Paul Feyerabend, Against Method,

The results obtained so far suggest abolishing the distinction between a context of discovery and a context of justification, norms and facts, observational terms and theoretical terms. None of these distinctions plays a role in scientific practice. Attempts to enforce them would have disastrous consequences. Popper’s ‘critical’ rationalism fails for the same reasons. (p. 147)

Even the most surprising stories about the manner in which scientists arrive at their theories cannot exclude the possibility that they proceed in an entirely different way once they have found them. But this possibility is never realized. Inventing theories and contemplating them in a relaxed and `artistic' .fashion, scientists often make moves that are forbidden by methodological rules. For example, they interpret the evidence so that it fits their fanciful ideas, eliminate difficulties by ad hoc procedures, push them aside, or simply refuse to take them seriously. The activities which according to Feigl belong to the context of discovery are, therefore, not just different from what philosophers say about justification, they are in conflict with it. Scientific practice does not contain two contexts moving side by side, is a complicated mixture of procedures and we are faced by the question if this mixture should be left as it is, or if it should be replaced by a more `orderly' arrangement. This is part one of the argument. Now we have seen that science as we know it today could not exist without a frequent overruling of the context of justification. This is part two of the argument. The conclusion is clear. Part one shows that we do not have a difference, but a mixture. Part two shows that replacing the mixture by an order that contains discovery on one side and justification on the other would have ruined science: we are dealing with a uniform practice all of whose ingredients are equally important for the growth of science. This disposes of the distinction.

A similar argument applies to the ritual distinction between methodological prescriptions and historical descriptions. Methodology, it is said, deals with what should be done and cannot be criticized by reference to what is. But we must of course make sure that our prescriptions have a point of attack in the historical material, and we must also make sure that their determined application leads to desirable results. We make sure by considering (historical, sociological, physical, psychological, etc.) tendencies and laws which tell us what is possible and what is not possible under the given circumstances and thus separate feasible prescriptions from those which are going to lead into dead ends. Again, progress can be made only if the distinction between the ought and the is is regarded as a temporary device rather than as a fundamental boundary line.

A distinction which once may have had a point but which has now definitely lost it is the distinction between observational terms and theoretical terns. It is now generally admitted that this distinction is not as sharp as it was thought to be only a few decades ago. It is also admitted, in complete agreement with Neurath's original views, that both theories and observations can be abandoned: theories may be removed because of conflicting observations, observations may be removed for theoretical reasons. Finally, we have discovered that learning does not go from observation to theory but always involves both elements. Experience arises together with theoretical assumption not before them, and an experience without theory is just as incomprehensible as is (allegedly) a theory without experience: eliminate part of the theoretical knowledge of a sensing subject and you have a person who is completely disoriented and incapable of carrying out the simplest action. Eliminate further knowledge and his sensory world (his `observation language') will start disintegrating, colours and other simple sensations will disappear until he is in a stage even more primitive than a small child. A small child, on the other hand, does not possess a stable perceptual world which he uses for making sense of the theories put before him. Quite the contrary – he passes through various perceptual stages which are only loosely connected with each other (earlier stages disappear when new stages take over – see Chapter 16) and which embody all the theoretical knowledge available at the time. Moreover, the whole process starts only because the child reacts correctly towards signals, interprets them correctly, because he possesses means of interpretation even before he has experienced his first clear sensation.

All these discoveries cry out for a new terminology that no longer separates what is so intimately connected in the development both of the individual and of science at large. Yet the distinction between observation and theory is still upheld. But what is its point? Nobody will deny that the sentences of science can be classified into long sentences and short sentences, or that its statements can be classified into those which are intuitively obvious and others which are not. Nobody will deny that such distinctions can be made. But nobody will put great weight on them, or will even mention them, for they do not now play any decisive role in the business of science. (This was not always so. Intuitive plausibility, for example, was once thought to be a most important guide to the truth; it disappeared from methodology the very moment intuition was replaced by experience, and by formal considerations.) Does experience play such a role? It does not, as we have seen. Yet the inference that the distinction between theory and observation has now ceased to be relevant, is either not drawn or is explicitly rejected.⁴⁾ 'Let us take a step forward and let us abandon this last trace of dogmatism in science!

4). ‘Neurath fails to give …rules [which distinguish empirical statements from others] and thus unwittingly throws empiricism overboard’, K.R. Popper, The Logic of Scientific Discovery, New York and London, 1959, p. 97.

(pp. 148 – 150)

Against Method, Verso, 1993. ISBN 0-86091-646-4 (Part 15, pp. 147, 148 – 150)