Holmes' discussion of the clinical implications of attachment theory is critically discussed. It is argued that the use of interpretation presented by Holmes to explain Bowlby's dispute with the psychoanalytic establishment is inappropriate and that the differences between psychoanalytic ideas and those proposed by Bowlby and arising from later observational studies of early development are substantial. Finally it is suggested that neither Bowlby nor psychoanalysis pay sufficient attention to the child's organisation of meaning in the early interaction with caretakers.