THE LOGICAL STATUS OF MENTAL FACTORS 133

failure of the utilitarians, any attempt to treat pleasures or values
as absolute magnitudes must be regarded as " prima facie fallaci-
ous." Nevertheless, he may, I think, legitimately reply that an
absolute or extensive measurement of values is not what he wants :
relative preferences can still be treated quantitatively, if they are
regarded as distensive magnitudes. If my guide book tells me that
Sir Richard Wallace paid ,£600 for a painting by Boucher and ^550
for one by Greuze, but declined a painting by Manet offered at
£300, I cannot deduce that his absolute love for Boucher was more
than twice as great as his absolute liking for Manet; but I can justly
infer that the gulf between his preferences for the impressionist style
and for the style of Louis Quinze was much greater than the minor
variations among his fancies for the ' school of pink and pale blue.'
However, purchase price is at best but an indirect measure of
what the economist calls c intrinsic' or subjective value. Hence,
it seems desirable to seek a more direct method of measuring the
latter. Without putting our examinees to the equivocal test of an
actual auction, we can hand them reproductions in postcard form,
and so elicit rankings or gradings, which will measure, if the experi-
ment has been properly planned, not indeed their feelings on any
absolute scale, but the differences between their preferences.
Similarly, we can get children to rank their preferences for different
school subjects ; and so compare their preferences for each with
their achievements in each.1 Formally, it would seem, all such
gradings are quite as valid as the more familiar graded judgments
on ability or skill.2

1 Burt [29], [69], p. 278. Here again, however, the proper procedure
has been the subject of some controversy: cf. Stephenson, Brit. J. Educ.
Psych., V, pp. 43 f.

2 As a means of studying aesthetic preferences, the e method of choice' is
as old as Fechner. Witmer seems to have been the first to apply Galton's
notion of ranking (or, as he terms it, * method of regular arrangement *) to
the aesthetic field (Phil. Stud., IX, 1894, pp. 96 f.). CattelTs work, however,
gave the ranking method its great popularity. What recent critics appear to
doubt is, not so much the validity of the ranking method in itself, as the
validity of factorizing correlations obtained from such data.

The view of value as a distensive magnitude seems fully in keeping with
modern economic theory. " The Utilitarians thought of absolute value as a
quantity—a simple sum of values viewed as atoms of pleasure, which they
treated arithmetically. Once the atomistic view is departed from, and
economic value becomes the expression of preferential relations, the measure-
ment of absolute value is no doubt impossible. . . . But the curious thing
is that, though absolute value is not conceivable as a quantity, or is barely
conceivable as a quantity, economic values are conceivable as quantities, and