J.—PSYCHOLOGY. 233 
Character has been defined as the sum-total of all those individual 
qualities which do not constitute, or are not pervaded by, intelligence ; 
to avoid the specifically moral implications that cling to the popular 
word ‘ character,’ I prefer to retain the old term ‘ temperament,’ and 
use it in the sense defined. ‘The qualities thus negatively grouped apart 
are not without a positive aspect shared by them all. Though they 
exhibit low correlations with intelligence, they yet show tolerably high 
correlations amongst themselves. Analytically, they are marked by 
affective and conative elements rather than by cognitive; by feeling 
rather than by knowledge; by will rather than skill. 
Temperament or character is always more difficult to assess than 
intelligence. Intellectual qualities are relatively constant. Emotional 
qualities are evanescent and evasive—hard to seize, and harder still to 
measure. It is significant to note that, though the idea of 
temperamental testing is almost as old as that of intelligence-testing, 
it has seen quite a different career. Hyery one has heard of Binet’s tests 
for intelligence. But most of us have forgotten his efforts to measure 
suggestibility, conscientiousness, and fidelity of report. Of late re- 
newed endeavours have been made to test the feelings and the will; 
and of these the most effective are the methods of associative reaction 
and the so-called psycho-galvanic reflex. Pressey has tried to detect 
fears aud repulsions by getting the examinee to pick out, from a pre- 
arranged list of words, those that have for him a special meaning, or 
suggest a special worry or dislike. Downey has tried to measure what 
she terms ‘ will-temperament’ by seeing: how far the candidate can 
modify at request his style of handwriting and manner of speech. 
Fernald measures self-control by the time the candidate can balance 
himself upon the ball of the foot. The Porteus mazes are to some 
extent a test of recklessness and impulsiveness. And the variability 
in repeated tests of almost any simple kind (as measured, for example, 
by the standard deviation) seems partly correlated with instability. But 
no tests of temperament can claim to have passed beyond the stage of 
tentative experiment.** 
In assessing temperament, therefore, we must fall back upon the 
method of observation in place of the method of experiment. The 
personal interview is one recognised device; and another is the collation 
of reports submitted by competent observers who have been acquainted 
with the examinee during a long portion of his life. Both interviewing 
and reporting has each its own technique; and in either case the 
technique is susceptible of great improvement by the application of 
simple scientific principles. Much, indeed, has already been done by 
drawing up questionnaires of facts to be noted and observed,” and by 
24 A good summary of the literature, with a detailed bibliography, will be 
found in Cady’s article on ‘ The Psychology and Pathology of Personality : A 
Summary of Test-problems,’ /. Deling., vii., 225 (1922). 
25 Of these perhaps the most suggestive are those given by Webb, ‘ Character 
and Intelligence,’ Brit. 7. Psych. Mon., I., and Hoch and Amsden, ‘ Guide to 
the Descriptive Study of Personality,’ Rev. Neur. Psych., xi., 577. Cf. Psych. 
Rev., xxi., 295 
