Abstract
The aim of our experiment was to determine whether context effects in the psychophysical judgments are produced by perceptual factors or by semantic ones. For this purpose we compared the functions of two different response languages. One, the response languages representing semantic factors are generally regarded as possessing the following three characteristics: a) they are related to a specific situation, b) they are novel and arbitrary linguistic conventions used exclusively within a laboratory, and c) they have a restricted number of response categories. Consequently, the category scales are to be considered as typical semantic languages. The other, the languages representing perceptual factors lack any of the abovementioned three chracteristics. The physical languages such as “cm” “gm” satisfy these requirements.
The task of the subject was to judge the length of white lines projected on a screen. A common “tracer” stimulus recurred in all phases of the experiment: in one phase the tracer stimulus interspersed among longer lines, and in another, among shorter lines.
One group of Ss was told to make judgment by the “cm” scale, the other by an E-defined category scale. The effects of the shifts in the context on judgments of the common “tracer” Stimulus 7 (13.7cm) were examined.
For the L (low) phase, the stimuli ranged from Stimulus 1 (4.0cm) to Stimulus 7. For the H (high) phase, the ranges was from Stimulus 7 to Stimulus 13 (23.6cm). In Exp. I, L-H-L and H-L-H groups were run for each combination of the “cm” scale and the 9-category scale. Transition phases (Phase 2 and Phase 4) interposed between the L and H phases and between the H and L phases. In Exp. II, only the “tracer” stimulus of the phases preceding a context shift was judged by the “cm” scale for the purpose of investigating the strengthened perceptual factors, In Exp. III, in order to investigate the strengthened semantic factors, the transition phase was omitted, and the 7-category scale was used instead of 9-category scale.
Results. Exp. I. The overall pattern of the context shift effect was quite similar between both kinds of the response languages. The degree of context effects decreased with frequent repetition of the context shifts, Refer to Fig. 1, Fig, 2 and Table 1. From a standpoint that these converging patterns were representative of semantic factors, the “cm” scale can be regarded as having assumed the function of a category scale. Exp, II. Refer to Fig. 3. Excepting the initial trial, the context effect resembled the phase in Exp. I. Exp. III. Refer to Fig. 4. The context effects of the Phase 1 and the Phase 2 are respectively similar to those of the Phase 3 and the Phase 5 in Exp. I. The abovementioned results show that stress should be laid on the semantic factors when we study context effects of visual length judgments. We cannot conclude directly, however, that the semantic factors produce context effects. On the contrary, our study indicates that it is not appropriate to regard naively the “cm” scale as representative of perceptual factors and category scales as semantic ones.
