142 SECTIONAL ADDRESSES 



to determine a structure of output and prices. If a change in these data 

 occurs, the theory professes ability to predict the consequences within 

 certain Hmits on the price-output structure. This professed abiHty to 

 predict impHes that we have available certain general laws concerning 

 the succession of events, causal laws in fact. Rigid demonstrability and 

 certainty, of an almost geometric kind, are claimed for them. Since the 

 laws concern the succession of phenomena they must have an empirical 

 basis ; and since the phenomena of economics are notoriously highly 

 complex and unamenable to scientific handling, it is a paradox that the 

 laws derived from their study should have the high degree of certainty 

 claimed for them. 



The paradox is resolved when we consider that the laws in question 

 are deducible from a single simple principle (Robbins), itself based on 

 experience, but on an experience far wider than that vouchsafed by the 

 study of markets and prices and extending back to the earliest phases of 

 man's self-conscious existence — namely, the Law of Diminishing Utility 

 or the Law of Demand, to be defined more precisely presently. The 

 experience is so broad that the principle may be taken as an axiom of the 

 highest possible degree of empirical probability. 



But against this very high degree of probability of the principle and the 

 laws deduced from it must be set their complementary degree of generality. 

 The degree of generality is indeed so great that, I shall submit, the power 

 of prediction vouchsafed by them is almost nugatory. 



Next, economists, even the most theoretical, have been prone to give 

 advice on the basis of theory. And I believe that economists would claim 

 that much of the advice so given since Adam Smith has been valid. A 

 type of the advice I have in mind, though this by no means covers the 

 whole field, is the recommendation of free trade. Now it will at once 

 occur to the critic to ask how, if it is true that the laws of value and dis- 

 tribution are so general that they yield but a nugatory power of prediction, 

 can a quite copious array of advisory propositions, admittedly based on 

 pure theory, be justified. 



The reply is that these prescriptions are based on the other department 

 of what is commonly regarded as pure theory. They are derived from the 

 maxim that productive resources should be so distributed among occupa- 

 tions as to yield an equi-marginal social net product. The nature and 

 justification of this maxim must be considered. 



In order to derive from it precepts, which are applicable in the real 

 world, certain knowledge about that world is necessary. This knowledge 

 does not, however, relate primarily to causal sequences, nor does it consist 

 of a bare enumeration of particular features and events. It arises rather 

 from a simultaneous chart or survey of the -economic field and the main 

 work of the cartographer is analysis and classification. This analytical 

 work is required both as a preliminary to the construction of the map and 

 to the derivations of specific causal laws from the law of demand. I 

 venture to submit that it is this identity of the preliminary groundwork 

 which has tended to obscure the fundamental distinction between the 

 set of conclusions which relate to causal sequences and involve predictory 

 power on the one hand and the comprehensive but simultaneous con- 



