F.— ECONOMIC SCIENCE AND STATISTICS 157 



pensable tool for further investigation, and the empiricist, however radical, 

 is likely to flounder if he is unable to use it. In the classificatory work I 

 include truisms like the quantity theory of money, and the wages fund 

 theory, which serve to give precision to the concepts. 



How shall I proceed into this unmapped territory ? At this stage 

 there should be no dispute on matters of principle. On the one hand, 

 for every proposition purporting to relate to the succession of events it 

 must be possible to point to the empirical evidence. Any attempt to 

 assume superior airs may be met with the rejoinder that if empirical 

 evidence is lacking, the proposition can be no more than a definition of 

 the terms which it employs. On the other hand, attention must be paid 

 to the mutual consistency of generalisations and each one must be valued 

 according to the extent to which it contributes to making the whole system 

 more coherent. 



One might draw up a methodological classification by reference to how 

 the investigator spends his day. There is armchair cogitation ; there is 

 the application of statistical technique to the great body of statistical raw 

 material already available, which may well require an elaborate apparatus 

 and assistant workers ; there is the compilation of fresh statistical material 

 by work in the field ; there is also the field work directed to gaining a 

 closer knowledge of how institutions actually work and the motives which 

 govern behaviour. It may safely be said that all these kinds of activity 

 have utility ; they may be regarded as ' factors ' in the production of 

 economic truth to be mixed in due proportions in accordance with the 

 general principles of production ; what is a due proportion depends in part 

 upon the abilities and temperaments of the workers available. I will only 

 add that the institutional arrangement whereby most professional econo- 

 mists are heavily burdened with teaching and administrative duties may 

 militate against a sufficient admixture of the more laborious forms of 

 statistical and field work. The remedy for this, now already in process 

 of application, is the endowment of full-time workers of the right tem- 

 perament and the provision of adequate laboratory equipment and skilled 

 assistants. It may be noticed with satisfaction also that statistical method, 

 on which economic advance depends, has recently displayed a great 

 vitality under the influence of such distinguished pioneers as Dr. Ragnar 

 Frisch. 



There is, however, a more fundamental difference between the outlook 

 of the more and the less empirically minded. This consists of a difference 

 of judgment as to the most hopeful source of clues for the future develop- 

 ment of the subject. On the one hand there are those — I believe that it is 

 fair so to represent the view of Prof. Wesley Mitchell — who believe that 

 clues are most likely to be obtained by the diligent scrutiny, arrangement 

 and rearrangement of the empirical data. The facts will one day speak 

 for themselves. By patient and continuous observation the investigator 

 will find the appropriate generalisation borne in upon him. On the other 

 hand, some believe that clues are more likely to be found by an inspection 

 of the existing body of theory. Close examination of it will reveal gaps, 

 and in those very gaps may be found clues suggesting new generalisations 

 which will render the theory more coherent, or even wider generalisations 



