Spatial Patterns and Residential Preference Systems of Mental Maps in Japan

Abstract

In his first research in America, Gould identified that there are mental maps shared in common among many individuals with respect to a specific perception point. This he termed as a specific mental map (S). In his later research in Britain, he found a general mental map (G) which was shared in common among all British school leavers. He explained the relationship between these two mental maps as follows: S=G+L (where L is local effects or local dome). This means that there are differences in every specific mental map. But he didn't consider differences of mental maps which derived from different residential preference systems. In mental maps, the residential preference systems or value systems are very important, because mental maps mainly depend on them. But there are many residential preference systems in a human group, so that the mental maps shared in common are not single but plural.In this papar, therefore, the existence of plural general mental maps is conceptualized and then the maps are extrected, their spatial patterns and their preference systems are analyzed, and their relationships are discussed.(Concepts)Prior to analysis, some concepts would be defined as follows.(a) Generality of mental maps and local effects of mental maps are features of spatial patterns. The former is a spatial pattern shared in common in a country (or study area) among the respondents. The latter is features of spatial patterns viewed from every specific perception point, which operate to deform the general mental map.(b) Dimensionality of mental maps is the variety of residential preference systems. Operationally, the dominant residential preference systems correspond to the dimensions of principle component analysis. The residential preference systems are interpreted by component scores on this scaling.(c) Homogeneity of mental maps is equivalent to the extent to which a particular residential preference system exists within a group. Operationally, it is measured by coefficient of determination.(Data and method)From six high schools in Aomori, Chiba, Fukui, Iwakura, Yao and Niihama, the residential preference ranking data are obtained. To these data, principal component analysis is used twice. In the first step, spatial patterns at every perception point are represented through principal component scores and their features are described. At the second step, principal component analysis is reapplied to the six component scores obtained (the first dimension and the second dimension, seperately).(Results)Fig. 2-(1-6) shows the spatial patterns of the first dimension at the six perception points respectively. Fig. 3-(1-6) shows that of the second dimension. Fig. 4 shows the spatial pattern of the general mental maps of Japan drawn by using the first dimension components. Fig. 5 shows the map drawn by using the second dimension components. The features of every specific mental map are summarized in the general mental maps.The features of spatial pattern in the first dimension are as follows; (1) prefectures known for sightseeing (Kyoto, Nara, Hokkaido, Shizuoka, Nagano, etc.) are preferect, (2) in general, the warm and urbanized prefectures have high score, and (3) Tokyo, capital of Japan, is not prefered. In specific mental maps, however, there are local effects, for example, the prefectures near the perception points have high scores as compared with other prefectures. The features of spatial pattern in the second dimension are as follows; (1) it is simpler than the first dimension with respect to spatial pattern, (2) rural prefectures have high scores, while urbanized prefectures have low scores. As seen from table 4, it can be said that a rather high correlation exists between the various perception points in both the first and the second dimension correlation matrix of component scores.