There is a predominant trend in the social sciences to study human behavior in terms of systems of analytically defined variables. In so doing the social sciences emulate the method of inquiry of the physical sciences. In this paper I do not intend to question the possibility of studying human behavior in this way or to question the validity of the knowledge that is thus obtained. This is not to say, however, that I see no limits or restrictions as to what can thus be analysed and explained. What I am here concerned with is to show how this prevailing method of analysing and systematising social phenomena affects our understanding of society and of ourselves as human agents, and to indicate how these effects can be remedied.