I907- 1 



TITCHENER AND PYLE— JUDGMENT OF DISTANCE. 107 



shows at least two instances of careless handling. If the two plates 

 on pp. 449, 451 are compared with the description on pp. 452 f., 

 it will be seen that the lower diagrams of each plate have been 

 interchanged: Figs. 7 and 8 should be Figs. 11 and 12, and con- 

 versely. ^ And again : if the table on p. 448 is scrutinised, two 

 mistakes will be noticed. The difference between + 2.65 and 

 -|- 1.75 is given as — .10; the difference between + i-o6 and + 0.78 

 is given as — ."ji. It is easy to read 2.65, 2.75, and 1.06, 1.78: but 

 then the differences, instead of being minus, are plus, — that is to 

 say, tell against Dunlap's conclusion. At all events, something is 

 wrong, either with the principal figures or with their dift'erences. 

 (3) Dtmlap's observers showed a progressive change of judgment 

 throughout the experiment. Whatever may be the explanation of 

 this change,^' he tells us that three of his observers overestimated 

 the right segment early in the experiment, and later underestimated 

 it ; while the fourth observer, '' with a single exception, overesti- 

 mated the right segment throughout the experiment, rather more 



^Dunlap says that 5, 8, 9 are indifferent; 6, 10, 11 faintly in accord, and 

 7, 12 in striking agreement with his hypothesis. If we read 5, 12, 9; 6, 10, 

 7; and II, 8, we bring the plates into accordance w'ith the text. These 

 changes, however, mean the replacing of the present 7, 8 by 11, 12, and con- 

 versely. 



2 The explanation must probably be sought in a general tendency of 

 judgment, complicated by preferential direction of attention: it will be re- 

 membered that the experiments were doubly one-sided, in that (i) the stand- 

 ard line was always shown on the left, and (2) the variable was always in- 

 creased from " shorter " to " equal," never reduced from "' longer." The 

 shift of judgment might have been checked by suitable instruction from the 

 experimenter (Titchener, "Experimental Psychology," II., 1905. ii-> 3^5 f-)- 

 How far practice was involved it seems impossible to say. — It is, of course, 

 theoretically possible that the minute values obtained by Dunlap for the 

 illusion-effect are due to a very high degree of practice with the Miiller- 

 Lyer figure. We have not seriously considered this possibility, (i) be- 

 cause Dunlap says nothing of preliminary practice; (2) because he gives 

 no intimation that the illusion-values of his Table I. were different in 

 kind from those of Table IV (see p. 450) : (3) because he defi- 

 nitely ascribes the small values to the " circumstances " of the experiment, 

 i. e., to the subliminal character of the shadows (p. 450) : and (4) because, 

 in view of Judd's and of our own results (our observer G, in particular, has 

 had extended practice with this illusion-figure), we do not consider that 

 Dunlap's experiments were numerous enough to reduce the illusion-average 

 to 0.5 mm. 



