This work is about a peculiar historical anomaly – the neglect and eventual abandonment of the rich and theoretically fertile conception of the social embraced by early American social psychologists – that I stumbled upon almost by accident.

Rom Harré and Paul Secord originally stimulated my interest in the social dimensions of human psychology and behavior and the special problems they generate for a scientific and experimental social psychology. Since my graduate days in Oxford, much of my professional career has been devoted to the exploration of these issues, developed in a number of books and journal articles. My more recent interest in the history of psychology came about as a result of having to substitute for a teaching colleague overtaken by motherhood. Although I immediately fell in love with the subject, which I have taught for the past fifteen years, for a long time the overlap with my metatheoretical work in social psychology was minimal.

However, some years ago I was asked to review Margaret Gilbert's book On Social Facts (Princeton University Press, 1991). In consequence, I was forced to recognize that I had been cheerfully talking about the social dimensions of behavior, emotion, groups, identity, and the like for many years without reflecting critically on my own conception of the social. As I explored this issue, I was pleased to discover that something very close to my own conception had been embraced by early American social psychologists.