METAPHYSICAL STATUS OF FACTORS 243

and even psychologists, are still very apt to assume that what
is called the * problem child' has only to be brought to the
psychologist's consulting-room, tested, interviewed, and
observed as it were in vacuo, and the examiner can then
pronounce what is wrong with the child.

In the past this tacit assumption was responsible for at
least half our failures with the neurotic and the delinquent.
The ' problem ' never lies in the ' problem child' alone :
it lies always in the relations between that child and his
environment. Neither the delinquent child, nor yet the
nervous child, nor even, as a rule, the backward child, can
be properly understood, unless the examining psychologist
has investigated, not only the child himself, but the
conditions under which he is living at home, at school, and
wherever he spends his leisure hours, and so is able to gauge
how the child, on the one hand, and his parents, friends, and
teachers, on the other, are constantly interacting.

This seems to me to be especially important in obtaining
reports on the child's character. Not only for purposes of
research, but also for purposes of clinical diagnosis and
vocational guidance, the common practice is to request
observers to grade persons according to their supposed
emotional or moral qualities. But, when we correlate
such gradings, we usually find that they throw light on
very little else besides the observers' own implicit views.
What we ought to grade is the overt behaviour of the per-
sons under review, not the presumable qualities of their
minds ; and behaviour consists essentially in relations—
in relations between the person and the conditions under
which he lives.

A striking example is to be found in the so-called human
instincts, about which so much controversy has been waged.
From the standpoint of the individual psychologist, in-
stincts are * factors/ not genetic factors but descriptive
factors, factors in the logical rather than in the biological
sense: they are " little more than convenient headings
under which certain reactions to certain stimuli can be
recorded." Whether we are to postulate an inherited set
of quasi-reflex mechanisms to explain such c instinctive'
reactions is a problem belonging to an entirely different