Category: The Logic of Scientific Discovery

Routledge: 2002.

The problem of the growth of knowledge

The central problem of epistemology has always been and still is the problem of the growth of knowledge. And the growth of knowledge can be studied best by studying the growth of scientific knowledge.

And yet, I am quite ready to admit that there is a method which might be described as ‘the one method of philosophy’. But it is not characteristic of philosophy alone; it is, rather, the one method of all rational discussion, and therefore of the natural sciences as well as of philosophy. The method I have in mind is that of stating one’s problem clearly and of examining its various proposed solutions critically. [Preface, 1959]

Popper on Duhem–Quine’s naive falsificationism

The falsifying mode of inference here referred to—the way in which the falsification of a conclusion entails the falsifi­cation of the system from which it is derived—is the modus tollens of classical logic. It may be described as follows:

Let p be a conclusion of a system t of statements which may consist of theories and initial conditions (for the sake of simplicity I will not distinguish between them). We may then symbolize the relation of derivability (analytical implication) of p from t by ‘t ➙ p’ which may be read: ‘p follows from t ’. Assume p to be false, which we may write ‘p’, to be read ‘not-p’. Given the relation of deducibility, t ➙ p, and the assumption p, we can then infer t  (read ‘not-t ’); that is, we regard t as falsified. If we denote the conjunction (simultaneous assertion) of two statements by putting a point between the symbols standing for them, we may also write the falsifying inference thus: ((t ➙ p).p) ➙ t , or in words: ‘If p is derivable from t, and if p is false, then t also is false’.

By means of this mode of inference we falsify the whole system (the theory as well as the initial conditions) which was required for the deduction of the statement p, i.e. of the falsified statement. Thus it cannot be asserted of any one statement of the system that it is, or is not, specifically upset by the falsification. Only if p is independent of some part of the system can we say that this part is not involved in the falsification.* With this is connected the following possibility: we may, in some cases, perhaps in consideration of the levels of universality, attribute the falsification to some definite hypothesis—for instance to a newly introduced hypothesis. This may happen if a well-corroborated theory, and one which continues to be further corroborated, has been deductively explained by a new hypothesis of a higher level. The attempt will have to be made to test this new hypothesis by means of some of its consequences which have not yet been tested. If any of these are falsified, then we may well attribute the falsification to the new hypothesis alone. We shall then seek, in its stead, other high-level generalizations, but we shall not feel obliged to regard the old system, of lesser generality, as having been falsified.

* Thus we cannot at first know which among the various statements of the remaining sub-system t ′ (of which p is not independent) we are to blame for the falsity of p; which of these statements we have to alter, and which we should retain. (I am not here discussing interchangeable statements.) It is often only the scientific instinct of the investigator (influenced, of course, by the results of testing and re-testing) that makes him guess which statements of t ′ he should regard as innocuous, and which he should regard as being in need of modification. Yet it is worth remembering that it is often the modification of what we are inclined to regard as obviously innocuous (because of its complete agreement with our normal habits of thought) which may produce a decisive advance. A notable example of this is Einstein’s modification of the concept of simultaneity. [55-6]

The untenability of induction

My own view is that the various difficulties of inductive logic here sketched are insurmountable. So also, I fear, are those inherent in the doctrine, so widely current today, that inductive inference, although not ‘strictly valid’, can attain some degree of ‘reliability’ or of ‘probability’. According to this doctrine, inductive inferences are ‘probable inferences’. ‘We have described’, says Reichenbach, ‘the principle of induction as the means whereby science decides upon truth. To be more exact, we should say that it serves to decide upon probability. For it is not given to science to reach either truth or falsity … but scientific statements can only attain continuous degrees of probability whose unattainable upper and lower limits are truth and falsity’.

At this stage I can disregard the fact that the believers in inductive logic entertain an idea of probability that I shall later reject as highly unsuitable for their own purposes (see section 80, below). I can do so because the difficulties men­tioned are not even touched by an appeal to probability. For if a certain degree of probability is to be assigned to statements based on inductive inference, then this will have to be justified by invoking a new principle of induction, appropriately modified. And this new principle in its turn will have to be justified, and so on. Nothing is gained, more­over, if the principle of induction, in its turn, is taken not as ‘true’ but only as ‘probable’. In short, like every other form of inductive logic, the logic of probable inference, or ‘probability logic’, leads either to an infinite regress, or to the doctrine of apriorism. [6]

The recklessly critical quest for truth

With the idol of certainty (including that of degrees of imperfect certainty or probability) there falls one of the defences of obscurantism which bar the way of scientific advance. For the worship of this idol hampers not only the boldness of our questions, but also the rigour and the integrity of our tests. The wrong view of science betrays itself in the craving to be right; for it is not his possession of knowledge, of irrefutable truth, that makes the man of science, but his persistent and recklessly critical quest for truth.

Has our attitude, then, to be one of resignation? Have we to say that science can fulfill only its biological task; that it can, at best, merely prove its mettle in practical applications which may corroborate it? Are its intellectual problems insolu­ble? I do not think so. Science never pursues the illusory aim of making its answers final, or even probable. Its advance is, rather, towards an infinite yet attainable aim: that of ever discovering new, deeper, and more general problems, and of subjecting our ever tentative answers to ever renewed and ever more rigorous tests.[281]

Our method is not to prove how right we were

[T]hese marvellously imaginative and bold conjectures or ‘anticipations’ of ours are carefully and soberly controlled by systematic tests. Once put forward, none of our ‘anticipations’ are dogmatically upheld. Our method of research is not to defend them, in order to prove how right we were. On the contrary, we try to overthrow them. [278-9]

Settled science

To obtain a picture or model of this quasi-inductive evolution of science, the various ideas and hypotheses might be visualized as particles suspended in a fluid. Testable science is the precipitation of these particles at the bottom of the vessel: they settle down in layers (of universality). The thickness of the deposit grows with the number of these layers, every new layer corresponding to a theory more universal than those beneath it. As the result of this process ideas previously floating in higher metaphysical regions may sometimes be reached by the growth of science, and thus make contact with it, and settle. [277]

Corroboration and timeless truth

In the logic of science here outlined it is possible to avoid using the concepts ‘true’ and ‘false’. …

Whilst we assume that the properties of physical objects (of ‘genidentical’ objects in Lewin’s sense) change with the passage of time, we decide to use these logical predicates in such a way that the logical properties of statements become timeless: if a statement is a tautology, then it is a tautology once and for all. This same timelessness we also attach to the concepts ‘true’ and ‘false’, in agreement with common usage. It is not common usage to say of a statement that it was perfectly true yesterday but has become false today. If yesterday we appraised a statement as true which to­day we appraise as false, then we implicitly assert today that ; that the statement was false even yesterday—timelessly false—but that we erroneously ‘took it for true’.

Here one can see very clearly the difference between truth and corroboration. The appraisal of a statement as corrobo­rated or as not corroborated is also a logical appraisal and therefore also timeless; for it asserts that a certain logical relation holds between a theoretical system and some system of accepted basic statements. But we can never simply say of a statement that it is as such, or in itself, ‘corroborated’ (in the way in which we may say that it is ‘true’). We can only say that it is corroborated with respect to some system of basic statements—a system accepted up to a particular point in time. ‘The corroboration which a theory has received up to yesterday’ is logically not identical with ‘the corro­boration which a theory has received up to today’. Thus we must attach a subscript, as it were, to every appraisal of cor­roboration—a subscript characterizing the system of basic statements to which the corroboration relates (for example, by the date of its acceptance).

Corroboration is therefore not a ‘truth value’; that is, it cannot be placed on a par with the concepts ‘true’ and ‘false’ (which are free from temporal subscripts); for to one and the same statement there may be any number of different cor­roboration values, of which indeed all can be ‘correct’ or ‘true’ at the same time. For they are values which are logically derivable from the theory and the various sets of basic statements accepted at various times.

The above remarks may also help to elucidate the contrast between my views and those of the pragmatists who pro­pose to define ‘truth’ in terms of the success of a theory—and thus of its usefulness, or of its confirmation or of its corro­boration. If their intention is merely to assert that a logical appraisal of the success of a theory can be no more than an appraisal of its corroboration, I can agree. But I think that it would be far from ‘useful’ to identify the concept of corrobo­ration with that of truth.* [273-5]

* Thus if we were to define ‘true’ as ‘useful’ (as suggested by some pragmatists), or else as ‘successful’ or ‘confirmed’ or ‘corroborated’, we should only have to introduce a new ‘absolute’ and ‘timeless’ concept to play the role of ‘truth’.

Step-by-step approximations to truth

The degree of corroboration of two statements may not be comparable in all cases, any more than the degree of falsi­fiability: we cannot define a numerically calculable degree of corroboration, but can speak only roughly in terms of positive degree of corroboration, negative degrees of corroboration, and so forth. Yet we can lay down various rules; for instance the rule that we shall not continue to accord a positive degree of corroboration to a theory which has been falsified by an inter-subjectively testable experiment based upon a falsifying hypothesis. (We may, however, under cer­tain circumstances accord a positive degree of corroboration to another theory, even though it follows a kindred line of thought. An example is Einstein’s photon theory, with its kinship to Newton’s corpuscular theory of light.) In general we regard an inter-subjectively testable falsification as final (provided it is well tested): this is the way in which the asymme­try between verification and falsification of theories makes itself felt. Each of these methodological points contributes in its own peculiar way to the historical development of science as a process of step by step approximations. [266-7]

Falsifiability and probability statements

How is it possible that probability statements—which are not falsifiable—can be used as falsifiable statements? (The fact that they can be so used is not in doubt: the physicist knows well enough when to regard a probability assumption as falsified.) This question, we find, has two aspects. On the one hand, we must make the possibility of using probability statements understandable in terms of their logical form. On the other hand, we must analyse the rules governing their use as falsifiable statements.

According to section 66, accepted basic statements may agree more or less well with some proposed probability esti­mate; they may represent better, or less well, a typical segment of a probability sequence. This provides the opportunity for the application of some kind of methodological rule; a rule, for instance, which might demand that the agreement between basic statements and the probability estimate should conform to some minimum standard. Thus the rule might draw some arbitrary line and decree that only reasonably representative segments (or reasonably ‘fair samples’) are ‘permitted’, while atypical or non-representative segments are ‘forbidden’. [197]

Why simplicity is so highly desirable

Above all, our theory explains why simplicity is so highly desirable. To understand this there is no need for us to as­sume a ‘principle of economy of thought’ or anything of the kind. Simple statements, if knowledge is our object, are to be prized more highly than less simple ones because they tell us more; because their empirical content is greater; and because they are better testable. [128]