430 THE FACTORS OF THE MIND

of the academic psychologist. In the intellectual field the notion
of discrete types has long ago been abandoned. By medical
writers, it is true, one still occasionally finds the mentally defective
described as " forming a special type apart to be sharply distin-
guished from the dull, who are merely unintelligent normals."
But careful surveys by means of mental tests have shown—I
think conclusively—that " the one group merges continuously into
the other; there is no gap. . . . Apart from rare pathological
cases, the mentally defective form simply the extreme tail of a
continuous normal distribution." l

In the temperamental field similar surveys are urgently needed.
The only extensive inquiry that I know is the oft-quoted investiga-
tion carried out over 30 years ago in Holland under Heymans
and Wiersma.2 Four hundred and fifty Dutch physicians were
induced to make exhaustive reports on 2,523 individuals in accord-
ance with a systematic questionnaire. The analysis of their data
led to an eightfold classification of temperaments, based on three
principles of division. Of these the most important was the
contrast between high and low ' secondary function '—a persevera-
tive tendency which was found to be high in persons of a melancholic
type and low in persons of a manic type. The distribution was
apparently bimodal: of the individuals studied less than 8 per cent,
remained unassignable to one type or the other—not because they
belonged to an intermediate category, but simply because in these
few cases the available evidence left the true assignment doubtful.
Suggestive as this research has proved, it would now, I imagine,
carry little conviction as a study in the frequency-distribution of
types. Its methods have been strongly criticized by Spearman ;
and its results can hardly be accepted as throwing genuine light upon
our present inquiry.

The investigation of special cases and special problems,
referred to me during my work as Psychologist for the Lon-
don County Council, has incidentally yielded a vast accu-
mulation of material that lends itself to analysis along the
lines described in the foregoing pages. During thirty

1 For the evidence, see my L.C.C. Reports on Tike Distribution and.
Relations of Educational Abilities, 1917, pp. 34 et seq., and on Mental and
Scholastic Tests, 1921, p, 163. Perhaps the most thorough criticism of the
traditional view of types is that of Thorndike Educational Psychology, III
(1923), chap, xvi: (temperamental types are not referred to as such, but they
are tacitly covered by Thorndike's general conclusions).

2 Zeitschr, /. Angezvandte Psychologie, " Beitrage zur Spezieller Psycho-
logic," I, 1908, pp. 313-83, et seq.