One of the most important tasks now faced by researchers in the field of developmental psychology is a way to deal with the concept of relationship. This symposium was planned to exchange views on such difficult problem the two speakers themselves had been involved with for the past few years. Miyake proposed that in developmental study, longitudinal and contextual approach to the interaction and relationship between an individual and his/her enviroment was to be consideres indispensable. He added that a prerequisite to such a kind of approach was to have adequate theory and method. Conerning methodology, he pointed out that thinking in psychology had been strongly conditioned, through both experimental paradigms and statistical tools, in order to deal with single behavioral events, and that any tool appropriate for dealing with multivariate problem was not available. He then emphasized the necessity of assessing interpersonal relation in a variety of contexts pointing out that an interpersonal relation in only one context would result in some misleading conclusions. According to Miyake, we have simplified our categories and techniques of observing and recording interaction and relationship in our search for objectivity, and then we have come to an adoption of an oversimplified view of them. Finally he proposed we first observe and film naturalistie situations, followed by an exact labassessment; and then a return to naturalistic situations in order to verify the validity of models derived from the experimental analyses. Nakajima pointed out the Problems of the relation between the development of linguistic function and several other functions. Pointing ont the barrenss in single-factor theories, he criticized research strategy dealing with single function for the reason that the mechanism of development will never be made clear by such a strategy. What he wanted to emphasize was the necessity of observing and recording child behavior in naturalistic situations. He then presented a model of the functional relationship in language development, explaining the relationship among linguistic, affective and action system. According to Nakajima the so-called experimantal analysis was ineffective as far as the study of relationship was concerned. Among the participants on the floor who made comment on the two speakers' proposal were Kojima(Nagoya University) and Takahashi(Kunitachi College of Music). Kojima took up the problem of conducting controlled laboratory assessment:he regarded it risky to attempt an experimantal study without having a clearly stated hypothesis. Takahashi's comment had to do with the naturalistic observations emphsized by Nakajima. According to her it would be impossible to generalize the finding from such observations unless a theory, though incomplete, was brought up in advance. Finally, Oka, the chairperson, made the following conclusion: To make a developmental research fruitful and of practical use, more attention to the complementary and reciprocal relations between experimantal and naturalistic approaches shuold be paid.