Loss: Sadness and Depression. Vol.  III of Attachment and Loss. John Bowlby. New York:  Basic Books, 1980.  xv + 472 pp.

Psychoanalytic theory, says Bowlby, provides a poorer model than cognitive theory and information processing theory with which to account for sadness, mourning, and loss in all their normal and disordered variants. Cognitive theory holds that our representational models of the world determine what information is selected and how it will be interpreted. The reader begins hoping that Bowlby's goal of broadening and deepening our understanding of the mourning process will be achieved, or at least that in the end we will be in a good position to evaluate whether or not a pure cognitive theory constitutes a better interpretive schema for the data than the models provided for us by psychoanalysis.

This last volume of the trilogy, Attachment and Loss, explores the implications for the psychology and psychopathology of personality development of the way in which young children respond to loss, particularly the death of a mother figure. The plan of the book is to study in vast detail a large variety of loss reactions to death, first in adults and then in children. All of the meticulously presented data come from previously conducted and published empirical studies by other investigators. Bowlby then offers for our scrutiny his own more or less systematic content analysis of this data, which takes the form, mainly, of the manifest content of interviews with the bereaved. Bowlby's main purpose is to demonstrate that traditional psychoanalytic theory is not sufficient to account for certain trends in the data. He calls, rather, for a theory drawing heavily on ethology, and, especially in this volume, cognitive psychology and the human information processing approach proposed by Peterfreund. It is, therefore, critically important for the reader who wishes to evaluate the extent to which Bowlby has succeeded in his aim of replacing psychoanalytic theory with a new theory to remember that the data presented do not derive from the psychoanalytic situation itself, but are, rather, of