This essay examines the theoretical and empirical bases of four analogical models, past and current, in psychosomatic medicine: conversion (Ferenczi); personality profile (Dunbar); typical conflict situation (Alexander); protective adaptive response (Wolff). The first three have inspired a good deal of research and thought, but are in many ways incomplete. The fourth is essentially untried. The authors also discuss the concept of physiological regression and its correlates (Michaels, Margolin, Szasz, Hendrick, and Grinker). They conclude: 'It does not seem that a new term like physiological regression affords the investigator a new vantage point from which to observe data … nor does it seem that this term contributes anything concrete to the understanding of psychophysiological phenomena. Neither does it seem to provide any useful insight to the therapist.'