2i4 THE FACTORS OF THE MIND

Now, though I do not identify the general factor g with
any form of energy, I should be ready to grant it quite as
much ' real existence ' as physical energy can justifiably
claim ; nor do I feel that such a concession would in any
way conflict with Thomson's own explanation of g in terms
of the sampling of a structureless aggregate of numerous
neurones or their bonds. The solution of factorial equa-
tions is not the only thing with which the factorial psycho-
logist is concerned : he may seek evidence for the ' reality '
of his concepts in many other fields. Hence I should rather
put the position in this way. The mathematical part of the
factorial argument is for the most part deductive ; and
therefore, like all deductive reasoning, is admittedly unable
to guarantee the reality of the results deduced. In such an
argument the factors are notions postulated, not things
directly observed or measured. As a mathematician the
factorist merely asks : what are the fewest and the simplest
postulates I can make in order to describe the phenomena I
am observing ? And from these postulates he tries to
reconstruct the facts observed. But, when he has completed
his reconstruction, he seeks to check and verify his inferences;
and so turns again to the empirical world.

Deduction, whether mathematical or non-mathematical,
can never provg the existence of the concepts with which it
deals. Induction is therefore required at two stages: first,
at the beginning, to suggest the initial postulates; and
secondly, at the end, to see whether the empirical facts
answer to the corollaries deduced.

This alternating procedure is not peculiar to factorial
psychology. It is followed in nearly every quantitative
science. In thermodynamics the physicist will first assume
a perfect gas or a perfect engine, described by a few simpli-
fied postulates ; he then deduces that the volume of a gas
varies inversely with pressure or that with constant volume
its specific heat is constant; finally he examines the results
of experiments to see how far his deductions hold true of
actual gases and actual apparatus. He would be surprised,

selective operator cannot guarantee any concrete 6 existence' for the class
c selected.' However, it would take us too far afield to discuss what precise
meaning (if any) can be attached to the vague term * existence/