DERIVATION OF CHIEF THEORIES 167

incidental and inconvenient consequences of our imperfect
methods of measurement. This implies that we must recog-
nize a marked difference of status between the general and
the group-factors (i.e. the ' common' factors), on the one
hand, as compared with the specific and the chance factors
(i.e. the c individual' factors), on the other. For the
former we may seek a concrete psychological interpretation;
the latter will have none.

Relative Nature of the Distinctions.—All through these
discussions, however, there is one point which psycho-
logists seem generally to have overlooked, but which to the
logician will appear self-evident. As I have endeavoured
to insist in all my writings, "the differences throughout are
principally differences of degree : the ' general factor ? is
simply the * group-factor* that is of the most widespread
occurrence ; and the ' specific factors' are simply the
6 group-factors ' that are most narrowly limited in their
operation." x Give me a list of tested traits which are
said to be governed by a general or universal factor only
such as Spearman's g : I can always add one or more tests
or traits which do not contain that factor, and so reduce
the general factor to a group-factor. Name any specific
factor, said to be peculiar to one tested trait alone : I can
always add one or more slightly different tests or traits,
guaranteed to contain that factor (or the constant elements
of it), and so convert that specific factor into a group-factor.
Similarly, as Thomson has pointed out, in many of the
investigations claiming to show that only a single general
factor exists, and no group-factor, the investigators have
frequently begun by " pooling all similar tests" (as they
put it) ; so that what their critics would claim as group-
factors are reduced to the status of specific factors. Thus
the distinctions between general, group-, and specific factors are
formal rather than material, relative rather than fixed.

The modern logician would be the first to remind the psychologist
that genus and species are " not absolute terms, but purely correla-
tive/' " The same term may be at tke same time a genus to the
lesser classes it contains, and a species of the next more general

], p. 19 J cf- [46]> P- 23°-