64 TWENTY-FIRST REPORT. 



possibility of adequately expressing them by any classification into types. 

 This would give data bearing on Thorndike's contention that we have either 

 but one type, or as many types as individuals. 



The tests used will be described in detail in a later publication. They 

 were arranged to give numerical scores for each kind of imagery for each 

 individual. 



Test A, for the clearness of concrete imagery, is a modification of the 

 Galton questionnaire, with a different method of scoring, in that each image 

 was compared not to the clearness of a percept of the object (which is impos- 

 sible in most cases), but to the clearest images of the series. This involves a 

 partial application of the order-of-merit method, which tends to eliminate the 

 tendency of some subjects to overestimate all their images. 



Test B, for dominance of concrete imagery, is a modification of what 

 Titchener calls the Secor method, which gives scores in terms of percentage. 

 This was modified so that the scores represent the actual dominance in atten- 

 tion of the different kinds of imagery instead of the mere frequency of each 

 kind. When more than one kind of imagery was aroused, the subject was 

 asked to state the relative importance or dominance of the two by a propor- 

 tionate "distribution of seven points." The number of points given to each 

 kind of imagery was reduced to percentages of the total number of points for 

 ail kinds. Seventy-five words and forty sentences were used. 



Te.st C, for dominance of verbal imagery, also gives percentage scores. 

 The verbal imagery in memory of both orally and visually presented material, 

 of poetry, in writing, silent reading, etc., was used as a basis for the "distri- 

 bution of seven points." It was found best to ask the subject to make a first 

 division into visual (verbal) and inner-speech, and a second division of the 

 inner-speech into its auditory and kinaesthetic elements, making a distribution 

 of seven points in each case. From these values scores for visual, auditory and 

 kinaesthetic verbal imagery were computed. 



Teit D is a visualization test, the scores being the number of seconds 

 required to successfully answer a series of problems including the well-known 

 "three-in^i-cube" problems. 



Test E is an objective test for the detection of inner-speech, and particu- 

 larly of the auditory element, in verbal imagery. Lists of six letters and 

 digits were read to the subject, who repeated them in reverse order. Half of 

 there lists contained several like-sounding letters, the other half contained 

 none. These are called "L" and "U" lists. The score, or the "U-L" is the sum 

 of the differences in time and in accuracy of reproduction between the two 

 kinds of lists. 



One hundred and twelve subjects who were beginning students in psychol- 

 ogy were tested. 



