II 



SCIENTIFIC RECONSTRUCTION 253 



prevent brilliant hypotheses from acting as mere will-o'-the- 

 wisps, nor does it yield a true account of those which have 

 had a more fortunate history. 



A brilliant critic of modern scientific hypotheses has 

 summed up the contrast between the historic fortunes of 

 two different classes of theory in the dictum that < laws of 

 nature are enduring, hypotheses are perishable.' Yet laws 

 of nature are themselves first attained in a tentative way ; 

 that is to say, they begin their career in hypothetical form 

 and they often undergo some modification before they set 

 into their permanent shape. What distinguishes them is that 

 though tentatively formulated on the inverse method, they 

 are proved, not by that method but by direct induction ; 

 that then (at least in the sphere of physics) they are capable 

 of being put in mathematical form, and that in that form 

 they can be corroborated by correlation with similar gener- 

 alisations. The hypothesis, which goes beyond that which 

 can be legitimately generalised from experience, has another 

 origin and a different fate. Ordinarily it contains some 

 sound generalisation within it, but at the same time it 

 endeavours to explain this result by means of some concrete 

 image which is intended to reconstruct the reality on which 

 the result depends. Thus, the same critic points out that 

 the old conception of light, as due to material or quasi- 

 material particles, emitted in straight lines and rebounding 

 from plane or curved surfaces, in accordance with the laws 

 of elasticity, gave a concrete representation of the behaviour 

 of light which embodied, suitably enough, the phenomena 

 in which investigators were then interested, but which, 

 outside the truth which could be accurately generalised 

 from observable data, possessed no validity. The facts of 

 refraction and polarisation necessitated fresh assumptions 

 to make the mechanical model agree with the working 

 experience, but at the same time suggested an alternative 

 image of an etherial medium capable of undulatory move- 

 ments. This theory was so successful as not only to 

 accord with known facts, but also to give rise to predictions 

 which tallied with subsequent observations, while a crucial 

 experiment yielded results which disposed of the rival 

 hypothesis and coincided with the conclusion deduced from 



