28 
POLITICAL ECONOMY 
the method of watching, directly or in records, 
multitudes of facts with a view to discovering 
their causal relationships, whether it is 
applied to existing conditions or historically, 
everybody is fully convinced in these days. 
But pure induction is not a flawless method. 
It labours under the limitation that, though 
it can declare what is, it cannot unaided 
explain why it is. To explain why a law holds 
we must always have recourse to analysis. 
Nevertheless, induction is indispensable, and 
no more subordinate than deduction. The 
former indicates what needs to be explained ; 
and to discover what needs explanation is 
sometimes far more difficult than to explain. 
Moreover, by separating from the seeming 
chaos of economic phenomena the sequences 
and co-existences which are repeated, it 
suggests uniformities, and thus gives birth 
to the soul of scientific explanation, the 
hypothesis. Further, its aid must be enlisted 
in the testing of hypothesis. 
The broad generalisations of economics, 
as those of every other science, are concerned 
with tendencies. We are not, therefore, 
trying to discover exactly what happens 
at any moment of time, and it would not be 
of much value merely to describe the economic 
world to those gifted with powers of observa-