VARIOUS USES OF FACTORS 67

a similar equation, and base their interpretation of the
hierarchy upon it ; as a rule, however, they prefer to speak,
not of causes, but of ' elements or factors? ([39], p. I/6).1

Accordingly, it is not surprising if psychologists should
treat the c common factors ' of one writer as synonymous
with the c common causes ' of another, or that they should
go on to identify the * factors' by which the future actions
of their testees can be predicted with concrete ' abilities?
or * mental energies ' which are conceived as effective causal
agencies determining such actions.2 Some of them ex-
plicitly declare, as we have already seen, that the ' funda-
mental tendencies' revealed by correlational analysis
" account for, explain, and are the cause of, all human
conduct."

Perhaps the most explicit expression of this view is
contained in CattelPs article cited above. The alternative
view—that c different kinds of properties, belonging to
human characters, can be distinguished, but not separated/
and that * the attribute is not a part of the concrete indi-
vidual, but only an aspect'—he emphatically rejects,
classing it, with the ' tenets of much Struktur and Gestalt
psychology,' as ' nothing less than a denial of the validity
of scientific method.3 Instead he argues that factors are
* psychological powers? which can be measured, " like the
power of a muscle," ..." in interactionist terms of the
real effects of those powers upon the physical and social
environment," and are " eventually expressible in terms of
energy transactions " (loc. cit,, pp. 127 et seq^).

The more recent statistical textbooks, it is true, abound
in warnings to the student not to accept mere correlation

1 Cf. Brown, Mental Measurement, 1911, p. 79 (where the equation is
deduced as a generalization of Weldon's experiment illustrating correlation
by combinations of dice). Cf. also p. 52 above.

2 The transition is made possible by the ambiguities that lurk in the word
' cause.' The more cautious statistical writers expressly state the broad
meaning they attach to the term. Thus Coolidge writes : " We shall
mean by the cause of an event any antecedent event whatever" (Pro-
bability, 1925, p. 88, his italics). Even Coolidge, however, goes on to talk
of causes as c operative ' and as ' producing results' (p. 88); and thus seems
to forget his own definition and to attribute to his ' causes' powers that
his initial definition does not include.