VII Improving clinical evidence

Chapter 14

Assessing strengths and limitations of clinical evidence in a psychoanalytic clinical material

Ricardo Bernardi and Luisa Pérez Suquilvide, with discussion by Charles Hanly

Introduction

This chapter aims to reflect on the criteria used to evaluate the degree of certainty or weight of the evidence offered by clinical material to support certain theoretical hypotheses or techniques related to a case. When analysts are faced with a clinical case, whether their own or of another person, they must answer a multitude of questions, both theoretical and technical. Therefore, they must decide which aspects of the material they can rely on to answer them. The material may provide support for more than one hypothesis, and we must then decide which is the most grounded, that is, evaluate the nature and degree of evidence offered by the material. By "evidence" we do not mean an absolute proof or an unquestionable truth. We refer to the clinical reasons that lead us to consider that an inference about a patient is better than some other inference (see Boesky, 2002). This evaluation involves examining not only the strengths of the evidence or findings that emerge from clinical observations but also the omissions or biases that limit its value.

Psychoanalysis, as a method of research, therapy, and theory, relies heavily on clinical observation. As Freud pointed out, speculative theoretical hypotheses "are not the foundation of science, upon which everything rests: that foundation is observation [Beobachtung] alone. They are not the bottom but the top of the whole structure, and they can be replaced and discarded without damaging it" (Freud, 1914, p. 77). This does not mean disregarding the theoretical conceptions, because they help us to bring order and clarity to the raw material of observations (Freud, 1933). However, theoretical conceptions very easily become paradigmatic models that, while helping to perceive and think about the clinical material based on certain premises, make it difficult or impossible to see it in the light of other approaches (Bernardi, 1989). For this reason, although current pluralism allowed the emergence of diverse theoretical and technical approaches, it did not manage to promote true scientific controversies (Bernardi, 2002).

Psychoanalysis is essentially a complex clinical discipline and encompasses both scientific and hermeneutic approaches (Bernardi, 2015). In the clinical field, it must address the questions raised by the formulation of a clinical case, namely, what the patient's problems are, what they are due to, what their treatment is, and the benefits obtained (Bernardi, 2016; Rodríguez Quiroga de Pereira et al., 2018). Therefore, they essentially refer to the therapeutic process and its results.

There is an agreement on the need to map evidence. However, emphasis on what the central aspects of evidence are vary, as shown in the controversy over "What does the presentation of case material tell us about what happened in an analysis and how does it do this?" (Blass, 2013; Boesky, 2013; Da Rocha Barros, 2013; Chabert, 2013). Boesky focuses on the interpretations and in explaining what the rationale is. Evidence understood as the information that the analyst can adduce to justify that his or her hypothesis gives a better account of the clinical situation than an alternative hypothesis that arises from the contextualization of the patient's associations in the development of central hypotheses. This brings certain aspects to the fore, such as the application by the pathologist of a particular dye determines the structures that can be made visible (Blass, 2013, p. 1131). Validating an interpretation based only on the patient's reactions immediately after its formulation, without considering the context, would be, for this author, to rely on a fallacious assumption. Given that some hypotheses explain the available data better than others do, one of the most serious problems we face is to resolve and refine the arguments for agreeing or disagreeing on how to make such comparisons. On the other hand, Da Rocha Barros emphasizes the emotional context, the transmission of the living experience of the analytical encounter, and the need for a poetics that can give an account through metaphorical expression. Nevertheless, he proposes that the repetition of specific patterns of mental functioning would be a useful initial form for mapping evidence. Chabert considers that what is important in a clinical material lies in expanding associative thinking, revitalizing a dialogue that would put us in touch with the dialogue that has been maintained in the analysis, a process that is inherently linked to the theoretical models from which we think. The 3-LM shows that some aspects of these three positions can be used as complementary perspectives in a sequential analysis of the clinical material.

Evidence-based practice

The search for evidence criteria is different in psychoanalytical clinical material and extra-clinical research. A systematic empirical investigation of the process and outcomes of psychoanalytic or psychotherapeutic treatments faces problems similar to those of evidence-based medicine

(EBM). This current orientation of health practices seeks the integration of evidence from the best available research with clinical expertise and patient values and preferences (Sackett et al., 1997). The APA Presidential Task Force on Evidence-Based Practice (2006) reaffirms these criteria in the field of psychotherapy and emphasizes that evidence can come from many different areas of both quantitative and qualitative methods. Practice-based evidence (PBE) complements EBM. PBE aims to gather scientific evidence from the clinician's routine practice settings (Barkham & Mellor-Clark, 2003).

A cornerstone of EBM is a hierarchical system of classifying evidence according to the level or weight of the available evidence. Therefore, a treatment's recommendation can range from high to moderate, low, or very low (or similar categories in other classifications) evidence. When the strength of a recommendation is weak, shared decision-making is especially important.

While the assessment of the levels of evidence was conducted with the utmost methodological rigour, many studies were found to be contradictory or non-replicable. This drew attention to the need to practise extreme care in exposing biases and omissions, intentional or unintentional, that could go unnoticed. In our field, Shedler (2015), Leichsenring et al. (2015), and other authors showed biases that could lead to favouring certain psychotherapies over others. This also led to highlighting the importance of the "Discussion" section in scientific work and the need to carefully outline not only the contribution of the undertaken research but also all the factors that may limit its scope (loannidis, 2007; Avidan, loannidis, & Mashour, 2019).

The degree of psychoanalytic clinical evidence

The path taken by the EBM suggests some topics of reflection for clinical psychoanalytic research. First, it is useful to make our criteria of evidence explicit when examining clinical material. Freud believed that not all clinical cases allowed the same advance of knowledge:

Something new can only be gained from analyses that present special difficulties, and to the overcoming of these, a great deal of time has to be devoted. Only in such cases do we succeed in descending into the deepest and most primitive strata of mental development and in gaining from there solutions for the problems of the later formations.

(1918, p. 10)

We would like to emphasize this idea: it is necessary to evaluate to what extent the analysis has managed to bring out the central problems of the patient.

Freud also believed that there was an inseparable bond between research and cure (juntkim is the word Freud uses to describe the relationship between research and healing). However, the relationship between increased knowledge and therapeutic benefit is complex. The effectiveness of a treatment can be effective as the result of factors other than those invoked by the psychoanalyst (e.g., "common factors," present in different psychotherapeutic approaches). There are "third variables," such as confounders or moderators, that influence both the assumed cause and the effects. In clinical research, we have to carefully examine the way we relate the changes in the analytic process to the changes in the internal world and in the life of the patients.

It is difficult to compare competing hypotheses in single cases. We cannot treat the same patient in two different ways. However, we can advance making our clinical reasoning and the assumed mechanisms of change traceable. In this way, we also expose the strengths and flaws of our hypotheses and "test" them through thought experiments and counterfactual thinking. Thought experiments are a conceptual tool that, although commonly used, has not been methodically employed enough. Thought experiments designate, in science and philosophy, the imaginary scenarios used to conceptually investigate inaccessible phenomena. This type of experiment does not provide factual results but allows us to understand the scope and consistency of our models better. They have always been used by philosophy and sciences such as physics, history, or economics, and the interest in them has increased recently in the philosophy of science (Brown & Fehige, 2019). However, in a rudimentary form they are present, for example, when we try to predict the effect of an interpretation. In that sense, they show the operational models that guide our actions. We will return to this point later, as we consider that their explicitness helps give replicability and traceability to our clinical reasoning.

Empirical research and EBM, as Tuckett says, forces clinicians to try to describe more rigorously and accurately what they see, and to draw out precisely both the viewpoint of observation and the inferential processes that follow (Tuckett, 1995,1998). It helps us to see more. "Evidence-based psychoanalysis, seen in this way, is an opportunity as well as an imperative" (Tuckett, 2001, p. 214).

We will focus on the criteria for evaluating clinical evidence. These problems are different, whether we consider a cross-sectional—i.e., one session or a few sessions—or a longitudinal study of a treatment and the changes in a long-term perspective. First, we will discuss evidence in both situations, and then we will refer to the problem of biases and omissions that limit the evidence provided by the material. Finally, we will propose, as an outline, a summary table that seeks to prioritize the main criteria set out in the chapter.

The clinical evidence in the session

When thinking about what evidence is available to evaluate the degree of certainty provided by clinical material, the interpretations and their effects present themselves as the observable elements that can bring us closer to the subject. In this sense, two aspects are closely related: the inferential process that leads the analyst to formulate an interpretation and the way in which the patient responds to such formulation, that is, how an interpretation is validated or not. Paul, quoted by Ramzy and Shevrin (1976), points out the need to make a distinction between the psychoanalyst's evaluation of a situation and the interpretation. The analyst's silent and uncommunicated evaluation of the situation constitutes a hypothesis, while the interpretation refers to the communication to the patient of all or part of this hypothesis. Sand (1983) raises the need to discriminate between the formation of hypotheses (how an idea is obtained) and what corresponds to their confirmation (how this idea is proven to be correct), in order to avoid confusion of categories.

Isaacs (1939) raised a controversial issue when she pointed out the need to discuss the tests and verifications to which analysts submit the views they hold in their interventions. At that time, she stressed the practical importance, for analysts and non-analysts alike, of establishing criteria to test the validity of interpretations when thinking about the progress or non-progress of the patients. However, the discussion has focused mostly on what, how, and when to interpret rather than on the validation of the interpretation (Schmidl, 1956). Ramzy and Shevrin (1976) raise with surprise the issue of the scarcity of publications on this subject (fewer than 20) in a period of 50 years. The strength, firmness, or veracity of the interpretations constitutes a central point in the psychoanalytic clinical work. Still, the issue of their confirmation (or not) has not been reflected in an in-depth discussion from the theoretical point of view (Sand, 1983).

Searching for truth in clinical psychoanalysis, Hanly (2009) states that the three criteria of the truth of philosophy—coherence, correspondence, and pragmatism—are not in disagreement with each other, but rather work together in psychoanalysis. In the session, the coherence criterion will bring the patient's previous clinical situations closer to the analyst's memory based on the similarity, consistency, and affection that unites them, and will guide our first ideas in the choice of interpretation. The correspondence criterion (the hypothesis that we formulate in relation to what is happening at that moment in the patient's mind and between analyst and patient) will guide the choice of interpretation. These criteria, together with the pragmatic criterion (if the interpretation is clinically useful), will help to decide on the adequacy of the interpretation based on the changes in associations, affects, and transference of the patient.

According to Isaacs (1939), an interpretation may be correct or incorrect. It can be incorrect in two ways: by being false (it is not related to what is in the patient's mind) or by being incomplete (some aspects are not taken into account by the analyst or one aspect is weighed too much to the detriment of another). If the interpretations have no relation to what is in the mind of the patient, they leave the patient cold or indifferent and are dismissed without any movement of affection. However, incomplete interpretations or those that emphasize one aspect of the material more than on another, "falsely emphasized," can provoke intense responses from the patient. Glover (1931) warns of the suggestive effect of inexact interpretation in the context of strong authority.

As Isaacs stated, the main points to confirm correct interpretations are:

• 1 The patient consciously elaborates images and meanings that were implicit in their associations or behaviours.
• 2 They contribute with other associations specifically linked to the interpretation, by amplifying the unconscious attitude or the interpreted fantasy.
• 3 The associations change, or the patient changes his or her attitude, actively rejecting the analyst's hypothesis by highlighting the defence.
• 4 In the following sessions, the patient can bring dreams that continue, elaborate on, and clarify the fantasy or unconscious aspect that was interpreted.
• 5 The patient recovers memories of real experiences from the past.
• 6 New material emerges that evokes a change at the level of the unconscious fantasy.
• 7 There is a reduction or transformation of specific anxieties (e.g., paranoid).

In our opinion, these criteria constitute a landmark. The new contributions, instead of questioning these criteria, aimed to broaden the perspective.

To determine the validity of an interpretation, Schmidl (1956) proposes to observe whether there is an analogy ("relationship identity" or at least "relationship similarity") between the proposition to be interpreted and the given interpretation. Each one has a specific gestalt, and both gestalts fit together so closely that there can be no doubt about their joint membership. First, it is necessary to ask ourselves if the gestalt of the proposition to be interpreted (e.g., a dream or a symptom) and the gestalt of the material used for the interpretation are sufficiently determined by concrete elements and their interconnection to justify their meaning. Second, we must ask ourselves if the gestalts fit one into the other—e.g., approach the implicit theory in the analyst's mind as closer to the real patient.

In order for the patient to test an interpretation for its content of truth or falsehood, it should be formulated precisely, clearly, and unambiguously and not include more than one hypothesis (Etchegoyen, 1986,2001). Long and complex interpretations, in addition to offering more than the patient may think, are difficult to test. Providing them in parts favours waiting for the answer at each step and adjusting the course accordingly. Etchegoyen warns about the risk of arbitrariness and whims of our theories and even our conflicts if the patient's response to the interpretation is not taken into account, not only in its informative and heuristic value but also in its evaluation. Verbal, para-verbal, and non-verbal associations and the insistence of certain signifiers of the patient allow the analyst to test hypotheses within the session. The repetition in the transference allows the analyst to do so throughout the process. The validity of the process is reached at the point of convergence where the findings of the session are prolonged in the slow and persistent changes that appear in it, an aspect that the 3-LM has been able to show.

From an intersubjective and a field perspective, M. and W. Baranger (2008) point out, as process and non-process indicators, the convergence of the variability of the patient and analyst languages, the enrichment of the patient's story, and the affective exchange and the alternation with blocking moments.

From a dyadic and interactive conception, Jiménez (2009) suggests that in the analyst's mind there is continuous decision-making influenced by the patient's reactions and actions and the "implicit" use of explicit theories as mini working models. The task of the analyst is not explanatory but predictive, i.e., to evaluate how the patient reacts if the analyst intervenes in one way or another. During the session, the analyst's mind moves between theoretical and practical reasons. Theoretical reasons allow the analyst to understand and explain the interaction on the basis of knowledge acquired—the implicit use of explicit theories. The practical reasons seek to answer, out of a series of alternatives, which would be the best in clinical utility. The singularity of the encounter in the here-and-now of the session marks the decision to intervene. It constitutes an "ideographic, creative, and ineffable" (p. 236) moment in which the analyst assumes a risk that might not be fully explained by theory. The validation or refutation of the analyst's interventions takes place in the course of this interaction where analyst and patient continuously interpret and negotiate, agreeing or constraining their differences in what they observe and what they consider "clinical facts" and their signification.

Then, the construction of an interpretation constitutes a complex silent and non-communicated evaluation, in which different aspects intervene. On the one hand, the unconscious and preconscious perception of the patient's mental states through free associations, transference, and counter-transference, change in affects, and para-verbal and non-verbal expressions intervene. On the other hand, the theoretical knowledge, the explicit theories, are used implicitly as mini operational models at the time of the session. It does not constitute an isolated process in the analyst's mind, but it occurs in an interactive, intersubjective context and within a dynamic field (Baranger & Baranger, 1961). In this way, the analyst elaborates a hypothesis in correspondence with what is happening in the patient's mind and formulates it so that the patient can evaluate it in its degree of truth or falsehood.

Summarizing, the criteria formulated above by Isaacs remain accurate as evidence to validate the interpretation. It has to be complemented with an interactional and field perspective focused on the evidence necessary to make decisions about what to interpret from a pragmatic and predictive clinical perspective.

Clinical evidence of changes over extended periods of analysis: The 3-LM

The efficacy of psychodynamic therapy in many mental disorders has been confirmed using the methodology of EBM (randomized control trials, meta-analysis) (Leichsenring et al., 2015). Systematic single case research (Kachele et al., 2008) uses a multi-level observational strategy using tape-recorded material. It combines traditional clinical observation with the application of assessment instruments applied by independent researchers, aiming at a systematic description of the aspects and dimensions of the psychoanalytic process. In this way, it combines traditional clinical understanding and the requirements of greater objectification.

The Three-Level Model (3-LM) to observe the patient's transformations (Altmann de Litvan, 2014; Bernardi, 2014a, 2014b) uses only the clinical perspective. We will not present this model here, but we will point out some aspects related to the issue of evidence. The 3-LM constitutes a guide or heuristic to refine and systematize observation of the patient's transformations from the perspective of multiple analysts with different theoretical approaches. Therefore, it seeks, in three successive steps or levels, to find in the clinical material the evidence that allows for identifying, conceptualizing, and formulating explanatory hypotheses about the patient's changes.

The empirical basis on which the 3-LM is based is given by the resonance of a transcribed or tape-recorded clinical material in a group of analysts who listen to it, making use of the "third psychoanalytic ear." This means making full use of the subjective and intersubjective empathic understanding. It is, therefore, an "observation of observation," where the participants take a "second look" at the patient, the analyst, and their own reactions in the group and in an observer-participant attitude. In this sense, it attempts to objectively describe phenomena that are essentially subjective and intersubjective. It moves "bottom up," from the experience as the analyst perceives it from an analytical phenomenological perspective, towards conceptualizations and explanatory hypotheses of a higher level of abstraction. The shared group resonance of the clinical material is the first source of evidence.

At the second level, evidence rises from the critical descriptive analysis of the clinical material, taking into account its group resonances in the previous level. The group is required to adopt a descriptive-analytical attitude in order to conceptualize the dimensions in which the patient's transformations occurred. The questions posed by the 3-LM are based on the diagnostic dimensions of patients' changes, which have been operationalized by manuals such as the OPD-2 and PDM-2. These dimensions encounter several problems. On the one hand, the operationalization is controversial inside psychoanalysis, which makes the use of shared diagnostic criteria difficult. On the other hand, data required by the assessment of the 3-LM dimensions are not always available in the clinical material. It is important to be also aware of other potential limitations. Initial psychoanalytic interviews are often of an exclusively associative type and do not include exploratory aspects, which does not allow for adequate investigation of aspects that the patient splits or omits. The limited number of sessions examined favours selection bias. The analyst does not always select sessions according to the criteria of qualitative research, showing both evidence in favour of the existence of changes and no changes. The greatest restriction appears when the material does not comply with Isaacs's criteria of clinical evidence and contextual information is insufficient.

The clinical material is convincing when the anchor points underlined by the discussion group converge with the foci of the interventions of the analyst, with the patients' structural changes, and with their assimilation and insight about their problematic experiences (see Chapter 4, the discussion of John's case). The clinical material is less convincing if the analyst's assumptions are not confirmed by relational episodes inside and outside the analysis, which show the changes in the patients' vision of their own self and of the others at both the real and fantasy levels. The importance of functioning at the level of the self and with others, highlighted in current psychodynamic manuals, has also been hierarchized by the World Health Organization to define personality disorders in its new classification of illnesses (ICD-11) (World Health Organization, 2018) and by the DSM-5 in its alternative model, section III (American Psychiatric Association, 2013).

The third level seeks to formulate explanatory hypotheses about the patient's changes or no changes. The aim is to find the best possible explanation of the changes observed in the analytic process with those that occurred in the functioning of the patient's personality. Since these hypotheses are speculative, it is important to compare them through thought experiments. In this way, the rationale of each is exposed, making it possible to examine the internal consistency of the analyst's working models and implicit theories and their correspondence with the clinical facts. The 3-LM long-term perspective has a positive effect on the search for evidence. On the one hand, participants make contact with the clinical material by following the way in which the events occurred, that is, from before to after. During the group discussion, it is also possible to examine this first impression a posteriori. The perspective from the "after" on the "before" allows participants to check the predictions and to examine factors that may have influenced the evolution of the analysis.

It is possible to extend this view over time. Often the 3-LM groups help analysts to change their interpretative strategy and to observe the consequences. Other times it is possible to discuss a material again after some time. In both cases, the new information facilitates reviewing the validity of the initial conclusions.

Systematic single-case research and the 3-LM have been used in a complementary way, which favours the increase of clinical and extra-clinical evidence available (Rodríguez Quiroga de Pereira, 2018). This type of research can contribute significantly to the development of psychoanalytic practice-based evidence.

Limitations, omissions, and biases

It is as important to discover and prioritize the evidence provided by clinical material, as it is used to identify where this evidence is not sufficient or not reliable. As we said in relation to EBM, something similar occurs in the field of health sciences. The study of errors in medical decision-making has increased recently, after the beginning of the patient safety movement (Henriksen & Brady, 2013). Of particular interest is the systematic error, when biases or omissions of significant information are recurrent.

In the field of psychoanalysis, perhaps the strongest criticism came from Grünbaum (1984), who argued that the patient's response to the analysis could be explained as an effect of suggestion and response to the analyst's expectations. This provoked multiple responses, some of which are very recent (Azcona, 2019). Phenomena of suggestion and indoctrination can occur in analysis, but this does not mean that they always occur, nor that it is easy to produce them in all cases. This suggests reversing the burden of proof and studying when and in what analysis suggestion occur. Evidence is greater when the material shows answers from the patient that confirm in an unexpected and surprising way the hypotheses of the analyst. However, multiple other factors can distort clinical evidence and oblige practitioners to adopt a cautious attitude regarding them.

Recently, research on EBM focused on situations when available studies provide only low or limited level of evidence (Detterbeck et al., 2018). In such cases, they suggest weighing the evidence according to an explicit framework and countering the temptation to think it is more robust than it actually is. They propose to use a framework that includes (1) the degree of confounding in non-randomized comparisons; (2) the uncertainty involved in extrapolation from indirect evidence; (3) the plausibility of a rationale; and (4) the systematization of clinical casuistic. These criteria are useful for clinical studies. Confounders, i.e., factors that influence both the intervention and the outcome, can mislead the attribution of causality. For example, better outcomes can be attributed to a high frequency of sessions, without examining if high-frequency patients are at the same time more motivated, have less stressful events, etc., which can also affect the outcome. Extrapolation using indirect evidence is frequently used, e.g., when a famous case is adopted as a general model for other cases without examining their similarities and differences. Sometimes a clinical conclusion rests on a rationale based on theoretical principles without consideration of reasonable alternative explanations. Clinical experience and clinical intuition are a valuable source of hypotheses, but they also are vulnerable to numerous biases.

A bias is an a priori tendency that leads to favouring a certain type of conclusion over other possible ones. They can occur intentionally or unintentionally. More than 50 types of cognitive biases and more than a dozen affective biases have been identified in the field of medical diagnosis (Croskerry, 2009).

Psychoanalysis has extensively studied counter-transference and the effects that the patient may exert consciously or unconsciously on the analyst, as well as the reactions or enactments of the analyst. The extent to which counter-transference can be an obstacle and/or an instrument for the progress of analysis has been also discussed at length. However, the study of other types of biases or omissions has not received similar attention, so, without intending to carry out a systematic study here, we would like to draw attention to their importance.

Biases are present in all fields of human knowledge, and the study of their role in decision-making in situations of uncertainty acquired greater notoriety from the work of D. Kahneman (2011), psychologist and economist, who received the Nobel Prize for his research on these topics.

The "confirmation bias" (also called "my side bias") that leads to looking at clinical phenomena from only one theoretical perspective is frequent in psychoanalysis. The influence of theoretical assumptions has been considered an insurmountable obstacle to achieve a clinical common ground between analysts with different theoretical affiliations (Green, 2005). Although the confirmation bias is usually considered a cognitive bias, from a psychoanalytic point of view, the intense emotions related to professional and personal identity that are usually associated with belonging to psychoanalytic schools cannot be ignored. This leads to the exclusion of hypotheses based on different theoretical premises. The groupthink phenomena reinforce an attitude in which freedom of thought is limited by the need to maintain agreement with the rest of the group. Confirmatory biases favour circular reasoning and "illusory correlations" or "apophenia," biases that lead to the establishment of an association between unrelated phenomena. For example, narratives that relate certain problems of the patient to specific childhood events or maternal attitudes are often considered an indisputable causal, although a closer examination shows that there are other equally defensible hypotheses.

The path followed by the 3-LM as it moves from experience to theoretical hypotheses and includes participants with diverse theoretical affiliations decreases this type of confirmatory bias. Instead, it promotes the search for specific models adjusted to the patient. It cannot prevent the emergence of other types of biases, such as groupthink, selection biases that lead to the omission of clinically meaningful material, and anchoring biases, which deem the information that arrives first more important, and to the "halo effect" that leads to the initial impressions spreading to other areas.

For that reason, it is key that clinical discussions are attentive to such biases. Kahneman (2011) used the acronym WYSIATI (What You See Is All That Is) to describe the belief that the information available is sufficient when it is not, because we see only a part of the problem. Some types of bias, such as "congruence bias," "selection bias," "memory bias," and "authority bias," are more frequent when setting aside a broader contextual perspective and attention is focused only on certain aspects of the material. Confirmation biases and omissions of information reinforce each other.

The importance and pervasiveness of biases and omissions have led several researchers in the field of psychotherapy to question whether the therapist can reliably assess a patient's progress. Lilienfeld et al. (2014) note that the combination of biases may lead to (1) perceiving changes that do not exist; (2) if they do exist, attributing them to the treatment when in fact they are due to other causes; or (3) relating them to specific factors in a therapeutic approach, when in fact they are due to common or nonspecific factors. For this reason, many authors render evaluation through tests and questionnaires more important than the opinion of the therapist (Hannan et al., 2005; Lilienfeld, 2007; Hatfield et al., 2010).

Without denying the usefulness of triangulating the therapist's opinion with other forms of assessment, our interest here is to improve clinical evidence, reducing biases and achieving wider and more reliable clinical information. In the field of health sciences, Wilson and Brekke (1994) consider cognitive bias "mental contamination" and de-biasing "mental correction." A crucial issue is the awareness of the direction and magnitude of biases and the motivation to correct biases. This is also the direction taken by this chapter. Kahneman suggests that there are two types of systems of thinking. "Fast thinking" is an automatic, frequent, emotional, stereotypic, and not conscious heuristic for problem-solving, used most of the time. "Slow thinking," on the contrary, implies logical, calculating, and conscious reasoning, which leads to more reliable although more effortful decisions. Clinical discussion groups force implementation of this second system of thinking. The 3-LM expands Baranger, Baranger, and Mom's (1983) "second look" concept by combining analysis of the analytical field during the sessions with a prolonged observation in time and with the contribution of participants with diverse theoretical approaches. This broadening of the perspective does not eliminate, as we said, the possibility of bias, but favours critical thinking.

The critical review of the strengths and limitations of the evidence offered by clinical material should be part of clinical papers. Their publishing in psychoanalytic journals does not usually include discussions of this kind. Usually authors present their perspective as the only one possible. When taking into account divergent authors, it is frequent that the work highlights the complementary aspects, without discussing the contradictory aspects or discussing the arguments for or against each of them. This situation does not promote the scientific development of the discipline.

The "Discussion" section of a scientific paper should fulfill two main functions. First, it should offer authors the opportunity to present the value of their results and proposals. At the same time, authors should acknowledge the limitations of their study. That means to mention where the results are inconclusive or present weaknesses or limitations in their scope. For editors and reviewers, the inclusion of a realistic and adequately self-critical description of these limitations is not a weakness but a strength of a clinical paper.

This is not an exclusive problem of psychoanalysis. loannidis (2007), in a paper provocatively entitled "Limitations are not properly acknowledged in the scientific literature," shows the consequences of this lack of critical appraisal. The acknowledgment of limitations, in his opinion, has to go beyond identification of systematic errors and validity problems. It requires an interpretation of its meaning and influence on the published findings. Knowledge and discussion of limitations are essential for credibility and genuine scientific progress. Insufficient acknowledgment of limitations contributes to the current crisis of conceptual reproducibility of scientific findings. However, sometimes authors are not aware of the limitations of their papers. For this reason, Avidan, loannidis, and Mashour (2019) suggest the inclusion of independent external reviewers in the "Discussion" section of scientific papers. In psychoanalytic publications, it would be ideal for authors to expose the inner path they took to reach their conclusions, showing both the arguments that convinced them of their conclusions and those that raised doubts or suggested alternatives. The 3-LM discussions are brainstorming processes where alternative hypotheses are confronted. The challenge is to include this kind of deliberation in the clinical papers.

Towards a framework for evaluating clinical evidence

What follows is a first attempt to offer a framework for evaluation of the level of evidence (high, moderate, or low) provided by clinical material. We are putting together the criteria that clinical psychoanalysis has always taken into consideration and we examined in the previous paragraphs. The result is a multidimensional assessment that helps critical reflection about clinical judgments and theory building.

Quality of the analytical process: Transformations in the session

The weight of evidence provided by an analysis increases when significant transformations occurred in the patient and in the therapeutic relationship.

This implies, first, evaluating the short-term effect of the analyst's interpretations. The criteria proposed by Isaacs (1939) remain unquestioned. They underline unconscious answers from the patient, which limit the most direct effects of the suggestion or the expectations of the analyst. In this sense, the importance of the manifestation of unexpected effects that are surprising for the analyst and for the patient themselves should be emphasized. When the clinical material meets the criteria indicated by Isaacs, with the required contextual information, we can affirm that the level of evidence it offers as support for hypotheses is high. On the contrary, the level of evidence is low or absent when we do not find answers of this type from the patient but rather conscious answers to the analyst's interpretations, in which the analyst's theoretical orientation is perceptible. When the analyst remains mostly silent, it is necessary to examine to what extent this leads to the patient's free association or to increase a defensive discourse.

The weight of the evidence increases if, following Baranger and Baranger's (2008) descriptions, we consider a dynamic field co-created by the patient and the analyst, which allows visualizing with more clarity of the progress of the analytical process and the shared resistances (bastions, enactments) or the sequences of ruptures and repairs of the link. The level of evidence increases when the analyst is able to identify these phenomena and predict the effect of the interpretations that will be effective. In this way, it is possible to verify that there is a real exchange and attun-ement between patient and analyst.

On the contrary, the level of evidence is low when the analyst's thread of argument is possible to discover only through interpretations, while on the patient's side we only find repetitive or conscious topics, which sometimes do not even clearly show what leads them, at a conscious and unconscious level, to continue the analysis.

Between clear cases of high and low evidence, there are intermediate degrees that can be classified as moderate evidence.

The transformations that take place during long-term periods of analysis

We find a second criterion of high-level evidence in changes at the patient's level of psychopathology and personality throughout the treatment. The second level of the 3-LM seeks to evaluate those changes. The weight of the evidence increases when there is a follow-up of the case or added evaluation tools from single-case research.

Evidence increases when patients' changes can be related to the analytic process through specific hypotheses about possible mechanisms of change. When there is no explanation of how the analytical process had an effect on the existence or non-existence of changes, evidence should be considered moderate or low. Evidence is also low when changes are described in terms of purely behavioural or symptomatic modifications without showing connections with changes in the patient's inner world and personality functioning. It is important to include contextual information about other factors that could have influenced patients' changes.

Discussion of alternative hypotheses

As already explained, clinical material can usually be interpreted from more than one theoretical perspective. Evidence is greater when different interpretative hypotheses are confronted and the strengths and limitations of each are critically compared. Each hypothesis should account for neglected aspects that are highlighted by competing hypotheses. This is the function of mental thought experiments. Although they do not provide direct factual evidence, they force the working models that seek to account for the situation to be formulated more precisely. This allows listeners to put their similar experiences into play, evaluating which hypothesis is more convincing, thus reinforcing the level of evidence. The discussion of alternative hypotheses helps to reduce omissions in the presentation of the case.

Table 14.1 Assessment of the weight of evidence provided by the clinical material for answering 3-LM third-level questions

Ricardo Bernardi and Luisa Pérez Suquilvide

Evidence levels decrease when the severity of the dysfunctions and split-off aspects of the self and the object are not investigated, or the duration, extension, and severity of the problems are not mentioned. Sometimes the clinical material does not allow for the distinction of a transitory state from a permanent personality trait, or problems related to unconscious conflicts from failures in integrating the structural aspects of psychic functioning. The assumption that exclusive attention to free association leads to better capture of unconscious phenomena leads to highlighting only aspects of the material in agreement with the previous theoretical premises, i.e., to a "my side bias."

Thoughtful discussion of alternative hypotheses led to a better understanding of the context. We agree with Boesky (2013) when he says, "We need to know how and why certain material was deemed meaningful" (p. 1138), and

How we think about "evidence" is inseparable from how we think about contextualization. At issue is a call for wider recognition of the problems we perpetuate in continually denying the profound epistemological penalties we have incurred by our neglect of this gap in our methodology, how we gather and evaluate our contextualizing criteria. . . . We cannot continue to delay the necessity of facing this problem

(p. H42)

Table 14.1 summarizes some of the characteristics that allow us to evaluate the weight of evidence of clinical material. This attempt to establish a gradation should be considered preliminary and necessarily incomplete.

References

Altmann, M. (2014). Time for change: Tracking transformations in psychoanalysis—The Three-Level Model. London: Kamac.

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th edn). Arlington: American Psychiatric Association.

APA Presidential Task Force on Evidence-Based Practice. (2006). Evidence-based practice in psychology. American Psychologist, 61(4), 271-285.

Avidan, M. S., loannidis J. P. A., & Mashour, G. A. (2019). Independent discussion sections for improving inferential reproducibility in published research. British Jotirnal of Anaesthesia, 122(4), 413-420.

Azcona, M. (2019). Hacia una redefinición del problema de la sugestión en psicoanálisis: minimizando el efecto Pigmalión [Towards a redefinition of the problem of suggestion in psychoanalysis: Minimizing the Pygmalion effect]. Perspectivas en Psicología, 16(2), 98-107.

Baranger, M., & Baranger, W. (1961). La situación analítica como campo dinámico [The analytical situation as a dynamic field]. Revista Uruguaya de Psicoanálisis, 4.

Baranger, M., & Baranger, W. (2008). The analytic situation as a dynamic field. International Journal of Psychoanalysis, 89(4), 795-826.

Baranger, M., Baranger, W., & Mom, J. (1983). Process and non-process in analytic work. International Journal of Psychoanalysis, 64,1-15.

Barkham, M., & Mellor-Clark, J. (2003). Bridging evidence-based practice and practice-based evidence: Developing a rigorous and relevant knowledge for the psychological therapies. Clinical Psychology & Psychotherapy, 10,319-327.

Bernardi, R. (1989). The role of paradigmatic determinants in psychoanalytic understanding. International Journal of Psychoanalysis, 70(2), 341-357.

Bernardi, R. (2002). The need for true controversies in psychoanalysis: The debates on Melanie Klein and Jacques Lacan in the Rio de la Plata. International Journal of Psychoanalysis, 83, 851-873.

Bernardi, R. (2014a). The assessment of changes: Diagnostic aspects. In M. Altmann (Ed.), Time for change: Tracking transformations in psychoanalysis—The Three-Level Model (pp. 263-278). London: Karnac.

Bernardi, R. (2014b). The Three-Level Model (3-LM) for observing patient transformations. In M. Altmann (Ed.), Time for change: Tracking transformations in psychoanalysis—The Three-Level Model (pp. 3-44). London: Karnac.

Bernardi, R. (2015). What kind of discipline is psychoanalysis? International Journal of Psychoanalysis, 96, 731-754.

Bernardi, R. (2016). La formulación clínica del caso: su valor para la práctica clínica [The clinical formulation of the case: Its value for clinical practice]. In R. Bernardi, B. Varela, D. Miller, R. Zytner, L. de Souza, & R. Oyenard (Eds), La formulación psicodinámica de caso. Su valor para la práctica clínica [The psychodynamic formulation of a case: Its value for clinical practice] (pp. 25-24). Montevideo: Grupo Magro Editores, Universidad Católica del Uruguay.

Blass, R. B. (2013). Psychoanalytic controversy. Introduction: What does the presentation of case material tell us about what actually happened in an analysis and how does it do this? International Journal of Psychoanalysis, 94, 1129-1134.

Boesky, D. (2002). Why don't our institutes teach the methodology of clinical psychoanalytic evidence? Psychoanalytic Quarterly, 71(3), 445-475.

Boesky, D. (2013). What does the presentation of case material tell us about what actually happened in an analysis and how does it do this? International Journal of Psychoanalysis, 94(6), 1135-1143.

Brown, J. R., & Fehige, Y. (2019). Thought experiments. In Edward N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy Archive (Winter 2019 Edition), https:// plato.stanford.edu/archives/win2019/entries/thought-experiment/.

Chabert, C. (2013). Psychoanalytic controversy. Response: What does the presentation of case material tell us about what actually happened in an analysis and how does it do this? International Journal of Psychoanalysis, 94,1153-1162.

Croskerry, P. (2009). A universal model of diagnostic reasoning. Academic Medicine, 84(8), 1022-1028.

Da Rocha Barros, E. (2013). Psychoanalytical controversy: What does the presentation of case material tell us about what actually happened in an analysis and how does it do this? International Journal of Psychoanalysis, 94,1145-1152.

Detterbeck, F. C., Gould, M. K., Lewis, S. Z., & Patel, S. (2018). Extending the reach of evidence-based medicine. Chest, 153(2}, 498-506.

Etchegoyen, R. H. (1986). Los fundamentos de ¡a técnica psicoanalítica [The basics of psychoanalytic technique]. Buenos Aires: Amorrortu.

Etchegoyen, R. H. (2001). Algo más sobre el testeo del proceso clínico [Something else about testing of the clinical process]. Subjetividad y Procesos Cognitivos, 1, 34-59.

Freud, S. (1914). On narcissism. S.E. 14, 67-102.

Freud, S. (1918). From the history of an infantile neurosis. S.E. 17,1-12.

Freud, S. (1933). New introductory lectures on psychoanalysis. S.E. 22,1-182.

Glover, E. (1931). The therapeutic effect of inexact interpretation: A contribution to the theory of suggestion. International Journal of Psychoanalysis, 12, 397-411.

Green, A. (2005). The illusion of common ground and mythical pluralism. International Journal of Psychoanalysis, 86(3}, 627-632.

Grünbaum, A. (1984). The foundations of psychoanalysis: A philosophical critique. Berkeley: University of California Press.

Hanly, C. (2009). On truth and clinical psychoanalysis. International Journal of Psychoanalysis, 90(2), 375-383.

Hannan, C., Lambert, M. J., Harmon, C., Lars Nielsen, S., Smart, D. W., & Shimokawa, K. (2005). A lab test and algorithms for identifying clients at risk for treatment failure. Journal of Clinical Psychology in Session, 61,155-163.

Hatfield, D., McCullough, L., Frantz, S. H. B., & Krieger, K. (2010). Do we know when our clients get worse? An investigation of therapists' ability to detect negative client change. Clinical Psychology and Psychotherapy, 17,25-32.

Henriksen, K., & Brady, J. (2013). The pursuit of better diagnostic performance: A human factors perspective. BMJ Quality & Safety, 22, Ü1-Ü5.

loannidis, J. P. A. (2007). Limitations are not properly acknowledged in the scientific literature. Journal of Clinical Epidemiology, 60,324—329.

Isaacs, S. (1939). Criteria for interpretation. International Journal of Psychoanalysis, 20,148-160.

Jiménez, J. P. (2009). Grasping psychoanalysts' practice in its own merits. International Journal of Psychoanalysis, 90(2), 231-248.

Kachele, H., Schachter, J., & Thomä, H. (2008). From psychoanalytic narrative to empirical single case research: Implications for psychoanalytic practice. New York: Routledge.

Kahneman, D. (2011). Thinking, fast and slow. New York: Farrar, Straus and Giroux.

Leichsenring, F., Luyten, P., Hilsenroth, M. J., Abbass, A., Barber, J. B., Keefe, J.

R., Leweke, F., Rabung, S., & Steinert, C. (2015). Psychodynamic therapy meets evidence-based medicine: A systematic review using updated criteria. Lancet Psychiatry, 2, 648-660.

Lilienfeld, S. O. (2007). Psychological treatments that cause harm. Perspectives on Psychological Science, 2(1), 53-70.

Lilienfeld, S. O., Ritschel, L. A., Lynn, S. J., Cautin, R. L., & Latzman, R. D. (2014). Why ineffective psychotherapies appear to work: A taxonomy of causes of spurious therapeutic effectiveness. Perspectives on Psychological Science, 9,355-387.

Ramzy, L, & Shevrin, H. (1976). The nature of the inference process in psychoanalytic interpretation: A critical review of the literature. International Journal of Psychoanalysis, 57,151-159.

Rodríguez Quiroga de Pereira, A. (2018). La formulación psicodinámica de caso. Su valor para la práctica clínica [Psychodynamic case formulation: Its value for clinical practice]. International Journal of Psychoanalysis, 99(1), 266-271.

Rodríguez Quiroga de Pereira, A., Borensztein, L., Corbella, V., & Marengo, J. C. (2018). The Lara case: A group analysis of initial psychoanalytic interviews using systematic clinical observation and empirical tools. International Journal of Psychoanalysis, 99(6), 1327-1352.

Sackett, D. L., Straus, S. E., Richardson, W. S., & Rosenberg, W. (1997). Evidencebased medicine: How to practice & teach EBM. New York: Churchill Livingstone.

Sand, R. (1983). Confirmation in the Dora case. International Review of Psychoanalysis, 10,333-357.

Schmidl, F. (1956). The problem of scientific validation in psycho-analytic interpretation. International Journal of Psychoanalysis, 36,105-113.

Shedler, J. (2015). Where is the evidence for "evidence-based" therapy? Journal of Psychological Therapies in Primary Care, 4,47-59.

Tuckett, D. (1995). The conceptualization and communication of clinical facts in psychoanalysis. International Journal of Psychoanalysis, 76,653-662.

Tuckett, D. (1998). Evaluating psychoanalytic papers: Towards the development of common editorial standards. International Journal of Psychoanalysis, 79, 431-448.

Tuckett, D. (2001). Evidence-based psychoanalysis. International Forum of Psychoanalysis, 10(3-4), 211-215.

Wilson, T. D., & Brekke, N. (1994). Mental contamination and mental correction: Unwanted influences on judgments and evaluations. Psychological Bulletin, 116, 117-142.

World Health Organization. (2018). International classification of diseases for mortality and morbidity statistics (11th rev.), https://icd.who.int/browsell/l-m/en.

Commentary: Perspectives from the Philosophy of Science—Charles Hanly

In their study, Bernardi and Pérez Suquilvide are seeking to strengthen two aspects of psychoanalytic knowledge: the objectivity of its clinical base and the implications of reliable clinical facts for psychoanalytic theory. This is an important step because psychic processes are often over-determined, e.g., Oedipal fantasies are phallic and narcissistic. As a result, theoretical differences have arisen over causal explanations that are compatible, or, with modifications, can be shown to be so. There are also theories or components of different theories that are incompatible, e.g., "the Oedipus complex is instinctual" as opposed to "the Oedipus complex is caused by narcissistically inadequate parenting." In assessing evidence for evaluating interpretations and the theories they imply, the sorting out of such issues is essential. Bernardi and Pérez Suquilvide have usefully explored the need to estimate the relevance and degree of evidence for and against alternative psychoanalytic theories. This undertaking is an essential aspect of the 3-LM's contribution to finding ways to integrate compatible theories and to find evidence to logically test contrary or contradictory alternative theories (Hanly, 2019). The authors take up Freud's empiricist premise that what happens and does not happen in an analysis is epistemologically fundamental to psychoanalytic knowledge. In this spirit, the "discussions" in this book entertain alternative assumptions and perspectives, not as a priori principles but as hypotheses to be tested against clinical verbatim material (Hanly, 1983). Bernardi and Pérez Suquilvide tackle the problems of how to find and evaluate the evidence for this important undertaking for the future of psychoanalysis.

Their basic position is addressed in the structure of the Three-Level Model. When it comes to validating (and falsifying) psychoanalytic hypotheses, there are three basic clinical phenomena to consider: the patient's presenting problems embedded in the patient's symptoms and inhibitions, and expressed (made evident) in the patient's feelings about the analysis and the analyst (referred to as "anchor points" in the model); their treatment in an analytic process of empathic interpretation; and the gains or lack of gains in these aspects of the treatment measured by the extent and durability of the amelioration of the symptoms and inhibitions after the treatment has ended. This is the core of the 3-LM method.

Psychoanalysis is a clinical discipline. Although it has powerful implications for the humanities and social sciences, the truth of its findings is vouchsafed only by clinical experience or what Grünbaum has called the tally argument. Grünbaum's critique of the tally argument can be shown to be fallacious, although I agree with his inductivism in the construction of knowledge. Grünbaum's critique should not deter us from being guided by the core idea of the tally argument (Hanly, 1988, 1992). In the extent to which psychoanalysis can genuinely and durably ameliorate symptoms, inhibitions, and related psychic incapacities is evidence that the interpretations that have brought about this amelioration do "tally" with what had caused them in the patient. Bernardi and Pérez Suquilvide tackle the fundamental problems of how to assemble and assess this clinical evidence.

The authors propose that a core problem of psychoanalytic clinical evidence is that analysts can be subject to bias. Although ideas are required for intelligently organized observation (Hanly, 1995), ideas need not, but may, fail to yield unbiased observation. An observational science must address and solve these problems of reliability as much as possible. For example, the personal analysis is supposed to facilitate a sufficient degree of clinical objectivity. But it may fail, for example, if the training analyst fails to analyze the candidate's idealizing transference, leaving the candidate predisposed to over-value the analyst's theoretical ideas and to devalue ideas that are contrary to them (Hanly, 1982). The unanalyzed idealizing transference turns theory into ideology and inhibits the open-minded skepticism needed to entertain ideas from alternate perspectives and to allow the details of what happens to the patient in the analysis to decide what theory is most probable. Ideological ideas are not subject to criticism on the basis of clinical facts, i.e., evidence. Nor do they favour the ability to carry out the thought experiments described by Bernardi and Pérez Suquilvide that allow analysts to check out the bearing of clinical facts on alternative theories by contextualizing the facts (i.e., in interpretation sequences that have brought about real symptom reduction). The importance of being able to employ alternative perspectives in the history of human knowledge is evident in Copernicus's revolutionary speculation about what we would see if the observer rotated and the sun stood still. Or much later Freud's question "What if sexuality had its onset with birth?" What Cooper (2008) identifies as "theoretical and technical dogmatism" may also be rooted in the analyst's idealizing transference to his or her own analyst and the aggression repressed by the idealization becoming attached to alternative theories causing the analyst to miss evidence in associations and transferences. The analyzed psyche of the analyst employing the methods of free association, transference observation, and interpretation is the only instrument we have to learn about the nature of the psychic unconscious of our patients. Along with their extensive scholarship, the authors explore how reliable evidence for deciding questions of theory can be evaluated even while facing this daunting fact.

In their search for the nature of authentic psychoanalytic evidence and the difficulties that stand in the way of gaining it, Bernardi and Pérez Suquilvide may attribute too much to Boesky's (2008) idea of contextu-alization. Boesky does strongly share the goal of finding reliable clinical evidence for and against psychoanalytic theories. Moreover, there is a fairness and a respect for others as well as an important cognitive value in making sure that we have understood an interpretation, or its absence, by clarifying and understanding its theoretical context. Contex-tualization is an essential part of the exploration of an analyst's explicit, public, formally espoused theories and the implicit theories that actually inform his analytic work (Sandler, 1992; Canestri, 2012). Bernardi and Pérez Suquilvide make good use in the chapter of Boesky's idea of con-textualization. The 3-LM requires us to use clinical detail to understand the meaning of the associations and transferences of the patient and of the analyst's interpretations communicated to the patient. For these purposes its value cannot be exaggerated. However, contextualization is not in itself adequate for meeting Cooper's (2008) wish for a method that can identify two or more logically inconsistent alternative theories and find out which interpretive strategy tallies best with the causes of the patient's neurosis because it heals symptomatic and inhibited functioning and restores normal functioning.

Boesky's idea of contextualization requires that we find the theory that makes an interpretive strategy fully meaningful, when any other contextualization would deprive the interpretations of their meaning or even distort their meaning. To this one must surely concur. But nothing in this procedure provides evidence for whether or not an interpretation tallies with what was causing the symptoms and inhibitions in a patient that have undergone change in an analysis. Contextualization establishes only that the interpretation is coherently related to essential components of a theory, be it Freudian, Kleinian, relational, self-psychological, etc., the effectiveness of which in any particular clinical situation is left up in the air. The results of contextualization, treated as evidence itself, leaves psychoanalytic knowledge divided in isolated mutually untestable, solip-sistic bodies of knowledge. The coherence involved is a necessary but not a sufficient criterion for reliable evidence. Contextualization does not provide the evidence we need to work toward a more integrated psychoanalytic theory composed of the most probable knowledge we now have, open to fresh evidence to further improve it and methods for doing so. For this work we need contextualization but also factual, objective evidence of significant beneficial changes in patients in analysis and the interpretative strategies that caused them. Only with these clinical facts in hand can we then test theories.

Bernardi and Pérez Suquilvide significantly address the problems of assessing the merit and weight of evidence for the efficacy of interpretations. Freud's tally argument and his use of it to falsify the universality of his initial seduction theory, and which also provided evidence for Freud's hypotheses of pre-pubertal sexuality and the power of unconscious fantasy, establishes the link between the therapeutic efficacy of interpretation and the probability of their correspondence with the psychological processes at work in the patient. This link is the pragmatic evidence for the probable correspondence of psychoanalytic interpretations with the patient's difficulties, and therefore evidence for theories they imply, to which Bernardi and Pérez Suquilvide refer in their chapter.

Given the importance of this pragmatic evidence, it might be useful to comment further on the philosophical and scientific definitions of pragmatism (Hanly, 2009). James (1907) held that a belief is true if it coheres with and enriches a person's other beliefs and experience of life. Hence, a belief in God that gives meaning and purpose to an individual's life is treated by James as evidence of its truth, although all that it proves is that a belief that there is a beneficent God is beneficial to the person whose belief it is. This definition relies in part on the idea that coherence is a sufficient criterion of truth. Freud's idea of pragmatic evidence is very different in being an idea of scientific pragmatism. Let us briefly consider the evidence that caused Freud to modify but not altogether reject the seduction theory. Release from the symptoms of hysteria caused by interpretations based upon a theory of their origin (sexual seduction theory) using a treatment method specific to the theory (cathartic method) would provide sufficient evidence that the theory is true. When the symptomatic relief does not occur or is not durable, although it does work for some cases of the neurosis, we have pragmatic evidence that the theory is true of some but not of all cases of hysteria. The theory is not, as Freud had believed, universal. Some other factors would have to be at work. Freud correctly inferred that patients must have experienced fantasies of being seduced and, therefore, these children must have had a sexual life and had known some form of sexual excitation and pleasure during their childhood. Bringing about, partially bringing about, and failing to bring about a specific change (in 3-LM in an anchor point) is a pragmatic test of the interpretations used and the theory that accredits them. The coherence of belief with other beliefs and experience is not enough.

Bernardi and Pérez Suquilvide have given us a useful (pragmatic in the scientific sense) perspective on assessing, among other sorts of evidence, pragmatic evidence in psychoanalysis. The 3-LM is on the path that leads to the use of scientific pragmatism to test interpretations and theories (here contextualization plays an important part) by realizing, despite difficulties, the potential for objectivity and intersubjectivity (competently trained observers see what is there to be seen for the most part and, hence, are able to agree about the facts) in the clinical observation of change, degrees of change, and no change in specified areas of a patient's functioning. The model is on its way to finding the causal links of sets of interpretations to ameliorative change and, hence, to the evaluation of theories.

The authors' thoughtfully skeptical and demanding study of psychoanalytic evidence is a welcome antidote to the subjectivistic and individualistic "arts and crafts" approach to psychoanalytic knowledge currently in vogue with some analysts. Theirs is a robustly demanding approach to the epistemological, methodological, and logical problems of evidence in a necessarily observational science.

References

Boesky, D. (2008). Psychoanalytic disagreements in context. Lanham: Jason Aronson. Canestri, J. (2012). Putting theory to work: How are theories actually used in practice?

London: Karnac Books.

Cooper, A. M. (2008). Psychoanalysis today: A plurality of orthodoxies. Journal of the American Academy of Psychoanalysis, 36, 235-253.

Hanly, C. (1982). Narcissism, defence and the positive transference. International Journal of Psychoanalysis, 63,427-444.

Hanly, C. (1983). A problem of theory testing. International Review of Psychoanalysis, 10,393-405.

Hanly, C. (1988). The foundations of psychoanalysis: A philosophical critique. Journal of the American Psychoanalytic Association, 36,521-528.

Hanly, C. (1992). Inductive reasoning in clinical psychoanalysis. International Jotirnal of Psychoanalysis, 73, 293-301.

Hanly, C. (1995). On facts and ideas in psychoanalysis. International Journal of Psychoanalysis, 76,901-908.

Hanly, C. (2009). On truth and clinical psychoanalysis. International Journal of Psychoanalysis, 90(2), 363-373.

Hanly, C. (2019). The problem of pluralism in psychoanalysis. In P. Cesar Sandler and G. Pacheco Costa (Eds), On Freud's "The Question of Lay Analysis" (pp. 135-167) Abingdon: Routledge.

James, W. (1907). Pragmatism: A new name for some old ways of thinking. Cambridge, MA: Harvard University Press, 1975.

Sandler, J. (1992). Reflections on developments in the theory of psychoanalytic technique. International Journal of Psychoanalysis, 73,189-198.