The results of this study are quite encouraging. At the very least, it demonstrates that a "graphic differential" consisting of "pictographic opposites" is a possible instrument: subjects can rate concepts on such scales with readiness and ease, and they do so, as the size of many loadings indicate, with reliability. But these data go much further as regards the development of such an instrument. This study succeeded in conclusively identifying several graphic scales that index the evaluative component of concepts. Intercorrelations of composite scores between the GD and SD evaluation (E) scales run for the most part in the 80's and 90's under less than ideal conditions for estimating equivalence. This is true for both indigenous and cross-cultural analyses.
The implications of this finding appear to me quite important and should not be underestimated. In the first place, the evaluative component of the semantic differential technique is by far the most important of the other affective components of words and is an excellent index of general attitude (9). Thus, we now have at our disposal a cross-cultural graphic instrument for measuring attitudes, viz. a culture-free instrument which surmounts the usual difficulties involved in measuring attitude with verbal scales.
In the second place, these findings corroborate the universality of the Gestalt laws of symmetry. Inspection of the graphic scales which correlate highly with the evaluation scales (Nos. 4, 10, 25, 31, 45, etc.; see Tables 4 and 7 and Fig. 1) indicates that, in every case, the pictographic alternatives involve a contrast between a symmetrical and a non-symmetrical figure. The symmetrical side is always good, pleasant, sweet, beautiful, happy, etc. while the non-symmetrical side is bad, unpleasant, sour, ugly, unhappy, etc. This is not to say that in art symmetry is always preferred to asymmetry, but only that, under conditions where a subject is confronted with a contrast such as this on a rating scale, he uses such a scale according to the laws of good Gestalt by apparently abstracting the common evaluative element between the concept that is rated and the symmetrical contrast of the pictographs. That untrained and naive subjects in five quite different cultures should exhibit this tendency with such high agreement is indeed a remarkable fact.
On the other hand, our success in identifying graphic scales that tap the Potency and Activity components Or affect is much less impressive. There may be two reasons for this. One is that these affective components are less stable cross-culturally, even when verbal scales are used. The other reason relates to the specific choice of pictographs included in this study. The "physiognomic quality" of these graphic scales is simple and unidimensional, and they are all black on white. We have evidence (3) that some dimensions of color are good correlates of Potency and Activity (saturation and brightness, respectively) and, before we give up the attempt, these additional stimulus dimensions should be tried. It may be, that visual form per se relates to evaluation only, and additional variations (three-dimensions, color) are needed to tab the other affects and to reduce denotative contamination. The question remains unsolved.
Nevertheless, the limited success obtained of indexing Potency and Activity in some cultures should not be overlooked. Thus, with selected indigenous scales, and with abstract concepts, the equivalence coefficients between the two instruments also runs in the 80's for AE, FF, and JP (for Potency) and for FF and GG (for Activity) (Table 8). These results, however, because they are not uniform, are to be taken with caution and need replication.
Previous research with the semantic differential technique has established the cross-cultural generality of three basic affective dimensions identified as Evaluation, Potency, and Activity. The purpose of this study is to develop a graphic equivalent of the pan-cultural short form semantic differential. This "graphic differential" is composed of bi-polar scales made up of pictographic opposites on which subjects rate a concept by placing a check mark on the seven-point scale. Such a non-verbal instrument would be culture free and would, under several conditions, have certain distinct advantages over the verbal semantic differential form. Data are reported for- a pre-test of the graphic instrument in five language/culture communities (American English, Delhi Hindi, Finnish, German, and Japanese). The study succeeds in identifying several graphic scales that provide reliable and effective indices of the Evaluation dimension of affect. These scales exhibit.' and conform to the symmetry law of Gestalt and could be used as a culture free attitude measure. With respect to the Potency; and Activity dimensions, the results are less satisfactory, although some of the graphic scales can be used for this purpose when chosen indigenously (i.e. independently for each culture). Evidence of visual denotative contamination is found and discussed.