A System of Pragmatic Idealism, Volume I - Human Knowledge in Idealistic Perspective [I] 0691073910

Citation preview

A SYSTEM OF PRAGMATIC IDEALISM VOLUME I HUMAN KNOWLEDGE IN IDEALISTIC PERSPECTIVE

Nicholas Rescher

PRINCETON UNIVERSITY PRESS

PRINCETON, NEW JERSEY

Copyright © 1992 by Princeton University Press Published by Princeton University Press, 41 William Street, Princeton, New Jersey 08540

In the United Kingdom: Princeton University Press, Oxford All Rights Reserved Library of Congress Cataloging-in-Publication Data Rescher, Nicolas. Human knowledge in idealistic perspective / Nicholas Rescher.

p. om. -——~(A System of pragmatic idealism: v. i) Includes index. ISBN 0-691-07391-0 (alk. paper)

1. Knowledge, Theory of. 2. Idealism. 3. Realism. I. Title. Il. Series: Rescher, Nicholas. System of pragmatic idealism ; v. 1. B823.R44 1991 vol. 1 [BDi61] 141 s—dc20 [121] 91-15575 CIP

This book has been composed in Linotron Caledonia Princeton University Press books are printed on acid-free paper and meet the guidelines for permanence and durability of the

Committee on Production Guidelines for Book Longevity of the Council on Library Resources Printed in the United States of America by Princeton University Press, Princeton, New Jersey

10987654321

For Dorothy

os

IN GRATITUDE

Contents

List of Figures Preface Introduction: The Thrust of Idealism PART I: A NORMATIVE THEORY OF RATIONALITY One

The Light of Reason Two

Cognitive Rationality and Risk: A Critique of Skepticism

15

Three Evolution as an Allocation Mechanism

36 ~

Four

Rational Inquiry as the Pursuit of an Unattainable Ideal

47

PART I: THE STATUS OF SCIENCE Five

An End to Science?

63

Six The Imperfectibility of Science

77

Seven

The Intelligibility of Nature

Eight

96

-

Our Science as Our Science

110

PART I: THE PURSUIT OF TRUTH Nine

Induction as Truth Estimation through Cognitive Systematization

129

Ten

A Coherentist Criteriology of Truth

157

Eleven The Pivotal Epistemic Role of Postulation

183

tae

CONTENTS

Twelve ; Truth as an Idealization

209

Thirteen Scientific Truth and the Arbitrament of Praxis

223

PART IV: REALISM AND IDEALISM

Fourteen

Cognitive Limits Fifteen Metaphysical Realism

255

Sixteen Scientific Realism

275

Seventeen

Science and Idealization

296

Eighteen Realism and Idealism

304

Name Index

329

Subject Index

333

List of Figures

1.1. Objects of Rational Deliberation 2.1

Approaches to Risk

20

2.2

Risk Acceptance and Misfortunes

22

2.3

A Problem of Choice between Belief Implementation and Inaction ,

28

3.1

Hypothetical Payoffs in a Situation of Competition

3.2

Hypothetical Distributions of Cognitive Ability

5.1

The “End-to-Science” Argument

6.1

The Goal Structure of Natural Science

9.1

Alternative Explanations

134

10.1

Foundationalism vs. Coherentism in Factual Inquiry

163

10.2

The Twofold Cycle of the Legitimation of Systematizing

Methodology

os)

77

175

10.3

Feedback Cycle of Legitimation

177

11.1

Communicative Negativities and Their Causes

203

11.2

The Relationship between Security and Definiteness

206

13.1

The Naive View and the Revised View of the

Inquiry/Reality Relation

230

15.1

The Retrojustification of the Presuppositions of Inquiry

269

18.1

Versions of Idealism

305

18.2

Mind and Matter in Conceptual Idealism

320

Preface

THIs BOOK is part of a larger project. In numerous and varied publications appearing over the last thirty years, I have developed a system of philosophy—a complex but (I hope) coherent position embodying a unified approach to a wide variety of interrelated philosophical issues. However, since these publications were individually self-sufficient—each standing on its own expository feet as a separate intelligible unit—it has generally not been clear to readers that a systematic whole is at issue. Not infrequently, reviewers have said things like, “This book is an interesting discussion of the problem area, but its treatment of topic X could use more detailed development”—not realizing that X is the subject of an entire book published a year or two earlier. The larger project here at issue seeks to present my philosophical ideas in a sufficiently comprehensive and coordinated form that their systematic interrelatedness becomes clearly manifest. Three books are contemplated. The first—the present volume—is dedicated to issues of epistemology, philosophy of science, and philosophy of nature. The second will deal with issues of value theory, ethics, and

practical philosophy. The third will deal with the nature of philosophical inquiry itself. As a whole, the trilogy is intended to exhibit and explicate the interactive coordination of pragmatic and idealistic themes that have throughout characterized my approach to philosophical issues. I am very much indebted to several people who have helped with the preparation of the book. Robert Butts and John Kekes offered constructive comments on the project. Marian Kowatch has produced the typescript with exemplary patience and competence. And Susan Sterrett and John Mandle have read my manuscript in a conscientious though not always availing effort to eliminate inadequacies of exposition and obscurities in formulation. And I also appreciate the caring helpfulness of Ann Himmelberger Wald at the Princeton University Press. Pittsburgh, Pennsylvania Summer 1990

Introduction The Thrust of Idealism

THE AIM of this book is to formulate and systematize the philosophical ideas that I have presented in various lectures and publications in the course of a long-sustained endeavor to develop an idealistic perspective on our knowledge and understanding of the world.. Idealism centers on the conception that reality as we understand it reflects the workings of mind. And it construes this as meaning that the inquiring mind itself makes a formative contribution not merely to our understanding of the nature of the real but even to the resulting character we attribute to it. Through its generality and indefiniteness, this sort of view is not so much a philosophical doctrine as a general strategy for _ developing one. This book will present a version of idealism that represents a particular instance of this general approach. It concretizes a generically idealistic perspective via two more specific commitments.

Interactionism. The product of our inquiries—the body of information that affords our picture of the real world—is not shaped one-sidedly by the nature of the object of investigation itself (objectivism), or shaped one-sidedly by the cognitive instrumentalities by whose means we address the problems at issue (subjectivism), but represents an interaction between the two in which the contributions of the two sides are inextricably intermingled. Both the object of inquiry and the inquiring mind that investigates it make a contribution to the character of the picture of nature at which our inquiries arrive, and the respective contributions of the two sides are so intermingled that one cannot drawaclear line to separate which is which. Axiological conventionalism. We manage our cognitive business via descriptive conventions whose nature as items of human artifice also informs the products that we produce by their means. These conventions are dictated in part by our values—by what we find convenient, efficient, and guided to our purposes. The idealism at issue here pins its fortunes on the circumstance that our picture of reality is—and has to be—constituted by means of the rep-

resentations of a mind that has its own characteristic sort of modus operandi and proceeds accarding to its own terms of reference. Its impetus to idealism lies in exploring the implications of the consideration that it is

One The Light of Reason Homo Sapiens: Intelligence as the Survival Mechanism of Our Species Rationality is a matter of the intelligently managed pursuit of intelligently adopted objectives. The ancients regarded the human being as “the rational animal” (zdon logon echén), set apart from the world’s other creatures by the capacity for speech and deliberation. Following the precedent of Greek philosophy,.Western thinkers have generally deemed the deliberate use of knowledge for the guidance of our actions to be at once the glory and the duty of Homo sapiens. It is not particularly surprising that we humans should have acquired intelligence to enable us to secure information about the world. This is only natural and to be expected because if we did not succeed in this cognitive venture, we would not be here as the sort of creatures we are. Intelligence arises through evolutionary processes because it provides one effective means of survival. It has evolved not because nature somehow favors intelligence but because intelligence favors the survival of its possessors within nature (at any rate, up to a point, since a benign out-

come to the spread ofnuclear-weapons technology is not yet a foregone conclusion). The long and short of it is that we rational animals would not be here as the sorts of creatures we are, and could not long continue in existence as such, if our rationality were not survival conducive. The explanation for our cognitive resources is thus fundamentally Darwinian. Indeed, the conception of knowledge as a tool for survival—cognitive Darwinism—is as old as biological Darwinism. The master himself put forward the idea that our human capacities and competences in the areas of language, reasoning, and theorizing are part and parcel of our biological endowment, emerging because these abilities were biologically advantageous in the. struggle for survival.1 And after Darwin this idea burst like a Roman candle across the firmament of nineteenth-century 1 “The small strength and speed of man, his want of natural weapons, etc., are more than counterbalanced, firstly, by his intellectual powers, through which he has formed for himself weapons, tools, etc., though still remaining in a barbarous state, and, secondly, by his social qualities which lead him to give and receive aid from his fellow men” (The Origin of

Species [1859; reprint, New York, 1952], pp. 443-44). Darwin elaborates on this greatly in The Descent of Man.

4

A NORMATIVE THEORY OF RATIONALITY

thought. The- concept of evolution was applied to explain our species: intellectual resources by various of the major philosophers of the day— . including Arthur Schopenhauer, Herbert Spencer, Charles Sanders

Peirce—and a great host of other thinkers of variant persuasions.also fol-_ lowed suit. . Intelligence is by no means an inevitable feature of conscious organic life. Here on earth, at least, intelligence is our peculiarly human instrumentality, a matter of our particular evolutionary heritage. Homo sapiens is also Homo quaerens. With us, the imperative to understanding is something altogether basic; we cannot function, let alone thrive, without information regarding what goes on about us. The knowledge that orients

our activities in this world is itself the most practical of things—a rational animal cannot feel at ease in situations of which it can make no cognitive

sense. The demand for understanding, for cognitive accommodation’ to one’s environment, for “knowing one’s way about,” is one of the most

fundamental requirements of the human condition. The discomfort of unknowing that drives us to inquire and investigate is a sentiment natural to any intelligent being and is readily understandable as such. We rational animals must feed our minds even as we must feed our

bodies. In pursuing information, as in pursuing food, we have to make do with the best we can get at the time. We have questions and need answers—the best answers we can gét here and now, regardless of their imperfections. This basic practical impetus to acquire coherent information represents the fundamental imperative of cognitive intelligence. Baf-

flement and ignorance (to give suspensions of judgment the somewhat harsher name that is their due) themselves exact a substantial price from us. The need for information, for cognitive orientation in our environment, is as pressing a human need as that for food itself—and more insatiable. We humans want and need our cognitive commitments to compose an intelligible story, to give a comprehensive and coherent account of things. Cognitive vacuity-or dissonance is as distressing to us as physical pain. Intelligence’s pursuit of information—of putative knowledge— is one more facet of evolution’s strategy of making what is useful to the~ species compelling to the individual by way of its own pleasure (want) or demand (need).

Intelligence and rationality are two sides of the same coin. Rationality consists in the appropriate use of reason to resolve choices in the best possible way. To behave rationally is to make use of one’s intelligence to

figure out the best thing to do in the circumstances. Rationali

i

Intelligence and Rationality

5

THE LIGHT OF REASON

sitersof:deliberately’ doing’‘the:‘best one’eanwith |

"Ration ity is not just Fimatter of thought, but of action as well. The person who proceeds unintelligently in figuring out the proper thing to

do thereby falls afoot of reason’s demands. But the person who figures out correctly the rationally optimal thing to do in the circumstances and then fails to do it thereby also strays from the path of reason. Whatever

sort of doing may be at issue—be it in belief, action, or evaluation—must be done appropriately if the demands of reason are. Sationality.cCalls:foxtheintelligent pursuit‘ofappropriate’ends. It pivots onthe e recognizably effective pursuit of appropriately appreciated benefits. Reason can ine should) ¢come> into Operation whenever a choice or

claim

Paarl of ELAN

ne h

vort.deed

rationality functions in all three of the very7 different domains of cognition, action, and evaluation, at bottom one.and the same thingis at issue throughout: effecting the resolution of choices in the best possible way—in line with the strongest reasons. Rational people are, ex officio, in a position to provide a rationale of good reasons for what they do— which is why the range of rationality has, since the days of the philosophers of ancient Greece, generally been deemed coextensive with the linguistically equipped intelligence needed for the articulation of warranting considerations—of reasons. Rationality thus involves the capacity Figure 1.1. Objects of Rational Deliberation 1. Cognitive rationality: What to believe or acceptP What to maintain re-

garding states of affairs in both the formal and the empirical domains. Product: factual contentions (beliefs). 2. Practical rationality: What to do or perform? What to decide regarding actions. Product: action recommendations (injunctions). 3. Evaluative rationality: What to prefer or prize? What to adopt regarding goals and ends. Product: evaluations (appraisals). 2 Note that optimization does not necessarily call for maximization. Usually the best use

we can make of our resources overall involves settling for enough of a given desideratum rather than striving for the obtainable maximum.

6

A NORMATIVE THEORY OF RATIONALITY

_to.give-an-account’#--to.useone’s intelligenceto,provide'a’tationale’for what one does thatestablishes itsappropriateness. It _Sucting. one’ vaffairs:‘respon ibk ree: “péliefor action or evaluation is accordingly rational if theagent¢cantell a story that succeeds in making sense of it by showing that (and how) it was a sensible thing to arrive at in the circumstances, due to its optimal conduciveness to appropriate ends. Where something is amiss in an agent’s management of beliefs, actions, or evaluations— where one’s information or decisions or appraisals are inappropriate in the circumstances—there is a failure of rationality. -Gognitiveredson:has historically been seen asthe capacity to grasp facts throughtheirunderlying ‘ation piendi;*

’

:

4The use of reason in-

e ationality of a particular belief—ofaccepting a particular contention——is of course a function of the evidence at one’s disposal. (The Siamese king who refused to believe that rivers solidify in northern coun-

tries at a certain season of the year was perfectly rational, the freezing of water into ice lying wholly outside his experience.)® Rationality is a matterof doingthings © in thebest at Gnost intelligent) way one can manage in

i ¢ "is‘to'do“itifor:good:and., ogent: this is not the same as just having some motive for‘doing it. All of us almost always act for motives, but valid reasons are (ex hypothesi) what motivate the rational agent, and most of us do not act rationally all of the time. All too often we are moved to what we do by desires or wants, and

these may or may not be rationally well advised. The crux is that it may 3 Logon didonai, as Plato puts it in the Republic. Note the logon ech6n in the Aristotelian formula. 4G, W. Leibniz, Nouveaux Essais, bk. 4, chap. 17, sec. 4.

5 John Locke, Essays concerning Human Understanding, bk. 4, chap. 15.

THE LIGHT OF REASON

7

or may not be in one’s best interests to get what one desires—that very much depends on exactly what it is that one happens to want. People

automatically have a motive whenever there is a desire, but they have a good reason for what they do only when it is recognizably in their best interests—when there are cogent grounds for thinkingthat what they do

is genuinely for their benefit. Desire;is:notaiitoiatically!r ings: Addictive gamblers have reasons for persevering in squandering alga th

Treasons... From the rationalpoint of view, our mere wantsass such have little significance. Their import can and should be outweighed by our

interests and our needs. A minimally rational agent does what he or she does for reasons— never mind for the moment whether they be good, bad, or indifferent. But a truly rational agent has good reasons here. Good reasons are those whose guidance optimally serves our real or best interests in the matters at issue. What.makes a reason a good reason is the fact that its implementation leads our efforts in appropriate directions, and the best reasons are those that achieve the most in this way. The rational thing to do is always the intelligent thing—that is, the adoption of the overall most cost-effective means toward the achievement of one’s appropriate goals whose realization is conducive to our “real” or “true” interests. ’ One can virtually always rationalize something one wants to do, putting it in a favorable light with respect to rationality by finding some reasons for doing it. But that by itself does not of course suffice to render the act in question a rational one. For despite its being supported by some reasons, further, still better, available reasons may well point an-

other way. Rationality is not just a matter of having some reasons for what one does, but of aligning one’s beliefs, actions, and evaluations effectively with the best or strongest available reasons. It pivots on doing that which, everything considered, one is “well advised” to do. The matter of

giving or following the course of intelligent and responsible advice is the crux of rationality. There is nothing complex or arcane about the sorts of considerations that determine good and cogent reasons in this regard. It is a matter of the sort of things that conduce to one’s real advantage, to one’s best interests. This is a matter of furthering a full and rewarding life, preeminently involving the sort of things that make us happier and/ or better persons in what relates to our benefit and the benefit of those

8

A NORMATIVE THEORY OF RATIONALITY

who do and should matter for us (our family, community, and fellows at large, and the advancement of our individual and communal values). Practical rationality thus calls for appropriate resolutions intelligently arrived at and sensibly implemented. It is geared to the sensible pursuit of appropriate ends. (Here “ends” serves as a broad generic term covering both concrete goals and more diffuse values.) Rationality is thus a two-sided, Janus-faced conception. On the side of means, it reflects a pragmatic concern for efficient process, while on the side of the appropriateness of ends, it reflects a value-geared concern for product. (Moreover, the acceptability of the means themselves also enters in.) To reemphasize: rationality is a matter of seeking to do the very best we can (realistically) manage to do in the circumstances. To proceed contrariwise by way of rationalizing—by putting reasons for what we want in the place of good reasons for what we ought to do—is an abuse of rationality. Rational persons are by definition those who, in effect, use their intelligence to maximize the (responsibly formed, subjective) probability that matters will eventuate favorably for the overall realization of their. real (or best) interests. Both cognition and evaluation are accordingly crucial. It is silly to dedicate shrewd means to inappropriate ends; it is foolish to pursue worthy ends by ineffective means. Rationality thus has a crucially economic dimension, seeing that the impetus to economy is an inherent part of intelligent comportment. Cost and benefits are the pivotal factors. Be it in matters of belief, action, or evaluation, rationality demands a deliberate endeavor to optimize appropriate benefits relative to the expenditure of available resources. Reason demands—and indeed consists in—the cultivation of intelligently adopted objectives in intelligent ways. Throughout, rational appropriateness is a matter of determining in which direction the best (or strongest) reasons point. All domains of rationality are thus united by the common mission of finding “the best reasons.” Rationality’s cardinal injunctions are: “Adopt what is, in the circumstances, the best available option! Pursue it in the way that is, in the circumstances, the best possible way!” The idea of optimization, of seeking for the best among the visibly available alternatives, lies at the verycore of rationality. The ancient Greek idea that nous, or reason, is a force

operating in the world to make things work out “for the best,” already codifies this critical linkage between rationality and optimality.® While rationality calls for striving for the best solution, this is frequently not a straightforward matter. Often several distinct parameters of merit are operative—in cognitive contexts, say, plausibility, probability, and naturalness (simplicity)—and they may yield rather different modes of optimality. Generally, we should speak of a rational (and a best) 6 See Plato, Timaeus, 41d4, 58b2—4.

THE LIGHT OF REASON

9

solution to our problems rather than simply the rational (or the best) solution. There often is no uniquely best alternative, and when this is so, rationality leaves us with options. (But of course to say that several alternatives may be eligible is not to say that anything goes.) It must be stressed that while rationality requires us “to do our very best,” this is a matter of our best. It is not a question of what, abstractly, is “the best that can possibly be done” as distinct from “the best that can be done by us, given the capacities and resources at our disposal.” Rationality does not make demands of us beyond the limits of what is genuinely possible for us. For rationality, no more is demanded of us than doing our realistic best to work efficiently and effectively toward the realization of our cognitive, practical, and evaluative goals. Rationality is,

indeed, a matter of optimization, of doing the best we can, but this must be understood as the best we can feasibly manage to do in the existing circumstances. Some problems are simply too complex to admit of a perfect resolution by the means actually at our disposal. Hence, rationality asks for no more than that we do the best we can manage with the means in hand. Being realistic is a part of being rational.

People as Rational Agents: The Presumption of Rationality Humanity may not be important for rationality; we humans may well be only one of many sorts of rational creatures on the universe’s wider stage—perhaps even a relatively eccentric one. But rationality is crucial to humanity, an integral part of what defines us as the sort of creature we are, or take ourselves to be. The claim to rationality is a crucial aspect of our self-image. Of all our characteristics, it is the most central and important. Do we humans have reason, or do we merely think we have reason?

The question is almost academic: it just does not matter all that much in the final analysis. What matters primarily is exactly that we do and must think ourselves to be rational—that this capacity features centrally in our self-definition. This circumstance that we invoke reason to define ourselves as the beings we take ourselves to be is quite enough to give the concept of rationality its pivotal place in the scheme of things. Yet, why fix on reason as our definitive trait? Why not another characteristically human capacity such as imagination or foresight or worriment or deceitfulness? Essentially, because reason is the crossroads where all these characteristic attributes (and many others) come together. Either these other capacities are essential resources for rationality (such as imagination or foresight), or the capacity for reason is involved in their _ operations (such as worry or deceit). All of the higher capabilities char-

10

A NORMATIVE THEORY OF RATIONALITY

acteristic of Homo sapiens are bound up in one way or another with our capacity for reason. But are we humans really all that rational—might we not well use that intelligence of ours to destroy ourselves? Unquestionably so! But consider. Plants are never foolish; only a creature that has the capacity for intelligent thinking can act unintelligently. Moreover, there is a crucial distinction between having the capacity for intelligent (rational) action and exercising this capacity intelligently (rationally). The very fact that we can use our intelligence unintelligently—self-destructively, for ex-. ample—betokens our status as rational agents. In principle, of course, rationality is not an exclusively human characteristic.? Other kinds of creatures may well be rational—if not here on earth, then elsewhere. But then the capacities that they must have to qualify as rational are just exactly those we ourselves lay claim to: intelligence (the ability to acquire knowledge through inductive learning and to process it through discursive, reason-governed thought) and free will (the capacity for decision and action in the light of evaluation on the basis of information). To recognize some as rational beings is to ascribe to them a capacity for the intelligence-guided pursuit of duly evaluated goals. This is an attribute that we standardly concede to our fellows, even as we wish them to grant it to us. It is bound up with the view that human beings as such have certain rights and claims, that by their very nature they have a peculiar status as rational agents (and morally responsible agents as well). In considering people’s actions, we generally proceed on the basis of a presumption of rationality with respect to our fellows. In the absence of any evidence to the contrary, we proceed on the supposition that people do what they do on the basis of good reasons, granting them the benefit of the doubt in point of rationality. We take the stance that there is an explanation on the basis of good reasons as to why they proceed as they do—that they act as they do under the impression that some good will

come of it, some benefit for themselves or ‘others. When asked to explain why someone did something, it suffices that we establish that this wa¥

the rationally appropriate thing for them to do—at any rate, in the absence of indications that put their rationality into question. Only reluctantly, as a last resort, do we judge that someone has proceeded on the basis of forces or motives outside the range of their rational control. Thus here, as elsewhere, a presumption of normalcy prevails. Realizing full well that people are not always rational, we nevertheless find that they (fortunately) are so generally and ordinarily, and we feel free to proceed -.. on this basis. In the ordinary course of things we grant people the benefit 7 See Jonathan Bennett, Rationality (London, 1964).

THE LIGHT OF REASON

ll

of any doubt and treat them as rational agents in the absence of convinc-

ing counterindications. This presumption of rationality is not just a matter of generosity but one of self-interest too. It affords us an important labor-saving device by allowing us to explain people’s actions simply by noting that they were, in the circumstances, rational. Insofar as people’s actions are of a sort that can be accounted for on principles of practical rationality, we may rest content and pose no further explanatory questions. Accordingly, we need not in general raise questions about someone’s motives when they do the rationally indicated thing. In the ordinary course of events—when things go “normally”—no further special account of the matter is called for. Inquiry can rest and explanation cease in the assurance that things are as they can be expected to be. In giving people the benefit of any doubt in point of rationality, we save ourselves an immense amount of explanatory labor. The assumption that others behave rationally is also highly useful in another regard. People need to be predictable to one another to coexist in fruitful interaction; we need to make ourselves intelligible to others to

coordinate our activities. Thus, by demanding that people act rationally and in conforming our own actions to this expectation, we facilitate the alignment of human actions. (We cannot align on unintelligent, haphazard choices because the possibilities proliferate unmanageably without the definiteness of optimization.) Even if the presumption of rationality did nothing else beyond rendering us this service of making our activities intelligible to one another, it would be eminently worthwhile. But explanatory convenience does not tell the whole story. A world in which we can communicate and collaborate with others is for this very

reason a safer world for us and our kind. Evolutionary pressures thus impel intelligent creatures toward increasingly sophisticated modes of communication and collaboration. The presumption of rationality is a crucial aspect of the concomitant processes of social interaction, since in its absence the basis of communication and collaboration is annihilated. We cannot communicate or collaborate with others in the absence of a supposition of (a fair degree of} rationality on their part. Its highly practical nature accordingly contributes importantly to validating the presumption of rationality.

Objectivity and Consensus Must genuinely rational minds ultimately reach agreement on meaningful issues? Does the fact that rationality is inherently universalistic in its objective and impersonal orientation mean that rational people have to

12

A NORMATIVE THEORY OF RATIONALITY

attain a consensus, that rationality leaves the scene where disagreement prevails? Not necessarily. For while in characterizing something as rational we are making a claim whose bearing is universal, it is perfectly conceivable that there might not be a universal consensus about the matter because only at a level of high generality is rationality absolute. Some theorists equate rationality with consensuality, deeming the prospect of attaining a consensus among all rational minds to be an indispensable requisite of reason. But there are deep problems here—not least in the cognitive case, where the oft-articulated idea that the lack of consensus undermines factuality is very questionable. No matter how plausible a contention on any significant issue (rationality included) may be, the prospect that some will (quite defensibly) dissent from it is ever present. Consensus appertains to rationality as an ideal, not as a realizable fact of life. The dismissive relativist argues, no consensus —> no objectivity.

This position is predicated on the principle that if different people can (justifiedly) think differently about some issue, then there just is no ob-

jective fact of the matter with regard to it. Such a person takes the line that where there is no relatively enforceable consensus (so that if two people disagree, then one of them is necessarily being unreasonable), then the issue is not an objective one: there is no “fact of the matter’; the issue is one of mere opinion, arbitrary decision, or arational (if not

outright irrational) allegiance. This attribution of dire implications to unavailable consensus dates back to the teachings of the skeptics of classical antiquity—to the last of the ten tropes, the arguments for skepticism inventoried by Sextus Empiricus in his Outlines of Pyrrhonism.® This argument pivots on the variation across the range of humankind’s culturally diverse views in matters of custom, manners, laws, and, above all, beliefs. Throughout the skep- |

tical tradition, this variation ofcustoms has been invoked to support a deconstructionism that takes a lack of consensus to betoken an absence of the objective factuality needed for meaningful deliberation. The prospect of different constructions, different interpretations or opinions, is

taken to annihilate the matter in view as an objective issue. The vitiating flaw of this position is that its salient mediating premise is completely unjustified. To validate a move from the thesis “People do not agree about X” to the conclusion “X does not represent a genuine issue of objective fact,” what is required is clearly an implicational link to... this effect: 8 See Sextus Empiricus, Outlines of Pyrrhonism, bk. 1, secs. 145-47.

THE LIGHT OF REASON

13

If X is an authentic factual issue, then people will, of necessity, ultimately come to agree about it.

But the closer one looks at this thesis, the less plausible it appears. Where is the Moses who has come down from the mountain with a stonegraven guarantee that whenever there indeed are facts, we imperfect mortals can come to discern them with an accuracy that precludes disagreement? Even perfectly rational people will agree only in those (very special) situations where the data at their disposal are both adequate and sufficient to sustain a uniform conclusion. With many kinds of clearly significant issues, we can readily account for failures to achieve consensus in ways that do not at all countervail the existence of facts of the matter. In the case of factual information, the

circumstance of evidential diversity, obtaining because different people have different bodies of evidence at their disposal, makes for a perfectly warranted—indeed rationally mandated—differentiation of belief. In the case of evaluation, the circumstance of experiential diversity (the fact that different people have different bodies of experience at their disposal) makes for a perfectly warranted—indeed rationally mandated—differentiation of appraisal. And the same holds for judgments about the appropriateness of actions. The validity of judgments of rationality is not destroyed by finding that there are some who dissent from them. The abstract thesis “Other things being equal, all rational people choose recognizably more effective problem-resolutions over less effective ones” is quite correct. But of course other things are not always equal. People can, quite defensibly, disagree about what sorts of measures are effective when different backgrounds of experience lead them to see matters differently.

Objectivity and Cognitive Consensuality Contrasted: The Import of the Ideal

Consensus among rational inquirers across the divides of time and culture is not something that often (let alone necessarily) happens in the real world. The link of consensus to factuality can be established tightly only at the level of idealization—only perfectly competent and well-informed inquirers need reach a consensus on factual matters. In this imperfect sublunary dispensation, consensus is too much to expect, given the variation of experience among intelligent beings. We are committed to the universality and objectivity of rational judgments because we see such judgments as implementing universal principles. But the sort of consen_ suality at issue with objectivity is a matter of ideal circumstances. It is

14

A NORMATIVE THEORY OF RATIONALITY

not something on which we should insist here and now. Objectivity stands coordinate not with actual but with ideal consensus. To insist on consensus as such is to maintain that in the actual circumstance people will reach agreement with respect to the issues. To insist on objectivity is to maintain that in ideal circumstances people would reach agreement. And these are quite different matters. The ideal demands of reason—its ultimate aspirations—are absolute. But in this world absolutes are never at our effective disposal. No human language, belief system, or thought framework is definitive and final—

delivered in unchangeable perfection to humankind by the world spirit from on high. We can do no more or no better than to use the local, particularized, diversified instruments that come to hand in the circum-

stances in which we find ourselves. Thus far, relativism is both inevitable and correct. But this emphatically does not engender an indifferentistic subjectivism. On the contrary, in rationally conducting our affairs, we become committed to trying for problem resolutions that ought to be binding on everyone. This normative dimension of the matter is crucial. It is the unifying cement of the ideal aspect of rationality (a crucial aspect

built into our determinative notion of what rationality is all about) which prevents rationality from becoming unraveled through a subjectivity engendered by the relativistic variability of our circumstantially conditioned rational proceedings. Rationality itself is, after all, a project that we are bound to pursue by variable means among the varied circumstances of a difficult world

where the consensus that objectivity ideally involves may well be unattainable in practice. (The extent to which reality cooperates with the demands of rationality is limited.) Rationality must, for us, remain something of an ideal that we can realize only to the limited extent that the circumstances of our situation permit. Nevertheless it is an ideal that, all considered, is eminently useful—and thus rational—for us to adopt and pursue. Here, as elsewhere, the validation of an ideal lies largely in the positive consequences that ensue from its adoption and cultivation.® ® This chapter’s theme of intelligence as a survival mechanism is further explored in my book A Useful Inheritance (Savage, Md., 1990). The topic of reason, rationality, and objectivity is treatedin Rationality (Oxford, 1988). Consensusand consensuality are the central issues of Pluralism (Berlin, 1992).

Two Cognitive Rationality and Risk: A Critique of Skepticism Is Cognitive Rationality Possible? A rational creature cannot manage its affairs intelligently in the absence of information. Reason’s commitment to the cognitive enterprise of inquiry is absolute, and it establishes an insatiable demand for extending and deepening the range of our information. As Aristotle observed, “Man by nature desires to know.” In this cognitive domain, reason cannot leave well enough alone but insists upon a continual enhancement in the range and depth of our understanding of ourselves and of the world about us. Cognitive rationality is a matter of using cogent reasons to govern one’s acceptance of beliefs—of answering one’s questions in the best feasible way. But is this project realizable at all? We must consider the skeptic’s long-standing challenge that it is not. For the skeptic—in his or her more radical moments, at any rate—insists that there just never is a satisfactory justification for accepting anything whatsoever. By rejecting the very possibility of securing trustworthy information in factual matters, skepticism sets up a purportedly decisive obstacle to an effective pursuit of the aims of reason. The skeptical challenge to the standard project of empirical inquiry based on established principles of cognitive rationality has a very plausible look about it. Our means for the acquisition of factual knowledge are unquestionably imperfect. Where, for example, are the “scientific truths” of yesteryear—those earth-shaking syntheses of Aristotle and .Ptolemy, of Newton and Maxwell? Whatever merits they undoubtedly had were flawed, and virtually no part of them has survived wholly unscathed. Given this past course of bitter experience, how can we possibly validate our present acceptance of factual contentions in a rationally convincing way? This question clearly indicates the need of coming to terms with a radical skepticism that maintains that rational cognition is unattainable—that the quest for knowledge is in principle a vain and vacuous pursuit.

,

To be sure, if it is indeed the case that rationally justified belief must always be based upon some prior, rationally prejustified inputs, then the skeptic prevails straight off. For then the process of rationally validating _ our accepted beliefs can never get started. To all appearances, we here

16

A NORMATIVE THEORY OF RATIONALITY

enter upon a regress that is either vitiatingly infinite or viciously circular. The rational justification of belief becomes in principle impossible —exactly as skeptics have always insisted. But this particular skeptical foray rests on.a false supposition. The rational justification of a belief does not necessarily require prejustifiedinputs. The idea that even as human life can come only from prior human life, so rational justification can come only from prior rational justification, is deeply erroneous. For the important distinction between discursive and presumptive justification becomes crucial here in.a. awaysthat skeptics conveniently overlook.

information-précessing terms, this ‘discursive. sortof justification is not

innovative but merelytransformatory as a production prdcess: ther 2 scursive ve justification is"not theonlysisortthere is.“Nor,will © eliefs do the trick, for there are-obviously tod”few-of them—and these few too feeble for the job, given the infeasibility ofany _ wholly un roblematic move from appearance to reality. However, ‘pre“Suimptive justification “uiilike.discursi sethrough:-the-evidential ditati 5

ifs:behilf-whilenno:valreadysi=jast ified When, after acareful look, I amunder the i impression that thereis a cat on the mat, I can (quite appropriately) base my acceptance of the contention “There actually is a cat on the mat,” which stakes a claim that goes beyond the reach of experience alone, not on any preestablished premises, but simply on my experience itself—on my visual impression. The salient consideration is that there just is no good reason why (in this case) ~ I should refrain from doing “what comes naturally” and endorse a visually

grounded belief of this kind as veridical. If there were counterevidence,

COGNITIVE RATIONALITY AND RISK

17

~ however (if, for example, I was aware of being in a wax museum), then the situation would, of course, be altered. ‘Presumptivély: stebeliefs:‘aréthe raw materials’ofcognition.They Otice,”thus permitting uus cation without a need for

overtummeéd; but'o ething «else yetm more'sectiré: some othétwise available conflicting’consideration. Theyare entitled to remain in place until:displaced by something superior. Accordingly, their impetus averts the dire consequences that would ensue if any and every cogent process of rational deliberation required inputs that themselves had to beauthenticated by a prior process of rational deliberation—in which case the whole process could never get under way. . This role of presumptions is absolutely crucial for cognitive rationality. Forthis mode:rati mpartinents;..the:diseursives(or conx tegoridal).The former iis a matter

d“principle . é pt their duly eévidentiated-conseqiiences. But‘of course this conditionalized principle cannot yield anything until one has already secured some acceptable theses from somewhere or other. And this is where substantive rationality comes.in, by enabling us to make categorical moves. Presumptions determineour “starter set” of initial commitments, enabling us to make a start on whose basis further inferential reasoning may proceed. - Presumptions accordingly playa crucial role in the cognitive sphere. “““FO¥-we cannot pursue the epistemic project—the quest for information about the world—without granting certain initial presumptions. Theyare _ reminiscent of Kantian conditions under which alone the securingof an-" swers to questions about the world is even possible. And prominent among these conditions is the consideration that we can take our data about the world as evidence, that a presumption of experiential veridicalityis typicallyin- order—a presumption that-is, to ‘be sure, defeasible, as when we enter a wax museum or indulge ourselvesin alcoholic drink.

In matters of sense perception, for example, we presume that mere appearances (the data) provide an indication of how things actually stand (however imperfect this indication may cventualy prove toobe. “Thatwe

oo

ive:éndeavors. If we systematically refuse, always and éverywhere,to accept seeming evidence as real evidence (at any rate, until the time comes when it is discredited as such), then we can get nowhere in the

18

A NORMATIVE THEORY OF RATIONALITY

domain of practical cognition. When skeptics reject any and all presumptions, they automatically block any prospect of anyone’s reasoning with them within the standard framework of discussion about the empirical facts of the world. The machinery of presumptions is part and parcel of the mechanisms of cognitive rationality; abandoning it aborts the entire project at the very outset, and standard probative practice of empirical inquiry stipulates a presumption in favor of such cognitive sources of information as the senses and memory. And the literature further contemplates such alternatives as:

iz Natural inclination: a “natural disposition” to accept (e.g., in the case of sense observation). 2? Epistemic utility in providing materials that would, if accepted, explain /— things that need explanation. Analogy with what has proved acceptable in other contexts. Tee

Fit: coherence with other accepted theses.

Even the weak sorts of grounding provided by analogy (the assimilation ~ of a present, problematic case to other similar ones) can prove itself an appropriate instrumentality of presumption through its utility in providing answers to questions.? All in all, presumption favors the usual and the natural—its tendencyis one of convenienceand ease of operation in cognitive affairs. Our cognitive proceedings accordingly incorporate ahostoffundamen-

tal presumptions of reliability, such as:

~“

~

Accept at face value the declarations of other people (in the absence of any counterindications and in the absence of any specific evidence undermining the generic trustworthiness of those others). , Trust in the reliability of established cognitive aids and instruments (telescopes, calculating machines, reference works, logarithmic tables, ete.) in i the absence of any specific indications to the contrary.

Principles of this sort are integral parts of the operational code of agents whotransacttheir cognitivebusiness rationally.

We'kriowthatvarious,highlyconvenient principles; of ductionzaresimply false. Whats ‘éms:to. 2. is.

What peoplésay's true. 1 This is why adherence to custom is a cardinal principle of cognitive as well aspractical | : rationality. See William James, “The Sentiment of Rationality,” in The Will to Believe and

Other Essays in Popular Philosophy (New York, 1897).

COGNITIVE RATIONALITY AND RISK

19

We realize full well that such generalizations do not hold, however nice it would be if they did.Nevertheless weaccept the theses at issue as

tie.”Our standard cognitive practices incorporate :ahost of of plane sible presumptions of initial credibility, in the absence of concrete evidence to the contrary. Two wo examples are the principles:

é(Ggains: rorabrenboounte adage)

Pasay

The validation of such presumptions is not the factual one of the substantive generalization “In proceeding inthis way, you will come at correct information and will not fall into error.”(After all,errors i

Our presumptivepracticesand policies arematters 0 ‘cognitive econ-

omy—of following the path of least resistance to a viablesolution of ou

paly.when.you,need.to,..al

_alution..of. lution..of-an.issue.* There is,vOf course, nothing sacrosanct about the result of such a procedure. The choice of the easiest way out may fail us; that which serves us well on first acquaintance may no longer do so in the end. But it is clearly the sensible way to begin. At this elemental levelof presumption we proceed by'‘doing what comes naturally.” Wha:

The. concession 0 presumptive status to our presystematic indications of credibility is the mostfundamental principle of cognitive rationality.

TLhezdétails“OF pre:

gousbles

light-ofevertialexperien 1 provab of-time; But without a programmatic policy for utilizing presumptions, “cognitive rationality cannot get under way at all. In the light of these considerations, the plausibility of skepticism takes on a rather different aspect.

20

A NORMATIVE THEORY OF RATIONALITY

Skepticism and Risk To put skepticism into a sensible perspective, it is useful to consider the demands of cognitive rationality in the light of risk taking. There are three very different sorts of approaches to risk, and three very different sorts of epistemic “personality” corresponding to these approaches. The general situation is summarized i4 n figure 2.1. ,

proach to riskisaltogether negative, Their motto n is something like, “Take no chances;playit safe;alwaysexpecttheworst.”

that a suitably Jarge gain beckons at safficlentiy auspicious odds.” It reflects the path ofprudence and caution.

The ‘Type2

‘(daringtIycaleulatin appe y

y large loss threatens at sufficientlyinauspicious odds.” It reflects the path of optimism and hopefulness. Types(risk-seeking): Osy-

7

nton thedelightful consequences ofafavorable issue of events: the sweet savor of success is alreadyin their nostrils. Risk seekers are chance takers and go-for-broke gamblers. They react to risk the way an old war-horse responds to the sound of musketry: with eageranticipation and positive relish for the fray. Their motto is, “Things will work out.’ . These basic attitudes condition both the extent to which people expose themselves to risk and the ways in which they respondin the situations of uncertainty in which they find themselves. In the conduct of practical Figure 2.1. Approaches to Risk Type 1: Risk avoiders Type 2: Risk calculators (1) cautious (2) daring

Type 3: Risk seekers

COGNITIVE RATIONALITY AND RISK

21

affairs the risk avoiders are hypercautious; with no stomach for uncertainty, they insist on playing it absolutely safe. In any potentially risky

situation, the mind of the risk avoider is given to imagining the myriad things that could go wrong. The risk calculators proceed with care; they take due safeguards but still run risks when the situation looks sufficiently favorable. The risk seekers, on the other hand, leap first and look later,

apparently counting on a benign fate to ensure that allwill be well; they dwell in the head atmosphere of“nothing can go wrong.” thr. ich’ tendencies ‘characterize people’s attitudes or personalities in i a way that is relatively stable over time. To be sure, armchair psychology teaches us that they are also predictably changeable within one’s life

span: as small children we tend to be risk avoiders; in youth we welcome risk; with maturity we become more calculating; in old age we incline to be risk aversive once more. (Actually, the situation is even more complex. For one and the same person will at any stage of the game differin his or her approach to different kinds of risk—for example, in being pre-

pared to take chances in their investments but not in their interpersonal relationships.)

attheother end of the spectrum. Like radical Popperians, such as P. K. _Reyerabend, “the: Srsynteretistsis,:inclined:to:think.,thatanythin »gae. The

lects only the chosen few”; ‘the syncretist inclinesevenly toward virtually everything. Somuch, then, for the three basic approaches to risk in various areas of endeavor. It is important to recognize that two fundamentally different sorts of misfortunes are possible in situations where risks are run and chances taken. . Misfortunes of the first kind: We reject something that, as it turns out, we should have accepted. We decline to take the chance and avoid running the risk at issue, but things tt urn out favorably after all, and we lose out on the gamble.

22

A NORMATIVE THEORY OF RATIONALITY

Misfortunes of the second kind: We do take the chance and run the risk at issue, but things go wrong, and we lose the gamble.

If we are risk seekers, we will incur few misfortunes of the first kind,

but—things being what they are—relatively many of the second kind will befall us. Conversely, if we are risk avoiders, we shall suffer few misfor-

tunes of the second kind but shall inevitablyincur relatively many ofthe first. The overallsituationisdepicted in figure22.2. _ These

eok adopt:themiddl he:positi can.to:balance:t

line is, “Neither avoid nor court risks, but manage them prudentlyinthe search for an optimal overall balance of fortune maximization and misfortune minimization,” and it insists on proceeding by way of carefully calculated risks. Figure 2.2. Risk Acceptance and Misfortunes Misfortunes of kind 1

Misfortunes of kind 2

‘

Number of (significant) misfortunes

| 50

0 Typel

(Risk

avoiders)

l 100

Type2.1 Type2.2

(Cautious

Type3

(Daring = (Risk

calculators) calculators) seekers)

Increasing risk acceptance (in % of situations) —->

COGNITIVE RATIONALITY AND RISK

23

In matters of cognition, skeptics succeed splendidly in averting misfortunes of the second kind. By accepting nothing, they accept nothing false. But of course they lose out on the opportunity to obtain any sort of information. Skeptics thus err on the side of safety, even as syncretists err on that of gullibility. In claiming that their position wins out because it makes the fewest _ mistakes, skeptics use a distorted system of scoring. While they indeed

make the fewest errors of one kind, they make the most of another. Once we look on this matter of “making mistakes” realistically, the vaunted advantage of skepticism vanishes. Skeptics are simply risk avoiders who are prepared to take no risks and stubbornly insist on minimizing errors of the second kind alone: After all, what we want in inquiry (the object ofthe whole enterprise) is information. What we seek is the very best overall balance between answers to our questions and the negativity of ignorance or misinformation.

We face a trade-off at this stage, however. Are we prepared to run a greater risk of error to sécure the potential benefits of greater understanding? The judicious cognitivist is a risk calculator who recognizes the value of understanding and is prepared to gamble for its potential benefits. In general, then, the course of wisdom is to position one’s standard of acceptance in the middle ground between the extremes—to try for the

optimal way to reduce errors overall. The details will of course hinge crucially on the exact form of the approximate curves of figure 2.2, as configured in the light of our standards for the appraisal of misfortunes. There is no single, uniquely rational specific choice among alternatives in the face of risk—some scope must even be allowed for the influence of one’s subjective constitution, the sort of person one is. But the course of

generic rationality clearly lies in a thoughtful endeavor to do the best one can overall, within the constraints of one’s situation.

The Deficiency of Skepticism The skeptic too readily loses sight of the very reason for being of our cognitive endeavors. The object of rational inquiry is not just to avoid error but to answer our questions, to secure information about the world. And here, as elsewhere, “Nothing ventured, nothing gained” is the operative principle. Granted, a systematic abstention from cognitive involvement is a surefire safeguard against errors of commission. But it affords this security at too steep a price. The shortcoming of that “no risk” option is that it guarantees failure from the very outset.

24

A NORMATIVE THEORY OF RATIONALITY

It is self-defeating to follow the radical skeptic into letting discretion be the whole of epistemic valor by systematically refusing to accept anything whatsoever in the domain of empirical fact. To be sure, when we

set out to acquire information, we may well discover in the end that, try as we will, success in reaching our goal is beyond our means. But we

shall certainly get nowhere at all if we do not even set out on the jour-

ney—which is exactly what the skeptic’s blanket proscription of acceptance amounts to.

:

In playing the game of making assertions and laying claims to credence, we may well lose: our contentions may well turn out to be mistaken. But in a refusal to play this game at all, we face not just the possibility but the certainty of losing the prize—we abandon any chance to realize our cognitive objectives. The skeptical policy of systematically avoiding acceptance is fundamentally irrational because it blocks from the very outset any plausible prospect of realizing the inherent goals of the enterprise of factual inquiry. In cognition, as in other sectors of life, there are no guarantees, no ways of averting risk altogether, no option that is totally safe and secure. The best and most we can do is to make optimal use of the resources at our disposal to manage risks as best we can. To decline to do this by refusing to accept any sort of risk is to become immobilized. The skeptic thus pays too great a price for the comfort of safety and security. If we want information—if we deem ignorance no less a negativity than error—then we must be prepared to take the gamble of answering our questions in ways that risk some possibility of error.

A middle-of-the-road evidentialism emerges as the most sensible approach. . Perhaps no other objection to radical skepticism in the factual domain

is as impressive as the fact that, for the all-out skeptic, any and all assertions about the world’s objective facts must lie on the same cognitive plane. No contention—no matter how bizarre—is any better off than any other in point of its legitimative credentials. For the thoroughgoing skeptic there just is no difference-relevant to rationality between “More than three people are currently living in China” and “There are at present | fewer than three automobiles in North America.” As far as the cognitive venture goes, it stands committed to the view that there is nothing to choose in point of warrant between one factual claim and another. Radical skepticism is an H-bomb that levels everything in the cognitive domain. The all-out skeptic writes off at the very outset a prospect whose abandonment would be rationally defensible only at the very end. As Charles

Sanders Peirce never tired of maintaining, inquiry has a point only if we._ accept from the outset that there is some prospect that it may terminate in a satisfactory answer to our questions. He indicated the appropriate stance with trenchant cogency: “The first question, then, which I have to

_

COGNITIVE RATIONALITY AND RISK

25

ask is: Supposing such a thing to be true, what is the kind of proof which I.ought to demand to satisfy me of its truth?”? A general epistemic policy that would as a matter of principle make it impossible for us to discover something which is ex hypothesi the case is clearly irrational. And the skeptical proscription of all acceptance is obviously such a policy—one that abrogates the project of inquiry at the very outset, without according it the benefit of a fair trial. A presumption in favor of rationality (cognitive rationality included) is rationally inescapable. It could, to be sure, eventuate at the end of the day that satisfactory knowledge of physical reality is unachievable. But until the end of the proverbial day arrives, we can and should proceed on the idea that this possibility is not in prospect. “Never bar the path of inquiry,” Peirce rightly insisted. The trouble with skepticism is that it aborts inquiry at the very start. The skeptic’s favored approach is geared to the level of individual contentions (propositions). Skepticism insists on addressing the issue of rational acceptance at the level of particular theses: are we really in a position to accept this or that p or not? And then the skeptic finds various seemingly plausible grounds for not doing so. But the real issue is methodological and relates to our general procedure with regard to acceptance. The skeptic’s mistake is one of omission, failing to see that the fundamental issue we face is not just that of particular acceptances but that of a choice between policies of acceptance. Initially, the question “Is the particular claim p justified in the prevailing circumstances?” is best approached indirectly. The sensible first move is to step back and begin with the question “What are the appropriate methods (standards, criteria) for acceptability determination?” Once this issue is resolved, we then have in hand the instruments for

deciding the justificatory status of p. To deal effectively with skepticism we had best begin by dealing with methods (standards, criteria) rather than particular theses. Our basic concern is—and should be—with the state of things at the level of cognitive policies, of general practices with regard to presumption and evidence.® We clearly require some broad policy or procedure of acceptance, some general stance that views particular claims in the light of general ‘ principles. And whatever policy or procedure we adopt is one that we must be prepared to justify. But there is a perfectly sensible nonskeptical approach, one that indeed is validated by the sort of systemic considerations that we have canvassed—namely, that of minimizing, as much as possible, the overall extent to which one runs into problems and difficul2C. S. Peirce, Collected Papers, 8 vols., ed. C. Hartshorne et al. (Cambridge, Mass., 1931-50), vol. 2, sec. 2.112.

3 This position is set out in fuller detail in the author’s Methodological Pragmatism (Oxford, 1977).

.

26

A NORMATIVE THEORY OF RATIONALITY

ties. And just this is the stance of a cautiously managed evidentialism of the sort at issue in scientific method. Of course, once we have an acceptance policy in place, then the issue of propositional acceptance in those particular cases against which the skeptic rails is now settled for us. Our hands are tied and no longer free: what it is appropriate to accept is now something that the general policy we have adopted requires of us, because that is exactly the sort of policy it is—an acceptable policy. At this point, with a general policy already in place, the option of skepticism lies behind us—on the far side of a-point

of no return.

.:

Such a refutation of skepticism does not proceed at the item level o: showing that the skeptic’s view of the matter is untenable in this or that particular case. Rather, it proceeds at the policy level, showing the pragmatic superiority of adopting a line of approach at variance with the skeptic’s. Viewing the issue in this pragmatic light, we see that the skeptic’s risk-avoidance policy is simply not one that is rational to adopt. It has to be recognized that any systematic cognitive policy that we could possibly implement in the real world is bound to allow some errors to arise. Given _ that errors can be of the two kinds we have been considering, there simply is no realistic way of averting errors altogether, all across the board. Now, the skeptic’s problem is that the preferred stance (“Accept nothing!”) represents a particular policy choice designed to avert errors of the first kind alone. Such an approach is not only not mistake-proof but rather a mistake-inviting, definite commitment of procedure—even though its only mistakes are those of the first kind. Viewed from the vantage point of representing a particular policy alternative, the skeptic’s problem is that this position avoids mistakes of commission only, without reckoning the overall price of maximizing those of omission.

Against Skepticism: The Pragmatic Dimension To be sure, even the most committed of skeptics have historically recognized that people must act to survive and thrive in the world. Like everyone else, the skeptics acknowledge that we humans find ourselves em~ placed in medias res in an environment that will not satisfy our needs,

wants, and desires automatically, without intervention on our part. This concession opens skepticism to the charge of immobilizing action, projecting a refutation of skepticism on the grounds that it makes the conduct of life itself impossible. David Hume put this point as follows: “But_ a Pyrrhonian . . . must acknowledge, if he will acknowledge anything, — thatall human life must perish, were his principles universally and steadily to prevail. All discourse, all action would immediately cease; and men ~

i

COGNITIVE RATIONALITY AND RISK

7

remain in a total lethargy, till the necessities of nature, unsatisfied, put

an end to their miserable existence.”* To this sort of charge, the ancient skeptics always replied that while we must indeed act, this action need not be based on knowledge at all.

They have insisted on the sufficiency of noncognitive guides for action— appearances, custom, the general consensus, inclination, instinct, or the

actu

is.or-that is.reall thet‘case. Life without

knowledge orrationally eevidencedbelief is certainly riot in principleimpossible: lower animals, for example, manage very well. Or again, a somewhat less radical strategy is available, one that puts acceptance (and belief) on a wholly unreasoned basis (say, of instinct, inclination, fash-

ion, or the like). The skeptic can accordingly hold—and act on—all the beliefs that people ordinarily adopt, with only this difference, that he or she regards them as reflecting mere appearances and denies that the holding of these beliefs is rationally justified. Hume’s charge seems to hit ' wide of the mark: skepticism need not immobilize action. Still, this sort of defense misses the real point. The core of the objection is not simply that skeptics fail to favor some assertions over others as a basis for action. It is rather that they insist that there just are no reasons for doing so. And this is deeply problematic. For while skeptics may indeed have guides to action (namely, the noncognitive guides of instinct or custom, for example), they simply cannot defend their actions. They cannot justify doing A rather than B. They can tell us that they eat food -to assuage hunger (rather than merely rubbing their stomachs) because this is the done thing, but this mere explanation of what they do-does not constitute any ground for it. Skepticism thus destroys the prospect of any rational basis for praxis. Given the skeptic’s total suspension of judgment, our behavior becomes not necessarily irrational but altogether unrationalizable. The linkage between action and rationality is disastrously severed. In fact, rationality must be abandoned altogether. (And of ’ course no cogent reason for taking this course can be provided.) In many real-life situations we are not so totally and unqualifiedly confident of our beliefs that we are willing “to bet the family farm” on them. These noncategorical beliefs are indeed beliefs; we accept them as true, 4 David Hume, An Enquiry concerning Human Understanding, sec. 12, pt. 2. Cf. John Locke: “He that will not eat till he has demonstration that it will nourish him; he that will

not stir till he infallibly knows the business he goes about will succeed, will have little else to do but sit still and perish” (Essay concerning Human Understanding, bk. 4, chap. 14, _ see. 1),

28

A NORMATIVE THEORY OF RATIONALITY

not just as likely or plausible. But our acceptance is a hesitant one. (And note the crucial difference between actually accepting something as true, however hesitantly, and viewing it as probably true.) Now, when it comes to the implementation of such guarded beliefs and unenthusiastic acceptance in situations that will exact real costs if we prove to be wrong, we would certainly want to hedge our bets. Consider the situation depicted in figure 2.3. Ina situation of this sort, an expected value calculation along the standard lines would engender the following result: EV(act) = pX + (1 — p)x = p(X ~— x) + 2, EV (play safe) = 0.

On this basis, it follows that: EV (act) > EV (play safe) iff p >

>

Thus on the standard principles of decision-theoretic rationality, it transpires that only if p is sufficiently large relative to the balance of the potential costs and benefits at issue will acting on the belief at issue be appropriate and advisable. Action on a (noncategorical) belief is not automatically validated by its status as such but hinges crucially on the detailed circumstances of the situation. The rational agent need not betimpelled to action by very weakly established beliefs. Accordingly, if all that a skeptic wanted was to remind us of the important fact that we should not automatically act on all our beliefs but should be mindful of the potential hazards involved—that we should not regard our beliefs as sacrosanct fixities—then these strictures would be altogether sensible and appropriate. But Pyrrhonian skeptics are much too radical for this cautious line, insisting that we should abstain altogether from belief and accept nothing whatsoever. And at this point, skepticism leaves mere prudent caution behind and slips into an unduly distrustful paranoia that simply turns its back on the life of reason. It is thus not difficult to discern the unsatisfactory nature of the skeptic’s noncognitive strategies for the guidance of action. For in being noncognitive, these approaches are also nonrational (which is, to be sure, | Figure 2.3. A Problem of Choice between BeliefImplementation and Inaction . Eventuations

Resulting payoffs “If Lact on my belief If I play safe

1. If matters are as I believe (probability p)

Xx

0

2. If matters are not as I believe (probability 1 — p)

x

0

COGNITIVE RATIONALITY AND RISK

39

something quite different from the irrational), While they indeed tell us what to do, they are stonily silent on the crucial issue of why it is to be

done. And this is eminently unsatisfying. The human being is an ineradicably rational animal, a creature that is committed to satisfying not only its physical hunger but its intellectual hunger to know “the reason why.” The insistence on noncognitive guidance is a line that skeptics themselves may willingly take, but there is no earthly reason why those of us ‘who are not already precommitted to their views should join them in doing so.

Justifying Presumptions To be sure, the risk of deception and error is present throughout our inquiries: our cognitive instruments, like all other instruments, are never

failproof. Still, a general policy ofjudicious trust in certain presumptions is eminently cost effective. In inquiring, we cannot investigate everything; we have to start somewhere and invest credence in something. But

of course our trust need not be blind. Initially bestowed on a basis of mere hunch or inclination, it can eventually be tested and can come to

be justified with the wisdom of hindsight. And this process of testing can in due course put the comforting reassurance of retrospective validation at our disposal. To be sure, the question remains: “What sorts of considerations validate our particular presumptive practices as such, and how is it that presumptions become entitled to this epistemic status?” The basis of the answer has already been indicated. A twofold process is involved. Initially, it is a matter of the generic need for true presumptions continued with the mere convenience of these presumptions. But ultimately, a thisor-nothing recognition of the validity of a presumption emerges ex post facto through the utility (both cognitive and practical) of the results it yields. Legitimation is thus available, but only through experiential retrovalidation, orretrospective validationin the light of eventual experience.®

something, we runa riskof being let down and disappointed. Nevertheless, it seems perfectly reasonable to bet on the general trustworthiness of the senses, the general reliability of our fellow human beings, and the general utility of reason. In such matters, no absolute guarantees can be 5 See the author's Methodological Pragmatism for a fuller development of this line of thought.

30

A NORMATIVE THEORY OF RATIONALITY

had. But one may as well venture, for if venturing fails, the cause is lost anyhow—we have no more promising alternative to turn to. There is little choice about the matter: it is a case of “this or nothing.” If we want answers to factual questions, we have no real alternative but to trust in the cognitively cooperative disposition of the natural order of things. We cannot preestablish the appropriateness of this trust by somehow demonstrating, in advance of events, that it is actually warranted. Rather, its rationale is that without it, we remove the basis on which alone creatures

such as ourselves can confidently live a life of effective thought and action. In such cases, pragmatic rationality urges us to gamble on trust in reason, not because it cannot fail us, but because in so doing, little is to

be lost and much to be gained. These considerations combine to indicate

that our cognitive practices have a fundamentally economic rationale. A general policy of judicious trust in our standard cognitive sources and resources is eminently cost effective within the setting of theproject of inquiry to which we stand committed (by our place in the world’s scheme of things). We proceed in cognitive matters in much the same way that banks proceed in financial matters. We extend credit to our informative sources and resources, doing so at first to only a relatively modest extent. When

and if they comport themselves in a manner that shows that this credit was well deserved and warranted, we proceed to give them more credit

and extend their credit limit, as it were. They improve their credit rating in cognitive contexts much as people and institutions do in financial contexts. The same sort of mechanism is at work in both cases: recognition of credit worthiness engenders a reputation on which further credit can be based; earned credit is like money in the bank, well worth the measures needed for its maintenance and for preserving the good name that is now at stake. The example of our senses-is a particularly important case in point. Consider the contrast between our reaction to the data obtained in sight and those obtained in dreams. Dreams, too, are often vivid and significant seeming. Why then do we accept sight as a reliable cognitive source-

but not dreams—as the ancients were generally minded to do? Surely not because of any such substantive advantages as vividness, expressiveness, or memorability. The predisposition to an interest in dreams is clearly attested by their prominence in myth and literature. The higher status we accord to sight is not a consequence of its intrinsic preferability but is preeminently a result of its success in building up credit in just the way we have been considering. We no longer base our conduct ofaffairs on dreams simply because we (our culture at large) have learned that it ~ does not pay. The methodological principles of presumption that characterize the

way in which rational agents transact their cognitive business are not

COGNITIVE RATIONALITY AND RISK

31

adopted because we can somehow establish their abstract validity, but because the concrete cost-benefit advantage of adopting them is so substantial. The justification of trust in our senses, in our fellow inquirers,

and in our cognitive mechanisms ultimately rests on such considerations of economic rationality. The pursuit of knowledge in rational inquiry can play a role akin to that

of a pursuit of wealthin business transactions. The financial markets in stocks or commodities futures would self-destruct if the principle “My word is my bond” were generally abrogated, since there could now be no way of telling whether or not a trade had actually been made. In just this way, too, the market in information would self-destruct if people’s truthfulness could not be relied upon. Thus in both cases, unreliable people have to be frozen out and exiled from the community. In cognitive and economic contexts alike, the relevant community uses both sanctions (ar-

tificially imposed costs and benefits) and positive incentives to put into place a system where people generally act in a trusting and trustworthy way. Such a system is based on processes of reciprocity that advantage virtually everyone. The harsh measures standardly employed to uphold the integrity of science (e.g., the destruction of careers through ostracism from the community for “breaking the rules”) are thus not devoid of rational justification. Cheating by data falsification, for example, is worth eliminating at great cost, because its toleration endangers and undermines the fabric of mutual trust, in whose absence the whole enterprise of collaborative inquiry in science becomes infeasible. Establishing and maintaining a community of inquirers united in common collaboration by suitable rewards and sanctions is a mode of operation that is highly cost effective. Individual probity and mutual helpfulness are virtues whose cultivation not only yields its own rewards but also pays ample dividends for the community of inquirers. In these matters, the cold iron hand of individual and com-

munal interest lies behind the velvet glove of etiquette and ethics. The commodity of information illustrates rather than contravenes the division of labor that results from Adam Smith’s putative innate human “propensity to truck, barter, and exchange.” The market in knowledge has pretty much the same nature and the same motivation as any other sort of market—it is a general-interest arrangement.

Skepticism versus Rationality The skeptic seemingly moves within the orbit of rationality by insisting that the knowledge we claim to get is never quite good enough—but only seemingly so. For in fact, skepticism runs afoul of the only promising ' epistemological instrumentalities that we have. Philosophical skeptics

32

A NORMATIVE THEORY OF RATIONALITY

generally set up some abstract standard of absolute certainty and then try to show that no knowledge claims in a certain area (sense, memory, scientific theory, and the like) can possibly meet the conditions of this standard. From this circumstance, the impossibility of such a category of knowledge is accordingly inferred. But this inference is totally misguided. For what follows is rather the inappropriateness or incorrectness of the standard at issue. If the vaunted standard is such that knowledge claims cannot possibly meet it, the moral is not “too bad for knowledge claims” but “too bad for the standard.” Any position that precludes in principle the possibility of valid knowledge claims thereby effectively manifests its own unacceptability. The skeptic’s argument is a double-edged sword that cuts both ways and inflicts the more serious damage upon itself. It is senseless to impose on something qualification conditions that it cannot in the very nature of things meet. An analogue of the old Roman legal precept is operative here—one is never obliged beyond the limits of the possible (ultra posse nemo obligatur). It cannot rationally be required of us to do more than the best that is possible in any situation—cognition included. But rationality also conveys the comforting realization that we are entitled to regard the best that can be done as good enough. Ina last-ditch defense, the radical skeptic may still take the following line: “Departing from your cognitivist view of ‘rationality’ and falling short by your standards is not something that I need regard as a genuine failing. Indeed, my very thesis is that your ‘rationality’ has no suitable credentials.” To meet this desperate but profound tactic, we must shift the ground of argumentation. It now becomes advantageous to approach the whole issue from a new point of departure, namely, that of the prospects of communication. For it turns out that not only does the refusal to accept claims preclude the radical skeptic from participating in the enterprise of inquiry, he or she is blocked from the enterprise of communica-

tion as well, All informative communication is predicated on the fundamental convention that normally and standardly what one declares to be so is something that (1) one accepts as true, and (2) one claims to be rationally warranted in accepting.® In rejecting the ground rules of our reasoning as inappropriate, the skeptic also abandons the ground-rules of our communication. In denying the prospect of any sort of rational warrant, 6 Peter Unger suggests a “reformation” of language in the interests of skepticism. “We want linguistic institutions and practices where our [universal] ignorance will not enjoin

silence” (Ignorance [Oxford, 1975], p. 271). But he offers no concrete proposals along these ~ lines, and this is very understandable. It is difficult to see how a language could be formed, taught, and above all used in which assertion played no role and declaration carried with it no claims to veracity. And even if (per impossible) one had such a language, what would be the point of using it in communicative contexts?

COGNITIVE RATIONALITY AND RISK

33

however tentative, the all-out skeptic embarks upona self-imposed exile from the community of communicators, seeing that the communicative

use of language is predicated on conceding the warranting presuppositions of language use. To enter into a discussion at all, one must acqui-

esce in the underlying rules of meaning and of information transmission that make discussion in general possible. (After all, if one suspends judg_ ment regarding whether “dog” means dog—rather, say, than cat or oyster—there is little one can gain from people talking about dogs.) But if nothing can appropriately be accepted, then no rules can be established——and thus no statements made, since meaningful discourse re-

quires an entry into the informative conventions.If skeptics deny us (and themselves) knowledge, or at least plausible beliefs, about the meanings of words and symbols, then there is just no way they can communicate with us (and perhaps even with themselves). A radical skepticism, in the final analysis, engenders the collapse of communication and deliberation, thereby leading to a withdrawal from the human community. Historically, to be sure, committed skeptics such as Sextus Empiricus have been prepared to grant that they are destroyers of logos (reason and discourse), insisting that they use reason and discourse simply as instruments for their own destruction. They have been prepared to grasp the nettle and admit—perhaps even welcome—the consequence that their position does not only gainsay knowledge but actually rejects the whole project of cognitive rationality (at any rate, at the level of our factual beliefs). But whatever satisfaction this drastic posture may afford the skeptic, it has little appeal to those who do not already share this position, since (on its own telling) no cogent reason can be given for its adoption. The collapse of the prospect of rational inquiry and communication is the : ultimate sanction barring the way to any rational espousal of radical skepticism. It is a price that a fanatically dedicated devotee of the skeptical position may be willing to meet, but it is clearly one that people who are not so precommitted cannot possibly pay. Rational people require cogent reasons for what they do, and only with the abandonment of a rigoristic all-out skepticism can such reasons be obtained. The only sort of critique of skepticism that makes sense to ask for is a rational critique. And viewed from this standpoint, the decisive flaw of skepticism is that it makes rationality itself impossible.

Skepticism and Praxis A viable theory of knowledge must somehow reconcile two factors: (1) a distributive epistemic fallibilism that acknowledges the fragility and defeasibility of our various particular claims to objective factual knowledge, - and (2) a collective cognitivism that realizes that we do actually have a

34

A NORMATIVE THEORY OF RATIONALITY

good deal of knowledge. The tradition that comes to us from C. S. Peirce via C. I. Lewis teaches that pragmatism provides the reconciling factor here: the systematic success we secure in practical implementation provides some assurance of collective adequacy in the face of distributive fragility. But how can such a pragmatic perspective be justified? How can pragmatism militate against a synoptic skepticism? Since the days of Greek antiquity, philosophers have often answered our present question “Why accept anything at all?” by taking the following stance: Human beings are rational animals. As such, they must act, since their very ©

survival depends upon action. But qua rational beings, they cannot act availingly except insofar as their actions are guided by what they accept.

This line of argumentation was articulated by the ancients and revived in modern times by a succession of thinkers from David Hume to William James.? Note, however, that the position of our present discussion does not run, “If you want to act effectively, then you must accept something.” Rather, its line is, “If you want to enter into ‘the cognitive enterprise — that is, if you wish to be in a position to secure information about the world—then you must be prepared to accept something.” Both approaches take a stance that is not categorical and unconditional but rather hypothetical and conditional. But in the classically pragmatic case, the focus is upon the requisites for effective action, while our present, cognitively oriented approach focuses upon the requisites for rational inquiry. In the present perspective, it is the frustration of our fundamental cognitive aims and aspirations (no matter how much skeptics themselves may be willing to turn their back upon them) that provides the salient theoretical reason for the rejection of skepticism. The argument against skepticism deployed here is thus at bottom also an essentially practical one. It-does not establish the internal inconsistency or theoretical untenability of a skeptical position. Rather, it shows that the price we would pay in taking such a position is so high as to outweigh any real benefit that could possibly accrue from it. But this price is a price exacted in the sphere of theory—in the area of understanding. That basic demand for information and understanding presses. in upon us, and we must indispensably adopt effective means to get.it

satisfied. Whatever the merits or demerits of skepticism as a theoretically tenable position, we are entitled on practical grounds to dismiss it unceremoniously. For our human practice is inherently cognitive. To reem7 As William James said, “[Someone] who says ‘Better to go without belief forever than believe a lie!’ merely shows his own preponderant private horror of becoming a dupe. . . but I can believe that worse things than being duped may happen to a man in this world” (The Will to Believe, pp. 18-19).

COGNITIVE RATIONALITY AND RISK

35

phasize an earlier point, cognitive deprivation is as debilitating for us as dietary deprivation. After all, merely theoretical argumentation cannot dislodge the skeptic from the stance of accepting no thesis at all. Argumentation is unavailing because any probatively cogent argument must proceed from conceded premises, and the skeptic can always simply refuse to make concessions. _All that argumentation can do is to forestall skepticism by showing the incompatibility of the skeptic with positions acceptable to sensible people in general. Such argumentation may not dislodge someone from a skeptical position, but it should prevent somebody who has not taken this position from doing so in the first place. The presumption of truthfulness with which we approach the cognitive enterprise is based on a certain idealization of the trustworthiness of our cognitive sources and resources that we realize full well is not actualized in the real world. But we nevertheless adopt this idealization and commit ourselves to it because doing so eminently serves ourcognitive purposes—indeed is, in a way, indispensably essential to them. Skepticism defeats from the very outset any prospect of realizing our cognitive purposes and aspirations. It runs counter to the purposive enterprise to which we humans by our very nature stand committed. It is ultimately this collision between skepticism and our goal-oriented position in the world’s scheme of things that makes the rejection of skepticism a rational imperative. A radical skepticism conflicts not only with the realities and practicalities of the human situation but also with its no less important idealities.® 8 This chapter’s themes are also the subject of my Rationality (Oxford, 1988) and Scepticism (Oxford, 1980). The importance of the practical turn was stressed inmy Primacy of Practice (Oxford, 1973) andin Methodological Pragmatism.

Three Evolution as an Allocation Mechanism

Why Are We So Smart? Knowledge acquisition lies at the very core of our being as Homo sapiens. But how did this come to be so? What sort of explanatory rationale can be. providedfor our cognitive abilities? Why are-we humans so smart? How is it that we possess the intellectual talent to create mathematics, medicine, science, engineering, architecture, literature, and other com-

parably splendid intellectual disciplines? What explains the immense

power of our cognitive capacities? The general direction of the answer to the problem of human intelligence is relatively straightforward. Human intelligence is the product of a prolonged process of biological evolution. There are many ways for an animal species to make its way in the world. Diverse alternative modes of coping within nature present themselves to biological organisms: the routes of multiplicity, toughness, flexibility, and isolation, among others.

But one promising pathway is clearly afforded by the route of intelligence, of adapting by the use of brain rather than brawn, of cleverness rather than power, of flexibility rather than specialization. There is a fertile ecological niche for a creature that makes its way in the world not by sheer tenacity or by tooth and claw but by intelligence—by coordinating its own doings and the world’s ways through cognitive foresight. “Humanity’s possession of intelligence and the capacity for reason are thus.to be accounted for on evolutionary principles. For these resources are clearly a means to adaptive efficiency, enabling us (sometimes at least) to adjust our environment to our needs and wants rather than the’ reverse. It is not all that difficult to visualize how intelligence—geared, as it is, to the pursuit of cogency, éfficiency, and optimality—can. be of. assistance in militating toward arrangements advantageous to life and its _ propagation. Reason-deploying intelligence—the use of our brains to guide action by figuring out the apparent best—is the survival instrument of Homo sapiens. Intelligence is our functional substitute for the numerousness of termites, the ferocity of lions, or the toughness of microorgan-_ isms. It constitutes our particular competitive advantage in the evolution- ~

ary scheme of things. As Darwin himself already stressed, in a competitive Darwinian world a creature that can understand how things

EVOLUTION AS AN ALLOCATION MECHANISM

37

work in its environment and exploit this understanding in action thereby secures an evolutionary edge. Basically, then, we are so smart because that is our developmentally assigned place in evolution’s scheme of things. Different sorts of creatures have different ecological niches, different specialities that enable

them to find their evolutionary way down the corridor of time. Some are highly prolific, some physically tough, some swift of foot, some hard to spot, some extremely shy. Homo sapiens is different. For the evolutionary instrument of our species is intelligence, with everything that this involves in the way of abilities and versatilities. This.capability of ours is not all just a matter of luck—of fate’s lottery bringing intelligence our way. Evolution’s bioengineering is the crucial factor. Bees and termites can achieve impressive prodigies of collective effort. But insects developed under the aegis of evolution could not become as smart as humans because the information-processing needs of the life-style opportunities afforded by their physical endowment are too modest to push them to the development of intelligence. The fact is that smarts are an inherent concomitant of our physical endowment. ~ Our bodies have many more independently movable parts (more “degrees of freedom”) than those of most other creatures of anything like our size.t And this circumstance has significant implications. For suppose a system with n switches, each capable of assuming an ON or OFFposition Then there are 2" states in which the system can find itself. With n = there are only 8 system states, but with n doubling to 6, there aresiready 64 states. As a body grows more complex and its configuration takes on more degrees of freedom, the range of alternative possible states expands rapidly (exponentially). Merely keeping track of its actual position is already difficult. To plan ahead is more difficult yet. Ifthere are m possible states that the system can assume now, then when it comes to selecting its next position, there are also m choices, and for the next two there are m X m alternatives overall (ignoring unrealizable combinations). So with a two-step planning horizon, the 3-state system has 64 alternatives, while the 6-state system has 4,096. With a mere doubling of states, the planning problem has become complicated by a factor of sixty-four. The degrees of freedom inherent in variable movement over time are pivotal factors here. The moment one walks upright and begins to develop the modes of motion that this new posture facilitates (by way of creeping, running, leaping, etc.), one has many more problems of physical management to solve. Considerations of this sort render it evident 1 The human skeleton has some 220 bones, about the same number as a cat when tail bones are excluded. A small monkey has around 120. Of course, what matters for present

purposes is independently moving parts. This demotes thousand-leggers and—thanks to fingers, among other things—takes us out of the cat’s league.

38

A NORMATIVE THEORY OF RATIONALITY

that a vertebrate having a more highly articulated skeleton, with many ’ independently operable bones and bone complexes, faces vastly greater difficulties in management and manipulation in what military jargon calls “command and control.” Physically more versatile-animals have to be smarter simply because they are physically more versatile. We humans are driven to devising greater capabilities in information acquisition and management by the greater demands of the life-style of our ecological niche. The complexity of our sophisticated surveillance mechanisms for friend-or-foe identification is an illustration. We can observe at a considerable distance that people are looking at us, discriminating minute differences in eye orientation in this context. The development of our sophisticated senses with their refined discrimination of odors, colors, and sounds is another example. Environmental surveil-

lance is crucial for our life-style. We have to know which feature of our environment to heed and which can safely be ignored. The handling of

such a volume of information calls for selectivity and for sophisticated processing mechanisms—for intelligence, in short. Not only must our bodies be the right size to support our physical operations and activities, but our brains must be so as well. The complexities of information management and control pose unre-

lenting evolutionary demands. To process the large volume of information needed to operate a sophisticated body, nature must fit us out with ~ a large brain. A battleship needs more elaborate mechanisms for guidance and governance than a rowboat. A department store needs a more extensive managerial apparatus than a corner grocery. But-a largé; sophisticated brain requires a large, complex body. The evolution of the human brain is the story of nature’s struggles to provide the machinery of information management and control needed by creatures of increasing physical versatility. A feedback cycle comes into operation—a complex body requires a larger brain for command and control, and sustaining a larger brain requires a larger body whose operational efficiency in turn places greater demands on that brain for the managerial functions required to provide for survival and the assurance of a posterity. As can _ be illustrated by comparing the brain weights of different mammalian species, the growing complexities and versatilities of animal bodies involve a physical life-style whose difficulties of information processing and management require an increasingly powerful brain. How one makes one’s living matters: insect-eating and fruit-eating monkeys have heavier brains, for their size, than those of leaf-eating ones.” 2 At any given time in evolutionary history, the then-current herbivores tended to have smaller brains than the contemporary carnivores. See Richard Dawkins, The Blind Watchmaker (London, 1981), p. 190.

EVOLUTION AS AN ALLOCATION MECHANISM

39

Here then is the straightforward (and rather trivial) answer to our question: We are as intelligent as we are because that is how we had to evolve to fill our place in nature’s scheme of things. We are so smart because evolution’s bioengineering needs to provide those smarts for us to achieve and maintain the life-style appropriate to our ecological niche. If we were not so intelligent, we would not be here as the creatures we are. We have all those splendid intellectual capacities because we require them in order to be ourselves.* But there remains the problem of why evolution would take this course? Surely we did not need to be that smart to outwit the sabertoothed tiger-or domesticate the sheep. Let us explore this aspect of the matteralittle. The things we have to do to manage our life-style must not only be possible for us, they must in general be easy for us (so easy that most of them can be done unthinkingly and-even unconsciously). If our problemsolving intelligence were frequently strained to the limit, often groaning under the difficult burden of the problems it is called on by nature to resolve in the interests of our life-style, then we just would not have it.

For evolution to do its work, the survival problems that creatures confront have to be by and large easy for the mechanisms at their disposal. And this fundamental principle holds just as true for cognitive as for biological evolution. If cognitive problem-solving were too difficult for our mental resources, we would not evolve as problem-solving creatures. If we had to go to as great lengths te work out the sum 2 + 2 as to extract the cube root of a number, or if it took us as long to discriminate 3- from 4-sided figures as it takes to discriminate between 296- and 297-sided ones, then these sorts of issues would simply remain outside our cognitive repertoire. The average problems for survival and thriving that are posed by our life-style must be of the right level of difficulty for us—that is, they must be rathereasy. And that calls for excess capacity. All of the ordinary problems of one’s mode of life must be solvable quicklyin real time, and with enough idle capacity left over to cope with the unusual. A brain that is able to do all that is needed to sustain the life of a complex and versatile intelligent creature will remain. under-utilized much of the time. To cope at times of peak demand, it will have a great 3 Of course here—as elsewhere—we cannot let matters rest with speaking of an evolutionary process in this rather anthropomorphic way. In the final analysis, we have to cash in these metaphors in terms of different groups (tribes, clans) of humanoids chancing to

produce a bumper crop of more than ordinarily intelligent individuals and finding themselves at a reproductive advantage thereby because of their comparatively greater and better ensured prospects, etc. But no imaginative student of recent demographic phenomena will find difficulty in envisioning an appropriate scenario here.

40

A NORMATIVE THEORY OF RATIONALITY

deal of excess capacity to spare for other issues at slack times. And so, any brain powerful enough to accomplish those occasionally necessary tasks will have the excess capacity at most normal times to pursue various challenging projects that have nothing whatever to do with survival. These considerations resolve the objection that evolution cannot explain our smarts because we are a lot smarter than evolution demands— that, after all, evolution does not set us examinations on higher mathematics or theoretical physics. The point is not that such disciplines somehow afford humans with an evolutionary advantage but, rather, that the capacities and abilities that make these enterprises possible are evolutionarily advantageous—that evolution equips us with a reserve capacity

that makes these activities feasible as a side benefit. An intelligent creature whose capacities do not allow of development in these directions just is not smart enough to pass evolution’s examinations in other more fundamental matters and so would not be able tomake intelligence its evolutionary specialty after all. The brain-computer analogy proves helpful in this connection. Very different things can be at stake with being “simple”: the simplicity of hardware at issue with comparatively less complex computers is one sort of thing, and the simplicity of software at issue with comparatively less complex programs is something quite different. And there are clearly trade-offs here. Solving problems of the same level of difficulty is gener-.. ally easier to program on a more sophisticated (more complex) computing machine. Something of an inverse relationship obtains: greater machine complication can make the actual use of the machine easier and less demanding. It is generally easier to program more advanced (i.e., complex) machines to do various sorts of tasks. This circumstance is reflected in the fact that a creature that makes its evolutionary way in the world by intelligence requires a rather powerful brain that enables it also torealize all sorts ofimpressive but evolution-irrelevant achievements. To be sure, evolution is not, in general, overgenerous. For example,

due to the costs involved, evolution will not develop creatures whose running speed is vastly greater than what is needed to escape their natural predators, to catch the prey that provides the staple of their diet, or to realize some other such strictly utilitarian objective. But intelligence and its works are to some extent an exception to this general rule, owing to its self-catalyzing nature. With cognitive capacities, the character of the issues prevents a holding back: once one can doalittle with calculation or with information processing, one can in principle do a lot. When biodesign takes the route of intelligence to secure an evolutionary advantage for a creature, it embarks ona slippery slope. Having started along — this road, there is no easy and early stop. Intelligence itself becomes a goad to further development simply because intelligence is, as it were,

EVOLUTION AS AN ALLOCATION MECHANISM

41

developmentally self-energizing. We desire knowledge for its own sake— for the benefit of its inherent satisfactions—and to some extent even need knowledge on this basis. But the causal origin of this desire unquestionably lies in pragmatic and evolutionary considerations. The result of these deliberations is straightforward. Intelligence is the evolutionary specialty of Homo sapiens. If we were markedly less smart than we in fact are, we would not have been able to survive. Or rather,

more accurately, we would not have been able to develop into the sort of creatures we have become. Intelligence constitutes the characteristic specialty that provides the competitive advantage that has enabled our species to make its evolutionary way into this world’s scheme of things. We are so smart because this is necessary for us to bé here at all.

Why Are We So Dumb? But at this point a very different question arises, one that points in the exactly opposite direction: Why are we not a great deal more intelligent than we in fact.are? Why can we not mastera foreign language within a single week or learn calculus in a fortnight? What explains our manifest cognitive deficiencies and limitations? Why are we so dumb? This question is also one that can in principle be answered in evolutionary terms.‘ But it has two importantly different aspects: (1) Why are we not comprehensively smarter by way of enhanced mind power for the species as a whole, and (2) Why are we not statistically smarter by way of an increase in the relative proportion ofsmart people within the presently constituted range of intelligence levels. Let us consider these issues one at a time.

First off, to be substantially smarter as a whole; we would, on prevailing bioengineering principles, need a much bigger brain. To manage this would require a larger but less agile body, forcing us to forgo the advantages of maneuverability and versatility. To process twice the information would require a brain of roughly four times its present size. But to quadruple our brain weight, we would need a body of sixteen times its present 4 It may be thought that there is something incongruous in asking for an evolutionary

explanation for something that has not happened. But the issue is rather one of using basic principles of natural process to explain why evolution does not take certain routes. In this regard, the situation with respect to intelligence (i.e., cognitive agility) is not dissimilar from that with regard to motion (i.e., physical agility). Explaining why evolution has not produced ahyperintelligentmammal is structurally akin to explaining why it has not produced a hyperswift one by outfitting creatures with organic wheels. For an interesting treat-

ment of this particular issue, see Jared Diamond, “The Biology of the Wheel,” Nature 302 :

(14 April 1983): 572~73.

42

A NORMATIVE THEORY OF RATIONALITY

weight.° Abody of so great a weight is not only extremely cumbersome but involves enormous demands for energy. The most plausible and probable move would then be to opt for a very different ecological niche and take to the water, joining our mammalian cousins the whales and dolphins. The intellectually stimulating surroundings of a land environment with its invitations to technological development and record keeping would all be denied us. That gain in brain power would have come at an awesome cost: the sacrifice of the collective intelligence of the social institutionalization of tool-using creatures. The price is one that an evolutionary development of intelligence cannot afford. There remains, however, the question of why we humans should not be smarter by way ofa statistical improvement in the relative proportion of very smart people within our existing species as actually constituted. With this shift of questions, we now move from the issue of bioengineering a more intelligent species to the development of a more intelligent population—one in which the percentage of people qualifying as superior in intelligence by the presently prevailing standards would be substantially enlarged. It is here that the social dimension of the matter comes into prominence. Consider the following sort of case. You and I interact in a competitive situation of potential benefit B that has a roughly zero-sum character, with one party’s gain as the other’s loss. Two alternatives are opento each of us: to collaborate with the other, or to try to outwit the other. If we collaborate, we shall share the resultant benefit (say, by each getting one-half of it). If we compete, then the winner takes all; whoever succeeds gains the whole benefit. The overall situation thus stands as depicted in figure 3.1. If I see my chances of winning to be measured by the probability p, then my expectations stand as follows: EV (collaborate) = p(0.5 B) + (1 — p)(0.5 B) = 0.5B, EV (compete) = p(B) + (1 — p)(0) = pB.

Figure 3.1. Hypothetical Payoffs in a Situation of Competition Fortune favors me

We collaborate Wefail to collaborate

-

_

Fortune favors you .

0.5 B/0.5B B/O

0.5 B/0.5 B o

O/B

Note: The table entry x/y represents the gains for the two parties x (for me) and y (for you), respectively.

~.

5 On this issue, cf. J.B.S. Haldane’s insightful and provocative essay “On Being the Right Size,” in his collection Possible Worlds and Other Papers (New York, 1928). -

EVOLUTION AS AN ALLOCATION MECHANISM

43

As long as p is less than one-half—that is, as long as my subjectively

appraised chances of winning are less than even—collaboration is the sensible course relative to the balance of expectations. But when p exceeds one-half, the balance moves in favor of noncooperation. If one views optimistically the benefits of self-reliance in situations of potential conflicts of interest, then decision-theoretic rationality inclines against cooperation; it favors going one’s own competitive way and taking one’s chances. Thus people who see themselves as comparatively more clever are—other things equal—less likely to collaborate. And a society in which this state prevails pays the price of a weakened impetus to cooperative effort. To picture the structure of the situation more graphically, consider the sort of situation poséd by the (purely hypothetical) statistical distributions presented in figure 3.2. If we suppose that people interact randomly, then in Case 1 well over half (54 percent) of pairwise interpersonal interactions are between cognitive peers, but in Case 2 well under one-half (46 percent) of the interactions are so. Given that it takes two to cooperate, the upshot for Case 1 is that the majoritarian social norm would here provide a rational impetus to cooperation, but in Case 2 it would militate toward noncooperation by trying to outwit. This illustrates the general point. If we humans were substantially more intelligent than we are, the greater success of our interactions with nature would doubtless incline us toward a yet higher estimation of our intellectual powers, but in consequence, the impetus to a collectively beneficial collaboration with others would become undermined. Astatistical enhancement of cognitive competence would tend to increase the frequency with which, trusting to their intelligence, people would try to outwit others rather than “playing by the rules’—which, after all, are largely designed to protect the numerous nonadvantaged. The natural outrage we feel, even as children, against noncooperators and our rational animus toward those who do not play fair is patently connected in the Figure 3.2. Hypothetical Distributions of Cognitive Ability

1. The actual distribution of cogni-

Highly able

Moderately able

Substandardly able

10%

70

20%

30%

60%

10%

tive ability (hypothetical) 2. A situation ofcontemplated improvement

Note: The ability levels at issue are to be taken as fixed by the presently prevailing standards.

44

A NORMATIVE THEORY OF RATIONALITY

evolutionary order with the fact that most of us draw substantial benefits from a system in which people play by the rules. The statistical inferiority of the group provides an impetus toward practices under whose aegis parties of different levels of ability can contend on a much more even footing. The social cooperation conducive to human well-being overall benefits from a suitable admixture of incapacity. We humans require intelligence to structure our interactions with nature into generally beneficial channels. But no less important, we are collectively so situated that we need to cooperate and collaborate with one another in ways that conduce to the general benefit. And if we were, on statistical balance, more intelligent than we are, such cooperation and collaboration would be more difficult to achieve. People would become more reliant on their own wits and retreat from reliance on others, deeming themselves as “above the common herd.” As the bee illustrates, the evolution of cooperation certainly does not

require individual intelligence. Quite to the contrary. As the numbers of clever people who pride themselves on strength of intellect increases, social cohesion becomes more difficult to obtain. University faculties are notoriously difficult to manage. Experts are thorns in the sides of popes and presidents alike: no sect manages to keep on easy terms with its theologians. (Anyone who is familiar with the ways of an intellectual avantgarde such as the Bloomsbury circle has some idea of the difficulties of socializing people who see themselves as more than ordinarily clever.) It is easy to envision how intelligence militates against cooperation in many circumstances. To some extent, then, dumbness too is evolutionarily advantageous.

For an individual’s prospects of surviving and thriving are generally bound up with how well or ill things are going in the society of which this individual is part (as wars, epidemics, and economic depressions indicate). And insofar as a society's well-being can be undermined by the pressure of too many intelligent individuals, evolution will (obliquely) select against individual intelligence. The interesting and perhaps surprising lesson thus emerges that if we humans were on balance to be substantially more intelligent than we actually are, then the rational impetus to socialization and interpersonal cooperation would be undermined.If people were bodily stronger than they are, they would have to be larger and heavier—and would thus be hampered physically by the resultant cumbersomeness. If people were mentally stronger than they are, they would be hampered socially by a resultant impetus toward trying to outwit one another. The prospect. of effective socialization in the service of communal interests and the general good would be diminished. And any such result would clearly be evolutionarily counterproductive. One very good reason why we are not

EVOLUTION AS AN ALLOCATION MECHANISM

45

a lot smarter is just that it would not be a very smart move for us to be

so. These deliberations yield the odd-sounding lesson that evolutionary pressure is a two-edged sword that can cut in opposed directions as regards the development of intelligence. Evolution is a process in which

the balance of cost and benefit is constantly maintained in a delicate equilibrium. And this general phenomenon is vividly illustrated in the particular case of our cognitive capacities. In its handling of intelligence, evolution, like a shrewd gambler, is clever enough to follow the precept “Quit while you're ahead.”

Conclusion: An Idealistic Prospect The evolutionary origination of our cognitive resources assures that our

performances in this regard are bound to be limited and fall far short of what we might like them to be. But this does not, of course, countervail against the fact that our aspirations are unlimited. On the side of wish and desire—of what we would ideally like to have—“the sky is the limit.” And the basic conceptions with which we operate in the cognitive domain—the ideas of knowledge, of truth, of science, and the like—are all geared to our ambitions rather than our capacities. All of these conceptions are of such a nature as to involve an element of idealization, a glancing of the mind’s eye toward a finality and a perfection that is not, in fact, actually realizable in this imperfect world. In this regard, then, our thought leaps beyond our reality-constrained capacities. The machinery of cognition that is at our disposal we owe to evolution. . But the conceptual mechanisms that we make use of in cognitive contexts—concepts that, like knowledge, truth, certainty, simplicity, com-

prehensiveness, system, and science, orient our thought about the issues of this domain—are in large measure idealizations, intellectual constructs in whose construction idealization plays a crucial role. The character of such conceptual mechanisms is not dictated by the realities. We project them against a background of a hypothetical ideal order in which the limitations of the actual are in hypothesis transcended. In matters of knowledge and science, as in matters of law or morality, our dealings with reality are always conditioned by our having in view certain ideas about what ought to be.. The restraints operative here are not set by faithfulness to reality alone but also by considerations of utility. The intelligence we possess is the finite resource of a limited creature. But like a sixteenth-century European monarch, it projects vast territorial claims throughout a wide world from its tiny homeland. Its intellec-

_ tual horizons are not bounded by the immediate realities of an imperfect

46

A NORMATIVE THEORY OF RATIONALITY

world but look beyond them to a perfected and idealized order. In characterizing our imperfect information in a descriptive framework that revolves about such demanding notions as truth, knowledge, and systematized science, we use these idealizations to orient ourselves in a thought order of our own making. Even as meridians of longitude are thought creatures that enable us to move about effectively in the surface of our planet, so these cognitive idealizations enable us to move easily and with good results through the realm of thought. Here once again our dealings

with the real are crucially conditioned by the use of thought instrumentalities that are themselves constituted with a view to the ideal.® § The issues of this chapter are elaborated upon in the author's Useful Inheritance (Savage, Md., 1990).

Four Rational Inquiry as the Pursuit of an Unattainable Ideal Against Convergentism, Approximationism, and the View That Later Science Is Lesser Science One of the leading ideas in the pragmatism of Charles Sanders Peirce is that ultimate science is correct science. He took the step of equating “the real truth” with “our ultimate truth’—conceiving of the truth as the result that science will approximate asymptotically in the long run. If §, is the state of science at time ¢, then we would, on this approach, project the idea of “ultimate science” or “science in the limit” S. = lim S, tre

and then see the real truth in scientific matters as given through the equation: ‘P’ is true = Pe §.,. Such an approach has the immense theoretical advantage of making it possible to see the real truth not as a relation between our domesticated human knowledge and something that lies wholly and altogether outside its range. Truth determination is now not a matter of relating the manifold. of our knowledge. to an altogether knowledge-external. reality. but merely one of having the successive stages of the cognitive state of the art themselves converge toward a limit. The truth becomes determinable as such within the domain of a developing manifold of knowledge. Despite its attractions, this Peircean approach faces grave, indeed insuperable difficulties. The first is the problem of the existence of such a

limit. For we confront the formidable difficulty: how can one plausibly secure a metric to measure the distance between bodies of knowledge?! Once this idea of distance is abandoned—as it seemingly must be—the idea of a measure can no longer be applied; and the prospect of a math-

ematical limit collapses.: But even if we had such a measure, the question of the existence of a

limit remains. Realistically, there seems precious little reason to think 1 This point was forcibly pressed by W. V. Quine in Word and Object (New York, 1960), p. 23. Quine argued that talk of the limit of theories is based on an inappropriate mathematical analogy. Various theorists have tried to resolve this difficulty, but their efforts generally suffer from the need to resort to restrictive assumptions and special cases.

48

A NORMATIVE THEORY OF RATIONALITY

that the actual course of scientific progress is such that the successive stages ofknowledge will come to be less and less distant from one another. As various writers across a spectrum from Thomas Kuhn to Paul Feyerabend have argued, there is no reason to think that the successive theories of science involve a historical consilience in the theoretical accounts they give. There is no indication in practice and no reason of general principle to think that scientific progress is a matter of diminishing differences in the meaning content of the successful theories at issue, moving toward an ultimate-limit theory. To be sure, the successive stages of science come to be more and more successful—in some sense of this idea they do an increasingly more adequate job. But this increase in efficacy is attained across a never-diminishing divide of content diversity. Our commitment to the view that later theories are better theories does not underwrite the idea of an ever-closer approach to a position of fixity. Improvements in evidential grounding or application efficacy do not somehow constitute improvements in depiction. Scientific progress thus correlates with warranted confidence. But this augmentation in our level of confidence moves—at the level of theory content—through a series of radical changes and discontinuities as regards matters of meaning. (The situation is not one of the Peircean picture of stable essentials with improvements only at the successive decimal points.) With the transition from a corpus of scientific knowledge to a successor, we neither get more of the truth nor draw nearer to it;

rather, we get a better-based estimate of the truth—one that gives us a firmer warrant for our claims. Accordingly, while we can make methodoriented claims about “improving correctness,” we should make no such claims at the contentual, substance-oriented level of “increasingly close approximation to a ‘correct picture of things.’ ” Our talk of progress cannot bypass the epistemic and regulative aspect and move directly ‘to a descriptive and constitutive result. To reemphasize: One must resist any

temptation to think of improvement in warrantin terms of improvement in approximation. Moreover, even if a limit existed, the question of its determinability still remains. We could nevertell that we have actually found it. For we live and breathe and have our own being in the short run, while limits

have to do with what happens in the eventual long run. And as the saying has it, “In the long run we are all dead.” Anything we can get hold of in the historical course of things is without firm implications, one way or the other, with respect to what will happen ultimately “in the limit.” If categorical truth is a matter of limits, then we cannot get there from here. Moreover, any talk of approximation is parasitic upon talk of limits, so once we abandon limits, we must also give up any temptation to discuss scientific progress in the language of approximation.

RATIONAL INQUIRY

49

In this regard, the oft-drawn analogy between empirical inquiry and geographic exploration is surely mistaken. It views scientific progress as a whole on the basis of one particular (and by no means typical) sort of cognitive advance, namely, the sequential filling-in of a roughly predetermined picture in greater and greater detail—the continual working

out of more decimal places to lend additional refinement to a generally known result. Scientific research is conceived on an analogy with terres-

‘trial exploration, whose ongoing pursuit yields findings of continually smaller significance, ‘since what is at issue is the removal of ever-more-

minute gaps in our information. We have here an accretional view of the

progress of science, subject to the idea that each successive accretion inevitably makes a comparatively smaller contribution to what has already come to hand. Progress, on this view, consists in driving questions down to a lesser and lesser magnitude, continually decreasing their inherent

significance to a smaller and smaller scale, descending to even more finegrained level of detail. (This at bottom is the Peircean version of ultimate convergence in scientific inquiry.) But such a picture combines two gravely mistaken ideas: (1) that the progress of science proceeds by way ofcumulative accretion (like the growth of a coral reef), and (2) that the

magnitude of these additions is steadily decreasing. If the former of these ideas collapses, then so will the position as a whole. And collapse it does. For science progresses not just additively but also largely subtractively. As.Thomas Kuhn and others have persuasively argued, today’s most significant discoveries generally represent a revolutionary overthrow of yesterday’s, the big findings of science taking a form that contradicts its earlier big findings and involves not just supplementation but even outright replacement on the basis of: conceptual and theoretical innovation. It will not do to take the preservationist stance that the old views were acceptable as far as they went and merely need supplementation and refinement. Significant scientific progress is genuinely revolutionaryin that there is a fundamental change of mind as to how things happen in the world. Relativity theory does not amend the doctrine of luminiferous ether but abandons it. The creative scientist is as much a demolition expert as a master builder. Significant scientific progress is generally a matter not of adding further facts (on the order of filling in of a crossword puzzle) but of changing the framework itself; substantial headway is made preeminently by conceptual and theoretical innovation. Moreover, the second thesis of significance diminution is also untenable. Our scientific picture of the natural world is built up not by way of cumulation but by way of substitution and replacement.? To all appear2 This shibboleth of the contemporary philosophy of science is not all that new. Already at the turn of the century, Sir Michael Foster wrote: “The path [of progress in science] may

50

A NORMATIVE THEORY OF RATIONALITY

ances, the progress of science is a matter of discontinuous jumps. Significant scientific progress is generally a matter not of adding further details to a preestablished picture but of changing the framework itself. And there is no reason to think that this process need ever come to a stop.? Given that science in the main develops not by way of cumulation but by way of replacement, it makes no sense to think of the development of science in terms of convergence or approximation. Not only are we not

in a position to claim to achieve the truth in science, we cannot even say

that we are drawing closer to it. For how could we possibly do so without knowing where it lies? We have no guarantee that we are not exchanging one somehowincorrectopinion for another that is at bottom no less so. Accordingly, we can claim neither that in inquiry we attain the truth, nor that we steadily draw nearer to this goal. Some theorists regard this prospect of science moving through an endless sequence of fundamental revisions as totally unacceptable. Since science is a pursuit ofthe truth, its progress must—they insist—increasingly converge upon some fixed and final result: initially we get the first decimal place worked out (the big picture), then we move on to the next (the issues of smaller detail).* But this doctrine of gradual convergence has its difficulties. There is no reason to think that nature will be cooperative in always yielding its most important secrets early on and reserving nothing

but the relatively insignificant for later on. No preestablished harmony— no metaphysical hidden hand—is at work to assure that the order of dis-

covery in our penetration of the secrets of nature replicates their order of importance in some ontological sense, so that later penetration of the not be alwaysa straight line; there may be swerving to this side and to that; ideas may seem to return again and again to the same point of the intellectual compass; but it will always be found that they have reached a higher level—they have moved, not in a circle, but in a

spiral. Moreover, science is not fashioned as is a house, by putting brick to brick, that which is once put remaining as it was put to the end. The growth of science is that of a living being. As in the embryo, phase follows phase, and each member or body puts on in succession different appearances, though all the while the same member, so a scientific.con- ception of one age seems to differ from that of a following age” (“The Growth of Science in the Nineteenth Century,” in Annual Report of the Smithsonian Institution for 1899 [Washington, D.C., 1901], p. 175 [as reprinted from Foster's 1899 presidential address to the

British Association for the Advancement of Science)). 3 E. P. Wigner puts the matter as follows: “In order to understand a growing body of phenomena, it will be necessary to introduce deeper and deeper concepts into physics and ... this development will not end by the discovery of the final and perfect concepts. I believe that this is true: we have no right to expect that our intellect can formulate perfect concepts for the full understanding of inanimate nature’s phenomena’ (“The Limits of Science,” Proceedings of the American Philosophical Society 94 [1950]: 224). 1 E.g., William Kneale, “Scientific Revolutions Forever?” British Journal for the Philosophy of Science 19 (1967): 27-42.

RATIONAL INQUIRY

51

“deeper” layers is always of lesser cognitive significance for the constituting of an adequate view of the world’s processes. In factual inquiry into the ways of the world, we can do no better than to pose questions and canvass the currently visible alternativessBut the

questions we can pose are limited by our conceptual horizons. And the “answers we can envisionare also limited by the cognitive state-of-the-art. (The Greeks could not have asked about continental drift; the Romans

could not have thought of explaining the tides through gravitation.) And of course the whole process of canvassing those answers can come to grief

because the very question being asked is based on untenable suppositions. “Which of the four elements (air, earth, fire, water) is the paramount arché, the fundamental type of stuff from which the whole of

physical reality originates?” asked the early Milesians. They contemplated just those four alternatives together with the fifth possibility of a neutral, intermediate stuff. It did not occur to them that their whole inquiry was abortive because it was based on a misguided conception of elements. Nor did it appear to be a realistic prospect to all those latenineteenth-century physicists who investigated the properties of the luminiferous ether that no such medium for the transmission of light and electromagnetism might exist at all. Investigators of an earlier era not

only did not know what the half-life of californium was, but they would not: have understood it even if this fact had been explained to them. Ongoing scientific progress is not simply a matter of increasing accuracy through improving by a few more decimal places the numbers at issue in our otherwise stable descriptions of nature. And this fact blocks the theory of convergence. Any theory of convergence in science, however carefully crafted, shatters against the conceptual innovation that continually brings entirely new, radically different scientific concepts to the fore, carrying in its wake an ongoing wholesale revision of “established fact.” ‘In any convergentprocess, later is lesser. But since scientific progress on matters of fundamental importance is generally a matter of replacement rather than mere supplementation, there is no reason to see the later issues of science as lesser issues in the significance of their bearing upon science as a cognitive enterprise—to think that nature will be cooperative in always yielding its most important secrets early on and reserving nothing but the relatively insignificant for later on. (Nor does it seem plausible to think of nature as perverse,leading us ever more deeply into deception as inquiry proceeds.) A very small-scale effect— even one that lies very far out along the extremes of a “range exploration” in terms of temperature, pressure, velocity, or the like—can force a far-

reaching revolution and have a profound impact by way of major theoretical revisions. (Think of special relativity in relation to ether-drift

52

A NORMATIVE THEORY OF RATIONALITY

experimentation, or general relativity in relation to the perihelion of Mercury. )® On this basis, one arrives at a view of scientific progress as an ongoing process in which every major successive stage in the evolution of science yields innovations of undiminished overall interests and importance. We are constrained to reject convergentism as a position that lacks the support not only of considerations of general principle but also of the actual realities of our experience in the history of science.

Truth Estimation Throughout our researches into nature we seek to determine the truth about things at the level of scientific generality and precision—this, beyond question, is our goal and our desire. But though we aim at this

target, we do so with the full realization that we generally miss the mark. And how—given that identification is not a realistic prospect here—is one to characterize the relationship between the actual fruits of our cognitive efforts and the real truth at which they are directed? What we manage to achieve in scientific inquiry is not to approximate the truth but to estimate it—to form, as best we can, a reasoned judgment of where the truth of the matter lies. With respect to our truth, we have to take the stand that it represents no more (but also no less) than

our best estimate of the truth. With the progress of science these estimates become less deficient: increasingly, the new estimates overcome various of the shortcomings of the old, both theoretical and practical.

Older theories confront the older problems. There is much to J.B.S. Haldane’s dictum that in science “a fact is a theory in which no one has made a large hole for a long time.” It is, accordingly, rational to prefer the new estimate to the old one. But the basis of this rationality does not lie in

“getting nearer to the truth” in a contentual sense akin to that in which a photograph made with a better camera is more accurate. At the level of . scientific theorizing, our present world-picture represents a better estimate than our past pictures only in the sense that it is, comparatively 5 The present critique of convergentism is thus very different from that of W. V. O. Quine. He argues that the idea of convergence to a limit is defined for numbers but not for theories, so that speaking of scientific change as issuing in a convergence to a limit is a misleading metaphor. “There is a faulty use of mathematical analogy in speaking of limit of theories, since the notion of a limit depends on that of a ‘nearer than,’ which is defined for

numbers and not for theories” (Word and Object, p. 23). The present deliberations are prepared to apply the metaphor of substantial and insignificant differences to theories but are concerned to deny that, as a matter of fact, the course of scientific theory-innovation must eventually descend to the level of trivialities.

RATIONAL INQUIRY

53

speaking, better based than they are in accommodating a wider range of data. But throughout this context, one must carefully bear in mind the

distinction between a better estimate (one that has fewer deficits, is based on securer or fuller information, or the like) and a closer estimate

(one that is “nearer to the real truth”). With scientific theorizing we must settle for a qualitative “better” because there just is no (science-independent) way of monitoring the issue of a measurable “closer.” This is something we must accept realistically and have no right to lament. The most we can ever realistically do in matters of theoretical science is to characterize what we do actually have as the very best that one can possibly obtain in the circumstances—as the best estimate of the truth (that estimate for which the best case can be made out according to the established standards of rational cogency). As concerns the substantive content of its theories, science moves

through a series of radical changes and discontinuities that invalidate any talk of successive approximation. The resultant view of scientific progress is emphatically not that of the Peircean picture of stable essentials with improvements even further and further beyond the decimal point. For while we can make content-external claims about an improved warrant for our truth-estimates, we can make no such claims at the content-internal level of “increasingly close approximation to a ‘true picture of things.’ ” In the circumstances of the case, there-is no way of moving from “improved warranted” to “more closely approximating to the truth.” We must resist any temptation to think of improvement in estimation in terms of increasingly closer approximation in this context of empirical inquiry. A better evidentiated view of things is not thereby one that approximates more closely to a preestablished fixity. The need for such an estimational approach is easy to see. Pilate’s question is still relevant. How are we imperfect mortals dwelling in this

imperfect sublunary sphere to determine where “the real truth” lies? The circumstance that we have no direct access to the truth regarding the world is perhaps the most fundamental fact of epistemology. We have no lines of communication with the Recording Angel. We live in a world not of our making where we have to do the best we can with the means at our disposal. And as regards our purported knowledge of this world, we have no prospect of assessing the truth—or presumptive truth—of our beliefs independently of the use of our imperfect epistemic mechanisms of inquiry and systematization. We have, and indeed can have, no direct access to “the real truth,” and it could be grossly foolish to hanker after it.6 Simple realism forbids us to regret that we can’t get outside our 5 Est ridiculum quaerere quae habere non possumus, as Cicero wisely observed (Pro Archia, 10)

54

A NORMATIVE THEORY OF RATIONALITY

thought and experience to compare it with reality. It involves the fallacious mis-analogy of picture taking to suppose that only in somehow direct or intuitive apprehension could we ever really know what things are like. And so we cannot bypass epistemic qualification in regard to the facts we ascertain by inquiry—we cannot get to “is true” except through the mediation of “is reasonably held to be true.” This latter must be accounted good enough—for the simple and sufficient reason that it is the best we can possibly do. In the circumstances of the situation, estimation affords the best means for doing the job. In the inescapable circumstances of the case, we have no alternative but to settle for the best avail-

able estimate of the truth of the matter. The lesson here is not that there is no difference between “what we think we know” and “the real truth,” but just that we ourselves are never entitled to do more than to view the former as our best estimate of the latter. There is nothing absurd about our holding ‘something to be the very best that we can possibly do to determine the truth of the matter while yet acknowledging the possibility, the probability—-nay, even the virtual certainty—of eventually coming to recognize its untenability and indeed falsity. Our accepted scientific truths can—and should—be viewed as representing nothing more than the best estimates of the real truth that we are able to make in the present state of the art.

Issues of Descriptive Particularity One important qualification is in order. In rejecting the approximationist view of scientific progress, the present position holds in effect that science moves through an endless sequence of fundamental revisions. But here it becomes crucial to distinguish between the level of theoretical generality and the level of descriptive particularity, between generalized theses and concrete particular quantitative facts, between issues oftheoretical explanation andissues of numerical description. Theoretical issues are in general open-ended—there is no predeterminable limit to the range of potential variation among the hypotheses at issue here. (Think, for example, of the shift from Newtonian to Einstein-

ian gravity, from gravity as a kind of force that produces action at a distance to gravity as an aspect of the spatiotemporal geometry structure of the universe.) With the descriptive characterization of concrete particulars, however, we generally confront a prespecifiable range of limited potential variation. Within the scientific framework we may inquire, for example, into

such matters as the age of humankind (or of the earth, or of the universe),

RATIONAL INQUIRY

55

the atomic weight of lead, the boiling point of water, the number of stars in the galaxy, or the relative positions of spectral lines of a certain substance. In each instance of such “constants of nature,” we are dealing with particularizations within a certain parametrized range. (A species must have an evolutionary origin at some time; an elemental substance

must have some atomic weight or other; and so on.) And if such a quality is well defined at all, ongoing inquiry should be able to determine it in an ultimately convergent manner. The concrete, particular, objective furnishings of “the real world” must by their very nature as such have certain delineable property-determinations. And when these are of a quantitative nature, then the Peircean idea of a “convergence of inquiry” takes on a more tractable aspect. Here, indeed, Peirce’s convergentist view is

fundamentally correct. Consider, for example, a particular descriptive natural constant that can be subject to meaningful inquiry: the acceleration due to gravity, for example, or the length of Saturn’s year, or the melting point of lead, or the population of India in a.p. 1,280. Then, as Peirce insisted, if this magnitude at issue is to be a well-defined quantity at all, then ongoing inquiry about it must discernibly converge to some definite value. No matter how wildly our initial efforts at evaluation may fluctuate, still as the volume of relevant evidence grows larger and larger, a definite result must emerge fairly soon, lest the quantitative constant at issue become dissolved as such. Here convergence is indeed coordinated with reality. We stand committed to the existence of a particular parameter-value lying within a certain determinate range—that the earth is an object that does have a certain age, which cannot be less than the roughly 1 x 108 | years that organic life has existed on the planet or greater than the roughly 2 X 10% years of the existence of the solar system of which it forms a part. And we stand committed (within the framework of our conceptual scheme) to the idea that fuller and more sophisticated inquiry will eventually pin this value down—or, rather, would in principle do so if inquiry were pushed far enough. All these commitments inhere straightforwardly in our conceptual scheme regarding the character of inquiry about quantitative facts. Our standard conception of objectivity and the nature of real things certainly does indeed assume or postulate that ongoing inquiry will converge to a definite result in just exactly the Peircean manner in such contexts of quantitative description. All the same, describing particulars is something very different from theorizing about generalities. The descriptive features of concrete particulars (of which we take the view that we can firmly fix certain of their descriptive features) and the explanatory law-structures we project at the theoretical level to account for the mode of operation of nature at large (where we neither can nor do claim such prior knowledge) stand on a

56

A NORMATIVE THEORY OF RATIONALITY

very different footing. At the level of quantifiable descriptive particularity, we can place firm limits about the range of conceivable variation. As long as we maintain our commitment to there being certain particular sorts of things at issue, we consider that the range of quantitative variation is in some ways predetermined, taking the view that the course of inquiry is thereby circumscribed as ever narrowing (as regards the future).” But at the level of scientific generality—of theoretical laws rather than particular parameter-values—we can indeed contemplate an endless sequence offundamental revisions in the course ofinquiry, an ever-changing scientific view of the process structure of the world’s lawful comportment.

Is a Skeptical Abandonment of Truth Possible? Some recent philosophers have been tempted by skeptical argumentation of the following sort: Neither can we claim to have attained the truth in scientific matters, nor can we even say that we are approaching it more and more closely. It follows then, that there simply is no such thing as “the real truth” in these matters. This whole absolutistic notion should simply be abandoned, and with it the idea that science involves “the pursuit of the truth.”

Accepting this view, some of the most prominent philosophers of science of the day have simply given up on truth. Rudolf Carnap teaches that science should never make flat assertions but only statements of probability. Again, Karl Popper has argued long and eloquently that we must abstain from staking claims to truth in the sciences, that scientists should never believe the theses they devise but should view them as merely instrumental conjectures, that they must try (and even hope) to falsify and must never regard as claims of substantive fact. Why then not bite the bullet and follow the-skeptical path in dropping all reference to “the pursuit of truth” in considering the aim of science? The answer is straightforward. Recourse to “the truth” in our delibér- | ations is justified pragmatically in. terms of the sort of useful work this notion is able to accomplish in the conceptual scheme of things. The characterizing aim of science, after all, is to provide appropriate answers to our questions about what goes on in the world. This obviously calls for providing information by endorsing various theses about how matters 7 To be sure, the items at issue can be involved with theoretical shifts, as, e.g., with the

speed of light or the quantity of mass in relativitytheory.

,

8 See Rudolf Carnap, “The Aim of Inductive Logic,” in Logic, Methodology, and the Philosophy of Science, ed. E. Nagel, P. Suppes, and A. Torshin (Stanford, Calif., 1962), pp. 308-18.

RATIONAL INQUIRY

57

stand. Accordingly, if one is to hold that human inquiry yields information about the world, one is constrained to deem oneself entitled to accept certain factual theses—with acceptance, of course, to be understood

as acceptance-as-true. The mission of science is to obtain answers to our questions about the world and to secure information about it. To secure information about the world, we must accept contentions about it. And given the standard Tarski equivalence, Piff ‘P is true,

we recognize that to accept ‘P’ is to accept it as true. So we certainly can-

not simply follow the skeptic-here and drop the reference to the issue of

a pursuit of truth. Scientific theories are doubtless no more than estimates, but they still and ineradicably are estimates of the true answers to the questions of the field. Without a reference to the truth, we would lose our hold on the teleology of aims and goals that define the very nature of the enterprise of scientific inquiry. Any view of scientific or of common-life claims as information-providing proceeds on an acceptance model of rational inquiry into the truth of things. We may, to be sure, not actually succeed in finding “the real truth in these matters.” But unless we are prepared to take a committal stance toward what we do find—unless we are prepared at any rate to claim truth for our findings and so to accept them (at least provisionally) as asserting what is actually the case, that is, as stating the putative truth— we must simply abandon an endeavor to take information-oriented cognitive stance toward the world. (It would not make sense to think of scientific inquiry as a project in truth-estimation if there were no truth to be estimated.) / , If science were not an attempt to get at the actual truth of things (an attempt that is, admittedly, imperfect and, as best we can tell, generally ends in failure), then the aim of providing information about the world and of answering our questions about how things actually stand would become ‘altogether untenable. Our conception of the very nature of the whole project of scientific inquiry would have to be abandoned—reference to science as a process of inquiry would go by the board. It becomes difficult to envision what we could be left with that bears any relationship to science as we know it. What we think of as scientific knowledge would become a matter of tribal ideology, a sociological issue of opinion research, of assessing the shared beliefs of the practitioners of scientific disciplines—an inquiry that would disintegrate into a matter of collecting opinions about opinions. Any doctrine able to avert such consequences automatically has much to be said in its favor. The pursuit of truth is the name of the game, the definite object of the

58

A NORMATIVE THEORY OF RATIONALITY

whole project of inquiry. Here as elsewhere, we must preserve a clear division between intent and achievement. It is manifestly the intent of

science to discern the real capital-T Truth about things. The “real Truth,” or authentic truth about reality, represents a conception to which we

stand committed throughout the whole project of rational inquiry because it constitutes the aspiration of the enterprise aim—its ultimate target or telos—though not, to be sure, its actual achievement. Admittedly, while we aim at “the real truth” in scientific inquiry, this is something that lies on the side of intention and purport. In point of actual achievement we realize (indeed presume) that our existing science falls short, that it gives us estimates of the truth rather than the truth itself. But this only means that we must treat its claims as representing provisional truths, not that we should abandon any and all claims to a relationship to truth. To do so would, as we have seen, be to throw out the baby. with

the bathwater: to abandon the definitive mission of science as a rational enterprise that is designed to answer our questions about the world. In abandoning our claims to achieve at least as much as plausible estimates of the real truth, we would be reduced to talking only of what we

(I, you, many of us) think to be so. The contrast between what we claim and what actually is so—the real truth of things—would no longer be. available. We would now only be in a position to contrast our putative truths with those of others, but could no longer operate with the classical

distinction between the putative and the actual, between what we think to be so and what actually is so. And at this point, the idea of inquiry— aimed as it is, at increasing our grasp of the truth of things—would also go by the board. It would be senseless to study an object whose very existence is in question and to endeavor to estimate something that is not there—or at any rate presupposed or assumed or postulated to be there. In abandoning truth—in refraining from assuming or postulating that there indeed is such a theory as “the real truth of the matter” in regard to how things work in the world sub ratione generalitatis—we would no longer be able to conceptualize the project of scientific inquiry as we standardly do. (The argumentation is, in effect, a transcendentalargument from the very possibility of science.) To be sure, we have learned in the school of bitter experience that, even as we must presume that“the real truthof the matter” exists, so

there is no alternative to acknowledging that our science as it stands here and now (or perhaps ever) does not achieve the real Truth. It is one thing to speak of truth realization in the language of aspiration and quite another to speak of it in the language of achievement. But the crucial contrast between the real truths of the perfected cognitive condition of things and the merely purported or ostensible truths of the cognitive state of the art as it stands would come apart at the seams if we aban-

RATIONAL INQUIRY

5S

doned our (regulative) commitment to the view that there is indeed such

a thing as the real truth. It is to the great credit of C. S. Peirce that he clearly saw—two millennia of skepticism from Pyrrho to Descartes notwithstanding—that the fallibility of our scientific truth claims does not mean that the conception of “the pursuit of truth” should be abandoned _

in this domain at the level of aim and aspiration.® The upshot of these deliberations is that in scientific contexts we must

think of the truth in terms of estimation rather than approximation. Deployed at the level of matters of scientific generality and precision, the conceptions of truth and knowledge are idealizations. They look to an order where all the things that we try to do by way of information acquisition—but only realize incompletely and imperfectly—have actually been done fully and adequately. In staking those claims to truth and knowledge, we implicitly claim to have done everything that is needful to ascertain the relevant facts. And this is never in fact the case. All we can ever accomplish is to do all that we can manage to do in this direction—and to suppose with some degree of hopefulness that this is enough.

The Idealistic Impetus of Inquiry We have to reckon in matters of scientific theorizing with Hegel's cardinal principle that only the totality contains truth,that definitive truth claims require complete evidence and total information, while in actual-

ity our claims to truth at the level of generality and precision at issue in theoretical science are always based on incomplete evidence and partial information. In consequence, the very idea of scientific knowledge is itselfan idealization. For scientific inquiry is the pursuit of an unattainable ideal—the ideal of a perfected science that affords us a true and adequate account of how things work in the world. We realize full well that this ideal is unachievable. We have no assurance that the cognitive venture will succeed in significant measure; we can only presume this in the light of our commitment to an ideal, a commitment whose justification comes

only ex post facto through its demonstrated utility. And we accordingly initially presume those conditions under which alone the pursuit. of this ideal makes rational good.sense and then, ultimately, manage with the wisdom of eventual hindsight to achieve cogent grounds for seeing our commitment to it as rationally justified. These aspects of the matter deserve closer examination. ® These deliberations are elaborated upon in chap. 10 of the author’s Empirical Inquiry (Totowa, N.J., 1982).

Part THE STATUS OF SCIENCE

Five An End to Science? -Is Scientific Discovery an Inherently Bounded Venture? Research in natural science makes substantial demands on the public purse nowadays—as is vividly illustrated in the recent American decision to build a multibillion-dollar supercollider particle-accelerator. Are such vast investments worthwhile in terms of potential intellectual results? Or is the wellspring of new discovery about to run dry? Some years ago, Bentley Glass made newspaper headlines with a presidential ‘address to the American Association for the Advancement of Science that posed the

question “Are there finite limits to scientific understanding, or are these endless horizons?” His answer ran as follows: ‘What remains to be learned may indeed dwarf imagination. Nevertheless, the universe itself is closed and finite. . . . The uniformity of nature and the general applicability of natural laws set limits to knowledge. If there are just 100, or 105, or 110 ways in which atoms may form, then when one has identified

the full range of properties of these, singly and in combination, chemical knowledge will be complete. There is a finite number of species of plants and animals—even of insects—upon the earth. . . . Moreover, the universality of the genetic code, the common character of proteins in different species, the ’ generality of cellular structure and cellular reproduction, the basic similarity of energy metabolism in all species and of photosynthesis in green plants and bacteria, and the universal evolution of living forms through mutation and natural selection all lead inescapably to a conclusion that, although diversity may be great, the laws of life, based on similarities, are finite in number and comprehensible to us in the main even now. We are like the explorers of a great continent who have penetrated to its margins in most points of the compass and have mapped the major mountain chains and rivers. There are still innumerable details to fill in, but the endless horizons no longer exist.!

This perspective sees the scientific project as an inherently bounded venture, subject to the idea that in the scientific investigation of nature, as in the geographic exploration of the planet, we are ultimately bound to reach the end of the road. The reasoning that underlies this analogy of terrestrial exploration is essentially that of the argument sketched in figure 5.1, which moves from 1 Bentley Glass, “Science: Endless Horizons or Golden Age?” Science 171 (1971): 24.

64

THE STATUS OF SCIENCE

the finitude of nature itself to that of natural science as we know it.? But alas, things are not quite so simple. There are, in fact, sound theoretical reasons for viewing this idea of an inevitable ending to scientific inquiry as thoroughly problematic. Let us consider more closely why the oft-contemplated prospect of an approaching end to science is in principle unrealistic and implausible.

Nature Might Exhibit an Unending Complexity of Physical Constitution To underwrite the prospect of endless progress in the development of natural science, some theoreticians deem it necessary to stipulate infinitude in the intrinsic makeup of nature as a physical structure. The physicist David Bohm, for example, writes, “At least as a working hypothesis science assumes the infinity of nature; and this assumption fits the facts much better than any other point of view that we know.” In line with Figure 5.1. The “End-to-Science” Argument The cardinal aim of science is to produce a complete inventory ofthe laws of nature.

“

The laws of nature are generalizations that specify how the constituents of nature interact.

Nature has only a finite number of constituents and finitely many constituents can interact in only finitely many ways. Therefore: There are only finitely many laws of nature.

Therefore: Science is a finite project that has a natural end. 2 This argumentation and the geographic-exploration analogy on which it is based are a

widely employed standby. “Sciénce cannot keep on going so that we are always going to discover more and more new laws... . . It is like the discovery of America—you only discover it once. The age in which we live is the age in which we are discovering the fundamental laws of nature, and that day will never come again” (Richard Feynman, The Character of Physical Law [Cambridge, Mass., 1965], p. 172; see also Gunther Stent, The Coming of the Golden Age [Garden City, N.Y., 1969], and S. W. Hawking, “Is the End in Sight for Theoretical Physics?” Physics Bulletin 32 [1981]: I5-17).

4 Marxist theoreticians take this view very literally—in the manner of Lenin’s idea of the inexhaustibility of matter in Materialism and Empirico-Criticism. Purporting to inherit from Spinoza a thesis of the infinity of nature, they construe this to mean, among other

things, that any cosmology that denies the infinite spatial extension of the universe must be wrong. . 4 “Remarks by David Bohm,” in Observation and Interpretation, ed. Stephan Koerner (New York, 1957), p. 56. For a fuller development of Bohm’s views on the “qualitative infinity of nature,” see his Causality and Chance in Modern Physics (London, 1957).

AN END TO SCIENCE?

65

this approach, Bohm postulates a theory of unending complexity in nature.

Such an infinity-of-nature postulate can take various forms. One of them contemplates a system of wheels within wheels, following the lines of Pascal's idea that nature might be an endless succession of Chinese boxes, with microscopic worlds emplaced within ever-more-minute microcosms.® Sometimes this shrinkage toward the submicroscopically small is inverted into an expanding sequence moving unendingly upward toward the supratelescopically large—emplacing macroscopic worlds within ever-more-macroscopic ones. (Perhaps what we see as a galaxy is simply. a subatomic particle of a macroworld, and our whole universe no more than an atom.) Following this line of thought, one might project an ascending sequence: subatomic particles, atoms, molecules, organic microorganisms, animals, populations, life systems, solar systems, galaxies, galactic clusters, worlds, world systems, and so on, each furnished with

its own characteristic modes of operation. Every system is potentially a member of a yet larger system. But on either view—with ongoing increase or decrease alike—the prospect of a theoretically unending progress of science is assured by the existence of an unending series of new worlds embracing (or embraced by) the old.® Science is potentially lim-

itless because there are ever-new domains of nature still to be explored. Other theorists take a less radical course and exchange the appeal to unending levels of physical inclusiveness for an ever-deepening succession of causal principles or “forces” within the physical makeup of nature. As one physicist puts it: Why should nature run on just a finite number of different types of force? May there not be an infinite hierarchy of types of force just as there is an infinity of structures that may be built up of matter interacting under the influence of one or more of those forces? . . . It is perfectly possible that there exist objects that interact powerfully with each other but only exceedingly feebly with the ob5 A nineteenth-century writer put it this way: “Go on as far as we will, in the subdivision

of continuous quantity, yet we never get down to the absolute point. Thus scientific method leads us to the inevitable conception of an infinite series of successive orders of infinitely small quantities. If so, there is nothing impossible in the existence of a myriad universe within the compass of a needle’s point, each with its stellar systems, and its suns and planets, in number and variety unlimited” (W. S. Jevons, The Principles of Science, 2d ed. [London, 1877], p. 767).

6 The idea of a succession of “layers of depth” in the analysis of nature, each giving rise to its own characteristic body of laws in such a way that each of these law manifolds is more encompassing than its predecessors and that their successive discovery represents a sequentially deeper penetration of the structure of nature, was, so far as I know, first mooted in contemporary physics by E. P. Wigner, “The Limits of Science,” Proceedings of the American Philosophical Society 94 (1950): 422-27. But it had already been contemplated by Charles Sanders Peirce, half a century earlier.

66

.

THE STATUS OF SCIENCE

jects with which we are familiar, that is to say with objects that interact strongly with ourselves, so that these unknown objects could build up complex structures that could share our natural world but of which we should be ignorant.”

Theorists of this general persuasion envision an ongoing succession of modes of natural process, each operating in its own parametric sphere within an unendingly varied cosmic manifold. Any such unfolding complexity in the physical constitution of nature would, of course, assure that, in principle, the well of potential scientific discovery need never run dry. It clearly suffices for this end. But it is by no means necessary. To see this, let us now turn our attention from the complexity of nature’s physical makeup to the nomic complexity of its lawful comportment.

Nature Might Exhibit an Unending Complexity of Lawful Comportment Some theorists—C. S. Peirce, for example—see the opening up of new realms of phenomena in evolutionary terms. Early on in world history, before the evolution of complex molecules, there was no place for biological laws; in the era of Neanderthal man there was no room for political

economy. As the cosmos grows older, new modes of natural organization gradually evolve to afford new phenomena that are governed by emerging laws that previously had no opportunity to come into operation. There is progression from laws of individual physical particles to laws of increasingly elaborate organized complexes thereof. Since the universe affords a varied panorama of modalities of physical process evolving over time, a science that reflects this will continue to find new grist for its mill. Still other theorists consider the element of chance and randomness at work in the world, which they see providing a potentially unending hierarchical order ofprocesses governed by principles that represent statistical summaries, as it were, of successively emergent tendencies. In this vein, the French physicist Jean Paul Vigier contemplates a succession of lawful orders conceived along Hegelian dialectical lines, reasoning as follows: At all levels of Nature you have a mixture of causal and statistical laws (which come from deeper or external processes). As you progress from one level to

another you get new qualitative laws. Causal laws at one level can result from averages of statistical behavior at a deeper level, which in turn can be explained by deeper causal behavior, and so on ad infinitum. If you then admit 7 Sir Denys Wilkinson, The Quarks and Captain Ahab, or; The Universe as Artifact, Schiff Memorial Lecture [Stanford, Calif., 1977], pp. 4-5.

AN END TO SCIENCE?

67

“| , ]

that Nature is infinitely complex and that in consequence no final stage of 4 |_- knowledge can be reached, you see that at any stage of scientific knowledge ‘~T" causal and probability laws are necessary to describe the behavior of any phe;

\)

nomenon, and that any phenomenon is a combination of causal and random

properties inextricably woven with one another.®

Such an approach envisions an unending hierarchy not of physical structures but of modes of lawful process characterizable through statistical

patterns in natural phenomena—and of the causal laws that result at successive levels. Two very different sorts of complexity hierarchies can in principle prevail. Oneis a hierarchy of material systemsestablished by physical inclusion: particles, atoms, molecules, macrolevel physical objects, stars and

planets, galaxies, galactic clashes, and so forth. (Or again: molecules, cells; Organs, organisms, coloni

.) However, such physical hierar-

chies of compositional structure stand in contrast to nomic hierarchies of physical operation rising from base-level laws that overn phenomena to higher-level laws coordinating laws, and upward to laws coordinating laws that coordinate laws, and so on. Even a worldthat is-fintteti the— ~ structural complexity of its physical constitution may nevertheless exhibit a hierarchy of lawful orders with an ongoing sequence of levels of higherorder principles of coordination. For in such a law hierarchy, any particular law is potentially a member of a wider family of laws that will itself exhibit some lawful characteristics and thus be subject to synthesis under still higher laws. We thus move from laws governing phenomena (firstorder laws) to laws governing such laws (second-order laws) and so on, ascending to new levels of sophistication and comparative complexity as we move along. Given such an unending exfoliation of levels of lawful operation, new higher-level metadisciplines can in theory always spring up to relate old disciplines. It is clear that such an infinite proliferation of principles of lawful comportment would also serve to block any prospect of completing science. For nature to have an unlimited cognitive depth, there is thus no need to suppose thatthe. physical complexity of nature is unlimited: After all,

the prime task of science lies in discovering the laws of nature, and the ongoing complexification of nature in point of lawful operation suffices for our present purposes of providing for potentially endless new discovery and deepened understanding in science.® A-supposition of the structural infinitude of nature is accordingly not 8 Jean Paul Vigier, “The Concept of Probability in the Frame of the Probabilistic and the Causal Interpretation of Quantum Mechanics,” in Observation and Interpretation, ed. Stephan Koerner (New York, 1957), p. 77. Cf. also E. P. Wigner, “Limits of Science.” § For an illuminating and suggestive analysis of “the architecture of complexity,” see Her- bert A. Simon, The Sciences of the Artificial (Cambridge, Mass., 1969).

68

.

THE STATUS OF SCIENCE

required to assure nonterminating progress in science. An unending

de lexity of nature’s laws would be quite enough to -tinderwrite the potential limitlessness of scientific inndVation. But oné can go still further.

The Phenomena of Nature Might Be Unendingly Diverse Even a system that is finitely complex both in its physical makeup and in its functional laws might nevertheless be infinitely complex in the phenomena that it manifests over time. For the operations of a structurally

_andlawfully finitesystem canyet exhibit an infiniteintricacy improdac> tive complexity,

manifesting this limitless diversity in the working out of ~

Fis arocesses tetherthan as regards Tt spattovtrtcteral-compositionoFthe— “nomic comporkment of is basiccomponents Evan ifthe number of con _stituents of nature were small,

the waysinwhich In they canbecombined~

to yield phenomena in space-time might yet be infinite. Think here of— the examples of letters, syllables, words, sentences, paragraphs, books, genres (novels, reference books, etc.), libraries, library systems. Even a finite nature can, like a.typewriter with a limited and operationally simple keyboard, yield an endlessly varied text. It can produce a steady stream of new products—‘“new” not necessarily in kind but in their functional interrelationships and thus in their theoretical implications. And on this basis our knowledge of nature’s workings can be endlessly enhanced and deepened by contemplating an unending proliferation of phenomena. There is no need to assume aceiling to such a sequence of levels of integrative complexity ofphenomenal diversity. The different levels each exhibit an order of their own. The phenomena we attain at the n-th level can have features whose investigation takes us to the (n + 1) level. New phenomena and new laws presumably arise at every level of integrative order. The diverse facets of nature can generate conceptually new strata .

of operation that yield a potentially unending sequence of levels, each giving rise to its own characteristicprinciples of organization, themselves quite unpredictable from the standpoint ofthe other levels. In this way, even a relatively simple world as-regards its basic operations can come to have an effectively infinite cognitive depth, once the notion of a natural phenomenon is broadened to include not just the processes themselves and the products they produce but also the relationships amongthem. Consider, for example, some repeatedly exemplified physical feature and contemplate the sequence of 0s and 1s projected according to the rule that the i-th entry in the sequence is 1 if this feature is exemplified

AN END TO SCIENCE?

69

on occasion number i, and 0 if not. Whenever two such feature-concepts, say C and C’, generate such sequences in the manner of C: 0100110100... C’: 1001011010...,

then we can introduce the corresponding matching sequence for C and

-C’, 0010010001 . . . , which is such that its i-th position is 1 if the two base sequences agree at their respective i-th positions, and 0 if they disagree: Such matching sequences will havea life of their own. For example, even if two base sequences are random, their matching sequences need not be—for example when those base sequences simply exchange Os and 1s. (Even random phenomena can be related: by laws of coordination.) Note, moreover, that one can always regard matching sequences themselves as further base sequences, so as to yield “second-order” phenomena, as it were. One can then proceed to examine the relationships between them—or between them and other base sequences. This process yields a potential hierarchy of laws of coordination—at level i + 1 we

have the laws of coordination between sequences at level i. Such a perspective illustrates how simple base phenomena can combine to bring more and more grist to the mill of study and analysis. Increasingly sophisticated mechanisms of conceptual coordination can lead us to regard the same phenomena in the light of different complexity-levels. And there is no reason why this sort of thing cannot go on and on. For the system always exhibits new patterns of phenomenal order over time, and

so there is always more to be learned about it. There will always be new levels of functional complexity of operation to be investigated with such a system. It will always be possible to discern yet further levels of structured relationship amongthe phenomena at issue. (And of course these emergent, higher-level phenomena are in no way reducible to the processes and procedures of the lower levels.) Even when we continue to occupy ourselves with the same species of object (e.g., symbols), and the same basic laws (e.g., combination rules, or grammars), we can nevertheless encounter increasingly complex operational modes that constitute new sorts of data. Consider an analogy. The chess master and the beginner make exactly the same sorts of moves—their realms are governed by exactly the same laws. An observer exclaims, “I know all about chess, for I have now discovered the rules

according to which those pieces move.” Splendid? But also naive, since this business of “the basic rules of the game” is of course merely Step One. It is in point of the complexity of the higher-level principles of tactics and strategy that masters and beginners differ decisively. To be sure, that chessboard is limited and finite, and so are the pieces

70

THE STATUS OF SCIENCE

sitting upon it. Moreover, so are the basic “laws of nature” of this do-

main—the moves that the different pieces may make. Nevertheless, the moves of a game of chess are covered by very different categories of rules—not just the rules of the game that define those basic legal chess moves, but rules of tactics and deeper-level principles of strategy. The

game can be played—and thus studied—at very different levels of depth and sophistication. As the analogy of chess makes clear, it is radically mistaken to think that we have got to the bottom of things when we merely grasp nature’s basic physical laws. The question of what sort of game nature is playing by those rules remains open. Even though nature might be of finite physical and nomic complexity as regards its physical structure and its basic

procedural laws, nevertheless it could be infinitely diverse in the unfolding operational complexity of its phenomenal products over time. To understand the world about us, we need departments of biology and economics as well as departments of physics. Regardless of the known character of a sequence of phenomena that we confront, we can never rule out the possibility that yet further pat-

terns of relationship exist. For there will be patterns of phenomena, and patterns of such patterns, and patterns of such patterns—and on and on. We can study letter sequences and move thence to paragraphs and on to chapters, to books, to book categories, to book systems (French vs. Chi-

nese literature), and so on. Every new level of consideration will afford phenomena of its own that will themselves admit of further study and analysis. Confronted with repetitive phenomena of any description, inquiry can always in principle find new grist for its mill. In theory it may occur that just the same relationship patterns simply recur from level to level—that the patterns of phenomena that we encounter at level i + 1 simply reduplicate those already met with at level i. (When physical configurations are at issue, then this situation of pat-

tern repetition at different levels of scale will yield the “fractal” structures made prominent by Elie Mandelbrot.) However, this is a very spe-

cial case that certainly does not obtain across the board. We have no good reason to think that our worldis fractalin the overall phenomenal structure of its processes. Butis natural science not bound by a principle of simplicity—is it not

committed to the idea that nature proceeds in fundamentally simple ways? By no means! We have no ground whatever for supposing the simplicity of nature. The so-called principle of simplicity is really the following principle of complexity management: Feel free to introduce complexity in your efforts to describe and explain nature’s ways. But “keep it simple!” Only introduce as much complexity as you

AN END TO SCIENCE?

71

actually (and minimally) need for your scientific purposes of description, explanation, prediction, and control.

Such an approach is eminently sensible. But of course sucha principle is no more than a methodological rule of procedure for managing our cognitive affairs. Nothing entitles us to transmute this methodological precept into a descriptive-ontological claim to the effect that nature is simple ‘in the sense of cognitive exhaustibility. The lesson of such deliberations is clear. There simply is no realistic prospect of establishing that scientific innovation will ever come to a stop. Suppose that we were to discover that: 1. the world is of limited structural complexity: that it is finitely stratified in the subordination and superordination of physical systems so that the world’s physical constitution is inherently limited; 2. the worldis of limited nomic complexity: that there are only finitely many (“basic”) laws of nature; and 3. theworld is of limited descriptive complexity: that at any given time there are only finitely many realizable states of nature.

Even if all this were to be so, the theoretical prospect of making new scientific discoveries nevertheless could not be precluded. Only if we lived in a Nietzschean world of eternal recurrence—a world whose fixed patterns of occurrence reiterate themselves in a vast periodicity of unchanging repetition—could the prospect of ongoing phenomenal diversity and thus of ongoing scientific innovation be decisively abolished. And there is, of course, no basis for viewing this prospect as even remotely plausible. On the contrary, for aught we can ever say with any confident assurance on the basis of our investigations, the world may exhibit all

three of these modes of infinitude.

The Basis for an Unending Prospect of Scientific Discovery Might Lie Wholly in the Character of Our Inquiry Processes But there is yet another, more fundamental, aspect to the matter. In its

most familiar form, recourse to an infinity-of-nature principle reflects a strictly one-sided approach that places the burden of responsibility for the incompletability of science solely on the shoulders of nature itself. The potential endlessness of scientific progress is taken to require limitlessness on the side of its objects, so that the infinitude of nature must

be postulated at either the structural or the nomic and/or operational level. But this approach is quite mistaken. What really matters for ongoing scientific innovation is limitlessness in point of cognitive rather

72

.

THE STATUS OF SCIENCE

than physical complexity. For scientific innovation—the cognitive exploration of the ways of the world—pivots on the interaction of the mind with nature, on the mind's exploitation of the data to which it gains access for the sake of penetrating the “secrets of nature.” And so the crucial fact is that scientific progress hinges not just on the makeup of nature itself but also on the character of the information-acquiring processes by which we investigators proceed in investigating it. Responsibility for the cognitive inexhaustibility of nature need not lie on the side of nature at all but can in principle rest one-sidedly with us, its explorers. Suppose a natural system to be such that a certain parameter value v, for essentially technical reasons, cannot be evaluated by us at the precise

point ¢ of the time continuum but only on average during an interval around ft. In such a case, the system can be very simple indeed—it need contain no complexities apart from those required to assure the preceding assumptions—and yet the prospect of endless cognitive progress is nevertheless available. For as our capacity to make v-determinations down to smaller and smaller time intervals increases from minutes to seconds to milliseconds to microseconds and so on, we can obtain an increasingly comprehensive insight into the modus operandi of the system and can obtain ever-fuller information about it that could not have been predetermined on the basis of earlier knowledge. Since averages at levels of larger scale cannot determine those at smaller ones, quite different modes of comportment—and thus laws—could manifest themselves at different levels. Again, suppose a (totally random) sequence of 2s and 3s, on the order of 23223 . . . , and suppose further a transformation that substitutes the pair 10 for 2 and 11 for 3 so as to yield: 1011101011...

We are now ina position to ascertain such “laws” as: 1. The sequence of even-numbered positions (01001 . . .) will be a random mix of Os and Is (which simply mirrors the initial random sequence of 2s and 3s).

2. All the odd-numbered positions are uniformly filled by 1s.

,

We here confront a peculiar mixture of randomness with regularity. But of course it is only by studying components of that 0-1 sequence (e.g., its odd-numbered entries) that we can discern its code—examining the statistics that ignore patterns of periodicitywillnot bring regularities of the sort at issue to our notice. And it becomes crucial in this context that higher-level patterns are not necessarily predictable from lower-level ones. Knowing the frequency with which individual-letters such as a and T occur in a text will not tell us much about the frequency with which a

combination such as aT occurs. When we change the purview of our

AN END TO SCIENCE?

73

conceptual horizons, there is always in principle more to be learned— novelty that could not have been predicted from earlier, lower-level information. Only by bringing appropriate coordinative concepts to bear can we discern the laws at issue. And no matter how far we push forward along such lines, the prospect of detecting further patterns of lawful regularity can never be eliminated—or downplayed by claiming that such . newfound laws are inherently less significant than the rest. The introduction of new conceptual perspectives is the key. Each time a new discipline opens up its novel characteristic perspective, a vantage point is gained for the reappraisal of old issues in other fields. (The body of Shakespeare’s work is finite, but this does not mean that Shakespeare scholarship will come to the end of the road; think of how psychoanalysis, for example, or sociology—disciplines that did not exist in the bard’s own day—have been used as launching platforms for innovations in Shakespearean studies.) The contemplation of such examples indicates that the complexity inherent in the ongoing exfoliation of further hierarchy levels need not inhere in the structural makeup of nature itself but may derive from the conceptual mechanisms being deployed to study it at ever-greater depths of sophistication. The case is similar to that of the geometer, whose hierarchy of definitions, axiomatic facts, lemmas, theorems, and so on does not reside in the geometer’s materials as such, but in the conceptual taxonomy that he or she chooses, for cognitive reasons, to impose on these materials. By itself, nature presumably lacks “depth” altogether. For depth (like difficulty) is an inherently relative matter, generated through the operation of a cognitive perspective. The endless levels at issue will not.be physical levels but levels of consideration that emerge from the modus operandi of inquiring beings. Complexity, after all, lies less in the objects than in the eyes of their beholder. (As Herschel ruminated long ago, particles moving in mutual gravitational interaction are, as we- human investigators see it, forever solving differential equations that, if written out in full, might circle the earth.)!° If our vision itself is imperfect, then even when the scene we examine is itself only finitely complex, an ever-ampler view of it will emerge as the resolving power of our conceptual and observational instruments is increased. When we make measurements at a given level of accuracy, the world may appear X-wise; and when that accuracy is improved by an order of magnitude, it may appear Y-wise. At each successive state-ofthe-art stage of increased precision in our investigative proceedings, the world may take on a very different nomic appearance, not because it 10 John Herschel, Familiar Lectures on Scientific Subjects (London, 1867), p. 458.

74

:

THE STATUS OF SCIENCE

changes, but simply because at each stage it presents itself differently to us. Accordingly, the reason for the cognitive inexhaustibility of nature need not rest unilaterally with the world alone. The character of our information-gathering procedures, as channeled through our conceptualtheoretical perspectives, is also bound to play no less crucial a part. Innovation on the side of data can generate new theories, and new theories can transform the very meaning of the old data. This dialectical process of successive feedback has no inherent limits and suffices of itself to underwrite a prospect of ongoing innovation in the study of nature. Accordingly, we cannot write finis to the book of scientific inquiry as long as the prospect of a change of mind on our part regarding the world’s workings cannot be precluded. Such considerations combine to indicate that an assumption of the structural infinity of the physical extent of the natural universe or of the functional infinity of its nomic or operational complexity is simply not required to provide for the prospect of ongoing scientific progress. Continuing discovery is quite as much a matter of how we inquirers proceed with our work as it is of the nature of the object of inquiry. The salient point is that it is cognitive rather than structural or operational complexity that is the key here, and that cognitive depth need not depend on these sorts of physical depth. There is no sound reason to think that the natural science of even a finite world is an inherently closed and terminable venture, and no adequate basis for insisting that the search for greater depth in our understanding must eventually terminate at a logical rock bottom.

The Regulative Rationale for Supposing the Cognitive Inexhaustibility of Nature Admittedly, it is conceivable (all that has been said here notwithstanding) that the world has a finite cognitive depth and that scientific inquiry into — its workings will some day come to a definitive end of the road as concerns the discovery of significant novelties. But it should also be stressed that it is difficult to see how one could ever establish this termination with significant assurance. How could we ever preclude convincingly the

prospect that, had we pressed just a bit harder and pushed our inquiries U This idea that our knowledge about the world reflects an interactiveprocess, to which

both the object of knowledge (the world) and the knowing subject (the inquiring mind) make essential and ultimately inseparable contributions, is elaborated in several of the au-

thor’s books, in particular in Conceptual Idealism (Oxford, 1973) and Methodological Pragmatism (Oxford, 1977).

AN END TO SCIENCE?

15

just a bit further, our efforts would have been repaid by new revolution-

ary discoveries—that new modes of order might have emerged within a setting that, heretofore, just looked like so much insignificant noise? Considerations regarding the mathematical tests for randomness serve to illustrate this point. In randomness screening there are various criteria. To fail any one of them is a decisive impediment to randomness. But surviving any given group of them is indecisive: there are always yet further tests that could still be posed and failed. The unsuccessful search for order up to a given level of sophistication does not establish actual randomness; it is no more than suggestive, seeing that the possibility of order at a deeper, yet-uncontemplated level of sophistication cannot be precluded. There is no effective procedure for determining randomness, and so the search for order can never be terminated with unalloyed confidence that further effort im the search for order is fruitiess. Accordingly, there can be no adequate justification for writing off the search for order in nature categorically once and for all. _ In the face of this prospect, we have no sensible alternative to taking it as a regulative principle—a methodological working hypothesis—that nature’s cognitive depth is not finite and that a sufficient further effort will, in principle, yield with further discoveries at any given stage. To be sure, our discussion of the incompletability of science has proceeded at

the level of plausible possibilities. Its argumentation moves as follows: The incompleteness of science—the availability of lawful regularities in nature extending beyond those we have been able to discern—can never be precluded as a theoretical prospect. There are good methodological (and thus practical) reasons for regarding our science as incomplete at any given stage.

,

Therefore: We are well advised to regard science as incompletable.

Such an argument is profoundly idealistic because it addresses itself not to the character of nature’s facts as such but to the suppositions that, in the circumstances, it is sensible to make about them. Its basis is fundamentally methodological rather than empirical. Only by being prepared to suppose that science-as-we-have-it is incomplete can we clear the way for an ongoing endeavor to move the frontiers forward. Of course, the enterprise advancing our scientific probes _ of nature is not simple, the work not easy or inexpensive. The call of convenience and the demands of other desiderata ever conspire to urge ® On the economic requirements of scientific innovation, see the author's Scientific Prog-

ress: A Philosophical Essay on the Economics of Research in Natural Science (Oxford, 1978).

76

:

THE STATUS OF SCIENCE

a slackening of effort in training scientists and in funding their endeavors. And the comforting thought that nothing substantial remains to be done provides a perfectly respectable-looking excuse here. The forces of antiintellectualism in society and in government are ever on the lookout for reasons to keep science on a short financial leash. Perhaps there indeed are such reasons. But the idea that science stands on the verge of com-

pletion—that all the great discoveries have been made and that only a few mere details remain to be filled in—is certainly not one of them. The analogy of terrestrial exploration is untenable. For in undertaking the cognitive exploration of nature, we are embarked on a quest that is, in principle, literally endless. Here, then, we have a consideration from which realism can draw

comfort because of its clear impact that there is more to reality than can be compassed by instrumentalities of mind. But this support for realism is of a somewhat peculiar sort, seeing that it rests on a deeply idealistic basis. For its rationale is not an investigation of reality—a duly secured view of the real as such. Instead, it emerges from an examination of how

we do actually think—and in the circumstances must think—about the

reality that our inquiries address. In pivoting the issue on the concept of the real as it figures in our thought, rather than on the nature of reality as it eventually emerges from inquiry, our deliberations obtain a markedly idealistic cast, notwithstanding their realistic tenor. 13 The theme of the imperfectability of our inquiries into the ways of nature is the topic - of several of my books: Scientific Progress (Oxford, 1978), Empirical Inquiry (Totowa, N-J., 1982), The Limits of Science (Berkeley and Los Angeles, 1984), and Scientific Realism (Dordrecht, 1987). The ideal aspect of inquiry is also stressed in Forbidden Knowledge (Dordrecht, 1987). The utility of unrealizable ideals is argued in Ethical Idealism (Berkeley and Los Angeles, 1987).

Six The Imperfectibility of Science ‘Conditions of Perfected Science How far can the scientific enterprise advance toward achieving a definitive understanding of nature? Can science attain a point of recognizable completion? Is the realization of perfected science a genuine possibility, even in theory, putting aside all merely practical obstacles as somehow incidental? These are challenging questions—and far from simple ones. What would perfected science be like? What sort of specifications would it have to meet? It would, obviously, have to realize in full the discharge of natural science’s characterizing mandate or mission. But just how is this to be conceived? Clearly, the goal structure of scientific inquiry covers a good deal of ground; it is diversified and complex, spreading across both the cognitive/theoretical and active/practical sectors. For it encompasses the traditional quartet of description, explanation, prediction, and control, in line with the picture of figure 6.1.

The theoretical sector concerns itself with matters of characterizing, explaining, accounting for, and rendering intelligible—with purely intellectual and informative issues, in short. By contrast, the practical sector

is concerned with guiding actions, canalizing expectations, and, in general, achieving the control over our environment that is required for the satisfactory conduct of our affairs. The former sector thus deals with what science enables us to say, and the latter with what it enables us to do; the one relates to our role as spectators in nature, the other to our role as active participants. Figure 6.1. The Goal Structure of Natural Science

Theoretical

goals

description (answering what and how questions about nature)

Explanation (answering why questions about nature)

Cognitive

goals

prediction (successful alignment of our ex-

Practical goals

pectations regarding nature) control effective intervention in nature to alter the course of events in desired directions

Manipulative goals

78

.

THE STATUS OF SCIENCE

It thus appears that if we are to claim that our science has attained a perfected condition, it would have to satisfy (at least) the following conditions: 1. Erotetic completeness: It must answer (in principle, at any rate) all the descriptive and explanatory questions that it itself countenances as legitimately

raisable and must accordingly explain everything it deems explicable. 2. Predictive completeness: It must provide the cognitive basis for accurately predicting those eventuations that are in principle predictable (that is, those that it itself recognizes as such). 3. Pragmatic completeness: It must provide the requisite cognitive means

for doing whatever is feasible for beings like ourselves to do in the circum-stances in which we labor. 4. Temporal finality (the omega condition): It must leave no room for the claim that further changes in the existing state of scientific knowledge can be achieved.

Each of these modes of completeness deserves detailed consideration. First, however, one preliminary stricture. It is clear that for a condition

of science to qualify as “perfected,” it would have to meet certain formal requirements of systemic unity. If, for example, there are different routes to one and the same question (e.g., if both astronomy and geology can inform us about the age of the earth), then these answers will certainly have to be mutually consistent. Perfected science will have to meet certain formal requirements of structural systematicity in the manner of its articulation: it must be coherent, consistent, consonant, uniform, harmonious, and so on. Interesting and important though they are, we shall, for the moment, leave these formal requirements aside, our present concern being with the more substantive issues.

The Infeasibility of Erotetic Completeness Could science ever actually achieve erotetic completeness—the condi- ~ tion of being able to resolve, in principle, all of our (legitimately posable) questions about the world? Could we ever actually find ourselves in this desirable condition?? In theory, yes. The available body of scientific information could conceivably be such as to answer all the questions that it allows to arise. But just how meaningful would this mode of completeness be? ! Chapter 9 below deals with these formal considerations. 2 Note that this is independent of the question “Would we ever want to do so?” Do we ever want to answer all those predictive questions about ourselves and our environment, or are we more comfortable in the condition in which “ignorance is bliss”?

THE IMPERFECTIBILITY OF SCIENCE

79

It is sobering to realize that the erotetic completeness of an existing state of science does not necessarily betoken its comprehensiveness or sufficiency, but it might instead reflect the paucity of the range of questions we are prepared to contemplate, incorporating a deficiency of imagination, so to speak. When the range of our knowledge is sufficiently restricted, then its erotetic completeness will reflect merely this impoverishment rather than its intrinsic adequacy. Conceivably, if improbably, science might reach a purely fortuitous equilibrium between problems and solutions. It could (possibly) eventually be “completed” in providing an answer to every question that one can appropriately ask in the then-existing (albeit still-imperfect) state of knowledge, without thereby being completed in the larger sense of answering the questions that would arise if only one could probe nature just a bit more deeply. And so, our corpus of scientific knowledge could be erotetically-complete and yet remain fundamentally inadequate. Even if realized, the erotetic mode of completeness need not be particularly meaningful. (To be sure, this discussion proceeds at the level of supposition contrary to fact. The exfoliation of new scientific questions from the answers given to old—the familiar process of question propagation—spells the infeasibility of ever attaining erotetic completeness.) The preceding considerations illustrate a more general circumstance. Any claim to the realization of a theoretically complete natural science would be one that affords “a complete, consistent, and unified theory of

physical interaction that would describe all possible observations.”* But to ascertain that the prevailing state of physics actually meets this condition, we would need to know exactly what physical interactions are indeed possible. And to warrant us in using the state of physics on hand as a basis for answering this question, we would already have to be assured that its view of the possibilities is correct and thus already have preestablished its completeness. The idea of a consolidated erotetic completeness shipwrecks on the infeasibility of finding any meaningful way to monitor its attainment.

After all, any judgment that we can make regarding the laws of nature—any model we can contrive as to how things work in the world—is a matter of theoretical triangulation from the data at our disposal. And we can never have unalloyed confidence in the definitiveness of our data base or in the adequacy of our exploitation of it. For one thing, there is always the problem of future observations. For another, different lawsystems can always, in theory, yield the same observational output: as philosophers of science are want to insist, observations underdetermine 3S. W. Hawking, “Is the End in Sight for Theoretical Physics?” Physics Bulletin 32

(1981): 15.

80

;

THE STATUS OF SCIENCE

laws. No matter how comprehensive our data or how great our confidence in the inductions we base upon them, the possible inadequacy of our claims cannot be dismissed. One can never obtain secure confidence in writing finis to inquiry on the basis of purely theoretical considerations.

We can plausibly estimate the amount of gold or oil yet to be discov-

ered, because we know the earth’s extent and can thus establish a proportion between what we have explored and what we have not. But we cannot comparably estimate the amount of knowledge yet to be discovered, because we have and can have no way of relating what we know to

what we do not. The very idea of cognitive limits has a paradoxical air. It suggests that we claim knowledge about something outside knowledge. But (to hark back to Hegel) with respect to the realm of knowledge, we are not in a position to draw a line between what lies inside and what lies outside—seeing that, ex hypothesi, we have no cognitiveaccess to that latter. One cannot make a survey of the relative extent of knowledge or ignorance about nature except by basing it on some picture of nature that

is already in hand—that is, unless one is prepared to take the deliverances of existing science as final and definitive, which we can recognize straight off as an inappropriate supposition. This process of judging the adequacy of our science on its own telling is the best we can do, but it remains an essentially question-begging and consequently inconclusive way of proceeding. The long and short of it is that there is no cognitively satisfactory basis for maintaining the completeness of science in a rationally cogent way. To monitor satisfactorily the theoretical completeness of science, we accordingly need some theory-external control on the adequacy of our theorizing, some theory-detached reality principle to serve as a standard of adequacy. We are thus driven to abandoning the road of pure theory and proceeding along that of the other goals of the enterprise. This circumstance lends special importance and urgency to the predictive pragmatic sector.

The Infeasibility of Predictive Completeness There are deep difficulties in trying to use prediction as a standard of completeness for natural science. For one thing, we can make predictions only about matters that lie, at least broadly speaking, within our present cognitive horizons. Newton could not have predicted findings in’quan-

tum theory any more than he could have predicted the outcome of recent American presidential elections. One can make predictions only about what one is cognizant of, takes note of, deems worthy of consideration.

THE IMPERFECTIBILITY OF SCIENCE

81

In this regard, one can be myopic either by not noting or by losing sight of significant sectors of natural phenomena. Moreover, science itself sets the limits to predictability—insisting that

some phenomena (the stochastic processes encountered in quantum .physics, for example) are inherently unpredictable. And this is always to some degree problematic. The most that science can reasonably be asked to do is to predict what it itself sees as in principle predictable—to answer every predictive question that it itself countenances as proper. And here we must once more recognize that any given state of science might have gotten matters wrong in various respects.

With regard. to prediction, we can unproblematically apply the idea of improvement, of progress. But it makes no sense to. contemplate the achievement of perfection. For its realization is something we could never establish by any practicable means.

The Infeasibility of Pragmatic Completeness The arbitrament of praxis—not of theoretical merit but of applied and practical capability—affords the best standard of adequacy for our scien-

tific proceedings that is available. But could we ever be in a position to claim on the basis of the success of its practical applications that science has been completed? On this approach, the perfection of science would have to manifest itself in the perfecting of control—in achieving a perfected technology. But just how are we to proceed here? Could our natural science achieve manifest perfection on the side of control over nature? Could it ever underwrite a recognizably perfected technology? The issue of control over nature involves much more complexity than may appear on first view. For just how is this conception to be understood? Clearly, it is in terms of bending the course of events to our will,

of attaining our ends within nature. But this involvement of “our ends” brings into prominence the factor of our own contribution. For example, if we are inordinately modest in our demands (or very unimaginative), we may even achieve “complete control over nature” in the sense of being in a position to do whatever we want to do, but yet attain this happy condition in a way that betokens very little real capability. One might, to be sure, invoke the idea of omnipotence and construe a

“perfected” technology as one that would enable us to do literally anything. But this approach would at once run into the difficulties of a kind already familiar to the medieval scholastics. They were faced with this challenge: “If God is omnipotent, then can he annihilate himself (contra his nature as a necessary being), or can he do evil deeds (contra his nature as a perfect being), or can he make triangles have four angles (con-

82

THE STATUS OF SCIENCE

trary to their definitive nature)?” Sensibly enough, the scholastics inclined to solve these difficulties by maintaining that an omnipotent God need not be ina position to do literally anything, but rather simply anything that is possible for him to do. Similarly, we cannot explicate the idea of technological omnipotence in terms of a capacity to produce any result, wholly without qualification. We cannot ask for the production of a perpetuum mobile, for spaceships with “hyperdrive” enabling them to attain transluminar velocities, for devices that predict essentially stochas-

tic processes such as the disintegrations of transuranic atoms, or for piston devices that enable us to set independently the values for the pressure, temperature, and volume of a body of gas. We cannot, in sum, ask

of a “perfected” technology that it should enable us to do anything that we might take it into our heads to do, no matter how unrealistic this might be. All that we can reasonably ask of it is that perfected technology

should enable us to do anything that it is actually possible for us to do— and not just what we might think we can do but what we really and truly can do. A perfected technology would be one that enabled us to do anything that can possibly be done by creatures of our kind. But how can we deal with the pivotal conception of “can” that is at issue here? Clearly, only science—real, true, correct, perfected science-——could tell us what in-

deed is realistically possible and-what indeed is inescapable and unalterable. Whenever our knowledge falls short of this, we may well ask the impossible by way of accomplishment (e.g., spaceships in “hyperdrive”) and thus complain of incapacity to achieve control in ways that put unfair

burdens on this conception. Power is a matter of the effecting of things possible—of achieving control—and it is clearly cognitive state of the art in science that, in teaching us about the limits of the possible, is itself the agent that must shape our conception of this issue. Every law of nature serves to set the boundary between what is genuinely possible and what is not, between what can

be done and what cannot, between which questions we can properly ask and which we cannot. We cannot unproblematically monitor the ade-quacy and completeness of our science by its ability to effect “all things. possible,” because science alone can inform us about what is possible. As

science grows, develops, and changes, it poses new issues of power and control, reformulating and reshaping those demands whose realization

represents “control over nature.” Forscience itself brings variant possibilities to view. (At a suitable stage, the idea of splitting the atom will no longer seem a contradiction in terms.) To see whether agiven’state of technology meets the condition of perfection, we must already have a body of perfected science in hand to tell us what is indeed possible. To validate the claim that our technology is perfected, we need to preestab-

THE IMPERFECTIBILITY OF SCIENCE

83

lish the completenss of our science. The idea works in such a way that valid claims to perfected control can themselves rest only on perfected science.

In attempting to travel the practicalist route to cognitive completeness, we are thus trapped in a circle. Short of having supposedly perfected science in hand, we could not say what a perfected technology would be like, and in consequence we could not possibly monitor the perfection of science in terms of the adequacy of the technology that it underwrites. Moreover, even if (per impossible) a pragmatic equilibrium between

what we can do and what we wish to do in science were to be realized, we could not be warrantedly confident that this condition will remain unchanged. The possibility that just around the corner things will become unstuck can never be eliminated. Even if we achieve control to all intents and purposes, we cannot be sure of not losing our grip upon it— not because of a loss of power but because of cognitive changes that engender a broadening of the imagination and a widened apprehension as to what “having control” involves. Accordingly, the project of achieving practical mastery can never be monitored in a rationally satisfactory way. The point is that control hinges on what we want, and what we want is unavoidably conditioned by what we think possible, which in its turn is something that hinges crucially on theory—on our beliefs about how things work in this world. Control is thus something deeply theory-infected. We can never safely move from

apparent to real adequacy in this regard. We cannot adequately assure that seeming perfection is more than just exactly that. We accordingly have no alternative but to presume that our knowledge (i.e., our purported knowledge) is inadequate at this and indeed at any other particular stage of the cognitive enterprise. One important point about control must, however, be noted with care.

The preceding negative strictures all relate to attainment of perfect control—of being in a position to do everything possible. No such problems affect the issue of amelioration-——of doing some things better and improv-

ing our control over what it was. While control does not help us with perfection, it is crucial for monitoring progress—the improvement at issue in doing more and better than before. It makes perfectly good sense to use its technological applications as standards of scientific advancement. We can determine when we have managed to enlarge our technological mastery, but we cannot meaningfully say what it would be to perfect it. With regard to technical perfectibility, we must recognize that (1) there is no reason to expect that its realization is possible, even in principle, and (2) it is not monitorable: even if we had achieved it, we would not be able to claim success with warranted confidence.

84

.

THE STATUS OF SCIENCE

Temporal Finality? Scientists from time to time indulge in eschatological musings and tell us that the scientific venture is approaching its end.*+ And it is, of course, entirely conceivable that natural science will come to a stop, and will do so not in consequence of a cessation of intelligent life but in C. S. Peirce’s more interesting sense of a completion of the scientific project: of eventually reaching a condition after which even indefinitely ongoing inquiry will not—and indeed in the very nature of things cannot—produce any significant change, because inquiry has come to “the end of the road.” The situation would be analogous to that envisaged in the apocryphal story of the commissioner of United States Patents during the middle 1800s, who resigned his post because there was nothing left to invent.5 Such a position is in theory possible. But here, too, we can never determine that it is actual. For there is no practicable way in which the claim can be validated that science has achieved temporal finality, that our current best efforts in inquiry cannot be improved upon. The question “Is the current state of science, S, final?” is one for which we.can

never legitimate an affirmative answer. The prospect of future changes of S can never be precluded, seeing that there is no practicable way of moving beyond the notion “We presently have (in S) no good reason to think that S will ever change.” To see S in the perspective of finality is to presuppose its completeness.® It is not simply to take the relatively unproblematic stance that that for which S vouches is to be taken as true but to go beyond this to insist that whatever is true finds its place within S. This argument accordingly embeds finality in completeness, and in doing so jumps from the frying pan into the fire. For to hold that if something is 4 This sentiment was abroad among physicists of the fin-de-siécle era of 1890-1900. (See Lawrence Badash, “The Completeness of Nineteenth-Century Science,” Isis 63 [1972]: 48-

58.) And such sentiments are coming back into fashion today. See Hawking, “Is the End in

Sight?”

:

-

5 See Eber Jeffrey, “Nothing Left to Invent,” Journal of the Patent Office Society-22 -

(1940): 479-81.

6 For this inference could be made only if we could move from a thesis of format — ~(Arir ¢ S & r => p) to one of the format (Ar\(r e § & r => ~p), where “=>” represents a grounding relationship of “furnishing a good reason,” and p is, in this case, the particular thesis “S will at some point require drastic revision.” That is, the inference » would go through only if the lack (in S) of a good reason for p were itself to assure the existence (in S) of a good reason for p. But the transition to this conclusion from the given premise would go through only if the former, antecedent fact itself constituted such a good reason; that is, only if we had ~(Ar\(r e S & r> p) > ~p. Thus, the inference would go through only if,

by contraposition, p > (Arr e S & r => p). This thesis claims that the very truth of p will itselfbe a good reason to hold that S affords a good reason for p—in sum, that $ is complete. We are, in effect, confronted with a circle.

THE IMPERFECTIBILITY OF SCIENCE

85

so at all, then S affords a good reason for it is to take a blatantly ambitious

(even megalomaniacal) view of S. Moreover, just as the realization of an erotetic or a pragmatic equilibrium can be the product of narrowness and weakness, so can that of tem-

poral finality. We may think that science is unchangeable simply because we have been unable to change it, but that is just not good enough. Were science ever to come to a seeming stop, we could never be confident that - it had done so not because it is at “the end of the road” but because we are at the end of our tether. We can never ascertain that science has attained the omega condition of final completion, since the possibility of further change lying just around the corner can never be ruled out finally and decisively; no matter how final a position we appear to have reached, the prospects of its coming unstuck with just a last-move effort can never be precluded. Future science is inscrutable. We can never establish the claim that science has settled into a frozen, changeless pattern once and

for all. We have no alternative but to proceed constantly on the assumption that the era of innovation is not yet over—that future science can and will prove to be different science.

The Dispensability of Viewing Perfection as a Realizable Prospect

As these deliberations indicate, the conditions of perfected science in point of description, explanation, prediction, and control are all unrealizable. Our information will inevitably prove inconclusive. We have no reasonable alternative to seeing our present-day science as suboptimal, regardless of the question of what date the calendar shows. One is thus brought back to the stance of the great idealist philosophers (Plato, Spinoza, Hegel, Bradley, Royce) that human knowledge inevitably falls short of recognizably “perfected science” (the Idea, the Absolute) and

must accordingly be looked upon as incomplete. The lesson of these deliberations is that the cognitive situation ofnatural science invites description in theological terms. The ambiguity of the human condition is only too manifest here. We cannot expect ever to reach a position of definitive finality in this imperfect dispensation; we do have knowledge of sorts, but it is manifestly imperfect. Expelled from the Garden of Eden, we are deprived of access to the God’s-eye point of view. Definitive and comprehensive adequacy is denied us: we have no basis for claiming to know “the truth, the whole truth, and nothing but the truth” in scientific matters. ‘We yearn for absolutes but have to settle

for plausibilities; we desire what is definitively correct but have to settle for conjectures and estimates.

86

.

THE STATUS OF SCIENCE

The cognitive absoluteness for which we may yearn represents an ideal that lies beyond the range of practicable realizability. We simply have to do the best we can with the means at our disposal. To aspire to absolutes—for definitive comprehensiveness—is totally unrealistic. In this imperfect epistemic dispensation, we have to reckon with the realities of the human condition. Age disagrees with age, era with era—different states of the art involve naturally discordant conceptions and incommensurate positions. The moral of the story of the Tower-of Babel applies across the reaches of time and place. It is sometimes maintained that such a fallibilist and imperfectionist view of science is unacceptable, because thinking of science as inevitably incompletable and regarding “the definitive answers” inscientificmatters as perpetually unattainable amounts—so it is said—to writing science off as a meaningful project.? But why should this be so? Surely in science, as in the moral life, we can operate perfectly well in the realization that perfection is unattainable. No doubt here and there a scientist may nurse the secret hope of attaining some fixed and final definitive result that will stand, untouchable and changeless, through all subse-

quent ages. However, such unrealistic aspirations are by no means essential to the scientific enterprise as such. In science, as in other domains of human endeavor, the sensible policy is to do the best we can with the tools that come to hand.

Technological Escalation The presumptive imperfectability of science means that a theoretical prospect of unending scientific progress lies before us. But its practical realization is something else again. One of the most striking and important facts about scientific research is that the ongoing resolution of significant new questions facés increasingly high demands for the generation and cognitive exploitation of data. Though the veins of gold run on,they become increasingly hard to mine. This matter of the practical impediments to scientific progress that op- | erate through economic pressures is generally neglected by philosophers of science, who tend to concentrate on theissue of theoretical rather than practical limits to science. They tend to overlook the fact that the instrumentalities of scientificinquiry can be enhanced not only on the side of theoretical resources but preeminently on the side of the technological instrumentalities of observational and experimental intervention. 7 See William Kneale, “Scientific Progress for Ever?” British Journal for the Philosophy of Science 19 (1967): 31. .

THE IMPERFECTIBILITY OF SCIENCE

87

The idea of scientific progress as the correlate of a movement through sequential stages of technological sophistication was already clearly discerned by the astute Charles Sanders Peirce around the turn of the cen-

tury.

Lamarckian evolution might, for example, take the form of perpetually modifying our opinion in the effort gradually to make that opinion represent the known facts as more and more observations come to be collected. . . . But this is not the way in which science mainly progresses. It advances by leaps; and the impulse for each leap is either some new observational resource, or some novel way of reasoning about the observations. Such a novel way of reasoning might, perhaps, also be considered as a new observational means, since it

draws attention to relations between facts which would previously have been passed by unperceived.®

This crucial role for “new observational means” has far-reaching implications for the perfectability of science. For only by operating under new and heretofore inaccessible conditions of observational or experimental systematization—attaining ever-more-extreme temperature, pressure,

particle velocity, field strength, and so on—can we bring new grist to our scientific mill. The impetus to augment our science demands an unremitting and unending effort to enlarge the domain of effective experimental intervention. Physicists often remark that the development of our understanding of nature moves through successive layers of theoretical sophistication.® But scientific progress is clearly no less dependent on continual improvements in strictly technical sophistication. Some of the most startling technological advances in our time are closely associated with basic research. As compared with 25 years ago, the highest vacuum readily achievable has improved more than a thousand-fold; materials can be manufactured that are 100 times purer; the submicroscopic world can be seen at 10 times higher magnification; the detection of trace impurities is hundreds 8 C. S. Peirce, Collected Papers, 8 vols., ed. C. Hartshorne et al., (Cambridge, Mass., 1931-50), 1:44-45 (secs. 108-9).

® “Looking back, one has the impression that the historical development of the physical description of the world consists of a succession of layers of knowledge of increasing generality and greater depth. Each layer has a well defined field of validity; one has to pass beyond the limits of each to get to the next one, which will be characterized by more general and more encompassing laws and by discoveries constituting a deeper penetration into the structure of the Universe than the layers recognized before” (Edoardo Amaodi, “The Unity of Physics,” Physics Today 261, no. 9 [September 1973]: 24). See also E. P. Wigner, “The Unreasonable Effectiveness of Mathematics in the Natural Sciences,” Communications on Pure and Applied Mathematics 13 (1960): 1-14; as well as his “Limits of

Science,” Proceedings of the American Philosophical Society 93 (1949): 521-26. Cf. also chap. 8 of Henry Margenau, The Nature of Physical Reality (New York, 1950).

88

.

THE STATUS OF SCIENCE

of times more sensitive; the identification of molecular species (as in various forms ofchromatography) is immeasurably advanced. These examples are only a small sample. . . . Fundamental research in physics is crucially dependent on advanced technology, and is becoming more so.!°

Our human acquisition of knowledge about the workings of nature is clearly a matter of interaction—a transaction in which both parties, human investigators and nature, must play a crucial role. Most writers on the limits of science operate predominantly on the human side and see the issue as the result—primarily and in the first instance—of human failings and deficiencies (in intellect, learning power, memory, imagination, willpower, etc.). It is too easily ignored that limits to scientific progress also reside in the physical limitations imposed upon us by the nature of the physical universe itself. The perspective afforded by a model of exploration is helpful in this connection. Progress in natural science has heretofore been relatively easy because we have explored nature in our own locality: not our spatial neighborhood, but our parametric neighborhood in the space of physical variables such as temperature, pressure, radiation, and so on. Here,

thanks to the evolutionary heritage of our sensory and cognitive apparatus, we have been able to operate with relative ease and freedom. But as we push out further and further from our home base toward ever more remote frontiers, scientific innovation becomes more and more citi cult—and expensive. Without an ever-developing technology of experimentation and observation, scientific progress would grind to a halt. The discoveries of today cannot be attained with yesterday's instrumentation and techniques. To secure new observations, to detect new phenomena, and to test new hypotheses, an ever-more-powerful technology of inquiry is needed. Throughout the natural sciences, technological progress is a crucial requisite for cognitive progress. We are embarked on an endless endeavor to improve the range of effective observational and experimental intervention. An acute observer has rightly remarked, “Most critical experiments [in physics] planned today, if they had to be constrained within the technology of even ten years ago, would be seriously compromised.” This situation points toward the idea of a technological level, corresponding to a certain state of the art in the technology ofinquiry in re-

gard to data generation and processing. This technology of inquiry falls into relatively distinct levels or stages in sophistication, correlatively with 10D. A. Bromley et al., Physics in Perspective, student ed., National Research Council/ National Academy of Science Publications, (Washington, D.C., 1973), p. 23. Xt Jbid., p. 16.

THE IMPERFECTIBILITY OF SCIENCE

89

successively later generations of instrumentation and manipulative ma-

chinery. These levels are generally separated from one another by substantial (roughly, order-of-magnitude) improvements in performance in regard to such information-providing parameters as measurement exactness, data-processing volume, detection sensitivity, high voltages, high or low temperatures, and so on.

The key aspect of this situation is that once the major findings accessible at a given data-technology level have been attained, further major progress in the problem area requires ascent to a higher level on the technological scale. Every data-technology level is subject to discovery saturation: the body of significant scientific findings realizable at any level is finite (and, indeed, not just finite but relatively small). Within a

given problem area and relative to a fixed level of data technology, one thus faces a situation of eventual exhaustion, thus confronting a localized version of the theory of progress based on the geographic-exploration

model of discovery.

7

But the exhaustion of discoverable phenomena at a given technological level does not, of course, bring progress to a stop. Further major findings

always become realizable when one ascends to the next level of sophistication in data-relevant technology. The need for new data forces one to look further and further from our familiar “home base” in the parametric space of nature. In science one is simply never called on to keep doing what was done before. An ever-more-challenging task is posed by the constantly escalating demands for the enhanced data that can be obtained only at new levels of technological sophistication. Thus, while scientific progress is in principle always possible (there being no absolute or intrinsic limits to significant scientific discovery), the realization of this ongoing

prospect demands a continual enhancement in the technological state of the art of data extraction or exploitation. In science, as in war, one cannot

fight the battles of the present with the armaments of the past. We arrive at a situation of technological arms race against nature, as it were.

Technical Requirements Spell Inevitable Limitations While we can confidently anticipate that our scientific technology will see ongoing improvement in response to continued expenditure of effort, we cannot expect it ever to attain perfection. There is no reason to think that we ever will, or indeed can, reach “the end of the line.” There is always

more to be done: every successive level of technical capability has its limits, whose overcoming opens up yet another more sophisticated level. The accessible pressures and temperatures can in theory always be increased, the low-temperature experiments brought closer to absolute

90

.

THE STATUS OF SCIENCE

zero, the particles accelerated closer to the speed of light, and so on. Any such enhancement of practical mastery carries along (so experience teaches) new phenomena and an enhanced capability to test yet further hypotheses and discriminate between alternative theories conducive to deepening our knowledge of nature. But while there is always more to be done, the doing of it becomes increasingly difficult. And the material resources at our disposal are limited. These limits inexorably circumscribe our cognitive access to the real world. There will always be yet-unrealized interactions with nature of so great a scale (as measured in energy, pressure, temperature, particle velocities, etc.) that their realization would require greater resources than we can commit. And where there are interactions to which we have no access, there are (presumably) phenomena that we cannot discern. It would be very unreasonable to expect nature to confine the distribution of cognitively significant phenomena to those ranges that lie within our reach. : Progress without new data is, of course, possible in various fields of

scholarship and inquiry. The example of pure mathematics, for instance, _shows that discoveries can be made in an area of inquiry that operates without empirical data. But this hardly represents a feasible prospect for natural science. It is exactly the explicit dependency on additional data— the ineliminably empirical aspect of the discipline—that sets natural science apart not only from the formal sciences (logic and mathematics) and also from the hermeneutic ones, which, like the humanities, address

themselves ceaselessly to the imaginative reinterpretation and re-reinterpretation of old data from novel conceptual perspectives. We must, accordingly, come to terms with the fact that we cannot realistically expect that our science will ever—at any given stage of its actual development—be in a position to afford us more than a very partial and incomplete control over nature. For the achievement of control over nature requires not only intellectual instrumentalities (concepts, ideas, theories, knowledge) but also, and no less important, the deployment of physical resources (technology and “power’). And the physical resources at our disposal are restricted and finite. It follows that our capacity to . effect control is bound to remain imperfect and incomplete, with much in the realm of the doable always remaining undone. We shall never be able to travel down this route as we might like to go. At each stage of scientific progress we can, to be sure, readily comprehend and explain what went before. We can always say, “Yes, of course, given that that is how things stand, it is quite understandable that earlier on, when we went at it in a certain way, we arrived at the sort of results

we did—wrong though they are.” But of course this wisdom is one of hindsight only. At no stage do we have the prospect of using foresight to

THE IMPERFECTIBILITY OF SCIENCE

91

predict what lies ahead in the process of data discernment. The impossibility of foreseeing the new phenomena that attend technological progress means that at no point can we prejudge what lies further down the explanatory road. Thus as regards both the observable regularities of nature and the discernible constituents of nature, very different results emerge at the vari-

ous levels of the observational state of the art. And at every stage we deal with a different facet or aspect of things. Given that we can learn about nature only by interacting with it, Newton’s third law of countervailing action and reaction becomes a fundamental principle of epistemology. Everything depends on just how and how hard we can push against nature in situations of observational and detectional interaction. And we cannot “get to the bottom of it” where nature is concerned. Nature always has hidden reserves of power. The ancient Greeks were certainly as intelligent as we are. But given the information technology of the day, it is not just improbable but actually inconceivable that the Greek astronomers could have come up with an explanation for the red shift or the Greek physicians with an account of the bacteriological transmission of some communicable disease. The relevant types of data needed to put such phenomena within cognitive reach simply lay beyond their range. Given the instrumentalities of the times, there just was no way for the Greeks (no matter how well endowed in brain power) to gain physical or conceptual access to the relevant phenomena. Progress in theorizing in these directions was barred, not permanently, but then and there for them, by a technological barrier on the side of data—a barrier as absolute as the then-extant technological barriers in the way of developing the internal-combustion engine or the wireless telegraph. , The Danish historian of science A. G. Drachmann closes his excellent book The Mechanical Technology of Greek and Roman Antiquity with the following observation: “I should prefer not to seek the cause of the failure of an invention in the social conditions till I was quite sure that it was to be found in the technical possibilities of the time.”!? The history of science, as well as that of technology, is crucially conditioned by the limited nature of the technical possibilities of the time. And this is as true for us as it was of the ancients. Technological dependency sets technological limits, first to data acquisition and then to theory projection. Limitations of physical capacity and capability spell cognitive limitations for empirical science. Where there are inaccessible phenomena, there must be cognitive inadequacy as well. 2 A. G. Drachmann, The Mechanical Technology of Greek and Roman Antiquity (Copenhagen, 1963), p. 218.

92,

;

THE STATUS OF SCIENCE

To this extent, at any rate, the empiricists were surely right. Only the most fanatical rationalist could uphold the capacity of sheer intellect to compensate for the lack of data. The existence of unobserved phenomena means that our theoretical systematizations may well be (and presumably are) incomplete. Insofar as certain phenomena are not just undetected but in the very nature of the case inaccessible (even if only for the merely economic reasons suggested above), our theoretical knowledge of nature must be presumed imperfect. Fundamental features inherent in the structure of our interactive inquiry into the ways of the world thus conspire to ensure the incompleteness of our knowledge. And in this domain, incompleteness leads to incorrectness. Scientific

theory formation is, in general, a matter of discerning a local regularity of phenomena in parametric space and then projecting it “across-theboard,” maintaining it globally. But the theoretical claims of science are themselves never localized—they are not spatiotemporally local, and they are not parametrically local either. They stipulate how things are always and everywhere. And so it does not require a sophisticated knowl-

edge of statistics to realize that inductive projection of the sort we make in science is invariably a risky enterprise. And it does not require a sophisticated knowledge of history of science to realize that our worst fears are usually realized—seldom if ever is it the case that our theories survive intact in the wake of extensions in our access to sectors of the parameter space of observation and intervention. The history of science is, after all, a story of successive episodes of leaping to the wrong conclusions.

“Perfected Science” as an Idealization That Affords a Productive

Impetus and a Useful Contrast Conception Reasons of general principle and practical prospect combine to block us from ever achieving the completion of natural science. Even if we achieve control over nature to all apparent intents and purposes—be it cognitive or practical—we cannot be sure of not losing our grip upon it, not because of a loss of strength, but because of cognitive changes that produce a broadening of the imagination and a widened apprehension as to what “having control” involves. We thus have no alternative but to

presume our knowledge (i.e., purported knowledge) to be inadequate at this and indeed at any other particular state of the game—though, of course we have no option of going anywhere else. Here, then, we are

once more brought back to the theme of the inevitable presumption of cognitive incompleteness and cognitive limits, and to the realization that

THE IMPERFECTIBILITY OF SCIENCE

93

the closing of the gap between attainment and objective, between purported and actual truth, is inevitably a matter of idealization. Perfection is simply not a realistic aim in the domain of scientific cognition. It is not a realizable condition of things but at best a useful contrast conception that keeps actual science in its place and helps to sensitize us to its imperfections. The validation of this idealization lies not in its future achievability but in its ongoing utility as a regulative ideal that

affords a contrast to what we do actually attain—so as to highlight its salient limitations. Ideally perfected science is not something we have in hand here and now. And it is emphatically not something toward which we are moving along the asymptotic and approximative lines envisaged by Peirce. Existing science does not and never will embody perfection. The cognitive ideals of completeness, unity, consistency, and definitive finality repre-

sent an aspiration rather than a coming reality, an idealized telos rather than a realizable condition of things. Perfected science lies.outside history as a useful contrast case that cannot be secured in this imperfect world—and yet serves us well by confronting us with an idealization that

does not let us rest content with what we have actually achieved. With respect to the moral aspirations of human will, Kant wrote: “Per-

fection [of the moral will] is a thing of which no rational being in the world of sense is at any time capable. But since it is required [of us] as practically necessary, it can be found only in an endless progress to that complete fitness; on principles of pure practical reason, it is necessary to assume such a practical progress as the real object of our will. . . . Only endless progress from lower to higher stages of moral perfection is possible to a rational but finite being.” Much the same story surely holds on the side of the cognitive perfecting of human knowledge. Here, comparable regulative demands are at work governing the practical venture of inquiry, urging us to the ever-fuller realization of the potentialities of the human intellect. The discontent of reason with achieved actuality is a noble discontent. The scientific project is a venture in self-transcendence; one of the strongest motivations of scientific work is the urge to go beyond present science, to “advance the frontiers.” Our commitment to such an ideal of perfected science in this pursuit of an unattainable systematic completeness is the epistemic counterpart of our commitment to moral ideals. It reflects a striving toward the rational ultimates of completeness, totality, and systematic finality—a striving that is all the more 48 On this issue,.cf. chap. 11 of the author’s Empirical Inquiry (Totowa, N.J., 1982), which develops lines of thought initially set out in Methodological Pragmatism (Oxford, 1973).

4 Immanuel Kant, Critique of Practical Reason, bk. 2, chap. 2, sec. 4; p. 122 (Akad.), trans. L. W. Beck.

94

.

THE STATUS OF SCIENCE

noble because it is not finally attainable. If the work of inquiring reason in the sphere of natural science were completable, this would be something utterly tragic for us, since it could spell the end of one of our most noble endeavors. The idea of perfected science is the focus imaginarius whosepursuit canalizes and structures our inquiry. It represents the ultimate telos of inquiry, the idealized destination of an incompletable journey, a grail of sorts that we can pursue but not possess. The ideal of perfection thus serves a fundamentally regulative function to mark the fact that actuality falls short of our cognitive aspirations. The ideal of perfection, though

unattainable, is nevertheless highly useful because it prevents us from sitting back in the complacent supposition that what we have is good enough. Contemplation of this ideal reminds us that the human condition is suspended between the reality of imperfect achievement and the ideal of an unattainable perfection. In abandoning this conception—in rejecting the idea of an ideal science that alone can properly be claimed to afford a definitive grasp of reality—we would abandon an insight that crucially shapes our view regarding the nature and status of the knowledge to which we lay claim. We would then no longer be constrained to characterize our view of things as merely ostensible and purported. We would then be tempted to the unrealism of regarding our picture of nature as real, authentic, and final, according to it a status that we at bottom realize it does not deserve. Itis useful in this context to heed the distinction between agoal ‘and an ideal. A goal is something that we hope and expect to achieve. An ideal is merely a wistful inkling, a “wouldn’t it be nice if,” something that

figures in the mode of aspiration rather than expectation; we look to its stimulation in the present rather than to its realization in the future. A goal motivates us in striving for its attainment; an ideal motivates and encourages. The former exerts an a fronte pull, the latter an a tergo push. An ideal does not provide us with a destination that we have any expectation of reaching; it is something for whose actual attainment we do not even hope, but whose contemplation renders us dissatisfied with what we have. Itis in this sense that“perfected science” is an ideal. Here, as elsewhere, we must come to terms with the standard gap

between aspiration and attainment. In the practical sphere—in craftsmanship, for example, or the cultivation of our health—we may strive for

perfection but cannot ever claim to attain it. And the situation in inquiry is exactly parallel with what we encounter in such other domains. The value of an ideal, even of one that is not realizable, lies not in the benefit of its attaimment (obviously and ex hypothesi) but in the benefits that accrue from its pursuit. The view that it is rational to pursue an aim only

THE IMPERFECTIBILITY OF SCIENCE

95

if we are in a position to achieve its attainment or approximation is mis-

taken; it can be perfectly valid (and entirely rational) if the indirect benefits of its pursuit and adoption are sufficient—if in striving after it, we realize relevant advantages to a substantial degree. An unattainable ideal can be enormously productive. And so, the legitimation of the ideal of perfected science lies in its productive impetus to the ongoing improvement of inquiry. In this domain, we arrive at the perhaps strange-seeming but nevertheless eminently sensible posture of an invocation of practical utility for the validation of an ideal.® 45 This chapter draws particularly upon the author's Scientific Progress (Oxford, 1978), but its themes are touched upon in all of the books cited on p. 76 above.

Seven The Intelligibility of Nature Explaining the Possibility of Natural Science How is natural science—and, in particular, physics—possible at all? How is it that we insignificant humans, inhabiting the satellite system of a mi-

nor star in one of the world’s myriad galaxies, can manage to unlock nature’s secrets and gain access to its laws? And how can our mathematics— seemingly a free creative invention of thehuman imagination—be used to characterize the modus operandi of nature with such uncanny effectiveness and accuracy? Why is it that the majestic lawful order of nature is intelligible to us humans in the conceptual terms that we have devised? This issue remained unproblematic as long as people thought of the world as the product of the creative activity of mathematicizing intelligence—as the work of a creator who proceeds mathematically (more mathematico) in designing nature. For then one could take the line that

God endows nature with a mathematically intelligible order and mind with a duly consonant mathematicizing intelligence. There is thus no problem about how the two get together—God simply arranged it-that way. But of course if this is to be the canonical rationale for mind’s grasp on nature’s laws, then in forgoing explanatory recourse to God, we also— to all appearances—lose our grip on the intelligibility of nature. Some of the deepest intellects of the day accordingly think that any such prospect of understanding is gone for ever, confidently affirming that there just is no way to solve this puzzle of nature’s being intelligible in a mathematically lawful manner. Erwin Schrédinger characterizes the circumstance that human beings can discover the laws of nature as “a miracle that may well be beyond human understanding,.”! Eugene Wigner asserts that “the enormous usefulness of mathematics in the natural sciences is something bordering on the mysterious, and there is no rational explanation for it,” and he. goes on to wax surprisingly lyrical in maintaining that “the miracle of the appropriateness of the language of mathematics for the formulation of the laws of physics is a wonderful gift which we neither understand nor deserve.” Even Albert Einstein stood in awe 1 Erwin Schrédinger, What Is Life? (Cambridge, 1945), p. 31. 2 Eugene P. Wigner, “The Unreasonable Effectiveness of Mathematics in the Natural Sciences,” Communications on Pure and Applied Mathematics 13 (1960): 2, 4.

THE INTELLIGIBILITY OF NATURE

97

before this problem. In a letter written in 1952 to an old friend of his Berne days, Maurice Solovine, he wrote: You find it curious that I regard the intelligibility of the world (in the measure that we are authorized to speak of such an intelligibility) as a miracle or an eternal mystery. Well, a priori one should expect that the world can be rendered lawful only to the extent that we intervene with our ordering intelligence. .. . [But] the kind of order, on the contrary, created, for example, by

Newton’s theory of gravitation, is of an altogether different character. Even if the axioms ofthe theory are set by men, the success of such an endeavor pre-

supposes in the objective world a high degree of order that we were a priori in no way authorized to expect. This is the “miracle” that is strengthened more and more with.the development of our knowledge. . . . The curious thing is that we have to content ourselves with recognizing the “miracle” without having a legitimate way of going beyond it.?

According to all these eminent physicists we are confronted with a profound mystery. They maintain that we have to acknowledge that nature is intelligible but have no prospect of understanding why this is so. The problem of nature’s intelligibility by means of our mathematical resources is seen as intractable, unresolvable, hopeless. All three of these distinguished Nobel laureates in physics unblushingly employ the word “miracle” in this connection. Perhaps, however, the very question is illegitimate and should not be raised at all. Perhaps the issue of nature’s intelligibility is not just intractable but actually inappropriate and improperly based on a false presupposition. For to ask for an explanation of why scientific inquiry is successful presupposes that there indeed is an explanatory rationale for this fact. But if this circumstance is something fortuitous and accidental, then of

course no such rationale will exist at all. Just this position is advocated by various philosophers—for example, by Karl Popper, who writes: [Traditional treatments of induction] all assume not only that our quest for [scientific] knowledge has been successful, but also that we should be able to ex-

plain why it is successful. However, even on the assumption (which I share) that our quest for knowledge has been very successful so far, and that we now know something of our universe, this success becomes [i.e., remains] miracu-

lously improbable, and therefore inexplicable; for an appeal to an endless series of improbable accidents is not an explanation. (The best we can do, I suppose, is to investigate the almost incredible evolutionary history of these accidents. . .).4 3 Albert Einstein, Lettres 4 Maurice Solovine (Paris, 1956), pp. 114-15.

4K. R. Popper, Objective Knowledge (Oxford, 1972), p. 28.

98

.

THE STATUS OF SCIENCE

Mary Hesse, too, thinks that it is inappropriate to ask for an explanation of the success of science “because science might, after all, be a miracle.”5

And so, on this grand question of how the success of natural science is possible at all, some of the shrewdest scientific intellects of the day avow themselves baffled and unhesitatingly enshroud the issue in mystery or miracle. On this sort of view, the question of the intelligibility of nature becomes an illegitimate pseudoproblem—a forbidden fruit at which sensible minds should not presume to nibble. We must simply rest content with the fact itself, acknowledging that any attempt to explain it is foredoomed to failure because of the inappropriateness of the very project. Clearly, however, such an approach has very questionable merit. Emi-

nent authorities to the contrary notwithstanding, the question of nature’s intelligibility through natural science is not only interesting and important, but it is also surely one that we should, in principle, hope and expect to answer in a more or less sensible way. Clearly, this important issue needs and deserves a strong dose of demystification.

A Closer Look at the Problem How can we make effective use of mathematical machinery to characterize nature’s modus operandi? The pure theorist seems to have a ready answer that runs, roughly, as follows: “Mathematics must apply to reality. Mathematical propositions are purely abstract truths whose validation turns on conceptual issues alone. Accordingly, they hold of this world because they hold of every possible world.” But this response misses the point of present concerns. Admittedly, the truths of pure mathematics obtain in and of every possible world. But they do so only in virtue of the fact that they are strictly hypothetical and descriptively empty—their “if-then” structure leaves them wholly uncommitted regarding the substantive issues of the world’s operations. Their ultimately conceptual status means that the theses of pure mathematics are beside the point of our present purposes. They say nothing whatsoever about the nature of the world, having no bearing at all upon its physical makeup or upon its mode of operation. It is not the a priori truth of pure mathematics that concerns us, its ability to afford truths of reason. Rather, what

is at issue is the empirical applicability of mathematics, its pivotal role in framing the a posteriori, contingent truths of lawful fact that render nature’s ways amenable to reason. After all, the circumstance that pure mathematics holds true in aworld 5 Mary Hesse, Revolutions and Reconstructions in the Philosophy of Science (Blooming-

ton, Ind., 1980), p. 154.

THE INTELLIGIBILITY OF NATURE

99

does not mean that this world’s laws have to.be characterizable in relatively straightforward mathematical terms. It does not mean that nature’s operations have to be congenial to mathematics and graspable in terms of simple, neat, elegant, and rationally accessible formulas. In short, it does not mean that the world must be mathematically tractable and “mathematophile” in being receptive to the sort of descriptive treatment it receives in mathematical physics. -

How, then, are we to account for the fact that the world appears to us

to be so eminently intelligible in the mathematical terms of our natural science? The answer to this question of the cognitive accessibility of nature to mathematicizing intelligence has to lie in a somewhat complex, two-sided story in which both sides, intelligence and nature, must be

expected to have a part. Let us trace out this line of thought one step at a time.

Our Side

-

.

Our human side of this bilateral story is relatively straightforward. After all, Homo sapiens is inseparably a part of nature. We are connected into nature’s scheme of things as an intrinsic component thereof, courtesy of the processes of evolution. Our experience is thus inevitably an experience ofnature. (That, after all, is what “experience” is—our intelligencemediated reaction to the world’s stimulating impacts upon us.) So the kind of mathematics—the kind of theory of periodicity and structure— that we devise in the light of this experience is the kind that is in principle applicable to nature as we can experience it. As Charles Sanders Peirce insisted, evolutionary pressures conform our intellectual processes to the modus operandi of nature. For nature not only teaches us (when we chose to study it) but also forms us (whether or not we chose to study it). And it proceeds in doing the latter in a way that is not, and cannot be, without implications for the former. -

Our mathematics is destined to be attuned to nature because it itself is a natural product as a thought instrument of ours: it fits nature because it reflects the way we ourselves are developmentally emplaced within nature as integral constituents thereof. Our intellectual mechanisms— mathematics included—harmonize with nature because they are themselves a product of nature’s operations as mediated through the cognitive processes of an intelligent creature that uses its intelligence to guide its interaction with a nature into which it is itself fitted in a particular sort of way. The very selfsame forces that are at work in shaping the physical world are also at work in shaping our bodies and brains and in providing the

100

;

THE STATUS OF SCIENCE

stimuli thatimpinge on our senses and our minds. Itis these interactions between thought and world that condition our sense of order and beauty—of regularity, symmetry, economy, elegance. Evolutionary pressure coordinates the mind with its environment. Even as we are destined to find healthy foods palatable and reproductively advantageous activities

pleasant, so nature's inherent order and structure are bound to prove congenial to our mathematical sense of elegance and beauty. The modes of order that attract the attention of mathematical theorists interested in structures (and that underlie their ideas of beautiful theories) are thus, unsurprisingly, also at work in the nature within which these conceptualizations arise. The mathematical mechanisms we employ for understanding the standard features of things themselves reflect the structure of our experience. In particular, the mathematics of an astronomically remote civilization whose experiential resources differ from ours might well be substantially different from mathematics as we ourselves know it. Their dealings with collective manifolds might be entirely anumerical—purely comparative, for example, rather than quantitative. Especially if their environment is not amply endowed with solid objects or stable structures congenial to measurement (e.g., if they were jellyfishlike creatures swimming about in a soupy sea), their “geometry” could be something rather strange from our point of view—largely topological, say, and geared to flexible structures rather than fixed sizes or shapes. Digital thinking might go undeveloped while certain sorts of analogue reasoning might be highly refined. For example, if the intelligent aliens were a diffuse assemblage of units that constitute wholes in ways permitting overlap, then social concepts might become so paramount in their thinking that nature would throughout be viewed in fundamentally social categories, with those aggregates we think of as physical structures contemplated by them in social terms. The processes that underlie their mathematicizing might be very different indeed. Admittedly mathematics is not a natural science but a theory of hypothetical possibilities. Nevertheless, these possibilities are possibilities as conceived by beings who do their possibility conceiving with a natureevolved and nature-implanted mind. It is thus not surprising that the-sort . of mathematics we contrive is the sort of mathematics we find applicable to the conceptualization of nature. After all, the intellectual mechanisms we devise in coming to grips with the world (in transmuting sensory interaction with nature into intelligible experience) have themselves the aspect (among many other aspects) of being nature’s contrivances in adjusting to its ways a creature it holds at its mercy. It is no more a miracle that the human mind can understand the world through its intellectual resources than that the human eye can see it through its physiological resources. The critical step is to recognize that

THE INTELLIGIBILITY OF NATURE

;

101

the question “Why do our conceptual methods and mechanisms fit ‘the real world’ with which we interact intellectually?” is to be answered in basically the same way as the question “Why do our bodily processes and mechanisms fit the world with which we interact physically?” In neither case can we proceed on the basis of purely theoretical grounds of general

principle. Both issues are alike to be resolved in essentially evolutionary terms. It is no more surprising that our minds can grasp nature’s ways than it is surprising that our eyes can accommodate nature’s rays or our

stomachs nature’s food. As we have noted from the outset, evolutionary pressure can take credit for the lot: they are part and parcel of what is mandated by attainment to our niche in nature’s scheme of things. There is nothing “miraculous” or “lucky” in our possession of efficient cognitive faculties and processes—effective “hardware” and “software”—for productive inquiry. If we did not, we just would not be here as inquiring creatures emplaced in nature, thanks to evolutionary processes. Nevertheless, it could perhaps be the case that our mathematics is useful only “locally”’—that we succeed in mathematicizing nature only as re-

gards the immediate local microenvironment that defines our particular limited ecological niche. The possibility still remains open that we secure a cognitive hold on only a small and peripheral part of a large and impenetrable whole. And so, our own one-sided contribution to the matter of

nature’s intelligibility cannot be the whole story regarding the success of science. For even if we do reasonably well in regard to our own immediate evolutionary requirements, this might still be very inadequate in the larger scheme of things. Nature’s receptiveness to our cognitive efforts remains to be accounted for—the fact of its being substantially amenable to reason and not just somewhat (and perhaps only very marginally) so. To clarify this issue, we must therefore move on to consider nature’s contribution to the bilateral mind-nature relationship.

Nature’s Side What needs to be explained for present purposes is why mathematics is not merely of some utility in understanding the world but actually ofvery substantial utility—that its employment can provide intelligent inquirers with an impressively adequate and accurate grasp of nature’s ways. We must thus probe more deeply into the issue of nature’s amenability to inquiry and its accessibility to the probes of intelligence. To be sure, the effective applicability of mathematics to the description of nature is in no small part due to the fact that we actually devise our mathematics to fit nature through the mediation of experience. But how can one get beyond this to establish that nature simply “must” have a

102

.

THE STATUS OF SCIENCE

fairly straightforward law structure? Are there any fundamental reasons why the world that we investigate by the use of our mathematically informed intelligence should operate on relatively simple principles that are readily amenable to mathematical characterization? There are indeed. For a world in which intelligence emerges by anything like standard evolutionary processes has to be pervaded by regularities and periodicities in the organism-nature interaction that produces and perpetuates organic species. And this means that nature must be cooperative in a certain very particular way—it must be stable enough and regular enough and structured enough for there to be appropriate responses to natural events that can be “learned” by creatures. If such appropriate responses are to develop, nature must provide suitable stimuli in a duly structured way. An organically viable environment—to say nothing of a knowable one—must incorporate experientiable structures. There must be regular patterns of occurrence in nature that even simple,

single-celled creatures can embody in their makeup and reflect in their modus operandi. Even the humblest organisms—snails, say, and even algae—must so operate that certain types of stimuli (patterns of recurrently discernible impacts) call forth appropriately corresponding types of response—that such organisms can “detect” a structured pattern in their natural environment and react to it in a way that proves to their advantage in evolutionary terms. Even its simplest creatures can maintain themselves in existence only by swimming in a sea of detectable regularities of a sort that will be readily accessible to intelligence. Their

world must encapsulate straightforwardly discernible patterns and periodicities of occurrence in its operations—relatively simply laws, in other words.® If an inquiring being, who is emplaced by evolution within nature and forms its mathematicized conceptions and beliefs about this nature on the basis of physical interaction with it, is to achieve a reasonably appropriate grasp of its workings, then nature too must “do its part”—it must be duly cooperative. Given the antecedent supposition, it must, obviously, permit the evolution of inquiring beings. And to do this, it must present them with an environment that affords sufficiently stable patterns to

make coherent experience possible, enabling them to derive appropriate information from those structured interactions that prevail in nature at large. Nature’s own contribution to solving the problem of its mathematical intelligibility must accordingly be the possession of a relatively simple and uniform law structure—one that deploys so uncomplicated a set of regularities that even a community of inquirers possessed of only § Conversations with Gerald Massey have helped in clarifying this part of the argument.

THE INTELLIGIBILITY OF NATURE

103

rather moderate capabilities can be expected to achieve a fairly good

grasp of the processes at work in their environment. Accordingly, a world in which intelligence can develop by evolutionary processes also must—on this very basis—be a world amenable to understanding in mathematical terms. It must be a world whose cognizing beings will find much grist to their mill in endeavoring to understand the world. Galileo long ago hit close to the mark when he wrote in his Dialogues that “nature initially arranged things her own way and subse-

quently so constructed the human intellect as to be able to understand her.”? And of course nature’s construction of mathematicizing mind has proceeded by evolutionary processes. The development of life and thereafter of intelligence in the world may or may not be inevitable; the emergence of intelligent creatures on the world’s stage may or may not be surprising in itself and as such. But once they are there, and once we realize that they got there thanks to evolutionary processes, it can no longer be seen as surprising that their efforts at characterizing the world in mathematical terms should be substantially successful. A world in which intelligent creatures emerge through the operation of evolutionary processes must be an intelligible world. On this line of deliberation, then, nature admits of mathematical depiction not just because it has laws (is a cosmos) but because as an evolution-permitting world it must have many relatively simple laws. And those relatively simple laws must be there because if they were not, then nature just would not afford the sort of environment requisite for the evolutionary development of intelligent life. An intelligence-containing world whose intelligent creatures came by this capacity through evolutionary means must be substantially intelligible in mathematical terms.® The apparent success of human mathematics in characterizing nature is thus nowise amazing, given our existence more or less as is. It may or may not call for wonder that intelligent creatures should evolve at all. But ex post facto, once they have safely arrived on the scene through evolutionary means, it is only natural and to be expected that they should be able to achieve success in the project of understanding nature in mathematical terms. A mathematicizing intelligence arrived at through evo7 Galileo Galilei, Dialogue 2, in Le Opere di-Galileo Galilei, Edizio Nazionale, 20 vols.

(Florence, 1890-1909), 7:298. (I owe this reference to Juergen Mittelstrass.) Kepler wrote: “Thus God himself was too kind to remain idle, and began to play the game of signatures, signing his likeness into the world. I therefore venture to think that all nature and all the graceful sky are symbolized in the art of geometry” (quoted in Freeman Dyson, “Mathematics in the Physical Sciencs,” in The Mathematical Sciences, ed. by the Committee on

Support of Research in the Mathematical Sciences [Cambridge, Mass., 1969], p. 99). §To say that such a world must be understandable in mind-invoking terms is not, of course, to claim that those intelligent beings must actually be there to understand it. This

problem raises different issues.

104

THE STATUS OF SCIENCE

lution must for this very reason prove to be substantially successful in achieving adequation to the world’s ways. The strictly hypothetical and conditional character of this general line of reasoning must be recognized. It does not maintain that by virtue of some sort of transcendental necessity the world has to be simple enough for its mode of operation to admit of elegant mathematical representation. Rather, what it maintains is the purely conditional thesis that if intelligent creatures are going to emerge in the world by evolutionary processes, then the world must be mathematophile, with various of its processes readily amenable to mathematical representation. It must be stressed, however, that this conditional fact is quite suffi-

cient for present purposes. For the question we face is why we intelligent creatures present on the world’s stage should be able to characterize its operations successfully in terms of our mathematics. The conditional story at issue fully suffices to accomplish this particular job. One brief digression to avert a possible misunderstanding is in order. Nothing whatever in our argumentation can properly be construed to claim that the development of mathematics is an evolutionary requirement or desideratum as such—that creatures are somehow impelled to develop mathematics because it advantages them in the struggle for existence. (This idea would put the cart before the horse, seeing that evolution produced human beings long before humans produced mathematics.) To say that intelligence, the precondition of mathematics, is of evolutionary advantage, is not to claim that this is the case with mathematics itself, All that is being maintained is (1) that intelligence is (in certain circumstances) of evolutionary advantage, (2) that any sufficiently intelligent creature can develop a mathematics (a theory of structure), and (3) that any sufficiently intelligent creature must be able to develop an effectively applicable “mathematics” in any world able to give rise to it through evolutionary means. |

Synthesis A brief review of the results of the preceding deliberations is in order. The overall question of the intelligibility of nature has two sides: 1. Why is mind so well attuned to nature? 2. Why is nature so well attuned to mind?

;

The preceding discussion has suggested that the answers to these-questions are not all that difficult—at least at the level of schematic essentials. The crux is simply this: Mind must be attuned to nature, since intelligence is a generalized guide to conduct that has evolved as a natural

THE INTELLIGIBILITY OF NATURE

105

product of nature’s operations. And nature must be substantially accessible to mind if intelligence manages to evolve within nature by a specifically evolutionary route. For nature to be intelligible, then, there must be an alignment that

requires cooperation on both sides. The analogy of cryptanalysis is suggestive. If A is to break B’s code, there must be due reciprocal alignment. _ If A’s methods are too crude, too “hit and miss,” A can get nowhere. But even if A is quite intelligent and resourceful, efforts to break the code

cannot succeed if B’s procedures are simply beyond A’s powers. (The cryptanalysts of the seventeenth century, clever though they were, could get absolutely nowhere in applying their investigative instrumentalities to a high-level naval code of World War II vintage.) Analogously, if mind and nature were too far out of alignment—if mind were too “unintelligent” for the complexities of nature or nature too complex for the capacities of mind—the two just could not get into step. It would be like trying to rewrite Shakespearein a pidgin English with a 500-word vocabulary or like trying to monitor the workings of a system with ten degrees of freedom by using a cognitive mechanism capable of keeping track of four of them. If something like this were the case, mind would not constitute an evolutionary advantage. It would then be better to adopt an alignment process that does not take the cognitive route to the guidance of action. Just as any creature that evolves in nature must find due physical accommodation within it (a due harmonization of its bodily operations with its physical environs), so any mind that evolves in nature must find due intellectual accommodation within it (a due harmonization of its intellectual

operations with its structural environs). In consequence, there must be a due equilibration between the mind’s mathematicizing operations and the world’s mathematical structure. The solution of our problem thus roots in the combination of two considerations: (1) a world that admits of the evolutionary emergence of intelligence must be sufficiently regular and simple (i.e., must be mathematophile), and (2) a sufficiently powerful intelligence must be able effectively to comprehend in mathematical terms any world in which it gains its foothold by evolutionary means. The possibility of a mathematical science of nature is accordingly to be explained by the fact that, in the light of evolution, intelligence and intelligibility must stand in mutual coordination. Two points areaccordingly paramount here: 1. Once intelligent creatures evolve, their cognitive efforts are bound to have some degree of adequacy because evolutionary pressures align them with nature’s ways. 2. It should not be surprising that this alignment can eventually produce a

106

.

THE STATUS OF SCIENCE

substantially effective mathematical physics, because the structure of the operations of a nature that engenders intelligence by an evolutionary route is bound to be relatively accessible to this intelligence.

No doubt, this somewhat schematic account requires much amplification and concretization. A long and complex tale must be told about physical

and cognitive evolution to fill in the details needed to put such an account into a properly compelling form. But there is surely good reason to hope and expect that a tale of this sort can ultimately be told. And this is the pivotal point. Even if one has doubts about the particular outlines of the evolutionary story sketched here, it must be acknowledged that some such story can provide a perfectly workable answer to the question of why nature’s ways are intelligible to us humans in terms of our mathematical instrumentalities. The mere fact that such an account is in principle possible shows that the issue need not be painted in the black on black of impenetrable mystery. There may indeed be mysteries in this general area. (Questions such as “Why should it be that life evolves in the world?” and—even more

fundamentally——“Why should it be that the world exists at all?” may plausibly be proposed as candidates.) But be that as it may, the presently deliberatedissue of why nature is intelligible to us, and why this intelligibility should incorporate a mathematically articulable physics, does not qualify as all that mysterious, let alone miraculous. There is simply no need to join Einstein, Schrédinger, and company in regarding the intelligibility of nature as a miracle or a mystery-that passes all human understanding.If we are willing to learn from science itself how nature operates and how we go about conducting our inquiries into its workings, then we should be able increasingly to remove the shadow of incomprehension from the problem of how it is that a being of this particular sort, probingan. environment of that particular type, and doing so by means of those particular evolutionarily developed cognitive and physical instrumentalities, manages to arrive at a relatively workable account of how things work in the world. We should eventually be able

to see it as only plausible and to be expected that inquiring beings should emerge in nature and get themselves into a position to makearelatively good job of coming to comprehend it. We can thus look to-science itself for the materials that enable us to understand how natural science is possible. And there is no good reason to expect that it will let us down in this regard. Admittedly, any such scientifically informed account of science’s ability to understand the world is in a way circular. It explains the possibility of our knowledge of nature on the basis of what we know of nature’s ways. Its explanatory strategy uses the deliverances of natural science retro-

THE INTELLIGIBILITY OF NATURE

107

spectively to provide an account of how an effective natural science is possible. Such a procedure is not, however, a matter of vitiating circularity but rather one of the healthy and virtuous self-sufficiency of our

knowledge that is in fact an essential part of its claims to adequacy.® Any scientific world-picture that does not provide materials for explaining the

success of science itself would thereby manifest a failing in its grasp of _ the phenomena ofnature that betokens its own inadequacy.

Implications But does suchascientific explanation of the success ofnatural science not explain too much? Does its validation of mathematical exactness in science not lead to the (obviously problematic) consequence that “science gets it right”—a result that would flyin the face of our historical experience of science’s fallibilism? By no means! Itis fortunate (and evolutionarily most relevant)that we are so positioned within nature that many “wrong” paths lead to the “right” destination—that flawed means often lead us to cognitively satisfactory ends. If nature were a combination lock where we simply “had to get it right”—and exactly right—to achieve success in implementing our beliefs, then we just would not be here. Evolution accordingly does not provide an argument that speaks unequivocally for the adequacy of our cognitive efforts. On the contrary, properly construed, it is an indicator of our capacity to err and “get away with it.” Admittedly, applicative success calls for some alignment of thought-governed action with “the real nature of things’—-but only enough to get by without incurring overly serious penalties in failure. 9 Just this approach is the salient feature of W.V.O. Quine’s program of “Epistemology Naturalized.” According to Quine, “Our question, ‘Why is science so successful?” [is to be] taken . . . as a scientific question, open to investigation by natural science itself. . . . Individuals whose similarity groupings conduce largely to true expectations have a good chance of finding food and avoiding predators, and so a good chance of living to reproduce their kind. ... I am not appealing to Darwinian biology to justify induction. This would be circular, since biological knowledge depends upon induction, and then observing that Darwinian biology, if true, helps explain why induction is so efficacious as it is” (“The Nature of Natural Knowledge,” in Mind and Language, ed. S. Gutemplan [Oxford, 1975], p. 70). My only reservation about this sensible passage relates to Quine’s feeling a need to apologize for using the products of induction in the course ofjustifying the use of induction. Such a procedure will, in suitable circumstances, not only not be vicious but even be appropriate (see Gerhard Vollmer, “On Supposed Circularities in an Empirically Oriented Epistemology,” in Evolutionary Epistemology, ed. G. Radnitzky and W. W. Bartley III [La Salle, Ill., 1987], pp. 163-200.) What is at issue is only a matter of getting a suitable starting point— even as we use a product of nutrition, namely energy, to carry on the process of nutrition

_ itself. (See also the author's Scepticism [Oxford, 1980].)

108

.

THE STATUS OF SCIENCE

The success of science should be understood somewhat on analogy with the success of the thirsty man who drank white grape juice, mistaking it for lemonade. It is not that he was roughly right—that such grape juice is “approximately” lemonade. Rather, the point is that his beliefs are not wrong in ways that lead us to his being baffled in his present

purposes—that such defects as they have do not matter for the issues currently in hand. The fact is that we simply cannot, in the circumstances, help laboring under the impression that our science is highly successful, even though subsequent experience repeatedly disillusions us in this regard. It follows that intelligence and the science it devises must pay off in terms of applicative success—irrespective of whether it manages to get things substantially right or not. We thus arrive at the picture of nature as an error-tolerant system. Consider the hypothetical situation of a species of behaviorally beliefguided creatures living in an environment that invariably exacts a great penalty for “getting it wrong.” Whenever the creature makes the smallest mistake, the least little cognitive misstep—bang, it’s dead! Our hypoth-

esis is not viable: any such creature would long ago have become extinct. It could not even manage to survive and reproduce long enough to learn about its environment by trial and error.If the world is to be a home for intelligent beings who develop in it through evolution, then it has to be benign-—it has to be error tolerant. For if nature were not error forgiving, a process of evolutionary trial and error could not work in matters of cognition, and intelligent organisms could not emerge at all.!° The applicative success of our science is not to be explained on’‘the basis of its actually getting at the real truth, but rather in terms of its being the work of a cognitive being who operates within an error-tolerant environment—a world setting where applicative success may attend even theories that are substantially off the mark. Its applicative efficacy undoubtedly requires some degree of alignment between our scientific world-picture and the world’s actual arrangements—but only just enough to yield the particular successes at issue. No claims to finality or perfection can therefore be substantiated for our science as it stands here andnow. os

Evolution is indeed the guarantor of the reciprocal attunement of mind and nature envisioned by classical idealism. But it by no means requires this attunement to be of a very high grade by some absolute standard of “

10 This unavoidable error-tolerant aspect of nature blocks any prospect of a naive “it works, therefore it’s true” pragmatism at the level of theses. To be sure, as regards largescale methods for providing action-guiding theses, the situation is different. Here “it works systematically, therefore it is cogent (as a cognitive method—-i.e., its deliverances are rationally credible)” is something else again. See the author'sMethodologicalPragmatism (Oxford, 1977).

THE INTELLIGIBILITY OF NATURE

109

perfection. It is one thing to be right and another to be so badly wrong that one is baffled of one’s purposes. And this sort of functional adequacy is something very different from truth. Such a perspective indicates that the success of the applications of our current science does not betoken its actual truth but merely means that those ways (whatever they be) in which it fails to be true are immaterial to the achievement of success— that in the context of the particular applications at issue, its inadequacies and incorrectnesses lie beneath the penalty threshold of failure. Accordingly, the success of natural science can be explained well short of the supposition that it manages to get at the definitive truth of nature’s

arrangements. Its success does not betoken science’s ultimate adequacy, seeing that evolution requires an error-tolerant nature that can afford success, despite falsity. If seeming success in intellectually governed operations could not attend even substantially erroneous beliefs; then we cognizing beings who have to learn by experience—by trial and error— just could not have made our way along the corridor of time. This critical fact that evolution requires an error-tolerant environment means that we can explain the impressive successes of our mathematicizing natural science without needing to stake untenable claims as to its definitive correctness.!!

Such evolutionary ruminations indicate that the capacity to develop an effective natural science and also the motivation to pursue this cognitive project of inquiry form a natural part of humanity’s evolutionarily developed natural heritage. But at the same time, they suggest that the scope, scale, and nature of such a science is bound to be essentially conditioned

by the cognitive resources and interests that we humans bring to bear on its development. The conceptual mechanisms we deploy in studying the

world’s ways are instrumentalities of our own devising, and in this case as in others, the sorts of tools we use condition the sorts of artifacts we can create. We are constrained to depict the world in terms of concepts, categories, and schematism to whose formation we ourselves make a decisive contribution. This important theme—which harks back to the classical idealist theme of the formative contribution of the knower to the character of the known—deserves closer scrutiny. i The present discussion draws on the author's Riddle ofExistence (Lanham, Md., 1984). Parts of Methodological Pragmatism are also relevant.

Eight Our Science as Our Science

Scientific Relativism

There is no cogent reason of general principle to think that our own human scientific view of the world is cognitively absolute, devoid of all situational gearing to the specific character of the reciprocal interaction between the world and its investigators. We must recognize that, even in cognition, process and product are coordinated—that our scientific concepts and theories are the product of an interaction in which both parties, both nature and ourselves, play a formative role. The results of our investigations of nature are accordingly such that the respective contribution of these parties simply cannot be separated from one another— at any rate by us. The abstract query “What is the discoverable character of nature— what are the constituents of physical reality and what are the discernible

regularities that govern them?” remains incomplete and indeterminate unless we first resolve the question “Detectable and discernible by whom>” For the issue is one that is inevitably relativized to the natureinteractive faculties and instrumentalities at the disposal of investigators. To be sure, the regularities of nature are something perfectly real and independent of the wants and wishes ofinquirers. Nevertheless their re-

alityis a relational reality—a matter of interaction between theworld.and its investigators. The answers that natural science gives to our questions

are indeed independent of our will, but not of our intellect. We must recognize our knowledge of the world as being, in inseparable part, a function of the manner of our evolutionary attunement to nature. Science indeed furnishes appropriate information about the world, but information that is appropriate from our own particular cognitive standpoint. What our empirical inquiries provide is not a picture of “reality in itself” but one of “reality as it presents itself to us inquirers of a certain particular sort.” On such a view, inquiry yields results that are inherently

relational. Of course, this does not mean that it is infeasible to contemplate such a thing as a self-subsistingly nonrelational reality. Rather, it means that reality as we picture it—both in fact and in prospéct—is bound to be a complex composite in whose constitution we ourselves play a formative role through the particular characteristic of our evolutionary attunement to nature.

,

OUR SCIENCE AS OUR SCIENCE

111

Factual (empirically based) theses regarding how things work in the world are always correlative with the mode of monitoring, of interaction/

detection, of what can be discerned about nature from the peculiar vantage point of a being emplaced therein and equipped with certain facilities for interaction with it. And this circumstance underwrites a certain sort of scientific relativism. For science is a matter of explaining how things work in the world, and we can explain only what we can discern. But our facilities for discernment reflect our mode of emplacement within nature. This inevitably conditions the kinds of scientific issues we can even address (let alone resolve)—relativizing them in part to the character of our evolved enmeshment in nature’s scheme of things. Natural science is in an important sense our science, and it describes reality not in a categorial and absolute way but by providing investigator-relative results that differ with different modes of interaction between investiga;

tors and their natural environment.

oH

The regularities of nature that can be discovered by us depends on. who we are. Our reality (reality as we can know it) is something whose nature is relativized to us humans—which is, of course, true of other cognizing beings as well. Reality-as-we-know-it is something relational, howeyer much“reality as such” may fail to be so. (The range of fact isalways broader than that of achievable knowledge. )! 3) das What is brought into question from this perspective is not the exis; tence of “the real world” that is self-subsistingly mind independent-but the status of our knowledge of it. For it emerges that any»realizahle knowledge of (“mind-independent”) reality is not itself mind indepéndent but represents information grounded in an interpretation of experi: ence. We arrive at the recognition that our information about theeworld is developed from the characteristically human perspective inthé overall disposition of things in nature and affords a view of reality as “experien:

tially accessible “from the human point of view.” It is species ‘reldtivizéd because the reality with which it deals is our reality—naturé asithe! cog: nitive instrumentalities of our own vnparticular evolutionary heritage'réveal it to us.? betsgo1g ed tiv To affirm this position is accordingly not to espouse skepticism“and 10 eotiee xelgaroo 6 10 1 This issue and its ramifications will be addressedin chapol4ivilivio oftitgsing A 2 Obviously, no sensible relativism can maintain that ‘fy AGUS ete YBio5 the; eby pulling the mat out from under its own feet. Relativism mustbe2 developed 1respect to a limited range; we cannot say, “No proposition is to bedesétted|Rohe b ‘But“ouily;SNS, proposition belonging to the range Ris to be asserted absolifalf'_Sw)HSI BIRD peataad itself should of course not belong to R. And thisis all we need for present purposes, seeing that the thesis “The contentions of natural science'aré qivéstigator'relativided??is of course itself not a contention of natural science. It is’a'thesis‘aboit the déilain?ivdther thantone lying within it.

of uL8@£) SS wisioad looimonotizd lpyoh oslt Yo

112

THE STATUS OF SCIENCE

deny the attainability of any and all duly evidentiated information about the world: Rather, it is to accepta (realistic) mode of contextualism or relativism—to recognize that the modus operandi of investigators must always crucially condition the sort of information that their science is in a position to provide about the world. One’s situation in point of data accessibility and information-processing capacities is bound to condition the scientific world-picture at which one will arrive.

The Problem of Extraterrestrial Science But can we actually come to terms with this idea that natural science is not something altogether investigator independent—not a body of selfsubsistent fact, wholly disconnected from the proceedings of its practitioners. Do not the findings of science represent a destination that all (sufficiently clever) inquiring intelligences are bound ultimately to reach in common? To elucidate this idea of investigator convergence, let us consider the prospect that a civilization of extraterrestrial aliens, living on a planet in some far-off galaxy, might also develop natural science and its concomitant technology.? This simple-seeming issue is in fact one of great complexity. And this complexity relates not only to the actual or possible facts of the matter but also—and crucially—tofundamental questions about the very idea of“natural science”thatis involved here. - To begin with, there is the matter ofjust what it means for there to be another science-possessing civilization. Note that this is aquestion tthat we are putting, one posedin terms of the applicability of our term \science.”It pivots on the issue of whether we would be prepared to recognize what those aliens are doing as a matter of forming beliefs (theories) about how things work in the world and to acknowledge that they are involved in testing these beliefs observationally or experimentally and in applying them ‘in practical (technological) contexts. We must, to begin with, be prepared to accept those alien creatures as (nonhuman) persons, duly equipped with intellect and will, and thereupon we must embark

on a complex series of claims with respect to their cognitive activities. A scientific civilization is not merely one that possesses intelligence and social’ organization but one that puts this intelligence and organization to workin’ a veryparticular way. And this opens up a rather subtle issue of priority. in. n regard to process versus product. Is the requisite for 3For. a usefulaccountofthehistorical backgrounds of this line of speculation, see Frank J. Tigler, “A Brief History. of the Extraterrestrial Intelligence Concept,” Quarterly Journal of the Royal Astronomical Society 22 (1981): 133-45.

OUR SCIENCE AS OUR SCIENCE

113

‘4 civilization’s “having a science” primarily a matter of the substantive content of their doctrines (their belief structures and theory complexes), or is it primarily a matter of the aims and purposes with a view to which their doctrines are formed (control, prediction, explanation, and the like)? The former issue of content turns on how similar their scientific beliefs are to ours, which is clearly something in which we would be ill advised

to put much stock. After all, the speculations of the nature theorists of pre-Socratic Greece, our ultimate predecessors in the scientific enter- prise, bear precious little substantive resemblance to our present-day sciences, nor does-the content of contemporary physics bear all that much resemblance to that of Newton’s day. Thus, in adjudgingscientific status,

we had best leave the matter of doctrinal content aside and give prime emphasis to matters of process and purpose rather than products-and results. Accordingly, the matter of these aliens “having a science” should be seen as turning not on the extent to which their findings resemble ours but on the extent to which their project resembles ours. The crux is whether we are engaged in the same sort of rational inquiry in terms of the sorts of issues being addressed and the observational/experimental ways in which they are approached. The issue is at bottom not one of the

substantive similarity of their science to ours but one of the functional equivalency of their projects to the scientific enterprise as we know it. Only if they are pursuing such goals as description, explanation, prediction, and control of nature, and do so by using something like “the scientific method,” will they be doing science. This perspective has farreaching implications.

The Potential Diversity of “Science” It is illuminating to approach the problems of different sciences from the angle of the question “To what extent would the functional equivalent of natural sciencebuilt up by the inquiring intelligences of an astronomically remote civilization be bound to resemble our science?” In reflecting on this question and its ramifications, one soon comes to realize that there is an enormous potential for diversity.

For one thing, the subject-matter orientation of the science of an alien civilization might be altogether different. All their efforts might conceivably be devoted to the social sciences—to developing highly sophisticated analogues of psychology and sociology, for example.In particular, if the intelligent aliens were a diffuse assemblage of interacting intelli-

114

THE STATUS OF SCIENCE

gences constituting wholes in ways that allow of overlap,‘ then the role of social concepts might become so paramount for them that nature would throughout be viewed in fundamentally social categories, with the sorts of aggregates we think of as physical structures contemplated by them in something like political terms. Accordingly, their natural science might deploy explanatory mechanisms very different from ours. Communicating by some sort of telepathy based upon variable odors or otherwise exotic signals, their science might envision a complex theory of empathetic thought-wave transmittal through an ideaferous ether. Again, the aliens might scan nature very differently, thus arriving at very different views about it. Electromagnetic phenomena might lie wholly outside the ken of variant life-forms; if their environment does not afford them lodestones and electrical storms, the occasion to develop

electromagnetic theory might never arise. The course of scientific development tends to flow in the channel of practical interests. A society of porpoises will probably lack crystallography but may develop a very sophisticated hydrodynamics; one of the molelike creatures might never dream of developing optics or astronomy. One’s language and thought processes are bound to be closely geared to the world as one experiences it. As is illustrated by the difficulties we ourselves have in bringing the language of everyday experience to bear on subatomic phenomena, our concepts are ill attuned to facets of nature different in scale or structure from our own. We can hardly expect a science that reflects such parochial preoccupations to be a universal fixture. The interests of creatures shaped under the remorseless pressure of evolutionary adaptations to very different and endlessly variable environmental conditions will doubtless be oriented in directions very different fromanything |that is

familiar to us.

\

The scientific laws that can actually figure in a body of science have to represent detectable regularities in nature. But detection of course varies drastically with the mode of observation—that is, with the sort of resources that different creatures have at their disposal to do their detecting. Everything will depend on exactly how nature impacts upon a‘creature’s senses and their instrumental extensions. Even if we were to discern everything that we can manage to detect, we would still be very far from having got hold of everything available to others. (And the reverse is equally true.) Since the laws we find are bound to reflect the sorts of data we can get hold of, the lawsthatwe (or anybody else) can manage to formulate will depend crucially on one’s place within nature’s scheme of things. , 4 Note the discussion in Goesta Ehrensvaerd, Man on Another World (Chicago, 1965), pp. 146-48.

OUR SCIENCE AS OUR SCIENCE

115

Supporting considerations for this position have been advanced from very different points of view. One example is a thought experiment suggested by Georg Simmel in the last century, which envisaged an entirely different sort of cognitive being: intelligent and actively inquiring creatures (animals, say, or beings from outer space) whose experiential modes differ substantially from ours.* Their senses respond quite differently to physical influences: they are relatively insensitive, say, to heat and light,

but substantially sensitized to various electromagnetic phenomena. Such intelligent creatures, Simmel held, could plausibly be supposed to operate within a significantly variant framework of empirical concepts and categories; the events and objects of the world of their experience might be radically distinct from those of our own: their phenomenological predicates, for example, might have altogether variant descriptive domains. In a similar vein, William James wrote: “Were we lobsters, or bees, it might

be that our organization would have led to our using quite different modes from these [actual ones] of apprehending our experiences.. It might be too (we cannot dogmatically deny this) that such categories unimaginable by us to-day, would have proved on the whole as serviceable for. handling our experiences mentally as those we actually use.”6 The science of a different civilization would thus inevitably be closely tied to the particular pattern of their interaction with nature as funneled through the particular course of their evolutionary.adjustment to their specific environment. The “forms of sensibility” of radically different beings (to invoke Kant’s useful idea) are going to be radically diverse from ours. The direct chemical analysis of environmental materials might prove highly useful, and bioanalytic techniques akin to our senses of taste and smell could be very highly developed, providing them with “experiences” of their chemical environment of a sort very different from ours. Moreover, the conceptualization of an alien science might also be very different. For we must also reckon with the possibility that a remote civilization might operate with a drastically different system of concepts in its cognitive dealings with nature. Different cultures and different intellectual traditions, to say nothing of different sorts of creatures, are bound to describe and explain their experience—their world as they conceive it—in terms of concepts and categories of understanding substantially different from ours. They would diverge radically with respect to what the Germans would call their Denkmittel—the conceptual instruments they employ in thought about the facts (or purported facts) of the world. They could, accordingly, be said to operate with different conceptual schemes, 5Georg Simmel, “Uber eine Beziehung der Selektionslehre zur Erkenntnistheorie,” Ar-

chiv fiir systematische Philosophie und Soziologie 1 (1985): 40—41. ® William James, Pragmatism (New York, 1907), p. 171.

116

.

THE STATUS OF SCIENCE

with different ideational tools used to “make sense” of experience—to characterize, describe, and explain the items that figure in the world as

they view it. The taxonomic and explanatory mechanisms by means of which their cognitive business is transacted could differ so radically from ours that intellectual contact with them would be difficult or impossible. To clarify this situation, consider the analogy between perception and conception. It is clear that there is no single uniquely correct and appropriate mode of perception. Different sorts of creatures have very different sorts of senses and perceive the world very differently by their means. The situation with conception is in many ways similar to this. Conceptualization too can clearly be done in very different ways. To be sure, there is also an important disanalogy. Conception lets us step be-

yond the domain of our idiosyncratic perception and enables us, as human scientists, to describe how very different sorts of creatures can sense

the world—for example, how they go about monitoring their environment in physical interaction in ways different from ours. We can represent their sensory framework within our conceptual framework. But this we cannot do with conception itself. There is no supraconceptual vantage point from which we (at any rate) can compare and contrast how different sorts of creatures might conceive the world differently from ourselves. It is a near-trivial truth that we ourselves must inevitably understand any sort of concepts in terms of our concepts. (Once we appropriate concepts,

we ipso facto make them ours.) A comparison of the science of different civilizations here on earth suggests that it is not an outlandish hypothesis to suppose that the very topics of alien science might differ dramatically from those of ours. In our own case, for example, the fact that we live on the surface ofthe earth (unlike whales), the fact that we have eyes (unlike worms) and thus can see the heavens, the fact that we are so situated that the seasonal positions of heavenly bodies are intricately connected with agriculture—all these facts are clearly connected with the development of astronomy. Since variant creatures would experience nature in ways radically different from us, they can be expected to raise very different sorts of questions. Indeed, the mode of emplacement within nature of alieninquirers might be so different as to focus their attention on entirely different aspects or constituents of the cosmos. And if the worldis sufficiently complex and multifaceted, they might concentrate upon aspects of their environment that mean nothing to us, with the result that their natural

science is oriented in directions very different from ours.” 7 Benjamin Lee Whorf’s anthropological investigations pointed him in much this same direction. He wrote: “The real question is: What do different languages do, not with artificially isolated objects, but with the flowing face of nature in its motion, color, and changing form; with clouds, beaches, and yonder flight of birds? For as goes our segmentation of the

OUR SCIENCE AS OUR SCIENCE

117

Epistemologists often insist that people whose experience of the world differs substantially from our own are bound to conceive of it in very different terms. Sociologists, anthropologists, and linguists talk in much the same way, and philosophers of science have recently also come to say the same sorts of things.® There is surely much to be said for this general position. It is (or should be) clear that there is no simple, unique, ideally adequate concept framework for “describing the world.” The botanist, horticulturist, landscape gardener, farmer, and painter will operate from

diverse cognitive points of view to describe one selfsame vegetable garden—and none of them has a monopoly on correctness here, though, of course, within each domain there are decidedly appropriate and inappropriate approaches. It is mere mythology to think that-the phenomena of

nature lend themselves to only one single correct style of descriptive and explanatory conceptualization. As we have noted even in the case of mathematics, different sorts of creatures are bound to make use of different conceptual schemes for the representation of their experience. To insist on the ultimate uniqueness of science is thus to succumb to an illusion. In theory—and perhaps in fact as well—there are very different cognitive perspectives for comprehending nature, no one of them more adequate or more correct than any other independently ofthe cognitive situation of their possessors. Science, after all, is an empirical enterprise. And as long as the fun-

damental categories for the characterization of experience—the modes of spatiality and temporality, of structural description, functional connection, and explanatory rationalization—are not seen unrealistically as necessary features of intelligence as such but as evolved cognitive adaptations to particular contingently constituted modes of emplacement in and interaction with nature, there will be no reason to expect uniformity. Sociologists of knowledge tell us that even for us humans here on earth, our Western science is but one of many competing ways of conceptualizing the world’s processes. And when one turns outward toward space at large, the prospects of diversity become virtually endless. The idea of a conceptually different science is usefully illuminated by casting the issue in temporal rather than spatial terms. In a way, the descriptive characterization of alien science is a project rather akin in its difficulty to that of describing our own future science. It is a key fact of life that progress in science is a process of ideational innovation that always places certain developments outside the intellectual horizons of earface of nature, so goes our physics of the cosmos” (“Language and Logic,” in Language, Thought, Reality, ed. J. B. Carroll [Cambridge, Mass., 1956], pp. 240-41). See also the

interesting discussion in Thomas Nagel, “What Is It Like to Be a Bat?” in Mortal Questions (Cambridge, Mass., 1976).

§ Thomas Kuhn, The Structure of Scientific Revolutions (Chicago, 1962).

118

;

THE STATUS OF SCIENCE

lier workers. The very concepts we think in terms of become available only in the course of scientific discovery itself. Like the science of the remote future, the science of remote aliens must be presumed to be such that we really could not achieve intellectual access to it on the basis of our own achieved position in the cognitive scheme of things. Just as the technology of a more advanced civilization would be bound to strike us as magic, so its science too would be bound to strike us as magic, as incom-

prehensible gibberish—until we had learned it “from the ground up.” They might (just barely) be able to teach it to us, but they could not explain it to us by transposing it into our terms. Differing radically in mode of formulation, in subject-matter orientation, and in conceptuali-

zation, their science could well be something that we could not begin to recognize as such.

The One-World, One-Science Argument One writer on extraterrestrial intelligence poses the question “What can we talk about with our remote friends?” and answers by remarking: “We have a lot in common. We have mathematics in common, and physics, and astronomy.”® Another maintains that “we may fail to enjoy their music, understand their poetry, or approve their ideals; but we can talk about matters of practical and scientific concern.”!° But to say this sort of thing is to presume a good deal. With respect to his hypothetical Planetarians, the ingenious Christiaan Huygens wrote, three centuries ago: Well, but allowing these Planetarians some sort of reason, must it needs’be the same with ours? Why truly I think ’tis, and must be so; whether we consider

it as applied to Justice and Morality, or exercised in the Principles and Foundations of Science. . . . For the aim and design of the Creator is every where the preservation and safety of his Creatures. Now when such a reason as we are masters of, isnecessary for the preservation of Life, and promoting of So" ciety (a thing that they be not without, as we shall show) would it not be

strange that the Planetarians should have such a perverse sort of Reason given them, as would necessarily destroy and confound what it was design’d to maintain and defend? But allowing Morality and Passions with those Gentlemen to be somewhat different from ours, . . . yet still there would be no doubt, but that in the search after Truth, in judging of the consequences of things, in reasoning, particularly in that form which belongs to Magnitude or Quantity about which their Geometry (if they have such a thing) is employ’d, there § See E. Purcell in Interstellar Communication: A Collection of Reprints and Original Contributions, ed. A.G.W. Cameron (New York, 1963), p. 184.

10 Paul Anderson, Is There Life on Other Worlds? (New York, 1963), p. 130.

OUR SCIENCE AS OUR SCIENCE

119

would be no doubt I say, but that their Reason here must be exactly the same,

and go the same way to work with ours, and that what’s true in one part will hold true over the whole Universe; so that all the difference must lie in the degree of Knowledge, which will beproportional to the Genius and Capacity of the Inhabitants.“

It is unquestionably tempting to reason, “Since there is only one nature, ~ only one science of nature is possible.” Yet, on closer scrutiny, this reasoning becomes highly problematic. Science is always the result of inquiry into nature, and this is inevitably a matter of a transaction or inter-

action in which nature is only one party and the inquiring beings another. We must expect alien beings to question nature in ways very different from our own. On the basis of an interactionist model, there is no reason

to think that the sciences of different civilizations will exhibit anything more than the roughest sorts of family resemblance. It is surely naive to think that because one single object is in question, its description must be uniform. For minds with different concerns and interests and with different experiential backgrounds are bound to deal with the selfsame items in ways that yield wholly disjoint and disparate results because different features of the thing are being addressed. Even where things are the same, their interpreted significance is going to be altogether different. One must reckon with the fact that while there indeed is only one world, nevertheless very different thought worlds can be at issue in the elaboration of a science. Perhaps it seems plausible to argue thus: “Common problems constrain common solutions. Intelligent alien civilizations have in common

with us the problem of cognitive accommodation to a shared world. Natural science as we know it is our solution of this problem. Therefore, it is likely to be theirs as well.” But this tempting argument founders on its second premise. The problem-situation confronted by extraterrestrials is not common with ours. Their situation must be presumed substantially different exactly because they live in a significantly different environment and come equipped with significantly different resources—physical and intellectual alike. The “common problems, common solutions” line does not work: to presuppose a common problem is already to beg the question.

Our alien scientific colleagues also scan nature for regularities, perforce using—to begin with, at any rate—the sensors provided to them by their evolutionary heritage. They note, record, and transmit those regu-

larities that they find to be useful or interesting and then develop their 1 Christiaan. Huygens, Cosmotheoros: The Celestial Worlds Discovered: New Conjectures concerning the Planetary Worlds, Their Inhabitants and Productions (London, 1698), pp. 41-43. :

120

.

THE STATUS OF SCIENCE

inquiries by theoretical triangulation from this basis. This is clearly going to make for a course of development that closely gears their science to their particular situation—their biological endowment (their “sensors”) and their cultural heritage (“what is pragmatically useful”). Where these key parameters differ, we must expect that the course of scientific development will differ as well. Admittedly, there is only one universe, and its laws and materials are, as far as we can tell, the same everywhere. We share this common uni-

verse with all life-forms. However radically we may differ in other respects (in particular, those relating to environment, to natural endowments, and to style or civilization), we have a common background of cosmic evolution and a common heritage of fundamental physical laws. And so, if intelligent aliens investigate nature at all, they will investigate the same nature we ourselves do and are subject to the same sorts of physical processes. All this can be agreed. But the fact remains that the available corpus of scientific information—ours or anyone’s—is an ideational construction. And the sameness of the object of contemplation does nothing to guarantee the sameness of ideas about it. It is all too familiar a fact that even where only human observers are at issue, very different constructions are often placed upon “the same” occurrences. As is clearly illustrated by the rival interpretations of different psychological schools, to say nothing of the court testimony of rival “experts,” there need be little uniformity in the conceptions held about one selfsame object from different perspectives of consideration. The fact that all intelligent beings inhabit the same world does not countervail the no-less-momentous fact that we inhabit very different ecological niches within it, engendering very different sorts of modus operandi. / The development of a science—a specific codification of claims vegarding the laws of nature—always requires as input some inquirer-supplied element of determination. The result of such an interaction depends cru-

cially on the contributionfrom both sides—from nature and from the intelligences that interact with it. With variable technologies of investigation and methods of assessment, the fruits of scientific inquiries depend on how inquirers proceed. A kind of chemistry is at work in

which nature provides only one input and the inquirers themselves provide another, with the outcome arising in such a way that we cannot dis-

entangle these respective contributions. The theses and theories of our science are necessarily based on “the available data” and are accordingly bound to reflect the character of our interactions with nature, through

which alone data can be acquired. This interaction is a two-sided process to which each party makes an essential contribution—and where the character of these respective contributions cannot be altogether distinguished and clearly separated. Here, the potential plurality of modes of

OUR SCIENCE AS OUR SCIENCE

2

121

judgment means that there is no single definitive way of knowing the world. There is no good reason for believing that intelligent creatures will think alike in a common world, any more than that they will act alike—that is, there is no reason why cognitive adaptation should be any more uniform than behavioral adaptation. Thought, after all, is simply a kind of action; and as the action of a creature reflects its biological heritage, so too does its mode of thought. No one who has observed how very differently the declarations of a single text (the Bible, say, or the dialogues of Plato) have been interpreted and understood over the centuries—even by people of a common cultural -heritage—can be unalloyedly hopeful that the study of a common object by different civilizations must lead to a uniform result. And this textual analogy is even overly generous to the case of uniformity. After all, the scientific study of nature is not a matter of decoding a preexisting text. There just is no one fixed basic text—the changeless “book of nature writ large”’—which different civilizations can decipher in different de-

grees. Like other books, it is to some extent a mirror: what looks out depends on who looks in. Things cannot of themselves dictate the significance that an active intelligence can attach to them. Human organisms are very similar, but there is not much similarity between the medicine of the ancient Hindus and that of the ancient Greeks. Given different interests and different procedures, it is not only inevitable but also ap-

propriate that different inquirers will arrive at different results. It is a striking phenomenon that throughout the earlier stages of humankind’s intellectual history, different human civilizations developed

their natural sciences (such as they were) in substantially different ways. The shift to an extraterrestrial setting is bound to amplify this diversity. The science of an alien civilization may be far more remote from ours than the language of our cousin the dolphin is remote from our language. We must face, however reluctantly, the fact that on a cosmic scale the “hard” physical sciences have something of the same cultural relativity that one encounters with the “softer” social sciences on a terrestrial basis. The ultimate reason why we cannot expect alien intelligences to be engaged in doing our sort of science is that the possible sorts of natural science are almost endlessly diverse. Sciences—understood as such in the functional-equivalency terms laid down above-—-are bound to vary with the cognitive instruments afforded by the physical constitution and mental equipment of their developers and with the cognitive focus of interest of their cultural perspective and conceptual framework. Our particular cognitive project is simply the intellectual product characteristic of one particular cognizing life-form. Even if one grants the (surely problematic) thesis of an ultimate uniformity of result in human inquiry, there _is no good reason to project this uniformity across species—to see prod-

122

.

THE STATUS OF SCIENCE

uct as independent of process in that inquiry by different species must ultimately arrive at the same results. Our human scientific truths are not necessarily those of other sorts of creatures. Natural science—broadly construed as inquiry into the ways of nature—is something that is in principle endlessly plastic. Its development will trace out a historical course closely geared to the specific capacities, interest, environment, and opportunities of the creatures that develop it. We are deeply mistaken if we think of it as a process that must follow a route generally parallel to ours and issue in a roughly comparable product. It would be grossly unimaginative to think that either the journey or the destination must be the same—or even substantially similar. It is experience that determines what sort of scientific findings can be arrived at, and different civilizations are bound to have different cases of experience. The science of the species, like the behavior of the individual, is

captive to a biological and cultural heritage. Factors such as capacities, requirements, interests, and course of development are bound to affect the shape and substance of the science and technology of any particular space-time region. Unless we narrow our intellectual horizons in a parochially anthropomorphic way, we must be prepared to recognize the great likelihood that the science and technology of a remote civilization would be something very different from science and technology as we know it. Our human sort of natural science may well be unique in its kind and adjusted to and coordinated with a being of our physical constitution, inserted into the orbit of the world’s processes and history in our

sort of way.

A Situational Realism

.

With respect to biological evolution it seems perfectly sensible to reason as follows: “What can we say about the forms of life evolving on these

other worlds? . . . [It] is clear that subsequent evolution by natural selection would lead to an immense variety of organisms; compared to them, all organisms on Earth, from molds to men, are very close relations.”!2 Exactly the same situation will surely also obtain with respect to

cognitive evolution. . These considerations point to a clear lesson. Mutually remote civilizations composed of different sorts of creatures must be expected to create diverse sciences. Each inquiring civilization must be expected to produce its own, perhaps ever-changing, cognitive products—allmore or less ad12 1. §. ShkKdovskii and Carl Sagan, Intelligent Life in the Universe (San Francisco, 1966), p. 350.

OUR SCIENCE AS OUR SCIENCE

123

equate in their own ways, but with little if any actual overlap in conceptual content. Though inhabiting the same physical universe with us and subject to the same sorts of fundamental regularities, they will unquestionably produce as cognitive artifacts different depictions of nature, reflecting their different modes of emplacement within it. The one-world, one-science argument is ultimately untenable. And so we are well advised to accept that the science as we know it is . merely our science—the projection on the screen of mind of a world picture devised in characteristically human terms of reference. The world is real enough, independently of our ideas about it, but the-world-as-weview-it on the basis of our inquiries—the only world with which we have cognitive as opposed to causal interactions—is a construction of ours correlative to our (characteristically human) place in the cosmic scheme. Being the product of our experience of nature, our empirical science is bound to reflect, at least in part, the peculiar character of our evolutionary heritage. Immanuel Kant’s insight holds: there is good reason to think natural science, in its mathematical structure as we know it, to be something that is not universally valid for all rational intelligences as such, but substantially a product of human artifice that is in crucial respects correlative with our specifically human intelligence. We have little alternative to supposing that our natural science is limited precisely by its being ours. The inevitability of an empiricism that accepts the fundamentality of experience for arriving at scientific knowledge of the world means that our science is bound to be relativized ultimately to the kinds of experiences we can have, being destined to reflect our nature—to be conditioned and delimited by the sorts of creatures we are with respect to our mode of sensory involvement in the world’s scheme of things. The “scientific truth” that we discover about the world is our truth—not in the sense that we make it up in an arbitrary way, but rather in the sense that, all-natural science being an empirical venture, science as we have it is bound to be conditioned by our human mode of emplacement within nature. ‘ There simply is no one unique standard body of scientific fact that different civilizations secure at different rates. There just is no unique course of scientific-technological development that different civilizations travel in common with mere differences in speed or in staying power (notwithstanding the penchant of some astrophysicists for the neat plotting of numerical “degrees of development” against time in the evolution of planetary civilizations).1* Even as cosmic evolution involves a spatial red shift that carries different star systems ever farther from each other, 3 John A. Ball, “Extraterrestrial Intelligence: Where Is Everybody?” American Scientist

68 (1980): 558.

=

124

.

THE STATUS OF SCIENCE

so cognitive evolution may well involve an intellectual red shift that carries different civilizations into thought worlds ever more remote from each other. Our science is our science—a human creation, an artifact like any other, that we have made with our tools and have formed with a view to our aims and interests. Even if we project this science of ours into the Peircean limit, it will still remain just that—the fully developed version of our science, a product that is bound to differ from the sort of science

that a creature of different, perhaps broader experience might very well devise. Our science correlates with our range of thought and experience. Like other created intellectual artifacts, it is destined to reflect the ca-

pacities and interests of its creator. The point is simply—and decisively—that what is relevant and important for us depends on how we go about experiencing the world, how we interact with our physical environment. The sort of scientific world picture that we obtain is clearly bound to reflect this circumstance. And of course the product of our inquiries—the world-picture at which we arrive by their views—cannot but be influenced by the character of those controlling values (goals, etc.) under whose controlling impetus we develop it. A science designed to fit the experiential and existential conditions of creatures whose physical investment in nature differs from ours (an advanced species of dolphin, say) would be very different from ours. After all, science is ineliminably empirical and experience bound. And the situational contextualization of different modes of experience is bound to pave the way to different sorts of science. To say all this is not, of course, to say that our scientific world-picture is a mere fiction—that it is something we simply make up as we go dlong. A restrictive reality-principleis clearly at workin any genuinely scientific endeavor. For any science (be it ours or the superdolphins) is, as such, formed under the aegis of controlling purposes such as prediction and control that involve issues over whose actual outcome we have no influence or power whatsoever. In these matters the ultimate decision lies

entirely with nature. But the fact remains that if we gave a different priority to the values operative in our science—if, for example, we prized explanatory analogy/uniformity a great deal more than simplicity/economy—then we would unquestionably obtain a very different scientific picture of things. On such a relational view, knowledge ofreality i is always (in some crucial respect) cast in terms of reference that reflect its possessor’s characteristic cognitive resources. There is, no doubt, a mind-independent reality, but cognitive access to it is always mind-conditioned. All that can ever be known of reality is mediated through conceptions shaped by the

OUR SCIENCE AS OUR SCIENCE

125

ways in which this reality affects us, given the sensory and cognitive endowments with which our evolutionary heritage has equipped us. This view of the matter is compatible with a scientific realism of sorts, but a realism that is contextualistic in its multifaceted recognition that any science will reflect its devisers’ particular slant on reality through their particular mode of interaction with nature. It sees our scientific picture of the world’s processes as an intellectual artifact that reflects the characteristic mode of perception and conception of the inquiring creatures that are its producers. The fact of its being an ideational construct of ours means that our science must, to be seen justly, be seen in an

idealistic perspective. All the information we ever have at our disposal is a matter of being putative facts—facts as we suppose'them to be—and, as such, artifacts. Our only cognitive access to reality is via the construction of a world picture, a model in whose construction our own intellectual resources play a crucially conditioning role. This view is of course deeply idealistic in its kinship with the classical mode of idealism that sees knowable reality as shaped unavoidably by the conditions of its knowability. The resultant view of science is thus complex and two-sided. On the negative side, we must recognize that any sort of natural science is something contextual, coordisiated to the cognitive situation of its producers. On the positive side, we have the patently observable fact that science is successfully applicable in practice and substantially informative in theory. The negative side speaks for realism. The very shortcomings of science reflect the subordinate status of the works of mind, since science’s

being imperfect, incomplete, and ever improvable means that thought is here in a subordinate position (“thought proposes, reality disposes”). But the positive side speaks for idealism. For it is, clearly, only by means of thought constructions that we can get a cognitive grip on the real: our only access to reality is via a thought-constructed estimate of it. This overall perspective indicates that, with regard to the long-standing quarrel between realism and idealism, the most sensible stance is that of a mixed,

halfway-house position that gives both realism and idealism their respective due—a theme that will recur in the concluding chapters of the book. 4 This chapter draws primarily on the author's Limits of Science (Berkeley and Los Angeles, 1986). Relevant themes are also developed in Scientific Progress (Oxford, 1978), Empirical Inquiry (Totowa, N.J., 1982), Scientific Realism (Dordrecht, 1987), and A Useful Inheritance (Savage,Md. , 1990).

Part Ti THE PURSUIT OF TRUTH

Nine Induction as Truth Estimation through Cognitive Systematization | The Teleology of Induction: The Search for Best-Available Answers to Questions At the basis of the cognitive enterprise lies the fact of human curiosity, rooted in the need-to-know of a weak and vulnerable creature emplaced in a difficult and often hostile environment in which it must make its evolutionary way by its wits. For we must act—our very survival depends upon it—and a rational animal must align its actions with its beliefs. We have a very real and material stake in securing viable answers to our questions as to how things stand in the world we live in. Our questions stand at the very center of the cognitive stage, providing the impetus for developing our knowledge (or putative knowledge) of the world. Questions arise most pressingly where the information in hand does not suffice—when they are not answerable in terms of what has already been established. This creates the requirement for an ampliative methodology of inquiry—“ampliative” in C. S. Peirce’s sense of going beyond the evidence in hand and thus transcending the secure reach of information extracted from the accreted experience already at our disposal. And so it becomes necessary to devise a means for obtaining the best available, the

“rationally optimal” answers to those questions we have about how things stand in the world whose definitive resolution simply transcends the information we currently possess. It is just here that induction enters upon the scene. The definitive task of the inductive enterprise is to provide an ampliative methodology for acquiring information inthe domain ofmatters of base larger contentions upon lesser data. Inductionis at bottom a mechanism for enlarging the stock of (putative) truths that we already accept about the world—a resource akin in this regard to observation and inference.

.

Induction is an instrument for question resolution in the face of imper“fect information. It isa tool for use by finite intelligences, capable of ielding not the bes i e rarified sense of this te but the best available answer, the best we can manage to secure in the - @xisting conditions in which we do and must conduct our epistemic la-

130

.

THE PURSUIT OF TRUTH

bors. Of necessity, it is restricted by the limits of our cognitive range: it obviously cannot deal with issues that might lie outside our conceptual horizons (as quantum electrodynamics lay beyond those of the physicists

of Newton's day). The available answers at issue have to be found within some limited family of alternative possibilities within our intellectual grasp. Induction is not an occult matter of a cognitive alchemy that transmutes ignorance into knowledge; it is a mundane and realistic instrument for doing the best we can to answer our questions in the epistemic circumstances in which we do (or with reasonable further exertion can) actually find ourselves. Dictionaries sometimes define induction as “inference to a general conclusion from particular cases.” But such inferences—for example, from “spaniels eat meat,” “schnauzers eat meat,” “corgis eat meat,” and

so forth, to “all dogs eat meat”—illustrate only one particular kind of inductive reasoning. Nor will it do to add merely those inferences to a particular conclusion that move from effects to cduses—from the smoke to the fire, or from the bark to the dog. For this does not go far enough. Inference from sample topopulation, from personal observation to objective fact, from part to whole (from the jaws to the entire alligator), from style to authorship, from clue to culprit, from symptom to disease, and so forth, are all also modes of inductive inference. The crucial thing about

induction is its movement beyond the evidence—from informatively lesser data to relatively larger conclusions. The crux lies in the erotetic (i-e., question-answer oriented) nature of induction as a matter of providing answers to questions that transcend the information in hand. Consider a question of the form “Are the F's also Gs?” The situation here is akin to that of a multiple-choice examination, where one can respond: i 1. 2. 3. 4,

Yes, all of them are. Never-——none of them are. No, some are and some aren’t. Don’t know; can’t say. —

This pretty well exhausts the range of alternatives. Now when in fact all . of the observed Fs (over a fairly wide range) are indeed Gs, our path seems relatively clear. Alternative 4 is not an answer but a mere evasion of the question that is usually appropriate only as a response of last resort, to be given (with respect to pressing questions, at any rate) only when all else has failed us. Alternative 2 is ex hypothesi ruled out in the circumstances. The choice is between Alternatives 1 and 3. And we naturally opt for the former. The governing consideration here is the matter of plausibility—specifically that of uniformity. For Alternative 1 alone extends the data in the most natural way, seeing that this response alone

INDUCTION AS TRUTH ESTIMATION

131

aligns the tenor of our general answer with the specific information we actually have in hand. It is, accordingly, this resolution that affords the

“inductively appropriate” answer in the postulated circumstances. The point of inductive reasoning is not to infer secure conclusions from given premises but fo take what we need to get answers to our questions,

- >‘jump to conclusions, albeit in the safest way relative to all we. know—to secure the information we want and need. Induction aims to

answer Gur questions thiaughthe optimal use of information—to make

the best manageable balance between the unknown leap required to get ans 0 Our questions and the safety net of an opum age to whatweCin alreadysubstantiate.— “TO be sure, in saying that induction involves the search for maximally plausible answers to questions, we are not saying that it does not (like all question-answeringdevices) enjoy the privilege of maintaining silence, of responding“can’t sa

in certain circumstances.

Quite the reverse. If we ask, “Which side of this (fair) die will come up?” this is exactly what induction would reply: we just cannot effect a rationally defensible resolution here. No inductively appropriate answer is available. (And this situation would still obtain even if the die were loaded in favor of one side.) Yet this sort of negativity is something the inductive enterprise seeks to minimize. But why not always opt for safety in answering our questions, systematically selecting the noncommittal pseudoalternative “none of the above’? Why not decline all risk of error and simply follow the path of skepticism? The answer is simple: Nothing ventured, nothing gained. The object of the cognitive enterprise is clearly to secure truth (and not simply to avert error!). This, after all, is a definitive task of inquiry, the venture of cognitive gap-filling—of securing information insofar as possible. Nevertheless, the “best available answer” at issue with induction is meant in a rather strong sense. Its acceptability claims must not only be stronger than those of the alternatives, but this difference in comparative strength must be substantial—in particular, more substantial than is reflected in any mere difference in probability, since the most probable cannot eo ipso be reasonably claimed as true. The quest for information hinges on the distinction between_good_ and bad answers, between an-

swers that have little or nothing to be said for them and answers for whose acceptance thereis adequate systematic warrant, everything taken into account. An inductively appropriate answer.mustqualify as our best estimate of the true answer in a ative sense that encompasses being a good answer pure and simple. Wewant not just an answer of some sort but an appropriate answer—an_ answer towhose tenability w

are willing to commit ourselves. Inductionis not a matter of mere guess

nn 1

132

.

THE PURSUIT OF TRUTH

work but of ible estimation, in a demanding sense of the term: it is not justsn estimate of é trac answer that we want but an estimate that is sensible and defensible: tenable, in short. The provision of reasonable warrant for rational assurance is the object of the enterprise.

Induction as Enthymematic Deduction: The Proliferation- | Reduction Model of Induction as Plausible Reasoning The term “induction” is derived from the Latin rendering of Aristotle’s epagoégé—the process for moving to a generalization from its specific instances.! Gradually extended over a wider and wider range, it has ultimately come to embrace all nondemonstrative argumentation in which the premises are supposed to build up a case of good supportive reasons “for the conclusion, while yet falling short ats ona lly pose) strative force of logical deduction (for it always remains logically possibl with inductive arguments to admit the premises and deny the conclusion).”oo

““This informational shortfall reflects a crucial facet of the matter. It means that_an inductive argument can always belooked upon as an aspiring but‘failed deductive inference—an enthymeme, or argumenti formulation some crucially necessary premise is lacking, so that a‘larger conclusion isbased on lesser premises. This is illustrated by such cases as the following (where the enthymematically tacit premises needed to make the argument deductively cogent have been bracketed): Spaniels eat meat. Schnauzers eat meat. Corgis eat meat.

y

{

/

i

(The remaining species of dogs (terriers, Dobermans, etc.) all eat meat. )

Therefore: All dogs eat meat.

~,

There is smoke yonder. (Where(ever) there’s smoke, there’s fire.) Therefore: There is fire yonder. Two-thirds of the items in the sample are defective. (The sample is representative of the whole.) Therefore: Two-thirds of the items in the whole population are defective. 1 See W. D. Ross, Aristotle’s Prior and Posterior Analytics (Oxford, 1949), pp. 47-51.

2 See the informative article “Induction” by Max Black in The Encyclopedia of Philosophy, ed. P. Edwards (1967).

INDUCTION AS TRUTH ESTIMATION

133

This enthymematic approach takes the step of transforming an inductive argument into a cogent deductive one, by projecting a missing premise that is required to make the argument fully (i.e., deductively) cogent, but suppressed in its actual formulation. We must accept the burden of David Hume’s critique of induction to this extent at any rate, that we acknowledge inductive reasonings to be incomplete and thus-+in requir_ing the addition of further, otherwise unexplicit premises—to fail as they stand to present a deductively cogent process of reasoning.®

The enthymematic view of induction sees it in the following light: We begin with a certain question Q (What do dogs eat? What does yonder smoke portend?) Within the setting afforded by the body K of the Qrelevant information that is already in hand, we then engage in a conjectural process of alternative proliferation to determine the alternative answers A;, Ao, ..., A, that (in the context of the relevant information)

are “worth bothering about” in that they exhaust or span the whole spectrum of the feasible alternatives. In each case, we then identify the Kcontextually most plausible enthymematic (because information-extending) premise E, that can underwrite a deductively valid argument leading from K to A;. We thus arrive at a series of arguments leading to the various conclusions A; that afford diverse answers to our inductive question

Q.

K

K

K

(Ey)

(Ez)

{E,)

Therefore: A,

Therefore: Ag

Therefore: A,

In each case, E, is the K-relatively most plausible supplementation to K that is capable of underwriting the deductive step from K to the particular A, at issue. See figure 9.1. . The inductive task is to determine which one of such alternative (and supposedly exhaustive) answers to the question “What does yonder smoke portend?” is to qualify as the most promising in representing the particular addendum F, that is, relative to the given data of K, the plausibilistically optimal alternative for Q-resolution that is at our disposal— where the plausibility at issue turns on the matter of “best fit” with respect to the cognitive commitments of K. The inductively appropriate answer to the question at issue corresponds to the outcome of this search 3 It might be objected that in physics and other branches of empirical science, one standardly solves one’s problems by wholly deductive reasoning. But while this is indeed true of textbook exercises, it'is clear that, in applying such calculations to reach contentions about the real world, information-in-hand transcendence always enters in. (Humean wor_ries regarding the continued operation of historically grounded laws are only one instance of this.)

134

THE PURSUIT OF TRUTH

Figure 9.1. Alternative Explanations

(1)

(2)

(3)

There is smoke there

There is smoke there

There is smoke there (and suitable back-

(and suitable back-

(and suitable back-

ground considera-

ground considera-

tions).

tions).

ground considerations).

(This smoke is being

(This smoke is being caused by a smoke

released from a stor-

caused by afire.)

flare there.)

age container.)

Therefore: There is fire there.

Therefore: There is a smoke flare there.

smoke-discharging

(This smoke is being

Therefore: There is a release there. _

for the enthymematic premise that is plausibilistically optimal (in the context of K)}—namely, the premise E, that (relative to the information in hand) represents the best answer providing enthymematic supplementation of the background information.‘ (In the case of the previous example, this will, of course, be the enthymematic premise of argument [1].) On this enthymematic analysis, inductive argumentation involves a characteristic two-step process.

=~

7

1. Possibility elaboration: the conjectural proliferation of the spectrum of alternative possible answers, accompanied by a process of finding the appropriaté enthymemes for each such answer by determining the best waysof clos“ 4 Jt deserves note that the “missing” enthymematic premise of an inductive argument need not be minimal in the sense of adding nothing over and above what is indispensably needed to move from premises to conclusion. We can, for example, make the inductive move of sampling generalization from “These (randomly selected) apples from the barrel are all sour” to “All apples in the barrel are sour” via the perfectly workable enthymeme: __ These apples from the barrel are sour (and the indications are that all the apples in the barrel are of the same kind).

(All the apples in the barrel are like these in taste-relevant respects.) Therefore: All apples in the barrel are sour. The enthymematic premise here is clearly stronger than what is minimally required for the move from premise to conclusion. But since it is no less plausible than its more restricted cousins, it serves just as well for present purposes. (See the author's Plausible Reasoning [Assen, 1976].) In general, it should be said that contextual plausibility—and neither con-

tent paucity nor probability as such—determines the inductive appropriateness of enthymematic premises. (To say this is not, of course, to deny that these other factors can play a role in the assessment of plausibility.)

SS 5

INDUCTION AS TRUTH ESTIMATION

135

ing the “epistemic gap” that separates those answers from the given data of the

problem. (This survey need not include all theoretically available alternatives but can be confined to those that qualify as “real” possibilities.) 2. Possibility reduction: the reduction of these alternatives through elimination of some of them. This is to be done by assessing the relative plausibility of the materials needed to close the enthymematic gap encountered en route to the solution in question. That is, we use an analysis of comparative plausi-

bilities as a reductive device for seeking out the optimal alternative(s) within this manifold of possibilities—that which is, all considered, the most plausible.§

On this enthymematic perspective induction comes to be seen as not so much a process of inference as one of estimation—its conclusions are not so much extracted from data as suggested by them. And clearly, we want to accomplish this gap-filling step.in the least risky, the minimally problematic way, as determined by plausibilistic best-fit considerations. Induction leaps to its conclusion instead of literally deriving it from them

through some extractive process. Long ago, William Whewell put the point nicely. “Deduction,” he wrote, “descends steadily and methodically, step by step: Induction mounts by a leap which is out of the reach of method [or, at any rate, mechanical routine]. She bounds to the top of the stairs at once.” We cannot pass, by any sort of literal inference or cognitive calculation, from the premises of an inductive argument to its conclusion because (ex hypothesi) this would be a deductive non sequitur—the conclusion (in the very nature of the case) asserts something regarding which its premises are altogether silent.” Clearly the paradigm mode of inference—of actually deriving a conclusion from the premises— is actual deduction, and this paradigm does not fit induction smoothly. As one recent writer has felicitously put it, our inductive conclusions are “not derived from observed facts, but invented in order to account for them.”® . Seen in this light, induction is not a method for extracting conclusions from given premises on the order of deduction but rather a vehicle for prejudgment, for injecting ab extra into a course of reasoning some more 5 This perspective indicates that it is desirable tc distinguish between an inductive argument (which is simply an enthymematic argument whose factual conclusion outruns the information provided by its premises) and inductive argumentation, considered as the general procedure of inductive reasoning, a complex process comprising very different sorts of

reasonings—including not only deductive inference but also plausibilistic argumentation and enthymematic conjecture. , § William Whewell, Novum Organon Renovatum (London, 1858), p. 114.

7 The force of Dickinson Miller’s principle must be acknowledged: “There are no intermediate degrees between following from premises and not following from them. There is no such thing as half-following or quarter-following” (Dickinson S. Miller, “Professor Donald Williams vs. Hume,” Journal of Philosophy 44 [1947]: 684). 8 Carl G. Hempel, Philosophy of Natural Science (Englewood Cliffs, N.J., 1966), p. 15.

136

.

THE PURSUIT OF TRUTH

or less plausible assumptions or presumptions. This enthymematic view of induction sees it as a matter of introducing substantive assumptions or suppositions into an informationally impoverished situation in which more data are needed for any definitive problem resolution. It is a matter of drawing upon our experience with analogous situations to inject from without into a given cognitive problem situation a plausible bit of conjecture or supposition that will suffice to provide an answer to the question at hand. Induction emerges as a method of truth estimation in questionresolving situations that must be handled enthymematically in terms of plausibilistic considerations. On such an approach, inductive reasoning is a process of leaping to conclusions by conjectural informational gap-filling, by supplementing the inferential mechanisms of ordinary (deductive) logic with substantive resources of plausibility and presumption.

Plausibility and Systematicity: Induction as Systematization Considerations of plausibility play a central role in the truth-estimation model of inductive reasoning, providing the eliminative process through

which we arrive at the best estimates of correct answers to our information-transcending questions. This raises the question of where these plausibilities come from. The answer is straightforward. Throughout inductive contexts, plausibility is a matter of cognitive systematicity: the standards of inductive plausibility inhere in the parameters of cognitive systematization. We must accordingly undertake a brief examination of the ideas at issue in the traditional concept of a system as anorganic unity of mutually coordinated and collaborative units. The principal factors at issue here—the parameters of systematicity, as one may dub them—include preeminently the following items: 1. completeness: comprehensiveness, avoidance of gaps or missing compo-

nents, inclusiveness, unity and integrity as a genuine whole that embraces and integrates all its needed parts; 2. cohesiveness: connectedness, interrelationship, interlinkage, coherence (in its unification-oriented sense), a conjoining of the component parts, rules,

laws, linking principles; if some components are changed or modified, then others will react to this alteration; 3. consonance: consistency and compatibility, coherence (in its consistencyoriented sense), absence of internal discord or dissonance; harmonious mutual

collaboration or coordination of components, “having all the pieces fall into proper place”; 4, functional regularity: lawfulness, orderliness of operation, uniformity, pattern conformity, normality (conformity to the “usual course of things”);

INDUCTION AS TRUTH ESTIMATION

137

5. functional simplicity and economy: elegance, structural economy, tidiness in the collaboration or coordination of components, harmony and balance, symmetry;?

¥

6. functional efficacy: efficiency, effectiveness, adequacy to the common task, versatility and range and power of operating principles.

These are some of the characterizing parameters of systematization. As their nature indicates, systematization is not just a matter of constructing a system, however jerry-built it may prove to be, but of constructing it under the aegis of certain standard criteria. A system, properly speaking, must exhibit all of these various parameters. (Think, for example, of the control system for a manufacturing process or the life-support system of a space capsule.) But a system need not exhibit all these facets of systematicity to an equal degree—tlet alone perfectly. They reflect matters of degree, and systems can certainly vary in the extent to which they embody these characteristics and in the manner of their embodiment as well, since the rather schematic nature of these parameters leaves a good deal of context-specific detail to be filled in. But no system can be found or constructed that lacks a substantial combination of these desiderata,

simply because they constitute the guiding standards that govern the process of systematization and determine the claims of its products to be characterized as a system. If a system (an economic or social system, for example) were to lack these characteristics in substantial measure—if its coherence, harmony of functioning, or end-realizing effectiveness were substantially diminished—then its very existence as a system would be compromised. !° To be sure, our present concern is not with systematicity per se but specifically with cognitive systematicity as based on the conception that our information about the world is to constitute part of a system of knowledge. Accordingly, the parameters of systematicity must, in this present context, be construed in a specifically cognitive sense. And the key fact for our purposes is that, thus construed, they afford our criteria of inductive plausibility—of the acceptability qualifications of our answers to information-transcending questions. This reliance upon the parameters of systematicity as inductive norms—as standards of plausibility and presumption for induction gapfilling—means that systematicity becomes the arbiter of acceptability-astrue, to use the well-chosen word of F. H. Bradley. For we seek to ® On the range of considerations at issue here, cf. Elliot Sober, Simplicity (Oxford, 1975). 40 For a further development of these issues and a fuller exposition of the parameters of systematicity, see the author’s Cognitive Systematization (Oxford, 1979).

4 See F. H. Bradley, “On Truth and Coherence,” in Essays on Truth and Reality (Ox-

138

:

THE PURSUIT OF TRUTH

articulate our knowledge by shaping its structure in the most systematic way, the way that endows the resulting manifold of cognitive commitments with the greatest attainable degree of systematic order. But systematicity is not merely ornamental; it is not just a matter of endowing the exposition of our knowledge with certain aesthetic features. Instead of merely representing a facet of the organization of our (otherwise preexistent) knowledge, systematicity is to provide an operative force in the very constituting of this knowledge. While inquiry is a process of enlarging the information at our disposal, of yielding new items to be added to the stock of our acceptances, such question answering is a matter of geiting not just an answer but a tenable answer—one that merits acceptance within the body of already-established information that provides the materials for our further systematizations. And systematicity itself furnishes us with the operative norms here, so that inductive acceptability becomes a matter of systematic fit—and indeed a matter of the tightness of that fit. In sum, we use system not just as organizer of what wé accept but as a Bradleyan arbiter of acceptability—a standard of what we are to accept, or at any rate endorse pro tem, until such time as discordant

counterindications come to view.

The connection between induction and systematicity is thus mediated through two governing ideas: optimality and plausibility: (1) induction seeks to present the bestavailable (i.e., contextually optimal) answers to our questions—specifically, our information-in-hand transcending questions about the world; (2) the idea of “best available” in relation to possible answers is to be understood on the basis of plausibility considerations, which themselves turn on considerations of systematicity, so that

the best available solution is that which is most smoothly attuned to the data and involves a minimum of implausibilities. The definitive features of the present analysis of induction are accordingly two. First, inductive reasoning is seen as part and parcel ofthe enterprise ofcognitive systematization—the systematic-structuring and rounding out of our. knowledge or presumptive knowledge of matters of fact. Second, in this process of systematic attunement, the parameters of systematicity (uniformity, simplicity, and the rest) are themselves taken to provide the requisite yardsticks of the plausibility assessment that guides deliberations of acceptability.12 And the second point is bound up with the first: it is beford, 1919), pp. 202-18, and esp. p. 210, where Bradley endorses “the claim of system as an arbiter of fact.” 22 In an oft-cited passage, John Stuart Mill wrote: “Why is a single instance, in some cases, sufficient for a complete induction, while in others, myriads of concurring instances,

without a single exception known or presumed, go such a little way towards establishing a universal proposition. Whoever answers this questions knows more of the philosophy of logic than the wisest of the ancients and has solved the problem of induction” (A System of

INDUCTION AS TRUTH ESTIMATION

139

cause induction implements the enterprise of cognitive systematization that the parameters of systematicity are to be taken as standards of plausibility in inductive contexts.

Induction as Truth Estimation on the Basis of Systematization

“with Experience” Induction accordingly emerges as a tool of cognitive system-building—a device for the plausibilistically optimal meshing of question-resolving conjecture with the available data, where the parameters of cognitive systematization themselves provide thé requisite standards of plausibility. Induction, thus regarded, is a mechanism of question resolution through the optimally systematic blending of conjecture with available information-in-hand. It is a process of dovetailing our experientially guided conjectures into their wider experiential context on the basis of plausibilistic best-fit considerations—a process in which support and systematicity are inextricably correlative. ‘This line of thought depicts induction as a particular sort of cognitive systematization with the data. For systematic “best fit” is always a matter of fitting with something, and this is where the matter of the data—of information already in hand to provide a background—plays its indispensable part.!* These data may stem from many sources: observations, records, or even (in such contexts as hypothetical exercises) mere assumptions. They are, in science, afforded by the body of our “funded experience” (as William James called it). And the drive to system embodies an imperative to broaden the range of our experience, to extend and expand the data base from which our theoretical triangulations proceed. In the course of this process, it may well eventuate that our existing systematizations—however adequate they may seem at the time—are untenable and must be overthrown in the interest of constructing ampler and tighter systems. Cognitive systematization is emphatically not an in-

delibly conservative process that looks only to what fits smoothly into heretofore established patterns, but one where the established patterns Logic, 8th ed. [London, 1895], bk. 3, chap. 3, sec. 2). But the matter is not all that arcane.

It is precisely the matter of systemic best-fit that distinguishes between some instanceindicated generalizations and others in inductive argumentation—the mere number of in-

stances is in itself secondary and insignificant. The decisive factors are fit, enmeshment, and the Whewellian “consilience” to which Mill never gave due credit. 3 Note, however, that our data are invariably fallible—that our sources of information

afford misinformation as well; the process of fitting things with the data also makes for a smoothing out of the data themselves. Cf. the author’s treatment of data in The Coherence _ Theory of Truth (Oxford, 1973).

140

.

THE PURSUIT OF TRUTH

are themselves ever vulnerable and liable to be upset in the interests of devising a more comprehensive systematic framework. Induction represents a cognitively conscientious effort at closing the gap between available information and needed answers in such a way that (everything considered) we can responsibly regard it as epistemically well advised to accept the indicated results. This quest for a cognitively optimal answer makes induction a matter of informational systematization on the basis of considerations of best fit within the setting of our cognitive commitments taken overall. On this approach, then, the assurance of inductively authorized contentions turns exactly on this issue of tightness of fit: of consilience, mu-

tual interconnection, and systemic enmeshment. Systematicity becomes our test of truth, the guiding standard of truth estimation. Under its aegis our “picture of the real” is taken to emerge as an intellectual product. achieved under the control of the idea that systematicity is a regulative principle for our theorizing. To be sure, our epistemological position moves from systematicity not to correctness itself but rather to the rational warrant of claims to correctness. The operative transition is not from “systematic” to “correct” but rather from “systematic” to “rationally claimed to be correct,” seeing that our inductive concern is with the estimation of truth. The role of systematicity is, in the first instance,epistemic (and only derivatively ontological). The“best availableanswer" at issue is so onlyin the sense of affording us the best available estimate of the truth. The present approach to induction differs radically from the widespread tendency to think of induction as a matter of characteristic inferential rules of ampliative inference for drawing larger conclusions from informatively lesser premises. It sees induction not as a characteristic body of rules of drawing conclusions but as an estimation technique, a procedure for obtaining answers to our factual questions through optimal exploitation of the information at our disposal. Thus regarded, induction is a conjectural (question-answering) rather than an inferential (conclusion-deriving) procedure. Instead of deriving “All Xs ‘are Ys” from a-set of premises of the form “X; is Y (i = 1, 2, 3, etc.)” by some sort of (obviously invalid) process of derivation, we take the view that the former is

the best answer to the question “What is the Y-status of the Xs?” available in the epistemic situation created by the premises. Induction thus conceived is the methodology of ampliative reasoning for securing the “best available answer” to our questions—for rational optimization in our quest for information that transcends the materials in hand. It accordingly represents a particular strategy of estimation—specifically, a general method for estimating the correct answer to a question as well as this can be done through cognitive systematization on the basis of the consid-

INDUCTION AS TRUTH ESTIMATION

141

erations of best-fit with information in hand. To be sure, given the information transcendence at issue in such truth estimation, we know that

induction does not guarantee the truth of its product. Since a process of truth estimation is at issue, inductive cogency as such is not purported to provide a theoretically failproof basis for answering our questions about how things stand in the world. Indeed, if the history of science has taught us any one thing, it is that the best estimate of the truth that we can make at any stage of the cognitive game is generally to be seen, with the wisdom of hindsight, as being far off the mark. Nevertheless, the fact remains that the inductively indicated answer does in fact afford our best available estimate of the true answer—in the sense of that one for whose acceptance as true the optimal qverall case be constructed with the instruments at hand. We live in a world not of our making where we have to do the best we can with the imperfect means at our disposal. We must recognize that there is no prospect of assessing the truth (or presumptive truth) of claims in this domain independently of the use of our imperfect mechanisms of inquiry and systematization. And here it is estimation that affords the best means for doing the job, though, to be sure, the fact that we have an inductively warranted answer in hand must never be taken as a basis for shutting the door to further inquiry. A cogent piece of inductive reasoning does not necessarily exhibit its conclusion as more probable than other concernable outcomes—if only because it is somewhere between difficult and impossible to establish a workable notion of propositional probability. What it does do is to establish its conclusion as more plausible—as better qualified for acceptance _ than any of the comparably informative available alternatives. The inductively appropriate answer to a question is that answer for which, all considered, the strongest, most cogent and sensible case can be made out. It is in just precisely this sense of affording the best-attainable assurance of rational cogency that we propose to justify induction in this discussion. It is certainly not a failproof, surefire instrument for generating certifiably correct answers, something that would in the very nature of the case be infeasible in these information-transcending cases. Rather, it is a method for doing the job at issue—that of truth estimation—as well as possible in the epistemic circumstances of the case.

Truth Estimation, Systematicity, and the Hegelian Inversion To obtain a clearer view of the rationale underlying the present conception of induction, let us glance back once more to the epistemological role of systematicity in its historical aspect. The point of departure for this

142

.

THE PURSUIT OF TRUTH

inductive approach was the traditional Greek view (present in Plato and Aristotle and clearly still operative with rationalists as late as Spinoza), which—secure in a fundamental commitment to the systematicity of the real—took cognitive systematicity (i.e., systematicity as present in the framework of our knowledge) as a measure of the extent to which our purported understanding of the world can be regarded as adequate. On this classical view, the principle of adaequatio ad rem is so applied as to yield the result that, since reality is systematic, an adequate account of it must also be so. A long line of theorists accordingly regarded system as a crucial aspect of truth. The tradition they represent has stressed the overall systematicity of the truth, holding that the totality of true theses must constitute a cohesive system. This classical approach saw systematization as a twostep process: first determine the truths, and then systematize them. (Think of the analogy of building: first assemble the bricks, then construct the house.) With the tradition from Leibniz through Kant to Hegel, however, we come to an inversion that leads to a single-step process: the determination of the right components through the very process of their assembling (the choosing ofjust those bricks that will fit together to make up a sturdy wall). Once one adopts the Hegelian Inversion of the implication relationship acceptable as true —> systematizable

/

into its converse systematizable — acceptable as true,

it becomes reasonable—and if not actually inevitable, then at any rate only natural—to construe the parameters of systematicity as themselves affording acceptability indicators. If systematicity is to be our standard of acceptability, then the sundry facets of system willthemselves count as acceptability-indicative. We thus arrive at the so-called Hegelian Inversion of thetraditional relationship of systematicity to truth. Beginning with the implication thesis that what belongs to science can be systematized (“If properly belong4 Cf. F. H. Bradley: “The test [of truth] which I advocate is the idea of a whole of knowl-

edge as wide and as consistent as may be. In speaking of system [as the standard of truth] I always mean the union of these two aspects [of coherence.and comprehensiveness] . . . [which] are for me inseparably included in the idea of system. . . . Facts for it [i.e., my

view] are true . . . just so far as they contribute to the order of experience.Ifby. taking certain judgements. . . as true, I can get some system into my world, then these ‘facts’ are so far true, and if by taking certain ‘facts’ as errors I can order my experience better, than these ‘facts’ are errors” (“On Truth and Coherence,” in Essays on Truth and Reality [Ox-

ford, 1914], pp. 202-3 and 210).

INDUCTION AS TRUTH ESTIMATION

143

ing to science, then systematizable”), we then transpose it into the converse: “If a proposition is smoothly systematizable with the whole of our (purported) knowledge, then it should be accepted as a part thereof.” From being a characteristic of science (as per the regulative ideal that a body of knowledge-claims cannot qualify as a science if it lacks a systematic articulation), systematicity was ultimately transmuted into a testing standard of (presumptive) truth. The key idea at issue is simply an extension of the old idea that simplicity—or now, rather, systematicity in all its various dimensions—is the identifying hallmark of truth (simplex sigillum veri). Accordingly, the Hegelian Inversion saw the transformation of systematicity from a framework for organizing knowledge into a mechanism for determining adequate knowledge-claims. Fit, attunement, and

systematic connection thus become the determinative standard for assessing the acceptability of claims, the monitors of cognitive adequacy. In effect this Hegelian Inversion proposes that we not judge a method of inquiry by the truth of its results, but rather judge the claims to truth of the results in terms of the merit of the method that produces them (assessing this merit by both internal [coherentist] and external [pragmatist] standards). We are not to evaluate an inquiry procedure by the truth of its results but, conversely, to assess the truthfulness of the results

in terms of rational merits of the procedure—which merits are (1) internal, systematic, and coherentist, and (2) external, applicative, and pragmatist. With an autonomous inquiry procedure—where an independent “quality control” (check on output) is impossible—one has to rest content with checks on systematic functioning, including controls of pragmatic efficacy at the methodological level. The inversion at issue thus proceeds by replacing the direction of reasoning demonstrably truer — worthier of acceptance

bya line of reasoning of an essentially inverted order: better (ie., more adequately) grounded -> rationally more acceptance-worthy — presumptively truer.

The direction of the reasoning thus does not proceed from greater truth to more adequate warrant, but the very reverse. Precisely because the later stages of the application of our inquiry procedure are more fully developed and more fully warranted, we take the stance that it is rational to view them as better qualified for endowment with the presumption of truth. On the standard, orthodox, and seemingly most straightforward view, an inquiry procedure is taken to acquire rational warrant on the basis of the truthfulness of its results. The reasoning is seen to proceed

144

THE PURSUIT OF TRUTH

from “greater truthfulness” to “greater rational warrant,” with truth as the independent variable and rational warrant for acceptance as the dependent variable in the linking equation between truth and warrant. But once we turn to our very mechanisms for determining where the truth may be taken to lie (which, in the factual area, is scientific inquiry), the matter can no longer be seen in this light. For one cannot avoid a vitiating circularity in seeking to validate the procedure in view through its capacity to lead to the truth, given that what is to count as truth is to be determined by this very procedure itself. To be sure, the linking equation between truth and warrant need not be abandoned, but it must now be viewed in a very different light. For at this stage, warrant must be seen as the independent variable and

truthfulness as the dependent variable; and so, in consequence, it

emerges that our inquiry procedures are not seen as warranted because truth producing but are presumed to be truth producing because of their greater rational warrant. On this approach, it appears that to validate the propriety of an inquiry procedure in terms of its truthfulness is simply to pick up the wrong end of the stick: truthfulness should be seen as the output of warrant rather than its input—one does not approach warrant by way of truthfulness, but truthfulness by way of warrant. The, whole thrust of the present approach is that the proper view is that one does not possess a superior methodology thanks to the greatertruthfulness of

its results, but rather that the greater (presumptive) truthfulness of the results—their superior status as truth estimates—derives from possession of a superior methodology (through the operation of a rational presumption connecting superior methodology with the rational warrant for truth claims). This perspective effectively reverses the natural interpretation of the Hegelian dictum that the real is rational. This is now not so mucha remark about the nature of the real as one about the nature of cognitive rationality. The thesis is that we are warranted in our claims to truth (accuracy, correctness) in matters regarding reality insofar as these claims proceed from adequate methods of inquiry. The “real truth” is thus rational precisely in that it is determined through the output of a rationally warranted methodology. The earlier discussions stressed the overall systematicity of “the

truth”—the fact that the totality of true theses must constitute a cohesive system—and presented system as a crucial aspect of truth. But with the Hegelian Inversion the bearing of systematicity is radically transformed. From being a hallmark of science (as per the regulative idea that a body of knowledge-claims cannot qualify as a science if it lacks a systematic articulation), systematicity is transmuted into a standard oftruth—an acceptability criterion for the claims that purport to belong to science.

INDUCTION AS TRUTH ESTIMATION

145

From a desideratum of the organization of our “body of factual knowledge,” systematicity is metamorphosed into a qualifying test of membership in it—a standard of acceptability. The effect of the Hegelian Inversion is to establish “the claim of system as an arbiter of fact,” to use F. H. Bradley’s apt expression. And we must recognize that there is no prospect of asserting the truth—or presumptive truth—of claims in this factual domain independently of our efforts at systematization in scientific inquiry. The Hegelian idea of truth assessment through systematization represents a realistic effort to adjust and accommodate to this fundamental fact.

Systemic Coherentism’s Reliance on Regulative Principles of Plausibility and Presumption Philosophers of science have frequently been exercised by the prominent but yet problematic role in science of such theoretical parameters of inductive reasoning as continuity, uniformity, regularity, conservation, and simplicity. They are generally approached from one of two directions. Sometimes they are regarded as objective tendencies of nature—constitutive facts regarding a world whose mode of functioning exhibits not the horror vacui of the medievals but an analogous principle like.amor simplicitatis. In contrast, they are sometimes cast in the role of principles that bear wholly on the subjects who do scientific theorizing rather than on the object of their theories, and that reflect the subjective intellectual

predilection of the working scientist rather than any objective features of the natural universe itself. But in fact neither side of this subjective-objective dichotomy is wholly appropriate. For all these various parameters of systematicity are most advantageously seen as principles of an essentially epistemological, or rather methodological, character. They represent regulative principles for the construction of adequate explanatory accounts: procedural principles ofplausibility that afford evaluative standards governing the provision of such accounts. A positive presumption of acceptability is therefore to operate in favor of all the traditional parameters of systematization: consistency, uniformity, regularity (causality, rulishness and lawfulness in all forms), simplicity, connectedness/coherence, unity/completeness, and so forth. These are now to function as principles of presumption, dictating where our epistemic preferences ought to lie, other things being equal. (This is readily exemplified in cases of attuning theory and data in curve fitting.) For the coherentist, systematicity thus comes to do double duty as a general criterion of acceptability-as-true and as a regulative principle of plausibility and presumption.

146

.

THE PURSUIT OF TRUTH

Accordingly, it would be an ill-advised and wholly unnecessary complication to regard such principles as representing fundamentally ontological factors, as indicating a straightforward fact about the world—an inclination on the part of nature itself toward certain principles of operation (to put it somewhat anthropomorphically). The probative methodology of scientific inquiry requires principles of this sort as part of the evaluative machinery of its own modus operandi. One should thus avoid treating such factors as the results of an inquiry for which they must in fact serve as inputs. It is best to view these parameters of systematicity as a (duly warranted) regulative or procedural or methodological facet of explanatory accounts, rather than a constitutive or descriptive (world-oriented) facet of nature. They represent not so much substantive findings about nature as procedural and regulative ground rules for the conduct of scientific inquiry. (Just this view of, specifically, the uniformity of nature has been advanced by several recentwriters. )*® We have seen that the coherence analysis sets out froma starting point of data, together with certain characteristic initial assessments of presumption and plausibility. Just what is the basis for these assessments? The answer to this crucial question is simply that these standards ofplausibility and presumption are, for the coherentist, provided by, the very conception of systematicity itself. \. Coherentism is, in effect, the very quintessence of a systematist’s approach. Taking systematicity as the standard of truth, it casts the parameter of systematization as prima facie determinants of the epistemic preferability of alternatives. The principles of systematicity now represent presumptive principles regulatively governing the conduct of inquiry. The regulative standing of such parameters as principles of epistemic preferability is reflected in the following schema: Other things being (anything like) equal, give precedence in acceptance deliberations to those alternatives that in the context of other actual or putative 13 Cf. T. S. Kuhn’s statement that “nature is vastly too complex to be explored even approximately at random. Something must tell the scientist where to look and what to look for” (“The Function of Dogma in Scientific Research,” in Readings in the Philosophy of Science, ed. B.A. Brody [Englewood Cliffs, N.J., 1970], p. 367). For Kuhn, paradigms provide the main source of such presumptions. Most writers on the subject simply invoke analogy to provide for an efficient heuristic. See, e.g., the interesting cases treated in G. Polya, Induction and Analogy in Mathematics (Princeton, 1954). The point at issue goes back at least to C. S. Peirce. See the author’s Peirce’s Philosophy of Science (Notre Dame, Ind., 1978).

46 See Stephen Toulmin, The Philosophy of Science (London, 1953), esp. sec. 5.2, “Physicists Work on Presumptions, Not Assumptions,” pp. 144-48; and J. P. Day, “The Uniformity of Nature,” American Philosophical Quarterly 12 (1975): 1-16. See also the author's essay “On the SelfConsistency of Nature,” in The Primacy of Practice (Oxford, 1973), pp. 88-106.

INDUCTION AS TRUTH ESTIMATION

147

commitments are relatively more uniform (or coherent or simple or complete,

etc.) than their alternatives.

On the coherentist approach, the choice between alternative rival systematizations in the same cognitive domain is not absolute—not a matter of yes-or-no, right-versus-wrong. Itis a matter not of forced choices but of preferential choices in the light of plausibilistic constraints. The basic process is one of cost-benefit analysis. The benefits are the information

afforded by the acceptance of the candidate theses (the “data”); the costs are the incurring of implausibilities—violations of simplicity, regularity, and so on. And these determinative plausibilities are themselves of an economic character: simplicity, regularity, and the like. The regulative resort to smoothness of fit is itself a matter of the rational economy of thought. The coherentist thus transforms the adequacy principles of system structure into selection principles for system inclusion. Theircrucial features from this perspective are: 1. They are regulative, that is, they guide our cognitive actions by telling us

how to proceed in system design. (“Of alternative accounts, adopt the most simple, uniform, etc.”) .

2. They are preferential, that is, they guide the issue of cognitive precedence and priority. (“Of alternative accounts, give precedence [priority, preference] to the most simple, uniform, etc.”) 3. They are essentially negative, that is, like the Old Testament rules implicit in the injunctions of the Pentateuch, they are to be construed negatively in terms of avoid! shun! minimize! with respect to such factors as complexity, disuniformity, ete.

C. I. Lewis has spoken of the “imperative force” of deductive logic in relation to the normative bearing of consistency subject to the governing injunction “Keep your assertions mutually consistent with each other!”!” The present approach contemplates an analogous imperative force for inductive reasoning (or its coherentist functional equivalent) in relation to the parameters of systematicity as instruments of inquiry in the factual domain. The underlying unifying injunction is simply, “Construct the best (most adequate) cognitive system that you can!” The regulative status of the parameters of systematicity is a consequence of this prime imperative, via the reasoning: Render yourknowledge systematic! The systematic is what conforms to the parameters of systematicity. Therefore: Conform your knowledge to the parameters of systematicity! 7 C. I. Lewis, Values and Imperatives, ed. John Lange (Stanford, Calif., 1969), p. 192.

148

.

THE PURSUIT OF TRUTH

The conception of a synoptic cognitive system embracing all of our knowledge of this world represents an ideal in the Kantian sense—a regulative principle of reason. Seen in this light, the regulative presumptions that provide coherentism with its mechanisms of plausibility assessment are closely akin to the “regulative ideas” (maxims) of Kant’s Critique of Pure Reason—such as those of uniformity (or “resemblance”), affinity, symmetry, and purposive orderliness. Such factors are not absolutes or even givens whose presence in our corpus of knowledge is a matter of established fait accompli—they represent cognitive desiderata that we try to inject into the framework of our knowledge insofar as this is conveniently and unproblematically possible.

Justfying Induction: Retrojustification as a Crucial Resource Our inductively formed picture of the real is built up by means of conventions—of postulations and suppositions—that are dictated not by the nature of things but by what we find to be convenient and useful for our purposes in the business of obtaining and transmitting information. These conventions of interpretation and communication by whose’ means we transact our cognitive businessare conventions of convenience whose

ultimate justification lies ratherin this retrospectively proven pragmatic efficacy than in any purely theoretical validation that could have been provided from the very outset." In the process of induction—of informational gap-filling—such factors as convenience, economy, elegance, and systematicity come into opera-

tion as our paramount guides. The standard cognitive practices by which we build up our picture of this world incorporate a host of fundamental principles of initial credibility, in the absence of concrete evidence to the contrary. We standardly operate with all sorts of principles of plausibility and presumption in favor of our established cognitive sources and resources, “Trust your own senses’not least among them.

But what justifies the standard. inductive methodology for informational gap-filling? The fundamental idea of providing an inductive justifi-

cation of induction is an attractive one that has long been contemplated, by various writers.!® Yet suchan approach has never managed to make much headway because of its vulnerability to the objection that it is circular. “Surely,” it is always urged by way of objection, “you cannot vali18 These considerations are elaborated in the author's Cognitive Economy (Pittsburgh, 1989).

18 For a survey of the literature, see the author’s Induction (Oxford, 1980), esp. pp. 11718.

INDUCTION AS TRUTH ESTIMATION

149

date the general policy of appealing to experience by an appeal to experience. Since this sort of ‘appeal to the experiential record’ is itself the object of justification—the very process whose legitimacy is under scrutiny—a justificatory argument that appeals to experience moves in a circle to beg the very question at issue.” This line of objection must be overcome by any acceptable inductive justification of induction. Some theorists have contended that a straightforwardly inductive justification of induction is feasible because explicit circularity can be avoided. Alice Ambrose, for example, has tried to meet the charge of

petitio principii—of question-begging circular reasoning—by arguing that the principle of induction is not a premise of inductive arguments but a principle of inference according to which inductive inferences are made.” What is one to make of this line of defense? Viewed at closer range, the argumentation envisioned by Ambrose appears to pivot on considering various particular inductive rules akin to the following: Rule R: To argue from “most have” to (probably) “most will.”

The question at once arises: How do we justify Rule R and validate the belief that this rule will (probably) guide us aright? Consider the contention that we can show this by Rule Ritself as per the argument, using RuleRitself to effect the inferential transition from Use of Rule R has for the most part guided us aright to:

Use of Rule R will (probably) guide us aright in general [and its ongoing use is thus justified].

But this argumentation clearly establishes the acceptability of its conclusion only if we already have in hand an adequate (and thus noncircular and Rule-R-independent) justification of Rule R. That is, the argument is cogent only if it is probatively redundant and is thus pointless because it can accomplish its task only if in fact it is altogether dispensable. As these considerations indicate, vitiating circularity arises in reasoning not only when premises already include the conclusion but also when a process of argumentation is used in its own support—when the very 9 Alice Ambrose, “The Problem ofJustifying Inductive Inference,” Journal ofPhilosophy 44 (1947): 260ff. See also R. B. Braithwaite’s argumentation to the same effect: “The proposition ‘presupposed’ in the predictionist justification of an inductive inference does not function in the inference as an additional premiss. The inductive inference to the proposition that induction by simple enumeration is an effective policy in the petitio principii sense of professing to infer a conclusion from a set of premisses one of which is the conclusion itself” (Scientific Explanation [Cambridge, 1955], pp. 275~76).

150

.

THE PURSUIT OF TRUTH

argument form at issue is question-beggingly invoked for its own substantiation, be it explicitly or tacitly. It is important to note, however, that this is not necessarily the case

with an inductive justification of induction. For we emphatically need not argue along the lines of the following inference: By induction itself (with reference to the success of our past inductive endeav-

ors), we establish that induction is a generally successful method of reasoning. A generally successful method of reasoning is eo ipso rationally justified (i.e., its deliverances merit acceptance. Therefore: Induction is rationally justified (i.e., its deliverances merit acceptance).

Such a line of argumentation would indeed be improper in that it invokes—in its first premise—the validity of the very process of argumentation whose legitimacy the argument is desired to establish. The key fact, however, is that a specifically methodological justification of induction can be altogether free from this sort of problem. Let us examine this important consideration more closely. A specifically methodological justification of induction involves a twostage process: (1) an initialjustification thatis wholly noninductive butis developed through faute de mieux argumentation, and (2) an experiential retrojustification that proceeds inductively on the basis of applicative success, seeing that the earlier stage is wholly silent in this regard. Now the crucial point is that the use of induction in the course of its own validation occurs only at stage (2), where we do indeed argue from experienced successes to generic efficacy, staking a strongly committal, positive knowledge-claim as to facts about the effectiveness of induction. There is, however, no begging the question here. For when it comes to

the substantiation of this positive knowledge-claim at issue in the second phase of argumentation, where the somewhat desperate reniedy of an initial option for the inductive method is duly retrojustified, it fortunately transpires that the results of the initial phase of justificatory argumenta- _ tion are already available to meet our needs. At this later stage.of its actual use, we already have induction in hand, duly endowed with its initial justification—a justification that, albeit weak, does carry a legitimate weight sufficient for this initial stage of the deliberations. (Just this was, after all, the very reason for being of the initial justification—to ren-

der the use of induction a reasonable proposition even in advance of its further justification through demonstrated efficacy.) The justification of our inductive presumptions does not lie in establishing as a secured fact the generalization “In proceeding in this way,

INDUCTION AS TRUTH ESTIMATION

151

you will obtain correct information and will not fall into error.” Rather, it is the methodological justification “In proceeding in this way, you will efficiently foster the interests of the cognitiveenterprise; the gains and benefits will, on the whole, outweigh the losses and costs.”We adopt our © inductive procedures not because we can somehow establish their validity from the very outset but because of their in-balance cost-benefit advantage. The justification of trust in our senses, in our fellow inquirers, and in our cognitive mechanisms ultimately rests on considerations of economic rationality—-of achieving our cognitive ends in the most efficient and cost-effective way. We initially use inductive methods because they are the most promising we can find, but we eventually trust inductive methods because experience has shown that they enable us, not necessarily to do well, but to do better than we can by any other method at our disposal. There remains, to be sure, the question “Why take that ‘natural step’ of allowing experience to guide us in our question-resolving generalizations? Specifically, why allow our favorable experience with induction to count in its behalf?” The reply here is that the adequacy claims of any cognitive method must be capable of quality control and that experience is, in the end, the only available sort of controlling monitor we can have

in this domain. The complaint that such a course would be circular, since induction is itself our experience-exploiting method, is to be met by the observation that we simply have no alternative—it is this or nothing! If one is not prepared to let our experience with induction count in its favor, then one cuts oneself off immediately and a priori from developing the only course of reasoning along which a fully adequate—and thereby efficacy-oriented—justification of such a method of inquiry could ever be developed. The literature of the justification of induction affords various instances ofthe following line of thought: Hume has shown decisively that we cannot establish that any predictive method—such as induction—will be successful. It follows that we must sever the question of establishing the rationality of the use of induction from the question of establishing its efficacy. Accordingly, effectiveness and rationality must be treated as altogether separate issues.”!

Such a position is ultimately untenable. With induction—as with any sort of method—we cannot sever the issue of rationality of its employment from that of its task-realizing efficacy. It would be foolish to adopt a 2! This position underlies P. F. Strawson’s contention that while the “success of induction” is (at best) a contingent fact about this world, the “rationality of induction” is a matter _ of conceptual necessity (Introduction to Logical Theory [London, 1952], pp. 261-62).

152

:

THE PURSUIT OF TRUTH

method whose effectiveness could not be presumed, and asinine to employ one whose effectiveness is counterindicated. But there is only one way in which the actual efficacy of the inductive method can be established—-namely, through an appeal to the tribunal of experience. (As Hume and his followers have rightly insisted, the success of induction is not and cannot be guaranteed a priori and in advance.) This being so— necessarily and as a matter of the general principles of the case—it is absurd to require that it should be accomplished: ultra posse nemo obligatur; “ought” implies “can.” So we must be content to establish this important aspect of the justification of induction through the only effective means at our disposal for doing so, namely, inductive ones. And there is no harm in this, no vitiating circularity as long as, at the eventual stage at which the relative success of the inductive method is inductively argued for, we already have in hand some preliminary noninductive justification. (The crucial merit of a methodological justification of induction is that it provides for such an essential prejustification possible by faute de mieux means.) There is no escaping the fact that the linkage between the rationality of inductive method on the one hand, and its efficacy (its prospects or promise of success) on the other, can be provided only ex ‘post facto through experience—that is to say, inductively. The trick is to render this sort of self-reliance probatively viable. And this is exactly what a strategy of two-phased argumentation makes possible. There is good reason for deploying such a multistage procedure. A cognitive method, just like a person, can establish credit only by being given credit. The first step of trust must be taken—-however small or tentative it might be. To become credited (or discredited) a method must be given enough rope to hang itself. Consider an example. A deliverance of memory can be checked and refuted only by some other item drawn from memory. Here, too, we can show (by this-or-nothing argumentation) that only an experiential-inductive substantiation can be provided. Now how could such an inductive argument proceed? Clearly it must take the form of arguing from the premise that “memory has generally proven reliable in the past” to the conclusion that “memory will generally prove reliable.” But how could this premise possibly be secured? Obviously only by memory! Whatever self-reliance there is here is clearly unavoidable, | and for that very reason cannot bé condemned as vicious. One cannot reasonably regard as a shortcoming what cannot-be helped in the very

nature of things. And this holds for induction as well. But once induction achieves even the smallest bit of initial substantiation—however tentatively and provisionally—its validation can grow apace, feeding on its own resubstantiation. Induction is part and parcel of the fundamentally conventional machin-

INDUCTION AS TRUTH ESTIMATION

153

ery by which we conduct our cognitive business. For our inquiry methodology itself is not something somehow preestablished (by what means?) through the use of previously validated (by. what means?) fact. It reflects a series of postulated conventions of question resolution—of interpretation and communication—whose initial status is merely presumptive and thus rooted in the order of “mere ideas.” But these presumptions eventually attain a very different status on the basis of proven success, so that their ultimate justification rests in their pragmatic efficacy.

Self-Substantiation

The presently envisioned inductive justification of induction does not rest on any fine distinction between good and bad circularity. Rather, it manages to avoid actual circularity altogether through its specifically methodological orientation. . It is certainly true that a thesis must never be employed in its own probative support. And as we have seen, this holds also for a particular rule of reasoning or inference as well. But this is not true of something as general as a method—and certainly not of a cognitive method. In showing the propriety of our cognitive tools, we may have no alternative but to put them to work in their own behalf. This is simply a part of that systemic self-supportingness that is a requisite of any adequate cognitive instrumentality. A defense of reasoning must use reasoning. What we

have hereis not a vicious circle but an essential feedback mechanism that reflects the ultimately systemic aspect of all justificatory argumentation. There is thus no viciousness here because one would expect—and indeed demand—that any rationally warranted process of reasoning emerge as justified on its own telling. Any fact-oriented verification method-of a sufficiently broad scope can always be turned upon itself and its performance evaluated from its own point of view. For if it is capable of serving as a test of general factual theses, then general claims about its own performance will also fall within its scope. As we have seen, our principles of inductive systematization have the feature that they are themselves monitored by conditions of systematic order. This aspect of fit, of uniformity, of a smoothly closing cycle of retrovalidation, is itself an aspect of systematicity. For it is inherent in the conception of the appropriateness of a cognitive principle as a principle that it must be of general applicability. The processes by which we carry on our ground-floor inductions regarding the empirical facts of nature must thus be the same as those by which we do higher-order inductions regarding our methods and procedures—including our cognitive methods. Our inductive methods should be self-sustaining in this way. Induc-

154

.

' THE PURSUIT OF TRUTH

tive self-substantiation is thus not a matter of vicious circularity but is itselfan aspect of that general coherence and systematicity through which alone cognitive adequacy can be determined. The sort of methodological self-substantiation that is at issue in these deliberations is part and parcel of the smooth systematic dovetailing of our cognitive resources that is the hallmark of their rational adequacy. The argumentation is comprehensively systematic, placing its several elements into a coordinative framework that unites them within one overall nexus of mutual substantiation. To clarify this recourse to systematic coherence and self-substantiation, let us consider the contrast case of an admittedly artificial (and indeed even perverse) inquiry procedure, that of counterinductivism. The counterinductivist adapts the following policy of validation for empirical generalizations: If a generalization has been applied unsuccessfully in prior applications, one is

to infer its (probable) success in the application presently in hand; and conversely if it has been applied successfully in prior applications, one is to infer its (probable) failure in the case in hand.

Its perversity notwithstanding, this counterinductivist approach to inquiry appears to be thoroughly self-sustaining; given its manifest failures in the past, we would be led—by the standards of the principle itself— to also apply this method in present cases. Yet in a somewhat deeper sense this method is not self-sustaining. For our interest in any method is always systematic; we care not simply about one application (or the next or the next two or three) but at least about the whole sequence of applications in the near term—and in the long run as well. (A procedure to be applied in one special case hardly qualifies as a method.) And from this perspective, the method of counterinduction does not really qualify as self-sustaining. For consider the question “What would lead the counterinductivist to feel increased confidence-—by the standards of the method itself—that this method is effective in its future application beyond the

next two or three cases?” The obvious answer is its failure in the nearterm cases—specifically including its present application. We thus reach the paradoxical result that the best support this method can provide for itself lies in its failure to work in the case at hand. The very fact that this method realizes its goals in a given case counts against its future employment. For if the counterinductivist is to have warrant—by the standard itself{—for the continued use of the method in cases beyond the horizons of the present moment, then he or she must refrain from endorsing the application of his method in the present case (any present case!) because its actual effectiveness in this case would only count against its effectiveness in further applications. Thus, the counterinductivist method fails to

INDUCTION AS TRUTH ESTIMATION

155

achieve self-sustainingness in the systematic sense and thereby emerges as ultimately incoherent.” . As C. S. Peirce stressed a century ago,‘ self-correctiveness is a crucial requisite for any cognitive method by which we claim to learn from ex-

perience. If a cognitive method does not work, this fact is something that the method itself should not blockfrom our view but, rather, should itself

enable us to discover. And if the specific way in which we are using the method is deficient and capable of improvement, this too is something the method should beable to bring to our notice. An adequate method of learning from experience must be self-monitoring—~it must be such that even the way in which we use it can be improved upon by its own means. The sort of self-monitoring at issue in the inductive monitoring of induction is a token of adequacy rather than a defect of circularity, precisely because of its capacity to set in motion a feedback cycle of selfconfirmation and self-improvement. These considerations show how a consistent empiricism becomés possible in the face of argumentation by Bertrand Russell and others that empiricism cannot provide a justification of induction. “The inductive principle, however, is . . . incapable of being proved by an appeal to experience. . . . All arguments which, on the basis of experience, argue as to the future or the unexperienced parts of the past or present, assume the inductive principle; hence we can never use experience to prove the inductive principle without begging the question. Thus we must either accept the inductive principle on the ground of its intrinsic evidence, or forgo all justification of our expectations about the future.” The situation in which we are left by Russell is one of a Hobson’s choice between the unworkable rationalism of a wholly a priori validation of induction on the one hand, and the emptiness of radical skepticism on the other. But the methodological-pragmatic justification of induction has the important merit of enabling us to evade the horns of this dilemma. It manages to provide us witha justificatory starting point that avoids dubious rationalistic presuppositions and proceeds to an inductive validation of induction in a way that manages to avert harmful circularity. In sum, our inductive practices rest on certain presumptions as to the

proper modes of question resolution in factual matters—certain principles of plausibility prejudgment to guide our informational gap-filling. And these presumptions themselves are validated retrospectively by a process of ex post facto retrovalidation, so that the overall structures of 2 See also A. Ofsti, “Some Problems of Counter-Inductive Policy as Opposed to Inductive,” Inquiry 5 (1962): 267-83. 33 Bertrand Russell, The Problems of Philosophy (Oxford, 1912), pp. 106-7.

156

.

THE PURSUIT OF TRUTH

rational question-resolution takes the form of a cyclic feedback, which— its circular appearance notwithstanding—is neither vicious nor vitiating.

Idealistic Intimations The preceding deliberations indicate induction to be a process of truth estimation conducted with a view to resolving our questions as to how things work in the world. The procedure that characterizes this process is one of the systematization of question-resolving conjecture with observation under the guiding aegis of the principles of systematicity (implicitly, uniformity, comprehensiveness, and the rest). Induction is thus a two-sided process—with observations conditioned by nature on the one side, duly amplified by principles of order imposed by inquirers on the other. An element of idealization is clearly involved in this inductive process of estimational question-resolution. For it is only under ideal conditions that we can suppose that experience-extrapolated systematization actually yields“the real truth” about nature’s operations. Our real-life inductive conjectures are—even at best—only estimates of the truth./Only under ideal conditions can we validly maintain that our standard reliance on the methodology of induction delivers the goods. Under the actually prevailing (inevitably suboptimal) conditions in which we labor, the most we can maintain on induction’s behalf in that by proceeding in this way, we do the very best we rationally can toward the realization of our cognitive goals. The element of idealization that inevitably conditions our inductively managed inquiries deserves closer examination.” °4 The issues of this chapter are treated in the author's Cognitive Systematization,in i Cognitive Economy, and esp. in Induction.

Ten A Coherentist Criteriology of Truth The Coherentist Approach to Inquiry Cognitive coherentism pivots on what we have called the Hegelian In-

version—the transition from system as organizer of what is accepted to that of system as arbiter of what is acceptable. The coherentist approach to epistemic justification results when this inversion is conjoined with the network theory of cyclic systematization. Coherentism thus views the systemic interrelatedness of factual theses as the criterial standard of their acceptability. But just how does such a coherence criteriology work? ; = ° The overall stance of the theory is to be articulated in terms somewhat as follows: Acceptance-as-true is in general not the starting point of inquiry but its terminus. To begin with, all that we generally have is a body of prima facie truths, that is, propositions that qualify as potential—perhaps even as promising-—candidates for acceptance. The epistemic realities being as they are, these candidate truths will, in general, form a mutually inconsistent set, and so exclude one another so as to destroy the prospects of their being accorded in toto recognition as truths pure and simple. The best that can be done in such circumstances is to endorse those as truths that best cohere with the others so as to “make the most” of the data as a whole in the epistemic circumstances at issue. Systemic coherence thus affords the criterial validation of the qualifications of truth candidates for being classed as genuine truths. Systematicity becomes not just the organizer but the test of truth.

A coherentist epistemology thus views the extraction of knowledge from the data by means of an analysis of best-fit considerations. Its approach is fundamentally holistic in judging the acceptability of every purported item of information by its capacity to contribute toward a well-ordered, systemic whole. It is important to stress that the need for input data is inevitable in the realm of factual knowledge. But this unavoidable recourse is not a concession to foundationalism. For the coherentist’s inputs are raw materials and not themselves finished products. They are not secure and solid; they have nothing “foundational” about them. In general terms, the coherence criterion of truth operates as follows. One begins with a datum-set § = {P,, P:, P3, . . .} of suitably “given”

158

.

THE PURSUIT OF TRUTH

propositions. To be sure, they are not given as secure truths in a foundationalist’s manner of theses established once and for all, but merely as

presumptive or potential truths, thatis, as promising truth candidates— and in general as competing ones that are mutually inconsistent. The task to which a coherentist epistemology addresses itself is that of using coherence considerations to bring order into S by separating the sheep from the goats, distinguishing what merits acceptance as true from what does not. The procedure that characterizes a coherenceanalysiscalls for the fol-

lowing epistemic resources: L Inputs

1. Data: theses that can serve as acceptance candidates |in the context of the inquiry, contentions that, at best, are merely presumptively true (like the“data of sense”). These are not certified truths (or even probable truths) but theses

that are in a position to make some claims upon-us for acceptance. They are prima facie truths in the sense that we would be inclined to grant them acceptance-as-true if (and this is a very big if) there were no countervailing considerations upon the scene. (The classical examples of“data” inthis| sense are those of perception and memory.) 2. Plausibility ratings: comparative evaluations of our initialasséssment (in the context at issue) of the relative acceptability of the data. This is a matter of

their relative acceptability “at first glance” (so to speak) and in the first analysis, prior to their systematic evaluation. The plausibility standing of truth candidates is thus to be accorded without any prejudgments as to how these theses will fare in the final analysis.

The governing injunction is, “Maintain as best you can the overall fit of mutual attunement by proceeding—when necessary—to make the less plausible competitors give way to the more plausible.” On this approach, a truth candidate comes to make good its claims to recognition as a truth through its consistency with as much as possible from among the rest of such data. The situation arising here resembles the solving of a jigsaw puzzle with superfluous pieces that cannot possibly be fitted into any — maximally orderly, systemically integrated picture representing “the correct solution.” . .The process of deriving uséful information from imperfect data is a key feature of the coherence theory of truth, which faces (rather than, like standard logic, evades) the question of the inferences appropriately to be drawn from an inconsistent set of premises. On this approach, the coherence theory of truth views the problem of truth determination as a matter of bringing order into a chaos comprised of initial data that mingle the secure and the infirm. It sees the problem in transformational terms: in-

COHERENTIST CRITERIOLOGY OF TRUTH

159

coherence into coherence, disorder into system, candidate truths into

qualified truths. Given an informational starting point of plausible data, the coherence analysis sets out to sift through these truth candidates with a view to minimizing the conflicts that may arise. Its basic mechanism is that of best-fit considerations, whichbrings us to the second main component of coherentism. Il. Machinery of Best-Fit Analysis The family embracing the truth candidates that are maximally attuned to one another is to count (on this criterion of overall mutual accommodation) as best

qualified for acceptance as presumably true, implementing the idea of compatibility screening on the basis of best-fit considerations. Mutual coherence becomes the arbiter of acceptability, which makes the less plausible alternatives give way to those of greater plausibility. The acceptability-determining mechanism at issue proceeds on the principle of optimizing our admission of the claims implicit in the data, striving to maximize our retention of the data subject to the plausibilities ofthe situation.

Against this background, the general strategy of the coherence theory lies in a three-step procedure: 1. To gather in all of the data (in the present technical sense of this term). 2. To lay out all the available conflict-resolving options that represent the alternative possibilities that are cognitively at hand. 3. To choose among these alternatives by using the guidance of plausibility considerations, invoking (in our present context) the various parameters of systematicity—regularity, uniformity, simplicity, and the rest—to serve as indices of plausibility.

The coherence theory accordingly implements F. H. Bradley's dictum that system (i.e., systematicity) provides a test criterion most appropri-

ately fitted to serve as arbiter of truth, The interaction of observation and theory provides an illustration of this process. Take grammar. Here one moves inferentially from the phenomena of actual usage to the framework of laws by the way of a best-fit principle (an “inference to the best systematization,” as it were), and one checks that the cycle closes by moving back again to the phenomena by way of their subsumption under the inferred rules. Something may well get lost en route in this process of mutual attunement—for example, some of the observed phenomena of actual language use may simply be . dismissed (say as “slips of the tongue”). Again, the fitting of curves to 1 The formal mechanism of best-fit analysis is described more fully in the author's books The Coherence Theory of Truth (Oxford, 1973) and Plausible Reasoning (Assen, 1976).

160

.

THE PURSUIT OF TRUTH

observation points in science also illustrates this sort of feedback process of discriminating the true and the false on best-fit considerations. The crucial point for present purposes is simply that a systematization can

effectively control and correct data—even (to a substantial extent) the very data on which it itself is based. The coherentist approach is thus quite prepared to dispense with any requirement for self-evident protocols to serve as the foundations of the cognitive system. The justification of a system-included thesis will not proceed by derivations from the axioms but comes to obtain through the pattern of its interrelationships with the rest. On a coherence approach, the truth is not seen as a treelike structure supported by a firm-rooted trunk, as it is on the foundationalist theory. Rather, it appears like a multitude of tied objects thrown into water; some of them rise to the surface themselves or are dragged there by others, and some of them sink to the bottom under their own weight or through the pull of others. The coherentist approach to cognition assumes an entirely inward orientation; it does not seek to compare the truth candidates directly with “the facts” obtaining outside the epistemic context; rather, having gathered in as much (presumptive) information about the facts aspossible, it seeks to sift the true from the false within this body. On this approach, — the validation of an item of knowledge—the rationalization if its inclusion alongside others within “the body of our knowledge’—proceeds by way

of exhibiting its interrelationships with the rest: they must all be linked together in a connected, mutually supportive way (rather than having the form of an inferential structure built up upon a footing of rock-bottom axioms). On the coherentist’s theory, justification is not a matter of derivation but one of systematization. We operate, in effect, with the equation “justified” = “systematized.” The coherence approach can be thought of as representing, ineffect, the systems-analysis approach to the criteriology of truth. One critic of coherentist epistemology has objected: “According to. .. [one] version of the coherence theory a certain given statement, of which we do not know whether it is true or not, is to be accepted as true if (and only if) it coheres with the statements we have previously accepted. This version has the effect of making our knowledge utterly conservative: “entrenched’ knowledge can hardly be overthrown.”? But this objection —

misses its mark if directed against the presently envisioned version of the theory. For “entrenchment” (acceptance, actual credence) is, as we shall

see, not all that decisive. Even a well-entrenched item can be dislodged in the face of more plausible data: little is better entrenched than the validity of “what one sees with one’s own eyes’—yet even here one can 2K. R. Popper, Objective Knowledge (London, 1973), p. 309.

COHERENTIST CRITERIOLOGY OF TRUTH

161

be brought to recognize that one has fallen victim to an illusion. Our “accepted facts” unquestionably have a “benefit of doubt,” but that is not to say that their lease on life is absolute.

The Contrast between Foundationalism and Coherentism A coherentist theory of factual knowledge stands in sharp contrast with the more restrictive foundationalist approach of the mainstream tradition of Western epistemology. Unlike foundationalism, coherentism dispenses with any appeal to basic, foundational truths of fact, diametrically opposing the view that knowledge of the actual, and even of the probable, requires a foundation of certainty. The coherence approach maintains that truth is accéssible in the extralogical realm on the basis of bestfit considerations. This entire procedure goes wholly counter to the classical epistemologists’ axiomatic quest for basic, or foundational, truths. For the coherentist, knowledge is not a Baconian brick wall, with

block supporting block upon a solid foundation; rather, an item of knowledge is like a node of a spider’s web that is linked to others by thin strands of connection, each alone weak, but all together adequate for its

support. Foundationalism might be caricatured as an essentially aristocratic view of truth: truths as such are not equal; there are certain “master” truths on which the other, “client” truths are totally dependent. Negating the need for any axiomatic truths, the coherence theory sets out to implement a rather more democratic concept of treating all the truth candidates not necessarily as equal but at any rate as all more or less plausible. The possibilities rendered available by the data at our disposal are treated with a complete “equality of opportunity”; truthfulness is determined from them only through a process of interaction—that is, by considerations of a best fit in terms of mutual accord and attunement (rather than their falling into the implicative captivity of certain basic prior truths). The coherentist’s approach effectively inverts that of the foundationalist. The foundationalist begins his or her epistemological labors with a very small initial collection of absolutely certain truths and from these proceeds to work outward by suitably additive procedures of supplementation to arrive at.a wider domain of truth. By contrast, the coherentist begins with a very large initial collection of insecure pretenders to truth and from thesé proceeds to work inward by suitably reductive procedures of elimination to arrive at a narrower domain of truth. The expansive approach of the foundationalist is the very opposite of the contractive approach of the coherentist. The foundationalist is forced to a starting

162

.

THE PURSUIT OF TRUTH

point of few but highly secure items and immediately faces the dilemma of security versus content. The coherentist bypasses this difficulty altogether. He or she begins with too many items—far too many, since the data generally stand in a conflict of logical incompatibility. But the coherentist proceeds to undo the damage of this embarrassment of riches by suitably reductive maneuvers. This approach avoids altogether the characteristic perplexity of foundationalist epistemology in finding appropriate candidates to supply the requisitely secure foundation. There are two profoundly different approaches to the cognitive enterprise that, for want of better choices, might be called the ampliative and the reductive. The ampliative strategy searches for highly secure propositions that are acceptable as “true beyond reasonable doubt.” Given such a carefully circumscribed and tightly controlled starter set of propositions, one proceeds to move outward ampliatively by making inferences from this secure home base of an entirely unproblematic core. . The reductive strategy, however, proceeds in-exactly the opposite direction. It begins in a quest not for unproblematically acceptable truths but for well-qualified candidates or prospects for truth. At the outset one does not require contentions that are certain and altogether qualified for recognition as genuine truths, but propositions that are no more than plausible, well-spoken-for, well-grounded candidates for endorsement. Of course, not all of these promising truth-candidates are endorsed or

accepted as true. We cannot simply adopt the whole lot, because they are competing—mutually contradictory. We have to impose a delimiting (and consistency-restoring) screening out that separates the sheep from the goats until we are left with something that merits endorsement. And here we proceed by way of diminution or compression by narrowing that overample range of plausible prospects for endorsement. While the paradigm instrument of ampliative reasoning is deductive derivation, the paradigm method of contraction is dialectical argumentation. To effect the necessary reductions, we do not proceed via a single deductive chain, but through backing and filling along complex cycles that criss-cross over the same ground from different angles of approach in their efforts to identify weak spots. The object of the exercise is to determine how well enmeshed a thesis is in the complex fabric of diverse and potentially discordant and competing contentions. We are now looking for the best candidates among competing alternatives—for that reso-

lution for which, on balance, the strongest overall case can be made. It is not “the uniquely correct answer” but “the most defensible position” that we seek in coherentist inquiry. The essential difference between the coherence methodology and any foundationalist approach to acceptance-as-true lies in the fact that on the latter line of approach, every discursive (i.e., reasoned) claim to truth requires truths as inputs. If a (presumptively) true result is to be ob-

COHERENTIST CRITERIOLOGY OF TRUTH

163

tained, the premises on which it rests must themselves be true (or assumed to be so). The only strictly originative provider of de novo truths is the process that yields the immediate truths of the starter set. The decisive difference of the coherence theory is its capacity to extract (presumptive) truths discursively from a basis that includes no conceded truths whatsoever—that is, from data that are merely truth candidates and not truths. The foundationalist requirement for basic truths is some‘thing that the coherence theory—proceeding as it does from a basis of data that need be neither compatible nor true—has been designed to overcome. The analysis seeks to provide a procedure for arriving at output truths without requiring any input truths as an indispensable starting basis. The motto “Truth without true foundations” may properly be inscribed on the banner-of the coherence theory of truth. The contrast between a foundationalist construction (Aufbau) of the domain of truth and the reductive approach of the coherence analysis is set out graphically in figure 10.1. This diagram makes plain the basic similarities between the two approaches but also brings out their substantial differences, which are as follows: 1. On the foundationalist approach there are two distinct sorts of knowledge, the immediate and the derivative, while for the coherentist all knowledge is essentially ofa piece. 2. On the foundationalist approach experience is called upon to provide ba-

sic knowledge (in the form of certain immediately evident truths), while for the coherentist it provides only the raw data for knowledge. Figure 10.1. Foundationalism vs. Coherentism in Factual Inquiry

The Nondiscursive Sector

Foundationalisi: Cummumuumemnadso Experience

The Discursive Sector

Derivative truths

Immediately evident | basic truths Inductive logic

COherentiSI.: (\-t_mummmmn= Experience

Raw data

>!

Coherentist truths

Data

Coherence

analysis

164

.

THE PURSUIT OF TRUTH

3. On the foundationalist approach all discursive (inductively or deductively inferential) processes require an input of known truths if truths are to be an output (which is exactly why an immediate, nondiscursive route to truth must

be postulated). The coherence analysis differs fundamentally in this regard. 4, On the foundationalist approach the initial givens are wholly nondiscursive and fixed invariants, while on the coherence approach the data represent a mixture of experiential and discursive elements. (The raw data are, to be sure, nondiscursive, but for the coherentist they are only one part of the total data and are by no means fixed and sacred but subject to a cyclic process of reappraisal and revision.) In consequence: 5. On the foundationalist approach nothing whatever that happens at the epistemically later stages of the analysis can possibly affect the starting point of basic truths, while on the coherentist approach there is a feedback loop through which the data themselves can be conditioned by the outcome of.a coherence analysis (in other contexts) and their status is subject to reevaluation in the light of new insights regarding their plausibility. 6. Unlike the foundationalist approach, the coherence analysis does not require a sharp disparity in the treatment of particular and general propositions (between “observation statements” and “laws”). Its data for factual inquiry are subject to no particularity stipulations, unlike the usual “directly evident” givens of the foundationalists.® 7. On the foundationalist approach the body of evidence from which the reasoning proceeds must be self-consistent. The coherence analysis has no need for this unrealistic supposition.

The diagram also brings to the fore one further facet of coherentist method that merits special emphasis. Foundationalist inductivism adopts the basically linear systematization of reasoning typical of mathematics;

here once a result is obtained, one simply passes on to other matters— there is no need ever to return to the reappraisal or resubstantiation of something that has already been “established.” But with coherentist inductivism the case is quite otherwise, as the feedback loop of the diagram illustrates graphically. Here there is a definite place for a dialectical process of cyclical structure, where one returns repeatedly to an item already “established.” For the process of confirmation is now more complex and cyclically interactive: a thesis might first appear in the status of a mere datum of low plausibility, later as one of higher plausibility, and ultimately even as a validated truth. The difficulty with any sort of foundationalism lies in the matter of foundations.If they are phenomenal (“I seem to see a steel dagger before me’; “I deem it highly likely that he will come”), then they are in the 3 Cf. R. M. Chisholm, The Theory of Knowledge, 2d ed. (Englewood Cliffs, NJ. 1977).

COHERENTIST CRITERIOLOGY OF TRUTH

165

final analysis autobiographical rather than objectively factual. If, however, they are factual (“There is an apple on the table’—which, of course, cannot have a golden interior—after all, apples just don’t—even though I have not checked this), then it is hard to see how they can be founda-

tional, since their own status is always somehow problematic. The foundationalist approach to factual knowledge thus faces a dilemma: the sort of claims at issue in objective knowledge lack cognitively unproblematic foundations, while cognitively unproblematic contentions do not yield objective facts. Coherentism sets out to resolve this problem. Rather than proceeding linearly, by fresh deductions from several premises, it proposes to cycle round and round the same given family of prospects and possibilities, sorting out, refitting, refining until a more sophisticatedly developed and more deeply elaborated resolution is ultimately arrived at. The information-extracting process developed along these lines is one not of advance into new informative territory but of cyclic reappraisal and revision of the old, tightening the net around our ultimate conclusion as we move round and round again, gaining a surer confidence in the wake of more refined reappraisals. This cyclic process of reappraisal is such that one can even—in suitable circumstances—dispense with the need for new data-inputs in an endeavor to squeeze more infor-

mation out of the old. It is readily observed that this repeated reappraisal of claims is in fact closer to the processes of thought one generally employs in scientific reasoning. , The coherentist unhesitatingly endorses the historic thesis that knowledge is “true, justified belief,” construing this as tantamount to claiming that the known is that whose acceptance-as-true is adequately warranted through an appropriate sort of systematization. However, since the systematization at issue is viewed as being of the network type, the impact of the thesis is drastically altered. For we now envisage a variant view of justification, one that radically reorients the thesis from the direction of

the foundationalists’ quest for an ultimate basis for knowledge as a quasiaxiomatic structure. Now “justified” comes to méan, not “derived from basic (or axiomatic) knowledge,” but rather “appropriately interconnected with the rest of what is known.” A coherence epistemology of truth-claims recognizes that the grounding of such claims cannot be evidential “all the way down” because evidentiation has to come to a stop. And it cannot be probabilistic “all the way down” because probabilities must ultimately come to rest on some categorically factual basis or other. Rather, so coherentism insists, the

validating process is not linear but cyclic. And coherentism’s network conception of grounding offers a perfectly viable alternative here. Our concern with the validation of cognitive claims is not (or need not be) an

166

.

THE PURSUIT OF TRUTH

issue of extracting “real knowledge” from “mere beliefs” by use of a philosopher’s stone of epistemic justification. This entire reductive or extractive impetus is absent from coherentism’s preoccupation with the rational structure of knowledge rather than with its heuristic origin. And here the network approach is not only a possible but an emphatically attractive alternative to the foundationalist’s axiomatic methodology.

The Central Role of Data for a Coherentist Truth-Criteriology The concept of a datum, whose role is pivotal in coherentist methodology, is something of a technical instrumentality. To be sure, the idea is one not entirely unrelated to the ordinary use of that term, nor to its (somewhat different) use among philosophers; yet it is significantly different from both. A datum is a truth candidate, a proposition tobe taken not as true but as potentially or presumptively true. It is a prima facie truth in exactly the sense in which one speaks of prima facie duties in ethics—a thesis that we would, in the circumstances, be prepared to class as true provided that no countervailing considerations are operative. A datum is thus a proposition that one is to class as true if one can, that is, if doing so does not generate any difficulties or inconsistencies. Such a datum is, in the traditional sense, a given. But a proposition may be given in two ways: . 1. as a truth or as actually true; to be classed as true unqualifiedly, sans phrase; 2. as a truth candidate; as potentially or presumptively true; to be classed as true, provided that doing so creates no anomalies.

A datum is a proposition that is given not in the first mode but only in the second: it is a pretender to truth whose credentials may well prove insufficient, a runner in a race it may not win. (And what is true of givens holds also for assumptions:-a proposition can be assumed to be a truth [i.e., assumed as true] or assumed to be a truth candidate [i.e., assumed as a datum]. A very different condition of things is posited in the two cases.) For a proposition to count as a datum is altogether different from its counting as a truth, just as a person’s being a presidential candidate is something far different from that one being a president. Presidential candidates are not presidents; data are not truths. Truth candidacy does not require or presuppose truth: quite different issues are involved. A potential truth is a truth no more than an egg-enclosed embryo is a hen. The “acceptance” of a proposition as a truth candidate is not acceptance at all

COHERENTIST CRITERIOLOGY OF TRUTH

167

but a highly provisional and conditional epistemic inclination toward it, an inclination that falls far short of outright commitment. A datum is not established as true; it is backed only by a presumption that it may turn out true “if all goes well.” It lays a claim to truth, but it may not be able to make good this claim in the final analysis. (A datum is thus something altogether different from the basic, protocol truths that serve as foundation for an Aufbau of truth in the manner of the logical positivists.) The following objection might be made: How can you speak of asserting a proposition merely as a datum but not asa truth? If one is to assert (accept, maintain) the proposition in any way at all, does one not thereby assert (accept, maintain) it to be true? The answer to be made here is simply a head-on denial. To maintain P as a datum is not to say it is true, but merely as true appearing, as true seeming “on the available indications.” Such characterizing as potentially or presumptively. factual is somewhat akin to maintaining P as possible or as probable: in no case are these tantamount to maintaining the proposition as true. Putting a prop-

osition forward as possible or probable commits one to claiming no more than that it is “possibly true” or “probably true.” Similarly, to assert P as a datum is to say no more than that P is potentially or presumptively true—that it is true appearing—but does not say that P is actually true, that it is a truth. As with assertions of possibility or probability, a claim of datahood definitely stops short—far short—of a claim to truth. We in general know that data cannot be identified with truths—that some of them must indeed be falsehoods--because they are generally incompatible with one another. Truth candidates—like rival candidates for public office—can work to exclude one another: they are mutually exclusive, and victory for one spells defeat for the others. Candidate truths are not truths pure and simple because it is of the very nature of the case that matters must so eventuate that some of them are falsehoods. In taking a proposition to be a datum, we propose to class it as true whenever possible, recognizing that this may not be possible because some data may contradict one another, so that we must consider some of them as nontruths.* A datum is to be carried across the line from datahood to truth automatically whenever such a transfer is unproblematic (i.e., in no way involves a contradiction). A member of a group of data

that meets with no rivalry from its fellows can immediately be accepted as true. Thus:® 4 The epistemic problem with data is thus not just that some of them are uncertain but

that, in general being collectively inconsistent, we know a priori that some of them must be false. 5 In this table, propositional variables p, q, r, etc. are to represent independent propositions.

168

;

THE PURSUIT OF TRUTH

We obtain as truths

Given the data 1

Pd

(relative to these data) Pq

2. p, ~p

—

3. p,q, ~q

p

4. 7.4.7, ~P\/~q, PV~p

r

Our stance toward data is unashamedly that of fair-weather friends: we adhere to them when this involves no problems whatsoever but abandon them in the face of difficulties. But it is quite clear that such loose attachment to a datum is by no means tantamount to no attachment at all. As coherentist methodologists see the process ofinquiry, all that we generally have to begin with is a body of data that qualify aspotential— perhaps even as promising—candidates for truth. Theepistemic/realities being as they are, these candidate truths willin general form a mutually inconsistent set so as to destroy their prospects of their being accorded in toto recognition as truths pure and simple. We are accordingly well advised to jettison what is necessary in an endeavor to make the most of the data as a whole. Coherence thus becomes the critical test of the qualifications of truth candidates for gaining endorsement as genuine truths. , The conception of a datum certainly does not open the floodgates in an indiscriminate way. Not everything is a datum: the concept is to have some logico-epistemic bite. To be a datum is not just to be a proposition that could conceivably be claimed to be true but to be a proposition that (under the circumstances) can be claimed to be true with at least some

plausibility: its claim must be well founded. A proposition will not qualify as a datum without some appropriate grounding. Data are propositions that have a proper claim upon truth, and we must distinguish between truth claims that can reasonably be made from those that are merely theoretically possible. (Not every human being is a possible winner in a race, but only those who are genuinely “in the running.”) A datum is a proposition that, given the circumstances of the case, is a real prospect for truth in terms of the availability of reasons to warrant its truth candidacy. A datum is not merely something that is “possibly true” or that is “true for all I know about the matter.” To class a proposition as a datum is to’ take a definite and committal position with respect to it, so as to say, “I propose to accept it as true insofar as this is permitted byanalogous and possibly conflicting commitments elsewhere.” That cautionary proviso here means thatin point of truth datahood is not an isolated feature of a single proposition but a contextual feature of a group of propositions, that is, a family of data. The reference to a “con-

COHERENTIST CRITERIOLOGY OF TRUTH

169

flict with others” gives to an imputation of datahood a relative and context-involving aspect. Virtually all writers on the subject take the position that, as one recent writer puts it, “acceptance as true is a necessary condition for acceptance as evidence.”¢ On this view, if something can count as genuine and actual evidence only when it is established as true, the weaker conception of potential evidence—as contradistinguished from

evidence as such (the actual evidence)—comes closer to our conception of the data. But although some basis is required for counting as a datum, this basis need not be very strong by the usual epistemic standards. To be a datum is to be a truth candidate. And to count as such, a proposition need be neither an actual truth nor a probable truth: it need only be a promising candidate oralive possibility for truth. Such propositions are not truths, but they make good a claim to truth that is at best tentative and provisional; by themselves they do not present truth but at most indicate it. Historically the tentativity of the experientially and mnemonically “given” has always been recognized, and the deliverances of our senses and our memory are the traditional examples of this circumstance of merely putative truth. ’A more ambitious claim to truthfulness characterizes such a family of data rather than any one member of it. No imputation of truth (as opposed to presumptive or candidate truth) attaches to any individual datum, but there is a definite implicit claim that the “logical space” spanned by the data as a whole somewhere embraces the truth of the matter. The winner of a race must be sought among those “in the running.” Taken individually, the data are merely truth presumptive, but taken collectively as a family, they are to be viewed as truth embracing: admitting that they do not pinpoint where the truth lies, we are committed to the view that once we have all the relevantdata i in hand, they define the area where it is to be found. One may well ask:“How can coherence with the data yield truth if the data themselves are not individually true? How can something so tentative prove sufficiently determinative?” The answer can be found only in the detailed development of the machinery; at this preliminary stage we can do no better than give a rough analogy. Consider again the earlier example of a jigsaw puzzle with superfluous pieces. It is clear here that the factor of “suitably fitting in” will be determinative of a piece’s place (or lack of place) in the correct solution. Not the (admittedly tentative) status of the individual pieces but their mutual relationships of systematic accord is the determinative consideration. Thus the issue is not that of 6]. Levi, Gambling with Truth (New York, 1967), p. 28. “To accept H as evidence is not merely to accept H as true but to regard as pointless further evidence collection in order to _ check on H” (p. 149).

170

.

THE PURSUIT OF TRUTH

how mere.truth candidacy itself can engender truth but rather of how it can help to determine truth in the presence of certain systematic considerations. Of course the exact nature of these considerations remains to be considered, but the key fact at this stage is that there is an important epistemic category of claims to presumptive or provisional verisimilitude that carry truth-indicative weight, while yet stopping well short of claims to truth as such.

Some Objections Removed Data—construed along some such lines as that of the preceding discussion—play a pivotal rolein the articulation of a coherence-basedcriterion of truth. The entire drama of a coherence analysis is played out within the sphere of propositions, in terms of the sorts-of relationships they have to one another. Now there is, in any event, enough merit in a correspondence account of truth that an appropriate consonance must obtain between “the actual facts of the matter” and a proposition regarding them that can qualify as true. To the criticism “Why should mere coherence imply truth?” one can and should reply that what is at issue here is not mere coherence but coherence with the data. It is not with bare coherence as such (whatever that would be) but with data-directed coherence that a truth-making capacity enters upon the scene. But of course datahood furnishes only the building blocks for truth determination and not the structure itself. Coherence plays the essential role because it is to be through its mediation that we move from truth candidacy and presumptions of factuality to truth as such. And the procedure is fundamentally noncircular: we need make no imputations of truth at the level of data to arrive at truths through application of the criterial machinery in view.

In his exposition of the coherence theory, Brand Blanshard has -written:

.

Granting that propositions, to be true, must be coherent with each other, may they not be coherent without being true? . . . Again, a novel, or a succession of novels such as Galsworthy’s Forsyte Saga, may create a special world of characters and events which is at once extremely complex and internally consistent; does that make it the less fictitious? . . . This objection, like so many other annihilating criticisms, would have more point if anyone had ever held the theory it demolishes. But if intended to represent the coherence theory as responsibly advocated, it is a gross misunderstanding. That theory does not hold that any and every system is true, no matter how abstract and limited; it holds that one system only is true, namely the system in which everything real and possible is coherently included. How one can find in this the notion that a

COHERENTIST CRITERIOLOGY OF TRUTH

171

system would still give truth if, like some arbitrary geometry, it disregarded experience completely, it is not easy to see.”

This key passage, intended to answer a basic objection, leaves matters in a badly muddled state. Just where is coherence to be operative? In “the system in which everything real and possible is coherently included”? But here—in this all-inclusive system—there is no difference drawn or to be drawn between the actually real and the merely possible with respect to coherence: this exactly is the force of the initial objection. Yet Blanshard’s position is at bottom correct. The coherence theory would indeed be deficient if it held “that a system would still give truth if. . . it disregarded experience completely.” Our recourse to data is intended to supply just this requisite of a recourse to experience. The concept of a datum along the general lines explicated in the preceding discussion is certainly no newcomer to epistemology. Coherence theorists and others have articulated conceptions of much this same sort,

F. H. Bradley himself being a prime case in point. In his essay “On Truth and Coherence,”® Bradley introduces the concept of a fact so as to have it play a role closely akin to that of a truth candidate in our sense. First ofall, Bradleyan “facts”—let us call them B-facts—differ from everyone else’s facts in not necessarily being factual (i.e., true). Typical, for Bradley, are the “facts of perception and memory,” which need not, of course, be true at all but are at best purportedly or presumptively veridical. These facts of perception [and memory], I further agree, are at least in part irrational [and so false]. . . . [Yet] I do not believe that we can make ourselves independent of these non-relational data. But, if I do not believe all this, does

it follow that I have to accept independent facts [i.e., facts true independently of all other considerations]? Does it follow that perception and memory give one truths which I must take up and keep as they are given me, truths which in principle cannot be erroneous? This surely would be to pass from one false extreme to another. . . . I therefore conclude that no given fact is sacrosanct. With every fact of perception or memory a modified interpretation is in prin-

ciple possible, and no such fact therefore is free from all possibility of error.®

Bradley espouses—with respect to the limited range ofperception, memory, and the testimony of others—a notion of “fact” according to which the facts do not automatically qualify as truths at all but at best as possible or potential truths.!° Despite its limitations in scope, Bradley’s concep7B. Blanshard, The Nature of Thought (London, 1939), pp. 275~76. 8 F, H. Bradley, Essays on Truth and Reality (Oxford, 1914), chap. 7, pp. 202-18. 8 Tbid., pp. 2034. See also discussion of Bradley’s position in A. C. Ewing, Idealism: A Critical Survey (London, 1934), esp. pp. 239-40. 10 Cf. C. I. Lewis's thesis that “whatever is remembered, whether as explicit recollection

_or merely in the form of our sense of the past, is prima facie credible because so remembered” (An Analysis of Knowledge and Valuation [La Salle, Hl., 1946], p. 334). This line of

172

.

THE PURSUIT OF TRUTH

tion of a fact is clearly a precursor of—indeed almost a paradigm for—our own conception of a datum. One acute critic has made the following charge against traditional formulations of the coherence theory: It is on this point particularly that the historical coherence theory appears to be ambiguous; it seems never possible to be sure, in presentations of that con-

ception, whether “coherence” implies some essential relation to experience, or whether it requires only some purely logical relationship of the statements in

question. Indeed, the so-called “modern logic”, associated with this theory, is such as totally to obscure the essential distinction between analytic truths of logic and those empirical truths we can only be assured by some reference beyond logic to given data of sense."

\

The presently contemplated version of the coherencecriteriology of acceptability is immune to this criticism. For the requisite “essential relation to experience” is provided by the essential reliance upon data (i.e., by restrictive use of only certain propositions as data). The “purely logical relationship of the statements in question,” in terms ofwhich the conception of coherence is implemented, comes into play only after these data are in hand, in providing the mechanism for determining some among them to be truths. The concept of a datum thus doesacritically important job for the coherence theory of truth. It serves to provide an answer to the question “Coherent with what?” without postulating a prior category of fundamental truth. It provides the coherence theory with grist to its mill that need not itself be the product of some preliminary determinations of truth. A reliance upon data makes it possible to contemplate a coherence theory that produces truth not ex nihilo (which would be impossible) but yet from a basis that does not itself demand any prior determiziations of truthfulness as such. A coherence criterion can, on this basis, furnish a mechanism that is originative of truth—that is, it yields truths as outputs without requiring that truths must also be present amongthe supplied inputs.

It is sometimes objected that coherence cannot be the standard of truth because then we may well arrive at a multiplicity of diverse but equally coherent structures, whereas truth is of its very nature conceived of as unique and monolithic. Bertrand Russell, for example, argues in this way: “There is no reason to suppose that only one coherent body of bethought was developed earlier in substantial detail in A. Meinong’s important essay “Zur erkenntnistheoretischen Wirdigung des Gedachtnisses,” in his Gesammelte Abhandlungen (Leipzig, 1933). Meinong there argues that memory-derived judgments must be accorded “immediate presumptive evidence” (p. 207). This conception of the presumptively evident is clearly yet another precursor of the conception of prima facie truth. 4 Lewis, Analysis of Knowledge and Valuation, p. 339.

COHERENTIST CRITERIOLOGY OF TRUTH

173

liefs is possible. It may be that, with sufficient imagination, a novelist might invent a past for the world that would perfectly fit on to what we know, and yet be quite different from the real past. In more scientific matters, it is certain that there are often two or more hypotheses which account for all the known facts on some subject, and although, in such cases, men of science endeavour to find facts which will rule out all the

hypotheses except one, there is no reason why they should always succeed.”!2 One must certainly grant Russell’s central point: however the idea of coherence is articulated in the abstract, there is something fun-

damentally undiscriminating about coherence taken by itself. Coherence may well be—nay, certainly is—a descriptive feature of the domain of truths: they cohere. But there is nothing in this to prevent propositions other than truths from cohering with one another: fiction can be made as coherent as fact; truths surely have no monopoly of coherence. Indeed “it is logically possible to have two different but equally comprehensive sets of coherent statements between which there would be, in the coherence theory, no way to decide which was the set of true statements.” In consequence, coherence cannot of and by itself discriminate between

truths and falsehoods. Coherence is thus seemingly disqualified as a means for identifying truths. Any viable coherence theory of truth must make good the claim that despite these patent facts, considerations of coherence can somehow be deployed to serve as an indicator of truth. But the presently envisioned theory averts these difficulties. It looks not to coherence in and of itself as a criterion of truth but to coherence with the data of experience. It thus renders Russell's objection effectively irrelevant. A further criticism is developed by pressing hard upon the question “Coherence with what?” Is this to mean coherence with everything— with all other propositions that can be enunciated? That is patently impossible: as a body, the totality of meaningful propositions is certainly inconsistent—and so incoherent. Is it merely coherence with something (i.e., with some other propositions) that is asked for? That clearly will not do. A novel or science-fiction tale or indeed any other sort of made-up story can be perfectly coherent. To say simply that a proposition coheres with certain others is to say too little. All propositions will satisfy this 2 B. Russell, The Problems ofPhilosophy (London, 1912), p. 191. Or cf. Moritz Schlick’s formulation of this point: “Since no one dreams of holding the statements of a story book true and those of a text of physics false, the coherence view fails utterly. Something more,

that is, must be added to coherence, namely, a principle in terms ofwhich the compatibility is to be established [sc. as factual], and this would alone then be the actual criterion” (“The

Foundation of Knowledge,” in Logical Positivism, ed. A. J. Ayer [Glencoe, Ill., 1959], p.

216).

. 3A, R. White, “Coherence Theory of Truth,” in The Encyclopedia of Philosophy, ed. P. Edward (New York, 1967), 2:131.

174

:

THE PURSUIT OF TRUTH

condition, and so it is quite unable to tell us anything that bears upon the question of truth. It would be quite senseless to suggest that a proposition’s truth resides in its coherence alone. Its coherence is of conceptual necessity a relative rather than an absolute characteristic. Coherence must always be coherence with something: the verb “to cohere with” requires an object just as much as “to be larger than” does. We do not really have a coherence theory in hand at all until the target domain of coherence is specified. Once this has been done, we may very well find that the inherent truth-indeterminacy of abstract coherence—its potential failure to yield a unique result—has been removed. Now it must be said in their defense that the traditional coherentists have never viewed truth as simply a matter of coherence per se but have insisted that it is specifically“coherence with our experience”thatis to be the standard of truth.'* The coherence theory of the British idealists has never abandoned altogether the empiricist tendency of the native tradition of philosophy. What is clearly needed is a halfway house between coherence with some (i.e., any) propositions and coherence with true propositions. Essentially, what is needed is coherence with somehow “the right” propositions. The coherence at issue in a coherence theory of truth must be construed as involving all in some way appropriately qualified propositions. This line of thought poses a task central to the construction of a workable coherence theory: that of specifying just what propositions are at stake when one speaks of determining the truth of a given proposition in terms of its “coherence with others.” Which others are at issue? It was, of course, for the sake of a satisfactory answer to this question that our approach made its crucial resort to the key concept of an experientially grounded datum that was addressed at some lengthin the preceding section.

Validating the Coherentist Approach The overall line of validation for a probative methodology of coherence involves a double circle, combining a theory-internal cycle of theoretical

self-substantiation with a theory-external cycle of pragmatic validation— M4 See A. C. Ewing, Idealism: A Critical Survey (London, 1934), p. 238, as well as his

later essay “The Correspondence Theory of Truth,” where he writes: “That coherence is the test of truth can only be made plausible if coherence is interpreted not as mére internal coherence but as coherence with our experience” (Non-Linguistic Philosophy [London, 1968], pp. 203-4). For an author of the earlier period, see H. H. Joachim, who writes: “Truth, we said, was the systematic coherence which characterized a significant whole. And

we proceeded to identify a significant whole with ‘an organized individual experience, selffilling and self-fulfilled’ ” (The Nature of Truth [Oxford, 1906], p. 78).

COHERENTIST CRITERIOLOGY OF TRUTH

175

as is illustrated in figure 10.2. Here Cycle I represents the theoretical/ cognitive cycle of intellectual consistency between regulative first principles and their substantive counterparts, and Cycle II, the practical/applicative cycle of pragmatic efficacy in implementing the substantive results of the first principles. Accordingly, the overall legitimation of a methodology for the substantiation of our factual beliefs must unite two distinctive elements: (1) an apparatus of systematic coherence at the theoretical level (a coherence in which factual presumptions and metaphysical presuppositions both will play a crucial part), and (2) a controlling monitor of considerations of pragmatic efficacy at the practical level. Neither can appropriately be dis-

pensed with for the sake of an exclusive reliance on the other. The proof of the theoretical pudding must, in the final analysis, lie in the applicative eating, by monitoring the adequacy of our procedures of cognitive systematization through an assessment of their applicative success in prediction and control over nature.'5 The legitimative process at issue thus relies on an appropriate fusion Figure 10.2. The Twofold Cycle of the Legitimation of Systematizing Methodology

Theoretical

. Theoretical

oo”

controls of self-substantiation

legitimation

(retrojustification) eo

Theoretical cycle -

I

.s Systematizing

~

methodology ~y

Utilization Pragmatic

legitimation (successful implementation)

System of idated

knowledge .

Applicative cycle

I y Pragmatic controls of applicative efficacy

18 The question “Why should pragmatic success of the applications of the products of a cognitive method count as an index of its cognitive adequacy?” though seemingly straightforward, in fact plumbs hidden metaphysical depths. The complex issues that arise here are examined at considerable length in the author’s Methodological Pragmatism (Oxford, 1976), where other considerations relevant to the present discussion are also set out at greater

length,

176

THE PURSUIT OF TRUTH

of considerations of theory and praxis. It is a complex of two distinct but interlocked cycles—the theoretical cycle of cognitive coherence and the pragmatic cycle of applicative effectiveness. Only if both of these cycles dovetail properly—in both the theoretical and the applicative sectors— can the overall process be construed as providing a suitable rational legitimation for the cognitive principles at issue. The symbiotic and mutually supportive nature of the enterprise is fundamental: its structure must afford a systematic union in which both methods and theses are appropriately interlinked. Legitimation once more inheres in an appropriate sort of systematization in which both cognitive and ontological factors play a role. On this approach, the strictly intellectual aspect of explanation and understanding stands coordinate in the teleology of science with the pragmatic aspect of “control over nature.” Indeed control—throughout the range from prediction as minimal control (the adequate alignment of our own expectations) to the more elaborately modificatory change in the course of nature through effective interaction—comes to be seen in the ultimately pivotal role of the final arbiter of adequacy. This aspect of the cognitive centrality of control over nature leads us to an interventionist theory of knowledge, one that sees the issue of monitoring the adequacy of our theorizing to reside ultimately on the side of efficacy in application. +6

The Evolutionary Dimension of System Development But how can it be shown that the specifically coherentist approach to cognitive systematization meets the demands of a pragmatic-efficacy standard of quality control? The argument here has two stages: (1) recalling our earlier thesis that the coherentist approach to cognitive systematization can assimilate the standard mechanisms of scientific method, and then (2) noting the dramatic efficacy of science vis-a-vis any even remotely available alternative candidate as a mechanism of prediction and

control over nature. One can thus invoke on coherentism’s behalf the pragmatic efficacy of science, holding that coherentist accommodation of 16 Tt is worth stressing an important aspect of the pivotal role of pragmatic efficacy in the quality control of cognitive systematization. Most of the theoretical parameters of systematizing adequacy (unity, uniformity, cohesiveness, etc.) exert an impetus in the direction of simplicity (economy, efficiency). Their operation would never in itself induce us to move without good reason from a system in hand that is relatively simple to one that is more complex. But the pursuit of applicative adequacy can reinforce the operation of completeness and comprehensiveness in counteracting the simplicity-oriented tendency of those parameters.

COHERENTIST CRITERIOLOGY OF TRUTH

Vi7

scientific method gains to the credit of a coherentist approach the dramatic success of science in realizing its conjoint purposes of explanation,

prediction, and control over nature. On this perspective, the pragmatic warrant of coherentism is seen to reside in its capacity to serve as organon

of scientific reasoning. But how can this capacity be made manifest? The picture in figure 10.2 of the interlocked circles of the theoretical and applicative validation of cognitive systems portrays the process of system validation in the essentially timeless terms customary in epistemological discussions. This static view of system validation needs to be supplemented—indeed corrected—by considering the issue in its temporal and developmental aspect. The atemporal relationships of probative justification must be augmented by examining the justificatory bearing of the historical dynamics of the matter—the evolutionary process of system development. After all, the articulation of cognitive systems is-a matter of historical development, of repeated efforts at improvements in systematizing in the light of trial and error. We are faced with a fundamentally repetitive process of the successive revision and sophistication of our ventures at cognitive systematization, a process that produces by way of iterative elaboration an increasingly satisfactory system, one that is more and more adequate in its internal articulation or effective in its external applicability. There are iterative cycles of tentative systematizations followed by resystematizations in the light of the feedback provided by its utilization for theoretical application and practical implementation. This dynamic process is depicted in figure 10.3, which presents the cycle at issue in a historical perspective, regarding it as a developmentally iterative feedback process. What is at stake is not just retrospective reappraisal in the theoretical order of justification but an actual revision or improvement in the temporal order of development. This sequential and developmental process of historical mutation and optimal selection assures a growing conformation between our systematizing endeavorsand “the real world.” In the final analysis our systematized cognition fits the world for the same reasons that our eating habits do: both are the product of an evolutionary course of selective development. It is this evolutionary process that assures the adaequatio ad rem Figure 10.3. The Feedback Cycle of Legitimation j Presumptive

Revision of

“trial assumption”

cognitive machinery

of certain cognitive machinery

36

Appraisal of success

pen

in the light of this appraisal

178

.

THE PURSUIT OF TRUTH

of our system-based claims to knowledge. The legitimating process at issue here is not only a matter of a static cycle of relationships in the probative order of rational legitimation, it also reflects a temporal and developmental process of successive cyclic iterations where all the component elements become more and more attuned to one another and pressed into smoother mutual conformation. This evolutionary development of intellectual methodologies proceeds by rational selection. As changes come to be entertained (within the society), it transpires that one “works out for the better” relative to another in terms of its fitness to survive because it answers better to the socially determined purposes of the group. Just what does “better” mean here? This carries us back to the Darwinian perspective. Such a legitimation needs a standard of survivalistic fitness. And this normative standard is provided by considerations of theoretical adequacy and applicative practice and is inherent in the use to which conceptual instrumentalities are put in the rational conduct of our cognitive dnd practical affairs. Our legitimation of the standard probative mechanisms of inquiry regarding factual matters began with the factor of pragmatic success and subsequently transmuted this into an issue of Darwinian survival. As the discussion has already foreshadowed at many points, it is clearly the method ofscientific inquiry that has carried the day here. The mechanisms of scientific reasoning clearly represent the most developed and sophisticated of our probative methods. No elaborate argumentation is necessary to establish the all-too-evident fact that science has come out on top in the competition of rational selection with respect to alternative processes for substantiating and explaining our factual claims. The prominent role of the standard parameters of systematizationin the framework of scientificthought thus reflects a crucial aspect of their legitimation. The methodological directives that revolve about the ideal of systematicity in its regulative role (“Of otherwise coeligible alternatives, choose the simplest!”; “Whenever possible invoke a uniform principle of expla-_ nation or prediction!” etc.) form an essential part of the methodological framework (the procedural organon) of a science. Experience has.shown these methodological principles.to be rooted in the functional objectives of the enterprise, being such as to conduce efficiently to its realization of its purposes. We have every reason to think that an abandonment of | these regulative principles, while not necessarily spelling an abandonment of science as such, would make hopelessly more difficult and problematic the realization of its traditional goals of affording intellectual and physical control over nature. Considerations of functional efficiency—of — economy of thought and praxis—militate decisively on behalf of the traditional principles of scientific systematization. The key considerations are effectiveness and efficiency, purposive ad-

COHERENTIST CRITERIOLOGY OF TRUTH

179

equacy and functional economy, acceptability of product and workability of procedure. (And systematicity is, of course, an ideal vehicle here in its stress on simplicity, regularity, uniformity, etc., all of which have to do with the minimizing of unnecessary complications and the pursuit of intellectual parsimony.) A quasi-economic dialectic of costs and benefits is operative here. And the question of system choice can ultimately be seen as a matter of survival of the fittest, with fitness ultimately assessed in terms of the theoretical and practical objectives of the rational enterprise. Legitimation is thus evidenced by the fact of survival through historical vicissitudes. To be sure, there are a variety of approaches to the problem of system-

atizing “how things work in the world.” The examples of such occult cognitive frameworks as those of numerology (with its benign ratios), astrology (with its astral influences), and black magic (with its mystic forces) indicate that alternative explanatory frameworks exist and that these can have very diverse degrees of merit. Thus the orthodox scientific approach to cognitive systematization is simply one alternative among others, and it does not have an irrevocably absolute foothold on the very constitution of the human intellect, nor indeed any sort of abstract justification by purely “general principles.” Its legitimation is not a priori and absolute, but a posteriori and experientially determined. It is not difficult to give examples of the operation of Darwinian processes in the domain of the instrumentalities of cognitive systematization. The intellectual landscape of human history is littered with the skeletal remains of the extinct dinosaurs of this sphere. Examples of such defunct methods for the acquisition and explanatory utilization of information include astrology, numerology, oracles, dream interpretation, the reading

of tea leaves or the entrails of birds, animism, the teleological physics of the pre-Socratics, and so on. There is nothing intrinsically absurd or contemptible about such unorthodox cognitive programs; even the most occult of them have a long and not wholly unsuccessful history. (Think, for example, of the long history of numerological explanation from Pythagoreanism, through Platonism, to the medieval Arabs, down to Kepler in the Renaissance.) But there can be no question at this historical juncture that science has won the evolutionary struggle among various modes of methods of cognitive procedure, and this more than anything else makes it manifest that the inherent coherentism of the orthodox scientific approach to cognitive systematization satisfies the requirement of pragmatic efficacy. It makes perfectly good sense to ask, “Why should our scientific deliberations proceed in the usual way—with reference to the pursuit of systematicity, and so on?” And it is possible to answer this question along two seemingly divergent routes:

180

.

THE PURSUIT OF TRUTH

1. the pragmatic route: it is efficient, effective, successful; “it works,” etc.

2. the theoretical route: it is rationally cogent, cognitively satisfying, “aesthetically pleasing,”!” conceptually economical, etc.

But the divergence here is only apparent, for Darwinian considerations assure that in the course of time the two maintain a condition of conver-

gent conformity. The merit of entrenched cognitive tools lies in their (presumably) having established themselves in open competition with their rivals. It has come to be shown before the tribunal of bitter experience—through the historical vagaries of a Darwinian process of selection—that the accepted methods work out most effectively in actual practice vis-a-vis other tried alternatives. The conformity between the regulative presumptions and other methodological instrumentalities of inquiry and its results is not guaranteed by a preestablished harmony. Nor is it just a matter of contingent good luck. It is the product of an evolutionary pressure that assures the conformation of our systematizing efforts and the real world under trial and error, subject to controlling constraint of applicative success (pragmatic efficacy). The evolutionary process assures the due coordination of our cognitive systematizing with the objective workings of a nature that is inher-

ently indifferent to our purposesandbeliefs.

,

The Contingency of the Governing Principles of Systematization At bottom, the legitimation of a coherentist approach to cognitive systematization turns on the implicational thesis: , If objective knowledge of nature can be obtained, then the coherentist methodology can afford a way to secure it.

And this thesis is itselfargued for through a combination of two premises: 1. objective knowledge = a body of appropriately systematized information (i.e., one systematized in line with the general strategy of coherentism) 2. If systematized information can be obtained by us at all, then the coherentist methodology can afford a way to secure it.-

Now (2) here is an a priori, necessitarian thesis, obtainable by conceptual analysis alone. Coherentist methodology based on the systematization of experience—with its epistemic canonization of the parameters of systematicity—is in its very nature the most direct route to cognitive sys-

tematization. To be sure, we cannot say a priori whether our information 17 On this terminology, see Michael Polanyi, Personal Knowledge (New York, 1964).

COHERENTIST CRITERIOLOGY OF TRUTH

181

(our purported knowledge) about the world can be effectively systematized. But we can make do with whatever comfort is afforded by the truism that only by running in the race do we stand a chance of winning it. It is this kind of a priori assurance that is at issue in the purely theoretical “this or nothing” argument that validates thesis (2). But thesis (1) has a very different footing. Its contention that properly systematized information is thereby such as to be true of the world is simply the Hegelian Inversion (as we have called it) all over again. And this thesis is certainly not a matter of a priori necessity. It is perfectly possible for the world—or, rather, for a world—to be such that our way of systematizing information about it (via a pivotal reliance on the usual parameters of systematicity such as simplicity, uniformity, and the rest) would not yield an appropriate cognitive modus operandi. It is just here—with the validation of (1)—that the complex dual cycle of theoretical and pragmatic retrovalidation comes into play. It is this crucial “closing of the cycles” that—on our theory—is needed to substantiate the status of our systematized information as objective knowledge. Validation here roots in the process of Darwinian conformation considered in the preceding section. It is clear that if the legitimation of the regulative principles of cognitive systematization is construed as proceeding along such pragmatic/evolutionary lines, then these principles come to stand on an ultimately factual footing—one that is a posteriori and contingent. The first principles by whose means we constitute our factual knowledge of nature (uniformity, simplicity, and the other parameters of systematization that make up our guidelines to plausibility) are themselves ultimately of an a posteriori and factual standing in point of their controlling force. Seemingly serving merely in the role of inputs to inquiry, they emerge in the final analysis as its products as well and accordingly have a contingent rather than a necessary status. For then the legitimation of our methodological guidelines to cognitive systematization in the factual domain is ultimately not a matter of abstract theoretical principle but one of experience. Our first principles of cognitive systematization have no claims to inherent necessity. Conceivably things might have eventuated differently—even as concerns the seemingly a priori “first principles of our knowledge.” (Why didn’t they so eventuate? The question “Why these principles rather than something else?” is not illegitimate—it is indeed answerable in terms of the overall double circle of methodological legitimation.) Are such first principles a priori and analytic (a part of the conceptual schema of our science), or are they a posteriori and synthetic (a product of scientific inquiry)? This question now looks naive, for they are both. The idea of a feedback cycle of evolutionary legitimation indicates that

182

:

THE PURSUIT OF TRUTH

we are ill advised to put the question in terms of a logically tidy yes-orno. Such first principles are thus “first” only in the first analysis. Their theory-internal absoluteness is deceptive—it represents but a single phase within the historical dialectic of evolutionary legitimation. They do not mark the dead end of a ne plus ultra. To be sure, all this is to say no more than that circumstances could arise in which even those very fundamental first principles that define for us the very idea of the intelligibility of nature might have to be given up. In other possible worlds, reliance on simplicity, uniformity, and so forth could conceivably prove misleading and cognitively counterproductive. But to concede the possibility is not, of course, to grant the likelihood— let alone the reality—of the matter as regards our actual world. The first principles at issue are so integral a component of our rationality that we cannot even conceive of any rationality that dispenses with them: we can conceive that they might have to be abandoned, but not how."® The very circumstance that these principles are in theory vulnerable is a source of their strength in fact. They have been tried and tested in the history of science—tried long, hard, and often—and yet not found wanting. They are founded onasolid basis of trial in the harsh court of historical reality. Forming, as they do, an integral component of the cognitive methods that have evolved over the course of time, it can be said that for them—as for all our other strictly methodological resources—die Weltgeschichte ist das Weltgericht. A quasi-Lamarckian process of rational selection is the key to quality control in the cognitive domain.° 18 On this fact-ladenness of the fundamental ideas by which our very conception of nature is itself framed, see “A Critique of Pure Analysis,” chap. 6 of the author's Primacy of Practice (Oxford, 1973). 18 This chapter encorporates ideas treated in author’s Coherence Theory of Truth (Oxford, 1973) and A Useful Inheritance (Savage, Md., 1990). Some of its key themes are further developed elsewhere: plausibility, in Plausible Reasoning (Assen, 1976); systematicity, in Cognitive Systematization(Oxford, 1979); retrospective justification, in Methodological Pragmatism (Oxford, 1977); cognitive efficiency, in Cognitive Economy (Pittsburgh, 1989).

Eleven The Pivotal Epistemic Role of Postulation Communicative Parallax

Real things have hidden depths—they are cognitively opaque. And this factis altogether pivotal for understanding how wethink and communicate about the world we livein. Any particular thing—the moon, for example—is such that two related but critically different versions of it can be contemplated: (1) the moon, the actual moon as it “really” is, and (2) the moon as somebody (you or I or the Babylonians) conceives of it. Now it is virtually always the former item—the thing itself—that we intend to communicate or think (i.e., selfcommunicate) about: the thing as it is, and not the thing as somebody conceives of it. We take the stance that all these different moon versions correspond to one selfsame moon. Yet we cannot but recognize the justice of Kant’s teaching that the “I think” (I maintain, assert, etc.) is an ever-present implicit accompaniment of every claim or contention that we make. No matter how loudly you shout “P,” all that we can ever secure unproblematically on the basis of your declarative action is that you maintain that P. This factor of cognitive reflexivity dogs our every assertion. And it reflects the unavoidable prospect of “getting it wrong,” since the gulf between “I think that P” and “P is true” can never quite be reduced to nothingness. Ambitious intentions or pretentions to the contrary notwithstanding, all that one can ever actually manage to achieve in one’s declarations is to deliver information about item (2)—to convey what one thinks or conceives to be so. I can readily distinguish the featues of (what I take to be) “the real moon” from those of “the moon as you conceive of it,” but I cannot distinguish them from those of “the moon as I conceive of it.” And when I maintain “The moon is roughly spherical,” all that I have successfully managed to deliver to you by way of actual information is, “Rescher maintains that the moon is roughly spherical.” And there is nothing that can be done to alter this circumstance—it does not matter how loudly I bang on the table. If you bind me by the injunction “Tell me something about the Eiffel Tower, but please don’t put before me your beliefs or convictions regarding it; just give me facts about the thing itself, rather than presenting any parts of your conception of it!” you condemn me to _the silence of the Lockean je ne sais quoi.

184

.

THE PURSUIT OF TRUTH

Let.us employ the phrase communicative parallax to indicate this circumstance that throughout one’s discourse about things, one always intends to convey information about “the actual thing itself” but only manages to disclose facets of one’s conception of the thing. With optical parallax, where you see something to be depends on where you stand in regard to it. With communicative parallax, how you see something to be depends on how you stand in regard to it. This parallax reflects an inevitable slippage between intention and accomplishment in all fact-stating or fact-purporting discourse. Now it is important to realize that it is not the case that two different things are at issue where we talk of parallax. It would be a grave mistake of illicit hypostatization to reify “the X as we see it” into a thing distinct from the real X. “Harry as I picture him” may be very unlike, and quite different from, “the real Harry,” but it is still the real Harry that is the intended object of my conception, however little it may do him justice. The world is not populated by many Harrys—the real one and mine and yours, and so forth, each answering to our respective (distinct) conceptions of him. “Harry as I conceive of him” may well not exist as such, but this does not block my conception from having Harry—the real Harry— as its object. The star we take ourselves to see (the star “as we see it”) is not a different entity—a thing distinct from “the real star.” It is the real star but seen as somehow displaced from its true position in the scheme of things. And much the same holds where communicative parallax is at issue. Here “the thing itself” that contrasts with “the thing as we conceive of it” is not a different thing: it is the very selfsame thing that our conception intends to capture. The distinction between our moon, say, and that of the Babylonians represents no difference in object. We have to espouse the view that only one thing—the moon itself—is at issue,

which is, as it were, “seen differently” by different individuals. To bring it on the stage of discussion is not to multiply entities by invoking the membership of a cognitively inaccessible transcendental realm. but simply to employ a distinguishing contrast to give convenient expression to the crucial fact of the potential inadequacy of our conceptions. To speak of parallaxis misleadingin one way. For we know what allowances to make for astronomical parallax—how to “calculate out” the displacement at issue. But we do not—and in the nature of the thing cannot—know what allowances to make for communicative parallax. We are never in a position to realize how our current conception of a thing is inadequate—we can only realize that it may well be so on the basis of retrospect. 1 While this account of communicative parallax is framed with a view to particulars (such as the moon or the Great Pyramid), much of it will also hold, mutatis mutandis, for diffused

THE EPISTEMIC ROLE OF POSTULATION

185

The Intentionality of Interpersonal Communication There is nothing unfortunate or regrettable about the fact of communicative parallax. Quite the reverse, it serves an important and positive function. For it is crucial to the achievement of intersubjective objectivity in discourse. After all, the teleology of language use is nothing mysterious and occult. Language is primarily a purposive instrument whose cardinal aim is the transmission of information for the sake of implementation in action. Language is designed to afford us resources for information storage and mechanisms for the interpersonal exchange of information needed for the coherent pursuit of individual goals and the coordination of effort in the pursuit of common goals. And only the accomplishment-transcending intention to discuss “the thing itself” makes communication possible. If my discourse were directed at my moon conception, we could never lock communicative horns. Two different objects would be at issue. Our discussions would be “out of touch”: the prospect of agreement and disagreement would vanish, and the prospect of interpersonal communication about a common object would vanish with it. Moreover, any transhistorical comparability of objects would go by the board. The sun and moon of the Babylonian priest-astrologers would be as disjoint from ours as are our respective deities. Communicative parallax would be overcome, but at an awesome price—communicative anarchy and immobilization. The exact configuration of information that I myself have about a thing at first hand is always something personal and idiosyncratic—based upon the contingencies of what I happen to have experienced about it and what I happen to have gathered about the experience of others. In making objective assertions about something, it is thus crucial that I intend to discuss “the thing itself” rather than “the thing just precisely as I conceive of it” relative to the body of information I have about it. Only the former is something that somebody else can get hold of; the latter certainly is not. The imputational move beyond the data at hand is indispensably demanded by that step into the domain of the publicly accessible objects, in whose absence interpersonal communication about a shared world becomes impossible. thing-kinds (water or copper) and thing-types (books or cows). What alone matters throughout is that there must be some objective item at the pre- or subtheoretical level of which

we can say that different accounts give different accounts of it. (This is why the language of everyday life is critical.) The approach does not work for creatures of theory, however. We cannot say of Democritean atoms that Rutherford’s theory is giving an alternative account

of them. Creatures of theory cannot exist outside the confines of their particular theoretical habitat; subtheoretical things, however, can survive changes in the theoretical environment.

186

THE PURSUIT OF TRUTH

We are able to say something about the (real) Sphinx because of our submission to a fundamental communicative convention or “social contract” to the effect that we intend (“mean”) to talk about it—the very thing itself as it really is—our own private conception of it notwithstanding. We arrive at the standard policy that prevails with respect to all communicative discourse of letting the language we use, rather than whatever specific informative aims we may actually have in mind on particular occasions, be the decisive factor with regard to the things at issue in our discourse. When I speak about the Sphinx (even though I do so on the basis of my own conception of what is involved here), I will nevertheless be taken to be discussing the real Sphinx in virtue of the basic conventionalized intention at issue with regard to the operation of referring terms.

Successful communication requires not only common concepts but common topics—shared items of discussion, a common world of self-sub-

sistently real “an sich” objects basic to shared’ experience. The factor of objectivity reflects our basic commitment of a shared world as the common property of communicators. Such a commitment involves more than merely de facto intersubjective agreement. For such agreement is a matter of a posteriori discovery, while our view of the nature of things puts “the real world” on a necessary anda priori basis. This stance roots in the fundamental convention of a shared social insistence on communicating—the commitment to an objective world of real things affording the crucially requisite common focus needed for any genuine communication.

Someone might object along the following lines to this sort of emphasis on the contrast between the positive and the actual: How can recourse to “the thing itself” possibly facilitate communication? My interlocutor cannot lay hold of this any more than I can, seeing that it has features transcending anyone’s conception of it. So how can it serve to establish contact between us? .

This objection misses the point. We do indeed “lay hold of” the thing itself—not by way of information, by inquiry or investigation, but by fiat or postulation. Objective knowledge ultimately roots in the power of positing something (etwas schlechthin zu setzen) that, according to Fichte, is a characteristic power of the self; the ego of an intelligent being. What links my discourse with that of my interlocutor is.our common subscription to the a priori presumption (a defeasible presumption, to be sure) that we are talking in common about a shared item, our own possible misconceptions of it notwithstanding. Communicative parallax assures us of being in touch with one another from the very outset. And it means that no matter how much we change our mind about the nature of a thing (the moon) or type of thing (the whale), we are still dealing with exactly

THE EPISTEMIC ROLE OF POSTULATION

187

the same thing or sort of thing. It assures reidentification across theories and belief systems. Again someone might object: But surely we can get by on the basis of personal conceptions alone, without invoking the notion of “a thing itself.” My conception of a thing is something I can convey to you, given enough time. Cannot communication proceed by correlating and matching personal conceptions, without appeal to the intermediation of “the thing itself”?

But think here of the concrete practicalities. What is “enough time”? When is the match sufficient to underwrite outright identification? The cash value of our commitment to the thing itself is that it enables us to make this identification straightaway by imputation, by fiat on the basis of modest indicators, rather than on the basis of an appeal to the inductive weight of a body of evidence that is always bound to be problematic.

Communication is something we set out to do, not something we ultimately discern, with the wisdom of eventual hindsight, to have accom-

plished retrospectively. We all have different experiences, and it is unclear—to say the least— how they could ever match up without the imputation of the real world underwriting the appearances. The objectifying imputation at issue here lies at the very basis of our cognitive stance that we live and operate in a world of real and objective things. This commitment to the idea of a shared real world is crucial for communication. Its status is a priori: its existence is not something we learn of through experience. As Kant clearly saw, objective experience is possible only if the existence of such a real, objective world is presupposed at the onset rather than seen as a matter of ex post facto discovery about the nature of things. What is at issue here is thus not a matter of discovery but one of imputation. The element of community, of identity of focus, is not a matter of ultimately learning from experience but of an a priori predetermination inherent in our approach to language use. We do not infer things as being real and objective from our phenomenal data but establish our perception as authentic perception of genuine objects through the fact that these objects are given—or rather, taken—as real and objectively existing things from the first.2 Objectivity is not deduced but imputed. 2 The point is Kantian in its orientation. Kant holds that we cannot experientially learn through our perceptions about the objectivity of outer things because we can recognize our perceptions as perceptions (i.e., representations of outer things) only if these outer things

are given as such from the first (rather than being learned or inferred). As Kant summarizes his “Refutation of Idealism”: “Idealism assumed that the only immediate experience is inner experience, ard that from it we can only infer outer things—and this, moreover, only in an untrustworthy manner. . . . But on the above proof it has been shown that outer experience . is really immediate” (Immanuel Kant, Critique of Pure Reason, B276); trans. Norman Kemp Smith [London, 1929]).

188

.

THE PURSUIT OF TRUTH

A closer look at the precise character of the justification of our evidence-transcending imputations is in order. The authorizing warrant for the imputational thrust of our objective categorical judgment ultimately resides in the purposive teleology of language use—the desire for successful communication. It was suggested above that this is simply a matter of this or nothing—that if we wish to achieve answers to our questions about the world and if we wish to communicate with one another about matters of objective fact, then we simply have no alternative but to undertake such evidence-transcending commitments. But we now see that this pivotal focus on this-or-nothing considerations is by no means the whole story. For the consideration that we must proceed in this way— the fact of practical necessity that there just is no alternative if our objective is to be reached—stops well short of achieving full adequacy in its justificatory force. It does not offer us any assurance that we actually will succeed in our endeavor if we do proceed in this way; it just has it that we won't if we don’t. The issue of actual effectiveness remains untouched. And here we have no choice but to proceed experientially, by the simple strategem of trying and seeing. Practical necessity remains a matter of a priori considerations, but efficacy—actual sufficiency—will be a matter of a posteriori experience. The justification of claims of efficacy emerges through pragmatic retrojustification—a retrospective revalidation in the light of experience. The pragmatic consideration that our praxis of inquiry and communication does actually work—that we can effectively . and (by and large) successfully communicate with one another about a shared world, inquiry into whose nature and workings proceeds successfully as a communal project of investigation—is the ultimately crucial consideration that legitimates the evidence-transcending imputations built into the praxis-governing use conditions of language. . In answering the question of what justifies our recourse to the evidence-transcending imputations, we thus proceed at two levels. On the negative side, we confront the realization that we must accept them; it is

this or nothing, given the goals of the enterprise. On the positive side, we involve a pragmatic retrojustification based on the fact that our. proceedingin this way underwrites an actually effective praxis.?

Objectivity as a Requisite of Communication and Inquiry Objectivity in its particular aspect of cognitive independence—of publicity and autonomy of epistemic status—is essential to our capacity to communicate about things. The information that we may have about a thing, 3 The issues that this paragraph treats telegraphically are developed in fuller detail in the author’s Induction (Oxford, 1980).

THE EPISTEMIC ROLE OF POSTULATION

189

be it real or presumptive information, is always just that—information we lay claim to. We recognize that it varies from person to person. Our attempts at communication and inquiry are thus undergirded by the information-transcending stance that we communally inhabit a shared world of objectively existing things, a world of “real things” among which we live and into which we inquire (but about which we do have and must _presume ourselves to have only imperfect information at any and every particular stage of the cognitive venture). This is clearly not something that we learn from the course of experience. The “facts of experience” can never reveal it to us. It is something we postulate or presuppose. Its

epistemic status is not that of an empirical discovery but that of a presupposition whose ultimate justification is a transcendental argument from the very possibility of the projects of communication and inquiry as we standardly conduct them. Our commitment to an objective reality that lies behind the data at hand is indispensably demanded by any step into the domain of the publicly accessible objects essential to communal inquiry and interpersonal communication about a shared world. We could not establish communicative contact about a common objective item of discussion if our discourse were geared to the substance of our own idiosyncratic ideas and conceptions. But the objectivity at issue in our communicative discourse

is a matter of its status rather than one of its content. For the substantive content of a claim about the world in no way tells us whether it is factual or fictional. This is something we have to determine from its context: it is a matter of the frame, not of what is depicted on the canvas. The factoriented basis of our information-transmitting exchanges is provided a priori by a conventionalized intention to talk about “the real world.” This intention to take real objects to be at issue—objects as they really are, our potentially idiosyncratic conceptions of them quite aside—is fundamental because it is overriding—that is, it overrides all of our other intentions when we enter upon the communicative venture. Without this

conventionalized intention we should not be‘able to convey information (or misinformation) to one another about a shared “objective” world. This fundamental intention of objectification, the intention to discuss

“the moon itself,” regardless of how untenable one’s own ideas about it may eventually prove to be, is a basic precondition of the very possibility of communication. It is crucial to the communicative enterprise to take the egocentrism-avoiding stance that rejects all claims to a privileged status for our own conception of things. In the interests of this stance we are prepared to discount any misconceptions (our own included) about things over a very wide range indeed; we are committed to the stance that factual disagreements as to the character of things are communica_tively irrelevant within very broad limits. The incorrectness of conceptions is venial.

190

THE PURSUIT OF TRUTH

An overriding concern for dealing with the objective order of this real world is crucial for communication.If our assertoric commitments did not transcend the information we have on hand, we would never be able to

get in touch with others about a shared objective world. No claim is made for the primacy of our conceptions, or for the correctness of our conceptions, or even for the mere agreement of our conceptions with those of others. The fundamental intention to discuss “the thing itself” predominates and overrides any mere dealing with the thing as-we conceive it to be. Certainly, that reference to objectively real things at work in our discourse does not contemplate a peculiar sort of thing—a new ontological category of “things-in-themselves.” It is simply a shorthand formula for a certain communicative presumption or imputation rooted in an a

priori commitment to the idea of a commonality of objective focus—a commitment that is allowed to stand unless circumstances arise to render this untenable. We could never establish communicative contact about a common objective item of discussion if our discourse were geared to the things as conceived of in terms of our own specific information about them.

The commitment to objectivity is basic to any prospect of our discourse with one another about a shared world of real things, to which none of us is in a position to claim privileged access. This commitment establishes a need to distance ourselves from things, that is, to recognize the prospect of a discrepancy between our (potentially idiosyncratic) conceptions of. things and the true character of these things as they exist objectively in the real world. The ever-present contrast between “the thing as we view it” and “the thing as it is” is the mechanism by which this crucially important distancing is accomplished. And maintaining this stance means that we are never entitled to claim to have exhausted a thing au fond in cognitive regards, to have managed to bring it wholly within our epistemic grasp. For to make this claim would, in effect, be to identify “the thing at issue” purely in.terms of “our own conception of it,” an identification that would effectively remove the former item (the thing itself) from the stage of consideration as an independent entity in its own right, by endowing our (own, current) conception with decisively determinative force. And this would lead straightaway to the unacceptable result of a cognitive solipsism that would preclude reference to intersubjectively identifiably particulars and would.thus block the possibility of interpersonal communication and communal inquiry.

The Communicative irrelevance of Conceptions Any pretentions to the predominance, let alone the correctness, of our own conceptions regarding the furniture of this realm must be put aside

THE EPISTEMIC ROLE OF POSTULATION

191

in the context of communication.If we were to set up our own conception as somehow definitive and decisive, we would at once erect a grave im-

pediment to the prospect of successful communication with one another. Communication could then proceed only retrospectively with the wisdom of hindsight. It would be realized only in the implausible case that extensive exchange indicates that there has been an identity of conceptions all along. We would then learn only by experience—at the end of a long process of wholly tentative and provisional exchange. And we would always stand on very shaky ground. For no matter how far we push our inquiry into the issue of an identity of conceptions, the prospect of a divergence lying just around the corner—waiting to be discovered if only we pursued the matter just a bit further—can never be precluded. One could never advance the issue of identity of focus past the status of a more or less well-grounded assumption. And then any so-called communication is no longer an exchange of information buta tissue of frail conjectures. The communicative enterprise would become a vast inductive project—a complex exercise in theory building, leading tentatively and provisionally toward something that, in fact, the imputational groundwork of our language enables us to presuppose from the very outset.* The fact that we need not agree on our conceptions of things means, a fortiori, that we need not be correct in our conceptions of things to communicate successfully about them. This points, in part, to the trivial fact that I need not agree with what you are saying to understand you. But it points also, more important, to the consideration that my having a conception of a thing massively different from yours will not prevent me from taking you to be talking about the same thing that I have in mind. Objectivity and referential commonality offocus are matters of initial presumption or presupposition. The issue here is not with what is understood but with what is to be understood (by anybody) in terms of certain generalized and communicative intentions. (The issue here is not one of meaning but only of meaningfulness.) Our concept of an objectively real thing is accordingly such that a thing is a fixed point, a stable center around which communication revolves,

the invariant focus of potentially diverse conceptions. What is to be determinative, decisive, definitive, (and so forth) of the things at issue in my discourse is not my conception or yours or indeed anyone’s conception at all. The conventionalized intention discussed above means that a coincidence of conceptions is not decisive for the possibility of communication. Your statements about a thing will convey something to me, even if my conception of it is altogether different from yours. To communicate we need not think alike—we need not take ourselves to share _

4 The justification of such imputations is treated more fully in chap. 9 of the author's Induction.

192

.

THE PURSUIT OF TRUTH

views of the world, but only to take the stance that we share theworld being discussed. In communication regarding things, we must be able to exchange information about them with our contemporaries and to transmit information about them to our successors. And we must be in a position to do this in the face of the presumption that their conceptions of things are not only radically different from ours but conceivably also rightly different. What is at issue here is not the commonplace that we do not know everything about anything. Rather, the key consideration is the more interesting thesis that it is a crucial precondition of the possibility of successful communication about things that we must avoid laying any claim either to the completeness or even to the ultimate correctness of our own conceptions of any of the things at issue. (This fallibilism is the concession that cognitivism makes to skepticism.) It is crucial that the mechanisms of human communication should lie

within the domain of human power—they could not otherwise be what they are. Now with respect to the meanings of words, this condition is satisfied because this is something that we ourselves fix by custom or by fiat. But the correctness of conceptions is not simply a matter of human discretion—it is something that lies outside the sphere of our effective control. For a “correct conception” is akin to Spinoza’s true idea, of which he stipulates that it must “agree with its object’*—in circumstances | where this issue of agreement may well elude us. (Man proposes but does not dispose with respect to this matter of idea/actuality coordination.) We do, no doubt, purport our conceptions to be correct, but whether this is indeed so is something we cannot tell with assurance until “all the returns are in’—that is, never. This fact renders it critically important that (and understandable why) conceptions are communicatively irrelevant. Our discourse reflects our conceptions and perhaps conveysthem, but it is not substantive about them. The conception of a thing may be the vehicle ofthought,but it is never the determinant of reference. By their very nature, conceptions are too personal—and thus potentially too idiosyncratic—for our communicative needs. For communication, interpersonal and public instrumentalities are indispensably requisite..And language affords this desideratum. It provides the apparatus by which the identity of the referents of our dis-

course becomes fixed, however imperfectly we ourselves perceive their nature.

How do we really know that Anaximander was talking about our’sun? He is not here to tell us. He did not leave elaborate discussion about his 5 Benedictus de Spinoza, Ethics, bk. 1, axiom 6.

THE EPISTEMIC ROLE OF POSTULATION

193

aims and purposes. How can we be so confident of what he meant to talk about? The answer is straightforward. That he is to be taken to talk about our sun turns, in the final analysis, on two very general issues in which Anaximander himself plays little if any role at all: (1) our subscription to certain generalized principles of interpretation with respect to the Greek language, and (2) the conventionalized subscription by us and ascription to other language users in general of certain fundamental communicative policies and intentions. In the face of appropriate functional equivalences, we allow neither a difference in language nor a difference of thought worlds to block an identity of reference.® The pivotal intention to communicate about a common object—resigning any and all claims to regard our own conceptions of it as definitive (decisive)}—is the indispensable foundation of all communication. And this intention is not something personal and idiosyncratic, a biographical aspect of certain particular minds—it is a shared feature of “social mind,”

built into the use of language as a publicly available communicative resource. The wider social perspective is crucial. In subscribing to the conventionalized intention at issue, we sink our own point of view in the interests of entering into the wider community of fellow communicators. Only by admitting the potential distortion of one’s own conceptions of things through communicative parallax can one manage to reach across the gulf of divergent conceptions so as to get into communicative touch with one another. We thus arrive at the key idea that these lines of thought contribute to the present deliberations. True enough, cognitive change carries conceptual change in its wake. But nevertheless—and this point is crucial—we have an ongoing commitment to a manifold of objective things, a world of objects that as such are themselves impervious to conceptual and cognitive change. This commitment is built into the very ground rules that govern our use of language and embody our determination to maintain the picture of a relatively stable world amid the ever-changing panorama of cognitive world-pictures. The idea of a continuing succession of the different states of science is unavoidably linked to a pre- or subscientific view of an ongoing real world in which we live and work, a world portrayed rather more stably in the lingua franca of everyday-life communication and populated by shared things whose stability amid cognitive change is something rather postulated than learned. This postulation reflects the realistic stance that the things we encounter in experience are the subject and not the product of our inquiry. ® See the interesting paper by Michael E. Levin, “On Theory-Change and MeaningChange,” Philosophy of Science 46 (1979): 407-24.

194

.

THE PURSUIT OF TRUTH

Building Up Communicative Trust: An Economic Approach _ In cognitive contexts, communication is the process of conveying infor-

mation from one person to another—or at any rate, endeavoring to do so. Its ways and means are governed by economic considerations to an extent that is not generally appreciated. The exchange of information is clearly a mutual-benefit process. It is far easier, cheaper, and- more convenient for people to get information by sharing than by themselves having to undertake the often laborious inquiries and researches needed to develop it de novo. To convey information—to communicate efficiently—we need not think alike. But we must, at least to some extent, trust one another. Some credence, however modest, must be given by each of us to the idea

that the others know what they are doing. The process through which mutual trust in matters of information de-

velopment and management is built up among people cries out for explanation by means of an economic analogy that trades on the dual meaning of the idea of credit. And this situation obtains not just in the management of information in natural science but in many other settings as well,

preeminently including the information we use in everyday-life situations. Moreover, we constantly rely upon experts, continually placing reliance on doctors, lawyers, architects, and other professionals. They too must so perform as to establish credit, not just as individuals but, even more crucially, for their profession as a whole.” Again, a not dissimilar story holds for our information-generating technology—for telescopes, microscopes, computing machinery, and so on. Here too we have to do with presumptions rather than guarantees. We initially extend some credit because we simply must, since they are our only means for a close look-at the moon, at microbes, and so on. But

subsequently we increase their credit limit (after beginning with blind trust) because we eventually learn, with the wisdom of hindsight, that it was quite appropriate for us to proceed in this way in the first place. As we proceed, the course of experience indicates, retrospectively as it were, that we were justified in deeming them creditworthy. To be sure, the risk of deception and error is present throughout our inquiries: our cognitive instruments, like all other instruments, are never

failproof. Still, a general policy of judicious trust is eminently cost effective. In inquiring, we cannot investigate everything; we have to start 7 See H. M. Vollmer and D. L. Mills, eds., Professionalization (Englewood Cliffs, N-J:, 1966). The sort of credit at issue, once earned, is generally safeguarded and maintained by institutional means, such as licensing procedures, training qualifications, professional societies, and codes of professional practice.

THE EPISTEMIC ROLE OF POSTULATION

195

somewhere and invest credence in something. But of course our trust need not be blind. Initially bestowed on a basis of mere hunch or inclination, it can eventually be tested and can come to be justified with the wisdom of hindsight. And this process of testing can in due course put the comforting reassurance of retrospective validation at our disposal. To derive benefit from the declarations of others, we must (1) listen to them, pay them heed;:(2) interpret (decode) what they say; (8) extend them cognitive credit. To be sure, none of these steps is cost free. Each

exacts from us an outlay of resources in point of time, effort, attention, and risk of error. All procedures for the acquisition of information—listening; watching, reading, and so on—involve expenditures of some sort. (No school student is ever wholly oblivious to the fact that learning can be painful.) And whether this outlay is warranted depends on the correlative advantages—preeminently including the cognitive benefits of acquired information. ° Suppose I tell you, “The cat is on the mat.” What information-do you now actually have? Is it (1) The cat is on the mat? Or is it (2) Rescher thinks (believes) that the cat is on the mat? Or is it merely (8) Rescher says that the cat is on the mat? In the circumstances under hypothesis, only the last item is wholly unproblematic. And it is clear that one cannot manage to get from (3) to (2) unless one adds something like: (4) When Rescher says something (in a serious tone of voice), he generally believes it. Moreover, you certainly cannot get from (2) to (1) unless you credit me with veracity and trustworthiness and impute to me a penchant for truth—that is, unless you also accept: (5) When Rescher believes something to be so (in such matters as cats and mats, at any rate), he is generally right. The entire process of communication—of deriving substantive information from the declarations of others—involves trust. But what validates this? To answer this question, it is best to look at the issue in economic perspective.

A communicating community is a sort of marketplace with offerers and takers, sellers and buyers. In accepting the declarations of others at their informative face value, we extend them credit, as it were. The prospect of informative communication is predicated on such principles as (1) Concede a presumption of veracity to the assertions of others, at any rate until such time as they prove themselves unworthy of credit; and (2) In communicative contexts, regard others as candid, truthful, accurate, and

the like, until proven otherwise. The rationale of such principles of epistemic procedure is largely or wholly economic. For here, as elsewhere, it is ultimately on the basis of considerations of cost-effectiveness that we decide how much credit to extend.

196

.

THE PURSUIT OF TRUTH

Afterall, why do we credit people with communicative capacity—with the power to provide information? Note that the purported fact, “When

he utters ‘the cat is on the mat,’ he is engaged in asserting that the cat is on the mat” represents a belief of ours, or at any rate a supposition on our part. We make this supposition initially in desperation, as it were, because it provides the only feasible way for us to derive any benefit from the content of someone’s assertions, but ultimately because we eventually accumulate evidence that indicates (with the wisdom of hindsight) that this supposition was well advised (warranted). In extracting information from the declarations of others, we rely on a whole host of working assumptions: (1) people mean to say what they apparently say (what we take them to be saying); (2) people believe what they say; (3) people have good grounds for their beliefs (i.e., there are such grounds, and they have cognizance of them). Essentially the same justification obtains throughout: unless we enter into such communicative presumptions, we deprive ourselves of any chance to extract information from the declarations of others. On this basis, we are rationally well advised (for example) to treat their declarations as epistemically innocent until proven otherwise, exactly because this is the most cost-effective thing to do. Our communicative procedures are motivated—-and justified—by the essentially profit-seeking objective of extracting the maximum benefit from our information-oriented interactions. Communication is accordingly predicated on conceding and maintain- “ ing credibility. Communication too is a commercial system of sorts. Credit is extended, drawn upon, and enlarged. And with communicative

and financial credit alike, one could not build up credit (prove oneself creditworthy) unless given some credit by somebody in the first place. For credit to be obtainable at all, there has to be an initial presumption that one is creditworthy. Clearly, such a presumption of innocent until proven guilty (i.e., fault free until shown to be otherwise) can be defeated; one can of course prove oneself to be unworthy of credit or credence. But initially it must be made. Several recent studies illuminate the extent to which we actually depend upon othersin our beliefs. The experiments of Solomon Asch have dramatized people’s tendency to conform to erroneous publicjudgments on matters where they would never make mistakes by themselves.° His 8 For a penetrating study of these and similarissues, see John Sabini andMaury Silver, Moralities ofEveryday Life (Oxford, 1982), esp. chap. 4. ® Specifically, Solomon Asch found that in certain situations of interactive estimation,

“whereas the judgments were virtually free of error under control conditions, one-third of the minority estimates were distorted toward the majority.” See his “Studies of Indepen-

THE EPISTEMIC ROLE OF POSTULATION

197

subjects had only to specify which of three lines was closest in length to a given line. People made this judgment unerringly, except when they knew that all the others who were asked the same question concurred in giving a different answer.!° Commenting on Asch’s experiments, Sabini and Silver report: “All (or nearly all) subjects reacted with signs of tension and confusion. Roughly one-third of the judgments subjects made were in error. Nearly 80 percent of the subjects gave the obviously wrong answer on at least one trial. The perception that a few otherpeople made an absurd judgment of a clear, unambiguous physical matter was a very troubling experience, sufficient to cause doubt, and in some cases conformity.”"' Such experiments actually reveal (in their own dishonest way) the extent to which people incline to trust others. A recent study of American juries arrived at very similar findings.’ On examining more than 250 jury deliberations, the investigators found that in no case was a hung jury caused by a single dissenter. Unless someone who disagreed with the majority found support by at least two others, the dissenters generally relaxed their reservations and came around to the majority view.

The rationale for this impetus to belief convergence is validated by sound economic considerations. Trusting others substantially reduces interaction costs. And knowing whom one can trust is worth a great deal. Outsiders who come as strangers into an established social framework generally have to pay for the benefit of learning which agents are trustworthy—and generally find this information well worth paying for. Such considerations militate for a universally advantageous modus operandi under whose aegis people can trust their fellows in a setting of communal cooperation. Such cooperationina conformity to communal

standards evolves because what is in the interests of most is, in most cases, in the interests of each. . As these deliberations indicate, our cognitive practices ‘of trust and

presumption are undergirded byajustificatory rationale whose nature is fundamentally economic. For what is at issue throughout is a system of procedure that assures for each participant the prospect of realizing the greatest benefit for the least cost. Our standard cognitive policies and procedures geared to a realm ofobjective things and to a presumption that others can be trusted in communicative commerce about it are sustained by considerations of economic rationality—that is, of cost-effectiveness. dence and Conformity: I. A Minority of One against a Unanimous Majority,” in Psychological Monographs: General and Applied, no. 70 (1956).

0 Tbid., p. 69.

Ut Sabini and Silver, Moralities of Everyday Life, pp. 84-85. 12 Harry Kalven, Jr., and Hans Zeisel, The American Jury (Chicago, 1966).

198

THE PURSUIT OF TRUTH

The Economic Rationale of Communication

We know that various highly convenient principles of knowledge production are simply false: What seems to be, is.

What people say is true.

;

The simplest patterns that fit the data are actually correct. The most adequate currently available theory will work out.

We realize full well that such generalizations do not hold, however nice it would be if they did. Nevertheless we accept the theses at issue as principles of presumption. We follow the metarule “In the absence of concrete indications to the contrary, proceed as though such principles were true.” Our standard cognitive practices incorporate a host of fundamental presumptions of initial credibility, in the absence of concrete

evidence to the contrary:

:

Believe in your own senses. Accept at face value the declarations of other people (in the absence of any counterindications and in the absence of any specific evidence undermining their generic trustworthiness). Trust in the reliability of established cognitive aids and instruments (tele-" scopes, calculating machines, reference works, logarithmic tables, etc.) in

the absence of any specific indications to the contrary. Accept the declarations of established experts and authorities within the area of their expertise (again, absent counterindications).

The justification of these presumptions is not the factual one of the substantive generalization “In proceeding in this way, you will come at correct information and will not fall into error.” Rather, it is the-methodological justification “In proceeding in this way, you will efficiently foster the interests of the cognitive enterprise; the gains and benefits will, on the whole, outweigh the losses and costs.” Such principles of presumption characterize the way in which rational agents transact their cognitive business. Yet we adopt such practices not

because we can somehow establish their validity but because the costbenefit advantage of adopting them is so substantial. The justification of trust in our senses, in our fellow inquirers, and in our cognitive mechanisms ultimately rests on considerations of economic rationality. And this sort of situation prevails in many other contexts. For example, the rationale of reputations for ability, as well as those for reliability, lies inthe

THE EPISTEMIC ROLE OF POSTULATION

199

cost-effectiveness of this resource in contexts of hiring, allocating one’s reading time, and so on.¥

It is clear that all such cognitive practices have a fundamentally economic rationale. They are all cost effective within the setting of the project of inquiry to which we stand committed by our place in the world’s scheme of things. They are characteristics of the cheapest (most convenient) way for us to secure the data needed to resolve our cognitive problems—to secure answers to our questions about the world we live in. Accordingly, we can make ready sense of many of the established rules of information development and management on economic grounds. By and large, they prevailbecause this is maximally cost effective in comparison with the available alternatives. To be sure, whenever we trust, matters can turn out badly. In being trustful, we take our chances (though of course initially in a cautious way). But one must always look to the other side of the coin as well. A play-safe policy of total security calls for not accepting anything, not trusting anyone. But then we are left altogether empty-handed. The quest for absolute security exacts a terrible price in terms of missed opportunities, forgone benefits, and lost chances. What recommends those

inherently risky cognitive policies of credit extension and initial trust to us is not that they offer risk-free sure bets but that, relative to the alternatives, they offer a better balance of potential benefits over potential costs. It is the fundamentally economic rationality of such cognitive practices that is their ultimate surety and warrant. If its cognitive needs and wants are strong enough, any group of mutually communicating, rational, dedicated inquirers is fated in the end to become a community of sorts, bound‘together by a shared practice of trust and cooperation, simply under the pressure of its economic advantage in the quest for knowledge. Only through cooperation based on mutual trust can we address issues whose effective resolution makes demands that are too great for any one of us alone. In the development and management of information, people are constantly impelled toward a system of collaborative social practices—an operational code of incentives and sanctions that consolidates and supports collective solidarity and mutual support. In this division of labor, trust results from what is, -to all

intents and purposes, a custom-consolidated compact to conduct their affairs in friendly collaboration. However, this cooperative upshot need not ensue from a moral dedi-

cation to the good of others and care for their interests. It can emerge for reasons of prudential self-interest alone because the relevant modes of 13 On this matter, see Thomas Sowell, Knowledge and Decisions (New York, 1980), esp.

_the discussion “Informal Relationships,” on pp. 23-30.

200

:

THE PURSUIT OF TRUTH

mutually helpful behavior—sharing, candor, and trustworthiness—are all strongly in everyone’s interest, enabling all members to draw benefit for their own purposes, the agent, too, specifically included. Cooperation emerges in such a case not from morality but from self-interested considerations of economic advantage. In science, in particular, the advantages of epistemic values such as candor, reliability, accuracy, and the like are

such that everyone's interests are well served by fostering adherence to the practices at issue. , The guiding principle ofjustification is that of cost-benefit calculation. The standard presumptions that underlie our communicative practices are emphatically not validatable as established facts. (For example, it is certainly not true that people say what they mean, except at the level of statistical generality.) But their justification becomes straightforward on economic grounds, as practices that represent the most efficient and eco-

nomical way to get the job done. For if we do not concede some credit to the declarations of others, then we lose any and all chance to derive informative profit from them, thus denying ourselves the benefit of a potentially useful resource. For the course of experience would soon teach us that, even where strangers outside the family circle are concerned, the benefits of trust, of credibility concession, generally overbalance the risks involved. In sum, we hold to the policy of believing what we are told in the absence of case-specific counterindications, because it is in our interest to do so by being highly cost effective vis-a-vis our informational aims and purposes.If playing safe were all that mattered, we would, of course, suspend judgment indefinitely. But it is simply not in our interest to do

so, since safety is not all. We adopt the policy of credencein the first instance because it is the most promising avenue toward our goals and then persist in it because we subsequently find, not that it is unfailingly successful, but that it is highly cost effective.“ Information exchange based on principles of cooperation is a process of mutual benefit, for everyone is advantaged by adopting a system of operation that maintains the best available balance of costs and benefits in this matter of creating a communally usable pool of information. Contrast two hypothetical societies of communicators: the Liars and the Deceivers. The Liars generally say the opposite of whatthey think. “Rotten day today,” they say when the weather is beautiful, and vice versa. The Deceivers, however, do not behave so reliably: they mix putative truth and falsity more or less randomly. The Liars can communi44 Usefully relevant discussions can be found in David Lewis, Convention: A Philosophical Study (Cambridge, 1969). But cf. Angus Ross, “Why Do We Believe What We Are

Told?” Ratio 28 (1986): 69-88.

“

THE EPISTEMIC ROLE OF POSTULATION

201

cate with us and with each other perfectly well. Once one has got the trick, one knows exactly how things stand in discussions with them. But the Deceivers are something else again. One never knows where one stands with them. And worse yet, they too have no idea where they stand with one another. Indeed, they could never even begin to communicate. Even if an initial generation of Deceivers came equipped with a readymade language (say, because they began as normal communicators and then turned en masse into Deceivers at some point), the fact remains that they could never teach language to their offspring. “That’s a lion,” the parents observe to Junior one day,pointing to a dog, and “That’s a cat,” the next time, and“That’s an elephant,” the time after that. Poor Junior would never manage to catch on. Contrast now two other communities: theTrusters and the Distrusters. The Trusters operate on the principle “Be candid yourself, and also accept what other people say as truthful—at any rate in the absence of counterindications.” The Distrusters operate on the principle “Be deceitful yourself, and look on the assertions of others in the same light—as

ventures in deceitfulness: even when people are ostensibly being truthful, they are only trying to lure you into a false sense of security.” It is clear once again that the policy of the Distrusters is totally destructive of communication. If exchange of information for the enhancement of our knowledge is the aim of the enterprise, then the diffusion of distrust is utterly counterproductive. To be truthful, to support the proper use of language and refrain from undermining its general operation is a policy of the greatest general utility, however beneficial occasional lapses may seem to be. Not only is the maintenance of credibility an asset in communication, but some degree of it is in fact a necessary condition for the viability of the whole project. The precept “Protect your credibility; do not play fast and loose with the ground rules but safeguard your place in the community of communicators” is basic to the communicative enterprise. From the sender's point of view, putting forth a message costs time, effort, energy, and the like. The rational agent will incur such costs only with a view to benefits—some sort of reward (if only in the respect or gratitude of others) or reciprocity, with a view to a quid pro quo. This point is simple but of far-reaching import. Given our need for information to orient us in the world (on both pure and practical grounds), the value of creating a community of communicators is enormous. We are rationally well advised to extend ourselves to keep the channels of communication to our fellows open, and it is well worth expending much for the realization of this end. And the same sort of story holds for the receivers’ point of view as well. They too must expend resources on taking in, processing, and storing

202

:

THE PURSUIT OF TRUTH

messages, Clearly, a rational hearer would be prepared to undertake this expenditure only if there were a reasonable expectation of drawing profit from it, be it by way of information obtained or pecuniary reward—an expectation that in general is amply warranted. In this regard, it is useful to consider H. P. Grice’s influential interpretation of communicative practice. Grice’s analysis focuses on the intentions of speakers, who aim to induce their auditors to accept (believe) something by getting them to recognize that this is what the speakers are seeking to do. This is correct enough as far as it goes. But Grice’s speakers’ perspective slights the fact that an equal burden is borne by hearers in communication situations. To begin with, auditors will not even enter into communicative contact unless convinced of the prospect of some benefit. If they did not think that the speakers were trying to convey something to them—f, for example, they regard the movement of speakers’ lips as we regard the twitching of a sleeping dog’s legs—they would give little heed. Moreover, hearers have to use whatever cues and clues corne to hand to construct an interpretation of what is (putatively) being said. Speakers’ intentions notwithstanding, the only information that is ever actually conveyed is what hearers accept. And this is, in general, a matter of the simplest, the most cost-effective construction that they can

place upon speakers’ declarations relative to what they take to be the questions at issue, using whatever data that come to hand, be they tacitly behavioral or overtly expressed. Speakers, of course, benefitin communicative situations insofar as the information conveyed influences hearers’ present and future actions in directions speakers see as desirable. And failures of transmission are generally of detriment no less to those who convey information than to those who receive it. In informative communication, hearers—-who are volun-

tary parties to the transaction—want to extract the maximumbenefit (in terms of answers to their questions) from the speakers’ assertions, subject to protecting the speakers’ credibility insofar as possible. This is a matter of a common interest between the two parties, for once speakers’ reliability is lost (in the hearers’ opinion), then their utility as a source of information is utterly destroyed. Hence speakers and hearers have a joint interest in protecting the credit of speakers. Consider just one example. Suppose you and I communicate (in the informative mode). In the first instance, for the economic reasons already canvassed, I extend you benefit of doubt and take your various declara-

tions at face value. I enter into the discussion with an established view of the world already in place, and I use your declaration to extend and am15H. P. Grice, “Meaning,” Philosophical Review 66 (1957): 377-88. Cf. also Jonathan Bennett, Linguistic Behavior (London, 1963), chaps. 1 and 7.

THE EPISTEMIC ROLE OF POSTULATION

203

plify it. You say, “The cat is on the mat,” and I proceed to readjust my world picture to place the cat on the mat. Your information supplements my information base and enlarges the details of my world picture. But the matter is very different when you present discordant information. If, for example, you say, “A unicorn is in the forest,” then I shall make no

change in my world picture save in relation to you and your state of mind. I endorse, “He is under the (mis)impression that there is a unicorn in the forest.” That is, when the result of integrating the substance of

your declarations into my world picture becomes too descrepant (when too great a cognitive cost would ensue from my endorsement of your claims, because I must abandon too much to accommodate them), then

instead of supplementing my register of endorsed propositions with your claim p, I merely add, “He has the (mis)understanding that p.” So when you say, “The present king of France is bald,” then rather than taking your declaration at face value, I instead endorse, “He has the mistaken

idea that there is currently a king of France who is bald.” And here no ontological novelty is introduced at all, because the claim I endorse is one about you, and you are already featured on my ontological agenda. The need for making special ontological provision for nonexistents simply does not arise. From the ontological point of view, we can deal with claims regarding nonentities wholly in terms of existents, namely, in terms of the beliefs, (mis)impressions, assumptions, suppositions, and the like of perfectly real people. On economic grounds, a new ontological category of nonexistent entities is quite dispensable. We can make do with real people and what they can‘know, think, and suppose. The manner in which nominalistic philosophers stress ontological economy in the context of talk about nonexistent possibles can accordingly also be understood in terms of quasi-economic costs and benefits. In general, when one endeavors to convey information to someone,

various sorts of unpleasant reactions can occur, as figure 11.1 shows. Throughout, effective communication is a matter of maintaining proper cost-benefit coordination. It is governed by such principles as: Figure 11.1. Communicative Negativities and Their Causes

Negativity One is disbelieved.

Sample Causative Etiology One speaks falsely too often (out of heedlessness or out of deceptiveness).

One is misunderstood.

One uses careless or inadequate formulations.

One is tuned out.

One speaks off the point (digresses) or speaks at undue length (even if it is to the point).

204

‘

.

THE PURSUIT OF TRUTH

Be sufficiently cautious in your claims to protect your credibility, but do not

say so little that people dismiss you as a useless source. Formulate your statements fully and carefully enough so as to avoid misunderstanding but not with so much detail and precision as to weary your auditors

and get tuned out. Make your message long (explicit, detailed, etc.) enough to convey your points but short enough to avert wasting everyone’s time, effort, and patience. Be sufficiently redundant that an auditor who is not intensely attentive can still get the point, but not so redundant as to bore or annoy or insult your auditors.

Keep to the point but not so narrowly that your message is impoverished by lack of context.

All of these principles are fundamentally economic principles of balance; they all turn on finding a point after which the benefit of further gain in information falls below the cost demanded for-its process of acquisition. Communication rests on postulation—on making assumptions and presumptions that lack any specific prior evidentiation. But these assumptions are not made randomly and willfully, without one sort of rational basis in good reasons. And these goodreasons lie in the order of economic rationality, of cost-benefit considerations relating to the systemic processes at issue.

Scientific versus Ordinary-Life Communication It is useful to bear in mind that different priorities obtain in different contexts of communication. In everyday-life communication where we are deeply concerned to protect our credibility, we value security over informativeness. Hence looseness and imprecision are perfectly acceptable. On the other hand; in science we value generality and precision over security. After all, natural science is not content with theses like“On the whole, larger Gbjects are heavier,” or again, “Most things made predominantly of lead generally melt at.temperatures around 330° C.” In science, we seek exactness and precision: we want to know accurately how all objects of exactly this or that sort always behave. Generality, precision, and detail are at a premium, and so in scientific discourse we prioritize these factors in a way that makes our scientific theories vulnerable. (The half-life of theories in frontier physics is relatively short.) The very reason for being of natural science is to characterize nature’s processes exactly and to describe how they operate always and everywhere, in full generality and precise detail. Technical science forswears the looseness of vague generality or analogy or approximation. It has no

THE EPISTEMIC ROLE OF POSTULATION

205

use for qualifiers such as “usually,” “normally,” or “typically”; universality and exactness are its touchstones. Natural science, accordingly, de-

clares not merely that roughly such-and-such generally occurs in certain sorts of circumstances but exactly what happens in exactly what circumstances. In science we always aim at the maximum of universality, precision, accuracy, and exactness. The law claims of science involve no

hedging, no fuzziness, no incompleteness, and no exceptions; they are strict: precise, wholly explicit, exceptionless, and unshaded. In making

the scientific assertion “The melting point of lead is 327.5° C,” we mean to assert that all pieces of (pure) lead will unfailingly melt at exactly this temperature. We certainly'do not mean to assert that most pieces of (pure) lead will probably melt at somewhere around this temperature. (And in this regard, there would be a potential problem, should it turn out, for example, that there is no melting point at all and that what is actually at issue is the center of a statistical distribution.) And this commitment to generality and detailed precision renders the claims of science highly vulnerable. We realize that none of the hard claims of present-day frontier natural science will move down the corridors of time untouched. Fragility is the price that we pay in science for the sake of generality and precision. Increased confidence in the correctness of our estimates can always be purchased at the price of decreased accuracy. We estimate the height of the tree at around 25 feet. We are quite sure that the tree is 25 + 5 feet. We are virtually certain that,its heightis 25 + 10 feet. But we are completely and absolutely sure ‘that its height is between 1 inch and 100 yards. Of this we are completely sure, in the sense that we deem it absolutely certain, certain beyond the shadow of a doubt, as certain as we

can be of anything in the world, so sure that we would be willing to stake our life on it, and the like. With any sort of estimate, there is always a characteristic trade-off relationship between the evidential security of the estimate on the one hand (as determinable on the basis of its probability or degree of acceptability), and its contentual definiteness (exactness, detail, precision, etc.) on the other. _ This relationship between security and definiteness is generally characterized by a curve of the general form of an equilateral hyperbola: s x d = c (where c = constant). (See figure 11.2.) The increased vulnerability and diminished security of our claims is the undetachable other side of the coin of the pursuit of definiteness. Science operates in the lower right-hand sector of the figure. Its cultivation of informativeness (definiteness of information) entails the risk of error in science: its claims are subject to great insecurity. No doubt the progress of science makes it possible to decrease the value of c somewhat, but the fundamental tradeoff relationship remains unavoidable. An information-theoretic uncer-

206

.

THE PURSUIT OF TRUTH

Figure’ 11-2. The Relationship between Security and Definiteness Sx d=c (constant)

Security

J

(Ss)

Definiteness (d)

tainty principle prevents our obtaining the sort of information we would ideally like.46 The exactness of technical scientific claims makes them especially vulnerable, notwithstanding our most elaborate efforts at their testing and substantiation. The situation in science accordingly differs markedly from that which prevails in everyday life. When we ordinarily assert that peaches are delicious, we are asserting something like, “Most people will find the eating of suitably grown and duly matured peaches a rather pleasurable experience.” Such a statement has all sorts of implied safeguards, such as “more or less,” “in ordinary circumstance,” “by and large,” “normally,” “if all things are equal,” and so on. They are nothing like scientific laws but. mere rules of thumb, a matter of practical lore rather than scientific rigor. And this enables them to achieve great security, for there is safety in vagueness: a factual claim can always acquire security through inexactness. Take the claims that there are rocks in the world and that dogs bark. It is virtually absurd to characterize such everyday-life generalizations as

fallible. Their security lies in their indefiniteness or looseness; it is unrealistic and perverse to characterize such common-life claims as defeasible. They say so little that it is unthinkable that contentions such as these should be overthrown. And this accords smoothly with the needs of the situation. For ordinary-life communication is a practically oriented

endeavor carried on in a social context. It stresses such maxims as “Aim for security, even at thepriceofdefiniteness”; “Protect your credibility”; “Avoid misleading people.” When, as in ordinary life, the preservation of credibilityis paramount, 46 Vagueness constitutes a context in which we trade off iriformativeness (precision) with probable correctness (security), with science moving toward the former, and everyday knowledge toward the latter. The relevant issues are considered in tantalizing brevity in Charles S. Peirce’s short discussion of the logic of vagueness, which he laments as too much neglected, a situation that has since been corrected only partially. (See C. S. Peirce, Collected Papers, 8 vols., ed. C. Hartshorne et al. [Cambridge, Mass., 1931-50], vol. 5;_secs.

5.446-50.)

THE EPISTEMIC ROLE OF POSTULATION

207

one wants to formulate one’s claims in as safe and secure a way as possible, and thus one resorts to vagueness and imprecision. In contrast, when, as in science, creativity and originality are paramount, then one would put one’s claims in the most ambitious and surprising way, accepting the risks inherent in universality, precision, and the like. Plausibly enough, the appropriateness of an epistemic policy hinges upon the nature of the governing desideratum of the domain (e.g., credibility vs. creativity). The aims of ordinary-life discourse are primarily practical, largely geared to social interaction and the coordination of human effort in communal enterprises that serve the common good. In this context, it is crucial that we aim at credibility and acceptance, that we establish and maintain a good reputation for reliability and trustworthiness. In the framework of common-life discourse, we thus take our stance at a point far removed from that of science. Very different probative orientations prevail in the two areas. In everyday contexts, our approach is one of situational satisficing: we stop at the first level of sophistication and complexity that suffices for our present needs. In science, however, our objectives are primarily theoretical and governed by the aims of disinterested inquiry. Hence the claims of informativeness—of generality, exactness, and precision—are paramount.

In science, we accept greater risks willingly because we ask much more of the project. We deliberately court risk by aiming at maximal definiteness and thus at maximal informativeness and testability. Aristotle’s view that terrestrial science deals with what happens ordinarily and in the normal course ofthings has long ago been left by the wayside. The theories of modern natural science have little interest in what happens generally or by and large; they seek to transact their explanatory business in terms of strict universality—in terms of what happens always and everywhere and in all kinds of circumstances. In consequence, we have no choice but to acknowledge the vulnerability of our scientific statements, subject to the operation of the security-definiteness trade-off. In ordinary life, we operate at the upper left-hand side of the curve in figure 11.2. The situation contrasts sharply with that of science, whose objectives are largely theoretical, and where the name of the game is rigorous understanding on a basis of unrestricted universality and extreme precision. The cost-benefit situation of the two domains is drastically different. Either way, however, an important fact remains constant. Communication about commonly accessible objects rests on certain assumptions and conventions and postulations (e.g., “There is a reality out there that in fact has certain characteristics, about which others are trying to tell as something”). The use of such postulations becomes justified retrospectively in due course—yet not by establishing their truth but rather

208

.

THE PURSUIT OF TRUTH

through considerations of utility, of efficiency and economy in the context of a pursuit of our own epistemic purposes. ON

Idealistic Implications The idealistic bearing of these deliberations is clear. In building up our picture of the world, we do and must make use of conceptual and communicative machinery to secure, record, and transmit information. And this machinery that we devise for the cognitive processing of information is based on conventions that reflect our purposes. Here form follows function: the substance of our world picture in some measure reflects the purposes that govern and determine the ground rules by which we process information. Thus once again the purposes established by minds serve to shape and condition the nature of the products that these minds create. We cannot escape the idealistic impetus of the fact that our knowledge is a cognitive artifact not only in its origination but in its make-up as well.” 17 The themes of this chapter are treated in the author's Plausible Reasoning (Assen, 1976), Cognitive Systematization (Oxford, 1979), Induction, and Empirical Inquiry (To-

may

towa, N.J., 1982).

:

.

Twelve Truth as an Idealization Ideal Coherence In general terms, a coherence epistemology of truth is one that proceeds on the basis of a truth criterion to the effect that a proposition-p is true iff it forms part of an appropriately formed set of optimally coherent propositions. On such an approach, the truth status that is accorded to propositions reflects nota strictly local aspect of their own but an ultimately global feature of their place in a larger context: it is not an isolated but a contextual characteristic. Yet how does this sort of approach to the issue of a criterion of truth relate to the question of the meaning at issue when we maintain the truth of a proposition? A common objection’to the coherence theory of factual truth is that the linkage of coherence to truth is simply too loose for coherence toprovide the definitive standard of truth. As one writer put it some years ago: “It is quite conceivable that the coherence theoryis a description of how the truth or falsehood of statements comes to be known rather than an analysis of the meaning of ‘true’. . . . One might agree that a given statement is accepted as true in virtue of standing in certain logical relations to other statements; still it would not follow that in calling it true one means to ascribe to it those relations.”! Here we have the standard reservation regarding a coherence theory of truth: “Coherence may perhaps be suitable as a criterion for the true, but certainly not as a definitional standard of truth.”? It can, however, be shown that if one is prepared to consider coherence in an idealized perspective—as optimal coherence with a per-

fected data base, rather than as a matter of manifest coherence with the actual data at our disposal—then an essential link between truth and coherence emerges. Supporters of a coherentist standard of truth ought surely to be able to 1 Arthur Pap, Elements of Analytic Philosophy, (New York, 1949), p. 356. Cf. H. H. Joachim, Logical Studies (Oxford, 1948). 2 The definition-vs.-criterion dichotomy was the starting point of the author’s Coherence Theory of Truth (Oxford, 1973). It also provided the pivot for the critique of the coherentism of Blanshard’s Nature of Thought in his contribution to The Philosophy of Brand Blanshard, ed. R. A. Schilpp (LaSalle, Ul., 1980). Several subsequent publications have kept the pot boiling, in particular Scott D. Palmer, “Blanchard, Rescher, and the Coherence Theory of Truth,” Idealistic Studies 12 (1982): 211-30, and Robert Tad Lehe, “Coherence: Criterion and Nature of Truth,” Idealistic Studies 13 (1983): 177-89.

210

THE PURSUIT OF TRUTH

establish that this criterion is duly consonant with the definitional nature of truth. For there ought rightfully to be a consonance between our evidential criterion of acceptability-as-true and the truth as definitionally specified. Any really satisfactory criterion must be such as to yield the real thing—at any rate, in sufficiently favorable circumstances. Fortunately for coherentism, it is possible to demonstrate rigorously that truth is tantamount to ideal coherence—that a proposition’s being true is in fact equivalent to its being optimally coherent with an ideal data base.* This circumstance has far-reaching implications. Given that “the real truth” is guaranteed only by ideal coherence, we have no categorical assurance of the actual correctness of our coherence-guided inquiries, which are, after all, always incomplete and imperfect. As the history of science amply substantiates, the discoveries about how things work in the world secured through the coherentism at work in the scientific method constantly require adjustment, correction, and replacement. It is only too obvious that we cannot say that our coherence-grounded scientific theorizing furnishes us with the real (definitive) truth, but just that it furnishes us with the best estimate of the truth that we can achieve in the circumstances. at hand. Let us begin, then, by considering what is meant when a contention is classed as true. In thus characterizing a claim, we indicate that what it states corresponds to the facts, so that its assertion is in order. But while this factuality (“stating what is the case,” “corresponding to the facts”) is what truth is all about, we cannot apply or implement it as such: it-does not provide a basis on which the truth of claims can be determined. Thus while we have the equation to be true = to be (assertible as) factually correct,

we also will have to be determinable as true = capacity to meet certain conditions that can be taken as indicators of factual correctness.

But these conditions of truth determination are in their very nature con-

ditions whose full realization is a matter of idealization—that is, the conditions are such as to obtain only in ideal cireumstances. , It warrants note that this approach makes it possible to meet the traditional objection to coherentism that the coherence theory is unable to deal with the problem of error—to be able to explain how it is that what is thought to be true might yet actually prove false. For the consideration 3 For the somewhat technical details of the argumentation that establishes this contention, see the appendix to this chapter.

TRUTH AS AN IDEALIZATION

211

that truth is a matter of ideal (rather than actually realized) coherence at once sweeps this difficulty aside. Definitive knowledge (as opposed to “merely putative” knowledge) is the fruit of perfected inquiry. Only here, at the idealized level of perfected science, could we count on securing the real truth about the world that “corresponds to reality,” as the traditional phrase has it. Factual knowledge at the level of generality and precision at issue in scientific theorizing is akin to a perfect circle. Try as we will, we cannot quite succeed in producing it. We do our best and call the result knowledge— even as we call that carefully drawn circle on the blackboard a “circle.” Butwe realize full well that what we currently call scientific knowledge is no more authentic (perfected) knowledge than what we call a circle in a geometry diagram is an authentic (perfected) circle. Our knowledge is in such cases no more than our best estimate of the truth of things. Lacking the advantage of a God’s-eye view, we have no access to the world’s facts except through the mediation of inquiry, which is inevitably incomplete and thus always potentially flawed. All we can do—and what must suffice us because indeed it is all/that we can do—is to do the best we can with the cognitive state of the art to estimate the correct answers to

our scientific questions.

Science and Idealization The fact that what we actually achieve in theorizing inquiry is, not the real truth as such, but only our best realizable estimate of it, means that

we must pursue this cognitive enterprise amid the harsh realities and complexities of an imperfect world. In deliberating about the truth of our scientific claims, as elsewhere, the gap between the real and the ideal must unavoidably be acknowledged. In the subideal, real-life conditions of an epistemically difficult world, an evidential gap indeed separates

presumptive from certifiable truth. But given an adequate criteriology of truth, this gap becomes closed in ideal circumstances. The continuity condition reflects the fact that inquiry aims at truth—that the real truth is the definitive aim and aspiration of the scientific enterprise. Given the facts of life in empirical inquiry, we have neither the inclination nor the justification to claim that the world is as our present science describes it to be. Nor, as we have seen, does it make sense to

identify “the real truth” with “the truth as science-in-the-limit will eventually see it to be.” All that can be done in this direction is to say that the world exists as ideal or perfected science describes it to be. The real, which is to say, final and definitive truth about nature at the level of scientific generality and precision is something we certainly cannot as-

212

.

THE PURSUIT OF TRUTH

sume to be captured by our science as it stands here and now (though, of course, this fact nowise destroys the rationality of our endorsement of

current beliefs). We cannot but take the stance that scientific truth is not something in hand but something that—so we must suppose—is attained only in the ideal or perfected state of things. With respect to scientific issues we thus arrive at the coordinating equation: the real truth = the truth as ideal (perfected) science purports it to be.

To be sure, in espousing this conception, we intend to make “ideal science” contingent upon truth, rather than the reverse: the former is the

independent, the latter the dependent variable. While one can never lay claim to have definitely secured “the real truth” in matters of scientific theorizing, this notion nevertheless serves an important role in providing a contrast conception that constitutes a useful reminder of the fragility of our cognitive endeavors. It establishes a contrast between our present science as we have it and a perfected ideal science that alone can properly be claimed to afford a grasp of real-

ity.

os

But what of the objection that we could not tell that we had arrived at the definitive truth even if we in fact had done so? Its resolution lies in the fact that the attainment of such an objectiveis simply not a liveissue. The goal of monitoring the correctness of our science independently of and, so to speak, externally from its own efforts is clearly infeasible in principle. In this imperfect epistemic dispensation, we have to reckon with the realitites of the human condition. In science, as in other do-

mains of human endeavor, it is a matter of doing the best we can with. the tools that come to hand. We can pursuea goal in the full realization that perfection is unattainable in its achievement. The unattainability of perfection does nothing to countervail against the no less real fact that improvement is realizable—that progress is possible. The undeniable prospect of realizable progress—of overcoming genuine defects and deficiencies that we find in the work of our predecessors—affords ample impetus to scientific innovation. The labors of science are not pulled forward by the mirage of (unattainable) perfection. We are pushed onward by the (perfectly realizable) wish to do better than our predecessors in the enterprise. We can understand progress in two senses. On the one hand, there is

O-progress, defined in terms of increasing distance from the starting point (the “origin”). On the other hand, there is D-progress, defined in terms of decreasing distance from the goal (the “destination”). Consider the picture:

TRUTH AS AN IDEALIZATION

origin O

213

destination the attained position

D

Ordinarily, the two modes of progress are entirely equivalent: we increase the distance traveled from O by exactly the same amount as we

decrease the distance remaining to D. But if there is no attainable destination—if we are engaged on a journey that, for all we know, is literally endless and has no determinable destination, or only one that is infinitely. distant—then we just cannot manage to decrease our distance from it. Given that in natural science we are embarked on a journey that is literally endless, it is only O-progress that can be achieved, and not Dprogress. We can gauge our progress onlyin terms of how far we have come by way of imposing upon our ventures prediction and control, and not in terms of how far we have to go. Embarked on a journey that is in principle endless, we simply cannot say that we are nearing thegoal. The upshot is straightforward. The idea ofimproving our sciencé can be implemented withoutdifficulty, since we can clearlyimprove our performance as regards its practical tasks of prediction, control, and the rest. But the idea of perfecting our science in terms of its theoretical adequacy and accuracy cannot be implemented. And no harm results from conceding this. Nor does it destroy the utility of the idea of perfected science. Perfected science—and .the definitive truth that it affords us—is an ideal, and an ideal is not something we encounter in experience but rather the hypothetical projection or extrapolation of what we encounter in experience. And the ideal status of “the truth” in scientific matters carries wider implications for how we can and should regard the actual deliverances of natural science. A salient example of this is the conception of “natural law” that lies at the very core of our conception of nature. By definition, a natural law is “a genuine truth regarding some universal feature of the workings of nature.” But insofar as, from the cognitive point of view, truth and universality are idealizations, so also is lawfulness, and with it such matters as space, time, element, and the other conceptual building blocks of our understanding of nature.‘ But ideal science is not something we have in hand here and now. And it is emphatically not something toward which we are moving along the asymptotic and approximative line envisaged by Peirce.> The asymptotic theory of scientific truth runs together two things that are by no means necessarily connected: there is in fact no warrant for identifying ideal or perfected science with ultimate science, or science-in-the-limit. Even if 4 For an elaboration of this issue, see the author’s Conceptual Idealism (Oxford, 1973). 5 See the author's Peirce’s Philosophy of Science (Notre Dame, Ind., 1978).

214

:

THE PURSUIT OF TRUTH

it made sense to contemplate the Peircean idea of an eventual completion of science, there would be no guarantee that this completed science (given its existence) would satisfy the definitive requirements of a perfected science—that it would, for example, achieve the completeness of erotetic equilibrium where every posable question is duly resolved. Peircean convergentism is geared to the supposition that ultimate science—the science of the very distant future—will somehow prove to be an ideal of perfected science freed from the sorts of imperfections that afflict its predecessors. But the potential gap that arises here can be closed only by old-style, substantive metaphysics of a most problematic sort.

Perfected science is not something that exists here and how, nor is it something that lies ahead at some eventual offing in the remote future. It is not a real thing to be met with in this world. It is an idealization that exists outside time—that is, it cannot actually come to realization at all. It lies outside history as a useful contrast-case that cannot be numbered among the achieved realities of this imperfect world. Existing science does not and presumably never will embody to perfection cognitive ideals of completeness, unity, consistency, and so forth. These factors rep-

resent an aspiration rather than a coming reality; a telos, not a realizable condition of things; a hypothetical condition from which any and all of the negativities of the realized actual ‘positions have been removed. The concept of science perfected—of an ideal and completed science that captures the real truth of things and satisifies all of our cognitive ideals (definitiveness, completeness, unity, consistency, etc.}—is at best a useful fiction, a creature of the fictive imagination and not the secured

product of inquiring reason. This ideal science is, as the very name sug-

gests, an idealization, something that involves the removal in thought of limitations that obtain in fact. It involves the use of mind to move from a concern with the order of things as they are into the order of things as they ought to be. -

Truth as an Idealization Definitive truth is realizable only by way ofidealization: actual inquiry ' presents us with estimates of truth. In matters of scientific theorizing, the real truth as such is realizable only under ideal conditions. And so we have no alternative to presuming that our science as it stands here and now does not present the real truth but only estimates it. “Our truth” in matters of scientific theorizing is not—and presumably never actually will be—the final truth. However confidently science may affirm its conclusions, the realization must be maintained that its declarations are provi-

TRUTH AS AN IDEALIZATION

215

sional, tentative, and subject to revision and even to outright abandon-

ment and replacement. But all this is not, of course, any reason to abandon the link to truth at the teleological level of aims, goals, and aspirations. The pursuit of scientific truth, like the pursuit of happiness— or, for that matter, any other ideal in life—is not vitiated by the consid-

eration that its full realization is not a matter of the practicalities of this imperfect world. The ideal of a state of the art in science that attains definitive finality in empirical inquiry is pie in the sky. It represents an idealization and not a matter of the practical politics of the epistemic domain. But it affords the focus imaginarius whose pursuit canalizes and structures our actions. It represents the ultimate objective (goal) of inquiry—the destination of an incompletable journey. The conception of capital-T Truth thus serves a negative and fundanjientally regulative role to mark the fact that the place we have attained falls short of our capacity actually to realize our cognitive aspirations. It marks a fundamental contrast that regulates how we do and must view our claims to have got at the truth of things. It plays a role somewhat reminiscent of the functionary who reminded the Roman emperor of his mortality in reminding us that our pretentions to truth are always vulnerable. Contemplation of this ideal enables us to maintain the ever-renewed recognition of the essential ambiguity of the human condition as suspended between the reality of imperfect achievement and the ideal of an unattainable perfection. We must suppose that science does not and cannot attain an omega condition of final perfection. The prospect of fundamental changes lying just around the corner can never be eliminated finally and decisively. A basic analogy obtains as per the following proportion: putative knowledge : actual inquiry :: genuine knowledge: ideal inquiry

Rational inquiry is the pursuit of an unattainable ideal—the ideal of “the real truth” about laws of nature as yielded by perfected science. Actual inquiry is no more than our best effort in this direction, and the information (the putative knowledge) it yields is no more than our best available estimate of the real truth of things. To abandon this conception of the truth as such would be to abandon an idea that crucially regulates our view as to the nature and status of the knowledge we lay claim to. We would then no longer be constrained to characterize our truth as merely ostensible and purported. And then, did our truth not exhibit any blatant inherent imperfections, we would be tempted to view it as real, authentic, and final in a manner that as we at

bottom realize it does not deserve. The lesson of these deliberations is clear. What we have in view in

speaking of “the real truth” in scientific matters is not the actual product

216

:

THE PURSUIT OF TRUTH

of current inquiry but the hypothesized product of idealized inquiry. The conditions of objectivity and definitiveness we have in view in relation to “the real truth” are not satisfiable in the circumstances in which we do and unavoidably must labor. In this regard we have no realistic alternative but to regard the truth in these matters as having an idealization. Conceptions such as definitive knowledge and truth in matters of scientific theorizing are idealizations geared to the idea of completed and perfected science. And this—all too obviously—is not something we have in hand. It is no more than a regulative ideal that guides and directs our efforts in question-resolving inquiry. Such an ideal is not (or should not be) something that is unhealthily unrealistic. It should produce not defeatism and negativity toward our actual efforts and their fruits but rather a positive determination to do yet better and fill a half-full barrel yet fuller. It should govern not our expectations of realized achievement but our aspirations, not our demands but

our hopes. Endowing us with a healthy skepticism toward what we actually have in hand, it should encourage our determination to make further improvements and act not as an obstacle but as a goad.

Appendix: Truth as Ideal Coherence For traditional idealists, coherence is the definitive standard of truth. But of course discerned coherence with our actual data does no more than provide for putative or presumptive truth. Only a step to idealization carries us further, seeing that only under ideal conditions (with optimal coherence with ideal data) is the claim to actual truth a valid one. For imperfect knowers, the gap from putative to genuine knowledge can never be securely closed. The idealistic coordination of cognition with truth unavoidably calls for idealization, but can indeed be achieved by its means. And this can be demonstrated in a logically rigorous way. Let us say that a factual proposition satisfies the condition of “ideal coherence” if it is optimally coherent with a perfected (or completed) data base. Given the nature of coherence, such a proposition will fit more _ smoothly and consistently with this idealized-data base than does its negation (and so consequentlyfits better than any other proposition that is incompatible with it). The ensuing discussion will argue that, when ideal coherence is construed in this way, then truth is demonstrably tantamount to ideal coherence. It will endeavor to demonstratedecisively that the linkage between these two factors is so tight that they in fact coincide. To establish this contention, it must be shown that two implication theses obtain with respect to any and all statements:

TRUTH AS AN IDEALIZATION (T1)

true — ideally coherent.

(T2)

ideally coherent —> true.

217

The idea of “ideal coherence” operative here should be understood as being a matter of optimal coherence (c) with a perfected data base (B). Deploying these abbreviations, it is clear that the two principles at issue can now be formulated as follows: (TI) (T2)

‘S’ is true > ‘S’c B. ‘Sc B—>‘S’ is true.

If the coherence theory is to be adequate, the validation of these two principles will be mandatory. They have to be grounded in the very nature of “optimal coherence (c) with a perfected data base (B).” To establish them we shall need to look more closely at the crucial ideas at issue: the conceptions “optimal coherence” and “perfected data base.” First a word about optimal coherence. Just what is it “to cohere optimally with a data base”? What does ‘S’ c B involve? The answer is provided by two conditions: 1. ‘S’ represents a member of some appropriate family of mutually exclusive and exhaustive alternatives: {S,, 52, Ss, ... , S,}2. ‘S’ is more smoothly cosystematizable with B than is any of its rival alternatives S,, singly or in combination. (Note that this means specifically that ‘S’ is more smoothly cosystematizable with B than is ‘~S.’) “To implement this second idea, we must have in hand some concrete

principles of cognitive systematization that determine a standard of systemic connection according to which ‘S’ is more smoothly coordinated with B than is the case with any of its available comparable and competing alternatives. However, since here is not the place for a full-scale presentation of a coherentist truth-criteriology, this issue is something that

we need not discuss in further detail within the orbit of present purposes.®

Let us now turn to the idea of a “perfected data base.” Perfection has two components: completeness (or comprehensiveness) and adequacy (or definitiveness). Thesé have the following ramifications: 1. Completeness: If D is to be a perfected data base, then it must be suffi-

ciently complete and comprehensive that, for any thesis ‘S’ within the domain of discussion at issue, either ‘S’ itself or its negation, ‘~S’ will cohere optimally with D: 6 For a further development of these ideas, see the author’s Coherence Theory of Truth and Cognitive Systematization (Oxford, 1979).

218

:

THE PURSUIT OF TRUTH

If perf-(D), then: either ‘S’ c D or ‘~S’ c D, for any and every statement ‘S’ of the relevant domain.

2. Adequacy: To acknowledge D as a perfected data base is to acknowledge it as actuality determinative. And so we must endorse:

If perf (D), then: if ‘S’ c D, then A(S), where A(S) — S.

Completeness requires decisiveness; adequacy requires facticity. These are conditions that inhere in the very notion of the perfection of a data base. To be sure, all this is not to say that we can ever actually find such a perfected data base. We doubtless cannot. The very idea of such a data base represents an idealization. The claim at issue is framed inastrictly hypothetical mode: “Ifa perfected data base exists, then it must ipso facto have the specified characteristics.” We are, in effect, dealing with the meaning postulates, or definitional requirements, for the idea of a perfected data base—certain explanatory stipulations for ‘what the idea of such a data base involves (in the context of “optimal coherence’). As a preliminary consideration, let us first establish the effective uniqueness of such a perfected data base in point of optimal coherence: To demonstrate this, let us make the assumption that both B, and B, answer to the characterization of.a perfected data base. We can then establish: If ‘S’ c B,, then ‘S’ c Bo, for any statement ‘S’.

oO De QR

This is accomplished by the following argument: . Suppose: ‘S’ c B,. . Suppose further that not: ‘S’ c Bg. Then ‘~S’ c B, follows from (2) by Completeness. Then A (~S) follows from (3) by Adequacy.

. But A(S) follows from (1) by Adequacy. . Since (4) and (5) are mutually contradictory given the law of excluded middle, we must negate supposition (2); we hence have: ‘S’ c Bz. QED

The converse, of course, follows by exactly the same line of reasoning. — And this shows that with respect to “optimal coherence,” there is in effect (at most) one perfected data base. Let us continue to designate this by B. By definition, then, B is the (unique) perfected data base, whose availability, as already observed, we can claim not as a matter of realizable fact but only as a matter of idealization. It follows immediately from the two stipulated requirements of Completeness and Adequacy that B must satisfy the conditions represented by the following principles:

TRUTH AS AN IDEALIZATION

219

(P1) By the condition of Adequacy, we have it that if ‘S’ does indeed optimally cohere with B, then this state of affairs must be actual: ‘Sc B— A(S). (P2) By the condition of Completeness, we have it that if ‘S’ does not cohere optimally with the perfected data base (B), then it follows that ‘not-S’ will be

optimally coherent with the perfected data base B. Symbolically:

~(‘S’ c B) > (‘~S’ cB). These two principles, (P1) and (P2), will furnish the materials on whose basis our two focal implication theses, (T1) and (T2), can and must be established. They are all we have; if the job is to be done, they must suffice to do it. Before proceeding to show this, however, the idea of factuality reflected in our actuality-indicator A deserves some comment. The claim at issue is one of factuality, of “adequation to fact” (adaequatio ad rem): to assert A(S) is to maintain that the state of affairs S is a constittiént of the real world, that existing reality is (in part) characterized by this state of affairs. Thus, to assert A(S) is effectively to assert that ‘S’ represents a bestehender Sachverhalt, an actual state of affairs, in the vocabulary of

Ludwig Wittgenstein’s Tractatus Logico-Philosophicus. The thesis at issue with A(S) is an ontological one: it claims that that is how things in fact are, whether or not people know or believe it. And this ontologically definitive aspect ofA means that we must have the law of excluded middle (LEM) represented by a tertium non datur principle:

(LEM)

~A(S)iffA (~S).

Actuality must “make up its mind” with respect to the A(S)-versus-A(~S) dichotomy. This condition inheres axiomatically in the very meaning of “actuality.” On this basis, let us now proceed to establish principles (T1) and (T2) as stated at the outset. The required demonstration is easily produced. Given that truth is (by definition, as it were) subject to the ancient principle of accord with fact (adaequatio ad rem):

(A)

'

‘S’ is true < A(S),

we have it that principle P1 immediately entails: ‘S’ c B—> ‘S’ is true.

This provides thesis T2, so that half of our task is already accomplished. To obtain thesis T1, let us consider principle P1 in the special case of _ the state of affairs ‘not-S’:

220

.

qd) -

THE PURSUIT OF TRUTH

~A(~S) > ~(‘~S’ cB).

.

By the law of excluded middle, namely:

(LEM)

A(S) ~A(~S),

we have it that (1) yields:

(2)

A(S) > ~(‘~S’ cB).

Now in view of (P2), this yields:

.

A(S) > ‘Sc B.

And given (A), this in turn yields: ‘S’ is true — ‘S’c B.

We have thus also provided for thesis (T1), thereby completing our task. It follows from the resultant equivalence of truth-as-adequation on the one side, with ideal coherence on the other, that an adequationist view

of the nature of truth affords no insuperable obstacles to coherentism. The coordinative linkage between truth and (idealized) coherence is grounded in the fundamental general principles of the matter. — It remains to be shown, however, that “the ancient principle of accord with fact (adaequatio ad rem)’—namely, thesis A above—is also available to the coherentist, who, after all, does not propose to define truth in this way, so that it is not to the coherentist a mere truism (as it is to the adequationist). Accordingly, we must show that this thesis is derivable on. coherentist principles, given that these principles consist not of (P1) and (P2) alone but also the favored truth-determinative axiom (or definition) that is obtained when we conjoin theses 1 and 2: (C)

~

‘S$’ is true — ‘S’c B.

Note that in view of this axiom, we have it that (P1) yields ‘S’ is true —> A(S).

To obtain the converse, consider theprinciple Pl in the special case of , the state of affairs “~S’:

~A(~S) > ~('~S’ cB). By the law of excluded middle, this will entail: A(S) > ~(~S’ cB).

TRUTH AS AN IDEALIZATION

221

By (P2) this yields:

A(S) > ‘S’ cB. By (C) this yields: A(S)— ‘S’ is true.

Together with its converse, as derived above, this provides for (A). QED It follows that an equating of “the (real) truth” with adequation to fact (with how matters actually stand in the world) is also an implicit consequence—in the idealized case—of a coherentist conception of the nature of truth. The coherentist accordingly has no need to renounce adequation. In defining truth in terms of ideal coherence, the principle of adequation (A) remains available to the coherentist as reflecting an essential feature of truth. The idea that principle A characterizes the essence of truth is as available to the coherentist as to anyone else. Let us recall that principle A encapsulates the correspondentist view of the nature of truth as adequation to fact: (A)

‘S’ is true — A(S).

On the other hand, principle C encapsulates the coherentist view of the

nature of truth as ideal coherence:

(C)

‘Sistrueo‘ScB.

The first stage of our argumentation has established: {(LEM), (P1), (P2), (A)} entails (C).

And the second stage of our argumentation has established:

/—

{(LEM), (P1), (P2), (C}} entails (A).

Putting these together, we arrive at:

{(LEM), (P1), (P2)} entails [(A) > (C)]. Given the explanation of “ideal coherence” at issue in the principles P1 and P2 (or, equivalently, the conditions of Completeness and Adequacy), it emerges that adequationism and coherentism are effectively coordinated. The coherentist criteriology of truth is also available to the adequationist. The adequationist view of the nature of truth is also available to the coherentist. The two positions can (under plausible suppositions) be coordinated with one another as deductively equivalent. The pivotal issue of this chapter's first section is thus resolved. The present deliberations indicate that the coherentist criterion of truth as optimal systematization is fully consonant with the definition of truth

222,

.

THE PURSUIT OF TRUTH

fact. Augearedto,the traditional conception of truth as adequation to

ed thentic truth may be characterized essentialistically in terms of idealiz of terms in lly ologica criteri ied identif coherence; putative truth may be

n the two manifest coherence. And the requisite coordination betwee logic.” s rigorou of matter a conceptions can be guaranteed as

en Knowledge (Dordrecht, 7 This appendix is based on chap. 2 of the author’s Forbidd ce Theory of Truth and Coheren The in treated also are 1987). Some of the chapter's ideas Cognitive Systematization.

Thirteen Scientific Truth and the Arbitrament of Praxis Monitoring the Adequacy of Science: The Arbitrament of Praxis The mission of science as a cognitive enterprise is to provide us with information about how matters stand in the world—to estimate the truth about nature and its workings. But this perspective at once poses the question “Can we say anything about how good an estimate of ‘the real truth’ science in fact provides?” Any sensible attempt to assess how effective our cognitive efforts are in their intended mission of getting at the truth of things must face the fact that there is no way to determine the adequacy of our theorizing by seeing how well inquiry is doing in getting at the real truth of things. Any attempt to appraise the adequacy of our theorizing on its own, purely theoretical terms is thus ultimately futile. For we cannot determine the truth about factual matters directly, independently of theorizing inquiry, and then monitor the adequacy of our science by comparing its deliverances with this independently determined truth.! There just is no science-independent means at our disposal for getting at the real truth of things at the level of scientific generality and precision. And this consideration establishes the need for a theoryexternal control on the correctness of our theorizing, some reality-principle that can, independently of our theorizing, serve as a standard of adequacy for our cognitive endeavors. Here, in the final analysis, lies the inseparable difficulty in implementing the otherwise plausible-sounding view that science makes progress insofar as it succeeds in its aim of discovering the actual truth about the world. However, we fortunately can monitor the adequacy of our science obliquely by considering how well we fare when it comes to applying and implementing its professed claims in the sphere of prediction and control. The guiding idea is that the adequacy of our cognitive instruments is to be assessed pragmatically in terms of their effectiveness in helping us to navigate successfully amid the shoals and narrows of a hostile (or at best indifferent) environment. The aim of science is represented by the traditional quartet of description, explanation, prediction, and control. The teleology of scientific inquiry is thus internally diversified and complex, extending across both 1 See the discussion of the so-called Wheel Argument (diallelus) in the author’s Method/ ological Pragmatism (Oxford, 1977), pp. 15-17.

224

.

THE PURSUIT OF TRUTH

the cognitive/theoretical and active/practical sectors. Our truth criteriol-

ogy thus comes to be endowed with a duality of objectives, and the relevant teleology of inquiry is both cognitive and practical.? Truth acceptance is, on the one hand, a determining factor for belief in purely

intellectual and theoretical regards, and on the other, a guiding standard for the practical conduct of life. The two are inseparably interrelated. Nevertheless, the adequacy of our predictive and interventionist praxis is the pivotal factor providing the quality-controlling monitor of our scientific claims. This is so, not because of an instrumentalistic stance that

abandons the pursuit of truth and sees prediction and control as the sole goals of science, but simply and solely because there just is no prospect of any more direct alternative, any immediate comparison of these claims

with the science-independent “real truth” of things. The practical and purposive aspects of cognition thus come to the fore. The governing quality controls of our mechanisms of inquiry—its methods, concepts, and so forth, the whole machinery by which we build up our world picture (knowledge, epistémé, science)—emerge as fundamentally pragmatic, a circumstance to which philosophical pragmatists have always accorded a central role on the epistemological stage. We.are led to the doctrine that effective praxis is the ultimate quality-control arbiter of acceptable theoria: that we must monitor the adequacy of our scientific knowledge by way of assessing the efficacy of its applications. In taking the praxis it underwrites as the salient index of the adequacy of our science, it is imperative to distinguish carefully between the ontological (or existence-oriented) and the epistemological (or cognition-oriented) dimensions. In ontological perspective, the aim of science is un-. doubtedly realistic—to offer us a true account of the what and why of the real world, or at any rate a best-available estimate of the correct account

of these matters. At the level of aim and aspiration, this is undenied and undeniable. Butit is no less undeniable that, epistemologically, the only feasible way to assess the functional adequacy of our efforts at reality

characterizations is through the efficacy of the praxis underwritten bythe methods of inquiry we use in constructing our world picture. The capacity of our cognitive tools to meet their theoretical goals can be controlled only obliquely, by appraising their realization of our practical goals. Thepivotal issue is that of “working out best.” But just what does “best” mean here?

In the Western tradition the governing standards of human rationality are implicit in the goals of explanation, prediction, and preeminently control. (And thus the crucial factor is not, for example, sentimental “at= 2 A further development of these considerations regarding the teleology of inquiry is given in the author’s Scientific Explanation (New York, 1970).

SCIENTIFIC TRUTH

225

oneness with nature’—-think of the magician vs. the mystic vs. the sage

as cultural ideals.) These standards revolve about considerations of practice and are implicit in the use to which our conceptual resources are put in the management of our affairs in the conduct of life. In the Western, Faustian? intellectual tradition, the ultimate arbiter of rationality is represented by a very basic concept of knowledge-wed-to-practice, and the ultimate validation of our beliefs lies in the combination of theoretical and practical success, with “practice” construed in its pragmatic and affective sense. To be sure, success in the domain of praxis is something very different from success in theoretical regards. For practical success lies in the affective order, ranging over the spectrum from physical survival and avoidance of pain and injury on the negative side, to positive satisfactions such as the satiation of physical needs on the other. Even merely intellectually rooted satisfactions, such as the achievement of “predictive success,” are

part of the picture—although only a small part. (After all, even.a worldexternal, disembodied spectator can make predictions about the world

and utter a pleased “aha—there it is” to itself when its predictions work out.) The key factor is that of the success of a creature who must intervene in the course of events to make matters eventuate so as to conduce

to its survival and well-being. Success in the pragmatic sphere is a matter of avoiding the affective mishaps and affective satisfaction by guiding satisfactorily the actions of a vulnerable being emplaced in medias res within a difficult and generally uncooperative world. Is success itself, however, not something theoretical, something that

also lies in the eye of the beholder? As one critic has put it: “But while the actual occurrence ofhappiness or unhappiness, pleasure or pain, etc., is indeed beyond Gur control . . . the same does not seem obviously to hold for our beliefs about such matters. . . . Thus beliefs or judgments about pragmatic success turn out not to constitute genuine input from

the world, but instead... can. . . be arbitrarily manipulated at will, so long as the other elements are appropriately adjusted.”* It would thus be contended that the actual occurrence of pain, suffering, disappointment, and so forth is itself simply a matter of belief. But this seems very dubious indeed. To begin with the extreme case, actual survival is not a matter of belief at all. Nor for that matter is actual bodily injury and physical harm something markedly susceptible to variation in the beliefs of individuals. The power of positive thinking is rather feeble in these regards. “Mind over matter” is a precept whose range of effectiveness has its lim3 Im Anfang war die Tat (In the beginning was the deed), as Goethe’s Faust puts it. 4 This objection is raised in Laurence BonJour, “Rescher’s Idealistic Pragmatism,” Re-

view of Metaphysics 39 (1976): 721.

226

.

THE PURSUIT OF TRUTH

its. In.human affairs, we are, after all, dealing with a conscious being for

whom the affective negativities of actual mishap—and indeed even of merely disappointed expectations—can neither be overlooked nor, in the normal course of things, appraised as other than the negatives that they are. The idea that—outside the laboratory setting of Pavlovian conditioning—we can manipulate arbitrarily and at will the sorts of things that cause people happiness or unhappiness, pleasure or pain is surely implausible. A powerful dictator can certainly get people to do things, but getting them to like it seems to be a good deal less open to manipulation. This line of thought may, however, still seem vulnerable to the following objection: Your pragmatic theory exhibits the characteristic flaw of all instrumentalism,

because, in effect, it simply abandons truth for utility in pivoting the issue of quality control on goal attainment in the practical rather than the theoretical area.

.

:

This objection is gravely mistaken. At the level of particular claims and theses, the present pragmatic approach is not instrumentalistic at all: it continues the orthodox commitment to the truth as such, duly construed as the pivotal object of the enterprise of inquiry and the subject of the usual truth-criteriology of the scientific method. Only at a remove, at the methodological level, do pragmatic considerations enter in. Successful praxis is seen as adequacy indicative with respect to our generalized criteria of acceptance and verification. The link between truth and praxis is not seen as direct, but as mediated by the methodology of inquiry. Our justificatory rule is, Truth claims via methods, and methods via praxis. The burden of those deliberations, then, is that the practical sector is

ultimately in the controlling position as regards cognition—that praxis affords our quality-control monitor over theoria, with applied science

constituting the ultimately commanding factor with respect to theoretical science. Control ranging from mere prediction as minimal (“merely intellectual”) control—the adequate alignment of our expectations with the actual course of events—to the more elaborately modifactory changé in the course of nature through effective intervention in the course of things thus comes to operate as the final arbiter of cognitive adequacy. This . insistence on the cognitive centrality of “control over nature” leads us to an interventionist theory of knowledge, one that sees the issues of monitoring the adequacy of our theorizing to reside ultimately on the side of efficacy in application.® / To be sure, one must give all due credit to the familiar spectrum of 5 The line of reasoning operative here is set out in “Why Relate Success and Thoughtfulness?” chap. 6 of the author's Methodological Pragmatism.

SCIENTIFIC TRUTH

227

theoretical/systemic desiderata for our belief structures: comprehensiveness, explanatory coherence, simplicity, and the other parameters of cognitive systematicity.° But this by itself is not enough: such theoretical factors count but do not decide. For our inquiry procedures not only provide neat systems of beliefs but also thereby lead us to stub our toes upon the harsh realities of implementing praxis. Everyone is familiar with the occasional surfacing, even today, of some occult or pseudoscientific methodology leading to factual claims of the strangest sort—flat-earth theories, for example. It is often striking here how beautifully everything meshes at the theoretical level—one bit of strangeness being supported by others. A seamless fabric of self-supportive absurdities is readily imagined. The crunch comes only with the tough question “Does this approach to the warranting of epistemic claims actually enable its proponents to navigate more successfully and effectively amid the rocks and shoals of this world by underwriting a complex and diversified body of successful applications?” A theory-external quality control upon cognition is ultimately vital to our justificatory deliberations because it blocks the prospect of a futile spinning around in reality-detached cycles of purely theoretical gyrations. Someplace along the line of justification there must be provision for a correlative contact with an uncooperative and largely unmanipulable reality—an agency that furnishes a “reality principle” quite independently of the drift of our theorizing. This crucially requisite principle is provided for in the framework of the present approach by the factor of the success consequent upon implementing action. Goal attainment— successful goal-pursuing praxis (regardless for the moment of the intrinsic character of these goals)—is the ultimate guarantor of validity of the products of humankind’s endeavors at the acquisition of empirical knowledge.

Why This Link between Pragmatic Efficacy and Truthfulness? But just why should the pragmatic success engendered by its deliverances be taken to justify an inquiry method and count toward establishing its legitimacy as a criterion of acceptability or verification? Why should its underwriting of a more effective praxis be pivotal for the theoretical issue of a method's greater adequacy in depicting the real world?

In embarking on this problem, it is useful to begin with the traditional objection against pragmatism that it might well prove highly successful § See chap. 1 of the author’s Cognitive Systematization (Oxford, 1979) for a fuller development of these issues. :

228

:

THE PURSUIT OF TRUTH

to act on some (quite incorrect) thesis. This perfectly possible prospect makes one rightly hesitant to maintain that the successful practical implementation of a thesis is in general an adequate basis for holding it to be true. After all, action that proceeds on beliefs that are false and that fail

to capture “what is actually the case” can on occasion—or even frequently—eventuate as pragmatically successful, due to chance or good luck or kindly fate or whatever. Individual theses may well manage to survive the test of praxis. However, the situation is radically different when what is at issue is not an isolated action or a particular accepted thesis (or even a cluster of them) but rather a general, across-the-board process or policy of belief validation. The comparatively greater effectiveness of a method in delivering more successfully implementable theses into our hands systematically and in a wide variety of contexts is clearly an indication of its comparative superiority as a means of inquiry. One critic has advanced the following objection: No doubta truth criterion which yielded results which were utterly discordant with reality would not lead to pragmatic success; but how close to the truth such results would have to be in order to be successful would seem to depend enormously on the particular world in question and the particular needs and

purposes of the beings which inhabit it. It is far from obvious that acriterion which produced only very approximately accurate results could not, in some circumstances, be more pragmatically successful than one whose results were more accurate.’

But this objection is farfetched unless the scope of the criterion at issue is very narrow indeed. Its force is blunted by the generality and universality of a methodology of inquiry. For the range at issue with scientific inquiry at a totally generalized systemic and methodological level is to be literally boundless: no factual issue is to lie outside its intended province. With an inquiry procedure of such comprehensive scope, it is.clear that probatively irrelevant side effects by way of fortuitous bonuses or disasters will become canceled out in the larger scheme of things. That a mistaken or unwarranted body of scientific commitments might prove really successful at this synoptic level of generality is a prospect so farfetched that it can be dismissed with confidence. It is also instructive to contrast our approach with that discussed by Hilary Putnam under the title “explanatory realism.”® On Putnam’s approach, the veridicality of the products of our inquiry methods turns on pragmatic argumentation of the form: 7 BonJour, “Rescher’s Idealistic Pragmatism,” p. 721. § See Hilary Putnam’s Meaning and the Moral Sciences (Oxford, 1978).

SCIENTIFIC TRUTH

229

Our epistemic criteria entitle us to assert the claims of the corpus S (i.e., of science). The products of our epistemic criteria are successfully implementable.

Therefore: The claims of the corpus S are indeed true. Now as Putnam sees it, this is an argument of an essentally inductive

“inference to the best explanation” mode. Its conclusion that the theses of science are veridical is seen as itself a quasi-scientific theory that effectively explains the successful implementability of these scientific claims in prediction and intervention. The presumptive truth of our scientific claims—their authenticity vis-a-vis the real world—is an inductive product, a scientific theory, an inference to the best explanation. The present approach is quite different, however, because it casts the second premise of the above argument in a very different role. It sees the conclusion as emerging from the first premise alone: if our epistemic criteria authorizes us to assert a thesis, then the truthfulness of this thesis

is something we stand committed to ipso facto, simply because that is exactly what is at issue in these epistemic criteria as standards for acceptance as true. But of course the further question of the appropriateness or validity of those epistemic criteria still remains. On this issue, the pragmatic efficacy of the second premise has a crucial bearing. The relative pragmatic efficacy of our implementation of the theses they validate is seen as an index of the relative acceptability of alternative systematic

methods of inquiry and substantiation. The argument thus breaks apart into two stages, with a crucial methodological level inserted as mediating link between pragmatic efficacy and truth. The mediation of methodology-—of a layer of considerations at the epistemological level—always intervenes between pragmatic considerations and considerations regarding the acceptability of factual claims. To reemphasize, on the methodological version of pragmatism, pragmatic considerations are never brought to bear in the acceptability of factual claims directly, but only in the issue of the methods, criteria, and standards that govern the acceptability of such theses. The resultant systemic situation is accordingly as depicted in figure 13.1, with the more complex picture at the bottom replacing the simple picture at the top. By their very nature as such, methods inherently function at a wholesale rather than a retail level. The success of a method of inquiry accordingly must be construed in systematic terms: working on some limited number of occasions does not entail working in general, and failing on some occasions is not necessarily completely invalidating. Its applicative success hinges on how our science fares in general over the whole gamut of its implementations. The range and versatility of an inquiry procedure

230

.

THE PURSUIT OF TRUTH

Figure 13.1. The Naive View and the Revised View of the Inquiry/Reality Relation 1. The Naive View Inquiry = ------ >

Reality

2. The Revised View Inquiry ------ >

Putative reality

—Implementation in action Successful praxis —Justificatory validation (via successful applications) of our claims to depict reality correctly.

as an instrumentality of opinion formation across the entire spectrum of human beliefs is too obvious to need much elaboration. Generality is here tantamount to open-endedness: our systematic commitments oper-

ate across the board in an enormous variety of areas of application and a literally innumerable proliferation of particular instances. Here all of the safeguards built into the statistical theory of the “design of experiments” come into play with respect to the probative significance of the number and variety of instances. It is improbable to the point of absurdity that a systematic success across so broad a range should be gratuitous. Fundamental mistakes at this level of generality are bound to have applicative repercussions across alimitless frontier that would not only be discernible but would ultimately prove catastrophic. The pervasiveness and open-endedness of an inquiry method is bound to wash out the influence » of fortuitous factors extraneous to the actual merits of a method. Yet another line of objection runs as follows: The possibility of a pragmatically successful exploitation of the instrumentalities afforded us by nature is surely compatible with the most profound ignorance of the workings of things. Think of an analogy: One can learn to drive a

car or operate a cyclotron most successfully and yet maintain a profound‘ignorance as to the workings of these devices. Successful manipulation requires

only the most superficial sort of knowledge: what happens when we pull which . levers, so to speak. Successful action may thus be compatible with massive cognitive inadequacy.

a

This objection is sound enough in its own way, but its way is not that of our methodological pragmatism. For this objection too is blunted by the tremendous generality of the pragmatic success at issue with an inquiry

procedure. Such success is not just a matter of a manipulative control of the surface of things; it is something that cuts through to the innermost internals. If one controls effective practice at the superficial level (know-

SCIENTIFIC TRUTH

;

231

ing merely how to start and drive and stop the car), one may indeed know

precious little of what makes it tick. But what is needed for the analogy to hold is a practical mastery that is totally comprehensive and systematic, one that includes the ability to repair all manner of breakdowns and malfunctions, a capacity to design, assemble, and maintain it and its components and their components, the ability to modify and improve it for more effective operation, and so on—in short, a capacity for pragmatically effective action in virtually endless comprehensiveness and detail. That all this should be afforded by a belief-underwriting inquiry procedure that would not on this basis be cognitively adequate is a circumstance that boggles belief. In the circumstances of the case, the systematic applicative success of their deliverances must be viewed as strongly indicative of the methodological appropriateness of an inquiring process. It is implausible to suppose—and eccentric to maintain—that a method of inquiry that outperforms available alternatives in facilitating the prediction and control of nature’s eventuations should yield results that do not merit preferential adherence.®

The Evolutionary Dimension The cognitive methods and substantiative procedures we deploy for structuring our view of reality evolve selectively by a historic, evolutionary process of trial and error, analogous to the mutations affecting the bodily mechanisms by which we comport ourselves in the physical world. Accordingly, cognitive methods develop subject to revision in response to the element of success and failure in terms of the teleology of the practice of rational inquiry. The central issue is a matter of survival of the fittest, with fitness assessed in terms of the practical objectives of the rational enterprise—particularly in the guidance of our practical affairs. Their evolutionary development proceeds by explicitly rational selection, rather than by some purpose-indifferent process analogous to the natural selection of the standard biological case.!° As changes are entertained (under the pressure of necessitating circumstance), one methodological instrument may eventuate as more fit to survive than another, because it ® For a further exploration of these issues, see the author’s Methodological Pragmatism and A Useful Inheritance (Savage, Md., 1989). 10 OF course, even in the most orthodox and rigoristic of biological evolutionisms, one

cannot but recognize that with the historical emergence of human beings, the purposive aspect must be given a role upon the evolutionary stage, since purpose and final causation

here overshadow efficient causation and introduce a Lamarckian aspect into evolution. See ’ G. G. Simpson, The Meaning of Evolution (New Haven, Conn., 1949), pp. 289-96.

232

:

THE PURSUIT OF TRUTH

answers better to the range of relevant purposes. An epistemological theory that emphasizes the pragmatic aspect of our cognitive resources is wedded to an evolutionary model of cognitive development through rational selection via variation and the selective retention on the basis of “the lessons of experience.” The major criticisms of evolutionary theories of epistemic validation all have a common core and purport. All proceed by dwelling heavily on the

distinction between the cognitive/intellectual side of human affairs and their affective/physical side. They insist that there is no good reason of theoretical principle why conceptions and beliefs conducive to human welfare thereby acquire substantiation. The question is pressed, “How—in the face of counterexamples of the sort readily constructed— can one validly maintain a linkage between practical advantage and theoretical truth?’ Survival-conducive beliefs are surely not invariably (and perhaps not even generally) correct; nor need correct beliefs necessarily prove survival conducive. This objection posesa serious difficulty for any program of evolutionary epistemology. It parallels exactly the usual attacks against linking the

practical utility of beliefs and their theoretical truth (as with the thesis pragmatism of William James). For essentially the same weakness indeed affects both these doctrines: their orientation toward theses. , But the telling force of such criticisms of an evolutionary epistemology

is blunted once we take the methodological turn. For then considerations of survival-conduciveness are no longer asked to militate for the adequacy 11 The fundamentals of this line of thought go back to Plato’s myth of the cave (in bk. 2 . of the Republic), with its moral that too close a heed of the realities and practicalities of this

world will impede those unconstrained excursions of intellect through which alone an insight into theoretical truth can be obtained. Henri Bergson developed at length this theme that only by separating the mind from the restraints of quotidian action-directedness can one achieve a proper understanding of reality, as there is no reason to think that a cognitive apparatus that has evolved for survival purposes should provide a handle on ‘theoretical truth (see his Essai sur les donnéés immédials de la conscience, 8th ed. [Paris, 1889], as

well as his Introduction a la métaphysique in the Revue de la métaphysique et de morale [1903], 2:1-183). In recent days this position—-that the aims of survival promotion and knowledge enhancement may even be antithetical—has been vigorously argued by Herman Tennessen. See his essays “Knowledge versus Survival,” Inquiry 16 (1973): 407-14, and “On Knowing What One Knows Not,” in The Psychology of Knowing, ed. J. Royce and W. Roseboom (New York, 1972), pp. 111-60.

12 Bertrand Russell’s criticisms of the pragmatists include many quaint examples, among them this: “Dr. Dewey and I were once in the town of Changsha during an eclipse of the moon; following immemorial custom, blind men were beating gongs to frighten the heavenly dog, whose attempt to swallow the moon is the cause of the eclipses. Throughout... thousands of years, this practice of beating gongs has never failed to be successful; every eclipse has come to an end after a sufficient prolongation ofthe din” (“Dewey's New Logic,” in The Philosophy ofJohn Dewey, ed. P. Schilpp [New York, 1939], p. 152.

SCIENTIFIC TRUTH

:

233

(i.e., truthfulness) of specific beliefs or belief systems, but for the adequacy of methods of inquiry. This methodological turn is crucial here, thanks to the fundamental difference between methods and theses as regards the proper modes ofjustificatory argumentation. For the argument rests on the presence of a cogent link between pragmatic success and theoretical adequacy (truthfulness), a linkage that is forthcoming in the case of methods, but not in the case of theses (as per the preceding argumentation). Accordingly, the presently envisaged strategy of legitimation resorts to a method Darwinism and not a thesis Darwinism. To be sure, if a bounteous nature satisfied our every whim spontaneously, without effort and striving on our part, the situation would be very different. For then the beliefs that guide and canalize our activities would generally not come into play—they would remain inoperative on the sidelines, never being put to the test. There would then be no need for active (and thought-guided) intervention in the natural course of things within an uncooperative (at best indifferent, at worst hostile) environment. But as things stand we are constantly called upon to establish varying degrees of control over nature to satisfy even our most basic needs (to say nothing of our virtually limitless wants). We. cannot reasonably view nature as a friendly collaborator in our human efforts, systematically shielding us against the consequences of our follies and crowning our cognitive endeavors with a wholly undeserved success that ensues for reasons wholly independent of actual adequacy vis-a-vis the intended range of purpose. The human condition is such that active intervention in the course of events is constantly required (and even the noninterventionist process of “letting things take their course” becomes, in such a context, simply another mode of control). Moreover, human agency produces a flow of consequences that rebound back upon the agent, ultimately producing satisfaction or frustration. This interactionism is a crucial part of the justificatory background. The facilitation of purpose-guided interaction with nature—of acting so as to produce successfully the intended results of our efforts—is a vital part of the cluster of factual considerations that establish the cognitive bearing of pragmatic success. The Darwinian aspect is particularly crucial here. It is no more miraculous that the human individual can understand nature more or less adequately than it is that the human eye should be able to see it more or less correctly. Peirce’s insight holds true: our cognitive methods are able to earn credit as giving a trustworthy picture of the world precisely because they evolve under the casual pressure of that world. Consider again J. S. Mill’s contention (advanced in his famous treatise On Liberty) that the competition among schools of thought is akin to that between biological varieties: the rivalry among ideas for acceptance _ amounts to a struggle for existence (i.e., perpetuation, or continued ex-

234

.

THE PURSUIT OF TRUTH

istence), a struggle in which those beliefs that are the fittest—namely, those that represent “the truth”—must finally prevail. The survival of beliefs within an intellectual community is viewed as parallel with biological survival. (C. S. Peirce’s theory of truth can be seen as a systematic development of this line of thought.) It is perhaps unnecessary to dwell, at this stage, on the shortcomings of this unduly optimistic view. Acceptance of theses is all too often governed by extrarational factors: familiarity, faddism, social pressures, bandwagons, propaganda (“thought control”), and so forth. And there can be no secure assurance that such perturbations will inevitably or even usually be eliminated in the course of time. The view that the inherent attractions of the truth assure its ultimate victory in a struggle for acceptance among the beliefs of imperfect humans is thus overly hopeful. However, its most serious shortcomings stem from its operation at the thesis level. An evolutionary approach seems to have substantially greater promise at the more general level of methods; where the force of the above-mentioned disabilities is removed or attenuated. The generality and open-endedness of an inquiry method furnish it with a capacity to wash out the influence of these fortuitous and extraneous factors. The survival of a cognitive method in a group of active inquirers who are rational and realistic is a cogent and powerful indication of adequacy:13 The standard objections to epistemic Darwinism all come to grief on the specifically methodological ground upon which our present approach takes its stand. To say all this is not to deny that some degree of evidential support is indicated by the survival of a thesis in a community of rational inquirers. But the link is relatively tenuous and oblique—not direct, but mediated through the argument that thesis survival over a considerable period is presumptively adequacy indicative. Method Darwinism does not carry any very strong or emphatic thesis Darwinism in its wake and is relatively secure against the sorts of objections applicable to it. ~ The process of rational selection implements an evolutionary model that renders pragmatic efficacy correlative with historical survival. Butof course people are not all that rational—they have their moments of aberration and self-indulgence. Might not such tendencies selectively favor _ the survival of the fallacious rather than the true and slant the process of ~ evolution in deteriorating directions? Peirce certainly recognized this prospect. “Logicality in regard to practical matters... . is the most useful 3 It is perfectly conceivable that in the course of intellectual history a quite untenable thesis may long prevail, lodged securely in its peculiar ecological niche simply because no one has found it worthwhile to dislodge it with some rival. One need not reckon with this sort of aberration in the methodological case under the conditions basic to our “metaphysical deduction.”

SCIENTIFIC TRUTH

235

quality an animal can possess, and might, therefore, result from the action of natural selection; but outside of these it is probably of more advantage to the animal to have his mind filled with pleasing and encouraging visions, independently of their truth; and thus, upon unpractical subjects, natural selection might occasion a fallacious tendency of thought.”* However, the methodological orientation of our approach again provides a safeguard against an unwarranted penchant for such fallacious tendencies. At the level of individual beliefs “pleasing and encouraging visions” miglit indeed receive a survival-favoring impetus. But this unpleasant prospect is effectively removed where a systematic method. of inquiry in concerned—a method that must by its very synoptic nature press deeply into the sphere of the pragmatically effective. This line of thought points toward what is in fact one of the greatest merits of the methodological approach. It is surely an obvious desideratum to meet the profoundly skeptical objection that human cognition is

fundamentally inadequate to the attainment of theoretical truth because it is the product of an evolutionary process attuned to the needs of practice. Darwin is himself the father of this thought, raising the question “Can the mind of man, descended, as I believe, from the lowest animals,

be trusted when it draws such grand [scientific and philosophical] conclusions?”’> One commentator observes that “Darwin’s ‘horrid doubt’ as to whether the convictions of man’s evolved mind could be trusted applies as much to abstract truth as to ethics.”!6 And another goes on to draw the inference that “there is surely something wrong with a theory which, at its very root, invalidates itself.”1”7 A whole host of variations have been

played on this theme of the possible or probable theoretical inadequacy of practice-oriented cognition.'® A-theory that casts pragmatic efficacy in the role of arbiter of cognition leaves no room for neurotic fears that the intellectual resources of praxis may be found wanting before the tribunal of theoria. 4 C, S. Peirce, Collected Papers, 8 vols., ed. C. Hartshorne et al. (Cambridge, Mass.,

1931-50), 5:366. On the broader aspects of Peirce’s views on evolution, see Thomas A. Goudge, “Peirce’s Evolutionism—after Half a Century,” in Studies in the Philosophy of Charles Sanders Peirce, 2d ser., ed. E. G. Moore and R. S. Robin (Amherst, Mass., 1964), pp. 323-41.

45 Charles Darwin, Autobiography (London, 1929), p. 93. Various post-Darwinian writers developed this theme at length. See, e.g., G. H. Howison, The Limits of Evolution (New

York, 1901).

16 David Lack, Evolutionary Theory and Christian Belief (London, 1957), p. 104. 17 Marjorie Grene, The Knower and the Known (New York, 1966), p. 200. 18 See, e.g., Tennessen, “Knowledge versus Survival.” This develops in epistemic terms “the Bergsonian position that a language with a conceptual frame . . . fashioned to fit a survival project, or other practical or quotidian objectives, is bound to fail when it comes

to unpractical, ‘useless,’ antibiological, trans-quotidian contemplations of what there is and _ what we can know” (p. 113).

236

.

THE PURSUIT OF TRUTH

To be sure, a satisfactory evolutionary pragmatism must be predicated upon the shift from a Darwinism of theses to a Darwinism of methods. For now we can supposea trial-and-error model that operates not with respect to individual theses but with respect to across-the-board methods of thesis substantiation. On such a model, the course of evolution, however slow in the initial stages, is able to provide for very rapid eventual progress when once any headway at all is made. A cognitive evolutionism that, like the presently contemplated methodological approach, is inherently oriented toward generalized instrumentalities can successfully pull its way out of the quicksands in which a thesis-oriented evolutionary theory becomes enmired.}9 In effect, the shift from theses to methods (namely, methods for thesis substantiation) enables us to have it both ways. We avoid occultism by relying at the methodological level uponastrictly trial-and-error mecha* nism of learning. And we avoid the rational impotence of an inability to account for the actual course of scientific progress. The combination of a

model of method learning based on the blind groping of trial and error and of thesis learning based on the use of methods makesit possible to have the best of both worlds.2° It means that the venture of thesis validation is not condemned to blind trial and error but is guided by heuristic principles of method, involving the use of methods that have:proven their effectiveness in the past, and whose application in present conditions thus embodies a fundamentally inductive commitment.” Cardano, a sixteenth-century mathematician with a Renaissance penchant for boasting, maintained that a monalphabetic cipher of his own invention was unbreakable because of the large number of possible solutions that must be tested. (A monalphabetic cipher is the simplest kind; _ 18 By the time of his retrospective contribution to The Philosophy of Karl Popper, 2 vols., ed. P. A. Schilpp (La Salle, Ill., 1974), Popper himself seems to have reached much the same conclusion. “But although the blind man who searches for a black hat may bring some order into his trials, the order is not given to him; he may choose or invent one order (method) first, and a different order later; and these choices will be trials too—even though

on a higher level. (They may, but need not, be influenced by his earlier experience in somewhat similar or in very different situations)”, (vol. 2, p. 1061). 2 The methodological approach can thus lay claims to resolving the issueperceptively posed by D. T. Campbellin the following terms: “Popper has, in fact, disparaged the com- mon belief in ‘chance’ discoveries in science as partaking of the inductivist belief in directly learning from experience. . . . That issue,and the more general problem of spelling out in detail the way in which a natural selection of scientific theories is compatible with a dogmatic blind-variation-and-selective-retention epistemology remain high priority tasks for the future” (“Evolutionary Epistemology,” in The Philosophy of Karl Popper, 2 vols., ed. P. A. Schilpp [LaSalle, Ill., 1974], 1:436). The present theory provides a natural basis for © combining a natural selection process at the level of theories with an epistemology of blindvariation-and-selective-retention at the level of methods. 21 Albeit one of a regulative sort. See chap. 3 above.

SCIENTIFIC TRUTH

237

each letter of the original message is replaced by one and only one cipher letter consistently throughout the whole message.) Since there are 26 letters, there are 26 factorial different ways of pairing plaintext letters with cipher letters, or approximately 1.1 x 108 different possible solutions, a consideration that led Cardano to feel the safety of numbers. Yet

with a few hours’ instruction in frequency analysis and anagramming, most amateurs can solve these ciphers in a matter of minutes. Such is the potency of method. One leading authority on cryptology, commenting on amateur inventors of cipher systems, has said: “Many inventors also invoke the vast number of combinations of keys afforded by their system as proof of its invulnerability. To exhaust the possible solutions would take eons, they contend, unwittingly using the same argument . . . as Cardano .. .—and with the same lack of validity. For, as Shannon has shown, the cryptanalyst does not go after these possibilities one by one. He eliminates millions at a time. Moreover, the trials progress from the more probable to the less probable hypotheses, increasing the cryptanalyst’s chance of striking the right one early.” It is thus important, even when we recognize the utility of a trial-anderror technique, that this be thought of not as a blind groping among all conceivable alternatives but as a carefully guided search among the really promising alternatives.“ Inquiry is not a process of setting a random generator to work to produce hypotheses for testing. Hypotheses emerge not from random combinations but from the detection of patterns in the empirical data. What is at issue is a “nested hierarchy” of heuristic processes that reflect the funded experience of the past. We thus need never confront the theoretically proliferated manifold of alternative hypotheses empty-handed but can filter it through these basic methodological presumptions that in themselves represent a heuristic protomodel of the world by whose means yet further knowledge can be secured. 2 David Kahn, The Codebreakers (New York, 1967), p. 766. / 23 This represents a crucial disanalogy between biological and cognitive evolution indicative of the quasi-vitalistic character of the latter. In biological evolution the mutations that

actually arise fall across the entire spectrum of possible alternatives with equal probability, and so the direction of evolution is not determined by thedirection of mutation. “It is emphatically selection, not mutation, that determines the direction of evolution,” and if

this were not so, then “it would be necessary to suppose that such mutations must be predominantly favourable” (Sir Gavin de Beer, “The Darwin-Wallace Centenary,” Endeavour 17 [1958]: 68). In the case of cognitive evolution viewed from the standpoint of thesis acceptance, the case is exactly opposite: the actualization of possible mutant alternatives is

probabilistically skewed, favorable mutations predominate, and the direction of evolution is governed as much by the inherent selectivity of mutation as by selection proper. But as we shall see, in the cognitive case—unlike the biological—there is nothing occult about any of this, because one can embed the vitalistic features of epistemological evolution at the thesis level within an orthodoxly randomized and blindly unguided evolutionary model at the methodological level.

238

.

THE PURSUIT OF TRUTH

Given the preceding critique of thesis Darwinism, someone may well object to the presently projected method Darwinism as follows: Your argument against thesis Darwinism is (put roughly) that the pace of progress of science is too rapid to be plausibly accounted for by a process of trialand-error selection among the available alternatives. But does this argument

not hold against your own position? The success of science certainly indicates very substantial sophistication on the side of methodology, and is this not every bit as difficult to account for along the lines of a Darwinian trial-and-error theory as is progress on the side of accepted theses?

Clearly if this objection held, it wouldbeserious; fortunately, however, it is seriously flawed. For the argument that not enough time is available for the accretion of observed progress by more or less random improvements tells very differently as between theses and methods. For thesis Darwinism requires a tremendous number of successive improvements,

given the immense range of the claims at issue in the buildup of the sciences. It is this enormous number of successive iterations of the selec-

tion process that demands so much time on the Darwinian account and that makes the rapid progress of science appear virtually miraculous from the Darwinian point of view. However, there is—by contrast—no reason why methodological improvement cannot proceed at a glacially sedate evolutionary pace. Then, when at last an even modestly effective method has finally been devised, any further development can clearly proceed with extreme rapidity. Think of an analogy: a vast long and stumbling process may have lain behind humankind’s ultimately successful development of the technological method for human flight, but once the rudimentary beginnings of the venture were in hand, the further development of sophisticated aerial transport proceeded with astonishing speed, and yet in an almost routine way. There is no reason to believe that the case of cognitive rather than technically manipulative methods should be any different in this regard. Once an even partially adequate method for the testing of factual theses has been contrived, there is every reason to think that human ingenuity will devise suitable occasions for putting it to use—to the great advantage of the rapid progress of knowledge. At the level of theses—of envisageable alternative explanatory hypotheses—one ~ faces an embarrassment of riches that makes effective progress through randomized selection-processes unintelligible on any basis that does not call for suppositions rationally unpalatable to any mind of empiricist inclinations. But the situation is very different in the methodological case. Once an even modestly satisfactory inquiry method is at hand, progress at the thesis level can be very swift because of the inherent power and generality of such a method. Moreover, thanks to the cyclically self-cor-

SCIENTIFIC TRUTH

239

rective aspect of such a method, further substantial progress in the methodological side itself becomes a real prospect.“

The “Hegelian Inversion” Once More The procedure labeled above as the Hegelian Inversion proposes that we not evaluate an inquiry procedure by the truth of its results but, conversely, assess the truthfulness of the results in terms of rational merits of procedure that underwrites them—as determined by factors that are in part internal, systematic, and coherentist, but that are predominantly external, applicative, and pragmatic. Such an inverted perspective sees any effort to validate the propriety of our inquiry procedure in terms of the truthfulness of its products as picking up the wrong end of the stick— one does not approach warrant by way of truthfulness, but truthfulness by way of warrant. Given this inversion, the linking equation between truthand warrant need not be abandoned, but it must now be viewed in a very different light. Warrant must be taken as the independent variable and truthfulness as the dependent variable, with our inquiry procedures not seen as warranted because truth producing, but rather presumed to be truth producing because of their rational warrant. Precisely because the later, “progressive” stages of the application of our inquiry procedures are more fully warranted on the basis of the ampler and more successful body of praxis that they underwrite, we can take the stance that it is rational to view their deliverances as better qualifiedfor endowment with the presumption of truth. It is on this basis alone that we can be increasingly confident that our currently accepted picture of nature affords a comparatively better estimate than our past pictures do. Accordingly, our scientific commitments are seen as having two interrelated aspects: on the one hand, they are mere estimates of the truth and not definitive demonstrations thereof, but on the other hand, they are responsible estimates of the truth that rest on an increasingly more simply authenticated methodological basis. Does such an approach indicate a “realistic” view of the products of our inquiries as revealing “the real truth”? Yes and no. The fact is that it constrains us to a certain view of inquiry, namely a commitment to the idea that the products of well-conducted inquiry qualify for acceptance as our best estimate of the real truth. But given that we have no inquiryindependent access to “the real truth,” the fact remains that we cannot % Although stated flatly and without much development here, this point is actually one of great importance.

240

.

THE PURSUIT OF TRUTH

of stricter rego beyond this estimational perspective to claim some sort tive article. In lationship between our putative truth and the real, defini and reality remains the best idealistic tradition, the relation of cognition

tely charactersubject to the epistemic mediation of a sort most adequa

return to this ized in terms of the idea of estimation. (Later chapters will issue of realism in greater detail.)*

ism. Some of its ideas 25 This chapter is based upon the author's Methodological Pragmat

1986), Scientific Realism (Dordrecht, are also treated in Empirical Inquiry (Totowa, N.J.,

A Useful Inheritance (Savage, Md., 1990). 1987), and

Part IV REALISM AND IDEALISM

Fourteen Cognitive Limits The Cognitive Inexhaustibility of Things Our thought about the real things of this world presses outward beyond the limit of any restrictive boundaries. From finitely many axioms, reason can generate a potential infinity of theorems; from finitely many words, thought can exfoliate a potential infinity of sentences; from finite data, reflection can extract a potential infinity of items of information. Even with a world of finitely many objects, the process of gaining information about these objects can, in principle, go on unendingly. One can inquire about their features, the features of these features, and so on. Or again, one can consider their relations, the relations among those relations, and

so on. Thought—abstraction, reflection, analysis—is an inherently ampliative process. As in physical reflection, where mirror images can reflect one another indefinitely, so mental reflection can go on and on.

Givena start, however modest, thought can advance ad indefinitum into new conceptual domains. The number of true descriptive remarks that can be made about any actual physical object is theoretically inexhaustible. For example, take a stone. Consider its physical features: its shape, its surface texture, its chemistry, and so forth. And then.consider its causal background: its genesis and subsequent history. Then consider its functional aspects as relevant to its uses by the stonemason, the architect, the landscape decorator, and so on. There is, in principle, no theoretical limit to the different lines of consideration available to yield descriptive truths about a thing, so that the totality of potentially available facts about any real thing whatever is in principle inexhaustible. The circumstance of its starting out from a finite basis does not mean that it need ever run out of impetus (as the example of Shakespearean scholarship seems to illustrate). All the same, language-bound thought has its limitations. It is helpful to introduce a distinction at this stage. If we adopt the standard conception of the matter, then a truth is something to be understood in linguistic terms—the representation of a fact through its statement in some actual language. Any correct statement in some actual language formulates a truth. (And the converse obtains as well: a truth must be encapsulated in a statement and cannot exist without linguistic embodiment.) A fact, on the other hand, is not a linguistic entity at all

244

:

REALISM AND IDEALISM

but an actual circumstance or state of affairs. Anything that is correctly statable in some possible language presents a fact. Every truth must state a fact, but in principle it is possible that there will be facts that cannot be stated in any actually available language and

thus fail to be captured as truths. Facts afford potential truths whose actualization as such hinges on their being given a linguistic formulation. Truths involve a one-dimensional possibilization: they include whatever can be stated truly in some actual language. Facts, on the other hand, involve a two-dimensional possibilization—they include whatever can be stated truly in some possible language. Truths are actualistically language-correlative, while facts are possibilistically \anguage-correlative.? Accordingly, it must be presumed that there are facts that will never be formulated as truths, though it will obviously be impossible to give concrete examples of this phenomenon.® Now propositional knowledge regarding matters of fact (including belief and conjecture) is always a matter of linguistically formulable information. But we have no alternative to supposing that the realm of facts regarding the real things of this world is larger than the attainable body of knowledge about them—regardless of whether that “we” is construed distributively or collectively. It is not very difficult to see why this is so. As long as we are concerned with information formulated in languages of the standard (recursively developed) sort, the number of actually articulated items of information (truths or purported truths) about a thing is 1 The position at issue thus has no quarrel with P. F. Strawson’s precept that “facts are what statements (when true) state” (“Truth,” in Proceedings of the Aristotelian Society, supp. vol, 24 (1950), p. 136). Difficulty would ensue only if an “only” were inserted.

2 But can any sense be made of the idea of merely possible (that is, possible but nonactual) languages? Of course it can! Once we have a generalized conception (or definition) of a certain kind of thing—be it a language or a caterpillar—then we are inevitably in a position

to suppose the theoretical prospect of items that meet these conditions are over and above those that in fact do so. The prospect of supposing the existence of certain “mere” possibilities cannot be denied—that, after all, is just what possibilities are all about. (in English, we have no single-word verb “to make possible,” akin to the German erméglichen. To adopt “possibilize” and the related noun and adverb forms would perhaps be sensible and certainly convenient.) . 3 Note, however, that if a Davidsonian translation argument to the effect that “if it’s sayable at all, then, it’s sayable in our language” were to succeed—which it does not—then the matter would stand on a very different footing. For it would then follow that any possible language can state no more than what can be stated in our own (actual) language. And then the realm of facts (i-e., what is [correctly] statable in some possible language) and of that of truths (i.e., what is [correctly] statable in some actual language) would necessarily coincide. Accordingly, our thesis that the range of facts is larger than that of truths hinges crucially upon a failure of such a translation argument. (See Donald Davidson, “The Very ~

Idea of a Conceptual Scheme,” Proceedings and Addresses of the American Philosophical Association 47 [1973-74]: 5-20, and also the critique ofhis position in chap. 2 of the author’s Empirical Inquiry [Totowa, N.J., 1982].)

COGNITIVE LIMITS

245

always, at any historical juncture, finite. And it remains denumerably infinite even over a theoretically endless long run.* The domain of truth is therefore, in principle, sequentially enumerable, but that of fact is not comparably manageable. Reality, as we conceive of it, outruns any such limitation. It is instructive to consider why this is so. Let us begin with the idea that an object X is “cognitively inexhaustible” if no finite list of truths about X exhausts the totality of true facts about it. Then it is easy to see that even an infinite list of distinct truths about X will fail to capture the totality of fact about it. For let the series Sis fa fs.-

represent a supposedly complete listing of facts. This listing is, by hypothesis, such that (1) each of its members represents a truth about X, and (2) each successive member adds something new, sothat, for any i, ~(C;—fiz.)

where by definition C; = fi, &KfL&... &f, and “—»” represents logical implication.

To show that the infinite f-list does not manage to register the totality of true facts about X, consider the infinite conjunction of all the f;

C=fehk&.... This clearly represents yet another true fact about X. But by the aforementioned condition of novelty, it is clear that this fact itself cannot occur anywhere on the f, list. This reductio ad absurdum of our hypothesis of enumerative completeness indicates that the facts about a thing are necessarily too numerous for explicit enumeration. Once we recognize that real things are cognitively inexhaustible in that the true facts about them cannot be encapsulated in any finite body of contentions, then we must acknowledge that even infinite listings will prove insufficient. It might be possible, however, to have latent or implicit knowledge of an infinite domain through deductive systematization. After all, the finite set of axioms of a formal system will yield infinitely many theorems. And so, it might seem that when we shift from overt or explicit to implicit or tacit knowledge, we secure the prospect of capturing an infinitely diverse implicit knowledge-content within a finite explicit linguistic basis through recourse to deductive systematization. The matter is not, however, quite so convenient. The totality of the

deductive consequences that can be obtained from any finite set of axioms is itself always denumerable. The most we can ever hope to encompass by any sort of deductively implicit containment withinafinite basis 4 See Philip Hugly and Charles Sayward, “Can a Language Have Indenumerably Many Expressions?” History and Philosophy of Logic 4 (1983): 73-82.

246

:

REALISM AND IDEALISM

of truths is.a denumerably infinite manifold of truths. And thus as long as implicit containment remains a recursive process, it too can never hope

to transcend the range of the denumerables, and so cannot hope to encompass the whole of the nondenumerable range of descriptive facts about a thing. Moreover, even within the denumerable realm, our at-

tempt at deductive systematization runs into difficulties: as is known from Kurt Gédel’s work, one cannot even hope to identify—by any recursive, axiomatic process—all of the inherently denumerable truths of arithmetic. It is one of the deepest lessons of modern mathematics that we cannot take the stance that if there is a fact of the matter in this domain, then

we can encompass it within the deductive means at our disposal. (Using ‘S’ to represent the knowledge of a fact by way of its deductive systematization, we have it that the thesis (Vp)(p —> Sp) is simply untenable.) To be sure, the following objection may still be attempted: A single suitably general truth can encompass infinitely many descriptive facts—-even a transdenumerable infinity of them. If I say of a particular spring that it obeys Hook’s law (over a certain range)—-assigning it the infinitely rich disposition to displace proportionally with imposed weights—I have implicitly provided for a transdenumerable infinity of descriptive consequences by means of the continuous parameter at issue. Accordingly, while it is true that the actual deductions that one can carry out from an axiomatic basis are denumerable, they can certainly manage to “cover’——at a certain level of implicitness— a transdenumerable range of descriptive fact.

The reply here lies in observing that the envisaged process allows for only one very limited sort of infinitism: the positing of a determinate value at a certain point of one and the same infinistic range of determi-

nation—the fixing ofparticular cases within a prespecified spectrum. The objection is thus transcended when one recalls the earlier point that there is in principle no theoretical limit to the lines of consideration available to provide descriptive perspectives upon a thing—thatthe range of descriptive spectra can always, in principle, be extended. But there is yet another way of substantiating our point. Forthe preceding considerations related to the limits of knowledge that can be rationalized on a fixed and given conceptual basis—a full-formed, developed language. But in real life, languages are never full formed, and a conceptual basis is never “fixed and given.” Even with such familiar things as birds, trees, and clouds, we are involved in a constant reconceptualiza-

tion in the course of progress in genetics, evolutionary theory, and thermodynamics. Our conceptions of things always present a moving rather thana fixed object of scrutiny, and this historical dimension must also be reckoned with. Any adequate theory of inquiry must recognize that the ongoing pro-

COGNITIVE LIMITS

247

cess of information acquisition at issue in science is a process of concep-

tual innovation, which always leaves certain facts about things wholly outside the cognitive range of the inquirers of any particular period. Caesar did not know (and in the then-extant state of the cognitive art could not have known) that his sword contained tungsten and carbon. There will always be facts about a thing that we do not know because we cannot even conceive of them in the prevailing conceptual order of things. To grasp such a fact means taking a perspective of consideration that we as yet simply do not have, since the state ofknowledge (or purported knowledge) has not reached a point at which such a consideration is feasible. Any adequate worldview must recognize that the ongoing progress of scientific inquiry is a process of conceptual innovation that always leaves various facts about the things of this world wholly outsidethe cognitive range of the inquirers of any particular period. The language of emergence can perhaps be deployed usefully to make the point. But what is at issue is not an emergence of the features of things but an emergence in our knowledge about them. Blood circulated in the human body well before Harvey; substances containing uranium were radioactive before Becquerel. The emergence at issue relates to our cognitive mechanisms of conceptualization, not to the objects of our consideration in and of themselves. Real-world objects must be conceived of as antecedent to any cognitive interaction—as being there right along, “pregiven” as Edmund Husserl put it. Any cognitive changes or innovations are to be conceptualized as something that occurs on our side of the cognitive transaction, not on the side of the objects with which we deal.®

The prospect of change can never be eliminated in this domain. The properties of a thing are literally open-ended: we can always discover more of them. Even if we were (surely mistakenly) to view the world as inherently finitistic (espousing a Keynesian principle of “limited variety” to the effect that nature can be portrayed descriptively with the materials of a finite taxonomic scheme), there will still be no a priori guarantee that the progress of science will not lead ad indefinitum to changes of mind regarding this finite register of descriptive materials. And this conforms exactly to our expectation in these matters. For where the real things of the world are concerned, we not only expect to learn more about them in the course of scientific inquiry, we expect to have to change our minds about their nature and modes of comportment. Be the items at issue elm 5 One possible misunderstanding must be blocked at this point. To learn about nature, we must interact with it. And so, to determine a feature of an object, we may have to make some impact upon it that would perturb its otherwise obtaining condition. (The indeterminacy principle of quantum mechanics affords a well-known reminder of this.) It should be clear that this matter of physical interaction for data acquisition is not contested in the _ ontological-indifference thesis here at issue.

248

:

REALISM AND IDEALISM

trees or volcanoes or quarks, we have every expectation that in the course of future scientific progress, people will come to think about their origin and their properties differently from the way we do at this juncture.

The Cognitive Opacity of Real Things It is worthwhile to examine more closely the considerations that indicate the inherent limitedness and imperfection of our knowledge of things.® To begin with, it is clear that, as we standardly think about things within the conceptual framework of our fact-oriented thought and discourse, any real physical object has more facets than it will ever actually manifest in experience. For every objective property of a real thing has consequences of a dispositional character, and the dispositions that particular concrete things inevitably have endow them all with an infinitistic aspect that cannot be comprehended within experience.’ This desk, for example, has a limitless manifold of phenomenal features of the type “Having a certain appearance fromaparticular point of view.” It is perfectly clear that most of these will never be actualized in experience. Moreover, a thing is what it does: entity and lawfulness are coordinated correlates—a good Kantian point. And this fact that things demand lawful comportment means that the finitude of experience precludes any prospect of the exhaustive manifestation of the descriptive facets ofany real things.® Physical things not only have more properties than they ever will overtly manifest, but they have more than they can possibly ever actually manifest. This is so because the dispositional properties of things always involve what might be characterized as mutually preemptive conditions of realization. This cube of sugar, for example, has the dispositional property of reacting in a particular way if subjected to a temperature of 10,000° C and of reacting in a certain way if emplaced for one hundred hours in a large, turbulent body ofwater. But if either of these conditions is ever realized, it will destroy the lump of sugar as a lump of sugar and ® For a useful survey of philosophical issues located in this general area, see Vincent Julian Fecher, Error, Deception, and Incomplete Truth (Rome, 1975). 7 To be sure, abstract things such as colors or numbers will not have dispositional properties. Being divisible by four is not a disposition of sixteen. But Plato got the matter right in book 7 of the Republic. In the realm of abstracta, such as those of mathematics,there are no genuine processes—and process is a requisite of genuine dispositions. : 8 This aspect of objectivity was justly stressedin the“Second Analogy” of Kant’s Critique ofPure Reason, though his discussion rests on ideas already contemplated by Leibniz, Philosophische Schriften, 7 vols., ed., C. I. Gerhardt (Berlin, 1875-90), 7:319-22.

COGNITIVE LIMITS

249

thus block the prospect of its ever bringing the other property to manifestation. The perfectly possible realization of various dispositions may fail to be mutually compossible, and so the dispositional properties of a thing cannot ever be manifested completely—not just in practice, but in principle. Our objective assertions about real things always commit us to claiming more than we can actually ever determine about them. The existence of this latent (hidden, occult) sector is a crucial feature of our conception of a real thing. Neither in fact nor in thought can we ever simply put it away. To say of the apple that its only features are those it actually manifests is to run afoul of our conception of an apple. To deny—or even merely to refuse to be committed to the claim-——that it would manifest particular features if certain conditions came about (e.g., that it would have such-and-such a taste if eaten) is to be driven to withdrawing the claim that it is an apple. The process of corroborating the implicit contents of our objective factual claims about something real is potentially endless, and such judgments are thus “nonterminating” in C. I. Lewis’s sense.® This cognitive depth of our objective factual claims— inherent in the fact that their assertoric content will always outrun the evidence we have for making them—means that the endorsement of any such claim always involves some element of evidence-transcending conjecture.

The very concepts at issue with “experience” and “manifestation” are such that we can only ever experience those features of a real thing that it actually manifests. But the preceding considerations show that real things always have more experientially manifestable properties than they can ever actually manifest in experience. The experienced portion of a thing is just the part of the icebergthat shows above water. All real things are necessarily thought of as having hidden depths that extend beyond the limits, not only of experience, but also of experientiability. To say of something that it is an apple or a stone ora tree is to become committed to claims about it that go beyond the data we have—and even beyond those that we can, in the nature of things, ever actually acquire. The meaning inherent in the assertoric commitments of our factual statements is never exhausted by its verification. Real things are cognitively opaque—we cannot see to the bottom of them. Our knowledge of such things can thus become more extensive without thereby becoming more complete. In this regard, however, real things differ in an interesting and important way from their fictional cousins. To make this difference plain, it is useful to distinguish between two types of information about a thing, 5 See C. I. Lewis, An Analysis of Knowledge and Valuation (La Salle, Ill., 1962), pp. 180-81.

250

.

REALISM AND IDEALISM

namely, that which is generic and that which is not. Generic information tells about those features of a thing that it has in common with everything else of its kind or type. For example, a particular snowflake will share with all others certain facts about its structure, its hexagonal form, its chemical composition, its melting point, and so forth. In addition, it will have various properties that it does not share with other members of its own “lowest species” in the classificatory order—its particular shape, for example, or the angular momentum of its descent. These are its nongeneric features. Now a key aboutfictional particulars is that they are of finite cognitive depth. In discoursing about them, we shall ultimately run out of steam as regards their nongeneric features. A point will always be reached when one cannot say anything further that is characteristically new about them—presenting nongeneric information that is not inferentially implicit in what has already been said.1° New generic information can, of

course, always be forthcoming through the progress of science. When we learn more about coal in general, then we know more about the coal in Sherlock Holmes’s grate. But the finiteness of their cognitive depth means that when fictional things are at issue, the presentation of ampliatively novel nongeneric information must by the very nature of the case come to a stop.

With real things, however, there is no reason of principle why the provision of nongenerically idiosyncratic information need ever be terminated. On the contrary, we have every reason to presume these things to be cognitively inexhaustible. A precommitment to description-tran-

scending features—no matter how far description is pushed—is essentialto our conception of a real thing. Something whose character was exhaustible by linguistic characterization would thereby be marked as fictional rather than real. This cognitive opacity-of real things means that we are not—and will never be—in a position to evade or abolish the contrast between “things as we think them to be” and “things as they actually and truly are.” Their susceptibility to further elaborate detail and to changes of mind regarding this further detail is built into our very conception of a real thing. To be a real thing is to be something regarding which we can always, in prin-’ 10 To deny inferentially implicit information the accolade of novelty is not, of course, to say that it cannot surprise us, in view of the limitations of our own deductive powers.

ll This also explains why the dispute over mathematical realism (Platonism) has little bearing on the issue of physical realism. Mathematical entities are akin to fictional entities

in that we can say about them only what we can extract by deductive means from what we have explicitly put into their defining characterization (though one may sometimes be surprised to discover what it is that we have put in). These abstract entities do not have nongeneric properties, since each is a “lowest species” unto itself.

_

COGNITIVE LIMITS

251

ciple, acquire further new information, which may not only supplement but even correct that which has previously been acquired. This view of the situation is supported rather than impeded once we abandon the naive cumulativist/preservationist view of knowledge acquistion for the view that new discoveries need not augment but can displace old ones. With further inquiry, we may come to recognize the error of our earlier ways of thinking about the things at issue. We realize that people will come to think differently about things from the way we do—even when thoroughly familiar things are at issue—recognizing that scientific progress generally entails fundamental changes of mind about how things work in the world. In view of the cognitive opacity of real things, we must never pretend to a cognitive monopoly or cognitive finality. This recognition of incomplete information is inherent in the very nature of our conception of a “real thing.” It is a crucial facet of our epistemic stance toward the real world to recognize that every part and parcel of it has features lying beyond our present cognitive reach——at any “present” whatsoever. Much the same story holds when our concern is not with physical things but with types of such things. To say that something is copper or magnetic is to say more than that it has the properties we think copper or magnetic things have, and to say more than that it meets our test conditions for being copper (or being magnetic). It is to say that this thing is copper or magnetic. And this is an issue regarding which we are prepared at least to contemplate the prospect that we have got it wrong. Certainly, it is imaginable that natural science will come to a stop, not in the trivial sense of a cessation of intelligent life, but in Charles Sanders Peirce’s more interesting sense of eventually reaching a condition after which even indefinitely ongoing effort at inquiry will not—and indeed actually cannot—produce any significant change. Such a position is, in theory, possible. But we can never know—be it in practice or in principle—that it is actually realized. We can never establish that our know!edge of reality, our science, has attained such an omega conditicn of final completion: the possibility of further change lying just around the corner can never be ruled out finally and decisively. Thus, we have no alternative but to presume that our science is still imperfect and incomplete, that no matter how far we have pushed our inquiries in any direction, regions of terra incognita yet lie beyond.

The Corrigibility of Conceptions It must be stressed that these deliberations regarding cognitive inadequacy are less concerned with the correctness of our particular claims

252

:

REALISM AND IDEALISM

about real things than with our characterizing conceptions of them. And in this connection it deserves stressing that there is a significant and substantial difference between a true or correct statement or contention on the one hand, and a true or correct conception on the other. To make a true contention about a thing, we merely need to get one particular fact about it straight. To have a true conception of the thing, however, we must get all of the important facts about it straight. And it is clear that this involves a certain normative element—namely, what the “important” or “essential” facets of something are. Anaximander of Miletus presumably made many correct contentions about the sun in the fifth century B.c.—for example, that its light is brighter than that of the moon. But Anaximander’s conception of the sun (as the flaming spoke of a great wheel of fire encircling the earth) was totally wrong. To assure the correctness of our conception of a thing we would have to be sure (as we very seldom are) that nothing further can possibly come along to upset our view of just what its important features are and just what their character is. Thus, the qualifying conditions for true conceptions are far more demanding than those for true claims. With a correct contention about a thing, all is well if we get the single relevant aspect of it right, but with a correct conception of it we must get the essentials

right—we must have an overall picture that is basically correct. And this is something we generally cannot ascertain, if only because we cannot say

with confidence what is really important or essential before the end of the proverbial day.

,

With conceptions, unlike with propositions or contentions, incom- pleteness means incorrectness, or at any rate presumptive incorrectness.

Having a correct or adequate conception of something as the object it is requires that we have got right all the important facts relevant to the issue. But since the prospect of discovering further importantfacts can never be eliminated, the possibility can never be eliminated that matters may so eventuate that we may ultimately (with the wisdom of hindsight) acknowledge the insufficiency or even inappropriateness of our earlier conceptions. Given the historical realities, conceptions based on incomplete data must be presumed to be at least partially incorrect. If wecari decipher only half an inscription, our conception of its overall content must be largely conjectural—and thus must be presumed to contain an admixture of error. When our information about something is incomplete, obtaining an overall picture of the thing at issue becomes a matter of theorizing, or guesswork, however sophisticatedly executed. And then we have no alternative but to suppose that this overall picture falls short

COGNITIVE LIMITS

253

of being wholly correct in various (unspecifiable) ways. With conceptions, falsity can thus emerge from errors of omission as well as those of commission, resulting from the circumstance that the information at our disposal is merely incomplete, rather than actually false (as would have to be the case with contentions). To be sure, an inadequate or incomplete description of something is not thereby false—the statements we make about it may be perfectly true as far as they go. But an inadequate or incomplete conception of a thing is for this very reason one that we have no choice but to presume to be incorrect as well,!2 seeing that where there is incompleteness, we cannot justifiably take the stance that it relates only to inconsequential matters and touches nothing important. Accordingly, our conceptions of particular things are always to be viewed not just as cognitively open-ended, but as corrigible as well. The conception of an objectivity that underlies ourdiscourse about the things of this world reflects a certain sort of tentativity and fallibilism— the implicit recognition that our own personal or even communal conception of particular things may, in general, be wrong, and that it is in any case inadequate. At the back of our thought about things, there is always a certain wary recognition of the possibility of error. The objectivity of real existents projects them beyond the reaches of subjectively conditioned information that we purport to have. Potentially there is always more to such things than we actually know about them. There is wisdom in Hamlet’s dictum “There are more things in heaven and on earth, Horatio. . . .” The limits of knowledge may be the limits of our world, but they are not the limits of the world. We do and must recognize the limitations of our cognition. We cannot justifiably equate reality with what can, in principle, be known by us, nor equate reality with what can, in principle, be expressed by our language. And what is true here for our sort of mind is true for any other sort of finite mind as well. Any physically realizable sort of cognizing being can know only a part or aspect of the real. We are led back to the thesis of the great idealist philosophers (Spinoza, Hegel, Bradley, Royce) that actual human knowledge inevitably falls short of “perfected science” (the Idea, the Absolute) and must be presumed deficient both in its completeness and its correctness. And our deliberations about the limited knowability of things have a further deeply idealistic aspect. For the fundamental fact of the unending cognitive depth of real things is not actually a discovery we make about 2 Cf. F. H. Bradley's thesis “Error is truth, it is partial truth, that is false only because

partial and left incomplete” (Appearance and Reality [Oxford, 1893], p. 169).

254

:

REALISM AND IDEALISM

in the course of experithem—not something that we learn about things an aspect of our very ential interaction with the real. Instead, it reflects lves

hing we ourse conception of what it is to be “a real thing”; it is somet we shape

whose terms provide—a feature of ideational instruments in re of our conception featu our understanding of the world. It is an integral edge of them.'* knowl our of the real that the nature of real things outruns author's Empirical Inquiry, The Limits of 13 The topic of cognitive limits is treated in the Scientific Realism (Dordrecht, 1987). Science (Berkeley and Los Angeles, 1984), and

Fifteen Metaphysical Realism The Existential Component of Realism Metaphysical realism is the doctrine that the world exists in a way that is substantially independent of the thinking beings that inquire into it, and that its nature—its having whatever characteristics it does actually have—is also comparably thought independent. In saying of something that it is “a real thing,” an object existing as part of the world’s furniture,

we commit ourselves to various (obviously interrelated) points: 1. Self-subsistence: being a “something” (an entity or process) with its own unity of being; having an enduring identity of its own. 2. Physicality, or reality: existing within the causal order of things; having a

’ place as a real item on the world’s physical scene. 3. Publicity, or accessibility: admitting universality of access; being something that different investigators proceeding from different points of departure can get hold of.

4, Autonomy, or independence: being independent of mind; being something that observers find rather than create and learn about rather than define in the course of their cognitive endeavors.

The fundamental idea of realism is that the existence and nature of the world are matters distinct from anyone’s thinking about it: that, minds themselves and their works aside, the world is what it is without any reference to our cognitive endeavors and that the things of nature are quite impervious, as it were, to the state of our knowledge or belief regarding them. One recent expositor formulates this idea as follows: “Even if there were no human thought, even if there were no human beings, whatever there is other than human thought (and what depends on that, causally or logically) would still be just what it actually is.”! Such a realism is based upon a commitment to the notion that human inquiry addresses itself to what is really and truly the condition of things whose 1 William P. Alston, “Yes, Virginia, There Is a Real World,” Proceedings and Addresses of the American Philosophical Association 52 (1979): 779. Cf.: “The world is composed of particulars [individual existing things or processes] which have intrinsic characteristics— i.e., properties they have or relationships they enter into with other particulars independently of how anybody characterizes, conceptualizes, or conceives of them” (Frederick Suppe, “Facts and Empirical Truth,” Canadian Journal of Philosophy 3 [1973]: 200).

256

.

REALISM AND IDEALISM

existence and character are altogether independent of our cognitive activities. Reality is not subject to the operations of the human mind; on the

contrary, the human mind and its dealings are but a minuscule part of reality. The objective facts about things obtain independently, regardless of what we or anybody else think of them. The issue of objectivity in the sense of mind independence is pivotal for realism. A fact is objective in this mode if it obtains independently of whatever thinkers may think about relevant issues, so that changes merely in what is thought by the world’s intelligences would leave it unaffected—that is, if it is thought invariant or thought insensitive. With objective facts (unlike those that are merely a matter of intersubjective agreement), what thinkers think just does not enter into it.

Realism has two indispensable and inseparable constituents—the one existential and ontological, and the other cognitive and epistemic. The former maintains that there indeed is a real world—a realm of mind-in-

dependent, objective physical reality. The latter‘maintains that we can to some extent secure adequate descriptive information about this mindindependent realm. This second contention obviously presupposes the first. But how can that first, ontological thesis be secured?

Metaphysical realism is clearly not an inductive inference secured through the scientific systematization of our observations, but rather represents a regulative presupposition that makes science possible in the first place. If we did not assume from the very outset that our sensations somehow relate to an extramental reality, we could clearly make no use of them to draw any inference whatever about “the real world.” The realm of mind-independent reality is something we cannot discover—we do not learn that it exists as a result of inquiry and investigation. How could we ever learn from our observations that our mental experience is itself largely the causal product of the machinations of a mind-independent matrix, that all those phenomenal appearances are causally rooted in a physical reality? This is obviously something we have to suppose from the very outset. What is at issue is, all too clearly, a precondition for empirical inquiry—a presupposition for the usability of observational data as sources of objective information. That experience is indeed objective, that what we take to be evidence is evidence, that our sensations yield information about an order of existence outside the experiential realm itself, and that this experience constitutes not just a mere phenomenon but an appearance of something extramental belonging to an objectively self-subsisting order—all this is something that we must always presuppose in using experiential data as evidence for how things stand in __. the world. : Accordingly, we do not learn or discover that there is a mind-independent physical reality; we have no alternative but to presume or postulate

METAPHYSICAL REALISM

257

it. Objectivity represents a postulation made on functional (rather than evidential) grounds: we endorse it in order to be in a position to learn by experience at all. As Kant clearly saw, objective experience is possible only if the existence of such a real, objective world is presupposed from the outset rather than being seen as a matter of ex post facto discovery about the nature of things. ; It thus emerges that the crucial existential (ontological) component of realism is not a matter of discovery, a part of the findings of empirical research. It is a presupposition for our inquiries rather than a result thereof. We have to do here not with an evidential discovery about the constitution of nature as such but rather with aformative assumption that undergirds our view of the nature of inquiry. Without subscribing to this idea, we just could not think of our inquiries as we actually do—as a cognitive (i.e., knowledge-generating) process. Our commitment to the existence of a mind-independent reality is thus a postulate whose justification pivots—in the first instance—on its functional utility in enabling us to operate as we do with respect to inquiry and deliberation. To be sure, once we have made astart by postulating an objective reality and its concomitant causal aspect, then principles of inductive systematization, of explanatory economy, and of common-cause consilience can work wonders in exploiting the phenomena of experience to provide the basis for plausible claims about the nature of the real. But we indispensably need that initial existential presupposition to makea start. Without a commitment from the very outset to a reality to serve as ground and object of our experience, its cognitive import will be lost. Only on this basis can we proceed evidentially with the exploration of the interpersonally public and objective domain of a physical world-order

that we share in common. Accordingly, that second, descriptive (epistemic) component of realism stands on a very different footing. Reality’s nature is something about which we can make warranted claims only through examining it. Substantive information must come through inquiry—through evidential validation. Once we are willing to credit our observational data with objectivity and thus with evidential bearing, then we can, of course, make use of

them to inform ourselves as to the nature of the real. But that initial presumption has to be there from the start. Let us consider this basic reality-postulate somewhat more closely. Our standard conception of inquiry involves recognition of the following facts: (1) the world (the realm of physical existence) has a nature whose characterization in point of description, explanation, and prediction is the object of empirical inquiry; (2) the real nature of the world is in the main independent of the process of inquiry that the real world canalizes or conditions; and (3) in virtue of these considerations, we can stake neither

258

.

REALISM AND IDEALISM

total nor final claims for our purported knowledge of reality. Our knowledge of the world must be presumed incomplete, incorrect, and imperfect, with the consequence that “our reality” must be considered to afford an inadequate characterization of “reality itself.” The crucial question is this: Assuming that there are objective facts, how can we possibly come to acquire knowledge of them? What sort of presuppositions must we make if our subjective experience—which is limited and episodic—is to provide a basis of legitimacy for maintaining objective and general claims? Two gulfs must be transcended: 1, From experiential appearances to objective facts (from “That looks like a red apple to me” to “That is a red apple”)—from phenomena to real things. 2. From particular cases to universals (from “These apples all contain seeds” to “All apples contain seeds”).

To effect these transitions, we must simply presuppose (for how could we possibly preestablish this?) that such ampliative moves can be made on the basis of available evidence: that subjective phenomena are (typically) indicators of objective realities and that particular cases are (typically) exemplifications of universal arrangements. The foundations of objectivity thus do not rest on the findings-of science. They precede and underlie science, which would itself not be possible without a precommitment to the capacity of our senses to warrant claims about an objective world order. Objectivity is not a product of inquiry; we must precommit ourselves to it to make inquiry possible. It is a necessary (a priori) input into the cognitive project and not a contingent (a posteriori) output thereof. The objective bearing of experience is not something we can preestablish; it is something we must presuppose in the interest of honoring Peirce’s pivotal injunction never to bar the path of inquiry. Our commitment to realism centers on a certain practical modus operandi encapsulated in the precept “Proceed in matters of inquiry and communication on the basis that you are dealing with an objective realm, existing quite independently of the doings and dealings of minds.” Accordingly, we standardly operate on the basis of the “presumption of objectivity” reflected in the guiding precept “Unless you have good reason to think otherwise (that is, as long as nihil obstat), treat the materials of

inquiry and communication as veridical—as representing the nature of the real.” The ideal of objective reality is the focus of a family of convenient regulative principles—a functionally useful instrumentality that en- _ ables us to transact our cognitive business in the most satisfactory and — effective way. Our endorsement of unobserved causes in nature is thus based not on

METAPHYSICAL REALISM

259

science but on metaphysics. What we learn from science is not that an unobservable order of physical existence causally undergirds nature as we observe it but rather what these underlying structures are like. Science does not (cannot) teach us that the observable order is explicable in terms of underlying causes and that the phenomena of observation are signs or symptoms of this extra- and subphenomenal order of existence; we know this a priori of any world in which observation as we understand it can transpire. (After all, observations are, by their very nature, results of interactions.) What science does teach us (and metaphysics cannot) is what the descriptive character of this extraphenomenal order can reasonably be supposed to be in the light of our experience of it.

Realism in Its Regulative/Pragmatic Aspect Our ventures into communication and inquiry are undergirded by the stance that we communally inhabit a shared world of objectively existing things, a world of “real things” among which we live and into which we inquire, yet about which we do and must assume that we have only imperfect information at any and every particular stage of the cognitive venture. This commitment to an objective reality that underlies the data at hand is indispensably demanded by any step into the domain of the publicly accessible objects essential to communal inquiry and interpersonal communication about a shared world. We could not establish communicative contact about a common objective item of discussion if our discourse were geared to the substance of our own idiosyncratic ideas and

conceptions of things. — Objectivityis not something w we infer from data; it is something we do and must presuppose. Itis something that we postulate or presume from the very outset of our dealings with people’s claims about the world’s facts—our own included. Its epistemic status is not that of an empirical discovery but that of a presupposition. whose ultimate justification is a transcendental argument from the very possibility of the projects of communication and inquiry as we standardly conduct them. Seen in this light, the key point may be put as follows: it is indeed a presupposition of effective communicative discourse about a thing that we purport (claim and intend) to make true statements about it. But for such discourse it is not required that we purport to have a true or even adequate conception of the thing at issue. On the contrary, we must deliberately abstain from any claim that our own conception is definitive if we are to engage successfully in discourse. We deliberately put the whole matter of conceptions aside—abstracting from the question of the agreement of my conception with yours, and all the more from the issue of

260

REALISM AND IDEALISM

which one of us has the right conception. This sort of epistemic humility is the price we pay for keeping the channels of communication open. Objectivity and the realism that goes with it ultimately root in the teleology of inquiry and communication. In the final analysis, they represent commitments we make because doing so serves our purposes. And this circumstance has far-reaching ramifications. Reality (on the traditional metaphysicians’ construction of the concept) is the condition of things answering to “the real truth”; itis the realm of what really is as it really is. But what legitimates this postulation of metaphysical realism that experience affords data regarding an objective and mind-independent domain and thus provides for viable information about the real? Given that the existence of such a domain is not a product of but a precondition for empirical inquiry, its acceptance has to be validated in the manner appropriate for postulates and prejudgments of any sort—namely, in terms of its ultimate utility. Bearing this pragmatic perspective in mind, let us take a closer look at this issue of utility and ask,

“What can this postulation of a mind-independent reality actually do for us?” / The answer is straightforward. The assumption of a mind-independent reality is essential to the whole of our standard conceptual scheme relating to inquiry and communications. Without it, both the actual conduct and the rational legitimation of our communicative and investigative (evidential) practice would be destroyed. Nothing that we do in this cognitive domain would make sense if we did not subscribe to the conception of a mind-independent reality. To begin with, we indispensably require the notion of reality to operate the classical concept of truth as “agreement with reality” (adaequatio ad rem). Once we abandon the concept of reality, the idea that in accepting a factual claim as true, we become committed to how matters actually stand (“how it really is’) would also go by the board. The very semantics of our discourse constrains its commitment to realism; we have

no alternative but to regard as real those states of affairs claimed by the contentions we are prepared to accept. Once we put a contention forward by way of serious assertion, we must view as real the states of affairs it purports and must see its claims as facts. We need the notion of reality to operate the conception of truth. A factual statement on the order of “There are pi-mesons’ is true if and only if the world is such that pimesons exist within it. By virtue of their very nature as truths, true state-

ments must state facts: they state what really is so, which is-exactly what it is to characterize reality. The conceptions of truth and of reality come | together in the notion of adaequatio ad rem—the venerable principle that to speak truly is to say how matters stand in reality, in that things

actually are as we have said them to be. That is just how these concepts

METAPHYSICAL REALISM

261

of truth and adequacy work, and they would not work out without that fundamental precommitment to the nature of the real. For in abandoning the commitment to reality as such and “an sich,” we would thereby also lose our grip on those regulative contrasts that canalize and condition our view of the nature of inquiry and indeed shape our conception of this process as it stands within the framework of our conceptual scheme, as an instrumentality of information transmittal about an objective order of existence. The very conception of inquiry as we conceive of it would have to be abandoned if the contrast conceptions of “actual reality” and “the real truth” were no longer available. In the second place, the nihilistic denial that there is such a thing as

reality would destroy once and for all the crucial Parmenidean divide between appearance and reality. And this would exact a fearful price from us: we would be reduced to talking only of what we (I, you, many of us) think to be so. The crucial contrast notion of the real truth would no longer be available: we would be able to contrast only our putative truths with those of others but could no longer operate the classical distinction between the putative and the actual, between what people merely think to be so and what actually is so. We could not take the stance that, as the Aristotelian commentator Themistius put it, “that

which exists does not conform to various opinions, but rather the correct opinions conform to that which exists.”? The third point is the issue of cognitive coordination. Communication and inquiry, as we actually carry them on, are predicated on the fundamental idea of a real world of objective things, existing and functioning “in themselves,” without specific dependence on us and so equally accessible to others. Intersubjectively valid communication can be based only on common access to an objective order of things. The whole communicative project is predicated on a commitment to the idea that there is a realm of shared objects about which we as a community share questions and beliefs, and about which we ourselves as individuals presumably have only imperfect information that can be criticized and augmented by the efforts of others. All our ventures at communication and communal inquiry are predicated on the stance that we communally inhabit a shared world of things available for contemplation independently of any single person’s ideas about them. Where we claim exclusive rights to conceptualization (where I conceive of my nightmare and you conceive of yours), the question of reidentification (of whether your dream matter is the same one as mine) cannot arise, and so the prospect of a sharing of experience about: them as such (i.e., as monsters rather than people’s ideas about monsters) becomes impossible. It is, to be sure, perfectly 2 Maimonides, The Guide for the Perplexed, bk. 1, chap. 71, sec. 96a.

262

.

REALISM AND IDEALISM

possible for two people to communicate effectively about something that is wholly nonexistent and about which they have substantially discordant conceptions (for example, X’s putative wife, where X is, in fact, unmar-

ried, though one party is under the misimpression that X is married to A, and the other under the misimpression that X is married to B). Given

a shared commitment to an objective framework of things, such aberrations are possible. But this possibility itself attests to our standard commitment to a commonality of communicative focus as the basis on which alone the exchange of information (or misinformation) and the discovery of error becomes possible. The shared (conventionalized) intention to talk about the same thing undergirds and sustains all our communicative ventures.

:

This points to a fourth important consideration. Only through reference to the real world as a common object and shared focus of our diverse and imperfect epistemic strivings are we able to effect communicative contact with one another. Our standard view of inquiry takes the line that the real nature of the world is in the main independent of the process of inquiry, one, however, that the real world canalizes or conditions. De-

pendency is a one-way street here: reality shapes or influences inquiry, but not conversely. Our knowledge of the world must be presumed incomplete, incorrect, and imperfect, with the consequence that “our reality” must be considered to afford an inadequate characterization of “reality itself.” With respect to our cognitive endeavors, “man proposes and nature disposes,” and it does so in both senses of the term: it disposes over our current

view of reality, and it will doubtless eventually dispose of it as well. Our view of the nature of inquiry and of the sort of process it represents is possible only because we stand committed from the very outset to the idea of ourselves as a minuscule component of a mind-independent real-

ity. We can act and affect a few things in it. But in the main it has the whip hand, and we merely respond to its causal dictates. And this is true in cognitive aspects as well—-where we must simply do the best we can with the relatively feeble means at our disposal. Lo The fifth consideration is that the very idea of inquiry as we conceive .

it would have to be abandoned if the conceptions of “actual reality” and . “the real truth” were no longer available to serve their crucial contrasting roles. We could no longer assert, “What we have there is good enough as far as it goes, but it is presumably not ‘the whole real truth’ of the matter.” Without the conception of reality we could not think of our

knowledge in the fallibilistic mode we actually use—as having provi-. sional, tentative, improvable features that constitute a crucial part of the conceptual scheme within whose orbit we operate our concept of inquiry. For our commitment to the mind-independent reality of “the real world”

METAPHYSICAL REALISM

.

263

stands together with our acknowledgement that, in principle, any or all of our present scientific ideas as to how things work in the world, at any present, may well prove to be untenable. Our conviction in a reality that lies beyond our imperfect understanding of it (in all the various senses of “lying beyond”) roots in our sense of the imperfections of our scientific world-picture—its tentativity and potential fallibility. In abandoning our commitment to a mind-independent reality, we would lose the impetus of inquiry. Inquiry, as we standardly conceive it, is predicated on the commitment to an inquiry-independent reality; it is a quest for information about “the real world” with respect to which our own conceptions of things are nowise definitive, and into which others can accordingly enter unproblematically. It is thus geared to the conception of an objective world: a communally shared realm of things that exist strictly “on their own,” composing an enduring and ‘impersonal realm within which—-and, more important, with reference to which—inquiry proceeds. We could not.operate the notion of inquiry as aimed at estimating the character of the real if we were not prepared to presume or postulate a reality for these estimates to be estimates of. It would clearly be pointless to devise our characterizations of reality if we did not stand committed to the proposition that there indeed is a reality to be characterized. The pivotal contrast in this regard is between “mere appearance” and “reality as such,” between “our picture of reality” and “reality itself,” between what actually is and what we merely think (believe, suppose, etc.) to be. And our allegiance to the conception of reality and to this contrast that pivots upon it root in the fallibilistic recognition that at the level of the detailed specifics of scientific theory, anything we presently hold to be the case may well turn out otherwise—indeed, certainly will do so if past experience gives any auguries for the future. Sixth and finally, we need the conception of reality in order to operate the causal model of inquiry about the real world. Our standard picture of humankind’s place in the scheme of things is predicated on the fundamental idea that there is a real world (however imperfectly our inquiry may characterize it) whose causal operations produce inter alia causal impacts upon us, providing the basis of our world picture. Reality is viewed as the causal source and basis of the appearances, the originator and determiner of the phenomena of our cognitively relevant experience. The real world is seen as causally operative both in serving as the external molder of thought and as constituting the ultimate arbiter of the adequacy of our theorizing. In summary, then, we need the postulate of an objective order of mind-independent reality for at least six important reasons.

264

:

REALISM AND IDEALISM

1. To preserve the distinction between true and false with respect to factual

matters and to operate the idea of truth as agreement with reality. 2. To preserve the distinction between appearance and reality, between our picture of reality and reality itself. 3. To serve asa basis for intersubjective communication. 4. To furnish the basis for a shared project of communal inquiry. 5. To provide for the fallibilistic view of human knowledge. 6. To sustain the causal mode of learning and inquiry and to serve as basis for the objectivity of experience.

The conception of a mind-independent reality accordingly plays a central and indispensable role in our thinking with respect to matters-of language and cognition. In communication and inquiry alike we seek to offer answers to our questions about how matters stand in this objectiverealm. It is seen as the epistemological object of veridical cognition, in the context of the contrast between “the real” and its “merely phenomenal” appearances. Again, it is seen as the target or telos of the truth-estimation process at issue in inquiry, providing for a common focus in communication and communal inquiry. (The real world thus constitutes the “ob-

ject” of our cognitive endeavors in both senses of this term—the objective at which they are directed and the purpose for which they are exerted.) And furthermore, reality is seen as the ontological source of cognitive endeavors, affording the existential matrix in which we move and have our being, and whose impact upon us is the prime mover for our cognitive efforts. All of these facets of the concept of reality are integrated and unified in the classical doctrine of truth as it corresponds to fact (adaequatio ad rem), a doctrine that makes sense only in the settingof a commitment to mind-independent reality. Towards transcendental reality itself, it is appropriate to take much the same stance that Kant took toward his “thing in itself.” As such, there is

nothing to be said about-it; its character is fundamentally a matter ofje ne sais quoi, because we recognize that definitive and error-immune ©

claims to knowledge cannot be substantiated at the level of scientific theorizing. We stand committed, however, to the idea that there must in-

deed be such a thing as knowledge-transcending reality, because only in this way can we operate our conceptual scheme with respect to inquiry and communication, with its inhérent dedication to the conception of an

objectivity that recourse to “the real world” ¢an alone sustain. Negative and regulative though the conception may be, we nevertheless require it as a tool of indispensable utility, for it alone enables us to operate our

standard concept-scheme in matters relating to inquiry. Accordingly, the justification for this fundamental presupposition of objectivity is not evidential at all; postulates are not based on evidence.

METAPHYSICAL REALISM

265

Rather, it is functional. We need this postulate to operate our conceptual scheme. The justification of this postulate lies in its utility. We could not form our existing conceptions of truth, fact, inquiry, and communication without presupposing the independent reality of an external world. We simply could not think of experience and inquiry as we do. (What we have here is a “transcendental argument” of sorts from the character of our conceptual scheme to the acceptability of its inherent presuppositions.) Consider the following objection: Let it be granted that this general approach is right—that the idea of a mindindependent reality is a presupposition basic to the conceptual framework that undergirds our project of inquiry and knowledge acquisition and communication. But why should one see this assumption as validated by its serviceability in this regard? After all, perhaps the entire project is simply unjustified. Consider the analogy of religion. God is essential to the project of religion and worship; the external world is essential to the project of inquiry and cognition.

But perhaps those entire projects are simply inappropriate.

In countering this considered objection with respect to cognition, we must stress the inappropriateness of the analogy. For the religious project is optional; one may simply decline to enter in. But the cognitive project is not so easily evaded. We must act to live: must eat this or that, move here or there, do something or other. And being the sort of creatures we are, our actions are guided by our beliefs. Is this substance edible? Is that place safe? Is that action goal-conducive?If we do not form views on these subjects and allow our actions to be guided by knowledge (or pretentions thereto), then there are but few alternatives (all duly noted and recommended by the skeptics of classical antiquity): to follow custom and “do what is generally done”; to follow instinct, or “hunch”; to follow our desires and inclinations.

But none of these noncognitive alternatives seem very promising; none have much appeal to a creature who demands good reasons for acting. (Even to begin to validate a reliance on probabilities, we need facts.) The impetus to inquiry for knowledge acquisition reflects the most practical of imperatives. Our need for intellectual accommodation in this world is no less pressing and no less practical than our need for physical accommodation. But in both cases, we do not want just some house or other, but one that is well built, that will not be blown down by the first

wind that sweeps along. Skeptics from antiquity onward have always said, “Forget about those abstruse theoretical issues; focus on your practical needs.” They overlook the crucial fact that an intellectual accommodation

266

.

REALISM AND IDEALISM

to the world is itself one of our deepest practical needs—that in a position of ignorance or cognitive dissonance we cannot function satisfactorily. The project of communal inquiry is not optional—at any rate, not for us humans. Its rationale lies in the most practical and prudent of considerations, since it is only by traveling the path of inquiry that we can arrive at the sorts of good reasons capable of meeting the demands of a rational animal. And given the mandatory nature of the cognitive project, we have no real choice but to “buy in” to its presuppositions. We thus arrive at an overall course of justificatory argumentation whose structure runs as follows: 1. We cannot survive and flourish in this world without effective action. 2. We cannot act effectively without rationally warranted confidence in our (putative) knowledge.

3. We cannot achieve confidence-inspiring knowledge without rational inquiry. . 4. Commitment toa real world is an essential requisite for rational inquiry. 5. Therefore realism (i.e., commitment to a real world that is the object of our inquiries) is a rational imperative on the side of practical reason—a sine qua non fora rational creature like ourselves to survive and to flourish.

Only in subscribing to such a fundamental postulate of reality can we take the sort of view of experience, inquiry, and communication that we in fact have. Without it, the entire conceptual framework of our thinking about the world and our place within it would come crashing down. The utility of the conception of reality is such that even if reality were not there, we would have to invent it.

Retrojustification: The Wisdom of Hindsight Can the functional utility. of our concept of reality by itself provide an adequate validation for it? A validation in terms of functional utility establishes our claims to mind-independent reality not by the cognitive route of learning but by the pragmatic route of an eminently useful postulation. Crucial though this may be, it clearly cannot be theentirestory. The consideration that we must proceed in the way of objectivity-presuming cognition as a matter of the functional requisites of our situation,

seeing that there is just no alternative if our aims are to be attained and our needs and purposes served, stops well short of being totally satisfactory. It does not offer us any assurance that we actually will succeed: in our endeavor if we do proceed in this way; it just has it that we will not if we do not. The issue of actual effectiveness remains untouched. But a nagging doubt still remains. It roots in the following challenge:

METAPHYSICAL REALISM

267

Let us grant that this line of approach provides a cogent practical argument.

All this shows is that realism is useful. But does that make it true? Is there any rational warrant for it over and above the mere fact of its utility?

At this point we have to move beyond presupposed functional requisites to address the issue of actual effectiveness. When it comes to this issue of actual efficacy, we have no choice but to proceed experientially, through the simple strategem of trying and seeing. We must now have recourse to the resources of actual experience. For what is learned by experience—and can only be learned in this way—is that in proceeding on this prejudgment our attempts do, by and large, work out pretty well vis-a-vis the purposes we have in view for inquiry and. communication. We want and need objective information about the real world. This, of course, is not to be had directly without the epistemic mediation of ex-

perience. And so we treat certain data as evidence—we extend “evidential credit” to them, as it were. Through trial and error we learn that some of them do indeed deserve it, and then we proceed to extend to them greater weight—we “increase their credit limit,” as it were, and rely on them more extensively. And of course to use those data as evidence is to build up a picture of the world, a picture that shows, with the wisdom of hindsight, how appropriate it was for us to use those evidential data in the first place. Charles Sanders Peirce put the issue with characteristic clarity: “It may be asked how I know that there are reals. If this hypothesis is the sole support of my method of inquiry, my method of inquiry must not be used to support my hypothesis.”? Peirce placed his finger on exactly the right question. Yet while this reality hypothesis is indeed not a product of inquiry but a presupposition for it, nevertheless it is one whose justification ultimately stands or falls on the success of the inquiries it facilitates. Its validation cannot be preestablished through evidence but can be provided only ex post facto through the justificatory impetus of successful implementation. What we began with was a basic project-facilitating postulation. But this does not tell the whole of the justificatory story. For there is also the no-less-important fact that this postulation obtains a vindicating retrojustification because the further we proceed on this basis, the more its obvious appropriateness comes to light. With the wisdom of hindsight we come to see with increasing clarity that the project that these presuppositions render possible is an eminently successful one. The pragmatic turn does crucially important work here in putting at our disposal a styie 3. §. Peirce, Collected Papers, 8 vols., ed. C. Hartshorne et al. (Cambridge, Mass., _ 1931-50), vol. 5, sec. 5.383.

268

.

REALISM AND IDEALISM

of justificatory argumentation that manages to be cyclical without vitiating circularity. Accordingly, the substantive picture of nature’s ways that is secured through our empirical inquiries is itself ultimately justified, retrospectively as it were, through affording us with the presuppositions on whose basis inquiry proceeds. As we develop science, there must come a “closing of the circle.” The world picture that science delivers into our hands must eventually become such as to explain how it is that creatures such as ourselves, emplaced in the world as we are, investigating it by the processes we actually use, should do fairly well at developing a workable view of that world. The validation of scientific method must in the end itself become scientifically validated. Science must (and can) retrovalidate itself by providing the material (in terms of a science-based worldview) for justifying the methods of science. (How this is to be done has already been sketched out in the story about feedback cycles toldin part 3 of this book.) The rational structure of the overall process ofjustification looks as follows: 1. We use various sorts of experiential data as evidence for objective fact. 2. We do this in the first instance for practical reasons, faute de mieux, be-

cause only by proceeding in this way. can we hope to resolve our questions with any degree of rational satisfaction. 3, As we proceed, two things happen: on the pragmatic side we find that we obtain a world picture on whose basis we can operate effectively (pragmatic revalidation); on the cognitive side we find that we arrive at a picture of the world that provides an explanation of how it is that we are encouraged to get things (roughly) right-—that we are in fact justified in using our phenomenal data as data of objective fact (explanatory revalidation). Accordingly, the success at issue is twofold—both in terms of under-

standing (cognition) and in terms of application (praxis). And it is this ultimate success that justifies and rationalizes,retrospectively, our 0 evidential proceedings. We arrive at the overall situation of dual retrojustification given.in figure 15.1, which shows that all the presuppositions ofinquiry are ultimately justified because of a wisdom of hindsight that enables us to see that by their means we have been able to achieve both practical success and a theoretical understanding of our place in the world’s scheme of things. This includes how our inquiry methods manage to succeed. The cycles must—and presumably do—close in smooth loops of systemic jus-: tification. And both loops are crucial. Successful practical implementation is needed as an extratheoretical quality-control monitor of our theorizing. And the capacity of our scientifically devised view of the world to

METAPHYSICAL REALISM

269

Figure 15.1. The Retrojustification of the Presuppositions of Inquiry Practical Loop of Retrojustification Successful practical implementation in action

v Presuppositions

of inquiry

\ pel

“|

Theconductof

inquiry

Ls.) The products of

inquiry

k

y Successful cognitive implementation in the explanation of how

Theoretical Loop of Retrojustification

;

inquiry works

underwrite an explanation of how it is that a creature constituted as we are, operating by the means of inquiry that we employ, and operating within an environment such as ours, can ultimately devise a relatively accurate view of the world is also critical for the validation of our know!edge.* The closing of these inquiry-geared loops validates, retrospectively, those realistic presuppositions or postulations that made the whole process of inquiry possible in the first place. Realism thus emerges as a presupposition-affording postulate for inquiry—a postulation whose ultimate legitimation eventuates retrospectively through the results, both practical and cognitive, that the process of inquiry based on those yet-tobe-justified presuppositions is able to achieve. Let us review the overall line of deliberation. Metaphysical realism— the doctrine that there is a mind-independent reality and that our experience provides us with a firm cognitive grip upon it—does not represent a learned fact but a presuppositional postulate. As such, it has a complex justification that comes in two phases. The first, initial phase is prospective, proceeding with a view to the functional necessity of taking this position, with its purpose-dictated inevitability. For this step alone renders possible a whole range of activities relating to inquiry and to communication that is of the highest utility for us—and indeed is a practical necessity. In possibilizing (i.e., bringing within the range of the feasible) a host of purpose-mandated activities, the postulate of metaphysical realism obtains its initial justification in the practical order of reasoning. Such initial functional justification is good, but not good enough. The second phase of justification goes further, albeit retrospectively. It proceeds by noting that after we actually engage in the goal-directed practice that the postulate in question possibilizes, our applicative and explana-tory efforts are, in fact, attended by success—that making the initial pos4 These issues are examined at some length in the author's Methodological Pragmatism - (Oxford, 1977).

270

.

REALISM AND IDEALISM

tulate has an immense pragmatic payoff. This issue of actual efficacy is ultimately crucial for the overall justification of the practical postulate at issue.

A notion that has such important work to do cannot be dismissed as vacuous or superfluous. As was observed above, the utility of the conception of reality is so great that if it were not already there, we would have to invent it. But the pragmatic success that ensues when we put this conception to work goes to show that we have not in fact done so in a frivolous or irresponsible way.®

The Idealistic Aspect of Metaphysical Realism Realism is a position to which we are constrained not by the push of evidence but by the pull of purpose. Initially, at any rate, a commitment to realism is an input into our investigation of nature rather than an output thereof. At bottom, it does not represent a discovered fact, but a methodological presupposition of our praxis of inquiry; its status is not constitutive (fact-descriptive) but regulative (praxis-facilitating). Realism is not a factual discovery but a practical postulate justified by its utility or serviceability in the context of our aims and purposes, seeing that if we did not take our experience to serve as an indication of facts about an objective order, we would not be able to validate any objective claims whatsoever. (To be sure, what we can—and do—ultimately discover is

that by taking this realistic stance, we are able to develop a praxis of inquiry and communication that proves effective in the conduct. of our affairs.) The ontological thesis that there is a mind-independent physical reality to which our inquiries address themselves more or less adequately—-and always imperfectly—is the key contention of realism and is supported by the deliberations set out above. But on the telling of the presenting anal- ysis, this basic thesis has the epistemic status of a presuppositional postulate that is initially validated by its pragmatic utility and ultimately retrovalidated by the satisfactory results of its implementation (in both practical and theoretical respects). Our commitment to realism is, on this account, initially not a product of our inquiries about the world but rather

reflects a facet of how we conceive the world. The sort of realism contemplated here is accordingly one that pivots on the fact that we think of reals in a certain sort of way, and that in fact the very conception of the real is something we employ because doing so merits our ends and purposes. 5 This discussion draws some relevant materials from the author’s Empirical Inquiry (Totowa, N.J., 1982).

METAPHYSICAL REALISM

271

Now insofar as realism ultimately rests on this pragmatic basis, it is not based on considerations of independent substantiating evidence about how things actually stand in the world, but rather on considering, as a matter of practical reasoning, how we do (and must) think about the world within the context of the projects to which we stand committed. In this way, the commitment to a mind-independent reality plays an essentially utilitarian role as providing a functional requisite for our intellectual resources (specifically for our conceptual scheme in relation to communication and inquiry). Realism thus harks back to the salient contention of classical idealism that values and purposes play a pivotal role in our understanding of the nature of things. And we return also to the characteristic theme of idealism—the active role of the knower not only in the constituting but also in the constitution of what is known. To be sure, this sort of idealism is not substantive but methodological. It is not a rejection of real objects that exist independently of mind and as such are causally responsible for our objective experience; quite the reverse, it is designed to facilitate their acceptance. But it insists that the justificatory rationale for this acceptance lies in a framework of mind-supplied purpose. For our commitment to a mind-independent reality is seen to arise not from experience but for it—for the sake of putting us into a position to exploit our experience as a basis for validating inquiry and communication with respect to the objectively real. “Reality as such” is no doubt independent of our beliefs and desires, but what can alone concern us is reality as we view it. But the only view of reality that is available to us is one that is devised by us under the aegis of principles of acceptability that we subscribe to because doing so serves our purposes. A position of this sort is indeed in business as a realism. But seeing ‘that it pivots on the character of our concepts and their modus operandi, it transpires that the business premises it occupies are actually mortgaged to idealism. The fact that objectivity is the fruit of communicative purpose allows idealism to infiltrate into the realist’s domain. This approach endorses an object-level realism that rests on a presuppositional idealism at the justificatory infralevel. We arrive, paradoxical as it may seem, at a realism that is founded, initially at least, on a fun-

damentally idealistic basis—a realism whose ultimate justificatory basis is ideal. And the idealism at issue cuts deeper yet. No doubt, we are firmly and irrevocably committed to the idea that there is a physical realm out there that all scientific inquirers inhabit and examine alike. We hold toa single, - uniform physical reality, insisting that all investigations exist within and investigate it: this one single shared realism, this one single manifold of physical objects and laws. But this very idea of a single, uniform, domain

272

REALISM AND IDEALISM

of physical object and laws represents just exactly that—an idea of ours. And the idea is itself a matter of how we find it convenient and efficient to think about things: it is no more—though also no less—than the projection of a theory devised to sort the needs and conveniences of our intellectual situation. We thus reach an important conjuncture of ideas. The ontological independence of things—their objectivity and autonomy of the machinations of mind—is a crucial aspect of realism. And the fact that it lies at the very core of our conception of a real thing that such items project beyond the cognitive reach of mind betokens a conceptual scheme fundamentally committed to objectivity. The only plausible sort of ontology is one that contemplates a realm of reality that outruns the range of knowledge (and indeed even of language), adopting the stance that character goes beyond the limits of characterization. It is a salient aspect of the mind-independent status of the objectively real that the features of something real always transcend what we know about it. Indeed, yet further or different facts concerning a real thing can always come to light, and all that we do say about it does not exhaust all that can and should be said about it. In this light, objectivity is crucial to realism; and the cognitive inexhaustibility of things is a certain token of their objectivity. Authentic realism can exist only in a state of tension. The only reality worth having is one that is in some. degree knowable. But it is the very limitation of our knowledge—our recognition that there is more to reality than what we do and can know or ever conjecture about it—that speaks for the mind independence of the real. It is important to stress against the skeptic that the human mind is sufficiently well attuned to reality that some knowledge of it is possible. But it is no less important to join with realists in stressing the independent character of reality, acknowledging that reality has a depth and complexity of makeup that outruns the reach of mind.

Realism and Incapacity Charles Sanders Peirce and others have located the impetus to realism in the limitations of human will—in the fact that we can exert no control over our experience and, try as we will, cannot affect what we see and

sense. Peirce’s celebrated Harvard Experiment of the Lowell Lectures of 1903 makes the point forcibly: I know that this stone will fall if it is let go, because experience has convinced me that objects of this kind always do fall; and if anyone has any doubt on the subject, I should be happy to try the experiment, and I will bet him a hundred to one on the result. . . . [I know this because of an unshakable conviction that]

METAPHYSICAL REALISM

273

the uniformity with which stones have fallen has been due to some active general principle [of nature]. . . . Of course, every sane man will adopt the latter hypothesis. If he could doubt it in the case of the stone—-which he can’t—and

I may as well drop the stone once and for all—I told you so!—if anybody doubt this still, a thousand other such inductive predictions are getting verified every day, and he will have to suppose every one of them to be merely fortuitous in order reasonably to escape the conclusion that general principles are really operative in nature. That is the doctrine of scholastic realism.®

No doubt, the ordinary person, as well as most philosophers, stands committed to the conviction that whatever happens in the world of observation, happens in line with the causally lawful machinations of an underlying mind-independent physical reality. , In this context, however, it is important to distinguish between mental

dependency and mental control. Peirce is clearly right in saying that we cannot control our conviction that the stone will fall: do what we will, it

will remain. Nevertheless, this circumstance could conceivablystill be something that depends on us—exactly as with the fearsomeness of heights for the person with vertigo. If the unconscious sphere of mind actually dictates how I must see something (as, for example, in an optical illusion of the Mueller-Leyeér variety), then I evidently have no control. But that does not in itself refute mind dependency—even of a very strong sort. There is always the prospect that we are deluding ourselves in these matters—that the limitations at issue appertain only to our conscious powers, not to our powers as such.

This prospect blocks Peirce’s argument in the way already foreseen by Descartes in the Meditations. “I found by experience that these [sensory] ideas presented themselves to me without my consent being requisite, so.that I could not perceive any object, however desirous I might be, unless it were present... . But although the ideas which I receive by the senses do not depend on my will, I did not think that one should for that reason conclude that they proceed from things different from myself, since possibly some faculty might be discovered in me—though different from those yet known to me—which produced them.”? We may simply delude ourselves about the range of the mind’s powers: lack of control 6 Peirce, Collected: Papers, vol. 5, sec. 5.64-67. Cf. vol. 2, sec. 2.138 and also: “Whenever we come to knowafact, it is by its resisting us. A man may walk down Wall Street

debating with himself the existence of an external world; but if in his brown study he jostles against somebody who angrily draws off and knocks him down, the sceptic is unlikely to carry his scepticism so far as to doubt whether anything besides the ego was concerned in

that phenomenon. The resistance shows that something independent of him is there. When anything strikes upon the senses the mind’s train of thought is interrupted; for if it were

not, nothing would distinguish the new observation from a fancy” (vol. 1, sec. 1.431). 7 Descartes, Mediations, no. 6, in Philosophical Works, 2 vols., ed. E. S. Haldane and -G. R. T. Ross (Cambridge, 1911), 1:187-89.

274

REALISM AND IDEALISM

notwithstanding, dependency may yet lie with the unconscious sector of mind. The traditional case for realism based on the limits of causal control thus fails to provide a really powerful argument for mind independence. However,a far more effective impetus to realism lies in the limitations of human intellect, pivoting on the circumstances that the features of real things inevitably outrun our cognitive reach. In placing some crucial aspects of the real together outside the effective range of mind, it speaks

for a position that sees mind independence asa salient feature of the real. The very fact of fallibilism and limitedness—of our absolute confidence that our putative knowledge does not do justice to the real truth of the matter of what reality is actually like—is surely one of the best arguments for a realism that pivots on the basic idea that there is more to reality than we humans do or can know about. Traditional scientific realists see the basis for realism in the substantive knowledge of the sciences; the present realism, by contrast, sees its basis in our realization of the inevi-

table shortcomings of our scientific knowledge.. This epistemic approach accordingly preempts the preceding sort of objection. If we are mistaken about the reach of our cognitive powers— if they do not adequately grasp “the way things really are’—then this very circumstance clearly bolsters the case for realism. The cognitive intractability of things is something about which, in principle, we.cannot delude ourselves, since such delusion would illustrate rather than abro-

gate the fact of a reality independent of ourselves. The very inadequacy of our knowledge is one of the most salient tokens there is of a reality out there that lies beyond the inadequate gropings of mind. It is the very limitation of our knowledge of things—our recognition that reality extends beyond the horizons of what we can possible know or even conjecture about it—that betokens the mind independence of the real. A brief addendum is in order here. One must be careful about what the presently contemplated sort of argument for realism actually manages to establish. For it does not establish outright that a stone—be it Peirce’s or Dr. Johnson’s or the geologist’s—is something mind-independently real. Rather, what it shows is that our conception of a “stone”—indeed

our conception of any physical object—is the conception of something

that is mind-independently real. And so the realism underwritten by these deliberations is not in fact a squarely ontological doctrine but a realism geared to our conceptual scheme for thinking about things—a realism that is idealistically grounded. (It is a certain sort of radical realism that is the target of these deliberations; radical skepticism is something that has to be addressed by other means.®) 8 The author’s Scepticism (Oxford, 1980) provides a further development of this issue.

Sixteen Scientific Realism The Fact of Scientific Fallibilism That there is a mind-independent reality—that our experience is “intentional” and inheres in a reality beyond itself upon which it has a merely cognitive rather than formative bearing—is not something that we can discover but something that we must presume from the very first. Accordingly, metaphysical realism is, if not a given, then a taken, something that we suppose from the very outset in managing our cognitive affairs.

However, the question not of whether but of what is something else again. What the descriptive character of reality is, is bound to be a matter of discovery, the result of empirical inquiry. And here scientific realism enters in via the idea that science informs us about the real—that it involves us in a realism of things as we conceive them to be on the basis of our (ultimately scientific) inquiries. The version of realism, then, is a realism of things as science conceives them to be. But does science actually furnish us with correct information about the reality of things? Exactly how are we to conceive of the adequacy relationship between the answers that natural science gives to our factual questions and the reality that they purport to depict? The history of science is not auspicious in this regard. It teaches that our discoveries in this domain constantly require adjustment, correction, and replacemént. We certainly cannot say that our current scientific theories regarding how things work in the world provide us with the real

(definitive) truth, but rather only that they provide us with the best estimate of the truth that we can achieve in the circumstances presently at hand. No one believes that there will be much substantive similarity between the scientific theories of the year 3000 and those of our own day. It is an important fact of epistemology—perhaps the single most important fact of the field—that science always involves guesswork, that we have no sure road to definitive information about the world at the level of scientific generality and precision. There is no realistic alternative to a cognitive fallibilism that acknowledges that our current body of scientific beliefs consists largely, and even predominantly, of false beliefs, embrac. ing various ideas and contentions that the wisdom of hindsight will ultimately reveal as quite untenable. We learn by empirical inquiry about empirical inquiry, and one of the key things we learn is that at no actual

276

.

REALISM AND IDEALISM

stage does science yield a final and unchanging result. All the experience we can muster indicates that there is no justification for viewing current natural science as more than yet another inherently imperfect stage within an ongoing development. We have no responsible alternative but to presume the imperfection of what we presently take ourselves to know in the scientific domain. With respect to natural science, as elsewhere, we must distinguish carefully between what we think to be true, “our (putative or ostensible) truth,” on the one hand, and on the other, “the real truth,” what actually is true—the (unqualified) truth of the matter. We realize in our heart of hearts that there is a strong prospect that we shall ultimately recognize that many or most of our current scientific theories are false and that what we proudly vaunt as scientific knowledge is a tissue of hypotheses—of tentatively adopted contentions, many or most of which we will eventually come to regard as quite untenable and in need of serious revision or perhaps even abandonment. ‘ = Committed to the unproblematic claim that reality exists, we are, nev-

ertheless, equally committed to the supposition that its nature is, in various not unimportant ways, different from what we ourselves think it to be. We can make no assured claims for our present-day science in this matter of “describing reality”: the most we can do is to see it as affording our very best estimate of nature’s descriptive constitution. We realize that science as it stands does not give us definitive knowledge. We know that we will eventually come to see with the wisdom of hindsight that each of the claims of current frontier science, taken literally in the fullness of current understandings and explanations, is strictly speaking false.1 The realities of the situation force us to accept the presumptive falsity of the claims made at the scientific frontier of the present day. We know that future scientists will think of the world in ways radically different from the way we do. To be sure, this realization is something of which we can make no effective tise—while we realize that many of our scientific beliefs are wrong, we have no way of telling which ones, and no way of telling just where error has crept in. We know or must presume that there are errors at the synoptic level, though we certainly cannot say specifically just where and how they arise.” 1 This, in effect, is the salient insight of twentieth-century philosophy of science from C. S. Peirce through K. R. Popper's Logik der Forschung (Vienna, 1935) to Nancy Cartwright’s How the Laws of Physics Lie (Oxford, 1984). 2 We occupy the posture of the so-called Preface Paradox—-standing in the shoes of the author who apologizes in the preface for those errors that have doubtless made their way into the work, and yet blithely remaining committed to all those assertions in the body of the work itself. Regarding the Preface Paradox, see Nicholas Rescher and Robert Brandom,

The Logic of Inconsistency (Oxford, 1980), pp. 47-49.

SCIENTIFIC REALISM

277

Accordingly, we have no alternative but to presume that our science as it currently stands does not present the real truth. All we can and should say is that current science affords us the best estimate of nature’s ways that we can make here and now. “Our truth” in matters of scientific theorizing is not—and may well never actually be—the real truth. However confidently science may affirm its conclusions, its declarations are

effectively provisional and tentative, subject to revision and even to outright abandonment and replacement. We must presume that science cannot attain an omega condition of final perfection. The prospect of fundamental changes lying just around the corner can never be eliminated finally and decisively. When it comes to answering our questions about the real, current science is unquestionably the best bet that we have. But our endorsement of its claims is not categorical but only tentative, provisional, and guarded. And this guardedness must, of course, also carry over to our acceptance of the theoretical entities that it envisions. Scientific progress is not of a character that encourages us to reify (hypostatize) the theory objects of science as they are presently conceived—

regardless of the date the calendar may show. Once we have takena realistic look at the history of science, it is scarcely an appealing proposition to maintain that our science, as it stands here and now, depicts reality

accurately and correctly—at best one can say that it affords an estimate of reality’s nature and mode of operation that will doubtless stand in need of eventual revision. Its creatures of theory may in the final analysis not be real at all in the form in which the theory envisions them. This feature of science must crucially constrain our attitude toward its deliverances. What decisively impedes the tenability of scientific realism is the fundamentally epistemological consideration that the world will doubtless eventuate as being very different from the way our best scientific theories currently represent it to be. Our inductive projections could be based on substantially atypical data, and the prospect of scientific innovation (and indeed scientific resolution) is in principle unending.

The Thesis of Scientific Realism What does the inescapable fallibilism of our scientific inquiries mean for the scientific world picture that we base upon them? To begin with, let us consider the plausible-sounding doctrine of scientific realism that holds that science describes the real world:3 that the world actually is as 3 This sort of descriptive realism has nothing to do with traditional ontological realism: the doctrine that the world exists independently of the thinking beings that inquire into it

278

REALISM AND IDEALISM

natural science takes it to be and that its furnishings indeed are as science envisages them to be. If we want to know about the existence and the nature of heavy water or quarks, of man-eating mollusks or a luminiferous ether, we are referred to the natural sciences for the correct answer. The

stance is that the theoretical terms of natural science refer to real physical

entities and describe their attributes and comportments: for example, the “electron spin” of atomic physics refers to a behavioral characteristic of a real, albeit unobservable, object-—an electron. On this realistic construction of scientific theorizing, the declarations of science are factually true generalizations about the actual behavior of objects that exist in the world as science takes them to be.* It is clear that this attractive-sounding position is deeply problematic. For even a casual glance at the history of natural science suffices to show that there is insufficient warrant for and little plausibility to the claim that the world indeed is as our science claims it to be, that we have got matters altogether right, so that our science—the science of today, which is the only science available to us—is correct science and offers the definitive last word on the issues. We really cannot reasonably suppose that science as it now stands affords the real truth as regards its creatures of theory. The theoretical entities of natural science are neither more nor less than the theory-creatures conceived. of by science as science conceives of them. And this means that they come equipped with the characterization that the scientific theories of the day affirm. They are indissolubly locked to the theories in which they root. And the truth is that our commitment to the correctness of those theories is far from categorical. We certainly do not (or should not) want to reify (hypostasize) the theoretical entities of current natural science—to say flatly and unqualifiedly that the contrivances of our present-day science adequately depict the actual furniture of the real world. We do not—or at any rate, given the circumstances of the case, should not—want to adopt categorically the ontological implications of scientific theorizing in just exactly the state-of-the-art configuration presently in hand. Scientific fallibilism precludes the claim that. and that its nature (having whatever descriptive characteristics it does have) is also comparably thought independent. Ontological realism contrasts with ontological idealism; scientific realism contrasts with scientific instrumentalism: the doctrine that science nowise describes reality but merely affords a useful organon of prediction and control. : 4 For some recent discussions of scientific realism, see Wilfred Sellars, Science, Perception, and Reality (London, 1963); E. McKinnon, ed., The Problem of Scientific Realism

(New York, 1972); Rom Harré, Principles of Scientific Thinking (Chicago, 1970); Frederick Suppe, ed., The Structure of Scientific Theories, 2d ed. (Urbana, Ill., 1977), Hilary Put-.. nam, Philosophical Papers (Cambridge, 1980), vol. 2; Jarret Leplin, ed., Scientific Realism (Berkeley and Los Angeles, 1984).

SCIENTIFIC REALISM

279

what we accept as scientific knowledge is in fact real knowledge, and accordingly it blocks the path toa scientific realism that maintains that the furnishings of the real world are as our science holds them to be. An epistemological realism to the effect that science gets it right (or nearly right) is counterindicated both by the facts of scientific history and by a theoretical analysis of the evidential processes on which empirical inquiry has to proceed. And the trouble with a scientific realism to the effect that the theoretical entities of science are a real part of the world’s furnishings is in the embarrassing—and ultimately invalidating—position that it really cannot be disentangled from this sort of epistemological realism. The world that we describe in science is one thing, the world as we describe it in science is another, and they would coincide only if our de-

scriptions were.totally correct—something that we are certainly not in a position to claim. The world as known is a thing of our contrivance, an artifact we devise on our own terms. Even if the data uniquely. determined a corresponding picture of reality and did not underdetermine the theoretical constructions we base upon them (as they always do), the fact remains that altered circumstances lead to altered manifolds of data. Our . recognition of the fact that the world picture of science is ever changing blocks our taking the chance that it is ever altogether adequate, let alone correct. When we say that “science describes reality,” we have to con-

strue this in the language of intent, not in that of accomplished fact. According to one expositor, the scientific realist “maintains that, if a theory has scientific merit, then we are thereby justified in concluding that ... the theoretical entities characterized by the theory really do exist.”® But this sort of position clearly has its difficulties. Phlogiston, caloric, and the luminiferous ether all had scientific merit in their day, but this plausibility relative to the evidence of the day did not assure their actual existence. Why, then, should things be all that different with us,

and our “scientific merit” now suddenly assure actual existence? What matters for real existence is clearly (and only) the issue of truth itself, not the issue of what is thought to be true at some particular stage of scientific history. And here problems arise. For its changeability is a fact about science that is as inductively well established as any theory of science itself. Science as we know it is not a static system but a dynamic process. The “that’s how it is” of our substantive scientific declarations must always be qualified by an “as we shall view the matter until further notice” at the metalevel of the epistemological stance that underlies them. 5 Keith Lehrer, “Review of Science, Perception and Reality, by Wilfred Sellars,” Journal of Philosophy 63 (1966): 269.

280

.

REALISM AND IDEALISM

The following objection might be attempted: In accepting something, one accepts it as true. And so once we accept the

contentions of science, there is no alternative to also accepting their correlative existential commitments, which, after all, are carried in their wake.

However, this argumentation’s basic premise ignores the fact that the matter of acceptance is not all that simple. It overlooks that scientific acceptance is not just acceptance (any more than a stuffed owl is not just an owl). What is at issue is not just acceptance as such but a tentative or provisional acceptance, predicated on the idea that what is being accepted, rather than being the definitive truth of the matter, is simply the best estimate of the truth that can be made in the present state of the art.

Against Homo Mensura: Is Man the Measure? In considering the question of what “reality” is all about—of what it is to be real—we do well to take the approach of a causal realism, which holds that to be real is to be part of the world’s causal commerce: to be a cause of or an effect of something that is patently and unproblematically real. But how, then, is reality related to knowledge? Whatever can actually be known by us humans to be real must, of course, for that very reason, actually be real. But does the converse hold?

Is the real for that very reason also automatically knowable? Is it appropriate to join C. S. Peirce, who, in rejecting “incognizables,” insisted that whatever is real must be accessible to cognition—and indeed must ultimately become known? Is humanly cognizable reality the only sort of reality there is? Some philosophers certainly say so, maintaining that there actually is a fact of the matter only when “we [humans] could in finite time bring ourselves into a position in which we were justified either in asserting or in denying _ [it].”° On such a view, reality is our reality. What we humans are not in a position to domesticate cognitively—what cannot be brought home to us by (finite!) cognitive effort—simply does not exist as a part of reality at all. Where we have no cognitive-access, there just is nothing to be ac- cessed. We are led back to the homo mensura doctrine of Protagoras: “Man is the measure of all things, of what is, that it is, of what is not, that it is not.” However, in reflecting on theissue in a modest mood, one is tempted

to ask: “Just who has appointed us to this exalted role? How is it that we 8 Michael Dummett, “Truth,” Proceedings of the Aristotelian Society 59 (1958-59): 160.

SCIENTIFIC REALISM

281

humans are established as the ultimate arbiters of reality as such?” There are two routes to establishing adequacy in this regard: the power of mind, and the poverty of reality. If “the (ultimately) known” and “the real” are said to stand in strict correlation, the question still remains open: on which side does the final responsibility for this coordination lie? Does it obtain because cognition is so powerful in its grasp that nothing real can permanently elude it, or does it obtain because reality is so restricted in its scope that inquiry cannot fail to exhaust it? Neither alternative seems all that attractive. On the issue of whether what is real must be knowable,. traditional

mo bop

realism takes an appropriately modest line. It insists on preserving, insofar as possible, a boundary line of separation between fact and knowledge of fact: between truth-status possession and truth-status decidability with respect to propositions and between entity and observability with respect to individual things. As the realist sees it, reality should be presumed to have depths that cognition may well be unable to plumb.. To be sure, it is possible to reduce the gap between fact and cognition by liberalizing cognizers. Consider the following series of metaphysical theses: For something to be real, it is necessary for it to be experientiable by: Oneself. One’s contemporary (human) fellow inquirers.

Us humans (at large and in the long run). Some actual species of intelligent creatures. 5. Some physically realizable (though not necessarily actual) type of intelligent being—creatures conceivably endowed with cognitive resources far beyond our feeble human powers. 6. An omniscient being (i.e., God).

The idea of an experiential idealism that equates reality with experientiality is one that can be operated on rather different levels. Specifically, the “n-th level” idealist maintains—and the “n-th level” realist denies— a view of reality geared to the thesis maintained at stage (n) of the preceding series. On this approach, the idealist emerges as the exponent of an experientiability theory of reality, equating truth and reality with what is experientially accessible to by us—with different, and potentially increasingly liberal, constructions ofjust who is to figure in that “us group” of qualified cognizers. Where one sets the boundary in interpreting the idealist thesis that to be real is to be cognizable will determine the sort of idealism that is at issue. No sensible idealist maintains a position as strong as the egocentrism of (1). No sensible realist denies a position as weak as the deocen-

282

.

REALISM AND IDEALISM

trism of (6). The salient question is just where to draw the line in determining what is a viable realistic-idealistic position. Let us focus upon case (3), the “man is the measure,” homo mensura doctrine. By this standard, both Peirce and the Dummett of the preceding quotation are clearly homo mensura idealists, seeing that both confine the real to what we humans can, in principle, know. But this is strong stuff. Of course, what people can know to be real constitutes (ex hypothesi) a part or aspect of reality at large. That much is not in question. The bone of contention between such homo mensura realism and a sensible idealism is the question of a surplus—of whether reality may have parts or aspects that lie beyond the reach of human cognition. Undoubtedly, a naturally evolved mind has a sufficiently close link to reality to be able to secure some knowledge of its features and furnishings. But the converse is eminently problematic. It. is dubious that the

linkage should be so close that only what is knowable for some actual being should be real—that reality has no-hidden reserves of fact that are not accessible within the cognitive reach of existing creatures (let alone one particular species thereof!). Accordingly, it seems sensible to adopt the idealistic line only at level (5), and to be a realist short of that. Essen-

tially this is the position of the causal-commerce realism that has been exposed in this discussion. To be real is to be in a position to make an impact somewhere on something of such a sort that a suitably equipped mind-endowed intelligent creature could detect it. What is real in the world must make some difference to it that is in principle detectable. Existence-in-this-world is coordinated with perceivability-in-principle. But of course such a position represents a realism that is relatively weak. The present experiential idealism is accordingly geared to the situation at level (5). In effect, it takes the line that to be real is to be causally active—to be a part of the world’s causal commerce. For since one can always hypothesize a creature capable of detecting a given sort of causal process, we need not hesitate to equate reality with experientiability in principle. We thus arrive concurrently at a level (5) idealism, one that. achieves its viability through its comparative weakness. At all the lowernumbered, more restrictive levels, our position is effectively realistic. The resultis a halfway-houseposition that combines an idealism of sorts with a realism of sorts. A qualified realism of this description holds that whatis so as a matter of factis not necessarily cognizable by us, no matter how far we extend

the boundaries of that “us-community” of inquiring intelligences, as long as we remain within the limits of the actual. Any particular sort of possi- ble cognizing being can know only a part or aspect of reality. (But the situation changes when we move to the level of collectivity; one cannot

SCIENTIFIC REALISM

283

make plausible sense of “such-and-such a feature of nature is real, but no possible sort of intelligent being could possibly discern it.”) The sort of descriptive realism to which we can subscribe appropriately is thus a rather shadowy one. For we can do no more than lay claim to an aspectival reaction that says that whatever reality may really and truly be like, we do have current knowledge of some of its aspects, namely, those dispositional ones relating to how it causally affects our senses and sensors. But of course this is a rather minimal sort of realism. It does not establish human beings as the measure of nature’s determinations, but only as incidental bystanders. And so, homo mensura realism is untenable. There is no good reason

to resort to a hubris that sees our human reality as the only one there is. Neither astronomically nor otherwise are we the center around which all things revolve. A homo mensura realism that equates reality and the nature of existence with what is knowable by us humans is totally implausible. After all, humans have the capacity not only for knowledge but also

for imagination. And it is simply too easy for us to imagine a realm of things and states of things of which we can obtain no knowledge because “we have no way to get there from here,” lacking the essential means for securing information in such a case. Be this as it may, the salient point of these deliberations remains—that being “knowable,” “identifiable,” or “imaginable” or otherwise subject to those pivotal cognitive potentialities operative in the realism-idealism controversy need not be relativized to us humans.

Realism and the Aim of Science ’ As these deliberations indicate, one must maintain a clear distinction be-

tween “our conception of reality” and “reality as it really is.” Given the equation Our (conception of) reality = the condition of things as seen from the standpoint of “our putative truth” (= the truth as we see it = the science of the

day), we realize full well that there is little justification for holding that our present-day science indeed describes reality and depicts the world as it really is. In our hearts of hearts, then, our attitude toward the declarations of our present-day science is—and should be—one of guarded affirmation. In science as elsewhere, there is a decisive difference between

achievement and aim—between what science accomplishes and what it endeavors to do. (Recall our strictures re: estimation.) Nevertheless, the posture of scientific realism (of a duly qualified sort)

284

.

REALISM AND IDEALISM

is built into the very goal structure of science. The task of science, the definitively characteristic mission of the enterprise, is to provide the very best answers to our questions about the world that we can manage to obtain. On the traditional view of the matter, this question-resolving concern is the very reason for being of the project. The characteristic doctrine of scientific realism is its equating of the resources of natural science with the domain of what actually exists. But this equation would work only if science, as it stands, has-actually “got it

right.” And this is something we are certainly not minded—and not entitled—to claim. The postulation as real of the commitments of our science is. viable only if done provisionally, in the spirit of viewing science as “doing the best we can now do, in the current state of the art” and giving our best estimate of the matter. The step of reification is always to be taken provisionally, subject to a mental reservation of presumptive revisability. We do and must recognize that we cannot blithely define the truth to be our truth. We do and mustrealize that the declarations of science are inherently fallible—that they are mere estimates and that we can accept them only with a certain tentativity, subject to a clear realization that they may need to be corrected or even abandoned.

It is accordingly useful to drawa clear distinction between a realism of intent and a realism of achievement. We are certainly not in a position to claim that science as we have it achieves a characterization of reality. In intent or aspiration, however, science is unabashedly realistic: its aim is

unquestionably to answer our questions about the world correctly and to describe the world “as it actually is.” The orientation of science is factual and objective: it is concerned with establishing the true facts about the real world: The theories of physics purport to describe the actual operation of real entities—the Nobel prizes awarded for discovering the electron, the neutron, the pi-meson, the antiproton, and so forth, were in-

tended to recognize an enlargement of our understanding of nature, not to reward the contriving ofplausible fictions or devising clever ways of relating observations. And so, the language of science is descriptively committal. At the semantic level of the content of its assertions, science makes firm claims as

to how things stand in the world. Scientific realism skates along a thin border between patent falsity and triviality. Viewed as the doctrine that science indeed describes reality, itis utterly untenable; viewed as the doctrine that science seeks to describe reality, it isvirtually a truism. For thereis no way of sidestepping the conditional thesis: If a scientific theory regarding heavy water or electrons or quarks or whatever is correct—if it were indeed to be true—then its subject materials would exist in the manner the theory envisages and would have the properties the theory

SCIENTIFIC REALISM

285

attributes to it: the theory, that is, would afford descriptively correct information about the world.

But this conditional relationship reflects what is, in the final analysis, less a profound fact about the nature of science than a near truism about the nature of truth as adaequatio ad rem. The fact remains that “our reality”"—reality as we conceive it to be—goes no further than to represent our best estimate of what reality is like. When we look to what science asserts, to the content and substance of its declarations, we see that these declarations are realistic in intent, that

they endeavor and indeed purport to describe the world as it really is.

But when we look to how science makes its assertions and note the tentativity and provisionality with which they are offered and accepted, we recognize that this realism is of an abridged and qualified sort—that we are not prepared to claim that this is how matters actually stand in the real world. Despite a commitment to realism at the semantic level of assertion content, there is no longer a commitment to realism at the epistemological level of assertoric commitment. Realism prevails with respect to the language of science (i.e., the content of its declarations); it is abandoned with respect to the status of science (i.e., the ultimate tenability or correctness of these assertions). What science says is descriptively committal in making claims regarding the real world, but the tone of voice in which it proffers these claims always is (or should be) provisional and tentative. All the same, commitment to a realism of intent is built into science

because of the genesis of its questions. The ultimate ground of the factually descriptive status of science lies in just this erotetic continuity of the issue of science with those of prescientific everyday life. We begin at _the prescientific level of the paradigmatic realities of our prosaic everyday-life experience—the things, occurrences, and processes of this experience. Given that the teleology of the scientific enterprise roots in the “real world,” which provides the stage of our being and action, we are committed within its framework to take the realistic view of its mechanisms.’ Natural science does not address itself to some world-abstracted realm of its own. Its concern is with this familiar real world of ours in which we live and breathe and have our being—however differently science may characterize it. However gravely science may fall short in performance, nevertheless in aspiration and endeavor it is unequivocally committed to the project of depicting the real world, for in this way alone could it realize its constituting mandate of answering our questions as to 7 Recall the discussion of chap. 3, which argued that the protocategories can never be entirely abandoned in favor of the scientific categories of thought because the latter get

their point through their exfoliative origin in our protoquestions.

286

.

REALISM AND IDEALISM

how things work in the world. This perspective reinforces the previous contention that we are irrevocably committed to viewing the lineaments of the world as depicted by science as an at any rate tentative or aspiring depiction of how things actually work in the world. One of the significant strengths of semantic realism roots in the weakness of epistemological realism. If we are fallibilists with respect to science—that is, if we believe (as we surely must!) that nature has repeatedly proved us wrong in the past and will probably continue to do so in

the future as well—then we must accept that our scientific theories at least purport to describe nature. It is clear that scientific claims can be undermined, let alone falsified, only if they have a substantive content

and at least purport to state what is true.

Instrumentalism and the Descriptive Purport of Science Should this fact that we cannot claim that our science, as it stands, cor-

rectly depicts reality be construed to mean that science has the status of a merely practical device—a mechanism of prediction and control that is (or properly should be thought of as being) devoid of any actually descriptive purport? Does science perhaps afford no objects of belief at all, but merely a guide to action? Such a stance is embodied in the doctrine of instrumentalism, which embraces roughly the following position: Science has no descriptive or existential import. It is simply an organon of prediction, a device for calculating what observational consequences will ensue (or will probably ensue) if certain things are done (or left undone)—above all, what results will be obtained when certain measurements are taken under certain conditions. It is an instrument for generating reliable predictions and guiding effective control. It is a “black box,” as it were, into which we put some information (data or assumptions) and get out predictions about events or instructions for modes of intervention. But the propositions that figure in the contents of this black box must be construed as devoid of any claims to describ- _ ing the world. As an instrumentality ofprediction and control, science is wholly free of commitment that certain sorts of things really exist and actually have such-and-such a nature. The theories adopted by science are not to be construed as assertoric propositions. ‘They are mere rules for drawing inferences to actual or possible observations. These rules themselves are neither true nor false: they are ways of dealing with the phenomena that work or do not work, are or are not successful and fruitful as guides to prediction and control. -

The instrumentalist sees our scientific theories as no more than a practical device—a set of rules for effective interventions and verifiable predictions.

SCIENTIFIC REALISM

287

Such a position gravely misconceives the character of the scientific enterprise. Admittedly, it is interested in prediction and control, but that is only a small part of it. What we seek above all—nay, what constitutes the very reason for being of science—is information, that is, description and classification and explanation, getting answers to our questions about how things go in the world. The statements of science are always made with descriptive and informative intent. And its descriptive orientation creates an unavoidable linkage between the assertions of science and objective reality—at any rate on the side of aim and aspiration. As the instrumentalist sees it, science must be emptied of descriptive substance: it is a mere instrumentality of prediction and control over nature, uninvolved in any descriptive or explanatory purport. But this selfdenying ordinance is clearly a position of last resort, to be adopted only ifno sense could be made of the phenomena (and in particular, those that

indicate the uneliminable prospect of scientific fallibility) without yielding up the idea of a commitment to truth. But the view of science as a venture in truth estimation neatly steps between the horns of this dilemma. _ Scientific inquiry, as we have argued, is a matter of achieving the very best estimate of the real truth that is available in the circumstances. And the idea of an “estimate of the truth” cannot be stripped of descriptive purport. The contentions of our science may not be—nay, presumably are not—adequate to reality itself, but they would not be what they are if one refrained from recognizing their descriptive intent and failed to acknowledge their purporting (although perhaps failing) to depict matters as they stand in the real world. The intention and aspiration to describe the world (however much our performance may fall short of its realization) is a crucial aspect of the goal structure of science. To fail to recognize this fundamental fact is to go badly awry in our understanding of the definitive nature of the enterprise—to fail to recognize what is the very reason for being of the venture (however far our actual performance may fall short of its realization). To be sure, science cannot be dogmatic in regard to its descriptive mission. It must, no doubt, refrain from claiming definitive correctness

for the descriptive picture of the world it puts before us. But while science cannot be dogmatic here, it can, nevertheless, unhesitatingly claim to do the very best that can be done in the circumstances at hand. And while we recognize that we cannot claim correctness for the descriptive claims of our science, we nevertheless do—and can—view them as

best estimate of the matter that can be made with the cognitive technology of the day. That, after all, is exactly what truth evaluation is all about:

it is the very essence of an estimate that it should seek to estimate the truth of the matter.

288

.

REALISM AND IDEALISM

A recognition that science is fallible—that the purported truths of current science may well be false and that the world’s furnishings may well not actually exist as our science conceives them to be—does not carry instrumentalism in its wake. We need not abandon the descriptive and explanatory intent of scientific declarations and pivot the whole matter on pragmatic issues of prediction and control. Scientific realism, as a semantic thesis about the inherent intent of science’s declarations, is not

abolished by a concession of fallibility: we can maintain our best estimates in the face of a recognition that they are mere estimates. The best available answer to a question is worth having, even where one realizes that a better answer is bound to come along. Instrumentalism draws wrong conclusions from the undoubted fact of the fallibility and corrigibility of science. This corrigibility does not mean that we cannot make existential and descriptive claims: it merely means that we must make them provisionally, talking in the hypothetical mode:

if our science is correct, then electrons exist and have such-and-such features, and so on. We must acknowledge that our science is ontologically committed in its existential and descriptive purport (in its aim or intention), though doubtless imperfect in its executions of this mandate. Admittedly, at the level of existing science, our explanatory and descriptive claims must be viewed as no more than corrigible estimates of the truth. But at the level ofperfected science, we clearly would want to make such claims, and this very fact shows how our intentions are directed. Instrumentalism insists that it is necessary to reorient the goal structure of science away from its traditional teleology of question answering. On its basis, our scientific theories are to be viewed as a contentless black box—a mere computing device that provides a calculating mechanism mediating between data inputs and data outputs, but yet itself a mere instrument of conveniences devoid of any descriptive import—a useful fiction that produces usable results. But why should one take this hypermodest stance? What we have here is clearly no more than a fallback position that would make sense only if the traditional descriptive intentions of science could be shown in some decisive way to be, in principle, -improper and illegitimate. But there is no good reason for taking this line. To abandon its informative aspirations is to abandon science as we know it. . A commitment to realism is thus inherent in the very teleology of sci- . ence. In abandoning realism, we would turn our back on the definitive aims and tasks of the scientific enterprise. In attempting answers to our questions about how things stand in the world, science offers (or at any rate, endeavors to offer) information about the world. The extent to which science succeeds in this mission is, of course, discussible. (And no doubt in this discussion the issue of success in prediction and control will

SCIENTIFIC REALISM

289

have to play a central role.)* But this does not alter the fact that science both endeavors and purports to answer our questions about the world by offering a realistically authentic description of what the world is actually like. To be sure, here as elsewhere one must be mindful of the distinction

between aim and achievement, between what we set out to do and the extent to which we actually manage to do it. But clearly, the circumstance that we can never appropriately claim to have realized in full science’s goal of a correct description of natural processes is something that does not constitute a sufficient reason for abandoning this goal altogether. In inquiry, as in the moral life, we can never attain perfection. But this is surely no reason for abandoning thé endeavor. The lesson of these deliberations is that an unabashed scientific realism is untenable, except in the context of a utopian ideal. The real truth about nature’s ways is not given in yesterday's science, nor in today’s, nor yet in tomorrow's. Scientific realism is not a plausible doctrine with respect to science as we even actually have it in hand—now or in the future. It is a plausible doctrine only with respect to ideal science. Ironically, realism is geared to idealism in this domain: only from the vantage point of an idealistic projection beyond the actually realized condition of things can we plausibly be scientific realists. Scientific realism takes the position that we should see the theoretical entities of natural science as real things—that reality indeed is as science takes it to be. But of course the things (entities, fields, forces, etc.) come laden with the description that science provides, with all of the detail that

this involves. And this is something that we are (rightly) bound to feel queasy about. That atoms are exactly as the science of the day envisions them to be is quite doubtful—at least if the history of science affects any indication. The only sort of serious scientific realism that we can unproblematically acknowledge is one that is geared not to real but to ideal science. But since we cannot arrive at the ideal, a halfway-house position affords something of a practical compromise.

Schoolbook Science as a Basis for Realism It is of the nature of technical natural science, as its practitioners actually pursue it at the research frontier, to state how things stand exactly and how they go always and everywhere—in full generality and precise detail. Technical science forswears any loose talk at the level of vague generality or analogy or approximation. It has no use for qualifiers such as § See the author's Methodological Pragmatism (Oxford, 1977).

290

.

REALISM AND IDEALISM

“usually” or “roughly.” Universality and exactness are its touchstones. This circumstance renders the claims of science vulnerable, with vulnerability being the price that we pay for generality and precision. We know that none of the hard claims of present-day frontier natural science will move down the corridors of time untouched. The contrast between the cognitive approach of science and that of common sense comes to the fore here. As we have seen, increased con-

fidence in the correctness of our estimates can always be purchased at the price of decreased accuracy. But in science we put ourselves at greater risk because we ask more of the project. Thus in this contrast between natural science and common sense, science wins out in point of

precision and detail, but common sense wins out in point of security. At the level of scientific generality and precision, we have no secure and invulnerable informetion, while at the level of common sense, we get no detailed understanding. As the aims of the enterprises are characteristically different, our inquiries in everyday life and in science have a wholly different aspect, with the former achieving stability and security at the price of sacrificing definiteness, a price that the latter scorns to pay. “But science is the best, most thoroughly tested knowledge we have—the ‘knowledge’ of everyday life pales by comparison. The theses of science are really secure and well established, those of everyday life causal and fragile.” Yet in fact the very reverse is the case: our scientific theories are vulnerable and have a shortish life span; it is our claims at the looser level of ordinary life that are very secure and stable. It thus appears that in managing the business of realism, we should put our trust in the bank of common sense rather than in that of natural science. And yet common sense too has its problems. For its concepts are loose, fuzzy, general. Its world is populated with things such as trees,

rocks, gases, and all those conceptions of a rudimentary knowledge that we deem too loose and sloppy for the purposes of explanatory understanding. The adequacy of this sort of knowledge in point of accuracy (rather than security) is something that science has undermined for good and all. The scientific project itself teaches us that the concepts of every- .. day life are too vague to characterize adequately the world as it really is—seeing that it was the unsatisfying character of the “stone-age phys-ics” of untutored common sense that drove us to develop science in the

first place.

.

.

Our ordinary, commonsensical concepts such as “tree” and “rock” are

useful enough in their way, facilitating the coordination of practical activities (e.g., working together to clear a field). But they lack the precision . required for satisfactory explanations. And yet, in seeking this precision, science courts error; the claims of the state of the art in science at any particular time doubtless need eventual revision and correction. To char-

SCIENTIFIC REALISM

291

acterize really adequately, we need theses that are both relatively secure and relatively precise. And yet we are caught betwixt and between. Common sense is capable of yielding secure claims, but only vague ones; science yields only insecure claims, but precise ones. What we need for realism is information about reality that is both tenable (secure) and highly informative (definite). But just this is something that the exclusion principle encapsulated in the security-definiteness trade-off renders problematic. From the angle of precision, science is in good shape, but

its condition is very problematic from the angle of truth. Common sense, on the other hand, stands on firmer ground in relation to security but is

so untidy conceptually that it lacks precision. Either way, the epistemic component of realism is in difficulty. But is there no middle ground? There is indeed! It is elementary school-level science-made-simple that comes to our rescue.

.

What we desire and require is a tenantable halfway house between an untenable scientific realism subscribing to the mistaken supposition that the theories of natural science correctly describe physical reality, and a description-abandoning instrumentalism subscribing to the bankrupting thesis that the theories of natural science should be construed nondescriptively, as merely coordinating the phenomena. Exactly such a halfway house is provided by what we have characterized as “schoolbook science, which is able to provide a realism of the middle range—the realism of the elementary science of popularizations and school texts. Are there atoms? Is the atomic theory true? It all depends. It is a matter of just what the “atomic theory” is—just what it is to be taken to assert. Taken at face value, the atomic theory that we now have is surely problematic. After all, exactly what are atoms like? Neither Rutherford, nor

Bohr, nor Weinberg-Salam, nor the nuclear physics of today provides a definitively correct picture of the atom. We clearly cannot trust the science of the day to the extent of saying that we know just exactly what atoms are like. All the same, atoms at large (or better, “at loose”) are altogether safe. There is no danger that atoms and molecules will cease to play a role in physics and chemistry. Even though we shall doubtless have to revise our conceptions of them, the things as such will not go away. The looser and detail-suppressive knowledge of popularized science— indebted though it is to technical science—is not itself scientific knowledge. Technical science scorns imprecision and operates at the outer limits of exactness and accuracy. It would not stoop to say the sorts of detailignoring things we find in science reporting or in introductory accounts for laypersons. But it is just this sort of popularized schoolbook science

292

REALISM AND IDEALISM

that is the focus of our confident assurance that we know how things stand in the world: One recent writer poses the following worry. Recent science abandons the luminiferous ether. What iffuture science did the same with the electron? “Then we will have to say electrons do not really exist. What if this keeps happening? What if all the theoretical entities postulated by one generation (molecules, genes, etc., as well as electrons) invariably do not exist from the standpoint of later science?”® But this worry. is simply another instance of the “what if” thinking of traditional skepticism. (“What if life were a dream?”) Such handwringing—abandoning electrons, molecules, and genes (which, after all, we can “observe” with contemporary

technology)—is hyperbolic. We can—and doubtless, will—come to think of these things very differently from the way in which we conceptualize them today. But given the conceptual plasticity and flexibility of what is at issue with “electrons,” “genes,” and so forth, the prospect of their total disappearance is on the same level as the prospect of the total disappearance of Julius Caesar. It is not that this is something logically impossible. It is simply that this is one of these unrealistic possibilities that need not worry us in conducting practical policies for our theoretical concerns. In succumbing to this sort of worry, we leave the issue of science behind and plunge into the generalized skepticism of farfetched “what if” speculation. , Moreover, the argument of the preceding passage has another defect when seen as an objection to scientific realism. Admittedly, scientific entities can become dislodged—phlogiston moved aside to make room for oxygen, the “gravitational attraction” of Newton gave way to the space warp of Einstein, radiation succumbed to fields and ceased to be a matter of waves-in-a-medium, thereby abolishing the luminiferous ether. But the disappearance of some scientific entities does not mean that they all disappear. On the contrary, they do not vanish at all; they are replaced: it takes one to dislodge one. -And so the preceding “argument from error” is not an argument against scientific realism but simply one against the ontological finality of science as we have it. At first people thought that ~ humankind’s protoancestor was akin to a gorilla, then the baboon came into favor, only to be displaced eventually by the chimpanzee. But such a series of changes does not support the induction that humans have no ancestor at all. Similarly, we cannot,in the case of science, construe the historical sequence of changes of mind about theoretical entities as providing an argument against the reality of theoretical entities at large. . We cannot simply rely on the present state of science to give us an ® Hilary Putnam, “What Is Realism?” in Scientific Realism, ed. Jarrett Leplin (Berkeley and Los Angeles, 1984).

SCIENTIFIC REALISM

293

unproblematically acceptable picture of the world, because we realize that this picture is defective. The frontiers of science will move on, and with the wisdom of hindsight it will come to be seen by our successors as every bit as inadequate as we ourselves deem the science of our predecessors of a hundred years ago. But our knowledge of the world nevertheless rests on scientific knowledge. It does so by taking the hard scientific information of the day and “loosening” it up: by removing all its fine-grained detail. We constantly get more information of high security, albeit information that is loose and imprecise. With the growth of our scientific knowledge (vulnerable though it is!), there is an ever-increasing number of things we can claim with high confidence. Certainly, our looser schoolbook science is not invulnerable—it too is

not absolutely secure and definitive but will doubtless undergo sea changes in the course of time. The point is rather that it is all relative— that schoolbook science is very secure in comparison with technical science and very informative in comparison with the stone-age physics of rational common sense. , This loosened-up popularization of science at issue in schoolbook science is, of course, generically scientific—it is indebted to technical sci-

ence and undergirded by it. But it is insufficiently detailed and therefore insufficiently deep and cannot answer adequately our demands for a detailed understanding of how things work in the world. (That is just what marks it as “popular” rather than “real” science.) All the same, it gains its crucial advantage of security precisely because of its loosensss. It achieves that middle ground of concurrent security and informativeness that is essential to any viable sort of information about the real. Absolutely definitive knowledge of nature being unattainable, we can and should be content to let our realism operate within the limits of the attainable represented by the middle range of the security-definiteness curve. Although the claims of our technical science will require eventual correction at the level of detail and precision at which they are cast, nevertheless the overall picture that emerges from science is doubtless right in its rough outlines. An increasingly adequate picture of nature emerges, not in technical science, but through technical science. We realize that we have to be prepared to revise any and every thesis of hard science— that none of its present-day theories will survive unscathed to the year 5000 wholly without revision. But this is not the case with the grosser materials of schoolbook science. (Atoms of some sort will be with us from here on in.) Schoolbook science not only claims that there are atoms, magnetic fields, and genes but obviously also involves some claims about what they are like. Yet they are only rather rough and inexact claims. Unlike sub-

294

.

REALISM AND IDEALISM

stantive science, with its aggregate of detailed theories, it involves no commitment to the precise details of any particular theory whatsoever. A realism of schoolbook science opens up the clearly plausible prospect of being realistic about the theoretical entities of science (holding they are real) without being realistic about the current theories of science (holding they are actually true and give a correct account of reality). Such a realism of schoolbook science is a physical realism that is neither merely a commonsense realism nor astrictly scientific realism but resides in a halfway house represented by popularized science. It is a softer scientific realism that draws on technical science not for its details but for the rough essentials. This approach puts us in the fortunate position of being able to hold that the objects discussed in our scientific theories have a life independent of those theories and do not stand or fall by their correctness.’° It is precisely our acknowledgment of the looseness of our knowledge about atoms and molecules that renders the existence of such things secure. If the existence of something is subject to the formula “to be is to be (exactly) as we now deem it to be,” then its existence would stand on very shaky ground indeed. The great advantage of a realism of schoolbook science is that it disconnects the issue of the existence of theoretical entities from the tenability of our particular theories, making it possible for such entities to have a life independent of our current beliefs about them. It makes it possible to say (surely rightly!) that we need not maintain the definitive truth of any of the current formal theories of natural science in order to maintain that science provides us substantial information about the workings of nature at the level of observables and unobservables alike. What we obtain on this basis is not a full-blown scientific realism that claims the entities discussed in technical theoretical science exactly as this science describes them. It does not maintain that we can read off the way the world is from the statements endorsed by natural science as it

stands. Rather, it is a realism that is indebted to science but does not _. involve reifying the theoretical entities of science in just the way currently envisioned by natural science. It is a realism that becomes available not in but through natural science as we have it—not a scientific realism, but a science-indebted realism. The realism of schoolbook science leaves us very much with a half-full barrel. It puts us well ahead of a bare-bones metaphysical realism that merely maintains that there is a mind-independent reality, but it is un10 Its far looser tie to theory makes experimentation a bulwark of scientific realism. Note Jan Hacking’s Representing and Intervening (Cambridge, Mass., 1983), esp., chap. 16, “Experimentation and Scientific Realism.”

SCIENTIFIC REALISM

295

able to provide detailed descriptive information about this reality. For on the basis of schoolbook science, we can know a good deal about nature—

and can come to know ever more about it as inquiry proceeds. All the same, what we learn at this level of schoolbook science is vague, imprecise, and general, rather than specific, exact, and accurate. It is gravely deficient in informativeness, giving us a picture of reality seen “through a glass, darkly” without the precision and detail of a scientifically exact idea that we would dearly love to have. Nevertheless, as technical science develops, the schoolbook realism it

pulls along in its wake enlarges as it grows both in scope and in detail. Such improvements never take us as far as we would like—of necessity the position we reach is always far from the cutting edge of precision and detail. Still, this half-full barrel is not empty. It combines a recognition of the fallibility of natural science with a robust realism that talks of truth

and real existence even when unobservables are involued—a realism that accepts electrons and genes (answering to rough-and-ready descriptions) without categorically affirming our current scientific conceptions about such things. This is where schoolbook science gains its utility for realism. Even as there is more to an apple or a piece of rock than our own potentially incorrect ideas about them, so there can be more to an electron or a gene

than the current theories of science-as-it-stands envision. It is thus perfectly possible to be an ontological realist about theoretical entities without being a semantic realist (i.e., truth endorser) with respect to the current theories in which they figure. We are relieved of any need to think that our theories fit their objects perfectly—that those objects stand or fall by the correctness of our present ideas about them. There is (and should be) more to our theoretical entities than meets the eye of current theorizing, seeing that our theories presumably get it wrong, doing no better than to have the right general idea. It is here that schoolbook science comes into its own and is able to provide a crucial support for realism. A realism based on schoolbook science is an attractive position because it is able to reconcile two facts—on the one hand, we cannot claim that

natural science as it stands characterizes reality correctly, but on the other hand, we cannot simply dismiss it as totally uninformative about the way the world is. In the choice between a naive scientific realism and a skeptical scientific fallibilism, recourse to schoolbook science affords a middle way enabling us to acknowledge the claims of realism despite the ambivalence of our attitude toward the deliverances of current science." 4 This chapter draws upon the author’s Scientific Realism (Dordrecht, 1987).

Seventeen Science and Idealization Scientific Realism at the Level of Ideal Science Once we distance ourselves from the particular cognitive commitments of our science by recognizing that they can and frequently do go awry,

we must also acknowledge that our scientific picture of reality is not fully accurate, admitting that we have neither the inclination nor the warrant for claiming that reality actually is just as the science of the day purports it to be. Given this circumstance, we have little alternative but to presume reality to have a character regarding which we are only imperfectly informed by natural science. Once we have taken a careful and realistic look at the history of science, it is scarcely an appealing proposition to maintain that our science, as it stands here and now, depicts reality actually and correctly. Success in providing the definitive truth about nature’s ways is a matter of intent rather than one of accomplishment. Correctness in the characterization of nature is achieved not by our science but only by perfected or ideal science—only by that (ineradicably hypothetical) state of science in which the cognitive goals of the scientific enterprise are fully and reliably realized. Science is imperféctable. We are constrained to acknowledge that it is not present science, or even future or ultimate science, but only

ideal science that correctly describes reality—an ideal science that we shall never in fact attain, since it exists only in utopia and not in this mundane dispensation. Scientific realism must thus come to terms with the realization that reality is depicted by ideal (or perfected or “completed”) science, and not by the real science of the day, which, after all, is the only one we have actually got—now or ever. As we saw in chapter.. 12, “the real truth” in matters of rational inquiry has to be seenin the light of idealization. Our science is constituted of putative knowledge that does no more than to estimate the truth as best we can discern it with the limited means at our disposal.

We must maintain a certain tentative and provisional stance toward our own scientific knowledge. We fully realize that what we take to be true or real here is not always true or real. It is just this consideration that constrains us to operate with the distinction between “our putative reality” and “reality as such.” We realize that what we think to be so—be it in science or in common life—frequently just is not so. We certainly can-

SCIENCE AND IDEALIZATION

297

not identify our achieved putative scientific truth with the real truth of the matter. No route except idealization can yield a sure and general connection between belief and the real truth. Only ideal or perfected science accurately and correctly depicts reality, and not science as we actually have it here and now. From the standpoint of epistemic status, truth is clearly an idealization—not what we do have (or ever will have), but what we could have if all the returns were in. It thus seems promising to contemplate a position along the lines envisioned in Fichte’s Wissenschaftslehre, which sees the ideal as providing the determining factor for truth and reality. On such a view, the real is not characterized by the science we actually have, but by the ideal science that is the telos of our scientific efforts. This position, which Wilhelm Wundt characterized as “ideal-realism’” (Idealrealismus),1 holds that

the knowledge that achieves adequation to the real (adequatio ad rem) by adequately characterizing the true facts in scientific matters is not the knowledge actually afforded by present-day science as we have it but only that of an ideal or perfected science. On such an approach (which has seenalively revival in recent philosophy), a tenable version of sci-

entific realism requires the step to idealization, and realism becomes predicated on assuming a fundamentally idealistic point of view. After all, we cannot justifiably claim that the world is such that the paraphernalia of our science actually exist as such. Given that we must recognize its claims to be tentative and provisional, one cannot appropriately maintain that our science depicts reality: all we can say is that it affords an estimate of it, an estimate that will, we may be sure, stand in need of eventual

revision, and whose creatures of theory may, in the end, not be real at all in anything like the currently contemplated ways. This feature of science must crucially constrain our attitude toward its deliverances. Depiction is in this regard a matter of intent rather than one of accomplishment.

It is not our science that correctly characterizes nature but only ideal science. And ideal science is not something we have got in hand here and now. Nor is it something toward which we are moving along the asymptotic and approximate lines envisioned by Charles Sanders Peirce.? For Peirce identifies ideal or perfected science with an ultimate condition of science that is fated to emerge in the eventual course of history. But there is, of course, no guarantee whatsoever for this cognitive millenarianism. Perfected science is not “what will emerge when” but “what would emerge if’—wherealot of (realistically unachievable) conditions of completeness and comprehensiveness must be satisfied. As far as the 1 Wilhelm Wundt, Logik, 2d ed., 2 vols. (Leipzig, 1895), 1:86-92.

2 Recall the discussion of this issue in chap. 4 above.

298

.

REALISM AND IDEALISM

actual course of history goes, we must recognize that even if it made

sense to contemplate the Peircean idea of an eventual completion of science, there would be no guarantee that this completed science (given it

existed!) would satisfy the definitive requirements of perfected science. Peircean convergentism is geared to the supposition that ultimate sci-

ence—the science of the very distant future—will somehow prove to be an idea or perfected science freed from the sorts of imperfections that afflict its predecessors. But the potential gap thatarises -here can be closed only by metaphysical assumptions of a most problematic sort.® The concept of science perfected—of an ideal and completed science that captures “the real truth” of things and meets all of our cognitive desiderata (definitiveness, completeness, unity, consistency, etc.)—is at best a useful fiction, a creature of the imagination and not the secured product of rational inquiry. This ideal science is, as its very name sug-

gests, an idealization. Here and now—regardless of the calendar’s date— we can only do the best we can in the cognitive state of the art to estimate the definitively correct answer to our scientific questions. This must suffice us because it is the very best that we can do. We recognize, or at any rate have little alternative but to suppose, that reality exists and that there is such a thing as “the real truth” about the mind-independently real things of this world. But we are not in a position to state any final and definitive claims as to just exactly what it is like. Only at the idealized level of perfected science could we count on securing the real truth about the world that “corresponds to reality,” as the traditional formula has it. Accordingly, here as elsewhere, the gap between the real and the ideal

must be acknowledged. What inquiry provides is “our purported truth,” contradistinguishable as such from “the real truth itself”’—a contrast that can, of course, be made only at the level of theoretical general principles and cannot be implemented in detail. The idea of the definitive truth functions as a guiding conception for us. It characterizes what we ideally aim at rather than what we actually obtain; it guides the direction of inquiry rather than describing its achievements.

Ideal-State Realism as the Only Viable Option There is only one world in existence: the real world as it actually is. But we will not be able to say just what it is really like until the day when natural science has been completed and perfected—which is to say, 3 Peirce repeatedly verges on seeing this point, but his latter-day congeners usually do not and try to get by with wholly transcendental arguments from the possibility of science. See Wilfrid Sellars, Science and Metaphysics: Variations on Kantian Themes (London, 1968).

SCIENCE AND IDEALIZATION

299

never. We must pursue the cognitive enterprise amid the harsh realities and complexities of an imperfect world. The thesis that science truly describes the real world must be looked upon as a matter of intent rather than as an accomplished fact, of aspiration rather than achievement, of

the ideal rather than the real state of things. Scientific realism is tenable only when it is the ideal state of science that is at issue. (That, ex hypothesi, is what makes that state into an ideal one.) But ideal-state realism,

while certainly correct, avails us less than we would like—we who find ourselves emplaced in the suboptimally actual rather than the perfected ideal condition of things. In forming a just appreciation of our scientific claims, the irremovable gap between the real andthe ideal must once again be acknowledged. It is ideal science alone that gets at the definitive truth of things to which authentic reality corresponds. Scientific realism is a viable position only with respect to idealized science, which, as we realize, we do not now have-—regardless of the“now”at issue. A viable scientific realism must therefore turn not on what ouractually developed science takes the world to be like but on what ideal or perfected science takes the world to be like. The thesis that “science describes the real world” must be looked upon as a matter of intent rather than as an accomplished fact, of aspiration rather than achievement, of

the ideal rather than the real state of things. We cannot justifiably be scientific realists—or rather, ironically, we can be so only in an idealistic manner, namely, with respect to an “ideal science” that we can never actually claim to possess. One cannot make the inferential transit from scientific knowledgeclaims to the objective characterization of reality without the mediating premise that these claims are substantially correct. And this mediating premise is simply not available with respect to existing science—science in the present state of the art—but only with respect to ideal science. Only with the idealized supposition of an unrealistic perfection can we unproblematically adopt the stance of a theory realism that holds that the world actually is as theorizing claims it to be. (Recall the discussion in Chap. 12 of the coordination of truth with ideal coherence.) Now as ever, existing science does not embody to perfection the desiderata of cognitive systematization: definitiveness, completeness, unity,

consistency, and so forth. These factors represent an ideal aspiration rather than an evolving reality: a telos or direction rather than a realizable condition of things. Accordingly, there is no warrant for joining Peirce in identifying ideal or perfected science with ultimate science. Perfected science is not something that exists here and how, nor is it something that lies ahead at some eventual offing in the remote future. _ It is not a real thing to be met with in this world. It is an idealization that

7

300

.

REALISM AND IDEALISM

exists outside time—an idealized hypothetical condition that cannot attain actual existence at all but lies outside history as a useful contrast with what we can actually manage to realize. Strange though it may seem, then, it is only at the level of idealization that fully adequate knowledge of reality can be expected: only that idealization presented by perfected science manages “to tell it as it is.” With respect to ideal science we can be realistic, but with respect to real science we must be idealists. The world picture of current natural science certainly need not characterize reality and cannot validly be presumed to do so. Its world—the world it envisions as it envisions it—may well lie only in the minds of the beholders and must, in fact, be presumed to do so. In this regard too, then, our realism has an idealistic mien. But if ideal science is our sort of science projected into ideal conditions, then what assures us that this projection actually captures reality— that such ideal science truly describes the real world? The answer is that this is just how, in the end, we have to understand what is at issue with “reality.” Whatever reality we can contemplate must, in the end, be our reality, reflected in our experience as duly systematized. (Recall the deliberations about possibly variant modes of “science” in chap. 8.) The uscorrelativity of this sort of “internal realism” effectively makes such a po-

sition at issue as a mode of idealism. Metaphysical realism as such need not be a Homo mensura doctrine, but a cognitively committed scientific realism has to pay the price of its gain in definiteness by homo mensura relativization.

The Validity of Idealization The cognitive enterprise is governed by ideals—in particular, those of knowledge/truth and of science/system. But in a community of rational agents, even ideals must pay their way by proving themselves to be efficient and effective in conducing to full realization of the goals and values in whose name they are instituted.

One has to distinguish between aims or purposes on the one hand, and ideals or aspirations on the other. The former are the objectives of conduct, the particular goals toward whose attainment we direct our actions.

They represent states of affairs whose realization we aim at in following out a particular course of action. Ideals or aspirations, by contrast, do not represent states of affairs that we anticipate actually bringing into being. _ We do not really expect to realize them at all. They represent general desiderata rather than particular targets. (For the police chief, “reducing the crime rate by 5 percent” may represent an aim, but the [realistically

SCIENCE AND IDEALIZATION

301

unachievable] desideratum of “eliminating crime” can be no more than an ideal.) Now in cognitive matters at the level of generality and precision at issue with the theories of natural science, conceptions such as “knowledge,” “truth,” “law,” “completeness,” “system,” and the like are all

geared to the ideal level. They are not potential results whose attainment we can (if realistic) see as achievable projects. All of them involve an element of idealization that renders their actualization unrealistic. The impracticability of its realization is thus no insuperable obstacle to the validation of an ideal. This issue of its feasibility or infeasibility is simply beside the point, because what counts with an ideal is not the question of its attainment but that of the benefits that accrue from its pursuit. As Max Weber observed with characteristic perspicuity, even in the domain of politics, that “art of the possible,” “the possible has frequently been attained only through striving for something impossible that lies beyond one’s reach.” To attain the limits of the possibilities inherent in our powersand potentialities, we must aim beyond them. And just here lies the great importance of the ideal realm. Ideals are visionary, unrealistic, and utopian. But by viewing the world in the light of their powerful illumination, we see it all the more vividly—and critically. We understand the true nature of the real better by considering it in the light reflected from ideals, and we use this light to find our way about more satisfactorily in the real world. To be sure, our ideals ask too much of us. We cannot attain perfection in the life of this world—not in the moral life, or in the life of inquiry, or in our economic, social, or religious life. Authentic faith, comprehensive

knowledge, genuine morality, or perfect markets are all idealizations, or destinations that we cannot actually reach. They are hyperboles that beckon us ever onward, whose value lies in their practical utility as a motivating impetus in positive directions. ' On this account, ideals, despite their superior and splendid appearance, are actually of a subordinate status in point of justification. They are not ultimate ends but instrumental means, subservient to the controlling aims and goals whose realization they facilitate. They are indeed important and valuable, but their worth and validity ultimately reside not in their intrinsic desirability (“wouldn’t it be nice if . . .”) but in their 4 “Nicht minder richtig aber ist, dass das Mégliche sehr oft nur dadurch erreicht wurde,

dass man nach dem jenseits seiner Kraft liegenden Unméglichen griff” (Max Weber, “Der Sinn der “Wertfreiheit’ in den soziologischen und ékonomischen Wissenschaften,” Logos 7 [1917-18]: 63; reprinted in Gesammelte Aufsdtze zur Wissenschaftslehre (Tiibingen, 1922), p. 476.

302

.

REALISM AND IDEALISM

eminent utility—in their capacity to guide and to facilitate the cultivation of the objectives of the enterprises to which they are affiliated. But what can possibly justify our adhesion to such unrealistic ideals? The answer is that they represent not goals but values. They orient and guide our action in certain particular directions. And it is in this actionguiding role that their legitimation lies. For their validation is ultimately pragmatic: their justification lies in their utility, their efficiency in promoting the furtherance of the enterprise at issue. Dedicating ourselves to those unrealizable ideals and staging our contact in their light enhances our performance through enabling the realization of results at the level of actual attainment superior to what we would otherwise be able to achieve. Paradoxical though it may seem, the validation of such ideals ultimately lies in the pragmatic domain. After all, perfected science is an idealization—asis the sort ofscientific realism that comes automatically in its wake. Now an ideal is not something we encounter in experience but rather the hypothetical projection or extrapolation of what we encounter in experience. And the legitimacy of our cognitive ideals as regulators inheres in their utility as guides to inquiry, and specifically in their capacity to guide our thoughts and efforts in constructive and productive directions. It is a fallacy to see the validity of goals and ideals as residing solely in the presumed consequence of their realization. Their validation may reside not in arriving but in the benefits we realize in the course of the pursuit itself. The striving after an ideal science that affords us “the ultimate truth” about the workings of nature seems to be a telos ofjust this sort. (We arrive at the perhaps strange-sounding posture of an invocation of practical utility for the validation of an ideal.) The validation of an ideal is thus something derivative. It does not lie simply in the nature of the (unrealizable) state of affairs that it contem-

plates—in the inherently unachievable perfection that it envisions. Rather, it lies in the influence it exerts on the actions of its human exponents through the mediation of thought. The justification and power of an ideal inhere in its capacity to energize and motivate human effort toward positive results—in its practical efficacy, in short. Ideals may involve unrealism, but this in no way annihilates their impetus or value, thanks to the positive consequences that ensue upon our adoption of them. Such an approach to theissue of legitimating ideals has a curious aspect

in its invocation of practical utility for the validation of our ideals. It maintains that the rational appropriateness of our commitment to an ideal lies in its practical utility for our dealings with the real through its capacity to encourage and facilitate our productive efforts. Such an approach does not adopt a Platonic view of ideals that sees them as valuable strictly in

SCIENCE AND IDEALIZATION

303

their own right as abstract embodiments of worth. Rather, their value is seen as instrumental or pragmatic: ideals are of value not for their own sake but for ours, because of the good effects to be achieved by using them as a compass for orienting our thought and action amid the shoals and narrows of a difficult world. The validation and legitimation of ideals thus lie not in their (infeasible) attainability but in their utility for directing our efforts—their productive power in providing direction and structure to our evaluative thought and pragmatic action. It is in this, their power to move the minds that move mountains, that the validation and legitimation of appropriate ideals must ultimately reside.® 5 The themes of this chapter are further developed in the author's books Empirical Inquiry (Totowa, N.J., 1982), The Limits of Science (Berkeley and Los Angeles, 1984), and Scientific Realism (Dordrecht, 1987). The utility of unrealizable ideals is argued in Ethical Idealism (Berkeley and Los Angeles, 1987).

Eighteen Realism and Idealism The General Nature of the Realism-Idealism Conflict Idealism, broadly speaking, is the doctrine that reality is somehow mind correlative or mind coordinated. Realism, broadly speaking, is the doc-

trine that reality is mind independent. Until these doctrines are more sharply formulated and concretized, they need not be contradictory; indeed, both contain a substantial element of truth. Nevertheless, it has been the idealistic side of the debate that has received primary stress in

the present discussion, and it is time to consider the reason why. Idealism centers on the conception that reality as we understand it reflects the workings of mind. And it construes this as meaning that the inquiring mind itself makes a formative contribution, not merely to our understanding of the nature of the real, but even to the resulting character we attribute to it. Bertrand Russell said that “idealists tell us that what appears as matter is really something mental.”! But that is rather stretching the matter. Idealism certainly need not go so far as to say that

mind makes or constitutes matter. Any such view of idealism ignores such versions of the theory as, for example, the explanatory idealism that holds merely that an adequate explanation of the real always requires some recourse to the operations of mind. Fora long time, a dispute raged within the idealist camp over whether the mind at issue in the position’s definition was a mind emplaced outside of or behind nature (absolute idealism), or a nature-pervasive power of rationality of some sort (cosmic idealism), or the collective impersonal social mind of people in géneral (social idealism), or simply the distributive collection of individual minds (personal idealism). Over the years, the more grandiose versions of the theory dropped increasingly from favor, and in recent times virtually all idealists have construed the minds at issue in their theory as a matter of separate individual minds equipped with socially engendered resources and forming part of the world, not

standing outside or behind it.

:

In contemplating the general thrust of idealism, it becomes clear that there is in fact a considerable variety of idealistic doctrines of significantly diverse orientation. The main alternatives at issue are outlined in figure 18.1.

As this survey indicates, realism and idealism are general orientations 1 Bertrand Russell, The Problems of Philosophy (Oxford, 1912), p. 58.

REALISM AND IDEALISM

305

pointing in directions that can lead to diverse destinations. In consequence of their generality, both approaches have many diverse versions, and some forms of either doctrine are perfectly compatible with some forms of the other. In particular, a conceptual idealism maintaining that we standardly understand the real in somehow mind-invoking terms of reference is perfectly compatible with an ontological materialism that holds that the human mind and its operations ultimately root (be it causally or superveniently) in the machinations of physical process. The three positions to the effect that real things just exactly are things as philosophy or as science or as common sense takes them to be—positions generally designated as scholastic, scientific, and naive realism, reFigure 18.1. Versions of Idealism Ontological Versions 1. Causal idealism. Everything there is, apart from minds themselves, arises

causally from the operations of minds. 2. Supervenience idealism. Everything there is, apart from minds themselves, is supervenient upon the operations of minds (that is, somehow inheres

in them, albeit in a way that is not necessarily causal but involves some other mode of existential dependency). Epistemic Versions

1. Fact idealism. To be asafact is to be a language-formulable fact—that is, a truth. Every fact can be semantically captured in a language-formulated

truth. 2. Cognitive idealism. To be asa truth is to be knowable. Every truth can (potentially) be cognitively captured as an item of knowledge. Truth stands coordinate with the cognitive potential of mind.

3. Strong substantival idealism. To be as a thing or entity is to be actually discerned (discriminated, identified, perceived) by some knower. (This is simply a restatement of Bishop Berkeley's idealistic thesis “To be is to be perceived.”)

4. Weak substantival idealism. To be as a thing or entity is to be discernible (discriminable, identifiable, perceivable). Any real thing (entity, object) can (in principle) be discerned by some knower; it must (in principle) be of a nature that admits cognitive access. 5. Explanatory idealism. Any adequate explanatiion of the existence of physical (“material”) reality must proceed in terms of verbal questions; explanatory recourse to mental characteristics or operations is required within the substantive content of the explanation. 6. Conceptual idealism. Any fully adequate descriptive characterization of

the nature of physical (“material”) reality must make reference to mental operations; some recourse to verbal characteristics or operations is required within the substantive content of an adequate account of what it is to be real.

306

REALISM AND IDEALISM

spectively—are in fact versions of epistemic idealism exactly because they see real things as inherently knowable and do not contemplate mind transcendence for the real. Thus, for example, the thesis of naive (com-

monsense) realism that “external things exist exactly as we know them” sounds realistic or idealistic according as one stresses the first three words of the dictum or the last four. It must be noted that the more traditional, ontological mode of idealism has not been at issue in the deliberations of the preceding chapters. Our discussion has focused on those versions of idealism here characterized as epistemic. Let us review how the land lies in this regard.

The Untenability of Fact Idealism The label “fact idealism” involves the doctrine that to be asa fact is to be a language-formulable fact, that is, a truth. It is not difficult to see that

this plausible-sounding thesis is ultimately untenable. For, as was argued at some length in chapter 14, the range of facts outreaches that of truths. Our conception of what is at issue with “genuine facts about the real” is so inclusive and extensive as to outrun the vast but nevertheless limited descriptive resources that language can (even in principle) put at our disposal. The realm of fact is, by its very nature, wider than that of statable

truth. In consequence, fact idealism faces insuperable obstacles: one must reject the idea that all of fact is cognitively tractable. An idealism that

seeks to domesticate fact through its absorption into formulable truth is destined to failure. We must, in the end, reject the appealing idea that language-deploying thought is in principle adequate to reality. And this spells the untenability of an idealism that equates objective fact with lin-

guistically domesticable truth. ~

The Complexities of Cognitive Idealism: Cognitive Idealism as an Idealization Strong substantival idealism—the classical idealist doctrine of Berkeley—

held that reality is mind correlative because “to be is to be perceived.” But if one is not prepared to join Berkeley in pivoting one’s metaphysics on God, and thereby divine intervention for the benefit of one’s theoriz-. ing, then it seems more plausible to hold merely that “to be is to be perceivable.”

REALISM AND IDEALISM

307

Against this background it seems plausible to contemplate a specifically cognitive idealism in terms of the implication principle to the effect that .to be true is to be a potential item of knowledge: p-> PKp

where K represents “is known”

and P is the modality of possibility.

The endorsement of such a thesis at the level of generality affirms the idea that truth stands correlate with the cognitive potential of mind. This sort of cognitive idealism is an ultimately tenable doctrine, although—as the subsequent discussion will show—only in a qualified sense that needs to be formulated with considerable care. ; Clearly, the flat-out identification of truth and knowledge is in diffi-

culty. The strong contention that p — Kp (that any truth is ipso facto known) is obviously untenable with respect to finite knowers: only God is an omniscient knower, not we imperfect mortals. To be “realistic” (in one of the various uses of this term) about science is to acknowledge its incompleteness and imperfection. The terminology of truth (and such of its cousins as “certainty,” “knowledge,” and “system”) cannot be applied categorically and without qualification in the context of the scientific information we actually possess. The appropriate application of such proud labels is not feasible in the context of real science but only in that of ideal science. An element of idealization is inherent in any claim to authentically “scientific knowledge.” The gap between putative and definitive truth is to be bridged not at the level of real knowledge but only at the level of ideal knowledge. All the same, a theory of the potential knowability of truth can be maintained and supported if formulated with sufficient care.

Ramifications of Substantival Idealism Old-style idealism was attracted to the following inference: 1. Reality does (and must) consist of things or substances. 2. A thing must be a unit, a single item (ens et unum convertuntur). 3. The only unity is the unity provided by thought: only:minds can individ-

uate, can carve nature up into discrete items; itemhood is something mind coordinated, its unity being something that lies in the conceptions of a beholder.

Therefore: Minds are the generators of the real; cognizable reality is mindcoordinated reality.

308

REALISM AND IDEALISM

Old-style realism rejected this conclusion because it rejected premise 3. It maintained the objective (i.e., mind-independent) integrity of things. But of course it is hard to deny that “to be (as an individual thing) is to be identifiable (as such).” A somewhat different view of the issue is desirable—and is available by looking a lot more closely at this matter of minds. Realism is a doctrine that holds the existence and nature of reality to be mind independent. But independent of whose mind? One can, in theory, hold that real things are mind independent inthe sense of existing independently of my conception of them; our current conceptions;

anyone’s conception (ever);

their conceivability-in-principle in the circumstances of an ideal knower (i.e., God’s conception of them).

As we move down this list, the position we take becomes increasingly more problematic. With the first thesis we start with a sensible rejection of megalomania. Well and good! But by the time we reach the last thesis, we arrive at something far more questionable—the denial of the idea that identity (individuality) consists in identification (individualability). To accept the existence of things that cannot in principle be conceptualized— things that by their very nature lie outside the range of any and every (hypothetically possible) intelligence—is implausibly strong stuff. Berkeley maintained that “to be (real) is to be perceived” (esse est percipi). And this does not seem particularly plausible because of its inherent commitment to omniscience. It seems more sensible to adopt “to be is to be perceivable” (esse est percipile esse). For Berkeley, of course, this

was a distinction withouta difference: if it is perceivable at all, then God perceives it. But if we forgo philosophical reliance on God, the matter looks different. We are then driven back to the question ofwhat an object of perception is for perceivers who are physically realizable in the real world. And so, something really exists if it is, in principle, experientiable: .. “To be (physically) real is to be actually perceivable by a possible perceiver—one who is physically realizable in the world.” Physical existence is seen as tantamount to observability in principle. The basic idea is that one can claim (legitimately or appropriately) that a particular physical object exists only if there is potential experiential access to it—access not necessarily for us humans but for some experience-capable sort of creature.

On this approach, one would endorse the idea that to be part of phys: ical realityis to be:

REALISM AND IDEALISM

309

not necessarily observed, but observable; not necessarily perceived, but perceivable;

not necessarily experienced, but experientiable.?

If something indeed exists in the world, then it must be observable in principle, detectable by some suitably endowed creature equipped with some suitably powerful technology. To exist (physically) is to be part of the world’s causal commerce—to be at the initiating or receiving end of a causal process that can, in principle, be detected and monitored by an attentive intelligence with sufficiently powerful sensors at its disposal. To exist is to admit of being observable—to be open to experiential confrontation by a cognitive being of some sort. And such observability (as contradistinguished from actual observation) is clearly something objective— a matter of what beings with mind-endowed capacities can encounter in experience, and not with what any particular one or more of them actually does encounter in experience. In this sense, detectability and discriminability in principle is an indispensable request for qualifying as part of the actual furniture of the world. A weak substantival idealism seems perfectly in order.

Explanatory Idealism and the Prospect of Axiological Explanation Explanatory idealism addresses not the constitution but the explanation of the real, and its orientation is not toward nature’s things but toward the explanatory principles and laws that govern nature’s comportment. It holds that an adequate explanation of the nature of material reality requires some recourse to mental characteristics or operations. How can

such a theory be made to work? One particularly intriguing—though certainly controversial—prospect along these lines is offered by the program of axiological explanation. Axiological idealism maintains that values play a causally explanatory role in nature, seeing the real as explicable in terms of value. Doctrines of this sort have been the stock-in-trade of philosophy from the days of Plato (think of Socrates of the Phaedo) to those of Leibniz, with his insistence that the real world must be the evaluative best possible. And this line of thought has recently surfaced once more in the controversial “anthropic principle” espoused by some theoretical physicists. 2 Note that we cannot say, “To be is to be describable” or “To be is to be identifiable,”

since purely hypothetical possibilities can also be described and identified. What they cannot be is experienced by some physically realizable intelligence.

310

REALISM AND IDEALISM

Axiological explanation thus does not prejudge the issue of purpose explanation—let alone reduce to it. The values involved in axiological explanation need not be somebody's values. No element of personalism, no reference to anyone’s aims or purposes, need be involved in axiological explanation. Purpose, on the other hand, necessarily requires a purposer—it must be somebody’s purpose. In this regard, value stands with order rather than with purpose. Order in nature does not presuppose an orderer, nor value a valuer. The maintenance or enhancement ofa value

can be a matter of “blind” operation of impersonal forces. Axiological explanation is therefore not ontologically mind invoking; it need not be grounded in the operations of a mind-endowed being. Rather, it is merely explanatorily mind invoking. It explains (aspects of) nature in terms of categories whose paradigm applications relate to the

dealings of mind. We do not have recourse to mind in providing a causal principle of operation; instead, we use the analogy of mind to provide a principle of explanatory understanding. Axiological idealism accordingly makes an epistemic rather than an ontological version of the theory. It is, perhaps, a somewhat strange doctrine, out of line with the spirit of the times, but this circumstance does not exclude it from the realm of possibility. All the same, there is something rather farfetched about it.*

Toward a Conceptual Idealism There is, however, yet another version of epistemic idealism—distinct

from cognitive idealism and far more tenable than it—that might be called conceptual idealism. This position holds that whatever is real is in principle knowable, and that knowledge involves conceptualization. In particular, human knowledge involves conceptualization that proceeds in terms of reference that reflect our mental dispositions. Such a view has it that we transact our cognitive business in terms of a series of conventions of interpretation and communication that bear the hallmarks of their mind-supplied origins. On its telling, it is not that the real as such is ~

mental but that the real as we know it is standardly interpretedinmindinvoking terms of reference. Conceptual idealism maintains that our knowledge of the real is grasped by us by means not merelyin mind-supplied but indeed even in to some extent mind-patterned terms of reference. Our knowledge of fact always reflects the circumstances of its being a human artifact. It is always formed through the use ofmind-made and indeed mind-invoking concep- © 3 Fora fuller discussion of axiological idealism, see the author’s Riddle ofExistence (Lanham, Md., 1984).

REALISM AND IDEALISM

311

tions, and its contents inevitably bear the traces of its human origins. Whatever we have any knowledge of, we know in terms of mind-construed terms of reference in whose conceptual content there is some reflection of its origin in operations characteristic of mind. At the heart of conceptual idealism lies the important distinction between conceptual mind-involvingness and explicit mind-invokingness, illustrated in. the contrast between a book and a dream. A dream or a worry is something explicity mind invoking. For a mental activity is explicitly at issue when we speak of a dream or a worry: dreaming and worrying are the sorts of things that, by their very nature, belong among the thought processes of mind-endowed creatures. Where there are dreams or worries, there must be mind-equipped ‘beings to do the dreaming and worrying. A book, by contrast, seems at first sight completely nonmental: books,

after all, unlike dreams or worries, are physical objects. If mind-endowed beings were to vanish from the world, dreams and worries would vanish

with them—but not books! Even if there were no mind-endowed beings, there could certainly be naturally evolved booklike objects, things physically indistinguishable from books as we know them. Nevertheless there could not be books in a world where minds had never been in existence. For a book is, by definition, an artifact of a certain purposive (i.e., communicative) sort with pages on which reading material is printed. Such

purposive artifacts all invoke goal-directed processes of a type that can exist only where there are minds. To be a book is to have writing in it, and not just marks. And writing is inherently the sort of thing produced and employed by mind-endowed beings. To explain adequately what a book is, we must thus make reference to writing and, in turn, ultimately to minds. It is not that the book is mind constituted as a physical object, but that to explicate what is involved in characterizing that object as a book, we must eventually refer to minds and their capabilties. The very conception of such things is mind inferential in such a way that the definition of book—or even a minimally adequate explanation of what it is to

be a book—requires a reference to mental processes. A world in which there neither are nor ever have been minds can contain objects physically indistinguishable from our books and nails, but books and nails they could not be, seeing that only artifacts created for a certain sort of purpose can correctly be so characterized. Physical objects are of course not mind invoking as such, but their conceptualization/characterization is nevertheless standardly cast in mind-involving terms of reference. But.what about sticks and stones, atoms and stars, and other such

products of nature rather than artifice? Surely, they stand in the way of conceptual idealism. By no means. To establish mind-involvingness at this level, a rather

312

.

REALISM AND IDEALISM

long and complex story will be unfolded—a story relating to the nature of the conceptions of causality and lawfulness.

Causality as Mind Involving Reality and causality are inseparably coordinated. To be physically real is

to be a part of the world’s causal commerce. And the conception of causality itself is obviously inseparable from that of causal principles or laws. But causal laws are more than mere generalizations. They have hypothetical force and characterize not just what does happen when but what would happen if. The element of causal necessity and counterfactual force

is inherent in their nature, and they encompass not only the actual but the hypothetically possible as well. But this realm of mere possibility— of what is not present as part of nature’s realities but accessible only via supposition, assumption, and hypothesis—is part and parcel of the realm of mind involving. For hypothesizing and its cognates (supposing, assuming, etc.) are one and all things that mind-endowed beings alone can do. Reality as such is simply factual and actualistic. Figuratively put, it talks the language of is and knows nothing of maybe. It contains no mere possibilities and admits of no element of:the merely hypothetical (over and above comprising mind-endowed beings who can project hypotheses). The only access route to the purely possible is that of supposition, assumption, and hypothesis—all of which call for mind-endowed beings to come upon the scene. One salient argument for the mind-involvingness of mere possibilities runs as follows: To explain what a mere possibility is, one has to contrast it with a real possibil-

ity. A real possibility is one that can come to be realized through developed processes producing changes in real things (the possible oak tree that may emerge from this acorn). A mere possibility is one that cannot come about (be ~ realized) through the machinations of the real; it can only be assumed (supposed, hypothesized). The concept of a mere possibility is accordingly mind invoking in the standard, explanatory ordered sense of that term. Possibilities as such can be either real or mere; it is the mereness of a mere possibility that is conceptually mind invoking. a

To explicate what it is to be (merely) possible, we thus need to involve the operation of minds. For the merely possible is not a part of the world’s furnishings; it is something that can only be imagined, supposed, conjectured, hypothesized: that is, it is by its very nature something

REALISM AND IDEALISM

313

whose status is correlative with something that only minds can do.* The concept of the merely possible is accordingly an instance of a mind-invoking conception. Its adequate explanation requires a reference to the sort of thing that only minds can do. Real possibilities are possibilities de re: “It is possible for Felix the cat to be on the mat.” Hypothetical possibilities are de dicto possibilities: “It is possible for a cat—not necessarily one of the actual ones—to be on the mat.” Their status and condition in the explanatory nature of things is based on assumptions, suppositions. To explain adequately what mere possibility: (in contrast to that of a real possibility) is all about, we must bring mental operations (assuming, supposing, hypothesizing) into the discussion. In view of this, hypothetical possibilities are inherently mind involving. And this has important consequences for laws. To be sure, mere regularity calls for no more than a universal generalization of the type: (1) Al X is Y.

But genuine lawfulness goes well beyond this to call for stipulations of the form: (2) AUX is Y, andifany z that is not an X were an X, then z would bea Y.

Here (1) is simply an issue of existential fact: if it is in fact the case that “all (pure) mercury solidifies at —39° C,” then this circumstance could continue operative even if one abstracted from all reference to minds, overt and covert alike. But a type (2) generalization—a lawful generalization—would fare differently. Its reference to things that are not Xs having to have certain characteristics if they were Xs involves claims outside the domain of ontological fact, claims of which sense can be made only by the use of mind-invoking conceptions of possibility. The mind’s ability to use its concept apparatus to project the specification of entities that, as it fully recognizes, have no standing in the world’s scheme of things is one of its characteristic capacities—one that is crucial to operate the conception of lawfulness as we have it. The implicit projection of merely hypothetical possibilities is crucial to lawfulness and represents an element of intellectual artifice that is among the most obviously self-provided constructions the human mind affords us. The position at issue here is not the old Parmenides/Plato doctrine that even nonexistents must “exist” in some sense because they can be talked about (assumed, supposed, etc.). Our theory does not propose to estab4 The point is not that only the merely possible can be supposed (assumed, conjectured, etc.); it is that the merely possible can only be supposed—it cannot be positioned in reality’s _ scheme of things in a mind-independent way.

314

.

REALISM AND IDEALISM

lish the link between fictionality and mind-dependency by using existence as a middle term as per the argument: Fictional objects must have an existence of some sort.

The only sort of existents that fictional objects can have is a mind-correlative one. Therefore: Fictional objects have a mind-correlation existence.

This sort of ontological argumentation just is not at issue in the conceptual idealism envisioned here. For the fact is that the linkage between a fictionality (i.e., assumptive or suppositional or hypothetical status) and mind-involvingness is not enthymematically mediated but an immediate one, because fictionality (and its assurative cognates) are inherently mind invoking in their conceptual makeup. The point is that to make proper use of “merely possible objects,” we must characterize them as fictions, thereby characterizing them in mind-referring terms of reference.

But does not conceptual idealism’s thesis that hypothetical possibilities are mind invoking not run into difficulty because ofa failure to distinguish between what is supposed (namely, the hypothetical possibility in question) and the supposing of it—this latter alone being mind invoking. Are we not committing with reference to supposition the same old error that Berkeley committed with reference to perception, in refusing to distinguish what is perceived with the perceiving of it? By no means! The cases are quite disanalogous. With existents we are (ex hypothesi) at liberty to distinguish the thing at issue from the perceiving (contemplating, etc.) of it. Precisely because it exists, it is (ex hypothesi) there to be distinguished. But with nonexistents this prospect is all too plainly absent. Berkeley was wrong: with existents we cannot maintain that to be is to be perceivable (let alone perceived), simply because there can or might be actual things existing entirely outside our perceptual reach (gravity waves, for example). But with nonexistents the case is very different; here to be is to be exactly as conceived because the item itself, being nonexistent, is no more (but also no less) than an object’ ~ of supposition and is constituted as the. thing it is through the assuming or supposing of it. In this domain, to be is indeed to be conceivable. _ To make the transit from the mind involvement of mere possibility to that of laws, we have to note that laws encompass the realm of possibility

at large. There are three distinct sorts of possibilities: 1. real potentialities of actually existing things (if circumstances are favorable, yon acorn will grow into an oak tree);

2. fictive possibilities for actually existing things (if yon acorn were a cat, it would have a backbone);

REALISM AND IDEALISM

315

3. mere possibilities (if a bolt of lightning hit here—not a repositioning of

any of the actual ones, but a somehow additional one—then a thunderclap would sound here).

But laws of nature apply to the sphere of possibility across the board, as the case-(3) example shows. And it is crucial to their status as genuine laws that their application outruns the limits of the real into the realm of the merely hypothetical. If a supposed law did not admit type-(3) implementation, we would have to reject it as such. If it is indeed a law of

nature that mice have tails, then this will still hold good when I imagine (assume, suppose) that there is a mouse in my pocket. For if it is indeed a (nonmutilated) mouse that is there, thena tail has to be there as well. But to explain and expound this absolutely essential feature of laws (namely, that they support counterfactuals relating to mere possibilities), we will have to bring assumptions (suppositions) into it—that is, we will have to make use of conceptions that relate quite explicitly to the operating of minds. Again, the thesis “Oak trees are (lawfully) deciduous” (i.e., “Oak trees have to be deciduous” in a sense that warrants “If that pine were an oak, then it would be deciduous”) does not deal with actu-

ality alone but frames its claims about the facts in such a way as to bring in a realm of the hypothetically possible that, by the very nature of the hypothetical, is inherently mind involving. But as another critic complains,® does not a law simply characterize what does happen in this world, irrespective of any reference to mere possibilities? The answer is neither yes nor no. A law of nature indeed characterizes what does happen in the world, but it does so in terms of what must happen in the world and thus does not effect its characterization without reference to possibilities. Laws are—and qua laws have to

be—more than mere generalizations. For the crucial point is that laws support counterfactuals; they hold not only from the real things of this world but also from the mere possibilities that we project into it by means of suppositions, hypotheses, and assumptions. Lawfulness is the product of the well-founded imputation to empirical generalizations of nomic necessity and hypothetical force. Both of these two factors—the essentially factual element of well-foundedness and the essentially decisional element of imputation—are necessary to laws. Well-foundedness is essential because the very spirit of the scientific enterprise demands reliance only upon tested generalizations that have a solid observational or experimental basis. But the element of imputation is also essential, since, as emphasized above, we can observe only what is, that is, forms parts of the realm of the actual, and not what corre5 Baruch Brody, “[Review of] Conceptual Idealism,” Philosophical Review 84 (1975): p. 581.

316

.

REALISM AND IDEALISM

sponds to the modally necessary or the hypothetically possible. The nomic necessity and hypothetical force characteristic of lawfulness thus represent factors that a generalization cannot conceivably earn for itself on the basis of observational or experimental evidence alone: it has to be

endowed with these factors. The basic aspect of a lawful proposition is on this view not the qualitative aspect of being-a-law but the relational aspect of being-maintained-as-a-law. Lawfulness is in the final analysis a relational rather than an absolute and purely descriptive feature, with minds playing a key part in the relationship at issue, although—to be

sure—a part only. For lawfulness involves possibility, and the gulf between the actual and the merely possible is one that can only be bridged by minds.®

Thereis a crucial difference betweena claim’s being a descriptivetruth about the real world as such and its being of the sort that—like a law—in its very meaning and purport goes beyond the real world. Descriptive

statements about the real are embodied in assertions that do indeed appertain only to the real world. But statements about what is necessary or what is (merely) possible have to be construed in terms of reference to the domain of mere possibility as well. That something is necessary (holds in all [physically] possible worlds) or that it is merely possible (holdsin some other [physically] possible world) are claims that, by their

very nature, transcend the descriptive characterization of the real. The truth conditions of such facts do not turn simply on the descriptive constitution of the actual world but invoke reference to the situation in other

possible worlds as well—worlds to which our only mode of entry lies (in the very nature of the situation) only through the route of supposition, assumption, and hypothesis. For it is, clearly, only through thought— through the imagination as deployed in assumptions, suppositions, and hypotheses—that we can get cognitive access to other possible worlds. It is necessary to face one rather immediate consequence of the doctrine of the mind involvement of lawfulness: namely, that causality will stand on exactly the same footing. To speak of causation is to invoke (however tacitly) the operation of causal laws. (That heating the water caused. -it to boil can be maintained only on a basis of some thesis of the general

type of the causal law “Water boils when heated to 100° C—at sea-level altitudes.”) That any discussion of causation in the scientific context is coordinate with an at least implicit reference to causal laws is a part of general agreement among present-day philosophers of science. The very idea of a causal connection between events is deployed in a generalized 6 While laws are indeed made by people, they are not thereby made as people want them to be. In maintaining that the necessity and hypotheticality of lawfulness are matters of imputation, we thus have no wish to suggest that the issue is one of indifferent conventions or arbitrary decisions.

REALISM AND IDEALISM

317

way that envisages their relationship as governed by a causal law. In this way, the argument inherent in our analysis of lawfulness as conceptually mind dependent leads inexorably to an acceptance of the traditional idealistic view that causality is mind dependent. Thus relative to our standard conceptual scheme, the mentalistic resource of counterfact is required, via lawfulness, for our conceptualized apprehension of the realm of causally interrelated fact. In somewhat Kantian terms, our experience of the real as a network of causality involves and presupposes a wider realm governed by the category of possibility. To consider a relationship among actual occurrences within the space-time framework in terms of a causal linkage is to regard it in a setting of powers and potentialities that inevitably ramify into the fact-transcending and mind-involving sphere of “what would happen if.” If lawfulness is conceptually mind involving, then so is causality. . These deliberations, then, enable us to address the “sticks and stones”

objection raised above (at the end of the section “Toward a Conceptual Idealism”). For a thing is what it does. To be a stick or a stone (or an atom or a star) is to be something that behaves in a certain kind of way, one that conforms to certain characteristic laws. And if such laws are conceptually mind involving, then so are the sticks and stones for whose

specification their invocation is indispensable. Conceptual Idealism Elaborated It cannot be emphasized sufficiently that such a conceptual idealism does not take the form of spiritualism or panpsychism; it need not and does not maintain that only minds exist and that all there is is somehow the work of mind in any productive or causative sense. Rather, this idealist position sees mental operations as providing a model for the categories in terms of which we think about real things—as not only devising these categories but doing so in mind-referring terms. The position implements in its own, conceptualistic way the idealistic thesis that mind contributes essentially to the constitution—as well as the constituting—of our knowledge of reality by characterizing things in its own terms, in using inherently mentalistic conceptions and paradigms. Conceptual idealism’s thesis is not the trivial one that mind makes the idea of nature. It is not open to Santayana’s complaint against Schopen-

hauer that “he proclaimed that the world was his idea, but meant only (what is undeniable) that his idea of the world was his idea.” Rather, what is at issue is that mind-patterned conceptions are built into our idea of nature: that the way we conceive of nature is in part patterned on our mind. The pivotal contention is that our world model, the world as we

318

.

REALISM AND IDEALISM

depict it in thought and cognition, is a mental artifact that is constructed (in part) in niind-referential terms—that the nature of the world as we conceive it reflects the nature of mind. Conceptual idealism holds that mind is cognitively “at home” in the world in such a way that the nature of the world is something it rationalizes for itself in its own terms. Conceptual idealism sees mind not as causal source of the furnishings of nature but as the basis of the interpretative archetypes in whose terms we understand them. The position has it that we come to cognitive terms with nature on our own terms—that is, in terms of concepts for which

our mental operations themselves are paradigmatic. Conceptual idealism is predicated on a view that reality as we standardly conceive it is in some degree modeled on the paradigm of mind. As this sort of idealism has it, the mind understands nature in a manner that in some ways reflects its own nature.

We must heed the crucial distinction between the conceptual order, with its essentially hermeneutic perspective upon the intellectual-explication of meanings, and the causal order, with its productive perspective upon the causal explanation of physical processes. The issue is not one of causal explanation at all but one of the understanding to be achieved

through an analysis of the internal meaning-content of concepts and of the semantic information conveyed by statements in which they are operative. Because of the fundamental difference between these two perspectives, any conflict in the dependency relations to which they give rise is altogether harmless from the standpoint of actual inconsistency. For once all the due distinctions are duly heeded, any semblance of con_ flict between the two opposed modes of priority disappears. (No doubt, this calls for a certain amount of care and subtlety—but then so do many

issues of intellectual life, and why should things be easier in philosophy than elsewhere?) The basic point is that altogether different sorts of dependencies or requirements are operative in the two theses: 1. that mind is causally dependent upon (i.e., causally requires) matter, in _ that mental process demands causally or productively the physical workings of matter;

:

2. that matter (conceived of in the standard manner of material substance subject to physical law) is explicatively dependent upon (i.e., conceptually requires) mind, in that the conception of material processes involves hermeneutically or semantically the mentalistic workings of mind.

The conceptual idealist’s thesis that one specific direction of dependence (namely, that of the physical upon that of the mental) is built into the

view of reality at issue in our standard conceptual scheme for thinking about reality must not be seen as conflicting with the debatable (but by .

REALISM AND IDEALISM

319

no means thereby negligible) prospect that the order of scientific explanation might lead to positing a reversal in the direction of dependence. When different perspectives are involved, seemingly conflicting theses are perfectly compatible. (I can say without conflict that my car is economical in point of gas mileage and uneconomical in point of maintenance costs.) The conceptualistic idealist sees mind as an explicative paradigm for our understanding of the real, rather than as a productive source in the causal order of its genetic explanation. This position does not see mind as the causal source of reality, but as making a formative contribution to the shape—as well as the shaping—of its characterization of the real. The conceptual idealist has no vested interest in arguing that mind and its functioning are not somehow causally emergent from the processes of .-matter. The point is simply that our standard rational account of the world—its sector of physical particulars specifically included—is given in terms of reference that are at bottom mind-involving. It is the analytical issue of how we think of the world, not the explanatory issue of its causal goings-on, that paves the road to mind dependence. This position does not need to be argued as an attack upon causal materialism: it is quite compatible with the idea that mental functioning has its material basis and causal origins in the realm of physical process. On conceptual idealism’s telling, there is something indelibly anthropomorphic (human patterned) or, more accurately, noomorphic (mindpatterned) in the way in which we think about the world. “But,” someone may say “this does not show anything about the world, but rather about us.” However, this objection loses the point of conceptual idealism, which is exactly that there is no clean separation between the world and the domain of thought, because our only possible route to cognitive contact with the world is through mediation of our conceptions about it, so that for us, “the world” is inevitably “the world as we can manage to conceive of it.” In consequence, how we conceive of the world (namely, in noomorphic terms) has to be seen.as a fact not just about us but about “the world” as well.

Objections to Idealism Even if one grants the preceding point and concedes that a materialist theory of mind is not inconsistent with an idealistic position, would its adoption not generate a vicious circle of the form in figure 18.2? The answer to this question is negative. For as the diagram itself indicates, the circle is not vicious, since different modes of dependency are involved in the noncausal versions of the doctrine. With conceptual ideal-

320

REALISM AND IDEALISM

Figure 18.2: Mind and Matter in Conceptual Idealism causal dependence

MIND

MATTER

conceptual dependence ism, for example, the circle is harmless because it breaks in two, with

matter hinging on mind in the conceptual order of understanding (of rationes cognoscendi, or rather concipiendi), and mind hinging on matter in the explanatory order of causation (rationes essendi). But even if no vicious circle arises, do we not arrive at an equally vicious infinite regress that altogether precludes understanding—that is, if an adequate understanding of mind requires reference to its causal origins in matter, and an adequate understanding of matter requires reference to its functional presuppositions of a mind-invoking sort? The an-

swer is no—no vicious regress is involved. Such a regress would arise only if one adopted an essentially linear model of understanding. But this is quite inappropriate in the case of coordinated concepts such as the present instance of mind/matter or the simpler case of cause/effect. To say that we cannot fully understand the cause until we understand its effect, and that we cannot fully understand the effect until we understand the cause, is not to show that there is a vitiating regress with the result

that we cannot understand either one. All it shows is that two such coordinated and interrelated concepts cannot be set out through a sequential explanation but must be grasped together in their systematic unity. An analogy may help. Take a knife and its blade.Ifthat object is to count as a knife, then that shiny thing attached to the handle must be a blade,. -but this thing cannot count as a blade unless the whole it composes to-

gether with that handle is a knife. The two items stand in conceptually symbiotic union: X cannot be properly characterized as X unless it is duly related to Y, and Y cannot be properly characterized as such unless it is duly related to X. We cannot pick up either end of the stick in separation from the other but must grasp the whole in one fell swoop. Just such a cognitive coordination of mentalistic and materialistic concepts holds with respect to our present analysis of their mutual interdependencies. By maintaining a sufficient set of due distinctions, any collapse into vicious circularity or vitiating regress can be avoided.

REALISM AND IDEALISM

321

Over the years, many objections to idealism have been advanced. Samuel Johnson thought to refute Berkeley’s phenomenalism by kicking a stone. He conveniently forgot that Berkeley's theory goes to great lengths to provide for stones—even to the point of invoking the aid of God on their behalf. On realism’s behalf, G. E. Moore pointed to the human hand as an undeniably mind-external material object. He overlooked that, gesticulate as he would, he would do no more than induce

people to accept the presence of a hand on the basis of the hand orientation of their experience. C. S. Peirce’s “Harvard Experiment” of letting go of a stone held aloft was supposed to establish scholastic realism because his audience could not control their expectation of the stone’s falling to earth. But an uncontrollable expectation is still an expectation, and the realism at issue is no more thanarealistic thought-posture. Immanuel Kant’s famous “Refutation of Idealism” argues that our conception of ourselves as mind-endowed beings presupposes material objects because we view our mind-endowed selves as existing in an objective temporal order, and such an order requires the existence of periodic physical processes (clocks, pendula, planetary regularities) for its establishment. At most, however, this argumentation succeeds in showing that

such physical processes have to be assumed by minds, the issue of their actual mind-independent existence remaining unaddressed. (Kantian realism is an intraexperiential “empirical” realism.) It is often said that idealism is predicated on a confusion of objects with our knowledge of them and conflates the real with our thought about it. But this charge misses the point. The only reality with which we inquirers can have any cognitive commerce is reality as we conceive it to be. Our only information about reality is via the operations of mind—our only cognitive access to reality is through the mediation of mind-devised models of it. Perhaps the most common objection to idealism turns on the supposed mind-independence of the real. “Surely,” so runs the objection, “things in nature would remain substantially unchanged if there were no minds.” This is perfectly plausible in one sense, namely the causal one—which is why causal idealism has its problems. But it is certainly not true conceptually. The objection’s exponent has to face the question of specifying just exactly what it is that would remain the same. “Surely roses would smell just as sweet in a mind-denuded world!” Well . . . yes and no. Agreed— the absence of minds would not change roses. But even the size of roses—let alone their fragrance and sweetness represents a factor whose determination hinges on such mental operations as are marked in measuring. _Mind-requiring processes are required for something in the world to be discriminated as being a rose and determined as being the bearer of certain features. Identification, classification, and property at-

322

REALISM AND IDEALISM

tribution are all required and by their very nature are all mental operations. To be sure, the role of mind is here hypothetical. (“If certain interactions with duly constituted observers took place, then certain

outcomes would be noted.”) But the fact remains that nothing could be discriminated or characterized as a rose in a context where all reference to performing suitable mental operations (measuring, smelling, etc.) is not presupposed. But what of an identity theory of mind that identifies mental processes with the operation of certain material configurations, namely, brains? Is our conceptual idealism not incompatible with such a theory? Not necessarily. It depends upon whether the identity is seen as a factual one (like the identity of the morning star with the evening star or thatof the tallest man in the room with the poorest man in the room), or as a con-

ceptually necessary one (like that of Smith’s only brother with Smith’s only male sibling). Our idealism will encounter no difficulties with a thesis of contingent identity. (Why, after all, should factual relationships not be superimposable upon conceptual ones?) An incompatibility will arise only if the identity theory of the mental with the material is taken to obtain in conceptual terms, as committed to the thesis that mentalistic

talk is eliminable, in that it can be translated without conceptually viable residue into talk about the behavior of matter. Such a conceptually eliminative reductionism is indeed incompatible with a conceptual idealism. For if “mentalese” were analytically altogether reducible to materialistic discourse, then mind could not be conceptually basic to matter in the

sense of our theory. But of course since this theory is based on an analysis of the ordinary conceptual scheme, this goes no further than to show that this ordinary scheme is incompatible with a conceptually reductive ma-

terialism. And this upshot is surely not surprising. (If we point out to reductive materialists that they violate the conditions of ordinary concep-

tual scheme, thereby effectively abandoning this scheme, they may well reply that they are only too ready to do so. In taking this stance they are, to be sure, not inconsistent, but they do cut themselves off from participation in the discussions that take place within the mentalistic framework - -of our standard conceptual scheme.)

Retrospect It is clear in the light ofthe preceding deliberations that the doctrinal stance of the present book has a mixed or mediative character. It is indeed arealism, since it acknowledges a realm of ontological mind-independent existence. But it is also a conceptual idealism that holds that this existence can and must be explicated in mind-referring terms of reference.

REALISM AND IDEALISM

323

Consider once more the various versions of epistemic idealism enumerated at this chapter's outset: (1) fact idealism, (2) cognitive idealism, (3) strong substantival idealism, (4) weak substantival idealism, (5) explanatory idealism, and (6) conceptual idealism. As regards these variant idealist doctrines, the position of the present deliberations stands as follows: 1. Fact idealism is untenable. Given what is at issue with facts (as features of the real) and truths (as linguistic formulations of fact), there is good reason

to deny that every fact can be linguistically accommodated as a truth. 2. Cognitive idealism is tenable. It is plausible to take the stance that “whatever is true is in principle knowable,” provided this thesis is construed in a sufficiently cautious, guardedly formulated way. . 3. Strong substantival idealism is clearly untenable. With respect to finite knowers, at any rate, it is inappropriate to say that all that is real is known to be so. ‘

4, Weak substantival idealism is tenable and indeed plausible. It seems perfectly sensible to take the Quinean line that “no entity is without identity.” To be an individual is to be in principle identifiable as such. But this version of causal idealism (as it might be called) is a very weak doctrine, given the gen-

‘erous scope of the idea of in-principle identifiability. 5. Explanatory idealism is an interesting, albeit deeply problematic pros-

pect. The idea that one must resort to minds in the causal explanation of nature’s physical processes is one that—at this stage of our understanding of how

things work in the world—is rather farfetched. 6. Conceptual idealism is plausible and indeed appropriate. Our knowledge of the world is—within the setting of our standard conceptual scheme for thinking about the real—a knowledge of it in our own, characteristically human terms of reference.

Our overall position with respect to the idealism-realism controversy is accordingly a mixed one. For while our deliberations have rejected some of the stronger versions of idealism as implausibly overambitious, they endorse and support some of the less ambitious, more defensibly constituted versions of idealist doctrine. Our position is a realism, since it acknowledges a realm of ontologically mind-independent existence. But in accepting a cognizability-in-principle standard of real/true/actual, it is also a cognitive idealism that holds that existence is in principle knowable—although known by us in terms of reference of a characteristically human orientation. What seenis right about idealism is inherent in the fact that in investigating the real, we are clearly constrained to use our own concepts to address our own issues—that we can learn about the real only in our own terms of reference. But what seems right about realism is that the an_ swers to the questions we put to the real are provided by reality itself—

324

.

REALISM AND IDEALISM

whatever the answers may be, they are substantially what they are because it is reality itself that determines them to be that way. Mind proposes, but reality disposes. But of course, insofar as one can learn about this reality, it has to be done in terms accessible to minds: our only access to information about what the real is through the mediation of mind. The salient point to emerge from such deliberations is that the realismidealism debate is rendered complex through the great variety of realisms and opposing idealisms that can be contemplated. When one considers the controversy in detached perspective, one is led to the recognition that there is no prospect of a one-sided victory here. The sensible move is to opt for the middle ground and to combine a plausible version of realism with a plausible version of idealism. The issue is not one of the dichotomous choice of either realism or idealism but rather one of a compromising synthesis in the interests of a fruitful collaboration between these historically warring positions.”

Postscript: Proposition Cognizability versus Entity Cognizability To deliberate sensibly about realism/idealism, it is necessary to note that the lay of the land is very different as regards propositions (truths) on the one hand, and entities (objects) on the other.

Approaching the situation from the propositional point of view, let us begin by acknowledging the following two, generally accepted knowledge-characterizing theses: (K1) Whatever is known must in fact be the case: Kp > p (Note: We here use “>” for material implication). (K2) Wherever a conjunction is known, so are its conjuncts:

K(p & q) > (Kp & Kq).

On the supposition that we are dealing not with God but with finite, imperfect knowers, we clearly want to reject the contention that all truths are known. But if (Wp)(p > Kp) is false, that is, if ~(Wp)(p > Kp),. — then obviously (Ap)(p & ~Kp) is true. Accordingly, let it now be supposed 7 For the history of idealism, see Otto Willmann, Geschichte des Idealismus, 3 vols. (Leipzig, 1894-97; 2d ed., 1907). For the German tradition, ‘see M. Krénenberg, Geschichte des deutschen Idealismus, 2 vols. (Munich, 1909-12), and Nicolai Hartmann Die

Philosophie des deutschen Idealismus, 2 vols. (Berlin 1923-29.) For British idealism, see J. Pucelle, L’Idéalisme en Angleterre de Coleridge a Bradley (Paris, 1955), and A. C.

Ewing, Idealism: A Critical Survey (London, 1934), as well as his collection of texts The Idealist Tradition (Glencoe, Il., 1957). Contemporary defenses of idealist doctrines are _ presented in N. Rescher, Conceptual Idealism (Oxford, 1973), John Foster, The Case for

Idealism (London, 1982), and Timothy Sprigge, The Vindication ofAbsolute Idealism (Ed-_ inburgh, 1983).

REALISM AND IDEALISM

325

that there is a particular value of p, say po, for which this holds, so that Po & ~Kpo.

Now, in virtue of (K1) and (K2), we already have it in general that (Vp) ~PK (p & ~Kp) or, equivalently, ~(Ap)PK (p & ~Kp), where ‘P’ represents the modality ofpossibility. This emerges from the following argument: 1. Suppose that: (Ap)PK(p & ~Kp). 2. (Ap)P(Kp & K ~Kp) from (1) by (K2). 3. (Ap)P(Kp & ~Kp) from (2) by (K1).

Since (8) is self-contradictory, its negation is established. And that of (1) follows in its wake. QED Now, combining the results of the preceding two paragraphs, we have: (py & ~Kpo) & ~PK (po & ~Kpy).

And this at once yields

(Ap)(p & ~PKp), since the proposition py) & ~Kypy itself represents a specific value of ‘p’ for which this thesis obtains. What does this rather striking finding mean? It means that if we want to reject (for finite knowers) the clearly untenable thesis p > Kp, then we must accept not just (Ap)(p & ~Kp) but even (Ap)(p & ~PKp). With finite knowers, there are not just unknowns but even unknowable truths. For if there indeed is an item of fact that is unknown—that is, if py & ~Kpy— then this particular thesis is one that cannot possibly be known, seeing that K(p) & ~Kp,) leads ad absurdum by way of the above demonstration. Thus, if we do want to uphold the thesis that “all facts are in principle knowable,” while yet rejecting the farfetched contention that ail facts are known—thereby endorsing ~(Wp)(p > Kp)—then we can maintain the aforementioned thesis in only the qualified form:

(W*p)(p > PKp), where V* represents a limitedly general (propositional) quantifier that confines its scope to values of the propositional variable in which the K-operator itself does not occur. That is, we must limit our purview of fact to propositions that deal with matters of ground-level fact, excluding higher-level facts about what is or is not known. To maintain consistently that “all truths are knowable” in the context of finite knowers to whom some truths are actually unknown, we must exclude certain cognitive

truths about what knowers do or do not know from the range that we take into view speaking of “all truths.” The preceding deliberations accordingly show that due caution must be used in the articulation of a propositional idealism holding that “to be

326

;

REALISM AND IDEALISM

(as a fact or truth) is to be knowable,” seeing that in its incautious formulation (Vp)(p > PKp), this thesis is inconsistent with the clearly appropriate contention ~(Vp)(p > Kp). But when due caution is experienced in its formulation, then the logical difficulties lying in the way of such an idealism can be sidelined. For this is no reason of fundamental fact why an idealistically minded theoretician cannot maintain, “All truths are, in principle, knowable” in the restricted, V*-correlative manner of “all

truth.”

,

It is, however, critically important in the context of our present delib-

erations to observe that this restriction imposes no hardship on the conceptual idealist who wants to maintain the mind correlativity of truth. For this position can still be supported by a two-pronged attack: 1. With respect to noncognitive truths, one can unproblematically maintain, “All (such) truths are in principle knowable”: (W*p)(p > PKp). 2. With respect to the remaining cognitive truths, the conceptual idealist can take comfort in the consideration that such truths will—by their very nature—wear their mind involvement on their sleeves. It is, moreover, important to note that the preceding line of reflection

does not mean that analogous qualifications must be adopted with a substantival idealism maintaining that “to-be (as a thing or entity) is to be knowable (identifiable, perceivable).” For our earlier qualification-indicating argumentation hinges in an absolutely crucial way on the fact that any and all propositional claims regarding knowledge are themselves also candidates for being known. But item knowledge of the form Dx (“x is a discriminated, cognitively identified item”) is something very different from propositional knowledge of the form Kp (“p is a known, cognitively ascertained fact”). The propositional substitution step Kp/p, which leads from Kp — p to such derivatives as KKp — Kp, has no appropriate analogue for nonpropositional objects, because Dx simply fails to fall into the range over which x itself ranges. With propositional knowledge, knowledge claims remain homogeneously at the level of propositions, but with © ~ thing-knowledge, knowledge claims are ‘no longer at the level of things. The (K1)-analogous thesis Dx > x is in dire difficulty.. The argumentation thus at once breaks down when transposed from propositions to things. Again, with thing-knowledge (unlike propositional knowledge), we have no analogue to (K2) of the form 7 D(x & y) — (Dx & Dy)

For clearly, the operator “&” ofpropositional conjunction cannot function between things. Precisely because x-items and Dx-items belong to differ-

REALISM AND IDEALISM

327

ent categories (are heterogeneous), our earlier argumentation cannot be carried over.®

These considerations indicate that the demonstration establishing the need to restrict (Vp)(p > PKp) falls apart when we shift from a propositionoriented approach to an entity-oriented approach. There is no need to qualify (Vx)(E!x > PDx). Accordingly, the circumstance that we cannot maintain an unqualified claim that to be true = to be knowable as being a true fact

does not mean that we cannot maintain an unqualified claim that to be real = to be identifiable as being an existing thing.

The need for internal subtleties in a propositional idealism of the knowability of truth does not in any way countervail against the tenability of a straightforward ontological or entity idealism. Romance languages have an inbuilt distinction between a fact-possessing information-knowledge (Lat.: scire) and acquaintance-involving item-knowledge (cognoscere). (Think of the French savoir/connaitre or the Spanish saber/conocer.) Our present deliberations indicate that this distinction has significant philosophical ramifications and implications.® 8 Observe that the preceding remark holds with unabated force even if one proposed to propositionalize D via some such definition as one based on the idea that to discriminate a thing is to know of its existence: D = KE! For then both DDx and D(x & y) both still continue to involve category mistakes. ® The themes of this chapter are further developed in the author's Scientific Realism (Dordrecht, 1987); Conceptual Idealism is also relevant.

Name Index

Alston, William P., 255n Amaodi, Edoardo, 87n Ambrose, Alice, 149, 149n Anaximander of Miletus, 192, 193, 252 Anderson, Paul, 118n Aristotle, 15, 132 Asch, Solomon, 196n, 197 Badash, Lawrence, 84n Ball, John A., 123n Becquerel, Henri, 247 Bennett, Jonathan, 10n, 202n Bergson, Henri, 232n Berkeley, Bishop, 305-8, 314, 321 Black, Max, 132n

Blanshard, Brand, 170, 170n, 171, 209n Bohm, David, 64, 64n, 65 Bohr, Niels, 291 Bonjour, Laurence, 225n, 228n Bradley, F. H., 85, 137, 137n, 138n, 142n, 145, 159, 171, 171n, 253, 253n Braithwaite, R. B., 149n Brandom, Robert, 276n Brody, Baruch, 315n Bromley, D. A., 88n Caesar, Julius, 292 Campbell, D. T., 236n Cardano, Girolamo, 236, 237 Carnap, Rudolf, 56, 56n Cartwright, Nancy, 276n Chisholm, R. M., 164n Cicero, 53n Darwin, Charles, 3, 3n, 235, 235n Davidson, Donald, 244n Dawkins, Richard, 38n

Day, J. P., 146n

de Beer, Gavin, 237n Descartes, Reni, 59, 273, 273n

Dewey, John, 232n Diamond, Jared, 41n Drachmann, A. G., 91, 91n Dummett, Michael, 280n, 282 _ Dyson, Freeman, 103n

Ehrensvaerd, Goesta, 1l4n Einstein, Albert, 96, 97n, 106, 292 Empiricus, Sextus, 12, 12n, 27, 33

Ewing, A. C., 171in, 174n, 324n Fecher, V. J., 248n Feyerabend, P. K., 21, 48 Feynman, Richard, 64n Fichte, J. G., 186, 297 Foster, John, 324n Foster, Michael, 49n, 50n Galilei, Galileo, 103, 103n Glass, Bentley, 63, 63n Gédel, Kurt, 246 Goudge, T. A., 235n Grene, Marjorie, 235n

Grice, H. P., 202, 202n Haldane, J.B.S., 42n, 52 Harré, Rom, 278n Hartmann, Nicolai, 324n Harvey, William, 247 Hawking, S. W., 64n, 79n, 84n Hegel, G.W.F., 59, 80, 85, 142, 253 Hempel, Carl G., 135n Herschel, John, 73, 73n Hesse, Mary, 98, 98n

Howison, G. H., 235n

Hugly, Philip, 245n Hume, David, 26, 27, 27n, 34, 133, 151, 152 Husserl, Edmund, 247 Huygens, Christiaan, 118, 119n James, William, 18n, 34, 34n, 115, 115n, 139, 232 Jeffrey, Eber, 84n Jevons, W. S., 65n Joachim, H. H., 174n, 209n Johnson, Samuel, 321 Kahn, David, 237n Kalven, Harry, Jr., 197n

NAME INDEX

330 Kant, Immanuel, 5, 93, 93n, 115, 123, 142,

183, 187, 187n, 248n, 257, 264, 321

Protagoras, 280 Ptolemy, 15

Pucelle, Jean, 324n Purcell, Edward, 118n

Kepler, Johannes, 103n, 179 Kneale, William, 50n, 86n Krénenberg, Moritz, 324n Kuhn, Thomas, 1170 Kuhn, Thomas S., 48, 49, 146n

Putnam, Hilary, 228-29, 278n, 292n Pyrrho, 59 Quine, W. V., 47n, 52n, 107n

Lack, David, 235n Lehe, R. T., 209n Lehrer, Keith, 279n Leibniz, G. W., 6, 6n, 142, 248n, 309 Levi, Isaac, 168n Levin, M. E., 193n Lewis, C. L., 34, 147n, 171n, 172n, 249, 249n Lewis, David, 200n Locke, John, 6n, 27n

Sabini, John, 196n, 197, 197n Sagan, Carl, 122n Santayana, George, 317

Sayward, Charles, 245n Schlick, Moritz, 173n

Malmonides, 261 Mandelbrot, Elie, 70 Margenau, Henry, 87 Massey, Gerald, 102n Maxwell, J. C., 15 Meinong, Alexius, 171n Mill, John Stuart, 138n, 139n, 233 Miller, Dickinson S., 135n

Mittelstrass, Jiirgen, 103n Moore, G. E., 321 Nagel, Thomas, 117n Newton, Isaac, 15, 80, 113, 130, 292 Ofsti, Arne, 155n

Schopenhauer, Arthur, 4, 317 Schrédinger, Erwin, 96, 96n, 106 Sellars, Wilfred, 278n, 279n, 298n Shakespeare, William, 73 Shannon, Claude, 237 Shklovskii, I. S., 122n Silver, Maury, 196n, 197, 197n Simmel, Georg, 115, 115n Simon, Herbert A., 67n Simpson, G. G., 231n Smith, Adam, 31 Sober, Elliot, 137n Socrates, 309 Solovine, Maurice, 97

Sowell, Thomas, 199n

Palmer, Scott D., 209n

Pap, Arthur, 209n Parmenides, 313 Pascal, Blaise, 65 Peirce, C. S., 4, 24, 25, 25n, 34, 47, 55, 59, 65n, 66, 84, 87, 87n, 93, 99, 129, 146n, 155, 206n, 213n, 233-35, 251, 258, 267, 267n, 272-74, 276n, 280, 282,

297-99, 321

Ross, Angus, 200n Ross, W. D., 132n Royce, Josiah, 85, 253 Russell, Bertrand, 155, 155n, 172, 173n, 232n, 304, 304n Rutherford, Ernst, 291

:

Plato, 6n, 8n, 85, 121, 232n, 248n, 309, 313 Polanyi, Michael, 179n Polya, George, 146n Popper, Karl, 56, 97, 97n, 160n, 236n, 276n

Spencer, Herbert, 4 Spinoza, Benedictus de, 64n, 85, 192, 192n, 253 Sprigge, Timothy, 324n

Strawson, P. F., 151n, 244n Suppe, Frederick, 255n ‘Tarski, Alfred, 57 Tennessen, Herman, 232n, 235n Themistius, 261 Tigler, F. J., 112n Toulmin, Stephen, 146n Unger, Peter, 32n

NAME INDEX Vigier, Jean Paul, 66, 67n Vollmer, Gerhard, 107n Weber, Max, 301, 301n Whewell, William, 135, 185n White, A. R., 173n Whorf, Benjamin Lee, 116

331 Wigner, E. P., 50n, 65n, 67n, 87n, 96, 96n

Wilkinson, Denys, 66n Willmann, Otto, 324n Wittgenstein, Ludwig, 219 Wundt, Wilhelm, 297n Zeisel, Hans, 197n

Subject Index

accessibility, 255 adaequatio ad rem, 260-64, 285

description, 253

adequacy, cognitive, 218, 223-40

detectability, 114

aims of science, 283-88 appearance vs. reality, 263-64 approximationism, 47-52 aspiration, 45-46 autonomy, 255

economy, cognitive, 8, 194-208 empirical inquiry, 49-52, 247-48 end-to-science argument, 64

axiological conventionalism, xi

descriptive particularity, 54-56

entity cognizability, 324-27

benefits, communicative, 202-9 best-fit analysis, 159-60

epistemic versions of ideation, 305 erotetic completeness, 78-80 ~ escalation, technological, 87-89 estimation of truth, 52-54, 129-56 evolution, 3-4, 11, 36-46; of systems, 176~

causality, 312-16 closing of the circles, 268-70 cognitive Darwinisn, 226-31 cognitive economy, 194-208

evolution, cognitive, 231-38

cognitive evolution, 231-38 cognitive idealism, 306-7

cognitive inexhaustibility of nature, 68-76 cognitive opacity, 248-51 cognitive rationality, 15-35 cognitive systematization, 129-56 coherence and coherentism, 145-48, 157— 82, 209-11, 216-22 communication, 185-208, 261-62; scientific, 204-8 communicative benefits, 202-9 communicative negativities, 203-9 communicative parallax, 183-84 completeness, 217-18; erotetic, 78-80;

pragmatic, 81-83; predictive, 80-81 conceptions, 190-93, 251-54 conceptual idealism, 310-24 consensus, 11-14

.

80 explanatory idealism, 309-10 extraterrestrial science, 112~25 fact idealism, 306

fact knowledge vs. thing knowledge, 32627 facts, 243-48 fallibilism of science, 275~77 finality, temporal, 84-85 foundationalism, 161-65 functional justification, 264-65 future science, 117, 275 goals of science, 77, 224-25 Hegelian inversion, 180-81, 239-40 hindsight: wisdom of, 266-70 homo mensura, 280-83

homo quarens, 4

conventionalism, axiological, ix convergentism, 47-52

cooperation, 42-45

corrigibility of conceptions, 251-54 ‘eriteriology of truth, 157-82 Darwinism, cognitive, 3-4, 36-46, 226-31;

of methods, 226-31; of theses, 236-38

data, 158-60, 166-69 Denkmittel, 115

ideal coherence, 209-11, 216-22 idealism: cognitive, 59, 305-7; conceptual, 310-24; explanatory, 309-10; factual, 306; metaphysical, 270-79; objections to, 319-22: and realism, 304—24; substantival, 307-9 idealization: and science, 210-14, 296-303; and truth, 209-29; validity of, 300-303 ideal-state realism, 298-303

334

INDEX

imperfectability of natural science, 77-95

particularity, descriptive, 54-56

importance, 50-51imputation, 187-88 incapacity, 272-74 induction, 129-56; retrojustification of, 148-56

perfected science, 85-86, 89-90, 212-14, 296-300

inductive inference, 129-36 infinity of nature, 64-76 inquiry, 47-57, 247-48; presuppositions of,

physicality, 255 physics: possibility of, 96-109 plausibility, 136-39 postulation, 183-208 power, 82

pragmatic completeness, 81-83

intentionality, 185-88

pragmatic success, 226-38 pragmatism, 26-29, 259-66 praxis, 33-35, 223-40 predictive completeness, 80-81 presumption, 16-19, 29-31; of rationality, &I11

interactionism, xi, 119 interests, 7-8

presumptive truth, 211-12 presuppositions of inquiry, 268-70

justification, functional, 264-65

progress, 212-13; scientific, 49-52; technological, 87-92 -

knowledge, limits of, 243-54, 324-27

proposition cognizability, 324-27 putative truth, 277

language, 243-48 lawfulness, 312-16 laws of nature, 66—74

rational inquiry, 47-57 rationality, 3-14, 31-33; cognitive,15-36; presumption of, 8-11

limits: of knowledge, 243-54; of science,

realism, 226-31, 239-40, 248-51; of

268-70 instrumentalism, 286-89 intelligence, 4-8, 36-46, 103-4

intelligibility of nature, 96-109 intent: realism of, 284

63-76

achievement, 284; and idealization, 298~

man as measure, 280-83

303, 304-24; of intent, 284; metaphysical, 255-79; regulative aspect of, 259-66;

metaphysical idealism, 270-79

scientific, 277-95, 296-300; situational,

metaphysical realism, 255-79 method Darwinism, 226-31

mind dependency, 312-16 mind-independence, 274 misfortunes, 22-23 natural laws, 66-74

natural science: imperfectability of, 71-95. nature: cognitive inexhaustibility of, 68~76; infinity of, 64—76; intelligibility of, 96—109 negativities, communicative, 203~9

nomic complexity, 68 objections to idealism, 319-22 objectivity, 11-14, 188-93, 262-64

ontological versions of ideation, 305 opacity, cognitive, 248-51 ordinary-life communication, 204-8

122-25

reality, 226-31; vs. appearance, 263-64 regulative aspect of realism, 259-66 relativism, scientific, 110-25 retrojustification, 268-70; of induction, 148-52 risk, 20-23

schoolbook science, 289-95 science: adequacy of, 223-40; extraterrestrial, 112-25; of the future, 275; goals of, 77, 224-25, 283-88; and idealization, 210-14, 296-303; limits of, 63-76; perfecting of, 85-86, 89-90, 212-14, 296~ 300

“science in thelimit, ”47, 299 scientific communication, 204-8 scientific fallibilism, 275-77

scientific progress, 49-52 scientific realism, 277-95, 296-300

parallax, communicative, 183-84

scientific relativism, 110-25

SUBJECT INDEX scientific truth, 223—40 self-subsistence, 255 self-substantiation of induction, 152-56 situational realism, 122-25 skepticism, 20-29, 31-35, 56-59

substantival idealism, 307-9 success, pragmatic, 226-38 systematization, cognitive, xii, 129-56 system-evolution, 176-80

335 thesis Darwinism, 236-38

think knowledge is fact knowledge, 326~27 truth, 56-59, 243-48; and idealization, 209-22; putative, 277

truth criteriology, 157-82, 223-40 truth estimation, 52-54, 129-56 “ultimate science,” 47, 299 unobservables, 295

technological escalation, 87-89

technological limits, 91-92 technological progress, 87-92 temporal finality, 84-85

validity in idealization, 300-303 wisdom of hindsight, 266-70