398 INTRODUCTION TO PHILOSOPHY OF SCIENCE 



problem reduces to an attempt to show that nothing in 

 science is incompatible with the conclusion thus established. 

 Eddington, for example, expresses himself as opposed to 

 any attempt to base religion on scientific discovery; all 

 that can be concluded is "that the recent changes of scien- 

 tific thought remove some of the obstacles to a reconciliation 

 of religion with science." 1 This destroys, definitely, the 

 intimacy of the relation between science and the larger 

 problems, and suggests that perhaps no inference from the 

 one to the other has any measure of justification. 



For this reason, finally, one should recognize the pre- 

 carious and conjectural character of the solutions to the 

 speculative problems. They are not certainties, nor are 

 they even very high probabilities — if one attempts to 

 measure them according to scientific criteria. There is 

 reason to doubt whether they ought even to be called 

 hypotheses. Logical positivism has denied that problems 

 of speculative metaphysics are philosophical problems; they 

 are either pseudo-problems, indistinguishable from poetry, 

 producing emotional responses but not asserting anything, 

 or else problems lying within the fields of the sciences, and 

 hence scientific problems. 2 This probably represents an 

 attitude as extreme as its opposite, which insists that prob- 

 lems of this kind can be answered with a definiteness and 

 finality which negates the possibility of debate. Both posi- 

 tions are probably wrong. It seems likely that speculative 

 problems are genuine problems, though their terms are 

 extremely obscure as to meaning, and the evidence on which 

 they are based very elusive. The fact that such problems are 

 being linked up with science, if not actually based upon 

 scientific data, argues both for the fact that they are mean- 

 ingful and for the fact that there are data relevant to their 

 solutions. These are important conditions for increasing 

 the probability that one solution rather than another is cor- 

 rect in each case. 



1 Science and the Unseen World (London: Allen and Unwin. 1930), p. 45. 



2 See above, pp. 11-12. 



