The Romantic Rationalist

Apart from his celebrated writings on the “open society” and its enemies, Karl Popper is chiefly known as a logician of science who has denied that science employs induction, and who has claimed that what demarcates science from nonscience, in particular metaphysics, is that scientists seek the truth by vigorously trying to falsify their theories. This has become one of the most celebrated and controversial views of science to have been put forward during this century. In a previous article, I noted that vigorous objections have been raised to Popper’s view, most recently in the book under review by Anthony O’Hear, and, as we shall see, there are strong reasons for questioning Popper’s thesis.¹

First, however, a common misunderstanding of Popper’s work should be clarified. His preoccupation with the logic of scientific method and his early association with some members of the Vienna Circle have frequently led readers to link him with logical positivism and a contempt for metaphysics. What is less well known, and perhaps surprising, however, is that Popper has devoted a number of years to outlining a metaphysics of his own, one drawing on evolutionary theory, that depicts the continuity of method between men and other organisms and more generally articulating the place of man and his intellectual products in nature. This vision, he reminds us, is not a scientific theory itself: it is not falsifiable. Popper regards his own “metaphysical” theory and his evolutionary epistemology as “conjectural” in character: while such conjectures are not empirically testable, he claims, they may, like other nontestable and nondemonstrable theories, such as realism and idealism, be helpful to us. Furthermore, they might be arguable.

What is this “vision”? Popper conjectures that “what is true in logic is true in psychology.” Induction is logically invalid, does not “exist” in logic and so no one (and no animal) has ever performed one. All organisms, according to Popper’s phrase, “from the amoeba to Einstein,” are problem solvers who use the method of trial and error. In nature, as Darwin taught, the plural forms of life evolve from a small number of simple forms by virtue of the mechanisms of heredity, variation through mutation, and natural selection. Of course, the lower organisms lack language and cannot formulate their hypotheses and guesses, but even they do something similar to what we do: they carry out “trial” solutions to problems of adaptation and adjustment, and the “errors” in these trials are eliminated by natural selection. As we solve problems of bridge building or scientific explanation, then, so spiders “solve” problems of where to build their webs and bacteria “solve” problems of overcoming antibiotics. Most organisms other than man incorporate the “solutions” to the problems confronted by their predecessors in their very anatomical design; they die off when these solutions are no longer successful.

With man, according to Popper’s “metaphysical” conjecture, things are different. Our invention and use of language creates a new “world,” an “ontologically distinct” realm, which he calls. “World 3” or the “third world.” He thinks that it exists alongside the “first” world of material objects like glasses and polar bears—objects for whose existence, as we have seen, he cannot advance any decisive empirical reasons, only a “metaphysical” faith—and the “second” world of purely mental states, feelings, emotions, dispositions to act. The third world is the world of our intellectual products, the world of documented theories, problems, errors, standards, rules, values, the world of “objective knowledge”—knowledge that is an object, not (as the old “subjectivist” theory held) something that is “in” you or me, an expression of some mental state like certainty or “justified true belief.”

This third world is composed of abstract entities—the discoveries of William Harvey, of Hilbert and Planck and Carnot, the imaginative worlds created by Pope and Swift and Flaubert; it contains the things we argue about when we discuss the truth of a prediction about money markets or the future of a political party. It contains the content, for example, of the passage in John 6 that Luther and Zwingli argued over, a content that was not identical with either the spoken or the written words used by these men or with their subjective states, a content that is not “subjective knowledge” but a public object. The third world is historical, for the ideas of men have arisen in time; it is also to a considerable degree autonomous, for it contains not only the ideas that men have proposed, but also their interrelationships and unintended consequences. For example, the existence of prime numbers was surely not intended by whoever invented the natural numbers. Indeed, the third world even contains logical truths about these unintended products, truths, as Popper says, we can do nothing about, such as the nonexistence of a greatest prime number.

The third world is deeply implicated in the themes of Popper’s mature philosophy: for example, he uses it to show that violence is not necessary. Unlike animals, we need not die if our theories are refuted, for theories are in the third world, not in our organisms or genetic system. In the democratic and pluralistic “open society” Popper cherishes, criticism of policies and institutions (which are World 3 objects) must therefore be protected and encouraged, and governments can be revised or replaced through critical discussion and without bloodshed, rather as hypotheses are replaced in science. This political system—as contrasted with systems that rely upon uncritical “non-falsifiable,” utopian, revolutionary, “total” blueprints for maintenance or change of the political order—is for Popper the embodiment of critical reason in human affairs.

The third world is even invoked in Popper’s philosophy of art. Just as we must give up the “subjectivist” view that human knowledge is an expression of some interior state, we must resist the theory that art is “self-expression,” an overflowing of the contents of the mind onto a medium, paper or canvas or marble, and the language of aesthetic criticism—of “authenticity,” “integrity,” “sincerity”—that goes with it. All great art. Popper says, is anchored in the third world, in inherited problems and stylistic traditions. “Subjective” artists like Beethoven or Wagner did not regard their art as a means of self-transcendence but as the expression of something “inner” or “private”: the attitude of “objective” artists like Back, on the other hand, conforms to the principle that “the best work in science or in the arts or humanities is done when we forget about ourselves, and concentrate on World 3 issues as much as we possibly can.”²

But more than this, the third world is connected with Popper’s view of the self and human freedom: he believes that our very minds and selves have come into existence in the evolutionary process because of the invention of language: we are self-conscious because we have learned language, and our minds are the evolutionary products or vehicles which enable us to grasp the abstract World 3. The recently published Post-script to The Logic of Scientific Discovery contains Popper’s most ambitious statement of a metaphysics of the universe.³ Like his theory of the three worlds, which it deepens and refines, this cosmological work defends metaphysical realism against idealism and subjectivism. One long volume of it argues against the intrusion of subjectivism in quantum mechanics. Many paradoxes in the theory of quantum mechanics derive, according to Popper, from the so-called “Copenhagen” interpretation of it promoted by Niels Bohr and his followers. This view holds that Heisenberg’s celebrated “uncertainty relations,” which follow from the theory, actually set “limits” to our knowledge in the realm of atomic physics because they imply that measurements of an elementary particle (such as an electron) “disturb” it, or “interfere” with it, thus making the measurement “dependent” on the “observer,” and rendering objective knowledge of matter impossible.

Popper rejects this view as a species of “subjectivism” and argues that it is based not only on an arrogant suggestion that quantum theory is the last word in atomic theory, but also on a misunderstanding of the essentially statistical character of the theory. According to him, quantum theory is just like any other statistical (but objective) theory in physics: the “uncertainty relations” have nothing more to do with “uncertainty” or “limits of our knowledge” than any other statistical predictions. They show, if anything, the limitations of a probabilistic theory like quantum theory.

But Popper does not contest the “Copenhagen” view of quantum theory as implying that nature is “indeterministic.” Indeed, he criticizes several forms of “scientific” determinism—the idea that if we know the present state of the world, all physical events, whether in the past or in the future, can be predicted (or retrodicted) with any desired degree of precision, including all the movements of human beings. He advances an array of logical arguments, some of them rather crude but others ingenious, against this famous principle. But his most passionately argued attack on it is based on considerations of human creativity and freedom. It is absurd, he writes, to suppose that “billions of years ago, the elementary particles of World 1 contained the poetry of Homer, the philosophy of Plato, and the symphonies of Beethoven as a seed contains a plant,” or that a physicist, by studying the bodies of Mozart and Brahms with meticulous care, could write scores which were not actually written by them, but which they could have written if, say, their diet had been different.

On Popper’s own metaphysical view the universe is “open”: it has “pockets” of causality—ranges of events that are fully determined—but also realms that are unpredictable. One important reason for this conclusion, he says, is that human beings have introduced, via World 3, completely novel ideas into the universe, ideas which through human action have made a genuine difference to the course of events. He asks us, however, to distinguish carefully: it is one thing to say that determinism in one or another version is false; it is another to ensure the possibility of human freedom. The denial of determinism is not sufficient to make room for human freedom or creativity, he says, for what “we want to understand is not only how we may act unpredictably and in a chancelike fashion, but how can we act deliberately and rationally,” how things like plans, purposes, arguments, and decisions can actually bring about modifications in the world. To do so, he continues, we need the extra notion that the world of particles and other material objects is incomplete, that it can be influenced by Worlds 2 and 3 and interact with them.

Popper’s reflections on quantum theory and indeterminism are woven together in a remarkable “Metaphysical Epilogue,” which adumbrates what he calls a “new and promising way of looking at the physical cosmos,” a piece of “speculative physics” which, like other research programs in the history of science—like atomism or the unified field theory—is not itself empirically testable, but which might assist scientists in coming up with fruitful ideas. As he explains his “dream program,” it is intended to preserve elements of the two views that created a schism in twentieth-century physics: the indeterminism of quantum theory on the one hand, and on the other the aspiration of Einstein and the rest to construct a “unified field theory” in which the opposition between matter and “field” (say, of force or energy) could be superseded and particles explained as “produced” by properties of fields or interactions between them.

In one interpretation of Einstein’s view, matter was a “form” of electrical energy. Popper also seeks a unified field theory, but of a more ambitious variety. For him, the properties of matter—as well as its creation or destruction—might be explained as arising from fields of what he calls “propensities.” But whereas earlier theories sought to explain no more than the disposition or propensity of particles to behave this way or that, Popper thinks that all physical properties of the world are propensities. Particles are just propensities to “realize” this or that behavior. But at the same time they are the result, or “actualization,” of other propensities that make up the physical world. These propensities or possibilities, moreover, are real, as real as the gravitational forces that lock the planets of the solar system in their orbits, although some of them are unpredictable, “open.” Popper only sketches his program, but he claims, or hopes, that if taken up by others, it will help resolve many problems: the metaphysical problems of matter and change and space and causation, but also the major difficulties that have been bedeviling quantum mechanics since its inception.

If nothing else, the Postscript will remind readers that Popper differs radically from those logical positivists with whom he is still sometimes identified, who held that nontestable cosmological speculation of this kind is a form of superstition or chicanery, if not simply nonsensical. The emphasis placed in these books on human creativity and the constant interaction of the three worlds should also highlight the oddity of O’Hear’s criticism, suitably accompanied by quotations from Lukács, that Popper’s third world is an untenable “reification” of language and criticism and has an “alienating quality,” encouraging in us the idea that human institutions and human knowledge are governed by “inhuman laws.”⁴

O’Hear fails adequately to convey the generosity and sweep of Popper’s ideas—ideas which must have appealed powerfully to the young English philosophers he encountered when, returning to Europe from New Zealand (where he spent the war years), he first taught at the London School of Economics students who had been starved on a diet of logical analysis and problems of sense perception. O’Hear’s book is a dry catalogue of the errors in Popper and does not communicate the heat of the vision that turned a good number of these students into disciples and evangelists, into members of a “Popperian” school, bound to their master by undeviating loyalty (and subject to expulsion if remiss in this respect), convinced that he had struck a serious alternative to certain fundamental beliefs (like induction) woven into the texture of human thought.

Still, O’Hear largely succeeds in reinforcing the widespread impression that something is deeply amiss in Popper’s philosophy of science. According to him, what is wrong is that Popper fails to appreciate how firmly entrenched induction is in our ways of thinking and acting; scientists “need” induction in order to perform actions—such as choosing between competing scientific theories—that arise in ordinary scientific research. And in any case, he continues, Popper himself cannot do without induction. For if we take him at his word, he is left with no resources for establishing reasons for rational choices between theories in science, or for supposing his method is in fact a means toward the goal of science, indeed with no good reasons for our “embarking on the scientific adventure” as he defines it. Such reasons, O’Hear assumes, could only arise from the admission of some determinate link between failed refutations, or “corroboration,” on the one hand and truth on the other. When the full implications of Popper’s rejection of induction are digested, we see that he is in fact committed to a vicious skepticism—a skepticism, says O’Hear, that Popper tries again and again to escape, sometimes by coming perilously close to admitting to the existence of induction after all.

This is true as far as it goes, but it does not go far enough. It does not really take Popper seriously enough. He is, as he writes, not interested in doing justice to “inductive intuitions”; he states again and again that induction is a myth; he is trying to replace our “inductive intuitions” with a different view of human knowledge and scientific method that dispenses with induction entirely. The problem is not just that Popper’s alternative may offend common sense and the requirements of action, and that his argument would be better off if an element of induction were admitted into it. It is that he has not offered a coherent alternative at all. For as he describes science, it is self-defeating to engage in it: if you adopt his description of the aim of science as the truth, it is pointless to pursue his method of conjectures and refutations, for he denies they can arrive at any rational claim to the truth; on the other hand, if you endorse his method, the aim of science, astonishingly enough, has no legitimate connection with scientific research. This is not, perhaps, a logical inconsistency in his thought, but it amounts to something like a “practical” one.

One way of illustrating this difficulty focuses upon the role of testing in Popper’s scientific method. For him, arriving at hypotheses—whether about the symptoms of a disease or voting patterns or astronomy—is a matter that is not susceptible to logical analysis; hypotheses are “free creations,” and indeed it does not matter much where they come from. What matters is the “severe” critical controls that are applied to them in the stage of testing. It is these sober efforts that make science rational and distinguish it from pure speculation. They explain, for example, why we in some sense “know” that iron is heavier than water or that air has pressure, and why, on the other hand, we no longer take seriously the views that music has magnetic effects or that disease is a function of a person’s “humoral economy.” Scientists want to contribute to the “growth of knowledge,” by which Popper means not just an aimless proliferation of hypotheses, but the critically controlled transformation of what we currently hold or take for granted—what he calls “background knowledge”—into a new body of conjectures which we hope might be nearer to the truth. Choice on the part of scientists in selecting theories to add to, or subtract from, “background knowledge” is therefore at the heart of Popper’s concerns; presumably, these choices can be rational because they are informed by the results of testing and criticism. As he has written, “It is the growth of our knowledge, our way of choosing between theories, in a certain problem situation, which makes science rational.”⁵

But what exactly do the critical controls amount to, according to Popper’s noninductive view? If we acknowledge, as he insists we should, that there is no proof that we will find the truth, how can they even assist us in searching for the truth? The answer is that they could not amount to very much, since Popper has, in effect, pulled the rug from under our feet. Our tests, he says, can provide no positive reasons for the truth of any of our empirical hypotheses. But neither can they give us worthwhile negative or “critical” reasons: they are simply a report of failed attempts to falsify our ideas and make no reference to how these theories might fare in the future. In neither case can testing—the “rational” element in science—bear upon considerations pertaining to our aim of truth or even something like the truth.

Why, then, should scientists bother to test their hypotheses at all? Why should they, if such tests could never in principle provide them with any good reasons for thinking their hypotheses true? Would it not in fact follow that any hypothesis would be as conjectural after any amount of testing is over as it was before testing began? And if this is so, what critical controls could govern the so-called “growth of knowledge”—the alteration of the existing or “background” knowledge we currently accept in the search for truth? What considerations could possibly control this alteration? Since background knowledge is “taken for granted” and used to direct and guide future scientific inquiry, how could we ever find good reasons for introducing new hypotheses into background knowledge, and then taking them “for granted”? And if there is no answer to this question, what significant reasons could there possibly be for ever assuming that we have altered background knowledge in a manner favorable to the search for truth? At least Popper is consistent when he remarks that in his own view the success he happens to believe science has enjoyed in the past is “miraculously improbable, and therefore inexplicable.”⁶

It might be said, however, that Popper does provide something like critical checks in science insofar as our tests, while they cannot help us in detecting the truth, might yet knock out or falsify many of our hypotheses. After all, finding a black polar bear does refute “All polar bears are white.”

But even this modest claim is undermined by Popper’s own view, for the same difficulty about testing we just examined breaks out when we consider the statements that falsify theories. It is for mally correct to say that the statement “This is a black polar bear” falsifies “All polar bears are white.” But then the latter statement also falsifies the former; logically, all we know is that the two clash. The issue turns on the claim that we have found a black polar bear. After seeing such a bear, most of us would no doubt say that our experiences have given us some reason for making such a claim.

But as I explained earlier, Popper does not hold this view. For him, “This is a black polar bear” is impossible to justify; it implies a vast number of testable consequences that “transcend” all observational experience. The word “bear” is a construction intended to apply to all bears; it is not derived from direct experience. While the “acceptance” of its use might be caused or “motivated” by experience, it cannot be justified by it—“no more than by thumping the table.”⁷ According to Popper, all we can do with any such statement is to “decide” whether to “agree” to accept it: “from a logical point of view, the testing of a theory depends upon basic statements whose acceptance or rejection, in its turn, depends upon our decisions. Thus it is decisions which settle the fate of theories.”⁸ And these decisions are “free,” ungrounded, and ungroundable in any decisive positive or negative reasons. “The acceptance of a refutation is nearly as risky as the tentative adoption of a hypothesis: it is the acceptance of a conjecture.”⁹

But we are once more in a bind: if what Popper says is true, if even banal elementary statements of observation like “This is a glass” are unjustifiable, so that experience provides no good reasons for claiming them to be true, why should scientists make observations at all? Especially if they are interested in the truths that falsify hypotheses? How could rejection of theories as “falsified” be critically controlled on the basis of these “free decisions”?

It would seem, then, that the rational element in science—its critical controls and “our way of choosing between theories”—is for Popper entirely detached from the aim of finding the truth about the universe. This is the result of a conflict of doctrines within Popper’s own view. If you seek the truth, the ban on induction prevents testing from making any determinate impact on your decisions to alter your “background knowledge” in the search for truth. From the point of view of seeking truth, you might as well just make guesses and not bother to test these guesses, since testing can in principle give you no good reasons for thinking you are making progress. On the other hand, if you adopt the method prescribed by Popper, you might as well give up the search for truth. If you adopt both Popper’s aim and his method, then in practice not only are the “trials” or hypotheses in science “free creations” but so also are the guesses that “weed out errors”—so that all of science consists in “trials.” But if this is so, then what of the “critical discussion” of science from Heraclitus to Einstein that Popper claims persuaded people that a “great step had been taken toward the truth”? Is this critical discussion distinguishable from a pointless multiplication or proliferation of guesses, a self-defeating enterprise checked only by the one thing we can be sure of, the inspection of logical relationships between statements?

While Popper can be admired for his stress on the critical nature of tests in science, and for his rejection of crude empiricist psychologies and the search for secure foundations of human knowledge, he has not succeeded in returning satisfactory answers to these simple questions.¹⁰ Insofar as he has addressed them at all,¹¹ his response is this: testing is an important critical control on our search for truth because while it is true that failed refutation, or “corroboration,” is not a measure or indicator of truth it does indicate how the truth appears to us in light of the present discussion; moreover, corroboration—whether of a theory like Newton’s or of a statement like “This is a glass of water”—does provide good reasons not for the truth of these hypotheses, but for a preference for one of them in the search for truth. It provides “logically inconclusive reasons,” that is to say, “for conjecturing that it is the most truthlike of the hypotheses competing at the time.”¹² And this “conjecture” or “preference” is a “guess” of a higher order or level, a “metaconjecture” that the hypothesis is truthlike or that “further tests will not lead to any deviating results.”¹³

But this solves precisely nothing. To say that we “prefer” “All polar bears are white” in the search for truth to “Some polar bears are white and some are black” in light of the present discussion just repeats what we already know, that we have not yet “decided” that we have found any black polar bears, and, contrary to Popper, does not provide any good reasons even for guessing that it is “the most truthlike of the hypotheses competing” at the present time.

But more disturbing than this, if induction really does not exist—if we cannot rationally rely on inconclusive evidence for our beliefs—how could the results of testing provide any good reasons for a “higher-order conjecture”? How could testing provide good reasons for the conjecture that testing the proposition “All polar bears are white” will not lead to its falsification in the future? The truth or falsity of this conjecture is given no rational warrant by any of the tests we have performed so far, so why couldn’t we have simply made this higher-order guess prior to testing? Why should we test a guess just in order to advance it as a guess all over again? Popper suggests, moreover, that it is rational to act on a well-tested hypothesis or conjecture—to use it—not because there are good reasons for thinking it true but because there are no reasons for supposing it is not true. But this is a desperate move. There are, on Popper’s skeptical premises, all kinds of reasons for doubting any hypothesis. Why should we accept the higher-order conjecture? Why should we not test it in turn?

Popper, in short, claims that there are no good reasons for thinking a theory is true, but that there are nevertheless good “critical” reasons for conjecturing it to be true; but these critical reasons are of no account unless we have some reason for believing that we are thereby getting closer to the truth, and Popper fails to supply us with a good reason for doing so. His “preference” theory covers up with a cloud of words the problem of how testing serves any determinate and rational purpose in his scientific method. The problem is unaffected. That it remains so hints at difficulties relevant not only to his philosophy of science but more generally to his theory of knowledge as a whole, and even to some of his justifications of the open society as a social order designed to promote unfettered criticism, criticism that will encourage in its turn the growth of knowledge and better, more informed decisions concerning the “piecemeal social engineering” that is intended to improve the lot of the citizenry. Why should there be unlimited criticism of this kind if the results are as pointless as we have suggested? At the very least, it is impossible to accept Popper’s grotesquely immodest claim to have completed the solution of the philosophical problem “whose more fundamental half was already solved by Hume,” a problem he describes “with a little generosity” as “the problem of human knowledge.”¹⁴

Popper’s philosophy of science is profoundly ambiguous: it is, he says, “empirical,” but it is left unclear why scientists should consult experience. It is called “fallibilism,” in which we “learn from our mistakes,” but it is really an ill-concealed form of skepticism. It claims to surrender the quest for certainty, but it is precisely the standards of this quest—that if one is not certain of a proposition, one can never be rationally justified in claiming it to be true—that underlie Popper’s rejection of induction (and the numerous doctrines that stem from this rejection). He asks scientists to search for new facts, new theories, new truths—but then adds that no matter how hard they search, all they will ever be able to know will be as risky when they are finished investigating it as it was when they began, and that the only things that can be other than “daring” guesses¹⁵ are empty logical truths. This is not abandoning the straitjacket of the quest for certainty but turning it inside out.

By a queer transition, Popper calls his skeptical view “critical rationalism” and “objectivism.” But it is small wonder that the most consistent of “Poperians” have been precisely those lapsed disciples of his, “anarchists” and “subjectivists” like Paul K. Feyerabend, who have persuaded themselves that psychological compulsions, habits, and whims are the actual, if well-nigh universally unacknowledged, motors of scientific change—an irrationalism similar to that which Popper deplored in Hume—and that science neither follows bodiless formulae nor yields genuine knowledge. Les extremes se touchent: as so often in the history of ideas, a formal “arch-objectivist” view offering precious little guidance to actual activities but trumpeting abstract formal principles leads through superficially persuasive steps of reasoning to a doctrine of “anything goes.”

Popper is led to these difficulties by a combination of doctrines implausible in themselves and even less plausible in combination. He wrote his first book in reaction to the view that the empirical sciences are “reducible” to our sense experiences, and rightly held that this view hardly did justice to objective science. But he jumped to the other extreme, and adopted a largely formal approach to the theory of scientific knowledge. From this approach he developed his main views about the priority of logic over psychology, the rejection of induction on logical grounds, and the discovery of the method of conjectures and refutations as the best, because most “rational,” scientific method.

But this unempirical approach, nourished by the horror of all “subjectivism,” yields a conception of science and its guiding methods which is utterly inappropriate to its subject matter. As Popper recognizes, science (like the law) is a human activity, involving a problem-solving intellectual community governed by critical habits and principles. But these principles have themselves developed by trial and error and may develop further. None of these principles is “intrinsically” or “essentially” rational: their rationality lies in their success in promoting the ends of inquirers and resolving the problems of scientists. This is clearly seen in the case of induction: even if there are no formal rules codifying this practice, it hardly follows that it is “arbitrary” or “subjective” (let alone that it does not exist), simply on the ground that it does not conform to the standards of rationality antecedently laid down by Popper.

Instead of espousing a view as riddled with internal flaws as Popper’s, we can reasonably continue to believe that there really are inductive practices, that people do discriminate differences of probative weight in evidence and do rely on logically inconclusive evidence for many beliefs they hold, and further that such evidence can, under certain circumstances, genuinely provide rational support for our claims to know a great variety of empirical statements. Popper’s fear of “subjectivist” views has led him to exclude from the analysis of science not only those judgments correctly described as infected with bias or subjective distortions, but also what is typical in science: decisions and cases of deliberation which are not strictly dictated by universal and exceptionless rules, but which require personal judgment; consequently, he leaves outside the scope of critical guidance the practices that most need it.

Popper is indeed a rationalist of sorts—a Romantic rationalist. Throughout his work we find the image of scientists trying to impose their theories on nature and then awaiting the voice of nature in response; an emphasis on fierce competition between theories; a stress on risks, bold conjectures, and imaginative criticism; a hatred of the view that science is nothing more than technological control; an image of science as a never-ending struggle whose mainspring is contradiction, in which we deliberately seek contradictions in our hard-won syntheses and solutions to avoid stagnation and erect fresh hurdles and challenges; an idea of science as a process that pursues an aim that is elusive, perhaps unreachable, a process that counts as much, if not more, than attaining the goal itself, a journey in which we heroically, impossibly, try to narrow the discrepancy between our finite grasp and our infinite aim. This picture has not enjoyed so distinguished (if not more convincing) an adherent since the days when it was applied to the moral realm, and sometimes to the whole of human existence, by such German Romantic writers as Fichte. Surely it is a powerful reason for Popper’s appeal, probably more so than the arguments that have made him the scarecrow of “inductivists.” The image Popper offers of our efforts to acquire the truth about the world may be momentarily intoxicating, and set up unusual and stimulating trains of thought, but, contemplated in a cool hour, it describes a wild-goose chase.

(This is the second of two articles on Karl Popper.)