Information science, or informatics, has almost from its beginnings been characterized by a seemingly inordinate self‐consciousness, exemplified by concern with its status vis‐à‐vis other disciplines, with its status as a science, and with the significance of its objects of investigation and the goals of that investigation. The bibliography by Port, and the survey by Wellisch, of definitions of information science, and the historical survey by Harmon, all give substantial evidence of this self‐consciousness. Some aspects of this attitude are of course due to the social and political problems facing any new discipline (or field of investigation aspiring to such status), such as indifference or hostility from the established academic community, the fight for a share of limited research and development funds, the inferiority complex associated with having no well‐defined methods of investigation in a social situation which requires them for acceptance, and so on. Other aspects of this self‐consciousness may, however, be more related to strictly internal, ‘scientific’ concerns; that is, to problems within the theoretical structure of information science which must be solved in order for substantial progress in solving its practical problems to be made. This review surveys contributions to one such problem: the question of a suitable concept of information for information science.