THE LOGICAL STATUS OF MENTAL FACTORS 121

weight, number, and perhaps a few other physical properties can
form " part of the subject-matter of science," because for these
judgements alone " universal and impersonal assent can be
obtained " l—a statement which might rather surprise lay members
of the British Association who had dropped in at certain meetings of
Section A.

Much the same reasoning has been used by psychologists them-
selves when discussing the assessment of persons or of personal
* behaviour—for example, in their criticisms of the attempt to rank
different individuals for the same mental qualities and (more often
still) the more recent attempts to rank different qualities for the
same individual. But with all such problems the first essential is
to make the experiment. And this is precisely what the factorist
proposes. He starts by denying the tacit assumption that the only
alternatives are either * no agreement' or ' universal assent.'z
Between psychologists, and presumably between other men of
science, agreement may be of varying degrees ; and these degrees
the factorist can easily assess by a coefficient of correlation. To his
surprise he finds, with a suitably chosen set of pictures, quite as high
a consensus of opinion among art critics about their relative beauty

1 Loc. cit., p. 273, Cf. Brit. Ass. Report on Quantitative Estimates of
Sensory Events, p. 16 : " physics is the discovery and study of those relations
between sense-perceptions concerning which, universal agreement can be
obtained."

2 It is not my object to criticize Dr. Campbell's admirable exposition.
From his later chapters on * errors of measurement' it is clear that the
statements quoted above from his chapter on£ the first conditions of measure-
ment ' were expressed in an unqualified form solely in the interests of lucidity.
Nevertheless, these statements have frequently been cited by others to
support the view that measurement in psychology is impossible. In an earlier
chapter still he explicitly recognizes " degrees of knowledge " ; and suggests
that the " degree of knowledge is measured by the subjective mental dis-
comfort we should suffer if we found it was not true " (p. 160). When
endeavouring to show the objective or impersonal character of judgements
of beauty, I carried out parallel experiments on judgements of truth. In
both cases, according to the introspections, £ degree of mental discomfort'
was a common criterion. But, even with experts in each of the fields, the
amount of agreement (measured by a so-called * reliability coefficient')
was often lower for scientific opinions than for aesthetic. If (as one reader
suggests) it might be better to " base our criterion of what is, or what is not,
the subject-matter of science, not on agreement between experts, but on
agreement with the facts, i.e. power of prediction," then I may add that the
psychologist's prediction of what a child will do in certain tests to-morrow is
far more accurate than the meteorologist's predictions of what to-morrow's
weather will be.