EMPIRICAL TOOLS AND EMPIRICISM 161 



define the "ordinary" or "natural." Lacking such a context everything 

 and nothing in our experience is a source of wonder and perplexity, 

 everything and nothing is a problem. Given such a context, a problem 

 becomes recognizable as a particular something that seems "unnat- 

 ural," "incomplete," "irregular," "unsatisfactory." Thus, for example, 

 the retrograde motion of the planets was a problem for the ancients 

 who believed in the perfect uniformity of celestial motions, and not so 

 very long ago the photoelectric eflFect posed a problem for those who 

 accepted the view of classical electrodynamics. To be sure, if our 

 theoretical preconceptions are wrong, they may lead us to study 

 problems as unrewarding for us as were for them the chemical prob- 

 lems studied by the alchemists. Outlawing preconceived ideas, 

 Method would save us from such error; but then we could not even 

 begin the work of science. The "illicit" context of presupposition dis- 

 charges the indispensable function of making the difficulty that con- 

 stitutes a problem, and so first permits the initiation of inquiry. 



Not all difficulties pose acceptable problems. Thus I refuse to treat 

 as a problem some "odd" datum I consider likely to prove no more 

 than a trivial experimental error. Making such peremptory dismissal, 

 I draw again on (potentially fallible) theoretical presuppositions. 

 And, if I am to advance any farther, I must now bring into play 

 not only these but some very highly speculative hypotheses as well. 

 Among many possible problems, I can first select one, as a good 

 problem, only as I exercise my imagination to guess its answer. That 

 premonitory intuition, or "educated guess," is then required to sup- 

 port all subsequent stages of inquiry. Thus, attacking my problem, I 

 seek first to collect the already-available data relevant to its solution. 

 Even to begin that collection, I must use a criterion of relevance that 

 finds its sole foundation in my own hypotheses about the form(s) 

 solution to my problem will take. Is the orientation of the tails of 

 comets relevant to the study of light? It first becomes so with the 

 conception of an hypothesis associating that orientation with a "light 

 pressure." 



THE DESIGN AND CONDUCT OF EXPERIMENTS 



Proceeding to experiment, I confront the problem of relevance in a 

 new and aggravating form. Abstracting a simple experimental system 

 from the complexity of natural events, what assurance have I that my 

 system in any way "represents" the situation in nature? Is the be- 



