NATURE OF THE FACTORIAL TECHNIQUE 83

Those who regard factor-analysis as the hobby of a special
school, which the ordinary student, with no taste for num-
bers, can safely ignore, may be reminded that almost
exactly the same procedure has found a reference in nearly
every psychological textbook ever since textbooks began to
incorporate experimental work. For the ordinary student
of psychology, perhaps the simplest, earliest, and most
familiar examples of this kind of analysis are (i) Wundt's
attempt to determine * the number of elementary feelings *
and (ii) the attempts of Helmholtz and other experimental-
ists to determine the number of elementary or ' primary'
colours. Though the phrase was not explicitly used, the
object of such efforts was to reduce the phenomena of feel-
ing and of colour vision to terms of ' orthogonal factors.'
Wundt, it will be remembered, " in developing a compre-
hensive theory of feeling, postulated three dimensions:
(i) excitement-quiet (c depression'), (2) tension-relief,
(3) pleasantness-unpleasantness. The total feeling at any
moment can be located by reference to each of these
dimensions, just as a point on the earth can be identified
by latitude, longitude, and altitude." 1 It is interesting to
note that the nature and independence of these three
6 dimensions? are in some measure confirmed by recent
endeavours to analyse the complex temperamental charac-
teristics of individuals, by means of a modernized factorial
technique,2 though Wundt himself based his analysis rather
on introspection and the study of emotional expression than
upon a formal mathematical analysis.

The determination of the ' laws * of colour mixture has a

1 Wundt, Grundriss der Psychologic (1896), pp. 98 f. Woodworth, Experi-
mental Psychology (1938), pp. 235 f.

2 In particular by the statistical evidence for excitable, repressed, and
cheerful c types' or tendencies and the reverse : cf. [114], [129]. Wundt's
own analysis of temperamental types was two-dimensional rather than tri-
dimensional; but his tri-dimensional theory of the feelings seems to have had
considerable influence on later German attempts to classify temperaments
(cf. [21], pp. 188, 483 f,). Wundt's pupil, Titchener, not to mention
many other critics, has adduced strong reasons for doubting whether
the " dimensions are really * independent' (as naively understood) " (Phil.
Stud., XX, pp. 382 f.); yet until recently no attempt seems to have been
made to test the * independence' by the obvious method of correlation.