This study is an attempt to investigate the affective meaning of 16 colors and 14 symbolic words as judged by Japanese and American subjects and to seek synesthetic correspondence between the colors and the words. Four groups of subjects (Japanese male college students, Japanese female college students, American female highschool students and American famale college students) rated the colors, and two groups of subjects (Japanese and American female college students) judged the words on a 35scale semantic differential. Color cards and semantic scales used, and two of the four color-groups of subjects were the same a those of the factor-analytical study previously reported (Jap. psychol, Res., 1962, 4, 78-91). First, in connection with colors, a good agreement was found among the four subject groups in their judgment of meaning of each color represented by mean scale scores (See Fig. 1-4 in the text). Secondly, it was noted that scales correlating highly with the Munsell Hue were mainly Activity scales (e.g. hot-cold, near-far, womanly-manly, unstable-stable, dangerous-safe, excitablequiet). Scales correlating highly with Value were in the main Potency cales(e.g. heavy-light, deep-shallow, full-empty, hard-soft, tense-relaxed, strong-weak, dry-wet.) Evaluative scales (e.g. beautiful-ugly, good-bad, clear-muddy, fresh-stale, healthy-unhealthy, new-old) appeared to have no such particular correlation with any of the Munsell dimensions. However, colors ranging from Yellow to Blue generally tend to be rated as positive on Evaluative scales(i.e., toward good), whereas colors approaching to and departing from the Purple-Red region tend to be rated negatively (i.e., toward bad). Black and Gray were judged as bad whereas White was rated as good. In connection with the 14 symbolic words, it was discovered first that the factor structures of both Japanese and American Word-spaces were reasonably similar to each other but quite different from the factor structure of the Color-spaces. In the former, the Evaluative factor was most salient while in the latter the Activity was most salient, although three factors-Evaluation, Potency and Activity-were as a group most dominant in each of four concept-culture group. Secondly, on the basis of 15 descriptive scales, 5 for each factor, it was discovered that symbolic words display similar affective meaning to certain colors (Table 2). For example, ANGER and JEALOUSY were found to be closely associated with Red, while SIN and FEAR were related to Black and Purple. ETERNITY, VIRGINITY, and QUIETNESS, on the other hand, were associated with Blue or White, while HAPPINESS, HOME and LOVE with Green. Thirdly, the whole scale profile of each word was correlated with that of each color. Basically similar associative relations were obtained between words and colors, as discovered above. The results obtained in this part of study were in general agreement with results obtained by Obonai and Matsuoka (J. gen. Psychol., 1956, 55, 229-239) who used direct word-color associations to assess the synesthetic relations. Differences between their study and ours were submitted to test in the last part of the present study by asking a new group of subjects to choose a color associate from 16 color-stimuli for each of the concept words, where the colors and words used were the same as in earlier parts of this study. The obtained data confirmed the Obonai-Matsuoka study. In general, it was concluded that the semantic differential provides a useful method for assessing the association between different classes of concepts.
Two experiments were conducted to examine the mediative effects of the sensory impressions of material words on their paired-associate learning. In experiment I, ten undergraduate students in each of three groups learned a list of nine pairs of three-syllabled nouns, in which each stimulus and response item gave high common association of sensory impression (Group 1), low common association of sensory impression (Group 2) or no common association (Group 3). These stimulus pairs were presented by a memory drum at random order and were learned by anticipation method to a criterion of three successive errorless trials. The results are summarized as follows: The facilitation effect of mediative sensory impression was observed both in terms of trials to different criteria and in terms of the number of items correctly recalled in each trial, But the difference in performance between the high association group and the low association group was not significant in terms of both measures. Experiment II was designed to compare the mediative effect o sensory impression with the effect of mediate association induced experimentally. The former was estimated in the following way: In the second learning, three groups of subjects (ten undergraduate students in each) learned the same set of three lists (different from one another) as were used in experiment I, The latter was estimated in the following way: In the first learning preceding the test learning, the subjects in each of three groups learned two lists sucessively. The first learning lists consisted of sixteen pairs, in the one half of which each of two stimulus items (three syllabled nouns) paired with a common response item (a letter of the alphabet), and in the other half of which each of two stimulus items was paired with each different response item. The second learning lists consisted of eight pairs of nouns which were all stimulus items of the first learning lists, consequently, the one half of which were paired in such manner as had common response items in the first learning lists and the other half of which were not paired in the ahove-mentioned way. These stimulus pairs were presented by cards (5cm×1.4cm) in random order and were learned by the anticipation method to a criterion of five correct response of each pair in the first learning and cf three correct response of each pair in the second learning. The results are summarized, as follows: The mediative effect of sensory impression and the effect of mediate association induced experimentally were both observed in terms of trials to different criteria but their interaction effect was not observed significantly.