1 I will be defending my view, that psychoanalytic theory attains ordinary scientific goals by ordinary scientific means, by appeals to analogies, maxims and piecemeal judgments of particular scientific achievements. It would be an enormously longer task to develop a general theory of scientific method out of such raw materials and to show that it is better than alternatives which would, in fact, deprive psychoanalytic theory of scientific justification. I develop this general account of scientific method in Fact and Method: Explanation, Confirmation and Reality in the Natural and the Social Sciences (Princeton: Princeton University Press 1987). This paper is, partly, recompense for that book's meager discussion of Professor Grünbaum’s work. What I say on p. 150 is much too brief and does not accurately convey his deep awareness of the problem of identifying Freud’s most important claims.

2 See Grünbaum, A. The Foundations of Psychoanalysis (Berkeley: University of California Press 1984), 161, 177–80Google Scholar.

3 See his book, An Elementary Textbook of Psychoanalysis (Garden City, NY: Doubleday 1973).

4 For some denials that psychoanalytic theory is a source of laws with predictive power, see ‘The Psychogenesis of a Case of Homosexuality in a Woman’ (1920), in Strachey, J. et al., Standard Edition of the Complete Psychological Works of Sigmund Freud (London: Hogarth Press 1953; 1974)Google Scholar [hereafter ‘S.E.’], v. 18, 167f.; ‘Analysis, Terminable and Interminable’ (1937), S.E., v. 23, 220f., 224-6; An Outline of Psychoanalysis (written 1938), ibid., 182, 187, 191. For some passages according to which psychoanalysis could not provide theory-neutral descriptions of the invariable antecedents of syndromes see ‘Character and Anal Eroticism’ (1908), S.E., v. 9, 175; Introductory Lectures on Psychoanalysis (1917), S.E., v. 15, 386-91; An Outline of Psychoanalysis, 226. Lanzer's treatment ended in 1908 and the case study was published in the following year as ‘Notes Upon a Case of Obsessional Neurosis’ S.E., v. 10. (My subsequent citations from this edition of the case study, which includes excerpts from Freud's personal notes on the case, will largely be confined to controversial or usually-neglected points, or direct quotations.)

5 For the hypothesis about love and hate see ‘Notes Upon a Case of Obsessional Neurosis,’ S.E., v. 10, 238-40. This hypothesis dominates the whole of the theoretical section (labelled ‘II - Theoretical’), none of which is concerned with the reconstruction of isolated precipating causes in infancy determining the content of obsessions.

6 Two of Darwin's most eloquent admissions that his theory yields no potentially predictive laws occur toward the respective conclusions of the chapters, ‘Struggle for Existence’ and ‘Laws of Variation’ in The Origin of Species, J.W. Burrow, ed. (New York: Penguin 1982), 129, 202. For detailed arguments that predictivity has not been achieved in modern evolutionary theory, see Lewontin, Richard C. The Genetic Basis of Evolutionary Change (New York: Columbia University Press 1974), especially 234–9Google Scholar, 266-71. I discuss these issues further in Fact and Method, Chapters 1, 3 and 4.

7 Silverman, Lloyd ‘Psychoanalytic Theory: Reports of my Death are Greatly Exaggerated’ in Stem, R. et al., Science and Psychotherapy (New York: Haven Publishing 1977), 262fGoogle Scholar.

8 See Ihde, Aaron The Development of Modern Chemistry (New York: Dover Books 1964), 98fGoogle Scholar;. ; Kuhn, Thomas The Structure of Scientific Revolutions (Chicago: University of Chicago Press 1970), 130–4Google Scholar.

9 See Kernberg, Otto Borderline Conditions and Pathological Narcissism (New York: Jason Aronson 1975)Google Scholar, especially Chapter 1.

10 See ‘Constructions in Analysis’ (1937), S.E., v. 23, 258.

11 A number of studies bearing out this clinical lore concerning the beginning, middle and end of analysis are summarized in 5. Fisher and Greenberg, R. The Scientific Credibility of Freud's Theory and Therapy (New York: Basic Books 1977), 330fGoogle Scholar., 335. These matters of time have an important bearing on assessments of the efficacy of psychoanalysis, since virtually all studies with formal control groups have investigated short-term treatment, generally four months or less.

12 ‘Original Record of the Case,’ S.E., v. 10, 287

13 See, for example, the increase with age of mental health problems reported in Srole, L. et al., Mental Health in the Metropolis (New York: Harper and Row 1978), 223–6Google Scholar; and the following long-term studies, in which low rates of ‘spontaneous remission’ were observed over periods lasting from five years to twenty years: C. Friess et al., ‘Psychoneurotics Five Years Later,’ American Journal of Mental Science (1942); Karush, A. et al., ‘The Response to Psychotherapy in Chronic Ulcerative Colitis’ I and II, Psychosomatic Medicine (1968 and 1969)CrossRefGoogle Scholar; Kringlen, E. ‘Obsessional Neurosis: A Long-Term Follow-Up,’ British Journal of Psychiatry (1965)CrossRefGoogle Scholar. Because nonpsychotic emotional problems tend to have an ebb-and-flow, with urgent symptoms burning out only to reignite years later, long-term studies are important. The failure to take this pattern into account produces overestimates of ‘spontaneous remission’ in otherwise well-designed studies, since most studies begin at a point at which distress was experienced.

14 This is the whole of the section entitled ‘The Cornerstones of Psychoanalytic Theory’ in ‘Psychoanalysis’ (1920), S.E., v. 18, 247:

‘The assumption that there are unconscious mental processes, the recognition of the theory of resistance and repression, the appreciation of the importance of sexuality and of the Oedipus complex - these constitute the principal subject-matter of psychoanalysis and the foundations of its theory. No one who cannot accept them all should count himself a psychoanalyst.’

15 See ‘Notes Upon a Case of Obsessional Neurosis,’ 206-8.

16 Ibid., 239

17 In the book that Professor Grünbaum cites, Freud and the Ratman (New Haven: Yale University Press 1986), Patrick Mahony accepts the finding that Lanzer's obsessive symptoms ended, and documents his improved capacity to assert himself successfully in love and in work. See pp. 17, 27, 216. Mahony's own interpretation of the case material is consistently - on occasion, extravagantly -psychoanalytic. The general drift of his criticism is, quite explicitly, that Freud did not extend psychoanalytic inquiry far enough, into additional, important areas of Lanzer's life: ‘Understandably, because of his lack of experience, his own complexes, and those of society in some instances, Freud could not yet see a whole series of issues that we grasp and contend with more readily today: narcissism, early parenting, body schema, sibling relationships, unresolved mourning, the transference in the here and now, and the resolution of transference neurosis’ (215). Like every psychoanalyst working today, Mahony is not prepared to believe Freud's grandest claim about the case, that all distortions rooted in Lanzer's personality were removed in a treatment that lasted barely nine months.

18 I discuss efforts to dispense with causal categories in laying down rules for probabilistic reasoning in Fact and Method, chapters 6 and 7. Professor Grünbaum's examples from Sober's recent work point to similar morals. My example about heart attacks and exercise is derived from Nancy Cartwright, ‘Causal Laws and Effective Strategies,’ Nous (1979).

19 I have adapted Professor Grünbaum's description of the way to test the hypothesis about obsessionals that he locates in the Lanzer case study to fit the somewhat different hypothesis that I find there.

20 See Smith, M. et al., The Benefits of Psychotherapy (Baltimore: Johns Hopkins University Press 1980), 88fGoogle Scholar. I use the figure for ‘psychodynamic therapy’ in Table 5-1, which is divided among eighteen modes of treatment. Another category, ‘dynamiceclectic’ seems to correspond to short-term treatments using psychoanalytic theory, especially those addressed to the symptoms typically measured in research on outcomes. Inclusion of these studies would raise the average effect-size to .74. The relevant definitions of kinds of treatment are given on pp. 70f.

21 See Sloane, R. Bruce et al., Psychotherapy versus Behavior Therapy (Cambridge, MA: Harvard University Press 1975), 87CrossRefGoogle ScholarPubMed. As Sloane and his associates make clear, the wait-list group were not a true control, since they had pulled themselves together, sought help, received the release and insight associated with an initial psychiatric interview, and obtained a promise of treatment after four months. So the efficacy of psychotherapy as against spontaneous remission was underestimated.

22 Duhrssen, A. and Jorswieck, E. ‘Ein empirisch-statische Untersuchung zur Leistungsfähigkeit psychoanalytischer Behandlung,’ Nervenarzt (1965)Google Scholar. Because of the wonderful abundance of data that the West Berlin health-delivery system routinely produces, Duhrssen and Jorswieck were able to match the treated and untreated neurotic groups according to recorded prognoses whose reliability could be verified independently. They were also able to determine that those not treated by psychotherapy received the same, high standard of medical care as those in treatment. For the latter reason, the study has a bearing on the question of ‘placebo effects’ in psychoanalysis. Physicians, quite as much as analysts, try to give sick people encouragement and hope.

23 The most influential study is Brill, N. et al. , ‘Controlled Study of Psychiatric Outpatient Treatment,’ Archives of General Psychiatry (1964)CrossRefGoogle Scholar. Its alleged placebo is a protocol in which patients receiving the familiar inert pill were seen, as needed, in brief but intense sessions of psychotherapy as often as once a week for as long as a year. The treatment by which psychotherapy was judged was administered, typically once a week for five months, by randomly assigned psychiatric residents, learning psychoanalytically oriented psychotherapy on the job from manuals and supervisors. Brill et al. rate those treated with psychotherapy as slightly more improved, but the results fall just short of conventional thresholds of statistical significance.

24 Sloane, Chapter 6. For example, assessors listening to tapes rated the behavior therapists as superior in ‘accurate empathy,’ and took their ‘depth of intrapersonal exploration’ to be only slightly shallower (148f.).

25 See M.L. Smith et al., 89.