little further. As I will note presently, it was widely recognized that there was a strong connection between the dictates of moral conscience and dharma. Life presents us with many moral conflicts, some of which even amount to genuine moral dilemmas. In genuine moral dilemmas, rational arguments in favor of either prescriptions or actions (where both cannot be done, for the doing of one is the undoing of the other) are equally balanced. Hence if the rational agent is forced to take action, it is usually under unresolved conflict, and the agent may suffer from such emotions as regret and remorse. The Indian epics, especially the Mahabharata, supply numerous illustrations of such genuine moral, or dharma, dilemmas, as I have argued elsewhere. 17 Some moral conflicts may be resolvable, but due to lack of enough information or informational constraints in the situation, a rationally arrived at decision may seem difficult or even impossible. But, since a decision has to be taken, for the situation demands it (and human beings cannot, like Buridan's ass, show an asinine preference for death by starvation over reaching a decision under unresolved conflict and picking either haystack, not neither), the agent may appeal to his own moral conscience, being impartial and not biased by any

200 B. K. Matilal baser emotions such as greed, and his mental inclination will reveal his preference. I believe many moral (as well as other) conflicts are resolved in this way, through the 'heart's approval,' and this is exactly what is implied in the last qualification found in the general definition of dharma in the verse quoted from the Manusamhita. It is easy to support the above interpretation by referring to the original text itself. In verse 6 of the same chapter of the Mamisamhita, five authorities on dharma axe cited (where, by the way, Manu did not forget to mention the Vedas): The roots of dharma are (1) the entire Vedas, (2) the DharmaSastras as well as the (3) virtues cultivated by the Vedic scholars, (4) the good conduct of the honest, and (5) satisfaction of the mind (of the agent). Six verses later, in verse 12, the same idea is repeated; this time, however, four authorities are mentioned instead of five, for, as the commentaries explain, the virtues cultivated by the Vedic scholars and the good conduct of the honest are merged into a single authority: The following, they say, are the direct characteristics of dharma: (1) the Scriptures, (2) the Dharma&astras, (3) the conduct of the good, and (4) satisfaction of the mind. In both lists, our interest lies, of course, in the last item, which is similar to the "heart's approval" of the previous general definition in verse 1. In his commentary on verse 6, Kulluka says clearly that this authority is appealed to where a conflict of dharma duties arises. He even quotes a line in support from Garga, another author of DharmaSastras: Satisfaction of the mind is the only authority in the cases of conflicting alternatives. Thus, it is clear that the tradition accepted several other authorities on dharma-morality besides the scriptures and the DharmaSastras. This openness bespeaks of the rational stream of the tradition as well as the lesser importance accorded to blind faith. Appeal to the mental inclination of the generally upright person in conflict-situations that are not rationally resolvable due to informational or other constraints, should not be confused

Dharma and Rationality

201

with an appeal to inherent bias or blind faith. In his play, Abhxjnana- Sakuntala,18 the poet Kalidasa put a similar argument in the mouth of his hero, King Dusyanta: In matters where doubt intervenes, the [natural] inclination of the heart of the good persons becomes the pramana, "authority" or the decisive factor.

Dharma and Skepticism I have said that dharma was a subject of rational inquiry, not a matter of blind faith, except in the case of the Hindus, where the Vedas were given the supreme authority. But even these Vedas were subjected to rational investigation (cf. mlmamsa). Two further points need to be made here. First, the Vedic injunctions cover only a very small part of our normal behavior at the social and personal level. Hence the necessity arose for guidelines — from the conduct of the good as well as from appeal to good conscience — to achieve rational resolution of conflict-situations. The second point is that medieval authors of the Dharmagastras such as Manu and Yqjnavalkya were fully aware of the role of rationality in determining various moral or dharma preferences. This may sound a bit odd to some of us moderners, since these authors achieved notoriety for their narrow mindedness, for 'irrational' rationalization of the same existing unjust social institutions such as inequalities in caste discrimination, and for resisting the change that was bound to come along with the change of time and environment. However, it is undeniable that these same authors also realized that the full extent of dharmamorality can be sustained only if it can be given a rational basis. Besides, the stream of critical rationality in the tradition was already alive and active. There were dissident voices not only from the low castes but also from the higher strata of the society, the brahmin priests, etc. We hear frequent stories in the epics and the puranas of a certain Carvaka — an adherent of a philosophy skeptical of religious tradition — who, having entered the assembly of the brahmin priests (he was sometimes described also as a brahmin), used to ask questions and challenge the validity of the Vedic rituals, using tarka or hetuSastra, "science of reasoning." Some of his questions were so radical that satisfactory answers were hard to find. The episode usually ended in turning him away

202 B. K. Matilal from the assembly. He probably represented the radical rebels of those days. In the Ramayana we read the story of Sage Jabali, who was one of the brahmin-advisers of King Dasaratha. He came to see Rama in exile along with Bharata and all others from Ayodhya, and tried to persuade Rama to return to his kingdom, for he claimed that it was a false dharma (according to Jabali) to abandon his kingdom in order to keep some old promise that his dead father had made to Kaikeyi. The sage expounded elaborately upon a philosophy of materialism and a hedonistic ethic in support of his argument. Of course, Rama rejected the advice, but that is not the point here. Even traditional authors of DharmaSastras occasionally referred to such a hedonistic ethic as well as to the arguments of those who challenged the authority of the Vedas. Manu, for example, recognized the persistence of such brahmins who used to condemn the Vedic rituals by using reasoning, when he said (verse 11, chapter 2): Such a "twice-born" [= brahmin, dvija] person as would condemn the root of dharma [i.e., the Vedas], having recourse to the "science of reasoning" [hetu-Scistra], should be turned away by the good people,19 for he vilifies the Vedas and is a "Negativist" [nastlka]. The commentators pointed out that logic, or the science of reasoning, can be used in two ways: (1) to support and give a rational explanation of the Vedas, and (2) to reject it and thereby overthrow its authority. Obviously, it was claimed that the first was 'right' or 'the good way,' which was acceptable to followers of the dharma-ethics, and the second was 'bad' and hence unacceptable. Towards the end of the ManavadharmaSastra, the rational basis of dharma was again emphasized. Chapter 12, verse 105, says in unambiguous language: The person who wishes to 'purify' the dharma-ethics must be very well-acquainted with such pramanas, "means of knowledge" as perception, inference and various Sastras and texts [sources of verbal testimony]. The commentators noted that the three well-known pramanas — perception, inference and verbal testimony — are mentioned here as essential requirements for determining any controversy regarding dharma. The next verse (106) makes it clearer:

Dharma and Rationality

203

Only he comes to know the dharma — he who examines in the light of reason {tarka} any instruction of dharma by the sages, and examines them so as not to contradict the Vedas — not anybody else. In the case of dispute over dharma, where it is not easy to decide which course of action should be followed, or by which action dharma-morality would be sustained, Manu suggested another method which undoubtedly foreshadowed the upholding of the 'rational-democratic' principle. In verse 110 of chapter 12, it was said: ["What is to be done?" if such a doubt arises with regard to a conflict of dharmas where the texts such as the scriptures or the DharmaSastra have not laid down anything — verse 108 supplies this context] an assembly of not less than ten persons, or [if ten are not available] not less than three persons, should deliberate and reach a decision on dharma, and that dharma [thus arrived at] should not be transgressed. In verse 111, a selection procedure for these ten members was given: The ten-member assembly will be constituted by three scholars versed in the three Vedas [Rg, Yajur, and Saman], one logician [haituka, one versed in hetuvidya], one dialectician or arguer [tarkl, one versed in tarka 'dialectics' or hypothetical reasoning], one expert in semantics and etymologies {nairuktdj, one scholar of the DharmaSastras, and three laymen from three different groups, one celibate student [probably a young man studying under a teacher], one house-holder [a married man or a man with a family] and one retired person [a considerably senior man who has retired into forest after leading a full family life]. This seems to be a good combination of people, whose combined wisdom will usually be an effective way to decide the dharma democratically in matters of dispute. Both terms, haituka and tarkl, may stand for "logician." Obviously, they were talking about two types of "logician," one who was expert in the science of evidence, or pramanaSastra, and the other who was expert in the prasanga or reductio type of argument. Kulluka, however, took one to mean a Naiyayika, versed in Nyaya logic, and the other a

204 B. K. Matilal Mimamsaka, versed in the Mimamsa-type cogitation. My own interpretation is, however, not very different. Dharma and Moral Weakness Admission of moral conflicts or genuine moral dilemmas (or dharma-dilemmas; they are given various names: darmasamkata, dharma-vikalpa, Krtyakrtya-viveka-nirnaya, kimkartauya-vimudhata) requires using some method for making a rational choice. It is obvious that some sort of pre-ordering or ranking of principles helps such rational deliberation. In matters of ritual-orientated dharmas, when conflict arises, the Mimamsa school has determined a fixed rule of pre-ordering, and has given a rational argument in favor of such ordering. Unfortunately, in all practical cases of value conflict or ordinary dharma-orientated conflict, it is extremely difficult to establish priorities in the same way. Many epic stories that illustrated such practical dharma-conflicts show that the practical resolution of such conflict does not always fix priorities according to the same pattern. It appears to me that this respect for the difficulties encountered in real life is not a mark of irrationality or inconsistency, but emphasizes that we sometimes face moral predicaments for which we cannot find a simply rational solution. One kind of moral conflict or dharma-conflict is the struggle against temptation, or what is called weakness of the will. It is typified by the oft-quoted verse that follows: I know what is dharma, but I cannot persuade myself to act accordingly. I know what is a-dharma [evil], but I am unable to refrain from it. {janami dharmam na ca me pravrttih/ janamy-adharmam na ca me nivrttih//) This type of struggle is well illustrated in the ancient epic, the Mahabharata, which tells of a struggle between two families, the Pandavas and the Kauravas. The leader of the Pandavas, Yudhisthira, was addicted to gambling, while the leader of the Kauravas, Dhrtarastra, had a blind affection for his son Duryodhana, an affection which led him to allow (by not discouraging) the latter to go to war against the P a n d a v a s .

Dharma and Rationality

205

Yudhisthira was called Dharmaraja, "the King of Dharma," for his righteousness and moral behavior. But his addiction to gambling was almost proverbial — a fatal flaw in his character. He lost everything, his kingdom, his four brothers and wife (and his wife was publicly humiliated while he himself was present), at the first of his gambling matches. Rescued from this situation, when the second invitation (or challenge) for gambling came, he had the option to refuse. Now, in those days, gambling was explicitly recognized as a vice among the princes, and it was not unknown to Yudhisthira that it was an a-dharma. Besides, he had the fresh experience of the humiliation and shame after his first defeat — not to speak of the immorality of placing his beloved and innocent wife, as well as his four brothers, in a morally unbearable situation of shame. But he was like the celebrated gambler of the Rg-Veda.20 This proverbial gambler said (verse 5, The Gambler's Hymn, 10th Mandala): Vainly I decide not to go gambling even when all my [gambling] friends left, but at that very moment I listen to noise made by the throw of dice, and then I have to rush to reach there just as a fallen woman runs [to meet her paramor]. Yudhisthira's behavior was in no way different. The second invitation came, and he ran to accept it. The temptation was great. Of course, he gave a reason in favor of this: as a prince, he must accept the so-called challenge. But a man under temptation can always argue himself into finding reasons that support his action. Thus, a bank-employee about to embezzle funds may indeed find reasons for his action (e.g., to counteract the injustice done by the capitalistic system to poor and middle-class people like himself). And there was no moral (or dharma-centered) obligation on the part of Yudhisthira in this case, as he himself admitted much later on, in the Vanaparva of the Mahabharata. In the Vanaparva, Bhlma once raised the question: "What is the use of our gaining the kingdom back even though we would have to fight for it? For, my dear brother, I know you well, and I believe you will again be tempted to gamble away everything if the challenge comes for the third or the fourth time, and so on." To this blunt accusation of Bhlma, Yudhisthira gave a significant reply: Yes, he would. He admitted that he would again be unable to check his temptation, for in this matter he had no control over himself!

206 B. K. Matilal There was a challenge all right — a challenge of a different kind — in the case of Yudhisthira. The challenge to him was to do what he himself (as Dharmaraja) recognized he ought to have done when desire, fear, temptation and the irrational hope that, this time he might win after all, all inclined him to do the opposite. There was not the slightest doubt about what he ought to have done. D h r t a r a s t r a ' s case was similar. He knew that his son Doryodhana was doing something completely immoral. Duryodhana's conspiracy to kill the Pandavas by setting fire to their house while they were asleep was quite serious, although even this might be ignored as an instance of the young prince's foolishness. Dhrtarastra always looked the other way. He never tried to control his son or teach him any moral lesson. The father knew very well what his son had been doing, and he did not lack knowledge of the dharma or moral understanding of the situation. Each time he felt the pangs of his conscience, he would call upon Vidura to explain the dharma-ethics to his son. But he seldom listened to Vidura's advice. Anyway, in the two cases of gambling described, Dhrtarastra made restitution to the Pandavas. He returned all their gambling losses to them. He also gave Yudhisthira Pandu's share of his kingdom after he learned that the Pandavas and their mother were alive and in exile. But Duryodhana could not reconcile himself to this outcome. The last act of Duryodhana was unpardonable. According to the conditions of the gambling match, he should have kept his part of the bargain and returned the Pandavas' share of the kingdom. But he wanted an all-out war with the Pandavas. Although there were several attempts to negotiate a deal in order to prevent a most devastating war between two lines of the same royal family, Dhrtarastra could not do anything to persuade his son to listen to the voice of reason. He just let it happen. His blind affection for his son led him to ignore the advice of all well-meaning persons. It was not until almost the end of the war, after almost everybody was killed except for ten warriors (seven on the side of the Pandavas and three of the Kauravas), that Dhrtarastra admitted his weakness and regretted it. Philosophically speaking, this moral weakness or weakness of the will may require some explanation, for it may be argued that it is impossible for rational beings to knowingly do wrong. In fact, Plato made such a claim in the Protagoras.21 And, in

Dharma and Rationality

207

Ethica Nicomachea, Aristotle raised the question again, to discuss it: "Now we may ask how a man who judges rightly can behave incontinently." 22 Aristotle argued that the Socratic (Plato's) claim that there is no such thing as "incontinence," and that people act so only by reason of ignorance, was indeed a puzzle that needed some philosophical explanation. For, according to Aristotle, this view plainly contradicted the observed facts. Of the several explanations suggested in Aristotle, one may be stated as follows: It is observed that a person behaves incontinently with knowledge, but perhaps, here "knowledge" is used in a slightly different sense — "for both the man who has 'knowledge' but is not using it and he who is using it are said to know."23 Thus, one might say that the incontinent person has "knowledge," but is not exercising it. Some philosophers put the matter more strongly. Human beings (and not lower animals, as Aristotle reminded us), 24 are able to, and often do, act against their sincerely held moral principles, with full knowledge and deliberation, and this is simply a fact about ourselves. As I have already noted, the predominant view in classical India is that people do, in fact, act against their moral convictions. Another point should be made before we leave the subject. In the New Testament, (Romans 7), our weakness of the will is seen as a consequence of our sin, and hence it is not philosophically puzzling. Indeed, the following quotation sounds almost similar to the Sanskrit verse I have quoted a few pages earlier: For I do not do the good I want, but the evil I do not want is what I do. Now if I do what I do not want, it is no longer I that do it, but sin which dwells within me. 25 Here, the first part of the sentence coincides with the meaning of the Sanskrit verse. But the attitude the sentence conveys is not prevalent in Indian tradition. Let me give an example. When the author of the Mahabharata dwelt upon the similar cases of Yudhisthira and Dhrtarastra, he did not refer to the question of sin. Also surprisingly, there was no explicit mention of the karma doctrine in this connection, to explain the obvious puzzle. Both Yudhisthira and Dhrtarastra were fully aware of the weakness of their own character and occasionally felt bewildered by it. It may be contended that, according to the classical Indian wisdom, weakness of the will is part of human nature

208 B. K. Matilal (svabhava evaisa bhutanam), and is not ascribed to sin. Here, we can refer back to the Manusarnhita for an insight: It is not a vice to eat meat, to drink liquor, or to have sexual intercourse, for they constitute the natural inclination of the creatures; but [judicious] refraining from such acts generates good consequences.26 Kulluka explains pravrtti as the "natural dharma." This, I believe, illuminates a great deal about the attitude of the classical Indians toward the issue. Philosophical Issues in Dharma Dilemmas We may now turn to the other kind of moral conflict: genuine moral dilemmas. Although several well-known philosophers today may disagree here, I believe that, in cases of moral dilemmas, weakness of the will does not have a large part to play in the philosophical analyses of such situations. Genuine dilemma arises when what ought to be done, all things considered, is as yet unsettled or even unsettlable. It may be argued that if the informational constraints are removed and impartial rationality is allowed its full play, then such dilemmas will never arise, provided the agent is not suffering from akrasia, or weakness of the will. As opposed to this, one may hold that there may be equally strong and equally admired moral principles which prescribe actions that are in conflict, so that the doing of one is the undoing of the other, so that there is no third choice, for example, the complete withdrawal from action. This, then, would be a case of genuine dilemma, where, we may add, decisions are no doubt made, but such decisions are ad hoc and not arrived at fully rationally. To go back to ancient examples: in the beginning of the Bhagavad-GIta, Arjuna faced a similar dilemma when he was torn between two dharma principles: the dharma of his Ksatriya caste to fight on, and his family dharma infused with such adorable moral sentiments as love, devotion and respect for his grandfather Bhisma, and adorable teacher Drona. Arjuna himself suggested the third way out — that of complete withdrawal from action — as he said, "I would rather accept the life of a mendicant" (cf. bhaiksyam apiha loke). But it was more than obvious that it was decidedly too late for such a course of action to be

Dharma and Rationality

209

undertaken. At Krishna's advice, Arjuna decided to fight, but still, he regretted it many times in the later part of the epic. Similarly, Yudhisthira was in moral dilemma on several occasions. One was when he had to decide between telling a lie (only once in his whole life) and thereby winning the war which he morally deserved to win, and not being untruthful (even for once in his life) and thereby suffering defeat and humiliation by letting Duryodhana win, even though the latter choice would have meant that overall justice would not have been maintained. On the one hand, there was the strong moral principle of telling the truth throughout his life, and on the other, was the equally strong principle of justice — justice against the utter humiliation of innocent Draupadl. Yudhisthira was torn between the two. At the persuasion of Krishna, and with great regret and reluctance, Yudhisthira decided to tell a lie. How did the tradition look upon such cases? I believe there is strong evidence that it regarded them as genuine dilemmas. Both Arjuna and Yudhisthira suffered from the moral emotions of guilt and shame for violating the dharma principles that they did violate. Both of them held the principles they violated very dear to themselves. Tradition believed that one met moral retribution if one violated any important moral principle. The author of the Mahabharata describes how Yudhisthira immediately lost the privilege he had enjoyed for upholding the high moral principle of truthfulness throughout his life. (During war, his chariot used to travel a few inches above the ground due to his merit or excellence in dharma. This was only one lapse; however, a lapse was a lapse and he had to pay for it. Thus, his chariot symbolically came down to the level of all the rest.) Even Arjuna was greatly humiliated toward the end. He lost his prized possession, the invincible Gandiva bow, which had been a gift from Agni. When the war was finally over, his famous chariot was reduced to ashes as soon as Krishna dismounted. Besides, toward the end (when Krishna died), this great hero of the famous battle of Kuruksetra suffered a crushing defeat and humiliation at the hands of ordinary tribal warriors and robbers, who kidnapped the women of the Yadava clan who were under his protection. When the battle of Kuruksetra was over, both Arjuna and Yudhisthira regretted the loss of millions of lives and doubted whether the throne had been worth fighting for. After all, what kind of kingdom was it in which only old people, widows and children were alive? Whatever

210 B. K. Matilal moral decision the moral agent might have taken in the case of genuine dilemmas, the violation of the conflicting principle was regarded as a violation by the tradition, and hence the talk about retribution seemed justified. >

Dharma and Moral Emotions I wish to regard the above cases as genuine dilemmas for a slightly different reason. I believe that, in the case of genuine dilemma, the agent has to suffer from certain appropriate moral emotions, such as regret and remorse, guilt and shame, since he cannot rationally justify his preference for one principle and the resultant violation of the other. In the epic stories, both of our moral heroes suffered from such appropriate moral emotions. Regret is not enough; the morally sensitive agent usually feels remorse as well. Both guilt and shame are appropriate feelings for the morally alert agent. The by-now-well-known Rawlsian distinction between guilt and shame may be mentioned here to make a minor point. 27 According to Rawls, guilt invokes the concept of right, while shame appeals to the concept of goodness. Guilt can be relieved by reparation, it "permits reconciliation" through forgiveness. Shame invokes aspiration and ideals, that is, certain forms of moral excellence the agent wishes to attain. In the case of our epic characters, they felt both the appropriate guilt and shame, although, as the texts testify, at some times, it was more guilt than shame and at others it was the other way around. For example, when Arjuna had to violate his principle of promise-keeping and refrain from killing his own venerable elder brother, Yudhisthira, he experienced more a sense of guilt than shame. 28 Hence when Krishna suggested a way for him to relieve himself of the guilt by reparation, Arjuna readily followed his advice. Rama, in the Ramayana, also felt more guilt than shame over his controversial acts, such as the abandonment of Slta. But Yudhisthira was a different sort of person. After lying (to Drona), he was overwhelmed with more shame than guilt; after the battle was over his self-shame, contempt and derision knew no bounds. He aspired to a sort of moral excellence and had the same ideals as Dhrtarastra, that is, the King of Dharma. But the war was finally won by not so glorious means, and, as Krishna made clear to him, it had to be done that way, for there

Dharma and Rationality

211

was no other possible alternative. Bhavabhuti's Rama, however (in his Uttara-Ramacarita29), was more like Yudhisthira, a tragic hero, than the original portrait of him by Valmlki. Dharma, Its Vulnerability and Elusiveness The epic stories presented the dharma morality in a novel light. They were dealing with practical problems of everyday life. I claim that sometimes there was more realism in these old epic stories than they are given credit for today. They underlined the two most prominent aspects of dharma: the vulnerability of moral virtues and the ever-elusive nature of truth in the moral domain. I shall conclude with another story from the Mahabharata. Yudhisthira had to participate in dharma discourses several times, under various odd situations. The following was one of them. On the last day of their exile, the Pandavas found themselves all of a sudden in a very strange situation. A brahmin-priest came and informed them that a deer had run away along with the two sacrificial sticks, used to light the sacrificial fire, stuck into its horns. The ritual would be stopped if the deer was not found in time, and dharma would be violated. Hence it was the moral duty [dharma) of the Pandavas, of the warrior caste (the Ksatriyas), to find the deer and recover the sticks. It was indeed a very small matter for warriors like Arjuna or Bhima to recover the sticks. But as luck would have it, the deer had disappeared. The Pandavas searched and left no stone unturned, but they only got exhausted, showing that, even with the best of intentions and sincerest of efforts, people may be unable to obey the dharma. This was the case with the Pandavas. They got thirsty in the middle of the forest, and Nakula (the fourth brother) observed from a treetop that there was a lake nearby. Each of the four brothers was sent by Yudhisthira to fetch water, one after another. But nobody returned. Hence Yudhisthira finally went near the lake to see what the matter was. He saw all his four brothers dead near the lakeside. As he was descending to get some water, a voice spoke: "Please do not be rash, O Prince. First you must answer my question and then touch the water. Your brothers did not listen and you can see what happened to them. The lake belongs to me." In utter surprise, Yudhisthira asked: "May I know who

212 B. K. Matilal you are?" "I am Yaksa," was the answer. Yudhisthira agreed and a session of question-and-answer followed. Of the many tricky questions asked by the Yaksa, the most important was: What is the way (to reach a decision about the dharma)? Yudhisthira's answer satisfied Yaksa, who then divulged his identity. Although he appeared in the form of a stork there, he said that he was in fact the Dharma, Yudhisthira's real father, and that he had stolen the sacrificial sticks of the brahmin in the form of a deer in order to teach Yudhisthira a lesson in dharma. This story illustrates the ever-elusive nature of dharma: One may fail in spite of everything, so no one should take pride in being a moral hero. Let me comment upon the verse cited by Yudhisthira in reply to Yaksa's question. Yudhisthira said: There are different Vedas, even the DharmaSastras vary from one another. There is not a single muni [teachersage] whose view is not different [from that of other teachers]. The truth of dharma lies hidden in the [dark] cave. [But] the way [leading to dharma] is the one that the mahqjana had followed. I have left the crucial term mahqjana untranslated on purpose. One (comparatively later) meaning of the term is "a great person," and plurality in such cases is implicit in the grammatically singular expression. Hence the meaning would be: the path taken by great men is the path in matters of dharma conflict. If this interpretation is accepted, then it is simply a reference to 'the conduct of the good people' as an authority on, i.e., a determining factor of, dharma, which was mentioned in the Manusamhita. I believe, however, that the second meaning of the term, "a great number of people," has a wider significance in the context. This may be a statement of a primitive form of a moral theory: The path is that one wherein lies the good for the maximum number of people. It seems to be a primitive proto-Utilitarian view, which need not be surprising; for a sort of primitive proto-Utilitarianism seems to be implied by the rational side of any ancient (and well-developed) civilization. Besides, the public rituals in classical India used to be prefixed by a benediction ritual-recitation (mangala), where the following phrase was commonly used: for the sake of happiness of many people, and for the sake of the good of many, (bahujana-sukhdya bahujanahitdya ca)

Dharma and Rationality

213

Pluralism and Dharma The more interesting part of the above Mahabharata verse is, however, the pluralistic ideals that it insists upon. For, the demand for rational decision is always greater when pluralism is admitted. Pluralism has several senses, some of them rather specific or technical in nature. In this context, we may take it in one specific sense in order to make some concluding comments. When pluralism is applied to a moral theory, it may offer the counterpoint to a monistic theory like Utilitarianism, which, even in some of its modern versions, tries to resolve all moral issues by relying upon one ultimate, uniform criterion: utility. A theory that refuses to reduce all judgement of (moral) preference to quantitative form in a single dimension (so that we can calculate), and that allows for diversity of (moral) goods which are sometimes incommensurable, is pluralistic. Pluralism raises several important questions. One is: How are we supposed to combine, in our lives, two or three or more conflicting goals, or virtues or principles, which we feel in our bones that we cannot repudiate? Conflict arises because they seem to demand 'incompatible' actions. If the conflict is only apparent (as it sometimes may be), then rational thinking will help resolve the issue. If, however, the conflict is real, we have cases of genuine dilemma, as I have already noted. Dharma-morality is pluralistic. Hence we must also face the other questions raised by Pluralism, which concern rationality in its formal aspects. The allure of Utilitarianism is that, among other things, it offers a neat model of maximization of a single homogeneous magnitude, and hence it is hoped that the formal criterion of consistency and completeness may be achievable here. Thus, compared to pluralistic theories, it seems to be the 'rational' moral theory par excellence. Pluralism, however, does not necessarily lead to irrationality. Even consistency seems achievable if we recognize the need for finding the consistent ordering of priorities in a pluralistic theory. Two offending principles may be put under strict logical scrutiny with regard to a particular situation, so as to discover whether one can be allowed to override the other. The cases of dilemma have already been discussed, where practical wisdom has to bear the occasional burden of moral emotions such as shame or remorse. They might signal the limits to the formal notion of consistency in a moral theory,

214 B. K. Matilal without necessarily rendering the theory inconsistent. As far as the requirement of completeness is concerned, I believe this may not be necessary in order to keep a pluralistic theory within the bounds of rationality. 30

Conclusion Dharma does not have a definitive form. It has an ever-elusive nature that has been well illustrated in the story from the Mahabharata. It is also open-ended and rational. Dharma does not rule, but (as Robert Lingat once put it), it reigns from above. It is a going concern of the society as well as the individual. It demands the best from our practical wisdom. I have tried to lay down the insights of the ancient writers of India with respect to dharma. In many ways, we must admit, their formulation of issues are primitive and dated. But the stories they have told sometimes have far-reaching significance and reveal their wisdom in a new light. Today's moral philosophy has become increasingly technical. One needs to talk about the decision-theoretic procedures, the Arrow-Sen impossibility theorem, as well as the Pareto optimality. In this background, exploration of the writings of the ancient Indians may not be relevant. I am, however, interested in cultural history, and I believe the historical understanding of the concept of dharma has some relevance today. I believe it is a widely misunderstood concept in the modern study of the history of Indian philosophy. I have tried to clear up part of this misunderstanding. Explanation of the traditional ethos of India has always been somewhat controversial among the Indianists of today. The sociologists or social anthropologists propagate one way of looking at it. The development economists favor another way of taking it. 31 Both, however, assume that to understand modern India some basic knowledge of classical India is absolutely necessary. My exposition of dharma and rationality has been partly aimed at this enterprise. Besides — and last of all — what has often attracted me to such a study is that, although the ancients did not always seem right from our modern point of view, what is surprising is that they also often got it right. Today they seem to us to have been mistaken in many ways, but that they did sometimes hit upon the right note is worthy of our notice and praise.

Dharma and Rationality

215

Notes 1. B.K. Matilal, "In Defence of Devious Krsna," in an anthology ed. by A. Sharma, in press. 2. M. Weber, The Religion of India (Glencoe: The Free Press, 1958), p. 144. 3. Ibid., p. 121. 4. Brhadaranyaka Upanisad 3.1, in The Principal Upanisads with SamkaraBhasya (Delhi: Matilal Banarsidass, 1964). 5. Ibid., 3.2. 6. Ibid., 3.2.13. 7. L. Dumont, Homo Hierarchicus: The Caste System and its Implications, Complete Revised English Edition (Chicago: The University of Chicago Press, 1980). 8. B.K. Matilal, "Images of India: Problems and Perceptions," in JV.V. Banerjee Commemoration Volume, ed. M. Chatterjee, in press. 9. Chandogya Upanisad 4.4, in The Principal Upanisads with Samkara Bhdsya (Delhi: Matilal Banarsidass, 1964). 10. B.K. Matilal, "Moral Dilemmas in the Mahabharata," in Moral Dilemmas and the Mahabharata, eds. B.K. Matilal and M. Chatterjee, in press. 11. See my "Images of India," ibid, (note 8). 12. The Mahabharata, Critical edition (different volumes by various editors), (Poona: Bhanderker Oriental Research Institute, 19281970). 13. See also B.K. Matilal, "Caste, Karma and the GIta," in Indian Philosophy of Religion, ed. R.Y. Perrett (Dordrechet: Martinus Nijhoff), in press. 14. Manusamhita, with Kulluka's Commentary, ed. Narayana Rama Acarya (Bombay: Nimay Sagar Press, 1946). For reference to the commentaries of Govindaraja and Medhatithi, see also Kulluka's commentary. 15. The Mlmamsa Sutra ofJaimini, with Sabara's Commentary, ed. M. Nyayaratna (Calcutta: Asiatic Society, 1889). 16. Manusamhiia, with Kulluka's Commentary, ibid, (note 14). 17. See my "Moral Dilemmas in the Mahabharata," ibid, (note 10). 18. Kalidasa, Abhjnana-Sakuntala, ed. Narayana Rama Acarya (Bombay: Nirnay Sagar Press, 11th Edition, 1947). 19. The Ramayana, Critical Edition (Baroda: Baroda Oriental Institute, 1960-1975). 20. The Rg Veda, with various Commentaries, ed. Vishva Bandhu (Hoshiarpur: Vedic Research Institute, 1965). 21. Collected Dialogues of Plato, eds. E. Hamilton and H. Cairns (New York: Pantheon Books, 2nd printing, with corrections, 1963), Protagoras, 352b-356c. 22. The Basic Works of Aristotle, ed. R Mckeon (New York: Random House, 1941), Ethica Nicomachea, 7.2. 23. Ibid., 7.3.

216 B. K. Matilal 24. Ibid. 25. The Holy Bible, Revised Standard Edition (New York: Oxford University Press, 1962). 26. Manusamhita V.56, ibid, (note 14). 27. J. Rawls, A Theory of Justice (Cambridge: Harvard University Press, 1971). 28. For full details of this episode, see my "Moral Dilemmas in the Mahabharata," ibid, (note 10). 29. Bhavabhuti: Uttara-Ramacarita, ed. P.V. Kane (Bombay: Nimay Sagar Press, 1964). 30. For a different treatment of pluralism, see my "Pluralism, Relativism and Confrontation of Cultures," paper for the Plenary Session of the 6th East-West Philosophers' Conference, 1 August 1989, Honolulu. 31. See my "Images of India," ibid, (note 8).

Indian Philosophical Tradition: The Theory of Pramana J. N. Mohanty

I There are two ways one may critically look at a tradition: one may do so from within or from the outside, and I intend to look at tradition the first way, even though raising the sort of questions that I do implies a certain estrangement from that tradition. I do so not in order to find faults or limitations in the Indian tradition, but with a view to continuing and creatively advancing the traditional modes of thinking. Living outside the country where the tradition in question developed and still has deep roots, and having been exposed to a powerful and temporally and culturally more relevant mode of thinking, exposes one to the risk of being an over-hasty, shallow and even arrogant critic of a long and hallowed tradition. One gathers the illusion of being free — free of all tradition — and thereby justified in criticizing one's own. But if that sense of freedom is illusory, this critique will be superficial. For, if the critic claims to be free of all tradition, he will be forgetting — what Gadamer has so poignantly pointed out — that he will be thinking from within a new tradition, e.g., the tradition of (modern) rationalism. In talking about the Indian philosophical tradition, I will be talking about the Indian darSanas, or classical philosophical schools, and only indirectly referring to the scriptures from which those schools derive their ideas and motivations. This decision, justified by usage of antiquity, leaves us with a less ambiguous discourse to reflect upon, and so makes it possible to avoid many familiar pitfalls. When talking about Indian philosophy, it is not uncommon, for example, to insist that it is deeply

218 J. N. Mohanty spiritual; that its goal is not simple intellectual juggling but the spiritual transformation of one's nature; that philosophy is a means to attain mok$a, or spiritual freedom. Such large claims are, to say the least, highly misleading; in a familiar construal, they may even be false. I hope that the following remarks will partly clear the way for a more fruitful reflection on the nature of Indian philosophy. In the first place, there is no doubt that the upanishads exhibit a strong spiritual motivation: knowing the atman is said to bring about an end to worldly suffering as well as a state of mok$a (however that is interpreted). One frequent mistake is the failure to distinguish between the spirituality of the upanishads and the alleged spirituality of the darganas, even when the latter trace their ideas and doctrines back to the upanishads. Second, it is a mistake to suppose that thinking about spiritual matters is itself spiritual. This assertion is not meant to degrade such thinking, but only to reiterate its nature qua thinking. As thinking, it may be thorough or superficial, adventurous or conventional, logically rigorous or logically lax, creative or merely critical; but, it is neither spiritual nor non-spiritual. Consider an analogous point: thinking about perception is not itself to perceive. Another thing often overlooked by those who argue for the spiritual character of the darSanas is that, although at least some of them recognize Sabda, or sacred words, as a pramana or means of true knowledge, they do not eo ipso identify Sabda with experience of any sort. This matter, about which the philosophical tradition was very clear, is misconstrued by those who argue that recognizing Sabda as a pramana is tantamount to according an authoritative status to the spiritual experiences of the "seers." I shall return to this confusion below. It is enough to say here that the same sort of confusion is characteristic of such cliches as the one that Indian philosophies make use of intuition rather than intellect. And here I should like only to remind those who revel in such cliches that, quite apart from the fact that "intuition" and "intellect" have many and muddled uses, none of the darSanas uses a pramana that suffers rendering into the much misused word "intuition." Without belaboring my point any further, I shall divide the rest of my (positive) remarks into three groups: those concerning pramana; those concerning prameya, or objects of true

The Theory of Pramana

219

cognition; and those concerning the over-all status of the pramana theory, its aim and its relation to other sorts of inquiry. II A philosophical theory not only elaborates a view about the nature of things, but also needs to back up this account with a theory of evidence, rational justification, and critical appraisal. It needs not only to use evidence, rational justifications and critical appraisals, but also to have a theory of these theoretical practices. It also needs to have generalized answers to such questions as: When is a cognitive claim valid? What sorts of evidences are acceptable in adjudicating the validity of a cognitive claim? What sorts of justifications of beliefs are acceptable? What criteria are admissible in the critical appraisal of rival claims? In cases where there are conflicting criteria, what are their relative strengths and weaknesses? These are the tasks to which pramana theory addresses itself. It is a singular sign of the high level of intellectual sophistication of the dar&anas that, at some time in the course of their development, all of them came up with their own theories of pramana . As is rather well known, these theories differed not only with regard to their definitions of pramana (and the implied concept of prama, or true cognition), but also with regard to the number of pramanas and their specific natures. I should like to draw attention to some obvious features that emerge in these discussions, which throw some light, however dim, on the Indian concept of rationality. To begin with, let us note an important difference in locution between Western and Indian philosophical traditions, which points to deeper substantive issues. In the Western tradition, it was usual until recently to ask whether knowledge arises from reason or from experience, the rationalists and the empiricists differing in their answers. And these answers, in their various formulations, determined the course of Western philosophy. In the Sanskrit philosophical vocabulary, on the other hand, the words "reason" and "experience" have no exact synonyms, and the epistemological issue was never formulated in such general terms. The question that was, however, asked — and which is

220 J. N. Mohanty likely to be mistaken for the question formulated in the Western tradition — whether perception is the only pramana, or whether anumana, or inference, is also a pramana. And here, too, it should be pointed out that in Sanskrit "perception" is not synonymous with the Western "experience;" nor is "anumana" synonymous with "reason." As all Indian philosophical schools recognize perception as a pramana, they did not always restrict perception to sensory perception. And when they did so restrict it, it was not always to the domain of sensible qualities such as color and material objects like sticks and stones. Among things that were taken to be perceived through the senses were: the self and its qualities such as pleasure, pain, desire and cognition; universals such as redness; natural-kind essences such as cowness; and relations such as contact and inherence (of a quality in a substance; of a universal in its instances). It is clear from its very etymology that anumana differs from reason (especially that of the rationalists), in that it follows upon perception. Now, if we exclude the Buddhists, no school of Indian philosophy ascribed a 'constructive' role to inference, since inference makes known what can be known otherwise; perception is always prior. Therefore there have been no Indian rationalists. Neither perception nor inference ever pointed to any specific faculty of the mind, as "experience" and "reason" did in classical Western philosophies. The same faculties or cognitive instruments — operating in different manners — resulted in perception in one case and in inference in the other. I have belabored this point in order to caution against the temptation to view pramana theories as near-kins of Western epistemologies. Moreover, my above remarks lead to another feature of the pramana theories: A pramana is the specific cause of an irreducible type of prama. There are two different sorts of reason why a particular pramana is not recognized by a certain school. One is that the sort of cognition that it causes is just not true cognition. This is why some Buddhists would not regard inference as a pramana, for, since an inferential cognition apprehends its object as an instance of a universal rule and not in its uniqueness, the cognition that results is not true to that object's own nature. But one may give a quite different sort of reason why a putative pramana is not really one. When the Vaisesikas deny that Sabda can serve as a pramana, they do not deny that the

The Theory of Pramana

221

putative linguistically generated cognition is true; what they do insist is that it is not of an irreducible variety, that, as a matter of fact, it is reducible to inference. There are thus three claims made in pramana theory: 1. Some cognitions are true, i.e., prama. 2. Some of these true cognitions belong to a type that is irreducible to any other type. 3. True cognitions belonging to an irreducible type are caused by a unique aggregate of causal conditions. Thus, a sort of causal theory of knowledge is built into the pramana theory1: a true cognition must not only be true to its object, or artha-avyabhicari, but must also be generated in the correct manner, i.e., by the appropriate causes. Expressed in a modern philosophical style, this amounts to saying: S knows that p iff S has a cognitive state that takes the form 'p'; this cognitive state is true, and it is brought about in the correct way. This last formulation in terms of a cognitive state leads me to the third difference between Indian and Western philosophical traditions. Western thought has been torn not only by the conflicting claims of reason and experience, but also — at least since Descartes — by the distinctions between mind and matter, between the subjective and the objective, as well as by one of the offsprings of the latter: the distinction between the private and the public. In more recent philosophy this has emerged as the problem of psychologism, which has haunted epistemologists and theorists of logic to the point where they have sought to banish all reference to inner mental states from their discourse. The consequence has been pure objectivism, of both the Platonic and the physicalist types. In contrast, Indian epistemologists make unabashed use of 'mentalistic' discourse and never really worry about problems such as psychologism or private language. Although it is possible to accuse them of uncritical naivety, their sharp critical acumen indicates that the reasons for this trait have to be sought elsewhere. It is well known that, for most Indian philosophers, manas, in the sense of "mind," is a subtle form of prakrti, or matter understood as a non-conscious inner sense organ rather than a domain of private experience. Cognitions and other experiences belong to atman, and — because they are self-manifested — can be 'perceived' only by their owner. But if S alone has an inner

222 J. N. Mohanty perception of his own experience, it does not follow that others cannot know this experience through one of the pramanas other than perception. What is more, even the experiences of a particular owner can have ideal intentional contents in common with numerically distinct experiences that belong to different owners and may have occurred at different temporal locations. I have shown elsewhere 2 how, by means of this conception of 'mental episodes' one is enabled to construct a logic of cognitions with appropriate logical rules for inference. Thus, to speak about cognitive events need not arouse the spectre of psychologism. I shall now take a closer look at the causal story that pervades Indian epistemologies. From about the time of Kant, it became usual to make a sharp distinction between questions of epistemic justification {quaestio juris) and questions of causal origin (quaestio Jactis). Although a causal theory of knowledge has become very much in vogue in more recent times, the causal theories of knowledge have to be able to accommodate such justificatory concepts as logical validity and truth. And in this regard, the Indian epistemologies can serve as an useful model. As Matilal has insisted in his recent book on perception, 3 the pramanas serve as both causes of and justifications for cognitive episodes. This seems to have been made possible by the separating out of noncontroversial instances of true from non-true cognitions and, at the same time, by the search for: (1) the marks that distinguish the former from the latter, and (2) the distinctive causal conditions that produce the former but not the latter, and finally by (3) combining (1) and (2) in the definition of pramana . In the case of theories that regard truth as svatah, or intrinsic to a cognition, the causal conditions of cognitions (of a certain type) and those of 'true' cognitions (of that same type) coincide. Although in the Western tradition causal theories are regarded as reductionistic, and therefore suspect for the logician-cumepistemologist, this is not true of the Indian tradition, in which they are regarded as descriptive and compatible with the uniqueness of cognitions and their claim to truth. There are two aspects of this liberalism. For one thing, reductionist causal laws are physicalistic and oriented to prevailing physical theory, while the causal laws used by the Indian epistemologists are formulated in terms of such heterogenous elements as physical contacts, revived memories and desires to have a certain sort of knowledge, even the activation of traces of past karma and the ubiquitous

The Theory of Pramana

223

passage of time. For another thing, the causal account used by Indians is not explanatory but descriptive, for this account is formulated by adapting it to the intuitive needs of a cognitive event rather than submitting it to the constraints of an available physical theory. The constraints, if any, are those of an embracing ontological theory. I will make only two more comments on pramana theory before passing on to the prameya, i.e., to ontology. These remarks concern anumana and Sabda, in that order respectively. Much has been said in the secondary literature about the psychologistic character of the Indian theory of anumana (for it tells a story about how inferential cognition arises) and its non-formal character (for it requires an instance in which the universal major premise is satisfied). While both these characterizations are correct, they can be misleading if they are not properly understood. I have said above that psychology and logic have been reconciled in Indian thought. The theory of inference provides a good illustration of this reconciliation: The rapprochement between psychology and logic was traditionally accomplished by logicizing psychology as well as by psychologizing logic. The former is accomplished by assuming that the psychological process of reasoning conforms to the logical, any seeming deviance, as in supposedly fallacious reasoning, being due to a misconstrual of the premises; and the latter by making logic a logic of cognitions rather than of propositions. It is not that the Indian theory of anumana does not recognize formal validity. In fact a formally valid mood can be abstracted from a valid Nyaya anumana. But, since the interest lies in cognitions (and not in either sentences or in propositions), and, in anumana as a pramana, or as a source of true cognition, the merely formally valid inference (as in tarka, or counterfactuals) was left out of consideration. This brings me to Sabda as a pramana, where the true foundation and deeper roots of the Hindu tradition really lie. For the very recognition that Sabda is a means of knowledge is a unique characteristic of Indian epistemologies. While Western epistemologies recognize perception, reasoning, introspection and memory as forms of knowledge singly or in combination, and, while many Western philosophers have come to understand the decisive role that language plays in shaping our knowledge, to the best of my knowledge, no Western philosopher recognizes either

224 J. N. Mohanty language or verbal utterance, by itself, as a means of acquiring knowledge about the world. And, yet, think how much we learn simply by hearing others speak, or by reading the words of others in books — not to speak of the religious and moral beliefs we derive from perusal of the scriptures. Consequently, the Indian epistemologies not only recognized Sabda (i.e., hearing the utterances of a competent speaker) as a pramana , but regard such words as the decisive source of our cognitions of all matters that transcend the limits of possible sensory experience. To bring out some of the unique aspects of the Sabda-pramana thesis, I should first make the following points. To begin with, Sabda, as a pramana , is not a mere word, but a sentence, and not a written, but a spoken sentence, as there the spoken and heard undoubtedly takes priority over what is written: as is well known, the Indian 'scriptures' have their primary existence not in writing but in their aural form. Second, with regard to language learning, most Indian theorists emphasize imperative rather than indicative sentences, i.e., utterances that primarily — if not exclusively — give orders, suggest courses of action to be undertaken or avoided, as opposed to sentences that present statements of fact. Third, in their theory of meaning — of both words and sentences — the majority of theorists subscribe to pure referential theory, and do not regard sense as different from reference. (In this regard, in the two decades since I made this diagnosis, there have been several attempts to indicate where to look for such a theory of sense. The most convincing of these is that of Mark Siderits. 4 However, while I believe that Siderits is correct in tracing a kind of sense theory to the Buddhist apoha theory, I do not think that this affects my general diagnosis.) This direct referential theory allows the theory of Sabdapramana to nullify the distinction between understanding and knowing. While translations of empty expressions like "hare's horn," a la Russell, abound in Nyaya literature, the real difficulty facing the theory is a reasonably acceptable account of exactly what it is to understand a false sentence. Thus, even for the Nyaya — if they are to be consistent — Sabdapramana must be intrinsically true. For false sentences cannot generate any Sabdabodha, or understanding, not to speak of prama, or true cognition. But, of course, according to the theory, Sabdabodha and Sabdqjanyapramd, or linguistic understanding and linguistically generated true

The Theory of Prarnana

225

cognition, are the same! The enormous problems generated by this identification are all too obvious.5 Fourth, there is one area of knowledge in which the claim that Sabda is an irreducible prarnana is strongest: the domain of what ought and ought not to be done. For, if it is possible to establish factual truths either by perception or by reasoning of some sort, it may reasonably be claimed that our only sources of knowing what ought and ought not to be done are the verbal instructions, either written or spoken, of moral teachers, elders or the scriptures. Finally, there is a strong claim, which underlies the tradition like a rock, that gruti, or the sacred, heard, scriptures (the vedas and the upanishads), are apauruseya, not composed by any human author. If accepted, this view grants freedom from possible faults. It grants an incorrigible authority that no text of human authorship could support. I shall return to this concept of 'apauruseyatvd below.

Ill What sort of theories of prameya, of possible objects of true cognition or knowledge, did the classical Indian philosophers hold? Given the great variety of ontologies — ranging from the pluralism of Nyaya-Vaisesika to the monism of Advaita Vedanta — we must begin by looking at some of the salient features of these ontologies. The first thing to be noted is that none of these Indian ontologies countenance abstract entities of the sorts admitted into Western ontologies; e.g., there are no full-fledged Fregean senses or propositions, numbers are reduced to gunas, or properties, of sets, and universals, although common, are not the rarefied entities of Western metaphysical tradition that are amenable only to the grasp of pure reason. In the Indian tradition universals are more concrete entities, perceived through the same sense organ by which their instances are. Moreover, there are no pure unactualized possibilities in this tradition, which is not surprising because God's mind, their habitat in the Western metaphysical tradition, does not play the role of creating out of nothing in Indian thought. In the absence of possibilia and of abstract entities such as

226 J. N. Mohanty propositions, some standard concepts of necessary truth and its opposite contingent truth cannot find any formulation in the Indian systems. Thus we have accounts of what the world does consist of, but not of what might have been or of what could not possibly not be. In this regard, we should point out that, the standard formulation of vyapti, or 'pervasion,' is not modal but extensional. ("It is never the case that in all those loci where smoke is present fire is absent," but not "It is impossible that ..."). One reason why in traditional Western metaphysics, the metaphysical scheme claimed a sort of necessity as contrasted with those features of the world that are studied in the sciences is that, in the West, metaphysics and science have been sharply separated ever since the beginning of Aristotelian metaphysics, which has always been concerned not with beings, but with being qua being — where 'being' is construed in such well-known ways as "the highest being"; "the most general predicates or categories"; "the meaning of 'being'." For Indian metaphysicians, on the other hand, science and metaphysics are one continuum that together undertakes to understand the structure of the world; they differ only in their order of generality. The one exception in this regard is Advaita Vedanta, in which the world, being unreal, is left to empirical science, while metaphysics (if, indeed, that is what paravidya should be called — which is doubtful), is the knowledge of the one Being that underlies all other beings. If creation out of nothing, or creation in the strict sense, has no place in Indian thought, this should not be taken to imply that creation is a marginal phenomenon for the darSanas. On the contrary, I believe that the issue of creation out of nothing determines some very central features not only of the Indian cosmologies but also of the metaphysical notions of God, substance, time and negation. But space does not allow me to undertake the task of showing this here.

IV In this concluding section, I would like to make a few remarks on the pramana-prameya structure in its entirety, that is, on the philosophical enterprise in general as illustrated in the darganas. While engaged in their highly sophisticated philosophical activity, the Indian thinkers did not explicitly or self-

The Theory of Pramana

227

consciously focus on the nature of their enterprise. It was while defending their enterprise in response to the sceptical challenges of a Madhyamika that they would sometimes remark on the nature of what they were doing. Without going into textual details, let me state some of the main issues. 1. The Madhyamika critique is not merely one of epistemology but also one of ontology, for the critic insists on their mutual dependence: but you cannot decide what the pramanas are unless you have already decided what things there are to be known, and you cannot settle the latter question unless you have the means of knowing at hand. Where, then, do you begin? If the circularity cannot be broken, why not give up the entire enterprise? 2. The pramana -prameya theorist's response to this challenge has been, in brief, that it presupposes an unnecessarily strong reading of the unity of the two parts of a darSana. There is no oneto-one relation between a pramana and its prameya. One and the same thing can be known by more than one pramana. One and the same system of ontology can be made to go together with different epistemologies: consider the Nyaya and the Vaisesika. The mutual dependence that threatens the relation between cognition in general and the object in general is broken by specifying both and thereby establishing a many-one or one-many relation between terms on each side. 3. What (2) entails is that a darSana is not such a seamless unity that parts of it cannot be taken out of the context of that system. My interpretation goes against the traditionalist view, which regards each darSana as a unique point of view. Among moderns, the Russian emigre David Zilbermann (whose untimely death was a serious loss to the cause of Indian philosophy) held such a holistic view of dargana, a view rejected in the present essay. 4. The reflective question, what sort of knowledge a philosophical system itself yields (or amounts to), or whether it can itself be appropriated into one or more of the pramanas recognized by the system, is not asked explicitly, but practice suggests that it is the latter alternative that has quite often been chosen. The reason for this — hinted at earlier — lies in not recognizing philosophical knowledge as knowledge that is, qua knowledge, distinct from the sorts of knowledge thematized within the system. An alternative way out, which would consist in distinguishing between understanding and knowing (i.e., that

228 J. N. Mohanty philosophy yields understanding, but no knowledge) was not open in view of the purely referential theory of meaning. When the Vedantin says that knowledge of Brahman brings about mok$a, this knowledge is such that both the knowledge and the entity of which it is knowledge are both thematized within the system. When the Nyaya sutra says that knowledge of the sixteen padarthas, or 'categories,' brings about the highest good, what sort of cognition is this? Is it by one or more of the pramanas? The answer seems to be 'y e s 5. Students of the dar&anas often wonder from where the early masters — the authors of the sutras and Bhdsyas — derived the list of pramanas and prameyas that constituted the framework which the later authors went on refining. To say that they elaborated a way of seeing — using the verbal root drs (=to see) — does not assuage that anxiety. In any case, it is not true that the later authors simply refined and clarified the framework suggested by the founding fathers. They also changed and modified within limits (such change also speaks against a strongly holistic reading of the darSanas). The more common response has been to trace the framework back to the Sruti, the heard texts with no human author. But if we consider the intellectual phenomenon that philosophical systems as diverse as Nyaya and Vedanta claim affiliation to the Sruti, how can the nature of §abda-pramana be construed to render this paradoxical situation intelligible? I suggest that, for this purpose, the nature of gabdapramana as applied to Sruti. be construed in a manner that is implicit in the tradition's understanding of itself, but not explicitly formulated as such. And it is here that I differ from the orthodox interpretation of the role of Sruti vis-a-vis the philosophies. The apauruseyatva of gruti means, for me, neither that the texts are not composed at all (thus I deny their literal construction), nor that these texts express some supernormal, mystic experience. Not the first, because there is enough internal evidence that the texts were composed and also because the literal construal makes no sense. And not the second, because, in my view, sentences express not experiences but thoughts. (Although I would like to defend this last thesis, this is not the place to do so.) Setting aside these two commonly held interpretations, I would like to suggest the following. First, in understanding the Sruti texts, it is utterly irrelevant and of no use to appeal to the intentions of their authors. For it is

The Theory of Pramana

229

the texts, the words themselves, that are primary, in that they are what is available to us and they are what define the tradition for us. We use these words and texts to interpret our experiences, our world and ourselves, and in so doing we also interpret the words themselves. Whereas orthodoxy ascribes to the words of the gruti what it takes to be "the" meaning, I leave open possibilities of interpreting them. It is this plasticity of meaning, this endless possibility of interpretation, the continuing challenge they face us with, that sets the texts of the Sruti apart from those of smrti, or 'the remembered texts.' The Sruti are foundational not because they express infallible truths, but because they define the parameters within which the Hindu philosophers asked questions, understood the concerns expressed in the texts and appraised their answers. In this sense, Sabda (as Sruti) is not itself a pramana , but underlies their applications. Apauru$eya$ruti is not the supreme pramana, infallible and raised above all the rest; it is, rather, the source of all those concerns and inquiries (not answers) in the solution of which the different pramanas exhibit their special philosophical relevance. Who is then thinking within that tradition? My answer is: To be thinking within the tradition, it is not necessary to subscribe to any or all of the answers of the schools; but what is necessary is to share the concerns as sources of philosophical problems. I define the tradition, then, in terms of concerns rather than of beliefs.

Notes 1.

Cf. B.K. Matilal, Perception. An Essay on Classical Indian Theories of Knowledge (Oxford: Oxford University Press, 1985), esp. p. 35. 2. Cf. my "Psychologism in Indian Logical Theory," in Matilal and J.L. Shaw, eds., Analytical Philosophy in Comparative Perspective (Dordrecht: Reidel, 1985), pp. 203-211. 3. Matilal, ibid, (note 1), esp. pp. 105, 135. 4. I am referring to Siderits' unpublished essay, "The SenseReference Distinction in Indian Philosophy of Language." 5. I have discussed this point in detail in my unpublished Presidential Address to the 61st session of the Indian Philosophical Congress, October 1986. See also my Foreword to Purasottama Billimoria's forthcoming book on Sabdapramana (Reidel).

Thinking on Empty: Madhyamaka Anti-Realism and Canons of Rationality M. Siderits

The Madhyamaka Buddhist doctrine of emptiness may be aptly if cryptically summarized by the slogan, "the ultimate truth is that there is no ultimate truth." To many of their opponents, this position seemed to commit the Madhyamikas to a kind of nihilism about rationality: if supreme rationality requires that we accept nothing as supremely rational, this can only mean that no canon of rationality is rationally acceptable. This became the basis of a standard objection to the doctrine of emptiness — that it is necessarily self-stultifying. Now the doctrine of emptiness may, I think, be viewed as a version of what is currently called anti-realism. And in recent disputes over anti-realism a stock objection to the view has been that it is committed to relativism about rationality: if what is rationally acceptable is partially dependent on human practice as shaped by human institutions, then given the cultural and historical variability of institutions, what is rationally acceptable cannot be canonized but must rather be relative to a concrete historical situation. Thus the question arises whether the anti-realist must in general be committed either to nihilism or to relativism about rationality. The Madhyamikas have a response to the charge of nihilism, but they are silent on the question of relativism. The cultural factors that make relativism a pressing issue for us were largely absent from the classical Indian context, so that the various forms of relativism do not receive philosophical scrutiny in the Indian tradition. What I wish to explore, however, is whether an anti-realist response to the charge of relativism might not be teased out from the Madhyamaka response to the objection that they are nihilists about rationality. To understand the force of the Madhyamaka claim that the

232 M. Siderits ultimate truth is that there is no ultimate truth, we must begin with the traditional Buddhist distinction between conventional truth and ultimate truth. This distinction seems to have originated in early Buddhist attempts to reconcile the Buddha's claim that there is no self, "I", or person, with various of the Buddha's pronouncements concerning persons and their properties. Such apparent conflicts were resolved by distinguishing between contexts in which the Buddha speaks the strict and literal truth, and contexts in which the Buddha adapts his speech to the capacities of an audience not yet philosophically prepared to understand the strict and literal truth. Thus, in discussing karma with the latter sort of audience, the Buddha might say, "The person who commits evil deeds in this life earns the status of a dog in the next life." Before a more philosophically sophisticated audience, however, the Buddha might instead say, "A causal continuum containing evil desires prior to death-consciousness also contains pain sensations subsequent to rebirth-consciousness." This distinction is sharpened when the implicit reductionism of early Buddhism is explicitly formulated and defended in the Abhidharma schools of Buddhist scholasticism. When the Buddha denied that "I" has a referent, he seems to have meant to deny both that there is some part of the person (the self) that T denotes, and that the person as a whole (the psycho-physical complex) is the referent of "I." At the hands of Abhidharma philosophers, the latter denial is defended on the basis of the reductionist principle that wholes are unreal — a chariot is a mere conceptual fiction, the word "chariot" being no more than a convenient designation for a set of parts assembled in a certain way. Conventional and ultimate truth may, then, be defined in the following way: • A statement is conventionally true if and only if it is acceptable to common sense. • A statement is ultimately true if and only if it corresponds to the facts and neither asserts nor entails that wholes exist. Thus, given that the person may be reductively analyzed into a space-time chain of causally related physical and psychological states, the statement, "persons dislike pain" is conventionally true but ultimately false (since it implies that persons exist). But, given that neither pains nor desires are reductively analyzable,

Thinking on Empty

233

the statement, "Under appropriate circumstances pain sensations serve as causal conditions for aversive desires," is ultimately true. 1 So formulated, the conventional-ultimate distinction clearly preserves the intent of the original hermeneutical distinction between popular and philosophically sophisticated contexts. For the intention of the early interpreters was to defend the Buddha against the charge of contradicting himself. With our new distinction in hand, we can say that when the Buddha makes assertions concerning persons and their properties, what he says is ultimately false. But such statements may be transformed into ultimate truths, namely by substituting for terms which appear to denote persons or other wholes the names of their irreducible parts and a description of the relations that obtain among them. Likewise much of what passes for truth in the world may, while ultimately false, be preserved through such a translation process. Not all, of course: "Persons can achieve enduring happiness" has no ultimately true correlate, for its truth requires the existence of an enduring experiencer, a condition that is not in fact met. But the vast majority of our common-sense beliefs must, it is held, be thus grounded, since otherwise we could not explain why their acceptance leads to successful practice. Believing that there is a bartender before me, I ask for a glass of stout. There are, ultimately, no bartenders, yet my thirst is satisfied. This is explicable, it is supposed, only if there is before me a set of causally interrelated physical and mental entities and events. This conception of the distinction between conventional and ultimate truth was common to both Abhidharma and Mahayana Buddhist philosophers. And it is, I think, possible for us to see both of them as thereby distinguishing between a conventional or worldly canon of rationality on the one hand, and philosophical rationality on the other. Thus the content of conventional truth is determined by what it would be rational to affirm according to the standards accepted by the world. And the content of ultimate truth is just those statements that are deemed acceptable according to the standards of philosophical rationality. But, clearly, philosophical rationality is here to be thought of as growing out of the worldly canon of rationality, through a process of refinement. Thus consider the argument whereby Abhidharma Buddhist philosophers sought to show that wholes are unreal:

234 M. Siderits With respect to the whole and its parts, there are just four possibilities: 1. The whole is real and the parts are also real. In this " case there are just two further possibilities: a. The whole is distinct from the parts in relation. But there is no empirical evidence for the existence, e.g., of a chariot which is not just evidence for the existence of chariot parts. b. The whole is identical with the parts in relation. Then the whole and the parts in relation should have all the same properties. But the whole has the property of having x parts, which the parts in relation do not have. 2. The whole is real, but the parts are unreal. In this case we could not explain the causal efficacy of the parts in producing the whole at the time of assembly. 3. Neither the whole nor its parts are real. Then nothing whatever would exist, which is absurd. 4. The whole is unreal, only the parts are real. Since the remaining possibilities have been proven false, (4) must be true. This argument is meant to persuade us of the truth of its conclusion by appeal to epistemic standards that are accepted by all persons, e.g., modus tollens (lb) and Ockham's Razor (la). But once we accept its conclusion we are required to revise those standards. For our common-sense standards of rational acceptability commit us to the existence of chariots, trees, armies and persons, entities which the argument shows not to exist. Philosophical rationality is just what remains when we have purged the worldly canon of those elements that lead to these and similar inconsistencies. How, then, does Madhyamaka seem to be committed to nihilism with regard to rationality? In his principal work, Mulamadhyamakakarika,2 Nagarjuna (the founder of Madhyamaka) seeks to demonstrate that there can be no consistent account of the ultimate nature of reality. He does this by subjecting the various metaphysical theories of his philosophical rivals to reductio argumentation. He thus seeks to show that there can be no coherent account of a distinct self-subsistent entity, of material causation, of time, of motion, and the like. The upshot of all this is expressed in Madhyamaka philosophical vocabulary as the claim that all things are empty [Sunyd), where emptiness is defined as lack of essential nature. Now on the view

Thinking on Empty

235

of ultimate truth developed by Nagarjuna's Buddhist predecessors, possession of an essential nature was the mark of that which truly exists. Thus the claim that all things are empty (in the Madhyamaka sense) is tantamount to the claim that the domain of ultimate truth is empty (in the ordinary sense). There can be no such thing as the one true theory that corresponds to the nature of reality. And, since this claim is itself arrived at through the employment of philosophical rationality, it is meant to represent the ultimate truth. Hence the Madhyamaka claim: the ultimate truth is that there is no ultimate truth. The obvious rejoinder to this seeming paradox is that such a view is necessarily self-stultifying, and possibly inconsistent. For the Madhyamika seems to be claiming that philosophical rationality is impotent to discover substantive truths concerning the ultimate nature of reality. Yet it is precisely philosophical rationality which the Madhyamika employs to arrive at this result — a result which would appear to represent a substantive truth concerning the ultimate nature of reality. Thus, if their view were correct, we would have no reason to believe it. And, if they are to be seen as advancing reasons for this claim, then they are inconsistent. Nagarjuna's response to this objection is that he himself has no thesis. 3 By this he appears to mean that the claim that all things are empty is not a substantive claim concerning the ultimate nature of reality. Nagarjuna admits that were he to make such a claim he would be inconsistent, but he feels he need not. All he need do is show the impossibility of constructing a coherent metaphysical theory, one which satisfies the demands of philosophical rationality; and this he takes himself to have done in MMK. To say that all things are empty is just to summarize this result; it is not to attribute to entities the ultimate nature of emptiness. 4 This response is, however, ambiguous. Nagarjuna might be taken as intimating that the ultimate nature of reality is not amenable to rational analysis, i.e., that no canon of rationality is adequate to reality. Or he might instead be seen as claiming that the very notion of an ultimate nature of reality is incoherent, that the contradictions that emerge from the exercise of philosophical rationality can be traced to a realist conception of truth as correspondence to a mind-independent reality. On the first interpretation Nagarjuna is not, of course, assert-

236 M. Siderits ing that reality transcends the limits of philosophical rationality, since to do so is to make a substantive claim concerning the ultimate nature of reality. But this is a fact nonetheless — perhaps one that, strictly speaking, cannot be stated but only shown. (It is sometimes suggested, by those who take this line, that this fact can be grasped in a kind of mystical intuition.) On this interpretation Nagarjuna is, after all, a kind of nihilist about rationality: no canon of rationality is rationally acceptable, since none is competent to arrive at the ultimate truth about reality. But his position is not inconsistent, and it is only selfstultifying in the weak sense in which any form of general skepticism is self-stultifying. Any reasons the skeptic gives to support his view cannot meet his conditions of adequacy. His reasons may nonetheless prove compelling to us. If so, we shall be moved to accept his claim that nothing can meet these conditions of adequacy. On the second interpretation of Nagarjuna's claim not to have a thesis, his view is akin to that of current anti-realism. He neither asserts nor intimates any claims about the ultimate nature of reality, for he takes the very notion of a way that the world is independently of our cognitive activity to be devoid of meaning. On this reading the arguments of MMK were meant to show that the project of philosophical rationality is misguided not because reality is beyond conceptualization but because that project presupposes a conception of truth as correspondence to mind-independent reality. Where the worldly canon of rationality is governed by the demands of human practice, philosophical rationality sets itself the goal of describing reality in a way that is not contaminated by human interests. (Thus our belief in the existence of persons is the product of 'mere' convenience.) Nagarjuna's MMK refutations of various metaphysical theories were meant to show us the implausibility of this project. That he should give reasons in doing so is not inconsistent, in that he is not a nihilist about rationality: the worldly canon of rationality, grounded as it is in human practice, is perfectly acceptable. The slogan, "the ultimate truth is that there is no ultimate truth," is merely a striking way of putting the point that an acceptable canon of rationality will have to reflect human needs, interests and institutions. Which of these two interpretations is correct? There is some evidence for the anti-realist interpretation in the strategy

Thinking on Empty

237

Nagarjuna employs in W. There he is faced with the objection that in order to substantiate his claim that all things are empty he must employ some accepted means of knowledge. But, since the use of a means of knowledge presupposes the existence of distinct objects of knowledge with determinate natures which a means of knowledge can adequately represent, any proof Nagarjuna can offer for the claim that all things are empty must have as a presupposition the denial of that claim. Nagarjuna responds by considering how one might go about constructing a theory of the means of knowledge. The only method that seems at all promising is one that seeks to achieve a kind of reflective equilibrium. We begin with a stock of basic beliefs that would share wide acceptance, and seek out the causal conditions that led us to have such beliefs. Having formulated a provisional account of the causal conditions for knowledge, we then look to see whether it accords with both our original stock of basic beliefs and with other beliefs we have about the world and about knowledge. If not, we make adjustments, either in our account of the conditions for knowledge, or in our stock of intuitions, or in both. This process continues until our account of the causal conditions for knowledge and our intuitions are in equilibrium, when no further adjustments need to be made. The resulting account of the causal conditions for knowledge is a theory of the means of knowledge. Now Nagarjuna would, I claim, hold that, if such a procedure is carried out properly, then one would be justified in holding any belief that was induced through some causal route that was identified by the theory as a means of knowledge. Such beliefs would constitute knowledge. What Nagarjuna denies is that such beliefs in any way "mirror" or "correspond to" a mind-independent existent. 5 Such beliefs may well deserve the title "knowledge"; if we have engaged in the process of reflective equilibrium with sufficient care, then such beliefs will no doubt meet the tests of worldly practice. But we cannot then claim that a belief induced by one of the causal routes we have identified as means of knowledge corresponds to reality. We cannot do this because, given the method of reflective equilibrium which we used to arrive at our theory of the means of knowledge, we cannot prise apart the respective contributions of "mind" and "world" to the beliefs warranted by our means of knowledge. For our theory of the means of knowledge was arrived

238 M. Siderits at on the basis both of a set of intuitions about the nature of the world, and a set of beliefs about the nature of knowledge. And the epistemic support this theory gives our resulting beliefs is coherentist in nature. Thus it is in principle impossible to distinguish between the support a given belief receives from the first set and that it receives from the second. But in order to give content to the notion of correspondence between knowledge and reality we must be able to make just such a distinction. Thus the notion of correspondence between belief and reality is without content. 6 Of greater relevance to the present purpose, however, is Nagarjuna's statement at MMK XXIV lOab: "The ultimate [truth] is not taught independently of customary practice." This would seem to bear out the anti-realist interpretation in that Nagarjuna seems here to be claiming that a truth that transcends human practices is unavailable. Unfortunately, not all Madhyamikas have so interpreted him. For instance, Prajnakaragupta takes this to mean that, since the ultimate nature of reality transcends conceptualization, the truth about reality cannot be stated but only intimated, namely, through developing the paradoxical implications of our conventional understanding of reality. 7 Bhavaviveka disagrees, however; he explicates the ultimate truth of MMK XXTV 10 as just the fact that nothing is attained upon the (philosophical) analysis of the categories of the conventional world-view. The point of Nagarjuna's dialectic is not that reality transcends conceptualization, but that truth must conform to human practice, that philosophical rationality is doomed precisely insofar as it seeks a truth free of all taint of human needs and interests. This is demonstrated by systematically displaying the paradoxes that arise when worldly categories are subjected to analysis in accordance with the strictures of philosophical rationality. On this point, Candrakirti agrees with Bhavaviveka. He describes his method to a realist opponent as follows: I, however, because of proficiency in the establishment of conventional truth, having settled just upon the worldly thesis, rendering ineffective by means of one proof another proof which is thrown up as contradicting a portion of conventional truth, like a world-elder I refute you, who are falling away from8 worldly conduct, but I do not refute conventional truth.

Thinking on Empty

239

The use of the term "world-elder" is telling. In the Buddhist context "elder" means one who has ascertained the truth and thus attained nirvana. Candrakirti's world-elder must then be one who takes the truths warranted by the conventional canon of rationality to be the only truth available, and on that basis undermines the pretensions of philosophical rationality. For both Bhavaviveka and Candrakirti the ultimate truth is just that there can only be conventional truth. Reality is determined through conformity to conventional practice (arthakriyasamarthycit) and as such dissolves upon analysis (avicaramaniya), i.e., cannot live up to the demands of philosophical rationality. Thus the concept of material causation has utility given our practices: pots are made from clay. Yet there is no philosophical analysis of the relation between material cause and effect. To Candrakirti and Bhavaviveka this is an indictment not of the worldly categories, nor of the conventional canon of rationality, but of philosophical rationality and its conception of realist truth. 9 The Madhyamika thus escapes the charge of nihilism about rationality by embracing a kind of conventionalism. Our conventional or customary standards of rational acceptance are the only game in town, and this is as it should be given that these develop out of the need to interact with our environment and with one another. But to modern ears this will have the ring of Feyerabend's "anything goes." Might not the game get played by different rules in different towns, or in the same town in different epochs? That is, if truth is just warranted assertibility, and warranted assertibility is determined in accordance with human needs and practices, then, insofar as the latter may change across cultures or historical periods, the true propositions and the general canon of rationality may likewise vary. But, then, if we find a culture that accepts beliefs that are irrational by our lights, and it turns out that those beliefs are warranted by that culture's standards of rational acceptance, we cannot criticize the members of the culture for holding those beliefs. This apparent consequence of anti-realism strikes many as clearly unacceptable. This "community standards" approach to truth and rationality would seem to entail that a proposition might be either true or false depending on the cultural milieu in which one finds oneself. And this is absurd, for what makes a proposition true or false is the facts, which do not vary across

240 M. Siderits cultures. Likewise the fact that we have more true beliefs and fewer false beliefs than did our ancestors can only be explained by supposing that we employ superior standards of rational acceptance, a notion that is incomprehensible on the present approach to rationality. There are three strategies that anti-realists have taken in responding to the charge that they are relativists about rationality. The first is the bullet-biting strategy of saying, with Feyerabend, that indeed anything goes, and accusing the opponent of engaging in mere table-pounding. Those who take this line would say that the relativistic consequences of anti-realism seem absurd only if one presupposes that truth is correspondence to a mind-independent reality — a presupposition the anti-realist explicitly denies. The second strategy is to deny the existence or perhaps even the possibility of radical divergence in canons of rationality and their resultant sets of accepted beliefs. Some who take this line argue that, while canons of rational acceptability depend on a set of shared practices, we have no evidence that such practices ever differ to the extent that would be necessary to give rise to incommensurable systems of belief. Others (e.g., Donald Davidson10) argue that we cannot even make sense of the idea that there could be such differences. The third strategy is to seek to make coherent a kind of non-relativistic pluralism — a view that affirms the existence of distinct canons of rationality but denies the relativistic claim that all such canons are equally acceptable. Those Madhyamika philosophers who took the anti-realist line were never confronted with the charge of being relativists about rationality. This is so for the simple reason that relativism was not the issue for classical Indian philosophy that it is for the modern philosophical community. It is interesting to speculate on why the issue did not arise for them. It might be suggested that relativism becomes a pressing issue only when a culture is confronted with significant changes in its science — a condition not present in classical Indian culture. But I would suggest that at least a part of the answer lies rather in the hegemonistic strategy which Brahmanic culture used to subsume the other cultures of the Indian subcontinent. (The various Indian philosophical schemes whereby an author shows the views of his opponents to represent partial insights into the truth expressed by his favored system are adaptations of this strategy to the philosophical

Thinking on Empty

241

context.) In any event, our authors are silent on the issue. This need not mean, though, that they must remain silent on the issue; perhaps a response to the charge of relativism might be teased out of what they say on other issues. In the last part of this paper I shall attempt to do just this. I will claim that we can discern two distinct answers coming from two wings of the Madhyamaka school, one embracing relativism, the other laying the groundwork for a non-relativistic pluralism. Tibetan Buddhist historians divided post-Nagarjuna Madhyamaka into two subschools, the Prasangika and the Svatantrika. While this is an imposed classification with a certain degree of arbitrariness, it does reflect discernible differences in philosophical practice. The Prasangikas are so called because of their reliance on reductio arguments, prasaAga being the absurd consequence of such an argument. The Svatantrikas, on the other hand, presumably advance and defend theses of their own, svatantra being an 'own-thesis'. Now, with the few possible exceptions noted above, all Madhyamikas advance theses, so if there is a difference here it is a difference in degree and not in kind. A deeper and more interesting difference lies in the respective attitudes of Prasafigikas and Svatantrikas toward conventional truth. Both agree that ultimately there is only conventional truth. They also agree that conventional truth accords with worldly practice, and dissolves upon analysis. They disagree, however, over how stringently to take the claim that conventional reals are not analytically findable. This difference can be brought out through their disagreements over the nature of the self and the proper response to subjective idealism. All Buddhists accept the claim that there is no self. They also agree that belief in the self is an important part of the commonsense world view. Where Prasangika and Svatantrika disagree is over the correct analysis of the term "self." The Prasafigika position is that the self is a mere name that is imputed to the aggregates (that set of entities and events, both physical and mental, that constitutes the life of a person); that is, the term "self is indiscriminately applied to any or all of the entities and events that make up a person, or to some entity supposed to be distinct from but related to these entities and events. The Svatantrika view is that the self is the continuum of mental consciousness, i.e., the continuous series of moments of consciousness that take

242 M. Siderits inner states as their objects. (In Buddhist psychology the consciousness that apprehends a datum lasts only as long as does the presentation of that datum to the mind.) This is not a dispute over our ordinary-language use of the term "self." It is agreed that by that term we appear to mean some self-sufficient entity that has our experiences and grounds our sense of personal identity over time. What is disputed is the extent to which we can make our use of the term coherent by constructing a theory of the self. When the Prasangikas claim that "self" is a mere name, they are denying that there can be such a theory. Thus Candrakirti argues that the self does not exist by demonstrating that the self could not be identical with the aggregates, distinct from the aggregates, the basis of the aggregates, based on the aggregates, the possessor of the aggregates, the relation that obtains among the aggregates, or the set of the aggregates. 11 This argumentative strategy presupposes that what we have in mind when we speak of the self is an entity that might exist in any of these seven ways. But, obviously, no entity could exist in all of these ways. Thus we must not have anything with definite identity conditions in mind: the self is a mere name that is imputed on the basis of the aggregates but cannot be analyzed in terms of the aggregates or anything distinct from the aggregates. The Svatantrikas, by contrast, maintain that we can identify a best candidate for the self from among the aggregates. To say that the self is the continuum of mental consciousness is to say that our practice with respect to the self can be made coherent by identifying it as that which is aware of our psychological states. This disagreement is important for what it reveals about the underlying views of conventional truth and worldly rationality. On the Prasangika view, conventional truth is a set of brutely given practices which must be taken at face value. To seek to analyze these practices is to introduce the standards of philosophical rationality, standards which the Madhyamika believes to be thoroughly discredited. If the only truth is that which accords with conventional practices, then we must resist all temptation to analyze and explain these practices. To give in to such temptation would be to confuse the worldly canon of rationality with the philosophical. The Svatantrikas, by contrast, seem to hold that conventional rationality contains within itself the resources to distinguish between better and worse theories. Given our prac-

Thinking on Empty

243

tices with respect to persons, we do better to identify the self as the subject of experiences than as, say, the totality of our physical and psychological states. What makes the former a better theory than the latter is that it better coheres with our practices — and coherence with our practices is one part of the worldly canon of rationality. So long as we understand the inherent limitations of the method of reflective equilibrium, we can use that method to construct theories which serve to articulate and extend our conventional beliefs and practices. We can see the same divergence at work in the respective responses of Prasangika and Svatantrika to the subjective idealism of the Yogacara school of Buddhist philosophy. The Prasangika Candrakirti takes the Yogacarin to task for his claim that cognition cognizes itself (a claim that the Buddhist idealist must accept, since he, unlike Berkeley, cannot posit mental substances). Candrakirti's criticism is that conventional linguistic practice requires a distinction between object of an action and instrument of the action. (We cut the tree with an axe, point to the tree with a finger, etc.) The Yogacarin will of course reply that the syntactic requirements of our language should not be taken as indicative of the ultimate structure of reality. Candrakirti's response is that there is no investigation of reality apart from our conventional linguistic practices. Thus the claim that reality does not conform to the structure of our language is incoherent. Here Candrakirti is rejecting the possibility that our linguistic practices might allow of alteration and improvement. Once again, conventional practices are thought of as brutely given and to be taken at face value. The Svatantrika Bhavaviveka criticizes Yogacara subjective idealism in a different way, focusing on the argument that Yogacarins used to demonstrate that the cause of sensory experience cannot be an external object. This argument relies in part on the claim that the cause of sensory experience cannot be a collection of atoms, since the individual atoms are by hypothesis too small to be perceived, and there is no such thing as a collection that exists over and above the individual constituents. Bhavaviveka attacks this claim by distinguishing between two kinds of collection: an aggregate, which is a collection of distinct and dissimilar substances related only through proximity (e.g., a junkyard); and a composite, which is a collection of distinct but similar substances that are related through more than mere

244 M. Siderits proximity (e.g., a pot, which is composed of interconnected earth atoms). Now, while the insensible properties of the atoms do not become sensible when the atoms form a mere aggregate, they may when atoms form a composite. Thus the rock will break a window, where a heap of sand with the same mass will not. 12 What is striking about this response is that it is made against the background of a long Buddhist reductionist tradition that explicitly denies any such distinction can be drawn in a principled way. Bhavaviveka must have known this, yet he says nothing to forestall this objection. Why is this? I suggest that it is because Bhavaviveka agrees: ultimately the distinction between aggregate and composite cannot be maintained. But Bhavaviveka also tells us quite explicitly that he is here concerned only with conventional truth. While such a distinction cannot be drawn in a way that survives the scrutiny of philosophical rationality, it is implicit in our conventional practice, and can be drawn out in a way that is consistent with the worldly canon of rationality. Although common sense does not actually recognize the distinction, common sense realism about external objects is nonetheless not without resources in its confrontation with subjective idealism. The strategies of the Prasangika and Svatantrika attacks on subjective idealism differ in an important way. Both start from the Madhyamaka position that there can only be conventional truth, that such paradigmatic products of philosophical rationality as the theory of subjective idealism are to be rejected insofar as they presuppose that there can be such a thing as the way the world is, independently of customary practice. Both seek to use this insight to refute subjective idealism, namely, by showing that the theory conflicts in important respects with conventional truth and thus undermines the only truth there could be. But the conflicts are developed in different ways. Candraklrti's criticism — that no one would ordinarily accept the claim that cognition cognizes itself — is derived by reading off the surface of our conventional practice (in this case, linguistic practice). He does not attribute to common sense the view that cognitions cannot cognize themselves; nor does he try to develop out of common sense the theory that cognitions cannot cognize themselves. Conventional truth is not to be seen as a theory, or as giving us the materials with which to construct a theory. It's just that we don't talk that way.

Thinking on Empty

245

Bhavaviveka's criticism, on the other hand, does involve developing a sort of theory out of our conventional practices with respect to, e.g., pots and heaps. Given how we do talk and act, it is possible to use some of the tools of philosophical rationality — here, the drawing of distinctions and the formulation of definitions, constrained by considerations of consistency and coherence — to mount a philosophical defense of common-sense realism about external objects. Of course we must be careful to bear in mind that the resulting theory is not ultimately true. All that can be in dispute in this debate is whether the belief in external objects is conventionally true. The Yogacarin, through his use of philosophical rationality, takes himself to have shown that it is not. Bhavaviveka's point is that the uncritical realism about external objects that we find in common sense can be developed into a theory that is better, as an articulation of conventional truth, than Yogacara subjective idealism. While no theory is ultimately true, one theory may still be better than another. This last point marks a crucial difference between Prasangika and Svatantrika. And it is in the Svatantrika position that we can discern the outlines of a response to the charge of relativism. Suppose the Madhyamikas were faced with the sort of evidence that is taken by many to support relativism: the existence of distinct cultures with what appear to be quite different canons of rationality. As anti-realists, the Madhyamikas would of course be barred from appealing to any ultimate standards in judging the relative merits of these two styles of reasoning. And the Prasangikas, with their no-theory approach to conventional truth, would be forced to accept the relativism about rationality that such evidence seems to suggest. But the Svatantrikas could, I think, be pluralists without being relativists: pluralists in admitting a plurality of possible canons of rationality no one of which is ideally suited to uncover the ultimate nature of reality; but not relativists in that one such canon may quite straightforwardly be said to be better than another. To see how this might be, we need to flesh out our tale of two cultures. First, as Davidson has taught us to see, there must be widespread agreement in belief between the two, since otherwise the members of each would have no reason to attribute any beliefs whatever to the other. But, in arriving at this common core of shared beliefs, we must also presuppose a common set of conventionally accepted principles of rational belief formation.

246 M. Siderits (Intuitively, we would not be inclined to say that they believe the same things we do unless we supposed that they form beliefs from evidence in the same way that we do.) So there will be substantial overlap in the canons of rationality current in the two cultures. How might they differ? Suppose that one culture includes the practice of homeopathic medicine, the other the practice of modern science-based medicine. Then, since each practice involves a distinct style of reasoning, some of the beliefs and practices of each will be unintelligible to members of the other. Thus the question whether a given sample of mandrake root is sufficiently similar to male genitalia to cure impotence makes no sense unless one has entered into the practice of homeopathic reasoning by similitude. Likewise questions about the representativeness of a sample in a population study aimed at isolating the causes of a disease would seem senseless to one who did not already accept certain forms of statistical reasoning. Now, for all the standard anti-realist reasons, there is no answer to the question which of these styles of reasoning is better suited to uncovering the ultimate nature of disease. There is no ultimate nature of disease. Nor is there any other external standard in terms of which we can judge the relative merits of these two styles of reasoning. But, just as the Svatantrika would say that a philosophical rationality that is non-reductionist, paradigmatist and sensitive to the issue of vagueness is better than a philosophical rationality that is reductionist, essentialist and committed to achieving absolute clarity — better in that it better serves needs and interests that arise out of the common core of conventional practice (e.g., in the case of a theory of the self, the need to resolve doubts about the practice of reward and punishment) without undermining conventional practice — so we can say in this case that one style of reasoning might prove better than the other at serving a generally accepted need (generating ways to cure our physical maladies) without undermining the worldly canon of rationality on which the whole enterprise rests. A cautionary note is in order concerning this kind of response to relativism. To say that one style of reasoning may be better than another in terms of adequacy to conventional practices and needs is not to say that we can expect to be able to anticipate in advance which styles of reasoning will prove superior. We can't say until we try, and even then, given our imperialistic tendencies concerning our own styles of reasoning, we might not be too

Thinking on Empty

247

good at telling. This is the point of pluralism. But neither should we be lulled into complacency about our own style of reasoning by the thought that there are no ultimate standards which could show them to be inadequate. This is the point of anti-relativism. 13 Pluralism about styles of reasoning is required in that what we seek are not just effective methods of finding cures for disease, but methods that also cohere with styles of reasoning in the common core of worldly rationality. The statistical reasoning characteristic of population studies is a formalized extension of a common epistemic practice; but, then, so is the reasoning by similitude characteristic of homeopathic medicine. That the ways of thinking of the homeopath may seem to us not to make sense should warn us that our culture's valorization of scientific rationality may have blinded us to the possibility of other ways of extending worldly rationality — ways that might turn out not only to be useful but also to fit in quite well with ordinary epistemic practice. Pluralism alone is not enough, however, for pluralism readily becomes a smug tolerance: they have their ways of reasoning, we have ours, and that's the end of the matter. This is tolerance without humility, which comes, I suspect, from mistaking antirealism for idealism. The idealist maintains that, in some sense, we "make it all up." But the anti-realist denial that there are ultimate standards of rationality is not equivalent to the claim that we are free to adopt any standards whatever. For there are constraints, some of which we must suppose to be independent of the particular practices of our culture. The denial of ultimate standards is just the denial that we can give content to this notion of independent constraints; it is not the denial of any such constraints. Their way of reasoning might well be better than ours. In the context of Madhyamaka, anti-realism has the soteriolgoical function of deflating the pretensions of philosophical rationality. It was Nagarjuna's insight that, in order to become truly selfless, one must become profoundly skeptical about our ability to arrive at the ultimate truth about reality. The notions of ultimate truth and ideal rationality breed clinging and attachment, and thereby a sense of self. Skepticism about these notions undermines these last refuges of the self. The Svatantrika attempt to bring elements of philosophical rationality into co-

248 M. Siderits herence with worldly rationality may be seen as an effort to temper skepticism with humility. When we give up the ultimate truth and resign ourselves to nothing more than conventional truth, we may still be tempted to suppose that we at least have that right. To see that there may yet be better and worse versions of conventional truth is to see that there, too, there is always room for improvement. 14

Notes 1. I am here treating in the formal mode what Buddhist philosophers uniformly discussed in the material mode. Thus, referring back to the chariot, a Buddhist philosopher would say that a chariot is a conventional truth but not an ultimate truth. 2. Nagarjuna, Mulamadhyamakakarika, ed. Vaidya (1960); hereafter MMK. 3. See Vigrahavyavarta.nl, v. 29, in Kamalswar Battacharya, "The Dialectical Method of Nagarjuna (Translation of the "Vigrahavyavartani"), Journal of Indian Philosophy, 1(1971), 217261; hereafter W; M. Siderits, "The Madhyamaka Critique of Epistemology I," ibid. 8(1980), 309-320. 4. On this point, see also MMK XIII 7-8, XXII 11. 5. See the commentary on W 51. 6. M. Siderits, "Nagarjuna as Anti-Realist," Journal of Indian Philosophy, 16(1988), 311-325. 7. Christian Lindtner, "Atisa's Introduction to the Two Truths, and its Sources," ibid. 9(1981), 185-187. On the problem of the two truths, see also Malcolm David Eckel, Jndnagarbha's Commentary on the Distinction Between Two Truths. (Albany: SUNY Press, 1987); C.W. Huntington, Jr., "The System of Two Truths in the Prasannapada and the Madhyamakavatara: A Study in Madhyamika Soteriology," Journal of Indian Philosophy, 11(1983) 77-106. 8. P.L. Vaidya, Editor, M a d h y a m a f c o ^ a s t r a of Nagarjuna (Mulamadhyamakakarikas) with the Commentary: Prasannapada by Candraklrti (Dharbanga: Mithila Institute, 1960), p. 23. 9. Ibid., p. 9. 10. D. Davidson, "On the Very Idea of a Conceptual Scheme," Proceedings and Addresses of the American Philosophical Association, 48(1974), 5-20; The Method of Truth in Metaphysics," Midwest Studies in Philosophy, 11(1977), 244-254. 11. Vaidya, ibid, (note 8), pp. 145-148. 12. On this point, see Donald S. Lopez, Jr., A Study of Svatantrika (Ithaca: Snow Lion, 1987), pp. 308-310; Malcolm David Eckel, "Bhavaviveka's Critique of Yogacara in Chapter XXV of the Prajnapradlpa," Miscellenea Buddhica, Indiske Studier V, ed. Chr. Lindtner (Copenhagen: Akademisk Forlag, 1985), pp. 65-67. See

Thinking on Empty

249

also Malcolm David Eckel, "The Concept of Reason in Jnanagarbha's Svatantrika Madhyamaka," in B.K. Matilal, ed., Buddhist Logic and Epistemology (Dordrecht: Reidel, 1985), pp. 253-278. 13. For further discussion of these issues, see Ian Hacking, "Language, Truth and Reason," in M. Hollis and S. Lukes, eds.. Rationality and Relativism (Cambridge, Mass.: MIT Press, 1982), pp. 48-66; Hilary Putnam, Realism and Reason (Cambridge: Cambridge University Press, 1983). 14. I have profited from discussing some of the issues addressed in this paper with Robert Steinman.

Contributors

Shlomo Biderman is Senior Lecturer and Chairman of the Department of Philosophy of Tel-Aviv University. He is interested in Indian philosophy, philosophy of religion and skepticism. His articles are predominantly in these fields. He is now working on the philosophy of Nagarjuna. A. C. Graham, for many years Professor of Classical Chinese at the School of Oriental and African Studies, London University, has translated and commented on two Taoist classics: The Book ofLieh Tzu (1960) and Chuang Tzu: The Inner Chapters (1981). He has also written Later Mohist Logic, Ethics and Science (1978); Reason and Spontaneity (1985); Studies in Chinese Philosophy and Philosopical Literature (1986); and Disputers of the Tao: Philosophical Argument of Ancient China (1989). Lenn E. Goodman is Professor of Philosophy at the University of Hawaii. His interests include Jewish and Islamic Philosophy, metaphysics and ethics. Among his publications are Ibn Tufayl's Hayy ibn Yaqzan (1972); Rambam: Readings in the Philosophy of Moses Maimonides (1977); Monotheism: A Philosophic Inquiry into the Foundations of Natural Theology and Ethics (1981); Saadiah Gaon's Book of Theodicy, A Tenth Century Arabic Commentary and Translation of the Book of Job (1988). John Kekes is Professor of Philosophy and Public Policy at the State University of New York at Albany. His interests include epistemology, ethics, and public policy. His main publications are A Justification of Rationality (1976); The Nature of Philosophy (1980); The Examined Mind (1988); and Moral Tradition and Individuality (1989). Joel Kupperman is Professor of Philosophy at the University of Connecticut. He is the author of Ethical Knowledge (1970) and 251

252

Contributors

The Foundations of Morality (1983). He has also published articles on ethics and metaphysics in Mind; Philosophy East and West; Proceedings of the Aristotelian Society and other journals. Bimal K. Matilal is Spalding Professor of Eastern Religions and Ethics and Fellow of All Souls College, Oxford University. His books include The Navya-Nyaya Doctrine of Negation (1965); Epistemology, Logic and Grammar in Indian Philosophical Analysis (1971); Logic, Language and Reality (1985); and Perception: An Essay on Classical Indian Theories of Knowledge (1986). He is the editor of Journal of Indian Philosophy and coeditor of the series Studies of Classical India. Jitenda N. Mohanty is Professor and Chairman of the Department of Philosophy, Temple University. His books include Edmund Husserl's Theory of Meaning (1964); Phenomenology and Ontology (1970); Husserl and Frege (1982); and Gangesas' Theory of Truth (1966). His latest book is Transcendental Phenomenology: An Analytic Account (1989), and he is now completing a two-volume work on Reason in Indian Thought. Roy W. Perrett is Lecturer in Philosophy at Massey University, Palmerston North, New Zealand. His fields of interest include metaphysics, ethics and philosophy of religion, Indian philosophy and philosophy and literature. He is the author of Death and Immortality (1987) and editor of Indian Philosophy of Religion (1989) Hilary W. Putnam, Professor of Modern Mathematics and Mathematical Logic in the Department of Philosophy of Harvard University, is a Fellow of the America Academy of Arts and Sciences and a Corresponding Fellow of the British Academy. His books include Meaning and the Moral Sciences (1987); three volumes of Philosophical Papers (1975-1982); Reason, Truth and History (1981); The Many Faces of Realism (1987); and Representation and Reality (1988). Ben-Ami Scharfstein is Emeritus Professor of Philosophy at TelAviv University. He is interested in comparative culture in general and comparative thought in particular. His most recent books are The Philosophers (1980); Of Birds, Beasts, and Other

Contributors

253

Artists: An Essay on the Universality of Art (1988); and The Dilemma of Context (1989). Mark Siderits is Associate Professor in the Department of Philosophy, Illinois State University. His work is primarily in Buddhist and orthodox Indian metaphysics and the philosophy of language, on which he has written numerous articles. Nathan Sivin is Professor of Chinese Culture and the History of Science at the University of Pennsylvania and a Fellow of the American Academy of Arts and Sciences. Among his major publications are Chinese Alchemy: Preliminary Studies (1968) and Cosmos and Computation in Early Chinese Mathematical Astronomy (1969). He co-edited Chinese Science: Explorations of an Ancient Tradition (1972) and edited Science and Technology in East Asia (1977). Richard Sylvan (formerly Routley) is Coordinator of the Ecological Trust. He is also employed at the Research School of Social Sciences, Australian National University. His main intellectual work has been on environmental philosophy and practice and on logic and metaphysics, as on the thought of Meinong. His books include Environmental Philosophy; The Fight for the Forests; and Relevant Logics and Their Rivals.

Index of Names

Aeschylus, 93 Alphonsus Liguori, 68 Anaxagoras, 85, 104, 109 Aquinas, St. Thomas, 184 Aristotle, xvi, 12, 30, 93-97, 101, 104107, 113, 114, 120, 155, 184, 207, 226 Arjuna, 208-211 Arrow, K.J., 214 Artabhaga, 195 Ash'ari, al-, 111 Athanasius, 58 Augustine, 58, 95 Averroes — see Ibn-Rushd Avlcenna — see Ibn-Sina Baier, A, 128 Bentham, J., 58, 59 Bergson, H., 81 Bharata, 202 Bhavavtveka 238, 239, 243, 244 BhIma,205, 211 Bishop, J., 69 Black, M., 37 Boutroux, E.-E.-M., xvil Brentano, F., 41 Buddha, 232, 233 CaiYong, 166, 170 CaiYuanding, 166, 170, 173, 176-178, 182 Candrakirti, 238, 239, 242-244 Cantor, G., 185 Carnap, R, 25 Chuang-tzu 87, 142, 149-151, 153 Chunqiu fan lu, 168 Cicero, 129 Clausewitz, C. von, 12 Confucius, xvi, 119, 120, 122-127, 129, 133-137, 141, 151, 154 Copernicus, 173 Dai Faxing, 171 DaiZhen, 166 Dasaratha, 202 Davidson, D., 240, 245 Derrida, J., 157-159

Descartes, R., xi, 81 Dhrtarastra, 204, 206, 207, 210 Dodds, E.R, 93 Dong Zhongshu, 168 Dumont, L., 195 Durkheim, E., 83 Duryodhana, 204, 206, 209 Eells, E., 38 Einstein, A., xv, 88 Elster, J., xi, 49, 64, 65, 67, 68 FangYizhi, 180, 181 Farabi, al-, 95, 103, 104, 106, 112 Feyerabend, P., 239, 240 Fingarette, H., 119, 122, 123, 125 Frege, G., 225 Freud, S., 94 Gadamer, H.G., 217 Galileo, 173 Garga200 Ghazali, al-, 95, 111, 112, 114 Govindaraja, 198 Granet, M., 153 Gregory of Nyassa, 58 HanFei, 141, 154 Harsanyi, J.C., 44 Hegel, G.W.F., 85 Henderson, J., 179, 182 Heraclitus, 183 Hilary of Poiters, 184 Hipparchus, 177 Hourani, G., 110 Hsun-tzu, 141, 154 HuiShih, 149, 150 Hume, D., xi, xvi, 58, 67, 111, 119, 120, 127-129, 133-135, 137 IbnRushd, 95, 105-110 IbnSina, 95, 104-106, 112 Ibn Tufayl, 94 being, 172 Jabala, 196 Jabali, 202 Jakobson, R, 155

256 Index James, W., 36, 122 JanCh'iu, 124-126 Janaka, 194 Jeffrey, R.C., 44 JiaKui, 169, 170 Jung, C.G., 159Kafka, F., ix Kaikeyi, 202 Kalidasa, 201 Kant, I., xvi, 6, 14, 19, 22, 81, 87, 119, 120, 130-137, 222 Kao-tzu, 141 Kepler, J., 155, 173 Keynes, G., 25 Kind}, al-, 95-97, 99-102 Krishna, 197, 209, 210 Kroeber, A., 83 Kuhn, T., 157, 158 Kulluka, 197, 198, 200, 203, 208 Kung-sun Lung, 142-144, 149 La Rochefoucauld, F. de, 124 Lao-tzu, xvii, 141, 158-161 Laplace, P.S. de, 133 Leeds, S., 24 Leibniz, G.W., xi., 85, 110 Ltngat, R, 214 Locke, J., 11, 34, 83 Lukasiewicz, J., 32 Maimonides, M., 95, 111-114 Manu, 201-203 Matilal, B.K., 222 Medhatithi, 199 Mencius, xvi, 119, 120, 125-127, 129, 133-137, 141, 178 Mill, J.S., 65 Milton, J., 95 Mo-tzQ, 141, 143, 148 Montaigne, M.E. de, xi, 87 Nagarjuna, x, 87, 234-238, 241, 247 Nakula, 211 Needham, J., 153, 159, 160 Nelson, B., 84 Neumann, J. von, 19, 20 Nietzsche, F., 58, 59, 153 Oakeshott, M., 37

Picasso, P., 93 Pindar, 93 Plato, ix, xiii, xvi, 57, 93, 94, 99, 100, 177, 178, 183, 185, 206, 207 Plotinus, ix, 57, 94, 114 Poincare, H., xii Prajnakaragupta, 238 Proclus, 106, 107 Ptolemy, 177 Rama, 202, 210, 211 Rauschenberg, R., 78 Rawls, J.,210 Razi, al-, 95, 100-102, 104, 106 Riker, G., 45 Ryle, G., 5, 157 Saadiah, 95, 96, 98, 100-102, 105 Santisagara, 59 Sartre, J.-P., 121 Satyakama, 196 Schopenhauer, A., 58 Sen, A., 46, 70, 214 Shen Kuo, xvii, 173-176, 178, 180 Siderits, M., 224 Sidgwick, H., 66, 67, 69 Simon, H., 32-33, 44, 71 Simplicius, 106 Slote, M., 71,72 Socrates, 93, 95, 109 Spengler, O., 84 Spinoza, B., xi Stendhal, 64 Stevenson, C, 136 Suppes, P., 44 Sylvan, R, 71 Tu Shun, 87 TychoBrahe, 177 Valmiki, 211 Vidura, 206 Weber, M., x, 83, 194, 195 Whitehead, A.N., 87, 88, 175 Wittgenstein, L., ix, 27, 82 Wolff, R.P., 135

Ordeshook, P.C., 45 Pareto, V., 214

Yajnavalkya, 194, 195, 201 YinZhiyi, 181, 182 Yin Zhongchun, 181 Yudhisthira, 204-207, 209-212

Parmenides, 106, 183, 185 Pascal, B., 22, 67 Paul, R, 135 Peirce, C, xii, xv, xvii, 23-26 Philoponus, 96, 99

Zeno, 149, 185 Zhang Jiebin, 181 ZhuXi, 173, 176-179, 182 Zilbermann, D., 227 Zu Chongzhi, 171, 172