P-, Q-, AND R-TECHNIQUES 207

is to be ascribed to a set of inherent, arbitrary, and inde-
pendent powers ; the second contention (which is put
forward as a consequence of the first) implies a faith, like
that which lasted from Aristotle to Linnaeus, in the defin-
iteness and stability of genera and species. And the
student who reads some of the most recent publications
comes away with ideas that are indistinguishable from the
ancient doctrines of mental faculties and mental types in
the crudest and most primitive form.

As regards the structure of the mind, the more cautious
investigators, factorial as well as non-factorial, seem
unanimous in agreeing that it is anything but fixed, sharply
localized, and rigidly defined. Thomson, as we have seen,
regards the mind as almost structureless • and few writers
would nowadays explicitly assert that it is an aggregate of
isolated powers. Similarly, with the exception of certain
psychiatric writers and the supporters of ' Q-technique/
few modern psychologists, and hardly any factorists, would
now maintain that the human population is divisible into a
number of sharply'distinguishable, mutually exclusive types,
each characterized by some specialized ability, or by some
specialized set of temperamental characteristics, which the
others do not possess: in the few instances in which the
statistical results exhibit some such discontinuous and
multimodal distribution, the effect can usually be explained
by unconscious selection in the initial data.

Those of us who have been directly engaged in practical
work on school children (the field of work which is at once
most freely accessible and most easily controlled) have always
emphasized the changing nature of the factorial picture,
as it is deformed by irrelevant or unintended selection.
The more general factors, which appear so conspicuous
when we are dealing with random samples or complete
age-groups, seem almost to vanish when we apply the
same tests to selected populations, such as pupils from
secondary schools or from homogeneous forms or classes.
On the other hand, the more limited group-factors, which
are so hard to demonstrate in small random samples, often
stand out in salient relief when we turn to a picked batch of
scholarship children or of mental defectives.