174 SECTIONAL ADDRESSES 
by which systematic principles and explanatory concepts come to be 
formed. And it is the most empirical of all the sciences, since the con- 
cepts of which it makes use are drawn directly from within experience 
itself, or, if inferences from it, are the least remote of all. I shall hope to 
illustrate this by reference to several of the explanatory concepts actually 
in use in physical and biological science and in psychology. But before 
doing this it will be useful to recall and distinguish the several stages by 
which science proceeds and in which such concepts are reached. 
DATA OF SCIENCE. 
The first step taken in any empirical science is to examine, describe, and 
classify the objects, or aspects of objects, with which it deals; such 
classification being made on the principle of similarity and difference, 
which, it may be noted, does not involve inference, but depends upon the 
immediate experience of relations. The first step in psychology will 
accordingly be to observe, describe, and classify mental processes.as such. 
Psychologists are fairly well agreed on the broad classification of these 
processes under the three heads of cognition, affection and conation, or 
knowing, feeling and willing, as aspects or actualisations of the self. But, 
though I have used the terms synonymously, I may note that there appears 
to me to be good evidence that conation (striving and doing) and willing 
(resolving, intending, choosing) cannot be included in the same general 
category ; and, accordingly, that there are at least four broadly irreducible 
kinds of mental event, which will require four groups of concepts to explain 
them. Classification, however, does not merely mean grouping together : 
it means separation as well. ‘Thus cognitive processes separate into 
sensory perception, conception, judgement, reasoning and remembering, 
for each of which a different explanatory concept may be needed. Though 
memory, for instance, may be involved in perception, we cannot explain 
remembering and perceiving in the same way. Incidentally, the postu- 
late of retentivity is a good example of the kind of inference made in 
psychology. It is evidently an inference, for we do not experience 
retentivity. But it is not a remote inference ; and does not become so 
until we further postulate some such thing as persistent brain traces to 
account for it. Similarly different concepts may be necessary to explain 
the experiences of desire, resolution, impulse and striving, whether they 
are classed under two heads or one, and no matter how closely one may 
be involved in the other. 
STRUCTURAL ANALYSIS. 
The next step consists in the finer structural analysis, so far as this is 
possible, of the phenomenological data. In psychology, this means the 
further splitting up of the products of mental processes. On analogy 
with the procedure of the chemist, who analyses a chemical compound 
into its constituent elements, or of the anatomist, who dissects out the 
fibres of a nerve trunk, the psychologist analyses a percept, memory, 
emotion or will-act. The proverbial seen orange yields in such an 
analysis sensory factors of an elementary kind—colour, odour, sapidity, 
smoothness, and the like. It cannot be said, however, that these are all 
