SECTION 26 DEDUCTION. 377 



too small to be detected with the delicacy of observation pos- 

 sible at the present day. Also, he calculated the magnitude of 

 the curvature of light rays passing the sun at grazing incidence 

 to be 17 seconds of -arc, which has been apparently confirmed 

 by observations recorded during the solar eclipse of 29th May, 

 1919. Finally, Einstein inferred from his theory "a displacement 

 of the spectral lines of light reaching us from large stars, as 

 compared with the corresponding lines for light produced in an 

 analogous manner terrestrially" (op. czY., pp. 103-104), with what 

 success remains yet to be seen. 



199. The extreme form of the deductive procedure, such 

 as Bacon condemned, is to venture on a bold conjecture and 

 to believe that the subsequent deductions will support or cancel 

 it. We have already dilated on the folly and the wastefulness 

 of this method. In scientific deduction we proceed from a 

 genuine hypothesis and endeavour to test it by noting its im- 

 plications. The first virtue, therefore, of a proper scientific 

 hypothesis about to be treated deductively is that it should be 

 in the form of an extremely definite statement, a statement 

 clothed, if possible, in mathematical garb; and, consequently, 

 the published inductive enquiry should not fail to contain such 

 a definite statement or statements. The inductive inquirer 

 must therefore pave the way for the deductive inquirer, and 

 deduction, like induction, should be regarded as a scientific 

 duty, which may not be neglected. Once the latter point is 

 admitted, a deductive code becomes a necessity, and it is prob- 

 able that this will demand that deductions, like observations 

 and generalisations, should be "graded, comprehensive, im- 

 portant, numerous, full, rational and relevant, original, auto- 

 matically initiated, and methodically developed". 1 By insisting, 

 then, that the inductive process should prepare the way for 

 the process concerned with theoretical deductions, as that should 

 prepare the way for deductions of a practical character; that 

 deduction is an integral component part of the scientific process 

 of investigation; and by assimilating its methods to those of 

 the other chief portions of scientific procedure observation and 

 generalisation, we round off our examination of the principal 

 methods employed in the sciences, we resist over-emphasis 

 or under-emphasis of any one of the principal methods, and 

 we secure an endless chain of investigations. 



200. In view of the difficulties ordinarily encountered in 

 deduction, it may seem extravagant to ask that the deductive 

 process should be governed by rules, as suggested in the head- 

 ing of this Conclusion. Yet, audacious as the proposal appears 

 at first sight, it may, we believe, be frequently realised to a 

 considerable extent. In asserting this, we are in a fair way 

 of robbing the process of its mystery and magic, and of ap- 



1 For the full significance of these adjectives, see Conclusion 25. 



