390 INTRODUCTION TO PHILOSOPHY OF SCIENCE 



is that science gives hints of facts which are not themselves 

 scientific. Accordingly, by a careful examination of the 

 data within science, one may be led by an inference of 

 greater or lesser cogency to a conclusion applying outside 

 of science. In all of the problems, therefore, three features 

 must be made explicit: (a) What, precisely, are the data 

 from which the inference starts? (b) What is the nature of 

 the inference itself? (c) What is the character of the con- 

 clusion inferred? With reference to (a), it will be found that 

 the data, in so far as they are drawn from the sciences, are 

 not usually the specific facts of science itself. For example, 

 it seems unlikely that the fact of gold melting at 1075° 

 would have any important implications outside of science. 

 It is rather the more general facts which seem to have this 

 inferential value, e.g., the general usefulness of mathematical 

 symbols in physics, or the inevitably analytic character of 

 scientific symbols, or the fact of an unavoidable uncertainty 

 in the precise location of very small particles. Yet the line 

 of demarcation between the general facts and the specific 

 facts cannot be drawn sharply, and it is likely that certain 

 relatively specific facts may have important speculative 

 implications. For example, the permanent deformations in 

 structure which certain metals undergo when subjected to 

 excessive strains — the phenomenon known as hysteresis — 

 may afford justification for the contention that there is in 

 the inorganic realm something analogous to memory, and 

 hence that the line between the organic and the inorganic 

 cannot be sharply drawn. With reference to (6), the char- 

 acter of the inferential route must be made clear. Is the 

 conclusion a definite implication of the data, i.e., deducible 

 from them, or is it merely an hypothesis having a greater 

 or lesser degree of probability? Is it obtained through the 

 method of construction, or through the method of hypoth- 

 esis? If it is obtained by the latter method, is it based upon 

 analogy, interpolation, serial extension, or some other recog- 

 nized inductive technique? With reference to (c), the 

 nature of the inferred entity must be made as clear as it can 



