Bill Bartley reviews the Popper volume in the Library of LIving Philosophers series edited by P A Schilpp

Bartley on Rationality, Criticism, Logic

An Ecological Approach

This was on line at the Critical Rationalism site, now replaced by the CR Blog.

CRITICAL STUDY

THE PHILOSOPHY OF KARL POPPER

Part III. Rationality, Criticism, and Logic

Philosophia 1982

W.W. Bartley, III

The Philosophy of Karl Popper, edited by Paul Arthur Schilpp, Two Volumes, Open Court, Library of Living Philosophers, La Salle, 1974. 1323 pp., $30.00.

The reasonable man adapts himself to the conditions which surround him. The unreasonable man persists in trying to adapt surrounding conditions to himself . . . All progress depends on the unreasonable man. George Bernard Shaw

I. Introduction

This is the third in a five-part critical study of the work of Sir Karl Popper, based on a review of the Schilpp volume in his honour. The first study dealt with biology, evolution theory, evolutionary epistemology, and Popper's doctrine of the "Three Worlds." 1 The second treated Popper's interpretation of quantum mechanics, probability theory, entropy, time, indeterminism, consciousness, and the body-mind problem. 2

This third instalment deals with rationality, criticism, and logic. Throughout, my goal is to contribute to creating a "body of informed and serious criticism" of Popper's thought. I aim to sketch the general problem situation within which Popper's thought has to be evaluated, and to indicate the current state of discussion of his theories. 4 In the present paper, I shall build a connected argument relating to rationality, criticism, and logic -- introducing Popper's views, and those of the contributors to the Schilpp volume, where they are relevant.

II. The Rational Way of Life

Although much of his written work relates to problems of rationality, Popper's most direct treatment of rationality dates to the mid nineteen-forties, and is found chiefly in Chapters 22 and 24 of The Open Society and Its Enemies (1945) and in several essays, "Utopia and Violence" (1948), "Towards a Rational Theory of Tradition" (1949), and "Humanism and Reason" (1951), reprinted in Conjectures and Refutations.

Rationality remerged as an important theme of Popper's work, and of the entire Popper school, in 1959-63, partly as a result of a running dialogue between Popper and myself. Out of this discussion, a number of books and papers were written almost immediately. There were for instance my own The Retreat to Commitment (1962) and "Rationality versus the Theory of Rationality" (1964) 5. Popper added an addendum on "Facts, Standards and Truth: A Further Criticism of Relativism" to the fourth (1962) and subsequent re-editions of The Open Society. At the same time, he made important revisions in his discussion of rationality in Chapter 24 of The Open Society, and added a new opening section on rationality to his unpublished Postscript. He also treated the matter in his Preface to Conjectures and Refutations (1962), and in his essay, "Truth, Rationality, and the Growth of Scientific Knowledge" (1962). lmre Lakatos applied these ideas in his "Infinite Regress and Foundations of Mathematics" (1962), as did J.W.N. Watkins in his "Negative Utilitarianism" (1963). Inspired by our discussion, Hans Albert, in Germany, began a long series of publications on rationality, culminating in his Traktat uber kritische Vernunft (1968). 6 From all this writing a large literature has grown.

The entire discussion touches issues of fundamental importance –- more important than those broached in the first two instalments of this series. Biology and quantum mechanics are two areas where Popper's ideas are applied. Theory of rationality, on the other hand, develops the fundamental ideas themselves. Nonetheless, rationality remains a comparatively little explored area of Popper's work - at least where the English-language readership is concerned -- despite its importance to his philosophy, which is, as a whole, often called "critical rationalism." Like physics and biology, the theory of rationality is largely neglected in the Schilpp volume, although one paper in it, A.E. Musgrave's "The Objectivism of Popper's Epistemology," reports some of the discussion of rationality to which I have just referred.

Many years ago, Popper himself used to complain about disregard of the issues of rationality. As an example, he would cite the work of an American philosopher who had made an extended study of different "paths of life" without even mentioning the rational way of life. Popper sees the rational way of life as consisting in (1) the quest for knowledge and truth, for "emancipation through knowledge," and "spiritual freedom";8 (2) the critical attitude that -- recognizing that any particular expression of the truth is fallible, limited, not final -- seeks undogmatically to subject all attitudes, ideas, institutions, traditions, so-called knowledge and so-called spiritual freedom, to critical examination and appraisal. 9 As Popper puts it: "Rationalists are those people who are ready to challenge and to criticize everything, including . . . their own tradition."10 (3) The rational way of life thus also involves the willingness to learn from others. Emphasizing how much we depend on others for knowledge, and the social character of language and reasonableness, Popper writes that "We must recognize everybody with whom we communicate as a potential source of argument and of reasonable information," and take the attitude that "I may be wrong and you may be right, and by an effort, we may get nearer to the truth." Thus is established the "rational unity of mankind." In a minor departure from Kant, the other person is recognized as an end in himself in that he is a source of criticism and correction.

Such talk may seem uncontroversial -- even insipid or needlessly hortatory. Its interest emerges as it is shown that virtually every traditional and contemporary western philosophy combines doctrines, assumptions, and practices that militate against such a way of life. For a brief preliminary example, take J. Bronowski's statement in A Sense of the Future (p. 4): "To listen to everyone, to silence no one; to honour and promote those who are right; these have given science its power in our world and its humanity." The phrase I have italicized conflicts sharply with Popper's approach, and with his understanding of rationality.

III. Four Problems of Rationality

I see at least four separate although related problems of rationality arising within Popper's thought -- each pertaining to the conditions and the possibility of the rational way of life. These are:

1) The problem of the ecology of rationality,

2) The problem of the limits of rationality,

3) The demarcational problem of rationality.

4) The problem of the limits of explanation or the limits of knowledge-achievement.

The terminology just introduced is mine. I aim to state the issues more generally than Popper himself does, so that his own contribution to the understanding of human rationality may be the better appreciated. Nothing hangs on the terminology: to suggest how little it matters, I refer readers to the quotation from Shaw with which this study begins. Generally, I mean by "reasonable" what Shaw means by "unreasonable": and I agree with what Shaw says. I do not think of reasonableness as "cooperativeness" or "submissiveness" to circumstances. Indeed, it is regrettable that common practice inclines one to call these problems ones of rationality, rather than of the limits of criticism and argument, and of the conditions furthering the growth of knowledge, awareness, and discovery. The term "rationalist," in particular, has been purloined by some thinkers who are foreign to Popper's way of thinking. Uncritical thinkers who are as imbued with dogmatism, partisanship, and rigidity as any of whom one might think. On the other hand, other thinkers who, because of their stance regarding the problem of the limits of rationality, may be called "irrationalists" -– Karl Barth for example - are on the whole wide-ranging, flexible, and critical individuals.11 I ask the reader kindly to bear this proviso in mind as he or she reads my later remarks about various kinds of rationalist and rationalism.

The first problem listed - the problem of the ecology of rationality – is implicit in some of Popper's writings but has never been stated explicitly. Once Popper's work is interpreted - as in the first instalment of this study - as a variety of evolutionary epistemology, this problem becomes a major focus of philosophical endeavour. It will be discussed in Sections IV and XX of this paper, and is implicit throughout.

The second problem -- that of the limits of rationality -- is one of the two or three major problems of philosophy as hitherto understood. I solved this problem in 1960, through a generalization of some of Popper's ideas.12 The core of my solution was my distinction between justificational and critical arguments. This distinction enables one to avoid, in principle, the ancient dilemma of infinite regress versus ultimate commitment. Commitment is not logically necessary; and it is logically, theoretically, possible to hold open all one's structures, theories, beliefs, to reexamination and criticism. The solution to this problem contributes to, and indeed licenses, the broader problem-programme of the ecology of rationality. This problem will be discussed in Sections V-- XII of this paper.

The third, or demarcational, problem of rationality concerns classification and preferential selection among competing positions. This problem has been clarified by Popper's work, much of which contributes directly to it. Although important historically, the demarcational problem falls to subordinate status within the broader programme of the ecology of rationality. This is discussed in Sections XXI--XXIV of this paper.

The fourth problem that of the limits of explanation or knowledge achievement is more widely understood, and examples are readily available. Some writers maintain, controversially, that certain limitations arise from the psychological and biological structures of the human mind (See Part I, p. 486). Less controversial is the contention that certain physical conditions in nature - e.g., the existence of radiation chaos and the velocity of light limit research in certain parts of the universe. Again, we can sometimes attain the principle on which phenomena of a certain class operate without being able to explain more concrete detail. For instance, while knowing the mechanisms whereby waves are formed on the surface of the water, we are unlikely ever to be able to predict the shape and movements of particular waves. Other limitations arise from our historical existence and the impossibility of predicting the historical future due to there being at least one factor in its shaping that we demonstrably cannot predict: namely, the future growth of human knowledge. This limitation Popper himself has stressed: we predict by reference to our present theories; we learn by refuting our present theories, by deriving predictions from them and trying to falsify those predictions. But we cannot derive or predict a refutation of these theories from these theories (provided they are consistent). Related limitations appear in physics and economics due to the existence of indeterminacy and feedback. Yet other limitations are connected with the necessarily selective character of description. Then there are practical limitations, such as those explored by Freud and Jung, due to human weakness, physical frailty, humanity. A different sort of limitation is stressed by F.A. von Hayek, who argues that any apparatus of classification must possess a structure of a higher order of complexity than that possessed by those objects which it is to classify; hence no examining agent can ever explain objects of its own kind or degree complexity. The human brain, for example, can never fully explain its own operations. 14

These limitations on explanation, knowledge achievement, and cybernetic operation relate importantly to issues of rationality. Many of them are illuminated in Popper's work. They will not, however, be treated in this study, which will focus, rather, on the first three areas: on the ecological, logical, and demarcational problems of rationality.

IV. The Problem of the Ecology of Rationality

Ecology is the theory of the interrelationship between an organism and the environment, and has to do with survival. The pertinent part of the human environment will contain - in addition, of course, to people, plants, animals, and things (a) a variety of vicarious representations of the environment (Part I, this series, p. 472) and (b) a variety of recommended ways of behaving within the environment. I call both of these positions. The pertinent environment will also contain: a variety of contexts for these positions; criticisms of and objections to various positions and contexts; various contexts of contexts, or metacontexts. The human econiche is one in which people hold conflicting positions in conflicting contexts and in terms of conflicting metacontexts.

We are concerned here with positions, contexts, and metacontexts which humans adopt and in terms of which they function and orient themselves -– whether these positions, contexts, and metacontexts be conscious or unconscious; whether they be subjective or objective; whether they be in Popper's World 2 or in his World 3. We are concerned also with the environment in which positions, etc., are taken, held on to, or abandoned; and with the alterations such positions, etc. -- such "exosomatic extensions of self" -- effect on the environment and vice versa.

An example of a position could be a simple statement purporting to be true or right: e.g., "The human soul is immortal," or "abortion is wrong." Similar-appearing positions may be embedded in quite different contexts. Christianity, Buddhism, Hinduism and Humanism, for instance, provide different contexts for the position-statement. "The human soul is immortal."

A context is not simply the sum of the positions it contains, but is also the framework and even the sensibility in which these positions are couched; and it weights positions with regard to importance and significance. Examples of contexts are belief systems, ideologies, traditions, institutions. To illustrate how a sensibility casts a context over a statement, one might notice the way in which I, in the comfort of my study, reach out to the overflowing bowl of strawberries, take and eat one, and comment, "This strawberry is so sweet." And one might contrast this with the Zen sensibility of the doomed Samurai, trapped on a collapsing bridge over a deep ravine and who, savouring the moment reaches out to pluck the strawberry growing wild on the steep bank, and says, "This strawberry is so sweet."

It is important ecologically to notice that many persons find personal identification or "identity" (in the sense of the psychologist and sociologist Erik H. Erikson 15) in contexts. They employ contexts to define themselves and their relations to others and to the world. Their allegiance is to the context; allegiance to particular positions sponsored within that context generally flows from allegiance to the context, rather than vice versa. The position may be able, more or less, to be predicted from the context, but not vice versa.

Rationality -- as embedding the search for knowledge and the critical attitude -- can hardly be a matter of positions or contexts. There is nothing intrinsically rational about any particular position or context. Positions and contexts may further or hinder the search for knowledge and the critical attitude. Rather, rationality would have to be a matter of the context of contexts, or the metacontext.

A metacontext differs from a context as understood here. A metacontext has to do with how and why contexts are held, subjectively and objectively. While there are endless positions and thousands of contexts, there are comparatively few metacontexts. Theory of rationality and ecology of rationality are metacontextual: they are theory about how and why to hold contexts and positions; and they depend in part on goals: e.g., is it one's goal to justify or defend a particular position? Or to attain a more adequate representation?

Robert Michels used to write that the humanist tradition possessed a myth of mission but lacked -- and needed -- a myth of origin. Popper may have provided just such a myth of origin for the rationalist tradition in one of his most brilliant essays. "Back to the Pre-Socratics," which incidentally illustrates the kind of thing I have in mind in writing of metacontexts.17 In that essay Popper writes of the two presocratic schools of the Pythagoreans and the Ionians. Contrasting the development of ideas in these schools, Popper pointed out that among the Pythagoreans a second-order tradition was developed of defending, preserving, and passing on to others the doctrines of the founder of the school. This second-order tradition had the effect of restricting development in the ideology: and of limiting changes to those that could he handled surreptitiously -- as, say, a restatement of the master's real intentions, or as a correction of previous misinterpretations. Whereas among the Ionians one had, according to Popper, the first recorded instance of widely different viewpoints being explicitly handed on, without dissent or schism, by the same school in successive generations. There a second-order tradition was developed of criticizing and of trying to improve upon the doctrines of the master. Within this second sort of school, Popper sees the origins of the critical tradition. Here a true history of ideas begins to develop, in which, along with the ideas of the leaders of the school, criticisms, and changes are also taught, respected, and recorded. There is a close connection between what I mean by "metacontext," and what Popper meant by "second-order tradition." For the purposes of this discussion, I shall restrict myself to three metacontexts:

1) The metacontext of true belief -- or justification philosophy. This aims to justify or defend positions and contexts: in Bronowski's words, as quoted earlier, "to honour and promote those who are right."

2) Oriental non-attachment. This aims to detach from positions and contexts.

3) The metacontext of fallibilism. This aims to create and to improve positions and contexts.

Most western philosophies -- philosophies of science as much as philosophies of religion - sponsor justificationist metacontexts of true belief. They are concerned with how to justify, verify, confirm, make firmer, strengthen, validate, vindicate, make certain, show to be certain, make acceptable, probabilify, cause to survive, defend, particular contexts and positions. Most such philosophies -- again, philosophies of science as much as philosophies of religion end in commitment and in identification. There may be a simple historical explanation for the entanglement of western philosophy of science in this metacontext lying in the fact that western science grew up in debate with a true belief religion, Christianity. Responding to a true-belief religion, western science became, in its philosophy, a true belief science. I shall return, in sections VI-X, to the justificationist metacontext of true belief, to examine it in detail.

At the other extreme: there is the oriental metacontext of nonattachment. In most varieties -- in Hinduism, Buddhism, and in the yogic tradition underlying both 18 - oriental philosophy opposes attachment: attachment to anything whatever: one's body, habits, wishes, lusts; cravings, aspirations, ideas, beliefs, ideologies, relationships, affiliations. From this perspective, it is as important not to be attached to particular beliefs as it is not to be attached to particular lusts and cravings. From the oriental perspective, the westerner erects positions and contexts -- belief systems -- for particularly ignoble ends, related to a kind of craving: the westerner commits himself to, identifies himself with, such systems in order to justify himself and invalidate others, to dominate and to avoid domination, to survive and make others fail to survive. By identifying with his positions, the westerner automatically becomes positional, he causes the positions with which he has identified to persist as part of his own combat for survival. This is seen as the source of that vaunted "hunger and thirst after righteousness" of which western moralities speak fondly and which orientals (and a few others, such as Nietzsche) see as a kind of vampirism, sapping the strength of western culture.

It is interesting and -- so far as I can tell -- hitherto unremarked that the oriental concept of attachment and the western idea of commitment are closely similar. Both, in turn, resemble the idea of addiction. The dictionary defines an addict as one who has "given himself over" to a practice, a habit, a pursuit. Thus it is not surprising that Supreme Court Judge Oliver Wendell Holmes, Jr., in explaining the force of his philosophical commitments or presuppositions, chooses the language of the addict: he calls these Can't Helps! 19 The chasm between the dominant trends of eastern and western thought is nowhere so evident as here: the metacontext of oriental philosophy sponsors non-attachment; the chief metacontext of western philosophy sponsors attachment, commitment, addiction.

It is alien to an eastern metacontext to see any position or contexts as a source of identification and commitment in the western sense. Yet oriental writers -- and some westerners -- sometimes speak of another, different, sense of identity: "true identity" or True Self. True Self is the context of all contexts (including metacontexts). Beyond any individual, identification, form, process, context, position, or econiche, true Self is the matrix that gives rise to them. No position, true Self is the space in which all positionality in life occurs.

Within his eastern metacontext, commitments, belief systems, ideologies, identifications, and so-called "ultimate values" provide contexts -- often valuable contexts -- for individual existence; but these are not who or what one is; nor does it make any sense to be attached or committed to such things. True Self, being the context of all Contexts, is the context in which things such as commitments, identifications, ideologies, metacontexts, and so on emerge, flourish for a time, and then decline. Thus one is the context in which content is crystallized and process occurs, and is not any individual content or process, not any individual form. Here there is a profound sense of the limitation inherent in form. Here identity as fixed identification is seen as a liability: the more fixed one's identity, the less experience of which one is capable; the less one is. The point is not to lack a position or context, but not be positional; not to be attached or committed to whatever position and context one does have at any particular moment. To adapt Jean-Paul Sartre's terminology: one may have a position or do a position, but may not be a position. 20

Two well-known western writers, Hermann Hesse and Hermann Keyserling, found in such eastern thought a "protean sensibility," writing of the "supple individual" of infinitely polymorphous plasticity who, "in order to experience enough must expose himself a great deal," and who "gains profundity from every metamorphosis." 21 As Keyserling's protean figure -- in the course of trying out different forms and experiences, different positions and contexts -- discovers how limited each is, and how one is linked to another, he 'passes beyond the danger of placing an exaggerated value on any single form, phenomenon, position or context. Personality and character, being forms, also imply limitation, "No developed individual," Keyserling writes, "can reverence 'personality' as an ideal; he is beyond prejudices, principles and dogmas." Such a supple individual, though perceived to be without character, may be as securely and firmly positive as any rigid individual.

This is not the confused and disordered state of one in the throes of "identity diffusion." The Yogi says "neti, neti: I am not that," to all nature, until he becomes one with Parabrahama. After that, as Keyserling says, "no manifestation limits him any more, because now each one is an obedient means of expression to him . . . A God lives thus from the beginning, by virtue of his nature, Man slowly approaches the same condition by passing through the whole range of experience."

On the other hand, such a god-like being may seem capricious. Like Proteus, the Greek sea god, the "old man of the sea," who not only had the power to assume any form he wished, but was also "as capricious as the sea itself." One finds such also in some Hindu accounts: the capriciousness, say, of credulity, which accepts all things, however contradictory, as vessels of the truth; which, regarding everything as holy, yet takes nothing seriously. Within limits, for to consider and discard another metaphor, the oriental approach does not commend the chameleon, that remarkable lizard with a greatly developed power to change the color of his skin. Lightning change artistry is a superficial sort of suppleness, skin-deep. for which the chameleon's characteristic slow power of locomotion is itself a metaphor.

All this may remind one somewhat of ordinary western scepticism, but it is hardly the same. Western scepticism, to be sure, also arches its brows at all attempts at knowing, all formulations, definitions, identifications --including any definition of its own position. But it is posed more as a position or context, not as a metacontext; and it is permeated by the spirit of justificationism. It rejects attachments not out of a positive quest for non-attachment but out of reaction to the internal contradictions of the western justificationist metacontext. Usually, Western scepticism is regretfully or resignedly non-attached, permeated by epistemological disappointment. None of the majesty of non-attachment appears in, say, Sextus Empiricus or Hume, as it does in the eastern writers. Scepticism is a defence against uncertainty.

Eastern thinking is then far removed from the metacontext of belief, identification, and commitment that one finds in most western philosophies. It is less distant, although still very different, from the fallibilism or critical rationalism of Xenophanes . . . or of Karl Popper. Popper's approach also allows for the fallibility, the distortion, of all forms, of all existing crystallizations in language, and yet maintains that one may, through form, through language, attain closer to the truth, measuring one's progress through . . . fallible criteria. It is precisely language, he insists, that permits one to dissociate from, to detach from, one's own positions and hypotheses: to make them into objects, not subjective states, not identified with ourselves, which may then be examined. While rejecting the identification, commitment, and positionality into which western justificationist philosophies are forced, Popper yet champions the growth of knowledge and of science, and the "rational way of life" as leading in this direction. In the interaction between ourselves and our intellectual products, he stales, we are most likely to transcend ourselves. Unlike most oriental philosophers, which are comparatively rarely interested in the growth of knowledge, Popper searches for a more adequate model or "vicarious representation" of the world; like the oriental, Popper gives no importance to "right belief," and searches for a pervasive condition of non-attachment to models and representations generally. 22 For one must detach from, must objectify, one's theories in order to improve them. The very asking of the Popperian question "Under what conditions would this theory be false?" invites a psychological exercise in detachment and objectification, leading one to step outside the point of view shaped by that theory. While Popper, however, emphasizes progress in knowledge and rationality, for the oriental the apparency of progress is illusory, and the pursuit of it is a manifestation of addiction. The oriental and the fallibilist also seek detachment for different reasons: the oriental, to attain distance from all models of the world, and thereby to win freedom from illusion, and peace, the fallibilist, in order to further the growth of knowledge, to attain a more adequate model of the universe. The products which they seek differ.

In a fallibilist metacontext, the ecological problem is to create the most lethal possible environment for positions, contexts, and metacontexts in which the production of positions, contexts, and metacontexts yet thrives (See Part I, p. 486). Popper made a step towards this understanding of the problem in The Logic of Scientific Discovery (1934), where he wrote: "What characterizes the empirical method is its manner of exposing to falsification, in every conceivable way, the system to be tested. Its aim is not to save the lives of untenable systems but, on the contrary, to select the one which is by comparison the fittest, by exposing them all to the fiercest struggle for survival" (p. 42) and "a supreme rule is laid down which serves as a kind of norm for deciding upon the remaining rules . . . It is the rule which says that the other rules of scientific procedure must be designed in such a way that they do not protect any statement in science against falsification" (p. 54).

Popper extended his approach, as applying not only to empirical science but also to political institutions, in The Open Society and Its Enemies, Chapter 7; and in 1960, in "On the Sources of Knowledge and of Ignorance" (Conjectures and Refutations, p. 25), he advocated replacing traditional questions about the source of knowledge with the question: "How can we hope to detect and eliminate error?". In The Retreat to Commitment and "Rationality versus the Theory of Rationality," I stated the problem more generally: "how can our intellectual life and institutions be arranged so as to expose our beliefs, conjectures, policies, positions, source of ideas, traditions and the like -- whether or not they are justifiable -- to maximum criticism, in order to counteract and eliminate as much intellectual error as possible?".

This formulation is still incomplete. First, it is stated too negatively in terms of the reduction of error. Whereas, an essential part of the task is the fertility of the econiche: the econiche must be one in which the creation of positions and contexts and the development of rationality, are truly inspired. Clumsily applied eradication of error may also eradicate fertility. Criticism must be optimum rather than maximum, and must be deftly applied. Second, my initial formulation overemphasizes matters intellectual such as beliefs, conjectures, ideas, and such like. I want explicitly to include for review not only beliefs, conjectures, ideas, ideologies, policies, programmes, traditions, but also institutions, and even etiquette, manners and customs, and unconscious presuppositions and behaviour patterns that may pollute the econiche and thereby diminish creativity, criticism, or both.

The ecological program of rationality is, then, how this is to be done. Before considering this question, we must first turn to the problem of the limits of rationality and to the contention that it cannot be done.

V. The Problem of the Limits of Rationality

"Judge not that ye be not judged. For with what judgment ye judge, ye shall be judged: and with what measure ye mete, it shall be measured unto you. And why beholdest thou the mote that is in thy brother's eye, but considerest not the beam that is in thy own eye? ... Thou hypocrite, cast out first the beam out of thine own eye..." Matthew 7: 1-5.

This problem, emerging from a justificationist metacontext, challenges the possibility of any Popperian -- fallibilist metacontext in which it is assumed that one can make progress towards a more adequate and objective representation of the world --claiming instead that, from a rational point of view, there can be no progress: that the choice between competing positions and contexts, whether scientific, mathematical , moral, religious, metaphysical, political, or other, is not reasoned but is arbitrary.

The core argument deployed to show this limitation consists in an analysis of the rational way to defend or justify ideas. The argument based on this analysis is called by various names including the argument about the limits of rationality, the dilemma of ultimate commitment, the problem of ultimate presuppositions.

The argument is a commentary on the fact that any view may be challenged by questions such as "How do you know?" "Give me a reason," or "Prove it!" When such challenges are accepted by citing further reasons which entail those under challenge, these may be questioned in turn. And so on forever. Yet if the burden of proof or rational justification is perpetually shifted to a higher-order premiss or reason, the contention originally questioned is never effectively defended. One may as well never have begun the defence: an infinite regress is created. To justify the original conclusions, one must eventually stop at something not open to question for which one does not and need not provide justificatory reasons. These are the halting points for rational discussion.

There are various possibilities with regard to these stopping points. They might be statements which are unanimously acclaimed as intrinsically self-evident or certain. In which case the only question would be whether they are really certain or only seem to everyone to be so. Even if all persons were subjectively to stop at the same place, the problem would remain of proving -- rationally -- that this universal subjective stopping point led to objectively true statements about the world.

Another possibility would be that these ultimate statements are matters of contention: that either they are perceived not to be certain or that different persons deem conflicting "ultimate" statements to be certain. Here ultimate relativism arises: some way of arbitrating rationally among competing ultimate stopping points by appeal to a common standard is now excluded in principle. For reason is now relativized to one's halting place or standards, and thus cannot arbitrate among different standards. Hence different halting places -- i.e., standards, criteria, presuppositions, conventions, dogmas, articles of faith -- are taken by different individuals and define irreconcilable communities. Whatever may explain how such differences arise, reason can never dissipate them.

The western sceptic and the Indian yogi would not dispute this argument thus far. They agree from the outset that truth is not to be attained through rational argumentation. Since no position can be defended by reason, they might say, so much the more reason for suspending judgement.

But it is the fideist who glories in the argument, and who challenges the sceptic and the yogi too. His claim is simple: since an eventual halt to rational justification is inevitable and cannot be made with objective and universal reason, it must be made with unreason, subjectively and particularly: One must choose or commit or attach oneself subjectively. Thus the fideist deliberately makes a final, unquestionable, subjective commitment to or choice of some way of life, or some presuppositions, or of some authority or tradition claiming to possess the competence or the right to make such decisions for him.

The limits of argument within any position come to be defined by reference to that object or claim in respect to which commitment is made or imposed, in regard to which argument is brought to a close. And if -- since the limitation is a logical one -- all men share it, if no one can escape subjective commitment, then no one can be criticized rationally for having made such a commitment, no matter how idiosyncratic.23

If one must, then . . . one may: any irrationalist thus has an excuse for subjective irrationalism, a secure refuge from any criticism of any subjective commitment: he has a "tu quoque" or boomerang argument. To any critic, the irrationalist can reply: "tu quoque," reminding him that people whose rationality is similarly limited should not berate others for admitting to and acting on the limitation. The limitation is the more telling in being accompanied by the remark that in those things which matter most, reason is incompetent: such things are beyond or above rational scrutiny; and that those things which reason can decide are of little importance.

The problem of rationality thus created has formed the controlling predicament of philosophy since the late sixteenth century, although its origins are earlier and its classical developments rather later. It had already been discussed in antiquity, and during the Hellenistic period became for a short time dominant in philosophical discussion, its most detailed statement at that period having been made by Sextus Empiricus.24 The problem was widely discussed again during the Renaissance and Reformation, when for the first time since the secular decline of Rome men had to choose amongst competing world views. Identical arguments about the limits of rationality could be used sceptically by Protestants, humanists, and scientists to attack the traditional authority and intellectual pretensions of the Roman Catholic Church; and fideistically by Roman Catholics to attack the Protestant and scientific rejection of the authority of the Pope and of Aristotle in the name of reason and science; and again fideistically by the Protestants to answer Roman Catholic allegations that Protestantism led inevitably to scepticism. Similar arguments about the limits of rationality were widely quoted In support of scepticism and fideism during Descartes's lifetime, and importantly influenced the emphases and themes of his philosophical program.

But do people really argue this way today? They certainly do. It is well to remember Michael Polanyi's remark that "it was a mistake to regard the Nazi as an untaught savage. His bestiality was carefully groomed ... His contempt for humanitarian ideals had a century of philosophic schooling behind it. The Nazi disbelieves in public morality in the way we disbelieve in witchcraft. It is not that he has never heard of it, but that he thinks he has valid grounds to assert that such a thing cannot exist." 25

It may however be insisted that there are no Nazis any more, and that the only serious irrationalists these days are theologians who, being theologians, cannot really be serious. It is therefore pertinent to quote a more recent political proclamation by a young student at Harvard, just as a reminder that this is no academic discussion. I shall quote a few philosophers (including some of this mere boy's teachers) later:

The only reason I wouldn't blow up the Center for International Affairs is that I might get caught. But the desire is there. As it is for the 7094 Computer, the Instrumentation Labs, and the Center for International Studies at M.I.T., draft boards, army bases, the Pentagon, the White House, the Capitol, New York City, Los Angeles, Las Vegas, and Harvard University ... Blowing up a bad thing will relieve much of that tension ... I may have learned only two things in my four years at Harvard. The first is that an equally intelligent, rational, and valid argument can be made on all sides of any question from any and all premises. The second is that those arguments have no relationship to anything but themselves ... The point is that arguments are based on reason and the valid laws of discourse can prove anything within their system. It is the feeling I have in my stomach against the war that matters. Any argument in favour of it does not ... We are told continually that knowledge will make us free. We are taught to ignore irrational consequences and to put our faith in reason ... We are fed reason in order to give an inferiority complex to the rest of our emotions and senses ... We are trapped in a philosophical system of cause and effect. Rationality binds the mind and restricts the soul. It might even destroy brain cells. We need to be liberated. 26

The young writer, if typical of his generation, may even think more highly of witchcraft than of "public morality" and rationality. Yet, quite apart from any tension-reducing effect his line of argument may have, an irrationalist, along with the rest of the public, pays a high price for using the rational excuse for irrationalism. As long as the tu quoque argument goes undefeated, rational criticism of subjective commitment does boomerang. But the tu quoque argument also has a boomerang effect on those who use it. To the extent that anyone employing it defends his own position by remarking on its parallel to his opponent's, to that extent he makes his opponent invulnerable to criticism from him. For the opponent, if criticized, may also use the tu quoque. Those who gain a refuge for themselves through appeal to the limits of rationality thereby provide a similar refuge for those whose commitments -- or the feelings they have in their stomachs -- differ from theirs. One gains the right to be irrational at the expense of losing the right to criticize. One gains immunity from criticism for one's own commitment by making any criticism of commitments impossible. Thereby the many criticisms that irrationalists have levelled at rationalism -- such as the one just quoted -- become as pointless as those that rationalists have directed at, say, fideistic theology. Ultimately, the use of the tu quoque makes nonsense of the idea of progress and of the historical development and change of ideas and social institutions in the face of criticism. Unless the tu quoque can be defeated, the fallibilist metacontext is at best a wish arising from an incoherent space. And "the ultimate Ratio," as Justice Holmes put it, "is force."

One way to describe the problem of induction, as understood by Hume and Popper, is that according to one's theory of knowledge, knowledge is impossible. A way to express the problem of rationality and its limits, as presented here, is that according to one's theory of rationality, rationality is impossible.

VI. Theories of Rationality: Panrationalism

Theories of rationality are accounts of rationality created within the rationalist tradition with the intention of avoiding or defeating the argument about the limits of rationality, and thereby scotching the fideist's challenge.

It is useful to contrast three such theories:

panrationalism (or comprehensive rationalism)

critical rationalism

pancritical rationalism (or comprehensively critical rationalism).

The history of modern philosophy can -- and ought to be -- told in terms of the history of panrationalism. The following is the merest sketch. 27 Traditionally dominant, panrationalism is still the most common conception of rationality. In a review of contemporary literature, W. P. Alston and R.B. Brandt call it the "Establishment" view. 28 It is explicitly stated as early as Epictetus, who wrote in his Discourses (Chapter 2) "To be a reasonable creature, that alone is insupportable which is unreasonable, but everything reasonable may be supported." Such definition combines two requirements:

1) A rationalist accepts any position that can be justified by appeal to the rational authority;

2) A rationalist accepts only those positions that can be justified in this way. 29 The second requirement forces a rationalist to be able to justify rationally everything he holds - including these two requirements. Rationalists have explicitly, but not necessarily rationally, embraced such requirements on many occasions. T.H. Huxley, for instance, comically and humourlessly claimed that his form of rationalism demanded "absolute faith" in the validity of the second requirement. 30

Most histories of modern philosophy are structured within panrationalism, and thus focus on subordinate questions that arise only when panrationalism is assumed to be correct. Of these, the most important is: "What is the nature of the authority to which a rationalist appeals to justify his claims?" The various modern theories of knowledge are the different answers to this question, and fall into two main categories:

1) Intellectualism 31 (or Rationalism -- with a capital "R"), according to which the rational authority lies in the intellect or reason. A Rationalist justifies his positions by appealing to intellectual intuition or the faculty of Reason.

2) Empiricism, according to which the rational authority lies in sense experience. An empiricist justifies his positions by appealing to sense observation.

All such approaches have failed, as may be seen by sketching a few important steps.

Descartes is one of the fathers of intellectualism. Bombarded by competing and doubtful theories on all subjects; horrified that many of his contemporaries had adopted a sceptical relativism; and struck by the need for some rational means of assessing competing theories, he set out to find something that would be intrinsically impossible to doubt and by means of which the worth of other theories might be assessed. If successful in his search, he would, he thought, cut off the infinite regress that produces the argument about the limits of rationality –- without resorting to dogmatic commitment. 32 If it were impossible in principle to ask for a justification of an adequate standard of rationality – that is, if the standard were indubitable, self-justifying as it were – the relativist could be answered.

Descartes concluded that a rationalist should base all positions on “clear and distinct” ideas presented to the intellect. Of these, the famous indubitable cogito ergo sum was a paradigm. Such ideas would not themselves need justification because to doubt them would be absurd. The Cartesian conception of rationality began to be eroded almost at once in the 17th century, not only by sceptics like Gassendi and Sorbiere and Aristotelians like Voetius, but even by those such as Leibniz who held views similar to Descartes’s. In the eighteenth century, Locke and Hume advanced further strong arguments against it; and Kant radically and conclusively undermined it. There were many difficulties in intellectualism -– not the least of which was that, far from being indubitable, intellectual intuitions are notoriously unreliable and variable. The basic objection to all intellectualist varieties of panrationalism was (and is) that even if clear and distinct intuitions are assumed to be indubitable, they still let in too much: they are too wide. As Kant showed with his antimonies, 33 clear and distinct ideas could lead to contradictory theories. It is thus impossible, on the basis of clear and distinct ideas alone, to decide among such theories. That one’s positions are deduced from clear and distinct ideas is a distinctly insufficient guarantee of their correctness.

The other main candidate for rational authority has been sense experience. Modern versions of this empiricist or “sensationalist” view, stemming from the work of Bacon, Locke and Hume, culminate in such movements as logical positivism. According to empiricist accounts, a rationalist derives all his positions from sense observations which are manifestly true, “incorrigible,” unable to be challenged. Whereas for Descartes an irrationalist was one who held positions that could not be derived from clear and distinct ideas, for the empiricist the irrationalist is one who holds positions that cannot be derived from sense observations.

Convincing arguments against empiricism have existed since antiquity. Those having the greatest impact on modern philosophy come through David Hume, who in his Treatise and Enquiry gave empiricism an empiricist routing. Developing and strengthening arguments of the ancient Pyrrhonian sceptics as recorded in Sextus Empiricus and revived in the sixteenth and seventeenth centuries, Hume reluctantly reached the conclusion that inductive (as he thought, scientific) reasoning was irrational.

Empiricists like Hume had hoped that on the authority of sense experience it would be possible to exclude ideas about God, demons, angels, and the like, which obviously could not be derived from observation, and whose existence could not be proved by appealing to the intellect either. Yet Hume's own arguments showed that -- apart from the question of the reliability and dubitability of sense experience itself -- the empiricist criterion was inadequate: it excluded not only claims about God and angels but also scientific laws, causality, memory, and claims about other people. None of these could be reduced to sense experience; empiricism in effect reduced to solipsism -- to a variety of radical subjectivism. Since it is implausible to say that it is irrational to believe that other people exist and have minds, empiricism evidently failed to provide an adequate characterization of rationality. Whereas intellectualism included too much, ascribing rationality to untenable views, empiricism excluded too much, rejecting obviously tenable views as irrational. It was too narrow for the purpose in hand.34

Whereas empiricism gained much of its support from its claim to provide an unimpeachable authority to counter traditional authorities seen as irrational, such as those of the Church and the State, submission to the so-called rational authority of sense experience became for post-Humeans irrational too. If scientific activity rested on such illogic, who could convincingly argue against a man with a different (il)logic? As Bertrand Russell put the problem more strongly, if there were no answer to Hume's argument, there would be "no Intellectual difference between sanity and

insanity."35

Yet, despite such difficulties, panrationalism was able to survive, particularly in its empiricist forms, by a number of repair measures. For example, to deal with the fact that many seemingly rational positions could not be justified by appealing to intellect or sense experience, it was proposed that positions are of differing logical types, some of which do not need to be justified. Those that say something about the world, which describe, do need to be justified by sense experience. Others do not describe, but perform some other job; while saying nothing descriptive about reality, they help "to get around" in it. It is concluded that they need be justified not on factual grounds, but only on such grounds as predictive or classificatory usefulness. They are instruments invented to simplify and organize experience. Hence any internal troubles in them hardly matter: they can be used when useful, and discarded for other tools when they break down.

Yet when this instrumentalist or pragmatist view is probed a bit, the old troubles of panrationalism reappear. Like the tu quoque argument, instrumentalism is an ultimate weapon: once the other side learns how to use it, there is no such thing as winning or losing any more.

Irrationalists were pleased to find empiricists being forced (by the inadequacies of their view) to justify instrumentally positions that were empirically unjustifiable. Irrationalists could readily take to instrumentalism since instrumentalism reduces the possibility of a clash between science and subjective commitment. The idea of intellectual instruments or tools, with its overtone of carpentry and plumbing, grew into the more dignified and aesthetically appealing notion of symbols. God, immortality, and even angels reappeared first as instruments and "heuristic fictions," and eventually as symbols. Contentions about them could not be justified empirically; they too were full of internal conflict. But they too might be useful; they too might simplify and organize experience. Another unintended consequence of pragmatism and instrumentalism is the opening they allow for the reductive correlation of positions with arbitrary factors like environment, social class, religion, financial position, geographical location, nationality, historical period.

Thus instrumentalism, far from resolving the problem, only aggravates it. And the tu quoque, incidentally, crops up again: to choose one set of instruments rather than another can be arbitrary.

Faced by such objections, a recalcitrant panrationalist might admit that he had indeed not yet adequately characterized rationality or provided a standard by which all rational positions could be justified and all irrational positions excluded. But this, he might judge, indicates only temporary lack of success.

Such optimism is however misguided. For, as Popper has shown in his argument against "comprehensive rationalism," panrationalism is unattainable in principle. Return for a moment to the two requirements cited above: that any position that can he justified by rational argument is to be accepted; and that only positions which can be justified by rational argument are to be accepted. The panrationalist accepts anything that can be rationally justified, and also is ready to justify rationally anything that he accepts.

Yet these two requirements cannot be held simultaneously. If we accept the second, we must justify the first. But this apparently cannot be done. The first requirement is not justifiable by sense experience, by intellectual intuition of clear and distinct ideas, or by any other rational authority ever proposed. Moreover, any such justification of the practice of accepting the results of argument, even if it could per impossible be carried out, would be pointless unless it were already accepted that a justification should be accepted at least here -- which is at issue. The argument would be effective only on those who had already agreed that argument should count. 32 It is pointless to try to prove something to a person who does not accept that proofs should be accepted. Thus it seems that an argument in favour of this requirement, to be effective, presupposes commitment to argument. So, if the first requirement cannot be justified, either theoretically or practically, the second requirement forbids that one hold it. In fact, the situation is worse. For most of what has been said about the first requirement can also be said of the second: it too, cannot itself be justified by appeal to rational criteria or authorities. Therefore it must, if correct, be rejected. It asserts its untenability.

So a panrationalist not only does not happen to exist, but is logically impossible. If the self-styled rationalist not only cannot justify his presuppositions rationally; if his position is basically self-contradictory, why should an irrationalist worry about the unjustifiability of his own position? He is actually better off than the rationalist, since he did not claim to be able to justify it. Panrationalism of the sort that I have discussed here was identified by Popper in The Open Society and Its Enemies, where it is referred to as "uncritical or comprehensive rationalism," and further criticisms of it are to be found there. 38

While the crisis engendered by the collapse of panrationalism remains unresolved, all would-be rationalists who are aware of their tradition are pitched into a crise pyrrhonienne. Whenever a rationalist accuses another of irrationalism, the irrationalist can reply that what is impossible cannot be morally demanded, and that the sort of panrationalism which the rationalist demands is, rationally speaking, impossible. In such a situation, the "moment is rather ill chosen for prophesying the extinction of a deep-rooted system of religion because your own studies make it seem to you incredible; especially if you hold a theory of knowledge that regards all opinions as arbitrary postulates."39 The panrationalists deified Reason, "and all she gave ... in return was doubt, insecurity, self-contempt, insoluble contradictions."40

VII. Critical Rationalism. Ayer, Putnam, Popper

These are serious difficulties, and drove some panrationalists into outright irrationalism. Others attempted to face and resolve them -- in effect, to create an alternative theory of rationality.

Their attempts vary widely in substance and sophistication. Here I shall examine three representative contemporary approaches to the problem. I call this sort of approach "critical rationalism," after Karl Popper. It is a critical rationalism in the sense that it begins by acknowledging that the principles and standards of rationality cannot be justified. It is, I suppose, in the tradition of what Henry David Aiken calls "a tradition to end all traditions, which is committed, at bottom, only to the principle of reasonableness itself, the principle, that is, that a reason may be properly requested for any proposition whatever."41 I shall review not only Popper's approach, but also those of Sir Alfred Ayer and Professor Hilary Putnam. This approach to rationality is also to be found in Wittgenstein, and that is some token of its influence. To bypass here the familiar problems of Wittgensteinian exegesis, I refer the reader to my discussion of that issue elsewhere, 42 and turn instead to the more lucid accounts of Ayer, Putnam, and Popper, in that order.

Ayer began with a panrationalist positivism of sense experience in his Language, Truth, and Logic (1935). But by the mid 'fifties he began to develop a variety of critical rationalism, as may be found in his book The Problem of Knowledge (I956). Ayer concedes that it is impossible to provide a rational justification for basic philosophical standards, principles, procedures. It is impossible to give a proof "that what we regard as rational procedure really is so; that our conception of what constitutes good evidence is right." (p. 74) But his strategy is to minimize the importance of the sceptical arguments against panrationalism whilst granting their cogency, to make the sceptic's victory "bloodless."(pp.75,80)

By his concession, Ayer seems to avoid claiming the ability to do more than he logically can. Yet simply to discard the demand that the standards of rationality be justified hardly suffices. Ayer must proceed to show how his approach, as a theory of rationality, can afford to dispense with the requirement that standards be justified. Initially, he seems alert to the scope of his task. He writes of the importance of showing "in a way that satisfactorily disposes of the sceptic's disproof," that the procedures "which sustain our claim to knowledge ... do not require a proof of their legitimacy." (p. 74)

Yet Ayer fails ever to show anything of the sort. Why on his account do our standards of rationality not need rational justification? Simply because any such standard "could be irrational only if there were a standard of rationality which it failed to meet; whereas in fact it goes to set the standard: arguments are judged to be rational or irrational by reference to it." (p. 75) "When it is understood," he explains, "that there logically could be no court of superior jurisdiction, it hardly seems troubling that inductive reasoning should be left, as it were, to act as judge in its own cause. The sceptic's merit is that he forces us to see that this must be so." (p. 75) "Since there can be no proof that what we take to be good evidence really is so," then "it is not sensible to demand one." (p.81).

Such a position, even if assumed to be coherent, fails as a theory of rationality. The nub of the attack on panrationalism was not simply that panrationalism is impossible, but that since it is impossible, the choice among competing ultimate positions is arbitrary. A theory of rationality that begins by admitting the unjustifiability of standards of rationality must go on to show that arbitrary irrationalism can be escaped without panrationalism. Thus Ayer's discussion begs the question and is itself a variety of fideism -- and hence no answer to it.

Consider his argument more closely. He contends that our standards of rationality enjoy an immunity from the demand for justification since it would be impossible to judge them to be irrational. For they set the standards on which any such judgement of their own irrationality would have to be based. Now an argument such as this could not be relevant, let alone valid, unless some particular standards and procedures of rationality, such as Ayer's own, which include "scientific induction,"are assumed to be correct. If some particular standards of rationality are correct, then there can exist no other rational standards which are also correct but which can nevertheless invalidate the former as irrational. This "if" marks a crucial assumption: this is precisely what is at issue. Criticisms of putative standards of rationality have always questioned whether they were correct. Alternative conceptions of scientific method, such as Popper's, which deny the existence of inductive procedure, let alone its legitimacy, do claim that there are standards of rationality which positions such as Ayer's fail to meet.

In the circumstances, it comes as no surprise to find that –- however radically their substantive positions may vary -- from a structural standpoint Ayer's position is closely parallel to that of a belligerently fideistic irrationalist, the late Karl Barth (1886-1968), the leading twentieth-century exponent of Calvinist "neo-orthodoxy." To appreciate the parallel, note that, as an alternative to justifying rational standards, Ayer sees the task of the philosopher as that of describing them. (This would be an important task if critical judgements are to be made by reference to accepted procedures of rationality. Accurate descriptions of accepted procedures would then be needed.) For Ayer, the business of the philosopher becomes to analyze, and state as principles, the patterns of accepted ways of speaking.

Following the same pattern of argumentation, Barth had maintained that although the content of the "Word of God" might be expounded or described, it is unnecessary for the theologian to apologize for it, to justify it before sceptical criticism. To do so is not only logically impossible, but is unnecessary and irreverent for anyone who is committed. There can be no assessment of the Word of God or of the Christian's ultimate commitment to it since the Word of God is the standard or criterion which any such assessment would have to use. In such circumstances, as Barth says, "There

is no justification for doubt itself ... No one should flirt with his unbelief or with his doubt. The theologian should only be sincerely ashamed of it." 44

Ayer did not put the point much differently: "Inductive reasoning should be left, as it were, to act as judge in its own cause" (Problem of Knowledge, p. 75). In Ayer's case as in Barth's, when we confront that to which commitment has been made, "it is we who are being judged.”

The shift made by Ayer from panrationalism to critical rationalism can even be described in traditional theological terminology. Both Ayer and the Christian fideists abandon apologetics (in theology the procedure which gives rational justification for commitment) and replace it with kerygmatics (which is devoted to the exposition and description of the fundamental message). While Barth begs the question of the existence and righteousness of God and his Word, Ayer begs the question of the existence and legitimacy of induction. Ayer's position bars in advance criticism of his fundamental standards by laying down a "persuasive definition" of "rational" in terms of which his own standards are so.

A position similar to Ayer's is stated by Professor Hilary Putnam, of Harvard University, in his contribution to the Schilpp volume, "The

'Corroboration' of Theories." After objecting to a sharp demarcation among scientific, political, and ethical ideas, Putnam speaks to the problem of ultimate justification, writing (pp. 238-9, my italics): "In this sense 'induction is circular'. But of course it is! Induction has no deductive justification; induction is not deduction. Circular justifications need not be totally self-protecting nor need they be totally uninformative: the past success of 'induction' increases our confidence in it, and its past failure tempers that confidence. The fact that a justification is circular only means that that justification has no power to serve as a reason, unless the person to whom it is given as a reason already has some propensity to accept the conclusion. We do have a propensity -- an a priori propensity, if you like -- to reason 'inductively,' and the past success of 'induction' increases that propensity ... Practice is primary." 45

To cater to a propensity -- inclination or predisposition -- to accept something which would, apart from such a predisposition, be unacceptable is precisely what is meant by "to be prejudiced" or "to act as judge in one's own cause." It is unclear from Putnam's discussion whether this 'a priori', prejudice is also supposed to be innate. In any case, the point of rationality is to review prejudices, attachments, propensities (a priori or otherwise), and to examine and criticize them. A good reason - as opposed to a rationalization -- is one that works independently of prejudice. If the problem of rationality cannot be solved, there are no good reasons; there are only rationalizations.

Sir Karl Popper, of course, has different standards of rationality from those of Ayer and Putnam. In particular, he denies the very existence, let alone the legitimacy, of "scientific induction." So there is some irony in taking Popper as my third example of a critical rationalist -- even though this is his own term for his position. Although the problem of the limits of rationality can, I think, eventually be solved only within the metacontext of a Popperian-style fallibilism, Popper's own, explicit first attempt to solve the problem is nonetheless as inadequate, and as fideistic, as Ayer's and Putnam's, and seems to operate within a justificationist metacontext foreign to the dominant themes of his own thought.

Popper's fideism is explicit and flamboyantly displayed in his proposal, in The Open Society and Its Enemies, to adopt a "minimum concession to irrationalism." He writes:

"Whoever adopts the rationalist attitude does so because he has adopted, without reasoning, some proposal or decision, or belief, or habit, or behaviour, which therefore in its turn must be called irrational. Whatever it may be, we can describe it as an irrational faith in reason , . . the fundamental rationalist attitude is based upon an irrational decision, or upon faith in reason. Accordingly, our choice is open. We are free to choose some form of irrationalism, even some radical or comprehensive form. But we are also free to choose a critical form of rationalism, one which frankly, admits its limitations, and its basis in an irrational decision (and to that extent, a certain priority of irrationalism).”46

This choice in which, as Popper says, we "bind" ourselves to reason, is for him not one between knowledge and faith, "but only between two kinds of faith. The new problem is: which is the right faith and which is the wrong faith."47 Thus Poppers discussion of rationality in The Open Society and Its Enemies utterly fails to solve the problem of the limits of rationality (if indeed an aim of an adequate theory of rationality is to escape fideism).

The position Popper takes here in The Open Society is, moreover, anchored in his earlier work. In his first work, Die beiden Grundprobleme der Erkenntnistheorie (1932),48 for instance, Popper had, in a different context, expressed himself similarly. Writing that "aims and goals can be quite various. I hold a rational decision among them as impossible" (p. 395), Popper went on to say:

"Wir stehen also mit dem Konventionalismus gemeinsam auf dem Standpunkt, dass die letzten Grundlagen alles Erkennens in einem Akt freier Setzung das heisst in einer Zielsetzung, die ihrerseits nicht mehr rational begrundet werden kann, zu suchen ist. Es ist in anderer Form Kants Gedanke vom Primat der praktischen Vernunft."

In Popper's defence one may of course point out that he is open and forthright about this fideism, whereas the fideistic character of Ayer's and Putnam's positions is not displayed, and may not even be recognized. This openness does not, however, solve the problem.

Moreover, such remarks as these seemed to me to be out of step with Popper's own approach. Thus in April 1960 I discussed these matters with Popper, and suggested how the problem could nonetheless be dealt with within the general framework of his own approach, in terms of my distinction between justification and criticism, to be introduced presently. In response, Popper altered the terminology of Chapter 24 of The Open Society (fourth and subsequent editions) to mute its fideism, and introduced a polemical addendum on relativism. In Conjectures and Refutations, Chapter 10, and in his still unpublished Postscript,49 he introduced my distinction between justification. and criticism, and this distinction is now routinely presented as a feature of Popperian thought. Despite these alterations, Popper's earlier fideistic approach has so far been corrected in a patch-work manner, dropping some of the old notions, but retaining the old terminology -- "critical rationalism" for instance -- and the old slogans.50 This results in a rather confused situation. In the following, I attempt to show how the problem can after all be solved.

VIII. Dogmas of the Justificationist Metacontext

The history of modern western philosophy is largely that of the failure of panrationalism to defeat scepticism and fideism. Most contemporary philosophies are forms of critical rationalism. Both panrationalism and critical rationalism arise within the polluted metacontext of justificationist philosophy of true belief. Critical rationalism, like panrationalism, is doomed by the structure of the justificationist metacontext. Within this metacontext there is no hope of solving the problem of the limits of rationality.

I have just written of justificationist philosophy as a polluted metacontext. I mean this precisely in the ecological sense, and refer to certain assumptions and doctrines that work against the critical enterprise, against what Popper calls "the rational way of life." In the presence of these pollutants, it is hard to sustain this way of life. I shall consider this whole question more broadly in section XX, where I turn again to the question of the ecology of rationality. Here I want to focus more particularly on three particular pollutants permeating the justificationist metacontext which have the effect of, preventing the problem of the limits of rationality from being solved within it. These are by no means the only ones, but seem to be fundamental to any discussion. In this and the two following sections, I shall expose these three pollutants. As they are exposed and eradicated, an alternative metacontext emerges. These three pollutants are:

1) the authoritarian structuring of the western justificationist metacontext;

2) the assumption that criticism is necessarily fused with justification;

3) the assumption that all measures and tokens of intellectual value, as properties of statements, pass through the relationship of logical deducibility from justifying premisses to justified conclusions (transmissibility assumption or "consequence condition").

The insight that the western philosophical tradition is authoritarian in structure is due to Popper (1960), as is the critique of the transmissibility assumption in its application to empirical science (1934). The unfusing of justification and criticism is due to me (1960), as is the application of the transmissibility assumption to the problem of rationality (1960).51

Let us begin with the first of these, which is simply an historical observation: namely, that our philosophical tradition -- the western justificationist metacontext is authoritarian in structure, even in its most liberal forms.

The authoritarian structure of philosophical questioning is obscured by the oversimplified way of presenting the rise of modern philosophy as part of a rebellion against authority. It was a rebellion against particular traditional authorities, not against authority as such. Far from repudiating the appeal to authority, modern philosophy has entertained seriously only one alternative to basing opinions on irrational and traditional authorities: namely, that of basing them on rational authorities.

At first this no doubt served as a response to an urgent need. Those challenging ecclesiastical and political authorities needed to be able to show that disputes could nevertheless be settled in an orderly way: that traditional, political, religious, and intellectual authorities could be displaced without producing social anarchy and intellectual chaos since they would be replaced by the authority of reason. Thus the various schools of modern philosophy whose careers we have sketched in our review of panrationalism arose to adjudicate amongst competing views by providing rational authorities to substitute for unwanted forms of traditional authority. Since then, however, this structure has been meticulously maintained. Each successive "philosophical revolution," each being a phase in the search for an adequate theory of rationality, has disclosed that the previous candidate for rational authority was unsatisfactory and has proposed a new, supposedly more satisfactory, rational authority. The church was to be replaced by intellectual intuition; intellectual intuition by sense experience; sense experience by a certain language system, and so on. The story is always the same: past philosophical error is to be given a positive explanation by attributing it to the acceptance of a false rational authority.

The traditional questions of philosophy even beg authoritarian answers. Questions like "How do you know?" "How do you justify your beliefs?" or "With what do you guarantee your opinions?" demand authoritarian answers, whether those authorities in particular cases be the Bible, the leader, the social class, the fortune teller, the Word of God, the intellectual intuition, or sense experience.

Yet what new questions would be asked philosophically if the old ones were

abandoned? Is it indeed possible to forsake the old questions? Is a nonauthoritarlan theory of rationality possible?

The importance of our historical observation is apparent, in that the questions just posed do not even arise until the authoritarian character of traditional philosophy is noticed. These new questions lead directly to the solution of the problem of the limits of rationality.

Exploring the implications of the historical observation, we may not only reject (as did Ayer) the demand for rational proofs of our rational standards. We may go further, and also abandon the demand that everything else except the standards be proved by appealing to the authority of the standards, or in some other way. Nothing gets justified. Instead of following Ayer in replacing philosophical justification by philosophical description, we may urge the philosophical criticism of standards as the main task of the philosopher. Nothing gets justified; everything gets criticized. I call this simply pancritical rationalism, instead of positing infallible intellectual authorities to guarantee positions, one may build a philosophical programme for counteracting intellectual error (ecology of rationality). Within such a programme, the traditional "How do you know?" question does not legitimately arise. And if it happens to be asked, one replies simply, "I do not know; I have no guarantees." Some of 'the positions one holds might in fact be true; but there will be no way to guarantee any of them.

A different, ecological question becomes paramount: "How can our lives and institutions be arranged so as to expose our positions, contexts, and metacontext (aims, beliefs, conjectures, decisions, policies, sources of ideas, traditional practices, etc.) whether justifiable or not - to maximum criticism, in order to counteract and eliminate as much error as possible? (See Section IV.)

The shift from a demand for comprehensive authoritative justification to a demand for comprehensive criticism is an innovation in philosophy whose importance can hardly he overemphasized. Yet the reader is no doubt blinking his eyes. It may be objected at once that any "innovation" is imaginary here: that the emphasis on the critical evaluation of competing views, far from novel, has been in the philosophical psalter since the presocratics.

Such a reaction is entirely understandable. Western justificationist philosophy of true belief does contain many theories of criticism; it pays lip service to progress; it avows the critical attitude. Yet within the polluted metacontext of justificationism, criticism can function only within the limitations set down by commitments and attachments. Western justificationist philosophy does not ecologize: it does not provide a metacontext in which avowals of criticism can be effectively pursued. An embryonic fallibilist critical metacontext may, for instance, be interpreted by -- and contained and stunted within -- a more developed justificationist metacontext. Hitherto, fallibilism has been largely confined to the level of well intentioned World 2 resolves, and has been contextualised within a justificationist World 3 institutional framework. Thus its limited success. The progress of criticism, and even the success of western science, has hitherto been in spite of the context in which they have been couched.

Thus the notion of criticism, far from being trite, appears to be one of the most unexplored and perhaps rewarding areas of philosophy. In the following, I shall bring out the crucial difference between the new idea of criticism being advocated here, and the familiar critical themes of western philosophy. To do this it will be necessary to expose the other two pollutants mentioned above.

IX. The Fusion of Justification and Criticism

We may as well begin by demanding an explanation of the historical situation just discussed. Why has an authoritarian structure been retained -- and gone unnoticed -- in modern philosophies that have been intentionally anti-authoritarian and critical in spirit?

The most important element in answering this question is this: In almost all traditional and modern philosophies -- those that have called themselves critical as well as those that have not -- the idea of criticism has been fused with that of justification. Since demands for justification are satisfied by the appeal to authority, the dilemma of ultimate commitment arises in regard to criticism within such philosophies; and authoritarianism remains inescapable. Those philosophies in which this fusion of justification and criticism occurs may be called justificational philosophies of criticism. The fusion of justification and criticism in Ayer's thought, for instance, explains why he turned to description when justification broke down. For criticism only appears as an alternative to justification after the two notions are separated.

What does it mean for justification and criticism to be fused? Justification and criticism have been fused in various ways. In most panrationalist accounts, the way to criticize a view is to see whether it can be logically derived from, i.e. "justified by," the rational criterion or authority. Thus in Hume's empiricism, to criticize a particular theory one shows that it cannot be justified by appealing to sense experience. Hume continually uses this justificationist strategy of criticism: he takes one idea after another -- the idea of God, of the soul, of memory, of other minds -- and asks whether it can be justified by being derived from sense experience, which he regards as man's only source of knowledge, or rational authority, if it can be so justified, he accepts it; if not, he either rejects it or implies that rationally it should be rejected. As he writes, "When we entertain ... any suspicion that a philosophical term is employed without any meaning or order (as is but too frequent), we need but enquire, from what impression is that supposed idea derived? And if it be impossible to assign any, this will serve to confirm our suspicion."52.

Descartes's method "for conducting the reason well and for searching for truth in the sciences" closely parallels Hume's in this regard. Ideas that cannot be reduced to clear and distinct ideas, and thus rationally justified, Descartes rejects, just as he accepts what can be so justified. For both thinkers, the way to criticize an idea is to see whether it can be rationally justified.

Another popular, although weaker, strategy of criticism also fuses justification and criticism. It is weaker in that it employs a kind of "elastic clause": what matters is not whether a position can be derived from the rational authority but whether it conflicts with it. That is, it is not irrational to hold a position that cannot be derived from, i.e., justified by, the rational authority unless its denial can be derived from the rational authority. The second strategy can be varied in subtle ways. Yet all varieties fuse justification and criticism in one way or another: to criticize a position, one must show either that it cannot be derived from, or else that it conflicts with, the rational authority, which is itself not open to criticism.

A semantic account of justification completely in line with what I have written here, and fusing justification and criticism, has recently been reported in The Journal of Symbolic Logic as capturing the "intuitive concept of justification"! Thus a sentence is justified on this account if it follows deductively from justified sentences. A sentence not justified at one time may become justified later, but once justified remains justified. The author notices how justification values assigned at present constrain future assignments. 0f those sentences which are not justified at one time, some are consistent with the justified sentences and are thus weakly unjustified, whereas other sentences are inconsistent with the justified sentences and are thus strongly unjustifiable.53

X. The Transmissibility Assumption. Lakatos

Why are justification and criticism fused in the ways described? Consider again what occurs during justificational criticism. In justificational criticism, the view to be criticized or evaluated is examined with regard to the question whether it can be derived from (justified by) the authority. Thus it is supposed to inherit logically whatever merit it possesses from the justifying authority whence it is derived. Thus if the justifying authority is true, the view being examined, if derivable from it, is true; if the justifying authority is probably, it is at least as probable; if the justifying authority is empirical, it is empirical. And so on. Without all this, here would be no justification.

This whole procedure is held in place by yet another hidden philosophical dogma: Most philosophical views take for granted that all properties, measures, and tokens of intellectual value or merit are transmitted from premisses to conclusion, in the same manner as truth, through the relationship of logical derivability or deducability.54 I call this the "transmissibility assumption." It is also called the "content condition," "the consequence condition," and other names.

Thus theories about the evaluation and criticism of competing views include (1) some more or less well-defined notion about the character of whatever properly (e.g. truth, probability. empirical character) is to be used in evaluating and criticizing, and (2) the assumption that this property, whatever its character, must be fully transmissible logically -- like truth and unlike falsity (which is retransmitted) Thus logical derivates inherit the quality and degree of merit of the premisses whence they are derived. This "common feature," writes Adolf Grunbaum, "should be an ingredient of

any theory of corroboration or rational credibility."55

This assumption is held with ferocious, if usually unexamined, tenacity -- a tenacity which can only be explained historically. The earliest attempted criteria of evaluation were criteria of truth, demarcating good ideas from bad ones coincident with the demarcation between the true and the false (section XXI). But criteria of truth proved to be either unattainable or practically inapplicable to the issues for which they were needed: and the search for criteria of truth was displaced by a search for weaker but more attainable measures. Probability (in the sense of the probability calculus) is most often used for this purpose. (Prior to the development of modern probability theory several different senses of "probability" were used in this connection; but probability in the sense of the probability calculus is now usually meant in such connections.)

Truth and probability do happen to be transmissible from premisses to conclusion through the deducibility relationship: the derivatives of a statement are true if the statement is true, and they are at least as probable as the statement whence they are derived. But most other evaluational properties are not like this. And various other properties of statements which have little if anything to do with evaluation are also not transmissible. One example of the latter would be the property of "being written in English." Truth and probability are indeed two of the very few characteristics which are transmissible. My historical conjecture is that these two concepts exerted such a determining influence over early developments within the justificationist metacontext that it is now unquestioningly assumed that other putative properties, however they might differ from truth and probability, nevertheless automatically share their logical transmissibility. Indeed, the demand for justification makes undesirable any property unable to justify its derivates by lending them its own respectability.

If the ability of truth and probability to be transmitted led, historically, to the general assumption that any indicator of merit or intellectual respectability is transmissible, the situation is now ironically also reversed: some measures, such as probability, are retained because they are logically transmissible. The transmissibility requirement itself is taken for granted. The self-reinforcing structure of the western justificationist metacontext so protects this assumption that today, when criteria of truth remain unavailable, and when, as Popper has shown, probability measures are unable to arbitrate rationally amongst competing scientific hypotheses, not to mention less precise ideas, logical transmissibility is still expected of other evaluatory properties and tokens without regard to their real logical capabilities: and it is also demanded that evaluation be made in terms of probability without regard to its evaluational capabilities. Hence the heroic yet futile attempts to retain probability as a positive evaluational property.

Take contemporary empiricist philosophy of science as an example. To comprehend it and its quandaries, it is important to notice that most theories relying on probability as an evaluatory measure also include "empirical character" as a further requirement for any acceptable theory. Legitimate statements must he reducible to something like individual "basic statements" reporting sense experience; and the logical derivates of a legitimate statement inherit not only its degree of probability but also its empirical character. Empirical character, however, is not transmissible. From every basic empirical statement both nonempirical metaphysical statements and all tautologies follow logically -- not to mention the problem of induction: that legitimate universal scientific hypotheses cannot be reduced to truth functions of a finite class of basic observation statements. Out of this conflict (between the transmissibility of probability and the non-transmissibility of empirical character) are produced several of the well-known "paradoxes" of induction and confirmation. When transmissible probability is mixed with the non-transmissible property of "empirical character," the results are indeed bizarre. Non-empirical consequences of empirical statements inherit the probability of the original empirical statement, consequently becoming respectable from the point of view of the probability standard. Yet, lacking empirical character, they remain disreputable from that standpoint. If meaning criteria are added as still further criteria of respectability, still more anomalies can appear.

Yet it is simply not necessary for criticism to be bound by this assumption. Alternative approaches to evaluation and criticism are possible which do not only not contain the transmissibility assumption but which are incompatible with it. A particularly appropriate example here is Popper's theory of testability or corroborability. It provides us with an example of nonjustifcational evaluation and criticism in broad terms (we shall get to some specific examples in sections XIII to XVII).

Popper's theory of testability assesses not the degree to which a theory is probabilified or confirmed or justified, but the degree to which it is testable or corroborable. In Conjectures and Refutations Popper describes this as a criterion of progress, expressing relative potential satisfactoriness of a theory, applicable in evaluating a theory even before that theory has been tested empirically (p. 217). (It should be noted that Popper does not describe this as a criterion of acceptability, since acceptance and acceptability play a marginal role in his methodology: see sections XIV and XXI.)

The measure of degree of testability, unlike truth and probability, is however not logically transmissible. Quite the contrary.

The difference between this theory and those referred to above can be defined thus. Whereas evaluational measures like probability and (Carnapian) degree of confirmation are transmitted in the same direction as truth, degree of testability, which is a measure of content, is, like falsity, retransmitted from conclusion to premisses. The difference between testability theory and various probabilistic theories of confirmation may be defined by reference to this irreducible difference in the ways the two properties are logically transmitted.56 (Some other evaluational measures and properties used in Popper's methodology - such as explanatory power are also not transmitted.)

The point here may be explained as follows. If Popper's theory of testability did share the transmissibility assumption, then any consequence of a hypothesis would have to be as highly testable as the original. But no such thing happens: since a hypothesis is testable (in the syntactical sense) by its consequents, the hypothesis must possess at least as high a degree of testability (and thus corroborability) as any of its consequents. But it may (and if logically stronger almost invariably will) possess a higher degree of testability. If a hypothesis can possess a higher degree of testability than its consequents, then a consequent does not inherit this particular property through the deducibility relationship. A theory does not bequeath its degree of testability to those theories it entails, its necessary conditions, which traditional accounts would represent it as justifying.

One paper in the Schilpp volume, the late Imre Lakatos's "Popper on Demarcation and Induction," touches on - and badly muddles - this issue, claiming (p. 262) that in 1971 Popper "gave up his long held and tenaciously defended doctrine that the degree of corroboration of an unrefuted theory cannot be smaller than the degree of corroboration of any of its consequences." Lakatos makes a triple error here (which Popper does not notice in his reply). First, Lakatos confuses corroborability and corroboration. Popper denies transmissibility assumption for degree of corroborability, not for degree of corroboration. For Popper, it is the degree of corroborability of a theory that cannot be smaller than that of its consequences.57 What is at stake here is a condition of content. Degree of corroborability is a measure of content; degree of corroboration is related to content but is not a measure thereof. Second, Popper believes that a consequence of a theory must only in so far as it is regarded as a part of the theory inherit the degree of corroboration of the theory (Objective Knowledge), p. 20). Thus, by contrast to the statement by Lakatos just quoted, Popper's view is that the degree of corroboration of a consequence of an unrefuted theory cannot be smaller than the degree of corroboration of the theory itself provided that the consequence is regarded as part of the theory. Lakatos had it the other way around, and omitted the essential proviso. One could of course challenge the proviso, and the assumption on which it is based: but Lakatos does not do this. Thirdly, in section 33 of his unpublished Postscripts to the Logic of Scientific Discovery, Popper explains in detail that, where b follows from a, there are some circumstances where C(a,d)> C(b,d) and other circumstances where C(a,d) <C(b,d). (C refers here to corroborability or degree of testability.) Fourthly, there is in fact no change of doctrine on this point in the paper to which Lakatos refers: Popper's "Conjectural Knowledge: My Solution of the Problem of Induction."58

If all measures of intellectual respectability resembled truth and probability in being transmissible, all criticism. would perforce be justificational. That degree of corroborability and testabiblity are not transmissible shows that not all criticism need be justificational. Hence a nonjustificational, nonauthoritarian theory of knowledge and rationality is possible. A fallibilist metacontext is possible.

The features of traditional philosophy discussed above - the authoritarian metacontext in which its questions are embedded, and the assumptions that are largely responsible for this structuring - deserve detailed exploration, as does the ecological question of how to construct critical institutions to nurture and inspire inquiry. In their present preliminary form, the matters discussed suffice to suggest the character of an account of rationality that escapes the dilemma of ultimate commitment and avoids both scepticism and fideism.

XI. Pancritical Rationalism

A new concept of rationality is implicit in the nonjustificational approach. Within this new metacontext, a rationalist may be characterised as one who holds all his positions, including his standards, goals, decisions, etc., and his basic philosophical position itself, open to criticism; one who protects nothing from criticism by justifying it irrationally; one who is committed, attached, addicted, to no position, context, or metacontext; one who is willing to entertain any position, but who holds (tentatively) only those positions which have been subjected to and have survived intense criticism. This characterization, which I call pancritical rationalism, shares with panrationalism its comprehensive aims, but not its justificationism.

Pancritical rationalism, also differs from critical rationalism, wherein the rationalist accepts the rational unjustifiability of his position but goes on to justify it in an irrational commitment to principles that art not themselves open to criticism. The pancritical rationalist does not justify at all.

If rationality lay in justification, eventual irrational justification or commitment-attachment would be inescapable. However, if rationality lies in criticism, and if everything - including the fallibilist metacontext itself - can be subjected to criticism and continuing tests without leading to infinite regress, circularity, the need to justify, or other such difficulty, then rationality is in this sense unlimited. If all justification - rational as well as irrational – is abandoned, there is no need to justify irrationally a position that is rationally unjustifiable. The position may be held rationally regardless of justification, provided it can be, and is, held open to criticism and survives it. The question of how well a position is justified differs utterly from the question of how criticizable it is, and how well it is criticized.

The general separation of justification and criticism can of course be extended to the examination of the rationalist position itself. Under previous conceptions of rationalism, the rationalist position, being unjustifiable, was itself not rational. But pancritical rationalism satisfies its own requirements: without any contradiction or other difficulty the very practice of critical argument can be criticized.

A pancritical rationalist who is uncommitted to the view that his position is correct, could be argued out of it (See section XVII). Continued subjection to criticism of his allegiance to rationality is explicitly part of his rationalism. Fideistic irrationalists (such as Barth), have, by contrast, argued that, even without actually abandoning allegiance to the object of commitment, merely subjecting the allegiance to criticism is to forsake the allegiance.

This spells defeat for the tu quoque argument that supports the sceptical and fideistic claims. The case for subjective commitment, and the rational excuse for it, rest on the argument that rationality is so limited logically that such commitment is inescapable. For the pancritical rationalist there is no such limitation. If he accuses his opponent of protecting some position from criticism through irrational commitment to it, he is not open to the charge that he is similarly committed. Since the tu quoque argument cannot be used against pancritical rationalism, criticism of commitments no longer boomerangs.

XII. Degree of Problematicality. Bernays, Musgrave

To avoid misunderstanding, several remarks are in order.

First, to say that a rationalist need not commit himself even to rationalism is not to claim that he or she will not or should not be able to make agreements or to have strong convictions on which he or she is prepared to act. One can be convinced of something without being committed (in the stronger sense) to its truth. As conceived here, a rationalist can, while eschewing commitments and attachments, retain both the courage of his convictions and the courage to go on attacking them. Yet, as Nietzsche of course says: 'A very popular error: having the courage of one's convictions; rather it is a matter of having the courage for an attack on one's convictions."59

Second, a pancritical rationalist, like everyone else, holds countless unexamined presuppositions, many of which may be false. His rationality consists in searching these out and submitting them to critical consideration when he discovers them or when they are pointed out to him. Charles Darwin's example is a good one here. 'I had," he wrote, "during many years followed a golden rule, namely, that whenever a published fact, a new observation or thought came across me, which was opposed to my general results, to make a memorandum of it without fail and at once; for I had found by experience that such facts and thoughts were far more apt to escape from the memory than favourable ones. Owing to this habit, very few objections were raised against my views which I had not at least noticed and attempted to answer."60

When one position is subjected to criticism, others have to be taken for granted ... including those with which the criticism is being carried out. The latter are used in criticism not because they are justified or beyond criticism, but because they are at present unproblematical. These are, in that sense alone, beyond criticism. We stop criticizing -- temporarily -- not when we reach uncriticizable authorities, but when we reach positions against which we can find no criticisms. If criticisms of these are raised later, the critical process thus continues. This is another way of saying that there is no theoretical limit to criticizability -- and to rationality.

One will not begin to question statements that seem to be true simply in the face of arguments that it is, say, logically possible that they are not! In that sense, one calls a halt to criticism. One will, however, begin to question this "halting place" when a concrete argument is produced to challenge it -- when an argument is produced that renders it problematical. In regard to standard sceptical arguments, all positions are equally problematical -- equally indefensible -- and equally defective because equally unjustifiable. In order to compare positions intelligently we need a theory of criticism in terms of which positions differ in problematicality. This becomes possible, once the aim of justification, which is responsible for equality of problematicality, is abandoned.

The fact that most of a person's positions are not being criticized at any particular time thus does not mean that any of them has to be beyond criticism all the time. Nor does it mean that the allegiance to criticism itself may not come up for critical review. Such a willingness unattachedly to hold open to revision even those positions supposedly most surely to be true is part of the spirit of Pancritical rationalism,

Pancritical rationalism is hence compatible with one kind of relativism. The survival of a position is relative to its success in weathering serious criticism. A position that survives at one time may be refuted later. This kind of relativism -- due to the fact that we are not gods, are ignorant, lack imagination, are pervasively fallible -- is harmless. It is an example of the way that learning proceeds by. Trial and error -- by guessing and trying to criticize guesses: the making and destroying of theories is part of the evolutionary process.

Third, it should be remembered that our problem is a logical one, and that the point being made here is logical too. The problem of rationality lay in the fact that, for logical reasons, the attempt to justify everything (or to criticize everything through justification) led to infinite regress or dogmatism. Nothing in logic prevents us from holding everything open to nonjustificational criticism. To do so does not, for instance; lead to infinite regress.

There may, of course, be other non-logical considerations which lead one to grant that it would be pointless to hold some particular view as being open to criticism. it would, for instance, be a bit silly for me to maintain that I held some statements that I might make e.g., "I am over two years old" open to criticism and revision.

Yet the fact that some statements are in some sense like this "beyond criticism" is irrelevant to our problems of relativism, fideism and scepticism. I may in fact hold some such views as beyond criticism; but I do not have to do so logically: I do not have to be dogmatic about any of these matters. In holding everything open to criticism, I of course do not deny that there are true statements and valid arguments; nor do I maintain that for every proposition there must exist some sound argument against it! Holding such statements as beyond criticism in a practical sense has nothing to do with stemming an infinite regress. What is needed for the effort to state a consistent theory of rationality, is to show that it is logically possible (without leading to infinite regress, vicious circle, or other logical difficulty) to hold such statements open to criticism. When this is done, no tu quoque can be mounted.61

Two other essays in the Schilpp volume deal explicitly with some of the issues of rationality that I have been discussing in scetions V-XII.

These papers, by Paul Bernays and A.E. Musgrave, are essentially no more than reviews of some of Popper's views. Bernays's paper -- notwithstanding Popper's effusive response to it -- is a disconnected and vague series of remarks on the "concept" of rationality.

Musgrave's paper appears at first to be more serious and relevant, in that it aims to show how the problem of the limits of rationality is to be solved. His answer, however, turns out to be particularly odd. First, he avoids mentioning pertinent background literature on the subject (e.g. Agassi, Albert, Bartley, Jarvie, Settle, Watkins). Then, after reporting Popper's distinction between logical/objective and psychological/subjective considerations in theory construction and evaluation (with which I have no quarrel), Musgrave announces-- but fails in any way to show -- that this distinction is sufficient to solve the problems of relativism. Finally, he misses the point of the entire discussion (p. 584) by conceding wrongly that scientists make justified choices among theories, and that a person's present choice of a theory to guide his practical actions needs justification.

XIII. Four Kinds of Non-Justificational Criticism

In the preceding section I have argued the possibility of non-justificational criticism, and suggested its importance in solving the problem of the limits of rationality. Not only is this sort of criticism possible; it turns out that most ordinary criticism carried on in scientific and intellectual practice (as opposed to theory about that practice, most of which is wholly justificationist) is quite nonjustificational. In the following sections I shall briefly review four important sorts of nonjustificational criticism, all of which play an important role in Popper's methodology.

One wants, in examining and criticizing positions and contexts, to use appropriate critical considerations. Sometimes this will mean checking or testing a theory against experience; sometimes, comparing it against other theories; sometimes it will mean pushing it against whatever problems it was intended to solve; and sometimes it will mean testing it logically for consistency. So, we shall look at nonjustificational criticism using (1) experience; (2) theories; (3) problems; and (4) logic. None of these checks is unproblematical; some of them give rise to particularly interesting problems.

XIV. The Check of Experience: Basic Statements,Fries's Trilemma. A.J. Ayer.

"You are deducing from things that you have seen. Nothing can be so misleading as observation."

Hercule Poirot, in

Agathe Christie: Death in the Air. 62

The first means of criticism just cited -- the check of empirical experience -- is crucial. In view of the goal of constructing accurate representations, theories have to be tested against accurate observation reports of the way the world is. This particular check has been extensively discussed within the tradition of logical empiricism, and also within the Popper school. The issue is joined again in the Schilpp volume, in the debate between Popper and Sir Alfred Ayer, as posed in Ayer's contribution, "Truth, Verification and Verisimilitude," and in Popper's reply. In his paper, Ayer disputes Popper's theory of "basic statements," as developed in Chapter V, "The Problem of the Empirical Basis," of The Logic of Scientific Discovery (and amplified in Conjectures and Refutations , pp. 386-8).

Popper's discussion of the empirical basis opens with, and revolves around; J.F. Fries's famous "trilemma," as presented in Neue oder anthropologische Kritik der Vernunf (1828-31). Fries had argued that, to avoid dogmatism, one must he able to justify the statements of science. But statements can he justified, logically, only by statements. This leads to infinite regress. To this the only alternative seemed to be what Popper calls psychologism, 64 the doctrine that statements can be justified by the "immediate knowledge" of perceptual experience as well as by other statements. Presented with this trilemma -- dogmatism vs. infinite regress vs. psychologism -- Fries had opted for psychologism.

In The Logic of Scientific Discovery Popper presented what he conceived as a resolution to Fries's trilemma. To summarize Popper's resolution is difficult. But put in a nutshell, and oversimplified, his view is that the universal statements of science are tested against "basic statements" asserting that an observable event is occurring in a certain region of space and time. Such basic statements must satisfy formal and material requirements. Formally, they must be singular existential statements; must be able to contradict a universal statement; and must be underivable from a universal statement alone without initial conditions. Materially, basic statements must be testable intersubjectively by observation.

Concerned to present science as an objective structure, Popper emphasizes that the material requirement of intersubjective testability through observation should not be taken psychologistically: one could as well omit any reference to observation, and refer instead to "an event involving position and movement of macroscopic physical bodies." He stresses that basic statements are not statements about our experiences and that it would be wrong to resort, as some positivists did, to 'Protokollsatze." The reason is simple: statements about experience are often of reduced testability. Thus, "I see that the table here is white," being about me, is less testable than "This table is white," which is about the table.

How does this resolve Fries's trilemma? Popper confronts each horn of the trilemma in turn. First, as to dogmatism. he states that we must, through a free act, decide by agreement to accept basic statements as satisfactory and sufficiently tested. Such a decision is necessary: for unless we reach a statement which we decide to accept, our test will have led nowhere. Thus, "It is decisions which settle the fate of theories" (p. 108). There is a similarity to conventionalism here, he concedes, but also a difference in that the conventionalist tends to accept universal, not singular, statements by agreement and does so for reasons of aesthetics, not of testability. These basic statements are, Popper also concedes, dogmas, in so far as we may desist from justifying them.

But, he cautions, the dogmatism here is innocuous since basic statements can always, if need be, be tested further.

Second, as to infinite regress, Popper concedes that the chain of deduction in terms of which universal statements are tested against testable basic statements is indeed infinite. But this, he contends, is also innocuous since there is no question of proving anything.

Finally, as to psychologism, Popper concedes that basic statements and sense experience are related in that sense experience may motivate a decision to accept a basic statement, but he emphasizes that such experience in no way justifies the basic statement.

Ayer disputes this account, taking a position which Popper would classify as psychologistic. Rejecting the view that statements can be justified only by one another, Ayer claims that basic statements are directly justified by experience. Experience, he states, is the source of claims to knowledge, and is authoritative, providing the "right to be sure" of one's knowledge claims. "If these observations give us no authority for accepting any basic statements," Ayer writes (p. 689), "they do not constitute a test of anything, so that there is nothing to be gained by examining the facts; if they do give us this authority, our claims to know basic statements are validated by their sources." As appears to be often the case in his work, Ayer wants to have it both ways, stating that these authorities are fallible and providing no explanation of how a "fallible authority" might function in such a justificational setting. (The Roman Catholic Church, in appreciating that their authority -- the Pope -- would have to be infallible, shows a better grasp of the logic of the situation than does the sometime Wykeham Professor of Logic.)

Ayer's position here, in and of itself, is uninteresting: it is stated flatly, not argued, and ignores objections to such a position stated by Popper and others. I agree with Popper's observation, in his reply, that Ayer's discussion misses the point. Yet the further details of Popper's reply to Ayer were so odd as to lead me to reread his original chapter on basic statements, and to notice that the entire discussion is a bit loose. I will cite only one example of what I mean, and then go on to reconstruct the problem.

Popper contends (p. 1110) that Ayer's criticism is based completely upon the mistaken assumption that every decision or convention must be arbitrary. Yet on the next page (p. 1111), he concedes that the sorts of decisions he is discussing may be called arbitrary … but are far from being "totally arbitrary." Three pages later (p. 1114), he states that the conventional or decisional element in the acceptance or rejection of a proposition generally involves "no element of arbitrariness at all." Two sentences later, he nonetheless states that decisions about the acceptance of basic statements are "somewhat arbitrary." In Logik der Forschung (p. 74), Popper had written that "die Entscheidungen uber die Basissatze nicht durch unsere Erlebnisse 'begrundet' werden, sondern, logisch betrachtet, willkurlichie Festsetzungen sind." That is. "decisions about basic statements are not 'justified' by our experience, but are rather, logically considered, arbitrary stipulations."

While I do not want to quarrel over words, and while it would no doubt be possible to reconstruct something of what is intended here, Popper is clearly using the word "arbitrary," which is after all a technical term, in an arbitrary way.

Yet there is no need here for any discussion of decisions or conventions at all -- let alone of arbitrary ones. The source of the problem here is the same one that we saw above (section VII) in Popper's first formulation of "critical rationalism," and also in Die beiden Grundprobleme der Erkenntnistheorie, where -- appealing to Kant and the primacy of practical reason -- Popper states that the final basis of all knowing is to be sought in an act of free postulation that cannot be justified rationally. This free and arbitrary postulating and deciding is no more needed with basic statements than it is with rationalism; and the fact that neither basic statements nor rationalism can be justified ought to be irrelevant from Popper's own wider point of view, provided that they are both criticizable.

All of this can be avoided from the start simply by eliminating Popper's requirements relating to decisions and agreements. To show this, let us first return to the problem itself, starting again with Fries's trilemma.

First, Popper's "Friesian" statement of the problem is unsatisfactory. For Fries's problem is no true trilemma, but is a dilemma-engendering dilemma.

If you are proving one proposition by another, either you go on forever - or you stop.

If you go on forever, there is no proof, only infinite regress.

Therefore there is no point to going on forever.

So suppose you stop. But can you stop? And if so, how?

There are a number of possibilities. It is a large number: I would guess that it is infinite. E.g., 1) You can just stop arbitrarily wherever you choose or decide to do so. (You can designate your stopping point as being those principles in terms of which you choose to lead your life.) In this case, you can either stick to your choice come what may or be prepared to give it up.

If you stick to your choice come what may, then you have a dogma. If you are prepared to give up your choice, to reconsider it, then you can do that under a number of different circumstances. E.g.,

a) You may revise your choice arbitrarily (and of course the same questions can be asked of the new choice: e.g., will you stick to your revised choice come what may or are you prepared to give it up?)

b) You may revise your choice under the weight of reason. (In which case, where did the reason come from, in case the first choice defined "reason" for you?)

2. You can stop at a self-evident proposition.

3. You can stop at a non-self-evident proposition -- and then justify it non-propositionally.

But there are a number of different "non-propositional" ways of justifying a proposition. E.g.,

a) By appealing to "immediate" experience. Here is where psychologism comes in at last!

b) By appealing to a decision -- which may in turn be arbitrary or otherwise, and revisable or otherwise, or "motivated" by experience or otherwise. In which cases ... And so on.

All of this is a toy, and, can be complicated almost ad libitum. The point here is not to play with the toy, but to see that one needs to pose the problem more clearly to allow a clear solution.

To solve the problem, Popper's talk of conventions, decisions, acceptance, agreement, and justification -- even his talk about the "empirical basis" of science -- is not fully to the point; it is indeed superfluous. There is not anything "basic" about "basic statements." This way of conceiving and presenting the problem is another remnant of that justificationism which Popper's own thought enables one to transcend

Popper's key correct insight here is that although the process of testing is potentially infinite, no infinite regress arises "since . . . there is no question of trying to prove any statement." (p. 105). This is so provided that the testing is indeed non-justificational. The moment that one presupposes that one can, with test statements, prove a theory to be false, or justify the contention that it is false, then one does indeed get an infinite regress. The test statements are intended here to be hypothetical, and criticisable and revisable, just like everything else in the system: there is no justification, no proof, no fixed point anywhere in the system.

Very important amongst those critical statements brought to bear in criticism of a theory are, as Popper realizes, those which report the occurrence of events involving position and movement of macroscopic physical bodies (what Popper calls "basic statements"). If such reports are incompatible with a theory, then the theory is false relative to them, and they are false relative to the theory. There is no question of theory proving reports wrong, or reports proving theory wrong. Both could be wrong: neither is "basic".

Suppose one wishes to test a particular theory. One will determine what sorts of events involving position and movement of macroscopic physical bodies would be incompatible with the theory, and will then set up an experimental arrangement to try to produce such events. One will then make or gather reports concerning the test. And, in order to maximise criticism and provide mutual testing of test reports, one will invite or permit more than one reporter. Note: one does not get these additional tests, pace Popper, to elicit an agreement.

Suppose these reports go against the theory. Stepping back from this situation, regarding the objective structure of circumstances and statements, one then makes a (metalinguistic, as it were) report on what has happened: that time theory is now problematic in that it is false relative to the test reports, and that the test reports are themselves, as may be at the moment, unproblematic.65

If the test reports had conflicted amongst themselves -- if some of them had gone against the theory and some failed to go against it – one would have a different situation: then one would more accurately report that although the theory was rendered problematical by some of the test results, the test was itself also problematical in having produced several conflicting reports.

One contributes nothing to this situation by adding, as Popper does, a requirement that one needs to decide by agreement which reports to accept. Moreover, he is wrong in claiming that the test will have led nowhere without an agreement: it will have led either to problematical or to unproblematical basic statements. These reports do not need to be backed up by "decisions" and "acceptances." One does not need to accept any of them, or even to consider them (however tentatively) to be true - only to give an account of what is happening, and to state how theory and report stand, hypothetically, in relation to one another. One may go on to conjecture about which of the reports are accurate: but this is a conjecture, not a decision, and may itself be tested accordingly.66 Hence a theory may be provisionally and conjecturally rejected because it conflicts with some less problematic view. Any theory is refuted only relative to critical arguments incompatible with it, which are themselves open to criticism by the testing of their own consequences. These in turn are criticisable forever.

This process in which one steps outside the positionality of the theory to comment on the state of examination of the theory, treating the theory as an object. Not as one's point of view, and oneself coming from beyond the theory - is virtually impossible from within justificationist metacontext, in which one is always, willy nilly, imprisoned within the bounds of one's position.

XV. The Check of Scientific Theory

A second type of nonjustificational criticism, on which I want to touch very briefly, comes not from the clash of an idea with observational experience, but from the clash of an idea with a scientific theory.

For it is possible for a claim to be compatible with all sense observation reports and yet in conflict with some well-tested scientific hypothesis in terms of which it may be criticized. "Every substance has a solvent" is irrefutable in principle in the sense that no empirical refutation is possible: it is compatible with "Gold has never been observed to dissolve." But it is incompatible with the theory that "Gold is insoluble."

There is some point to arguing, as J.O. Wisdom does, that such statements -- although irrefutable and nonscientific in Popper's sense -- should nonetheless be classified as scientific. For a number of theories of this character -- Wisdom calls them "theory-refutable" (as opposed to "observation-refutable") -- play an important role in science. Some other examples are: "All bodily changes are due to physical causes," "All mental changes are due to physiological causes," and "Energy occurs in all possible quantities." Although these theories conflict with no possible basic statement, they may conflict with testable scientific hypotheses. For instance, Schrodinger's discovery of the wave equation, involving as it does discontinuities as consequences refutes the (observation-irrefutable) theory that energy occurs in all possible quantities: i.e., is continuous.67

I agree with Wisdom that such theory-refutable theories play an important role in science and that they conflict with Popper's demarcation between science and nonscience (although not with the spirit of his demarcation). Where such theories are brought into clash with scientific theories, and thus are criticizable in terms of these scientific theories, one must not assume too readily, however, that the irrefutable statement is wrong, and the scientific hypothesis right. Since no scientific theory can ever be fully verified by experience it remains possible that any particular such hypothesis may be falsified by experience at some later date. Thus, in the case of a conflict between a scientific theory and an irrefutable statement, the latter could in principle be correct.

However, I disagree with Wisdom's contention that "this kind of refutation is hypothetical in a way that refutation by observation is not, for the refuting theory, though tested and confirmed, may later be falsified; then the programme it had refuted becomes 'derefuted'." This is a misunderstanding of the situation that obtains with observation-refutation. For as we saw in the last section, observation-refutations are and remain quite hypothetical; and theories refuted by observations may also be "derefuted" if the observation is revised in further testing.

XVI.

The Check of the Problem: Goodman's Paradox.

Quine, Settle, Maxwell, Levison

Empirical testability, in and of itself, is hardly sufficient in criticism, either in or out of science. Even within science, considerations other than testability must be brought to bear, and can play a more important role in deciding a conflict among positions. The particular consideration I have in mind here I call "the check of the problem" which, as we shall see, also has nothing to do with justification. The importance of this check, and the significance of neglecting it, is illustrated by Nelson Goodman's famous "paradox" about "grue" emeralds.

It is particularly appropriate to consider Goodman's paradox in a review of the Schilpp volume. For it lurks troublingly in the background of the discussion. Although Goodman's paradox is often treated together with Hempel's paradoxes of confirmation, the contributors to the Schilpp volume clearly see matters differently. Thus W. V. Quine, in his incisive short essay. "On Popper's Negative Methodology," concedes that Popper's methodology avoids Hempel's paradoxes. Yet both Quine and another contributor to the volume, T.W. Settle, mention Goodman's paradox as a difficulty unresolved by Popper's work. In another paper, Grover Maxwell's “Corroboration without Demarcation," it is suggested briefly that Goodman's theory can be resolved only within Maxwell's own un-Popperian framework. And Arnold Levison, in "Popper, Hume, and the Traditional Problem of Induction," cites Goodman's paradox as a direct challenge to Popper suggesting that even if Popper could solve the traditional problems of induction (as Levison thinks he cannot) he would fail to solve Goodman's paradox. Popper, in his "Replies to My Critics," fails to answer these particular challenges. Meanwhile, Goodman himself, in his preface to the new edition (1973) of Fact, Fiction and Forecast (p. vii), explicitly denies (without argument) that a Popperian approach can resolve his problems.68

What is Goodman's paradox? In his lectures at the University of London, published as Fact, Fiction and Forecast, Goodman describes a hypothetical situation in which all emeralds examined before a certain time t are green. He introduces a special artificial predicate, "grue," which applies to all things examined before t just in case they are green but to other things just in case they are blue. Thus, at time t we have, for each evidence statement asserting that a particular emerald is green, a parallel evidence statement asserting that that emerald is grue. Whatever has confirmed the statement that all emeralds are green will also confirm the statement that all emeralds are grue. This is not really a paradox, but a variant of the possibility, known from the works of Hume, that the next observation may be different; it is also a vivid illustration that the same "evidence" may be compatible with different and incompatible theories. Goodman claims that "we are well aware which of the two incompatible predictions is genuinely confirmed," but searches unsuccessfully for a definition of the difference between "lawlike" or "projectible" and "non-lawlike" or "non-projectible" properties which could provide a basis for sorting the correct statement from the incorrect, to demarcate the good statement from the bad.

An initial Popperian approach to Goodman's paradox could run as follows. To attempt a decision between the two statements on the basis of confirmation theory, which would involve an inductive logic, would be impossible because an inductive logic is impossible; to attempt a decision on the basis of a definition of "lawlike" or "projectible" character of predicates would be Cartesian or "essentialistic," not empirical, despite the fact that an empirical study of the actual history of the predicates in the language may be required to discover which is the "better entrenched". Rather what is needed is a crucial experiment between these two theories. But no crucial experiment of an empirical character is possible before time t. Hence one must wait until time t.

Goodman would hardly be satisfied with such an approach, even if he accepted Popper's assumption about the impossibility of an inductive logic. Rather, Goodman wants some explanation of why it happens that even before time t the claim that "All emeralds are grue" is not taken seriously. ("We are well aware of which of the two incompatible predictions is genuinely confirmed ..."). Such an explanation is, it seems to me, provided readily, and in a way fully compatible with a Popperian approach. The reason the statement that all emeralds are grue is not taken seriously in science has little to do with the evidence at hand. It is not taken seriously because there exists no problem in mineralogy to which it responds. Whether a counter-intuitive hypothesis is taken seriously depends not simply on an analysis of the predicates used in the hypothesis, not on psychological dispositions, not on past experience, not on the degree to which certain predicates are "entrenched", but rather on the question whether the hypothesis is directed to an existing intellectual problem, scientific or otherwise. Theories arise not in a vacuum, but in situations where some problem exists, and where some attempt is being made to solve it. The grue hypothesis appears odd, implausible, unacceptable not because of the peculiarity of the properties used, which are no more so than the ideas of mass, force, and simultaneity as they occur in Einstein's theories - "peculiarity" being defined by reference to the properties of late nineteenth-century Newtonian theory. Take the idea of mass: in Newtonian theory mass is a property of an object, independent of its behaviour in coordinate systems. Whereas in relativity theory, mass is a relation, involving relative velocities, between an object and a coordinate system.

One may add another apparently counter-intuitive hypothesis about emeralds to indicate further the context and problem-dependence of theories. Suppose one says: "All emeralds are young." The proverbial man in the street will hardly pause to listen further. And the sophisticated philosopher, if he pauses, will probably do so with the surprised expectation that he is actually witnessing a "category mistake" like "Socrates is identical". It may take an alert mineralogist to recognize that "All emeralds are young" is true. If one thinks comparatively in terms of the geological age of stones, emeralds are young and diamonds are old.69 But this background is needed to respond intelligently to, to assess, such a statement.

In sum, hypotheses like "All emeralds are grue" are not entertained seriously because no problem is given towards the solution of which both "All emeralds are green" and "All emeralds are grue" are directed. On the other hand, Einstein's theory - which is as queer and unintuitive – is taken seriously because it was posed as a solution to certain problems which Newtonian theory appeared unable to solve. Einstein's general theory of relativity, incidentally, was taken seriously before its "time t." Several years passed before it could be tested in the famous crucial experiment conducted by Sir Arthur Eddington at Principe, off West Africa, on 29 May 1919. That it could be tested so soon was due to a quite exceptional coincidence. Writing of his famous experiment, Eddington muses:

"In a superstitious age a natural philosopher wishing to perform an important experiment would consult an astrologer to ascertain an auspicious moment for the trial. With better reason, an astronomer today consulting the stars would announce that the most favourable day of the year for weighing light is May 29. The reason is that the sun in its annual journey round the ecliptic goes through fields of stars of varying richness, but on May 29 it is in the midst of a quite exceptional patch of bright stars - part of the Hyades - by far the best star-field encountered. Now if this problem had been put forward at some other period of history, it might have been necessary to wait some thousands of years for a total eclipse of the sun to happen on the lucky date. But by strange good fortune an eclipse did happen on May 29, 1919."70

A further aspect to Goodman's paradox, going beyond the utter neglect by Goodman of the problem situation, should be mentioned briefly here. It is commonly the case in scientific progress that the new theory is corroborated by the evidence of the older theory and yet also contradicts the older theory. Thus Einstein's theory, which contradicts Newton's, is from the outset nonetheless richly corroborated: for all evidence for Newton's theory also corroborates Einstein's: i.e., it corroborates a contradictory theory.71 The same happened again with Friedmann's improvements on Einstein's cosmology. Friedmann's account contradicted Einstein's: and yet it was corroborated by all corroborations of Newton's and Einstein's theories. Such can always happen.

But consider: For any given theory, one that contradicts it and is as well confirmed can be routinely invented. Yet the sum of the probabilities of pairwise exclusive theories must be equal to or less than 1. Thus if we have a method or routine of drawing up an infinite sequence of such corroborated and competing theories (i.e., any two of which contradict one another), almost all would have to have probabilities as near as might he wished to zero. As Popper has pointed out to me in correspondence, Goodman's approach is just such a method or routine. It is also an utterly trivial, and even silly, exercise, unconnected with any scientific problem, and permeated by scholasticism and justificationism. (Incidentally, such an approach also shows, into the bargain as it were, that probabilistic induction is impossible, and that support or corroboration cannot be a probability. This result, shown trivially through Goodman's approach, is of course quite contrary to Goodman's intentions.72

My solution of Goodman's paradox is intended to illustrate the importance of the third nonjustificational critical check cited above, the "check of the problem," in evaluating conjectures; and how it may often be more important than the check of sense experience reports. Although some philosophers have emphasized the importance of criticizing philosophical positions by comparing them historically against the problems they are intended to solve, this is a much-neglected matter amongst philosophers, and even amongst some scientists, who seem all too willing to chatter on long after the problem has been forgotten by all parties to the discussion.

In addition to using this check just now in solving Goodman's paradox, I used it implicitly still earlier. I argued that panrationalism, in failing to solve its problem, led to intellectual crisis; that critical rationalism avoided crisis at the expense of ignoring the problem; and that pancritical rationalism solved the original problem.

The check of the problem may nonetheless seem vague. One lesson for this is that contemporary philosophers tend to devalue the importance of the history of philosophy. To tell which philosophical view best solves important philosophical problems, it is necessary to go to historical texts and examine what were those problems and how they have changed. The historical study of philosophical problems, of the way they have been approached, and of the way in which they themselves have altered - a "natural history of philosophies," as it were - becomes of crucial ecological importance for even the most analytical of philosophers once a nonjustificational approach to criticism is adopted.

XVII. The Check of Logic. Lejewski

By far the most controversial critical check is the fourth: logic.

Important in any discussion of the dispensability, revisability, criticisability of logic are the views of W.V. Quine, as expressed in his treatment of the traditional distinction between analytic and synthetic. According to Quine, when a body of belief is brought to the test of criticism, any part of it may be revised. No segment of it, such as the set of analytically true statements, including logic, is so insulated that we could say in advance that "the mistake could not be here."

CONTINUED IN THIS FILE