持続音の継時比較における時間錯誤

抄録

The purpose was to ascertain the characteristics of the P-and E-functions in the comparison of the temporal length given as the duration of tone (substantial time), where the former function represented the time-error as a function of the length of pause between tones, and the latter indicated that as a function of the length of time under comparison.
The changes in the length either of pause (P) (Tables 1 and 2) or of duration (E) (Tables 3 and 4) revealed that both functions were similar in tendency, i.e., the time-error was negative and remarkable at the shorter Ps or Es, but the error gradually decreased its amount and approached zero as P or E became longer. Within a single series, however, the changes in both P and E (Table 5) yielded a disagreement with the preceding results in the amount of error. The disagreement was more remarkable and the individual differences were larger when each combination of P and E constructed a separate series (Table 6). These disagreements are possibly due to the difference in the manner of constructing the stimulus series. A continued experimentation on a single stimulus condition showed no specific transition of error (Table 8). On the other hand, the repetition of a series, which was constructed by various stimulus conditions, increased or decreased the amount of error in that series as a whole (Table 9). The effect of the distribution of variable stimuli was hardly observed within the range of the present experiment (Table 10). The linear representation of the relative proportion of E-P-E did not result in any specific tendency in P- or E-function. By introspective method the following impressions were reported: a rhythm in binary time (R), aggregation of Es into a single unit (A), segregation of Es into pieces (S), a phrasing composed of two bars in binary time (P), and a wave-like image of E or oppressive impression (W) (Table 11). And the reports changed from R, through A, A and P, to W according to the elongation of either P or E. Such a complementary relation between P and E was indicated by the modification of E-function in the order of the length of P, in another detailed experiment (Table 12). If the effects of series construction and repetition were eliminated, E-functions for respective Ps, and P-function for respective Es also, could be brought to coincidence with each other by transformation of coordinate system (Fig. 1).
These experimental results suggest that P- and E-functions are divided into three phases. Phase 1: at shorter Ps or Es the amount of error varies largely, though being always negative. Phase 2: at moderate Ps or Es the negative error decreases as P or E gets length, and both functions appear in a similar form. Phase 3: at longer Ps or Es, P-function is horizontal, while E-function descends to the negative side as E increases. The first phase, corresponds to the impression R, the second to A, S and P, and the third to W. Many kinds of factors would be complicated under the phenomena of time-error. Therefore, further analysis of conditions affecting the phenomena would be required rather than the attempt at an integrated explanation.