equal periods, and grouped the pairs of associates according to the period in which they were learned; or, in other words, in respect to the degree to which they were memorized. The first group to be memorized represents the highest degree of learning, and the last group memorized represents the least degree of learning. In order to ascertain the relation between the disintegrative effect of a change of contextual conditions to the degree of learning, we computed the amount forgotten under the changed and unchanged contextual conditions for each of the three groups of associates. The comparative data are graphically represented in Fig. 2. With the unaltered context, the amounts forgotten were 20 per cent., 34 per cent. and 48 per cent. for the first, second and third groups respectively. With the altered context, the corresponding values were 24 per cent., 60 per cent. and 79 per cent. Under both conditions, the amount forgotten varies inversely with the degree of learning. For all three groups, or for all degrees of learning, more was forgotten when the context was altered, but the difference between the two contextual conditions increases as the degree of learning decreases. In other words, a change of contextual conditions exerted the most disintegrative effect upon the material which was the least learned. CONCLUSIONS Both learning and recall were influenced by the contextual conditions. The nature and amount of the effect varied with the nature of the conditions. A word-context logically related to the response word exerted a beneficial effect upon learning, and its removal or interchange during recall was highly detrimental. When the context was varied during learning, the effect was lessened. The introduction of such a context in recall when none was present in learning also exerted a facilitating effect. The substitution in recall of a new context of the same kind exerted about the same effect as the introduction of such a context when none was present during learning. A word-context logically related to both the stimulus and the response words facilitated learning to about the same degree as did one that was logically related only to the response word. Its removal in recall, however, exerted a less detrimental effect than did the latter. A word-context logically related to the stimulus word exerted a detrimental effect upon learning. Its removal in recall did not exhibit any appreciable influence. A word-context logically unrelated to both the stimulus and the response words was also detrimental to learning. Its removal in recall was apparently beneficial. The presence of a number-context during learning was slightly detrimental, and its removal in recall exerted a similar effect. The substitution of a new context of the same kind (picture post-cards) also exerted a detrimental effect upon the recall of the face and name associations. The degree of the effect varied inversely with the degree of learning. The explanation of these results is obvious in most cases. A logical relation between two words presupposes some sort of a priori associative connection between them, so that the presence of one naturally tends to arouse the other. The introduction in recall of a word-context that is logically related to the response word, although it was not present in learning, will facilitate the arousal of the proper associate because both the stimulus-word and the context-word are associated with it. Likewise, the presence of such a contextual word during learning will exert a facilitating effect upon the learning process, and naturally the retention of this word in recall will be distinctly favorable. Even the variation of this type of context during learning and recall will exert some beneficial effect. The interchange of such a context in recall will exert a detrimental effect, for in any given situation the stimulusword will tend to arouse one associate and the context-word another. When the context-word is unrelated to either the stimulus or response words, the subjects may ignore it; their attention may be distracted from the task in hand, or they may attempt to establish a double association. In the two latter cases, the context will presumably exert a detrimental effect. As a matter of fact, this context did interfere with learning to some extent, and its presence in recall may have been slightly detrimental. Practically the same results were obtained when the context-word was logically related only to the stimulus-word. Apparently, the presence or absence of an associative connection between the stimulus- and the context-words does not materially influence the subjects' reactions to the situation. From the preceding results, we should expect a context that is related to both the stimulus-word and the response word to exert an effect like that of a context that is related only to the response word, for a connection between the context and the stimulus-word does not seem to be effective. The results justify this assumption. This type of context is distinctly favorable to learning, and its presence in recall is also quite effective. In fact, the ability to recall seems to be more dependent upon this type of context than upon one that is related only to the response word. The number-context was detrimental to learning but beneficial in recall. We must, as it appears, assume that the attempt to utilize these clues in the initial presentations was distractive; while the subjects were able to employ them with some degree of success in the final presentation, so that their presence in recall was beneficial. Our results thus indicate that the recall of any material is favored by the presence of an environmental factor which has some associative connection with that material. In the absence of such an association, the environmental situation is likely to be unfavorable to recall. When the material is learned in a contextual situation with which it has no associative connection, such an association is likely to be established with a sufficient number of presentations. Such an association was established in the case of the face-name material, and the strength of the connection increased with the number of the presentations. In the case of the unrelated verbal context, three presentations were apparently insufficient to form an effective connection. REFERENCES I. CARR, H. Maze studies with the white rat. I. Normal animals, J. Animal Behav., 1917, 7, 259-75. 2. HUNTER, W. S. Some labyrinth habits of the domestic pigeon, J. Animal Behav., 1911, 1, 278–304. 3. KJERSTAD, C. L. The form of the learning curves for memory, Psychol. Monog., 1919, 26 (no. 116). 4. PORTER, J. P. Further study of the English sparrow and other birds, Amer. J. Psychol., 1906, 17, 248–71. 5. SMITH, S., & GUTHRIE, E. R. General psychology in terms of behavior, New York, 1921. 6. Watson, J. B. Kinaesthetic and organic sensations; their role in the reactions of the white rat to the maze, Psychol. Monog., 1907, 8 (no. 33). 7. WATSON, J. B. Behavior of noddy and sooty terns, Carnegie Instit. Wash. Publ., 1909, no. 103. 8. WONG, H., & BROWN, W. Effects of surroundings upon mental work as measured by Yerkes multiple choice method, J. Comp. Psychol., 1923, 3, 319–31. VARIABILITY OF JUDGMENTS ON MUSICAL INTERVALS BY HELEN MORAN AND CARROLL C. PRATT Harvard Psychological Laboratory A musical interval, or any separation of two impressions on the pitch-continuum, is generally defined or recorded in terms of the ratio of the vibration-frequencies of the two generating physical bodies. A fifth, e.g., is that interval whose components stand in the ratio 2/3; a minor sixth has the ratio 5/8, etc. But these definitions are wholly in terms of physical properties. Nothing has been said about the interval considered as a datum of psychological experience. There is not a great deal, in point of fact, which can be said as yet concerning the phenomenological aspects of auditory intervals. Stumpf singled out certain aspects of musical intervals for exhaustive study and detailed discussion. Among other descriptive matters he observed differences among them with respect to the degree to which they approximate a single auditory impression, and he found it possible to work out uniformities in the degrees of fusion among intervals. But Stumpf was not always clear as to just what he meant by fusion.1 And later investigations have shown that observers may pass comparative judgments on intervals under at least half a dozen totally different criteria of judgment, which would seem to indicate that in spite of whatever uniformities may be secured in the order of intervals when judged by degrees of fusion, there are other characteristics, such as smoothness, simplicity and affective tone, which must be considered as entering into, or at least as related in some way to, the 1 For ambiguities in Stumpf's definitions of fusion see Tonpsychologie, 1890, vol. 1, 127, 128; Konsonanz und Dissonanz, 1898, 35, 44; and cf. M. Bentley, Amer. J. Psychol., 1903, 14, 65–70. * W. Kemp, Methodisches und Experimentelles zur Lehre der Tonverschmelzung, Arch. f. d. ges. Psychol., 1913, 29, 139-181; C. C. Pratt, Some Qualitative Aspects of Bitonal Complexes, Amer. J. Psychol., 1921, 32, 490–515. |