Bertrand Russell The Relation of Sense - data to

PHYSICS is said to be an empirical science, based upon observation and experiment. It is supposed to be verifiable, i.e. capable of calculating beforehand results subsequently confirmed by observation and experiment. What can we learn by observation and experiment ? Nothing, so far as physics is concerned, except immediate data of sense: certain patches of colour, sounds, tastes, smells, etc., with certain spatio-temporal relations. The supposed contents of the physical world are prima facie very different from these: molecules have no colour, atoms make no noise, electrons have no taste, and corpuscles do not even smell. If such objects are to be verified, it must be solely through their relation to sense-data: they must have some kind of correlation with sense-data, and must be verifiable through their correlation alone. But how is the correlation itself ascertained ? A correlation can only be ascertained empirically by the correlated objects being constantly found together. But in our case, only one term of the correlation, namely, the sensible term, is ever found: the other term seems essentially incapable of being found. Therefore, it would seem, the correlation with objects of sense, by which physics was to be verified, is itself utterly and for ever unverifiable. There are two ways of avoiding this result. (1) We may say that we know some principle a priori, without the need of empirical verification, e.g. that our sense-data have causes other than themselves, and that something can be known about these causes by inference from their effects. This way has been often adopted by philosophers. It may be necessary to adopt this way to some extent, but in so far as it is adopted physics ceases to be empirical or based upon experiment and observation alone. Therefore this way is to be avoided as much as possible. (2) We may succeed in actually defining the objects of physics as THE RELATION OF SENSE-DATA TO PHYSICS IO9