160 EMPIRICAL TOOLS AND EMPIRICISM 



FINDING THE PROBLEM 



At every stage, eflFective scientific investigation demands more than 

 Method can supply. Beginning at the very beginning, observe that 

 empirical devices are themselves but rarely the fruits of pure empiri- 

 cism. The first creation of an NMR spectrometer demanded highly 

 sophisticated theoretical work. But even the humble Dewar flask is an 

 insulating device shaped to its purpose by a theoretical conception of 

 three modes of heat transfer. A similarly abstract pedigree stands be- 

 hind the startling simplicity of modern chromatographic techniques. 



Waive all such considerations. Suppose ourselves presented at the 

 outset with an abundance of empirical tools. We have then to decide 

 "only" how those tools shall be deployed. In a brand new field, random 

 casts may be sufficiently rewarding: we learn what there is to be seen, 

 and even to learn that certain effects are absent is some gain of knowl- 

 edge. But ultimately we reach the stage of having seen enough, and 

 then we have to ask what is worth observing. Random experiments 

 now become vastly inefficient; from the vast majority of them we learn 

 nothing of value. Thus we come to seek "promising" deployments of 

 our empirical resources. Shall I use my machines to count the 

 pebbles on the beach? Absurd! But why absurd? Because I judge 

 their number irrelevant to any significant problem. 



How identify such a problem? Often a highly individualized ca- 

 pacity seems here involved. Others before Newton had seen, and dis- 

 missed, the elongation of the prismatic spectrum: to Newton that 

 elongation represented a major problem, one through which he ar- 

 rived at a new conception of light and color. For millennia men knew 

 of "fanciful" dreams and "trivial" slips of the tongue: in just these 

 Freud saw a substantial problem. So to detect a significant problem, 

 long concealed under pseudo-explanations or in the trivial, demands 

 intuition. Indeed, often it is the discovery and formulation of the 

 problem that demands insight, whereas the solution demands no 

 more than routine mathematical or experimental skill. The identifica- 

 tion of a problem both solvable and worth solving is then a matter of 

 prime importance. In this identification "experimental method" can- 

 not help us: what we seek is necessarily prior to our experiments. Nor 

 can we look for support to any Method that denies us the use of pre- 

 conceived ideas. 



Problems can emerge only within contexts of presupposition that 



