In response to:

The Romantic Rationalist from the December 2, 1982 issue

To the Editors:

Dr. Lieberson writes eloquently and at length [“The Romantic Rationalist,” NYR, December 2] on Popper’s philosophy of scientific knowledge, especially his solution to the problem of induction. But he succumbs, alas, to the same conclusion as most of Popper’s other critics, that “if you adopt his [Popper’s] 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” (p. 53). Dismissing falsificationism accordingly as “riddled with internal flaws” (p. 55), Lieberson endorses instead that worthless product that is still being offered to an innocent public under the trademark of “induction.” That “there are no formal rules codifying this practice” (Ibid.) and no standards of rationality to which it appears to conform, seems not to have lessened his esteem for something that critical rationalists regard as simply an intellectual fraud.

Lieberson’s complaint against falsificationism is the familiar one that it provides us with no good reasons for thinking that its methods will help us to attain the truth, or even to approach it. His article resounds with requests for good reasons for this and that, and with regrets for Popper’s resolute refusal to rest science on reasons that do not exist. But Lieberson doesn’t anywhere face the question of what such supposed good reasons are supposed to be good for. In my contribution “Conjectural Knowledge: Popper’s Solution of the Problem of Induction” to In Pursuit of Truth, the Popper Festschrift edited by Levinson, I argue that actually they are good for nothing at all (pp. 20-22, 42-44). Good reasons not only are, as Hume showed, impossible to obtain, and, as Popper showed, unnecessary anyway for rational thought and action; they would, as Socrates indicated at the end of the Meno, be utterly useless even if they could be obtained. It is therefore not easy to see what criticism it is of falsificationism that “it fails to do something that it has no plan or purpose to do: to justify its own success” (“Conjectural Knowledge,” p. 44). Lieberson’s strictures, that is to say, like those of most other critics of falsificationism, miss their mark; in no way do they respond to my challenge “to say which of its [falsificationism’s] recommendations or injunctions, its methodological rules, are unsuited to the business of sorting out what is true about the world from what is false” (Ibid.). Lieberson badly misreports me here as merely asking falsificationism’s critics “to try to be more careful and to formulate their criticisms more clearly” (“The Romantic Rationalist,” p. 54). But, to be honest, I’m not really worried about the clarity of these criticisms, for they are already clear enough. What’s sad about them is that they are so waywardly misdirected.

I think that Lieberson will find in my paper answers to pretty well all the objections he levels against Popper’s philosophy of conjectures and refutations. He will find there answers even to the questions he chides me for failing to answer; for example, the question of why “scientists bother to test their hypotheses at all” (“The Romantic Rationalist,” p. 53). Far from ducking this question I answer it, as Popper does, in the obvious way (“Conjectural Knowledge,” pp. 22-25): scientists test their hypotheses in order to detect falsehoods, and so to eliminate as contenders for the truth those hypotheses that are false. That this is a deeply uncertain undertaking, that “tests could never in principle provide them [scientists] with any good reasons for thinking their hypotheses true” (“The Romantic Rationalist,” p. 53) is to my mind quite beside the point. For scientists are not seeking good reasons to think that they are eliminating falsehoods and retaining truths; they are seeking to eliminate falsehoods and retain truths. If Lieberson (or anyone else) thinks that testing can’t help them to do this, I should be interested to know why.

David Miller

University of Warwick

Coventry, England

Jonathan Lieberson replies:

Mr. Miller’s strongly worded letter claims that my “complaint against falsificationism is the familiar one that it provides us with no good reasons for thinking that its methods will help us to attain the truth, or even to approach it.” I was presumably asking “falsificationism” to “justify its own success,” a goal that is, in the words of his essay, “utterly pointless.” But this was not my intention: no less than Mr. Miller did I demand that Popper’s method certify itself, although I also claimed that asking for reasons (not proof) for espousing a method is not “utterly pointless” and that these reasons are not always formal. Indeed, the history of science suggests that our appraisal of methods depends on the extent to which they aid us in our inquiries; as such, our assessment of methods is a quasi-empirical question. At any rate such an assessment cannot be founded exclusively (as Popper’s seems to be, despite his celebrated “anti-essentialism”) on conformity to principles that are presumed to definitively characterize rationality in all circumstances.


But my central criticism was different from what Miller claims. I did not ask that Popper’s method be judged against a standard of proof foreign to it. I argued that it is internally flawed. Popper attempts to provide a non-inductive account of scientific method, which regards the aim of science as the truth. We try to falsify our theories and so get closer to the truth. I asked for grounds internal to this view for supposing that it makes sense to engage in it at all. In particular, I asked whether it makes sense to say that testing is necessary at all, given what Popper’s views imply about the epistemological status of such testing. Popper primarily (if not exclusively) locates the “critical control” of science—what distinguishes it from myth—in severe and rigorous testing. Yet he also implies that any hypothesis would be as conjectural after any amount of testing is completed as it was before such testing began. Such a view is certainly at odds with itself.

His method is self-defeating because if you adopt the aim of science as he describes it, his method seems pointless, for you are blocked at the outset from arriving at any rational claim to the truth. Truth—the goal of your research—has, it seems, no determinate connection to the very efforts you undertake in order to secure it, and your efforts enjoy no significant bearing on your goal. In practice, Popper’s method would amount to a skepticism that would trivialize science as we know it. I added that in light of this criticism, I saw, despite Popper’s claim to have solved the problem of induction, no reason why we should not continue to accept one of the contrary views that presuppose the existence and rationality of induction. I did not say that induction is perfectly understood, or that it has been exhaustively described, or that anyone has succeeded in formulating rules, even maxims, of inductive reasoning comparable in scope and clarity to those in formal logic. But, on the other hand, I do not regard this lack as a decisive objection to “inductivism.”

This criticism, I hope it is clear, is quite different from asking for the self-certification of falsificationism. It is one thing to ask a method to prove that it will attain its goal; it is quite another to inquire whether it prevents us in principle from making any rational claim that it has made the slightest progress toward it. That Mr. Miller thinks that my statement that “tests could never in principle provide them [scientists] with any good reasons for thinking their hypotheses true”—I might have added, “or truthlike”—is “quite beside the point” seems to me to be of a piece with the fanaticism that leads him to characterize induction as a deliberate perversion of the truth, as “intellectual fraud.”

It is indeed a sobering thought to imagine Mr. Miller addressing the following advice to a scientist: “You have created a hypothesis designed to solve a significant scientific problem. You must now test your idea by trying to falsify it; induction is nonexistent, and this is the sole ‘rational’ way of trying to get closer to the truth. But you must also be aware that your testing, however prolonged, will never provide you with a scintilla of rational evidence for thinking you are advancing toward, or receding from, the truth, the goal of your inquiry; it cannot even do so for the claim that you have found out your hypothesis is empirically false. Nevertheless you must continue testing. And when you stop testing, remember that your tests will have given you no help in rationally deciding whether your hypothesis is true or not; you must simply guess that it is true, or discard it, because it contradicts some other hypotheses which you have decided are true, once more without rational grounds of any kind.”

In view of Mr. Miller’s indignant accusation that I misreported the “final challenge” of his essay—which, I must repeat, does not seem to me to cast any light on the objections I have advanced—I should have been wise to make my meaning unambiguous. Many defenses of Popper by his supporters and epigoni claim that he has been egregiously misunderstood by his critics; the critics are asked to read what Popper has actually written and reformulate or clarify their criticism. This is how I described Mr. Miller’s challenge, and it seems to me correct: Mr. Miller asks why the objections he has “refuted” arise so frequently; “the answer,” he continues, “must lie in the failure of most critics to see what falsificationism is trying to do.” He then goes on to address these critics: he challenges them to say what is wrong with the ideas Popper actually holds, “which of its recommendations or injunctions, its methodological rules, are unsuited to the business of sorting out what is true about the world from what is false.” I took up this challenge, for if I am right, then there is some doubt whether Popper’s central methodological rules are suited to this task.


The cleverest of Popper’s disciples, Mr. Miller is adept at devising ingenious buttresses for the cathedral of Popper’s thought; and his writing displays the aggressive hauteur he has inherited from his master. But expertly aware as he is of points of detail, he seems to me blind to flaws in the larger design of Popper’s philosophy. As I tried to say in my articles, Popper’s philosophy of science contains many suggestive and stimulating ideas. But for all its breadth and sweep and occasional flashes of insight, it still seems to me fundamentally disfigured, in large part by its initial rejection of induction. It is, in words that were used by Santayana in another context, “capable of assimilating a great deal of wisdom, while its first foundation is folly.”

This Issue

April 28, 1983