66 THE FACTORS OF THE MIND

outward and visible effects : the factors (in Spearman's
phrase) are the " hidden underlying causes." l

No doubt, this causal language, which we all to some
extent favour, arises partly from the irrepressible disposition
of the human mind to reify and even to personify whatever
it can—to picture inferred reasons as realities and to endow
those realities with an active force. But, in this particular
reference, it is still further strengthened by the phraseology
of most statisticians, who have nearly always discussed
probability in terms of hypothetical * causes/ From
Laplace onwards, scientific writers on the theory of proba-
bility have regularly represented it as a procedure for
c inferring events from causes? or inversely for * ascending
from events to their causes.3

Thus, Bravais's original deduction of the product-
moment function proceeded on this basis. Or again, to cite
one of the earliest textbooks of statistics to introduce the
subject of correlation, Bowley,2 after expounding the

general conception, formally deduces the equation r = —2—

and then concludes that " expressed in words the formula
shows that the correlation coefficient tends to be the ratio
of the number of causes common in the genesis of the two
variables to the whole number of independent causes on
which each depends." More recently, Fisher has given an
identical formula to illustrate the ( analysis of variance'
into * portions contributed by the two causes'—the
' common ' cause and the non-common : " in such cases,"
he writes, " the correlation merely measures the relative
importance of two groups of factors catising variation "
([50], pp. 212, 210). Brown and Thomson have deduced

1 Cf. loc. cit. sup., p. 4. Actually Spearman's own system of factors
has been criticized by Thomson because, though it " gives an admirable
description of correlation data" (i.e. of " certain types of normal and
abnormal persons "), " it does not give the causes " ([87], p, 64 : italics as
in the original). The latter, it is suggested, are supplied by the more
numerous and more elementary components, which are ** not new entities,
but things we already know of," and in themselves far more simple " aspects
of the causal background.'7

2 Elements of Statistics, vol. II, p. 336 (my italics). See Keynes' trenchant
criticism of this passage [43], p. 425.