^56 



SCIENCE. 



[Vol. XVI. No. 39S 



somewhat different phase of this influence appears in the 

 results of Vintschgau. In multiplying the numbers from 

 IXl to 9X9, the smaller number was always announced 

 first. Accordingly, when the first nine was announced, the 

 subject practically anticipated the result, and had the prod- 

 uct ready; when eight was announced, he knew that it was 

 one of two results ; when seven, one of three; and so on. 

 Accordingly we find these to be the shortest processes (9X9, 

 only leOff); but there is another factor at work counteract- 

 ing this efl'ect, viz., the familiarity of certain multiplications, 

 making the products by one short, and those by four and 

 five long. 



(6) The next type of "question and answer" will be one 

 in which, in addition to the act of memory, a comparison, 

 or a, judgment, is involved. The result of the comparison, 

 though not always the same for all individuals (and in this 

 sense the question is not limited to a single answer), will 

 probably always be the same ia the same individual. The only 

 experiment of Cattell's that seems properly to belong here is 

 that in which the subject decided which was the greater of 

 two eminent men (558(7). Munsterberg finds the average 

 time for answering a miscellaneous group of such compari- 

 sons 9470', or 99ff longer than the process without compari- 

 son: comparisons rapidly made (eOOff-SOOff) being, "Which 

 has the more agreeable odor, — cloves or violets?"' "Who is 

 greater, — Virgil or Ovid?" "What is more beautiful, — woods 

 or mountain?" and difficult questions (1200(?-1500(5') being, 

 "Which is healthier, — swimming or dancing?" "Which do 

 you like better, — Goethe's drama or his lyric?" "Which is 

 more difficult, — physics or chemistry?" The comparison 

 may be among more than two objects. Thus, in asking which 

 is the finest of Goethe's dramas, the process of formulating 

 the reply may include the calling to mind what the various 

 dramas are, and a choice among them; not, of course, a con- 

 siderate judgment, but the selection, under the necessity of 

 an immediate answer, of one deciding motive. On the other 

 hand, among the several possibilities, a certain one may, by 

 habitual association or for other reasons, have become so 

 prominent that virtually no comparison ensues; and the 

 relatively slight excess in time of this type of association 

 above the former ones (1049<7) suggests that this was often 

 the case. To decide which is the pleasantest odor (rose), or 

 which the most important German river (Rhine), required 

 only between 600(T and TOOff; to decide which was the most 

 difficult Greek author (Pindar), or your favorite French 

 writer (Corneille), from 1400(7 to 1600(7. 



Munsterberg has ingeniously modified this form of experi- 

 ment to show the influence of the foreknowledge or pre- 

 paredness of the subject. He precedes the asking of the 

 question by a dozen or so words of the category within 

 which the comparison is to be made. Thus, "Apples, pears, 

 cherries, peaches, plums, grapes, strawberries, dates, figs, 

 raisins: which do you like better, — grapes or cherries?" 

 Although the comparison cannot be begun until the last 

 word is heard, still the subject has in a way anticipated the 

 general nature of the question, as well as the scope of the 

 comparison, and has reduced the time considerably (676c>, as 

 compared with 9470'), — certainly a striking result. 



(2) Questions with More than a Single Answer. In the 

 class of re-actions to which we now pass, the question admits 

 of several answers. The answer at 6ne time may and need 



not be the same as at another time; and the determining 

 factors in the particular character of the answer are the 

 peculiar mental habits and tendencies of the individual. 

 The question thus changes from a specific to a general one, 

 the answer being any member of a more or less extended 

 class answering to such and such a description. In some 

 the choice may be somewhat limited. This is true of Cat- 

 tell's experiments in which, given a country, we are to name 

 a city in it (3460') ; given a season, to name a month in it 

 (435(7); given a language, to name an author writing in 

 that language (519ff) ; or, given an author, to name any work 

 of his (763(7). In all these cases we are apt to have in mind 

 only a very few prominent instances under each head among 

 which individual preference is exercised. In the following 

 series the classes are more general, and accordingly the 

 scope for individual preference much larger: given a general 

 term to name a particular instance under that term (587(7); 

 given a picture to name some detail of it (447(7) ; given the 

 word instead of the picture, to make a similar association 

 (439(7) ; given the picture or the name to mention some prop- 

 erty of it (372(7 and 337(7) ; given a quality to name an ob- 

 ject to which it can be applied (351(7) ; given an intransitive 

 verb to find an appropriate subject (527(7), or a transitive 

 verb to find an appropriate object (379(7). Miinsterberg has 

 a series including a miscellaneous collection of such re-ac- 

 tions, and finds a time of 1036(7. Trautscholdt has investi- 

 gated a similar series in which a specific instance of a gen- 

 eral term had to be given, and finds a time of 1020(7 (average 

 of three subjects), 155(7 of which must be deducted to get the 

 pure association time. 



Here, again, we may stop to consider a few generalizations 

 which these results seem to sustain. The processes involved 

 vary very considerably in the diiTerent experiments. Miin- 

 sterberg cites as quick responses (450(7-600(7) the instancing 

 of "a German wine (Riidesheimer)," "of a number between 

 ten and four (six)," " of a Greek poet (Homer);" as slow 

 ones (1200(7-1500(7), "a beast of the desert (lion)," "a French 

 author (Voltaire)." Trautscholdt names " mast " as "a part 

 of a ship" in 391(7, but requires 1899(7 to name "art" as 

 " an aesthetic activity of man." These difi'erences should ap- 

 pear in the average variations ; that is, the average divergence 

 of the re-action times from their mean. When the process 

 is simple and constant, the average variation is small; when 

 the processes are complicated and variable, the average 

 variation is large. While in simple re-actions it is often less 

 than 10 per cent of the re-action time, it is not infrequently as 

 high as 30 per cent in the re-actions just considered. If may 

 have been noticed that in certain cases the process in (2) was 

 the reverse of that in (1). The one was a step from the 

 whole to the part, the general to the special; while the other 

 was from the part to the whole, the special to the general. 

 In Cattell's case the former is the longer (433(7 and 374(7). 

 In Trautscholdt's results the conclusion comes out more 

 clearly, the pure association time of an association of part to 

 whole is 608(7; of whole to part, 901O'; of special to general, 

 754(7; of general to special, 947(7. It is thus easier to refer 

 an individual object or quality to its class than to give an 

 instance of a general concept. A similar result (namely, 

 that the bond of association between two concepts is not 

 equally strong in both directions) is derived from observing 

 that it takes longer to recall that May precedes June than 



