238 THE FACTORS OF THE MIND

conceived as moulded into sporadic masses whose forces
introduce irregularities into the movements of other masses
in the neighbouring spatial field; the irregularities in the
spatial field are matter. The same change of standpoint, the
same enlargement of our ' scope of apprehension ? till it can
think in terms of complex patterns instead of isolated units,
is required in psychology ; for there, as elsewhere, what the
observer must look for and judge are not the spots of paint,
but the picture—the canvas behind the picture remaining
for ever concealed from his vision.

I conclude, therefore, that in describing mental life the
psychologist is driven to the same position as the physicist
has reached in investigating the material world.1 Once we
have left the field of applied psychology for that of psycho-
logy as a pure but empirical science, we are no longer justi-
fied in assuming a universe of individual ' objectsJ or
stimuli acting on an individual ' subject? or mind. The
objects or stimuli prove to be merely Gestalten ; so are the
minds; so is the experience which the minds have of the
objects. We are reduced to the study of a changing structure
of relations linking two sets of systems; and these systems
themselves are for the scientist simply structures of relations.

Of isolated minds, then, in and for themselves, the em-
pirical psychologist can know nothing; equally he can do
nothing with isolated test-stimuli such as the early psycho-
physicists supposed they were using; and even the relations
between minds and test-stimuli can be described in their
formal aspect only. Here it is that we find at once the
reason and the justification for using that peculiar mode of
mathematical analysis which the psychologist, like the physi-
cist, has recently been led to adopt. This is why, as I have
already hinted, we have been gradually driven to measure
both stimuli and reactions not by single figures, but by
matrices of figures : for a matrix is essentially a device for

1 This position had already been reached before the advent of relativity
and quantum-theory had led to its general acceptance: e.g. by Russell
(Principles of Mathematicsy 1902,1, p. 468); " The only relevant function of
a material point is to establish a correlation between all moments of time and
certain points of space. . . . Thus a material point may be replaced by a
many-one relation." . , . ' ' \ ,.