128 PSYCHOLOGICAL EFFECTS OF ALCOHOL. 



as the first letter appeared, the visual word was complete. We exposed 

 the words backward instead of forward because the final letter of a word 

 is the least definite cue that could be given. Final letters are much less 

 suggestive than initial letters. Moreover, when the end letter is shown 

 first, the correct pronunciation can not begin unless the whole word 

 is revived in consciousness. If the initial letters were shown first, 

 articulation might start correctly before the word was fully known. 



In spite of its theoretical plausibility, its relatively small demands 

 on the active cooperation of the subject, and the probability that it 

 corresponds more or less closely to a common practice of extra-labora- 

 tory life, our method for obtaining quantitative expression of changes 

 in retentiveness is the one of our techniques which developed most 

 serious defects in use. We incline to believe, however, that the prin- 

 ciple of the technique is valid and useful, and that only the form in 

 which we used it is faulty. Its defects are not peculiar to our technique. 

 They are due to the difficulty of securing really homogeneous material 

 for memorizing. This constitutes a universal source of difficulty in all 

 quantitative investigation of the memory process. Vogt 1 remarks it 

 in his poetic material. Differences in the associability of material 

 exists even in the M tiller-Schumann 2 development of Ebbinghaus non- 

 sense syllables. Only the elaborate scheme of Miiller and Pilzecker 3 

 for presenting the material in every possible order completely obviates 

 this source of error. In the effort to equalize the associability of the 

 memory material, Ebbinghaus, 4 and, following him, most of the Ger- 

 man investigators emphasize the value of nonsense syllables. In his 

 excellent review of the various memory techniques, Pohlmann 5 regards 

 nonsense syllables as the best experimental material. In English, 

 however, nonsense syllables are not so satisfactory as they are in 

 German. As Miss Gamble 6 has pointed out, special rules are necessary 

 for the construction of English nonsense syllables, while "no devices 

 seem to make the reading and spelling of English nonsense syllables a 

 simple matter for all American college students." Our vowel signs 

 are relatively few and superlatively ambiguous without arbitrary rules. 

 There are very few nonsense syllables of three letters that may not be 

 pronounced so as to resemble or to recall a significant word in some 

 language that the subject may know. Differences in pronunciation in 

 successive repetitions may completely change the series. There is no 

 way by which the spelling can be guaranteed to give the same series of 

 words to two different subjects except by the adoption of artificial rules. 



l Vogt, Norsk Magazin f. Laegevidenskaben, 1910, 8, p. 605. 



J Muller and Schumann, Zeitschr. f. Psychol., 1894, 6. 



'Muller and Pilzecker, Experimentelle Beitrage zur Lehre vom Gedtichtnis, Zeitschr. f. Psychol., 

 Erganzungsband, 1, 1900. 



4 Ebbinghaus, Ueber das Gedachtnis, Leipsic, 1885. 



5 Pohlmann, Experimentelle Beitriige zur Lehre vom Gedachtnis, Berlin, 1906. 



8 Gamble, Wellesley College Studies in Psychology, No. 1. Psychological Monograph No. 43, 

 1909; esp. pp. 18-23. 



