If Popper’s style puts off many, however, it also attracts some fervent disciples most of whom are represented in The Philosophy of Karl Popper. For some years now there has been a little hand of enthusiastic Popperians, carrying forward the banner of Popper’s ideas. Now and again one will fall from grace, and another will appear. Magee is obviously a recent convert; on the other hand, Popper’s reply to the essay by Imre Lakatos in The Philosophy of Karl Popper clearly marks the excommunication of that once lofty follower, for the series of papers in which this former student and colleague of Popper’s has tried to guide the reader through Popper’s writings is now declared, by the ultimate authority himself, to be “unreliable and misleading,” Poor Lakatos.
Popper’s knack of attracting disciples is an intriguing phenomenon, although one that cannot be discussed here. The irony is that Popper, the biting critic of petty, scholastic wrangling, now has to admit that his own works have become the subjects of scholastic disputes. It must be galling for Popper to find himself divided by his supporters into Popper, Popper1, and Popper2 with consequent endless possibilities for debates over interpretations. On the other hand, Popper complains so frequently of being misunderstood—the intentions of the editor of The Philosophy of Karl Popper have been seriously thwarted by the fact that on several occasions Popper and his critics simply fall to engage because, according to Popper, the essays are directed against positions that he never held that one begins to suspect that the fault may lie with the author as much as with the expositor or critic.
Finally, so far as style is concerned, Popper’s desire to swamp his opponents with criticism results in a failure to distinguish good arguments from bad. While one may applaud Popper’s conviction that real argument is preferable to the kind of suggestive observations that Wittgenstein and his followers used to throw out, Popper himself has debased the currency of argument by his indiscriminate employment of any argument that comes to hand. Does Popper really think, for instance, that it is an argument against the impossibility of doubting one’s own existence that Kipa, a Sharpa who went further up Everest than was good for him, afterward thought he was dead? Or even that Popper himself had the same experience when struck by lightning in the Austrian Alps (Objective Knowledge, p. 36)? Any undergraduate philosopher would reply that believing one is dead is very different from believing that one doesn’t exist.
Some of Popper’s replies to his critics in The Philosophy of Karl Popper contain arguments almost as bad—for instance, in reply to the ease for determinism presented by Feigl and Meehl, Popper remarks that they were unable to predict the form his reply would take, although Feigl and Meehl had explicitly disclaimed the ability to make such predictions.
What has Popper achieved in the philosophy of science, and how does his achievement relate to the problem of induction?
Science, according to a tradition going back to Francis Bacon, proceeds from the open-minded accumulation of observations. When the scientist has collected enough data he will notice a pattern beginning to emerge, and he will hypothesize that this indicates some natural law. He then tries to confirm this law by finding further evidence to support it. If he succeeds he has verified his hypothesis. He has discovered another law of nature.
Popper’s challenge to this view starts from the simple logical point that a universal statement like “All swans are white” cannot be proved true by any number of observations of white swan—we might have failed to spot a black swan somewhere—but it can be shown false by a single authentic sighting of a black swan. Scientific theories of this universal form, therefore, can never be conclusively verified, though it may be possible to falsify them.
Hence Popper says that it is wrong to begin by accumulating observations, and it is wrong to seek confirming instances of a theory. Instead we should advance bold conjectures—derived from intuition, or creative genius, or any way we like—and attempt to refute them. Of two competing theories, the one that has run the greater risk of falsification, but has not been falsified, is the better corroborated. This does not mean that it is true—it may be falsified in the future—but it is likely to be a closer approximation to the truth than its rival. We can never, in science, know that we have discovered the truth although there is such a thing as truth, it is a regulative idea which we try to approach, but can never be sure of reaching.
There is an objection to this, urged by both Hilary Putnam and Thomas Kubn in their contributions to The Philosophy of Karl Popper; it is always possible to deny that a theory has been falsified by an observation that at first foes seem to falsify it. One can deny, for instance, that the reported sighting of a black swan was authentic; or one could say that if the bird was black, then by definition it just wasn’t a swan, no matter how much it resembled swans in other respects. In general, scientific theories are not tested in isolation, but in conjunction with other assumptions; therefore it is possible to save the theory, and explain away an observation that contradicts it, by claiming that one of the other assumptions was at fault.
Popper has not overlooked this objection; indeed, he mentioned it in his earliest writings. His reply is that as a methodological rule we should avoid “immunizing” our theories in this way, although he admits that there will be times when it is worth trying to preserve a theory despite anomalous observations.
What Popper says on this point is hardly precise, and perhaps for that reason it may not satisfy his critics; at the same time, Popper warns against the search for precision in places in which it is not to be found. We must allow ourselves to be guided by the circumstances of each case. In this way, Popper is able to retain his central point; the asymmetry of verification and falsification.
It may be helpful to illustrate this by an example, and the example that actually influenced Popper most decisively serves well. When Einstein conjectured that light rays passing close to a heavy body like the sun would be deflected from their normal path, this effect had never been observed. Newtonian physics predicted no such effect. When the observation was made and Einstein’s prediction confirmed, Newton’s “laws” were shown to be false, despite the immense amount of “verification” they had received over the centuries. It might in fact, have been possible to cling to Newton’s theory by introducing some ad hoc hypotheses to explain the observations, but to do so seemed implausible when the new theory explained matters more simply. The point is that Einstein could account for all of Newton’s successes, plus one of his failures. Newton’s theory could not stand; Einstein’s must be a closer approximation to the truth. So it survives to face further testing.
Although, as The Philosophy of Karl Popper shows, this view of science may not be unanimously accepted today, it has much support. I regard it as a huge advance upon the previously accepted idea of scientific method. Is it also a solution to the problem of induction? In two separate essays in Objective Knowledge and again in the course of his replies to his critics in The Philosophy of Karl Popper, Popper claims that it is. He complains, however, that other philosophers have not recognized his solution and he says that his critics have not understood him.
What is this notorious problem of induction? Hume’s logical problem of induction is whether we are justified in reasoning from instances of which we have experience to instances of which we have no experience. For example, we have observed the sun rising on numerous past occasions. Does this justify our belief that it will rise on a future occasion, say tomorrow?
The assumption that this belief is justifiable is, Hume said, absolutely basic to our ideas of rational belief and rational action. Without it the most carefully derived expectations are ultimately no more defensible than the bizarre fancies of a madman. Without a justification of induction, the distinction between rationality and irrationality appears to be in peril.
How does Popper answer Hume’s question? First, we must note how he interprets it. He regards it, correctly, not merely as a problem of generalizing from single cases to all cases, but as one about reasoning from past cases to a single future case. He also—perhaps less soundly, but we cannot go into that here—agrees with Hume that describing our belief about tomorrow’s sunrise as probable, rather than certain, makes no difference to the argument.
Popper then says that Hume’s question has to be reformulated. The main effect of his various reformulations is to turn the issue from a question about single cases to one about universal explanatory theories. On that reformulated issue his answer, predictably, is that observations cannot justify the claim that a universal theory is true, but they can allow us to say that it is false, and they can justify a preference for some theories over others.
Popper’s critics point out that this does not seem to answer Hume’s question. The theory that the sun rises every day may have survived falsification yesterday, but is this a reason for believing that it will survive falsification tomorrow? That was what Hume wanted to know.
Popper retorts that he has answered this question. His answer is “No.” Induction is not justifiable. That a theory has been corroborated in the past “says nothing whatever about future performance.” The best corroborated theory may fall tomorrow. “It is perfectly possible that the world at we know it, with all. Its pragmatically relevant regularities, may completely disintegrate in the next second.”
Of course universal disintegration is possible, but we appear to be justified in gambling heavily against it occuring in the next second, if only because the world has not disintegrated In any of the previous seconds of which we have knowledge, and we have no grounds for believing that disintegration is more likely in the next second than in any other second.
Hume showed, however, that this plausible argument simply assumes that past observations do justify future predictions, and if we try to defend this assumption on the grounds that those who, in the past, assumed that the future will be like the past have turned out to be right, we have again assumed what we wanted to prove. So we have an assumption that appears impossible to defend without circularity, but equally impossible to avoid making.
Popper wants to say that it is possible to avoid assuming that the future will, or probably will, be like the past, and this is why he has claimed to have solved the problem of induction. We do not have to make the assumption, he tells us, if we proceed by formulating conjectures and attempting to falsify them.