Methodologic Grounding of Clinical Interpretations
by PHILIP RUBOVITS-SEITZ Hillsadle, NJ; The Analytic Press, 1998. 480 pp.
I want to share with you someone else’s work. I reviewed a book called
Depth-Psychological Understanding: The Methodologic Grounding of Clinical Interpretations published by The Analytic Press. It was hard work, because it is particularly dense, because I needed a dictionary at hand, because it is written from a philosophical frame and easiest read by someone with a similar grounding.
The author was Philip Rubovits-Seitz, an analyst from Chicago, who combined that with an academic career specialising in psychiatric research. It is a scholarly book surveying and synthesising the leading thinkers of the century. The references occupy 100 pages, many by Horst Kaechele, and one by Don Grant dealing with one of Grunbaum’s arguments.
Rubovits-Seitz casts a challenging eye over a central part of analytic method. He inspects a cornerstone of psychoanalysis. He turns analytic investigative skill onto analysis, finding the flaws and contradictions and trying to explore them.
My contribution is to have read and re-read the book. To have felt challenged by it and to find it stuck in my mind. I found the book opened new vistas for me, gave another basis for understanding psychoanalysis and the debates of our time about how it can be justified to a sceptical world. I have summarised, paraphrased and rejigged, and tried to make it digestible, as ultimately it raises important issues for analysts to think about. To help summarise dense material I have constructed a number of tables. I have also left out sections on Grunbaum’s Critiques and the particular experimental design used by the author.
Let me give a CONTEXT in which to place this book
In International Psychoanalysis (1996) Robert Wallerstein talked of the conceptual (and methodological) difficulties of trying to do properly scientific research in full harmony with the vital ‘spirit’ of psychoanalysis, ie. research that is simultaneously faithful both to the highly subjective and complex data of the psychoanalytic consulting room, and to the so-called objective canons of the empirical scientific inquiry.
He goes on to say how science is not defined by its content, but by its method of inquiry and the challenge of psychoanalytic research is to accomplish its necessary work by methods devised in such a way as not to do violence to the nature, or ‘spirit’, of the enterprise being studied.
In the same Issue, and at the other end of the spectrum is Andre Green who points out that the great contributors to psychoanalytic theory did so from their single minds and the experience of working with their patients. That there is no single major discovery for psychoanalysis, which has emerged from research.
He adds that we do not have the method for research, which is coherent, not with the content of psychoanalysis, but with the type of thinking, which is its true object. The more the object which science dealt with is in the inanimate world, the more precise the scientific knowledge. When we get closer to living substance, the questions become more obscure, and the enigmas more numerous.
These days there is a conference on psychoanalytic research; there is an empirical research section in the IJPA, which tries to address the beneficial effects of psychoanalysis employing methods using questionnaires, DSM diagnoses, symptom and personality measures.
This summary of Rubovits-Seitz’s work is very much on the side of the Wallerstein/American and Kaechele/German approach to psychoanalysis and research. That something exists therefore it must be researchable. Many of you will have heard Kaechele’s approach - taped sessions, number of dreams counted, verbal activity compared between analyst and patient over time, and so on.
However the focus in this book is not on some test or questionnaire but on interpretations and whether and how they can be used in research. His approach is at the microstructures of conflict-defence and are retrospective studies. But it is the background thought to his research method that I want to tell you about.
What I’m going to tell you about is related to Kaechele’s topic, but he was an expert and knew what he was talking about. He was a research insider - I’m an outsider trying to grapple with what it is all about and it’s relevance to my everyday clinical work.
BACKGROUND TO THE BOOK
Glover identified the "Achilles heel" of psychoanalysis as the uncertainty of inferring latent meanings and determinants in clinical data. The uncertainties of clinical inference and interpretation did not begin to surface in psychoanalysis until the 1950s when Glover and French, working independently and employing different methods of investigation, reported that clinical interpretation may not be as easy or reliable as Freud claimed. French found how individual analysts react differently to the same clinical data. Glover expressed alarm at the variability of conclusions based upon interpretations.
Those reports led to a group of analysts in Chicago to undertake a systematic investigation of the consensus problem. Seitz coordinated a team but could never achieve satisfactory consensus on the blind interpretation of the same case material.
Of some comfort is that the consensus problem is not confined to us but applies equally to interpretive judgements in other clinical fields, and to the human and social sciences generally. That is, interpretive methods in all fields of human study suffer from limitations of reliability.
All branches of science, including the physical sciences, employ interpretive methods of some kind and to some extent. In the human and social sciences, interpretive methods are the commonest approach to investigating human beings. Interpretive process is first and foremost a form of inquiry - an attempt to gain depth psychological information and understanding. Interpretations moves from ‘how things are’ to ‘how things are known’.
This book is about the methodology of interpretations, ie., how latent contents are sought, construed, formulated and justified. This is different from the technique of interpretation which is about the criteria of communicating information to patients, ie., the whether, what, when, how to communicate information. So what we are trying to address today are the methods clinicians employ in the process of interpretive inquiry.
HISTORICAL BACKGROUND
Freud would have preferred to use purely observational rather than interpretive methods, for the scientific goal of positivism is the discovery of absolutely certain knowledge, obtained by rigorous observational methods applied to strictly empirical data. Despite his positivist ideals Freud found it necessary to develop and use nonpositivist interpretive methods.
Probably when Freud stopped using suggestion and relied completely on his patient’s free associations, he needed more than a strictly observational method.
Freud actually minimised the role of interpretation and talked of the "ease" of making them and the "certainty" of them. He rarely used the word interpretation, instead talking of constructions, information and outlines.
In the Wolf Man and Papers on Technique, Freud was uninformative regarding his method of interpretation. In the Interpretation of Dreams it was about the theory of dream formation, not interpretation. He blamed therapeutic failures on patient resistance, not incorrect interpretation.
He attempted to make interpretations appear objective and certain, and while championing observation did not always distinguish it from inference.
Freud was a positivist as it was understood during his early career. Positivism advocated the application of scientific method not only to physical sciences but to human studies also. It was employing the methods of mathematics, logic, observation and controlled experimentation. It was important to restrict science to the observable.
It was not until the 1920s that Logical Positivism appeared. This insisted that no hidden entities or causes should be postulated, a principle that was antithetical to Freud’s thinking and approach. Freud was with the prior version of positivism - that if we do not attach our observations to certain principles we would not be able to recognise them.
However Freud’s flexibility and farsightedness was crucial. He did not apply strict rules to interpretation, leaving a great deal to the clinician’s tact and skill; he acknowledged one may wrongly surmise and never be able to discover the whole truth; he accepted the need to employ circumstantial evidence; he accepted most scientific propositions have a variable degree of probability. But throughout, his writings continued to be tinged with hopes of certainty.
Freud struggled with a methodological conflict between his positivist ideals, on one hand, and his clinical need for flexible, nonpositivist interpretive methods, on the other. He attempted to make psychoanalysis both a clinical art and a positivist science.
The search for knowledge in the human and social sciences (including psychoanalysis) continues to employ two very dissimilar methodological approaches
1) a research-oriented, positivistic, quantitative, objective and cumulative approach, (in the direction of Wallerstein) and
2) a clinically-oriented and predominantly humanistic, subjective, qualitative, interpretive, pluralistic and non-cumulative approach. (a la Green)
Freud moved from a Positivist perspective with assertions about confirming constructions he made, to a Post-Positivist perspective where he could be tentative and conjecturing. The Post-Positivist approach is an attitude - that there is no one correct method to follow.
The clinical controversy regarding objectivity touches on the nature of observations in science. Analysts cannot stand outside the analytic (or therapeutic) relationship and make neutral observations because of
1) the presuppositions we bring to the clinical situation,
2) we participate and therefore change as a result of the experience,
3) we are susceptible to counter transference reactions.
Consequently, our interpretations are never able to be limited to one fixed reading and so the exchange between patient and analyst is always open and continuous.
There can however be some objectivity of scientific (including clinical) observations and interpretations, but it is variable.
There is Relativism (anything goes; one interpretation is as good as another; truth is only what is experienced as true) at one extreme, and the Positivist Perspective (only absolute certain knowledge is valid) at the other extreme. Between absolute relativism and absolute certainty is a middle ground, which builds on the best beliefs available, and leaves all aspects of those beliefs open to revision or rejection. It is this middle ground that a theory and methodology of interpretation tries to develop.
MAJOR FORMS OF METHODOLOGICAL LAG IN CONTEMPORARY MODELS OF INTERPRETIVE INQUIRY
1) Monistic and Method Oriented - rather than pluralistic and problem oriented (Monism - that only one approach exists and there is a particular method of proceeding) (Pluralism - recognises more than one ultimate principle)
These approaches are based on the Positivist model. They employ prescribed methods which can be replicated, and believe that problems can be solved by one correct method.
Example: Kohut’s insistence that empathy is the only useful method of understanding latent mental contents.
Responses: Introspections and empathic responses aren’t transparent and require interpretation.
The subject matter of depth psychology is too complex for a single method to construe.
A single method used intermittently is useful, but if applied constantly it won’t get at the plurality and interrelations of latent meanings and determinants in clinical data.
A Postpositivist approach would be pluralistic and problem oriented.
These interpretive approaches are monistic in emphasising a particular interpretive method, perspective, or variety of data. (eg., Searl’s exclusive focus on resistance analysis; Gray’s preferential focus on the immediacy of the patient’s defensive processes; Lang’s claim that all of a patient’s associations are a commentary on the therapist’s behaviour; Etchegoyen’s view that the only way to test an hypothesis is to communicate it; and many other examples.)
- The post-positivist position is that methods are appropriate only in relation to questions being asked and no one method is the correct method for investigating human beings.
- That science is not fixed or finished - it’s agenda is it’s own improvement.
- Most methods and models of interpretative inquiry are pluralistic. If we believe in only one avenue to truth then it is truth which suffers.
- Freud’s interpretive methodology was decidedly pluralistic, comparable with casting a wide net rather than dropping a single line into a large elusive school of fish.
2) Doctrinal rather than Data - Driven approaches.
Freud’s positivist convictions led him originally to apply theoretical understandings to patient associations and he believed this was proper interpretive methodology.
He moved on, so in 1912 he wrote - the most successful results occur when one allows himself to be surprised by any new developments in the data and can respond with an open mind free of preconceptions. By 1915 he wrote - theory should remain a stranger not invited into one’s house.
Freud shifted from a ‘top down’ (theory driven) to a more ‘bottom up’ (data driven) strategy. But he lapsed and didn’t recognise the indoctrinating nature of his "suggestions operating in an educative sense" nor "giving the patient the conscious anticipatory idea of what he may expect to find".
We need freedom in thought for good interpretations, not pre-existing knowledge.
A Doctrinal approach employs theories that precede and are external to the data. A different approach is an ad hoc, ex post facto methodology which remains focused on and within the data and allows interrelations among associations to dictate meanings retrospectively.
-we want the analyst to discern the red thread, not spin it and weave it in.
-we must avoid selectively filtering data to fit favourite theories.
-we must not ignore issues of evidence and justification.
-we try to avoid simplistic, limited, cliche-ridden meanings rather than multiple meanings and determinants that change over time.
Freud’s own development showed how feeling lost and insecure in the complexities of clinical data, people turn to theory.
An initial grasp of theory is a necessary starting point so we can know, to some extent, what to listen for and how to think about it. But theoretical preconceptions compromise free floating attention and close off possibilities of new discovery.
A Doctrinal approach employs theories that precede, and are external to the data, leading to meanings being imposed rather than discovered.
Example: Kleinian theory uses a particular model of infant’s early life to guide understanding of patient’s associations, and Klein and Kohut’s approaches both are doctrinally driven by respective theories of pathogenesis.
Responses: If theory is imposed on data you get a doctrinal interpretation.
Theory driven clinical interpretations often are not applicable to the individual patient and interfere with the discovery of unique personal meanings and determinants. They also tend to generate only single constructions.
Core concepts underlying the psychoanalytic interpretive system are generalities and not highly specific clinical theories.
Core concepts underlying the psychoanalytic interpretive system -
1) The Unconscious Mind
2) Continuity
3) Meaning
4) Determinism
5) Overdetermination
6) Instinctual Drives
7) Conflict
8) Defence
9) Repetition
10) Transference
11) Importance of Childhood Experiences
These are generalities and not highly specific clinical theories.
At this point we could also add that research which shows patients improve at the same rates whether the practitioner is Kleinian, Winnicotian, whatever.
3) Over confident and Dogmatic rather than Skeptical, Error-correcting Approaches.
Methodological lags are related to the positivist emphasis on certainty. This interferes with the necessary willingness to search for negative evidence and for alternative hypotheses.
Post-positivist science has a skeptical and error correcting approach. It does not claim to discover "truths" it seeks approximations to move forward but never fully achieve "the truth".
LANGUAGE BASED MODELS OF CLINICAL INTERPRETATION
Freud employed a language based model, postulating a series of linguistic transformations underlying meaning and mechanisms of paranoid symptoms. From the Schreber case he proposed that the four clinical types of paranoia result from specific grammatical transformations of the latent homosexual statement "I (a man) love him".
But it wasn’t sufficiently related to the main body of analytic theory to become an established method or doctrine. To Freud words were part of the preconscious not the unconscious.
To take two of these language based models to give you the flavour of the approaches (and exclude Edelson’s Linguistic model; Schafer’s ‘Action Language’; Labov and Fanshel’s Rule-Governed Discourse model; Clippinger’s Simulated Non Rule-Governed model)
I Lacan’s Structural-Linguistic Model.
Lacan drew on Saussere’s structural-linguistic concept, that language is a system of relations consisting of differences in both the sound images (signifiers), and in the mental concepts (signified) that the sounds signify. But he inverted that view.
Lacan believed that searching for the signified, the mental concept or unconscious thought, is futile. He did not see it as an object of clinical interpretation. He stressed the signifier - the sound, not the meaning, of a word. Bits of speech sounds (words) are idiosyncratic to the individual, and have been repressed and disguised by the unconscious operations. He insisted the signifying chain is the proper object of study for linguist and psychoanalyst.
Lacan’s unconscious is entirely signifiers (linguistic chains and networks) and resembles Freud’s preconscious. For him primary process condensation and displacement are equivalent to metaphor and metonymy.
Lacan believed he discovered something universal and claimed there was no mental structure except that of language. But he failed to recognise the limitations of language as a model of mind. He overlooked that the syntax of dreams is too compressed and discontinuous to be viewed as analogous to the syntax of natural language. Lacan dismissed unconscious thought as an object of interpretation and gave no place for the unconscious in antedating the acquisition of language.
From the perspective of interpretive methodology and the probity of clinical interpretations, Lacan’s approach has little to recommend it. It exhibits extensive "methodologic lag" in a number of positivist features: it seeks foundations in alleged linguistics universals, is monistic, method-oriented, absolutist and theory driven. (the seven deadly sins of philosophy)
The argument here is -
Psychoanalysis is a theory of language (Lacan’s position) VS.
The Unconscious can’t be equated with language (the mental activity of dreams and other unconscious processes is different from language).
II Narrative Models
Like discourse, narrative is also universal, transcultural, transhistorical, a largely language based human activity not governed by specific rules.
Narration involves reasoning; it informs action by attempting to construct true or likely stories about events represented in perception and memory.
It preceded science as an attempt to make the world intelligible. (Aboriginal dreaming is just that)
It is acquired early in life, and narrative competence functions as an assimilating structure for organising and comprehending discourse, reading and experience generally. Narratives try possible outcomes in fantasy or thought. Typically narratives are combinations of actual happenings and imaginings.
Narratives contain some historic truth, but like a fable, permits the expression of how one feels, how one would like to feel, to be. Its ambiguity is its strength. Narratives suggest themes and organise data into coherent stories. The therapist contributes to narrational process by "redescribing, reinterpreting, recontextualising and reducing" the patients life and experiences in terms of certain story lines and subjective experiences along psychoanalytic lines. The danger of the approach being theory driven and indoctrinating is evident. That is, if we can stay with the person and their issues rather than superimpose analytic theories, a meaningful exploration is possible.
The narrational approach deemphasises the criterion of accuracy (perhaps it can be regarded as constructed psychoanalytic fictions). Clinicians seek as much dynamic and historical truth as we can find, then extrapolate to other clinical findings using interpretations - but acknowledge gaps and uncertainties rather than glossing over by using narrative smoothing.
These language models help us see what they contribute to a flexible model of interpretive methodology.
NON CLINICAL METHODS OF INFERRING LATENT CONTENT
Looking at interpretive procedures in non-clinical fields we may learn more about the nature, problems and possibilities of improving the methodology of clinical interpretation.
All branches of science employ interpretation, but vary
-in the nature and purpose of the interpretive tasks,
-in the degree of similarity between the interpreter and what is interpreted,
-in the competence required in interpreters, and
-in the content of their interpretations.
However interpretive method in diverse disciplines have some features in common and the more we know about how other scientists work and what methods they employ, the more we can spell out our own methods and realise our own methodologic assets and liabilities.
NON CLINICAL METHODS OF INFERRING LATENT CONTENTS
A. Relation, and Pattern, Oriented Methodologies
1) Pattern Model Approach
When the analyst detects relations in the form of repetitions, he weaves them together into larger patterns. Analogies entail imaginative relations among the data which then suggest restructurings into unexpected patterns.
2) Gestalt Psychology
A gestalt is the total structure and the basic principle in this approach is that a whole cannot be understood by an analysis of parts; a whole is more than the sum of its parts.
3) Systems Theory
A system is a set of interdependent entities that function together at a higher level of organisation. Understanding depends on recognition of relations, esp. relations between relations, eg., internal object relations.
4) Hermeneutics
The art and science of interpreting meanings in texts of all kinds (including spoken communications) It focuses on interdependent relations of part and whole meanings, eg, we may pick up a predominate conflict or theme say about sexuality or loss or control and then scan the associations for possible relations between that postulated conflict and the actual data. But there must be a checking - revising - rechecking built into this process.
5) Structuralism
Originated in Saussere’s concept that language is a system of relations between sound images and mental concepts that comprise the linguist sign. More on this when I mention Lacan.
All five of the methodologies in the pattern seeking approaches share with clinical interpretation an emphasis on identifying relations rather than things, relations which serve as clues to unobservable entities.
Relational patterns take various forms; repetitive, analogical, sequential, thematic, part-whole, causal, oppositional, and others.
B. Language Related Disciplines
1) Linguistics
Syntax provided clues to UCS content. Devices like "foregrounding" giving greatest importance to material at the beginning of a communication. But clues are so subtle their detection may be subliminal rather than conscious.
2) Psycholinguistics
Process the language we hear so we fill in gaps and produce coherent meaning and infer from what said, from what not said and how it is said or not said.
This processing is like clinical data processing in the necessity to construct rather than merely extract meanings from what is observed. We tend to think of meanings as tangible entities in the mind of the patient, but the meanings we interpret are actually abstractions, hypotheses of the therapist’s, attempts to infer or to guess what might be occurring in the mind of the patient.
3) Communications Theory
Emphasises the listener thematising continually to surmount the limitations of working memory (capacity 6 or 7 items). Perceiving and assimilating involves recoding (putting into ones own words or images). Also includes Kinesics (non verbal communication and gestures), Rhetoric (communication that attempts to persuade) (a criticism of analysis).
4) Semiotics
Science of signs (smoke a sign of fire) PSA uses signs to indicate mental processes. This is the most relevant of this group to clinical interpretation.
5) Literary Theory and Criticism
Lit. scholars study the nature of texts and also the methods and models of interpreting texts. Lit. theory requires interpretations to correlate with specific data and to enrich and illuminate the particulars of those data and account for as many details in the narrative as possible. (the comparison with psychoanalysis is clear)
6) Archaeological Decipherment
Freud compared therapeutic process with arch. excavation and restoration, and, clinical interpretation with deciphering ancient scripts. Requires making inferences, weighing imponderables and using scientific imagination.
C. Cognitive Science
1) Representations in Psychodynamics & Cognitive Science
2) Relation of Schema Theory to the Problems of Interpretative Methodology
3) Investigations of Problem Solving: Relations to Interpretive Methodology
These share with psychodynamics the concept of mental representations and use schema theory which shares a lot with transference, unconscious fantasy, self and object representations and inner working models.
D. "Commonsense" (Intentional) Psychology
This is knowledge of everyday life based on experience rather than formal education. It has predictive power and the knowledge of everyday life is shared by all members of a culture. It focuses on explaining the particular, which is consonant with psychoanalysis (unlike scientific explanations based on laws to explain generally).
Intentionality (different from intention) is defined as consciousness of or about something, ie., the "object-directedness" of the mental. That requires representations of objects. This was from Brentano who was Freud’s teacher in philosophy. Freud:
- extended commonsense psychology and intentionality with his interest in the psychological states for symptoms, dreams, parapraxes and phantasies;
- extended from desire satisfying actions to the psychological states imaging wish fulfilment;
- extended the commonsense concept of mind by adding unconscious desires and beliefs.
These are slightly foreign fields for the clinician because its about the philosophic positions taken by various theories and methods and the world of ideas rather than a personal world we are accustomed to exploring.
After reading this section I felt like the man who discovered he’d been speaking prose all his life and didn’t know it. That so much of this is familiar, in daily use, but I never knew where it came from having just picked it up from teachers and colleagues over the years.
JUSTIFICATION OF INTERPRETATIONS
Critics say Freud and his disciples act as if they had no conception of the meaning of proof; that failure to develop and use justifying procedures is destroying the science of psychoanalysis. In a study of sixty frequently cited papers in the psychoanalytic literature, none offered direct evidence for the claims being made.
But it isn’t a problem limited to psychoanalysis. It resonates throughout most of modern philosophy and more generally, modern thought. This is the perennial problem of interpretation. How do we know whether we are interpreting from or into associations? Aren’t we dignifying the analyst’s thoughts and the analyst’s associations by the pretentious title of interpretations? If interpreting is just entertaining options then it may be of service in making further discoveries. The danger lies in claiming premature validity for interpretations.
One of the basic reasons that systematic methods of justifying interpretations, are essential scientifically, is because interpretations are the first-level inferences, the lowest level of theoretical statements of psychoanalysis and dynamic psychotherapy; and as such they are the only propositions that can be tested by direct empirical evidence, ie, evidence from the case being studied.
Evidence consists of empirical data and logical arguments that support a conclusion.
The best arguments for clinical inferences
1) argument by analogy - eg., you respond to me like I am your mother, ... where a resemblance, a parallel, is shown between two things.
2) by consilience (different data converging to the same conclusion) (program about global catastrophe around 500 AD, got writings from Chinese, English, and others, got ice core samples from Arctic and Antarctic, got tree wood samples, etc and pulled them all together to prove a massive explosion of Krakatoa caused climate change and thence social change)
3) by abduction (argue to the best explanation).
The best arguments for clinical inferences are these nonexperimental ones. The evidence need not be manifest, it may be symbolic. Evidence applies not only to number of data but also the quality eg, relevance.
The distinction between evidence and proof is a matter of degree. Proof implies convincing support of a proposition. The goal of clinical justification isn’t merely to confirm or disconfirm a particular hypothesis but to provide discriminations between alternative interpretations, ie., to draw objective conclusions about the relative probabilities of competing hypotheses.
PROBLEMS WITH CLINICAL EVIDENCE
We are trying to get good evidence in order to justify our interpretations, whether that be for a study for Prof. Kaechele or for reviewing the work we do with a patient.
I have extracted and compiled a list of problems with clinical evidence:
1) the patient cannot report completely.
2) the therapist cannot directly observe the patient’s inner experience.
So if the evidence can neither be reported nor observed to the degree that would meet scientific standards from other disciplines, there needed to be a quantum leap.
To surmount the disjunction between observable and non-observable, psychoanalysis combines three methodological concepts: the unconscious, determinism and meaning -to postulate latent meanings common to observable and non-observable evidence. But to understand and justify a specific latent content one must evaluate the ever unfolding context of interpreted events. Hence Freud insisting full interpretation of any clinical fragment must await completion of analysis.
3) we cannot demonstrate the latent meanings and determinants that unify observational and non-observational evidence in any simple , direct, empirical way. Again requiring interpretation rather than simple observation of empirical data.
4) the patients associations are not entirely free and self-generated, but are influenced to an extent by the therapist.
5) our reliance on relations and patterns in clinical data which can lead to "soft pattern matches" that may not derive from the patient’s inner mental states. Since meanings differ depending upon context, and because the number of possible contexts is unlimited, we often find ourselves confronted with a bewildering array of alternative interpretations.
6) The irreducible subjectivity of interpretive work introduces further problems of inference and evidence. The therapist’s assumptions, values, and perceptions are crucial data that may contribute significantly to interpretive conclusions. To really assess an interpretation we would need to know all about how the analyst arrived there.
7) Intraclinical justification of interpretations utilises the same pool of clinical data which gave rise to the therapist’s interpretative hypotheses in the first place - a potentially circular, self-confirming process. So investigating interpretations might merely rediscover and reconfirm its own basic methodologic concepts.
8) Methods of testing are necessarily paradigm specific, ie., they apply only to the data and interpretations generated by a specific conceptual system of psychoanalysis. So the interpretive justification derives from the principles, perspectives and varieties of data on which the conceptual system is based. (can’t test a Lacanian interpretation using Kleinian tests)
9) The therapeutic function of interpretations compromises their hypothesis stating function
Freud noted this with Little Hans that PSA is not an impartial scientific investigation, but a therapeutic measure. Its essence is not to prove anything but merely to alter something.
Taken with all the problems of clinical evidence, it is little wonder that clinical interpretations are error-prone, uncertain, and a major source of disagreement among clinicians.
Despite all the problems in justifying interpretations, investigators, methodologists and philosophers of science believe that clinical interpretations are potentially testable.
These problems of clinical evidence raise the following basic methodologic QUESTIONS about the justification of interpretations.
Must interpretations be justified?
Interpretive hypotheses are private and personal to the patient about whom they are constructed. Higher level theoretical propositions can be thought of as meta interpretations (interpretations of interpretations) which apply to many or all patients.
So interpretations stand between observations and higher level theories. If they are so close to observations must they really be justified?
There are arguments for and against.
If first we take the arguments for justifying interpretations:
1) interpretations might appear to be, but never are, purely descriptive and atheoretical.
2) we find what we are looking for and we must attend to that bias.
Then the arguments against the necessity and possibility of justifying interpretations run:
1) interpretations are to provide "perspective" on the patients unconscious and past,
2)they can never be designated as true or false,
3) they aren’t concerned with truth but altering the meaning of the patient’s experiences,
4) the truth of the proposition isn’t the focus, it is whether it stimulates new material and furthers the analytic/therapeutic process.
The author comes down on the need to justify because clinicians must choose between alternative hypotheses and probably some choices are better than others.
Can interpretations be justified?
Some philosophers conclude that clinical interpretations can be justified though not with the precision or completeness of the physical sciences, and the principal criterion of justification is change in the patient that can be predicted (within limits)
Whether clinical interpretations can be justified depends also upon when confirmation is undertaken, ie., during sessions or after completion of the treatment.
Justification during consists of the most plausible among alternative constructions with checking, revising, rechecking and modified hypotheses. The principal selection criterion is which of the alternative hypotheses accounts most consistently, coherently and comprehensively for the largest number of data.
This derives from the three principles of confirmation employed in human and social sciences:
- comparison of alternative hypotheses,
- tests of competing hypotheses against all data,
- use of disconfirming tests.
Justification after, based on the record of the entire, completed, therapeutic process can employ multiple procedures and methods.
Can interpretations be justified intraclinically?
Most analysts including Freud, have assumed that psychoanalytic propositions can only be tested within the analytic situation, because no other clinical or experimental setting adequately replicates the subtleties, complexities and specific dynamic features of the relationship and process.
Critics from Popper and Eysenck to Grunbaum scoff at the idea of objective testing within the therapeutic situation.
Besides the issue of suggestibility only lowest level theoretical statements, interpretations, can be justified by direct empirical (observation and experiment, not theory) evidence.
So interpretations can be justified intraclinically but higher level clinical theories are validated by other, including extraclinical, methods.
In testing interpretations, multiple validity checks reduce the need for a single, grand test. Multiple checks reduce errors and two proofs are better than one.
Freud did not spell out or apply a systematic methodology of justifying interpretations, his basic approach was empirical and pluralistic. He employed all three major theories of truth:
- correspondence
- coherence
- pragmatic.
Freud proposed multiple methods of empirical testing
1. Internal evidence (mainly coherence)
2. Patient’s responses to interpretation (mainly indirect)
3. Postdiction
4. External validation
5. Similar results in other cases
6. Cross validation and convergence of evidence.
INDIVIDUAL METHODS OF JUSTIFYING INTERPRETATIONS & THEIR RELATIVE PROBITY
1) Pluralistic Methods.
Any combinations of the following, using guidelines: that multiple controls reduce errors, and, two proofs are better than one.
The more justifying methods one uses, and the more systematically and integratively the various methods are applied and compared, the greater the likelihood of justifying the most plausible interpretation at a given time.
2) Cross validation and convergence of evidence.
Cross validation compares different samples of data. Convergence compares the results of different justifying methods. These are cross-checking methods. eg. 1. formulations during a mothers analysis, 2. observations of the mothers child during child therapy, 3. empirical infant research, - all sources of data and the different ways of collecting it converged to conclude the child’s psychopathology was due to a form of maternal neglect.
3) Demonstration of organised, interlocking microstructures underlying interpretations.
This follows Freud’s jigsaw model. These aren’t feasible during treatment but are to retrospective studies. Studies using this method focus on repetitive microstructures in the therapeutic process whether they be emotions, types of relationships, verbal language, etc
4) Indirect prediction and postdiction.
Prediction of individual human behaviour, lacking general laws, are very limited. Theoretical statements don’t carry the force of predictions in natural science. That is why we are an explanatory rather than predictive science, like Darwin not Newton.
Researchers cannot postdict exactly the content of earlier events but can make cogent indirect postdictions regarding the class of configurations that may be found.
5) Repetition of themes and patterns.
According to Hume’s Principle of the Uniformity of Nature "the observation of past regularities provides grounds for predicting the same regularities in the future". Allied to core concepts of continuity and the compulsion to repeat we assume unconscious conflicts and defenses will repeat throughout the therapeutic process.
Not all consider repetition a reliable method of justification, "repetition only confirms the hypothesis that the phenomenon tends to be repeated", ie., seeking repetitive patterns and defining them as valid because they recur is tautological.
6) Internal evidence: coherence.
Observations and inferred meanings fitting well together in an interpretive hypothesis. A popular concept for the alternative is the unsettling effect of ambiguity and incompleteness. It is like the pleasure of a "good hour" when so many are murky, confused and uncertain where it is not clear what has happened. One of the criticisms is that it is applied unsystematically to macroscopic aspects of data, rather than in a detailed micro way like with justifying interpretations.
7) Patients responses to interpretations.
That a directly relevant criterion for justification of an interpretation is also a psychical phenomenon, namely, the degree to which an interpretation changes a patient’s psychical reality, as indicated by responses to an interpretation. Freud went further and suggested the furthering and deepening of the therapeutic process.
Freud’s list of indirectly supportive responses to interpretations included expressions of surprise; associations containing something similar to the interpretation; a parapraxis which contradicts a denial; and either improvement or exacerbation of symptoms. Many other patient responses have been suggested by other analysts since along the lines of the patient justifying or not the interpretation by examining their personal responses to it.
Researchers however pose the question of can it be known if the response confirms or refutes an interpretation? Especially as the response itself requires interpretation.
8) Quantitative Methods.
The pluralistic perspective of post positivist science favours use of all methods, not only sensitive clinical judgement. In this grouping are computer assisted content analysis; statistics; psychometric procedures; etc.
9) External Methods.
Freud always included the possibility of external validation, like old diaries, hospital records, new information from relatives, etc.
10) Confirmation by observation.
An interpretation makes assertions about the patient, if it is supported by actual observations it may be thought to be partially justified. Problem is any observation may have a number of hypotheses to explain the same data. Freud tended to confuse a replicated observation with a verified theory of the phenomenon in question.
11) Justification by implication.
Patient might not remember the first day at school but it can be presumed.
12) Ruling out the improbable.
Inferences by ruling out what seems impossible or improbable
Various methods of confirmation differ in their probity, but certain types of clinical data tend to be more probative than others. The patient’s thoughts, feelings, fantasies, and enactments towards the analyst, ie., data pertaining to the transference, are generally considered the most decisive for purposes of both discovery and justification.
Dreams, emotions, urges, impulses, and wishes also rate highly as clinical data. Then conscious fantasies, imagery, slips of speech and metaphors not far behind in probity.
Conscious memories and details about everyday events tend to be less revealing and probative.
Clinical interpretations are the first and most basic level of inference about clinical data. They are the only propositions tested by direct empirical evidence and thus justification of interpretations is crucial to the science of analysis and therapy.
Interpretations must and can be justified though not as completely or accurately as theories in physical sciences. Like the interpretive process as a whole, the justification of interpretations is methodologically pluralistic. The most probative methods of justifying interpretations include cross-validation and convergence of evidence, demonstration of organised interlocking microstructures underlying interpretations, indirect prediction and postdiction (of classes of events) and repetition of themes and patterns.
Psychoanalysis gives us insight and an understanding of paradoxes. We need them:
To meet the current challenges to psychoanalysis and getting a grasp of what is said by our critics.
To learn from both the gems and the misconceptions, handed down from Freud and his successors, and rise to the challenge, of defining what analysis is and it’s place in the world today.
To hold in mind the paradox of analysts championing scientific research and analysts condemning it.
This is a version of a presentation to the Melbourne Institute for Psychoanalysis June Conference, 2001. It was in turn based on a review of a book for the Journal of the Royal Australian and New Zealand College of Psychiatrists published in August 2000. The conference fortuitously followed presentations and a public lecture by Prof. Horst Kaechele about research in psychoanalysis.
