Studying places of knowledge practices: the laboratory and beyond

Since the pioneering laboratory studies (Knorr-Cetina Reference Knorr-Cetina1981, Latour and Woolgar Reference Latour and Woolgar1986), STS has paid close empirical attention to the question of how knowledge is produced in practice, in specific sites and institutions, and with specific instruments, tools and settings. Much recent work has taken this approach out of the scientific laboratory, narrowly defined, to include an examination of practices in others sites, including knowledge practices in the social sciences (Williams and Klemmer Reference Williams, Klemmer, Downey and Joseph1997, Maynard Reference Maynard, Houtkoop-Steenstra, Schaeffer and van der Zouwen2002) and work by experts outside universities, such as judges (Latour Reference Latour2002) bankers, economists, and traders (Knorr-Cetina and Preda Reference Knorr-Cetina and Preda2005; MacKenzie, Muniesa, and Siu Reference MacKenzie, Muniesa and Siu2007) or architects (Jacobs and Strebel forthcoming).

This attention to the places and tools of knowledge-production in different empirical sites might afford new opportunities to compare knowledge practices with regard to the relationship between knowledge and the setting, in which it is produced. But these opportunities are not always taken advantage of in this tradition. Too often the vocabulary used to describe work in the natural sciences (“laboratory”, “experiment”) is transferred to other settings without reflecting on the specificity of either the scientific laboratory or the new context (Guggenheim Reference Guggenheim2012).

If the word “laboratory” becomes a metaphor for everything a scholar influenced by STS happens to study, we lose many of the insights gained in the treatment of the laboratory as a place constructed in particular ways (Kohler Reference Kohler1994, Krohn and Weyer Reference Krohn and Weyer1989, Reference Krohn and Weyer1994). The laboratory in the natural sciences is constructed to produce “placelessness”, or better contextlessness, and the set-up is designed so that nothing varies but the object and the conditions that are of interest. In this way the knowledge produced is thought to be independent of its location (Kohler 2008). This is not simply a matter of labelling a place as a “laboratory”; much practical and material work goes into producing this space. Whether or not we think this work is successful in producing the conditions for generalisable knowledge is an entirely different question.

To understand the specificity of psychoanalytic – or any type of – knowledge-production, we need to combine STS’s attention to the places and tools with a comparative agenda drawn from institutional and differentiation-theoretical approaches in the sociology of the professions and of fields.

Sociology of psychoanalysis: taking knowledge-spaces and practices seriously

Previous work on psychoanalysis in the tradition of the sociology of knowledge and the professions has yielded important insights. We have learned about the institutional history of psychoanalysis in the US for example, including its rise (Hale Reference Hale1971) and fall (Strand Reference Strand2011), as well as the success and failure of specific intellectual currents (McLaughlin Reference McLaughlin1998).

But the concrete ways in which knowledge is produced has rarely been the explicit target of studies in this tradition. Rather, knowledge is examined on the way to studying other things, such as forms of governance, or the distribution of status and recognition. Most sociological work on psychoanalysis has looked at it in the context of other ways of establishing authority and making money in the sociology of professions (Abbott Reference Abbott1988), or other ways of governing selves (Foucault 1979; Rose Reference Rose1999), not in the context of other ways of producing knowledge. This work mirrors the strengths and weaknesses of the sociology of professions and fields at large. It has tended to provide an external account that foregrounds power and institutions, and to some extent ideas, but has not tended to make the practices and settings of knowledge-production the focus of the investigation.

For combining the focus on practices and settings inspired by STS with the attention to institutional contexts and comparisons more common in the sociology of professions and fields, some developments in the sociology of professions have been particularly useful: Andrew Abbott has built on earlier studies of the professions and has distinguished between knowledge-related practices on the one hand and their ecologies in the state and the university, on the other (Abbott Reference Abbott2005). This has allowed him to ask new questions about the relationship between the professions and academic disciplines (see also Stichweh Reference Stichweh and Stichweh1994). It allows us to ask how a particular practice, psychoanalysis, relates to these institutional contexts.

We also build on some work in the sociology of professions that has paid attention to the locations and settings of professional work (Abbott Reference Abbott1988; Stichweh Reference Stichweh and Stichweh1994), as well as specific studies of psychoanalysis in the sociology and history of science (Forrester Reference Forrester1996; Lorenzer Reference Lorenzer1973, Reference Lorenzer1974; Rustin Reference Rustin1985, Reference Rustin1987, Reference Rustin, Braddock and Lacewing2007, Reference Rustin, Midgley, Anderson, Nesic-Vuckovic and Urwin2009; Stengers Reference Stengers and Stengers1997; Davies Reference Davies2009; Sealey Reference Sealey2011). Karin Knorr-Cetina has to be credited with the first comparison of the psychoanalytic couch with the scientific laboratory in the sociology of science (Knorr-Cetina Reference Knorr-Cetina and Pickering1992). In our effort to gain an understanding of the couch and the psychoanalytic relationship as the particular material location of psychoanalysis we are especially indebted to Abram de Swaan’s (1980) pioneering work on the psychoanalytic setting and to Andreas Mayer who has shown how this emerged in conversation with hypnosis in the late 19^th century (Mayer Reference Mayer2006).

Research, diagnosis, treatment: Stabilising differentiation with objects, places and practices

Many fields of expertise today distinguish between different aspects of practice, such as the distinctions between research, diagnosis, and treatment. The distinction between these aspects of practice is not just an analytical distinction but is often stabilised with organisations, roles, and objects. Differentiation is, to use the words of ethnomethodology, an ongoing socio-material accomplishment, which is not stable but continually renegotiated and stabilised (Gieryn Reference Gieryn1999). Differentiation creates places, practices and tools that are purified, purpose-built, and more or less unambiguous. Such differentiation creates the advantage of clearly defined goals, the specialisation of tasks and thereby more fine-grained expertise and the ability to ignore many non-essential tasks. The downside of differentiation is the increasing problem of misunderstandings between the different specialisations and the need for complex skills and tools to overcome such divisions (Star and Griesemer Reference Star and Griesemer1989).

This differentiation is in each case the result of specific and contingent historical developments. Since the late 19^th century, medicine, for example, has established the laboratory as the privileged place of knowledge production, through which facts must pass (Whooley Reference Whooley2010a). Today, the research lab is different from the research hospital and from the GP practice. Various kinds of medical instruments exist for different kinds of diagnosis and treatment. Individuals may combine research and treatment roles, but this does not change the fact that these are imagined and evaluated as distinct aspects of practice.

Some fields of knowledge prioritise one of these dimensions of practice over others, but can still be described using these distinctions: architecture, as a profession, for example, has a stronger emphasis on diagnosis (problem analysis) and treatment (buildings) than research; sociology has a stronger emphasis on research than diagnosis or treatment; its professional arm is relatively weakly developed.

Psychoanalysis operates in an environment where neighbouring practices are in various ways defined by such differentiation. Like these practices psychoanalysis also covers the ground of research, diagnosis, and treatment, but unlike them does not distinguish between them. Freud insisted on a “junktim” or link between research and treatment (Freud Reference Freud1958a: 4396, see also Dreher (2000)), which implied not just an admonition to not neglect either aspect of psychoanalytic practice but also a refusal to separate them from each other.^{Footnote 1} Psychoanalysts do not invest in an internally or externally visible differentiation of these tasks by either persons, places or objects.

Andrew Abbott has analysed the state and the university as different ecologies of expert practices and coined the term avatar for parts of a practice related to a core, which extend into another setting across the boundary between profession and discipline (Abbott Reference Abbott2005). This account provides a good description of, for example, the role of criminology in the academy: criminology is an avatar of crime-related professions within the university. Criminology operates in the university and organizes research, yet unlike the traditional disciplines it is not self-reproducing: candidates with PhDs in other disciplines are hired for faculty positions in criminology.

What we see in the case of psychoanalysis is something less than an avatar, which we call a phantom-limb; if an avatar is an actual institutional arm, a phantom limb is the mimicking – in claims and gestures – of an institutional arm that does not exist. Because psychoanalysis does not make a distinction between research, diagnosis, and treatment, many of its surrounding activities – such as the publication of journals or the issuing of bills to health-insurance companies – can be read as a form of empty institutional isomorphism building on DiMaggio and Powell’s original analysis (DiMaggio and Powell Reference DiMaggio and Powell1983). Indeed, psychoanalysis raises the possibility of an ensemble of phantom limbs without a body to which they truly belong. Let us now look in more detail at the setting of psychoanalysis and how it relates to each potential form of practice.

Research: the couch as laboratory?

The sciences have invented the laboratory as a specific place to do research (Krohn and Weyer Reference Krohn and Weyer1989; Kohler 2008; Guggenheim Reference Guggenheim2012), the central feature of which consists in taking the object of knowledge out of place and time (Kohler 2008). This de-contextualising distinguishes the “laboratory” from the “field”, where objects are studied in their “natural environment”. Conveniently, when material is taken to the laboratory, the researcher does not have to wait for a phenomenon of interest to “just happen”; he or she can manipulate the research object under controlled conditions (Knorr-Cetina Reference Knorr-Cetina and Pickering1992). A second, but related, feature of the research laboratory is its insulation from the environment. The laboratory is designed to contain the consequences of the actions inside in order to allow “action without consequences” or “trial action” for the researcher (Krohn and Weyer Reference Krohn and Weyer1989).

Psychoanalysis – like laboratory science but unlike philosophy, literature, economics or stratification research, for example – also has a theory of how a specific place is essential to the production of its knowledge. Indeed, Freud quite consciously sets out to construct the psychoanalytic setting in analogy to the chemistry lab (Karin Knorr-Cetina Reference Knorr-Cetina and Pickering1992; De Swaan Reference De Swaan1980). The psychoanalytic setting as a technology combines space, relationship and analyst into an instrument to produce a unique kind of knowledge.

Psychoanalysis builds on a mechanism of placelessness or decontextualization similar to that of the scientific laboratory: The psychoanalytic setting also takes the patient out of the context of his normal life and thus creates a distinction between "analysis" and "life" in analogy to the distinction between "laboratory" and "field". In the words of Abram de Swaan the setting creates a “social null situation” (De Swaan Reference De Swaan1980: 389-408). Psychoanalytic practice occurs in a place that has no other role in the life of the patient, and the relationship with the analyst is carefully separated from other dimensions of the patient's life. Prohibitions on personal relationships even after treatment reinforce the insulation of the setting. Patient and analyst are not supposed to make arrangements to meet outside the setting and there are awkward provisions about what to do if the patient and analyst meet by chance.

Psychoanalytic practice thus takes the patient out of the place and time of his biography. The psychoanalytic relationship, produced in its intensity through the setting, leads the patient to reproduce interpretative and behavioural patterns that are central to his or her psychic structure; because of this, the analyst does not have to travel back to the patient’s childhood or wait for a pattern to manifest in the patient's present life outside the setting in order to analyse it. He or she can analyse “scenes” that take places within the setting (Lorenzer Reference Lorenzer1973).

Also, in some similarity to the laboratory, the psychoanalyst, as a kind of instrument, is standardised through training and continues to be standardised through supervision. The development of specific feelings towards patients is carefully controlled and reflected upon in the form of a discourse on counter-transference (Freud Reference Freud1958c: 2343). While some later psychoanalysts try to make counter-transference useful for knowledge-production in the therapeutic setting (e.g. Racker Reference Racker1968), the distinction between personal counter-transference and diagnostic counter-transference maintains the effort to avoid pollution of the professional by the subjective and the personal. In these efforts to control for the observer's individuality, psychoanalysts go much further than, for example, ethnographers do in sociology or anthropology. Ethnographers might write and talk about how their personality shapes what they find but do not generally subject themselves to a process of standardisation that involves as much ongoing commitment and that is as methodical as psychoanalytic supervision.^{Footnote 2}

But while psychoanalysis shares with laboratory science the attempt to stabilise its setting, it does not exploit this stabilisation in the same way. In the scientific laboratory, context is excluded but much effort is put into adjusting the laboratory depending on which relevant context one wants to examine or hold constant. The laboratory is not a given and stable apparatus; rather it is adjusted from experiment to experiment. In contrast, psychoanalysis insists that the place of analysis is not arbitrary but highly specific, but it does not adjust the setting to different objects of knowledge or to the different questions it seeks to explore. Psychoanalysis standardises its setting in the same way for all conditions and for all patients. Observers are also all stabilised in the same way, independently of the kind of patients they treat.

The setting is only adapted for those patients deemed “too ill” for psychoanalytic treatment proper. According to Kurt Eissler’s famous essay on psychoanalytic technique, it can be appropriate, for example, to have the patient face the analyst instead of lying on the couch with one's back to the analyst. However, this is not recognized as a legitimate form of variation but rather as a necessary deviation; the aim in these cases is to gradually progress towards the standard setting (Eissler Reference Eissler1953).

The laboratory in the natural sciences and psychology produces contextlessness in order to generalise about specific conditions and factors. The procedures of psychoanalysis produce contextlessness but do not use this contextlessness to generalise beyond the individual patient. The importance of the case study in psychoanalysis has been much remarked upon and there are those in other disciplines who look to psychoanalysis as a model for justifying qualitative and interpretative forms of inquiry (Forrester Reference Forrester1996; Creager, Lunbeck, and Wise Reference Creager, Lunbeck and Wise2007). However, it is important to note that, in contrast to the use of case studies in medicine or sociology, psychoanalysis uses the singularity of the case to show the universality of the theory. It does not develop intermediate categories, such as for groups of patients or conditions. Sociologists are trained to ask what a case is a case of (e.g. Ragin and Becker Reference Ragin and Becker1992). Psychoanalysts are reluctant to embrace new categories that could provide answers to this question and are ambivalent towards formally recognizing those categories identified by Freud as a basis for studying variation.

Analysts may adjust to individual patients in the course of analysis but they do not adjust to groups of patients. Orthodox psychoanalysis tends not to be interested in cultural variations but it has not even adapted the setting for those variations which it itself does consider relevant. It has not developed a theory, for example, as to whether patients with intrusive mothers have to be studied in a specific way, or whether the couch is better suited for neuroses or for anxiety. Even in cases where there were debates about the technology of psychoanalysis, as for instance in the debate between Lacan and other French psychoanalysts about the duration of a meeting, this was not a debate about different groups of patients. Rather, as Langlitz (2005) points out, Lacan wanted to keep the duration flexible, in deviation from the standard 50 minutes duration, to allow psychoanalysis to move away from a concept of therapy towards an idea of the temporality of the subject and of memory, based on structuralism, cybernetics and linguistics.^{Footnote 3} The flexibilisation of the meetings led to a variability, which was not controlled according to the patient but according to the inner logic of analysis. Even the analysts who were interested in cultural context and historical variation – such as Horney, Fromm, and Sullivan (see McLaughlin Reference McLaughlin1998) – did not adjust the setting of knowledge production.

Psychoanalysts gesture towards research through conferences and journals that report on ongoing work and findings. However, most scholarly papers proceed in such a way that they use the stability of the couch to apply the stability of the theory to the unstable single case (or a few cases). If they do not find a confirmation of the theory, they tend to ascribe the deviation to the unique case. It is rarely officially acknowledged within psychoanalysis that one might find in the conversation with the patient a piece of data that one could then take to other contexts and slowly harden to form a new theory. Rather theoretical knowledge always already exists as a concept (e.g. as a neurosis, say the Oedipal complex) and is then newly adapted for each specific case. New categories for groups of patients often emerge from concerns outside of psychoanalysis, as in the case of anorexia, but these categories are then related back to the existing Freudian categories. Psychoanalysis has not found it necessary to adapt its theoretical apparatus for anorexia. In that sense, psychoanalytic theory is extremely resistant to irritation by evidence.

Psychoanalysis does not design the setting of knowledge-production with a view to generalisation beyond the individual patient, but that does not mean that psychoanalysts do not generalize – generalization in claim-making happens, and it happens mostly via the texts of Freud, who had already taken the step towards generalisation for all future psychoanalysis. Every generalisation in psychoanalysis arrives back at Freud – if it does not arrive back at Freud, it is no longer psychoanalysis.

Freud himself established this pattern of arbitrarily excluding deviation from the masters’ theory, rather than accepting multiplicity or change. In “New Introductory Lessons on Psychoanalysis” he writes: “When the differences of opinion had gone beyond a certain point, the most sensible thing was to part and thereafter to proceed along our different ways – especially when the theoretical divergence involved a change in practical procedure” (Freud Reference Freud1958b: 4793). Freud famously went on to say that such different ideas “may be a school of wisdom” but are “no longer analysis” (4793).

When forms of generalization that do not proceed via Freud, such as those based on either conditions of origin (dominant fathers, clingy mothers) or symptoms (narcissism) are hinted at, this allows for a starting point from which to differentiate research and treatment. However, psychoanalysis will not explicitly endorse movement in this direction when it comes into conflict with Freud, and will refuse to be held accountable on the ground of research or treatment. Indeed, psychoanalysts openly dispute the idea that something like research is possible in psychoanalysis (see for example Green Reference Green, Sandler, Davies and Sandler2000; Kernberg Reference Kernberg2006; Wallerstein Reference Wallerstein2009).

Diagnosis: the couch as stethoscope?

In professions, diagnosis is the task of cognizing the unique case in relation to the general with a view to intervention. One might say, following Abbott, that diagnosis is what distinguishes professions, such as medicine or law, from academic disciplines such as economics or chemistry (Abbott Reference Abbott1988). Psychoanalysis implies the need for diagnosis. It claims it is important to learn about the individual person in relation to general hypotheses about psychic structures. Indeed, some psychoanalysts respond to questions about the status of psychoanalytic knowledge by insisting that it is not a scientific discipline but rather a profession (e.g. Buchholz Reference Buchholz2000, Reference Buchholz2006).

The features of the psychoanalytic setting are central to the way diagnosis of individual cases is conceived. We have said that the psychoanalytic setting takes the patient out of the place and time of his or her biography, leading the patient to reproduce interpretative and behavioural patterns that are central to his or her psychic structure. Psychoanalysis does not claim that it can generalise from one patient to other patients, but it does claim that it can generalise from behaviour in the psychoanalytic setting to the psychic structure of the patient in general.

For psychoanalytic diagnosis, the patient comes to the analyst and not vice versa (see De Swaan Reference De Swaan1980).^{Footnote 4} This centralisation of space presupposes that the object of diagnosis is mobile. Psychic disturbances or the psyche in general have to be independent of location. If symptoms would not appear, or if they were different on the couch, psychoanalysis would have a problem (this is where Nathan's critique takes its starting point, see Nathan and Stengers Reference Nathan and Stengers1995). Psychoanalysts do not usually compare behaviour outside the practice with behaviour inside. They do not collect their own data about what the patients’ mothers or partners are like.

This has parallels in the history of the professions: Rudolf Stichweh has pointed out that historically the process of professionalization goes hand in hand with a stabilisation of the setting in which the profession is practised (Stichweh Reference Stichweh and Stichweh1994). The medical doctor provides a particularly striking example. In the past, the general practitioner might have spent a large part of his time on route to his patients (for an illustration see Gabriel Garcia Marquez’ Love in the Time of Cholera, 2005). Doctors’ professionalization is based on the establishment of a surgery, to which his patients come. Through this, he can spend more time seeing greater numbers of patients and his knowledge increases. This also allows for the delegation of lesser tasks to non-professionals, who are also stationary. It is instructive to compare this to diagnosis in social work. Here, a problem is diagnosed, which is assumed to be situated and connected to a given context and comprised of a multiplicity of people and objects. This is usually done “in the field”, where problems originally occur, rather than in a space constructed by the profession. It is part of the habitus as well as the epistemology of social work that social problems are inherently linked to the place in which they emerge and that therefore the social worker has to move about. Diagnosis in the laboratory or on the couch runs the risk of not understanding a problem because it ignores its context. In contrast, diagnosis in the field runs the risk of not being able to attribute a problem correctly, and thus not being able to find an angle for treatment.^{Footnote 5}

As long as psychoanalysts limit themselves to claims about the psychic structure as revealed in the psychoanalytic encounter, the specificity of the location of knowledge might not be a problem. Indeed, psychoanalysts have developed two responses to the particular constraints and limits of knowledge generated “on the inside” (see also Gellner Reference Gellner1993: 179-199): one is a long tradition of radical constructivism and reflection on language. In this tradition, it is not important to find out “what really happened”; rather the aim is to construct and reconstruct a narrative that works for the patient. The other strategy is the denial of the importance of external reality – as opposed to internal fantasies or drives. In this tradition, it does not matter in principle what happened to the patient in his childhood – all behaviours can be interpreted in relation to universal facts about drives or fantasies. The controversy between the Kleinians and John Bowlby is instructive in this regard (van Dijken et al., 1998). Bowlby was interested in the effect periods of long separation from their parents had on young children, and was shocked to find the Kleinian establishment was very explicitly not interested in the actual experiences of young children but was only interested in fantasies claimed to be shared by all people according to general psychoanalytic theories.

The unique construction of the setting does become a problem when psychoanalytic therapy implies a simple application of its truth to the world outside the setting or leads to consequences outside it. The cases of fictional child abuse that have made the headlines since the 1980s is revealing here (Hacking 1992).^{Footnote 6} In this case it emerged that the protected space of the laboratory-like setting produced data that were only the result of the setting. Without psychoanalysis patients would not have “discovered” the abuse. In terms of the psychoanalytic process this does not necessarily matter. But in this case the psychoanalytic setting, insulated against the outside like the scientific laboratory to enable trial action, was not insulated enough and produced huge consequences outside its context.

Psychoanalysts gesture towards diagnosis in communications with health insurance providers. But they are somewhat resistant towards recognizing diagnosis as a discrete task (Shedler 2002). Psychoanalysts are famously reluctant to tell patients “what they have”; orthodox psychoanalysts are highly critical of the labels imposed by the DSM for example (for a typical critique see Galatzer-Levy and Galatzer-Levy (2007), see also Whooley (2010b)). Psychoanalysis in its orthodox form goes further than that: beyond a set number of neuroses identified in Freud’s writing psychoanalysts reject all newer labels such as anorexia or addiction and do not produce their own. With regard to the diagnosis too, psychoanalysts do not recognize anything between the individual case and the universal properties of psychic structure identified by Freud.

Treatment: the couch as a pill?

While Freud tried to construct the psychoanalytic setting in analogy to the chemistry lab, he also frequently compared psychoanalytic treatment to a medical surgery (Knorr-Cetina Reference Knorr-Cetina and Pickering1992). Freud often claims to work towards a “cure”, and today many psychoanalysts see psychoanalysis as a form of healing.

The psychoanalytic setting is again claimed to be essential to the way treatment is conceived. We have observed that the psychoanalytic setting is insulated from its environment, in some similarity to the laboratory in the natural sciences. For the researcher in the natural sciences, this insulation of the laboratory allows for “trial action”, or action without consequences (Krohn and Weyer Reference Krohn and Weyer1989). The laboratory is a closely observed and controlled space, which allows researchers to do things the effects of which are uncertain. “Trial action” takes a specific form in psychoanalysis: while the research laboratory constructs what the researcher does inside as a trial, what the analyst does in the practice is real; it is not a test. The focus is not on an implied benefit for science, or for other, future patients, as it is in research, but rather the focus is on benefits for this patient.

In psychoanalysis, the insulation of the setting allows trial action, not for the researcher but for the patient. The patient can say things which he or she does not usually feel able to say. He can, for example, share that he sometimes wants to kill his boss or his partner. Psychoanalysis relies on the hypothesis that this kind of trial action reduces the pain that is at the origin of the fantasy and goes some way towards relieving the patient from the need to do so in reality. The patient can try out actions, such as expressing anger or love in a safe setting. This type of trial action is important in order to destabilise expectations that have their origins in the parent-child relationship.

Like other forms of professional treatment, psychoanalysis distances itself in strong terms from the diffuse expectations of “interpersonal helping” associated for example with friends or family members (see Parsons Reference Parsons1939). Psychoanalysis stands out, however, by insisting that treatment happens in the same practice and the same space as research and diagnosis. Psychoanalysis also goes further than other forms of professional treatment in its insistence that treatment is internal to the professional setting. The treatment is in its own space, not the patient’s. Psychoanalysis insists that its treatment happens on the premises, yet it then assumes that it travels with the patient.

Some forms of medical treatment also happen on the premises, but many others can be taken home, such as, most notably, drugs.^{Footnote 7} Most professions project outside the encounter between professional and client by giving out some form of advice. Social work, in addition to providing advice, also provides access to welfare benefits, restraining orders, and other services. Lawyers provide letters, petitions and submissions to court. Cognitive-behavioural therapy encourages the development of strategies for implementation outside the space of treatment. One might be given concrete instructions as to what to think or do when feeling the onset of a panic attack. Psychoanalysts famously resist giving out advice.

Psychoanalysts invoke claims about curing patients when they issue bills to individual patients, and especially when they issue bills involving reimbursement from health insurance companies. As in psychoanalytic research and diagnosis, there has been no differentiation of treatments according to different kinds of conditions patients may have. An analyst might adapt to individual patients, but there are again neither categories nor strategies that would formalise the conversation between the generalities of the theory and the specificities of individual cases.

A lack of specialization as treatment or healing is also evident in a relative lack of interest among psychoanalysts in verifying outside the practice whether it is effective. There are different ways in which professions and organizations think about their own success or failure. Companies measure success by whether they sell something. Government organizations often measure success by whether money has been spent, or by ex-post evaluations. Cognitive behavioural therapy observes itself through large-n correlational research, like much of contemporary medicine and pharmaceutical research.

Psychoanalysis does not tend to ask itself questions about its own effectiveness. Recent attempts to measure the effectiveness of different forms of therapy have often been denounced by psychoanalysts. For example, when Klaus Grawe attempted to compare the effectiveness of various forms of therapy in a large metastudy (Grawe Reference Grawe, Donati and Bernauer1994), psychoanalysts not only answered that his criteria failed to capture the meaning of psychoanalysis, but also that such attempts at measuring effectiveness are futile in principle (Mertens Reference Mertens1994, Tschuschke Reference Tschuschke1997). While psychoanalysts often critique each other, the charge is rarely that a competing approach does not work, but rather that it misunderstood Freud, or the human condition.

Psychoanalysts invoke claims about cure, and will privately say they want to “help” their patients. But they never fully embrace that task: treatment is a phantom limb which only exists under certain circumstances and for certain rhetorical purposes. In some versions of self-justification, psychoanalysis is all about diagnosis – a custom-made diagnosis, where the result is not being better but knowing more about oneself (see Bader Reference Bader1994).

What psychoanalysis does sometimes do is provide evidence for efficacy in the form of success stories (e.g. Yalom and Elkin Reference Yalom and Elkin1991; Yalom Reference Yalom1989). Case histories of patients illustrate the progress a patient has made over the course of sessions with the therapist. This is to some extent similar to what state organizations and politicians do; in development aid, for example, stories circulate about this or that intervention with evidence about how the lives of beneficiaries have changed: it might be said for example, that children in a certain village did not have a school, but now they do (see Watkins, Swidler, Hannan 2012). Even within the genre of “success story”, psychoanalysts are unusual in that the evidence is taken entirely from within the psychoanalytic setting; psychoanalysts do not usually go to check if they have had an effect outside the practice.

On the divisions in psychoanalysis

Much work in the sociology of science (Gieryn Reference Gieryn1999) and the sociology of fields of cultural production points to the importance of boundaries and symbolic distinctions (Bourdieu Reference Bourdieu1995, see also Lamont and Molnar Reference Lamont and Molnar2002). The case of psychoanalysis raises the question as to how we might explain variation in the forms divisions take in different fields. As Freud himself already noted, psychoanalysis displays an extreme tendency toward division into competing schools, which accuse each other of betrayal (Freud Reference Freud1958b: 4793).

A number of explanations have been offered for this feature of psychoanalysis: participants in these symbolic struggles tend to explain these divisions with reference to the theoretical or political “mistakes” of their adversaries. Psychoanalysts are also, unsurprisingly perhaps, inclined to read divisions in psychoanalytic terms: their prevalent mode of self-observation is itself psychoanalytic – psychoanalysts attribute narcissistic injuries to each other, diagnose parent-child relationships and so on (see Malcolm Reference Malcolm1981). In an interesting account by an insider, Stephen Frosh attributes the prevalence of schools and divisions to properties of the object of psychoanalysis, the unconscious, which is “too hot to handle” (Frosh Reference Frosh1997: 232).

Sociologists might try to explain these divisions as a result of institutional factors external to the production of knowledge itself. In contrast, our analysis suggests that the intensity of the divisions in psychoanalysis can be explained as a result of the insistence on the centrality and undifferentiated role of the psychoanalytic setting.

The comparison both to other disciplines and to the professions is instructive: disciplines often develop a division of labour according to a focus on different topics of research; professions differentiate peacefully according to specializations focused on different problems or treatments. Psychoanalysis’ insistence on the unity of research, diagnosis and treatment closes off both of these paths: because psychoanalysis does not acknowledge research as a distinct practice, it cannot peacefully differentiate according to different topics or theoretical approaches. Every challenging finding tends to lead to a new fraction of psychoanalysts, not a reformulation of theory. The only exception to this is feminism, which is a legitimate object of contestation within most schools of psychoanalysis because gender is a distinction inherent to Freud's original theory.

Because psychoanalysis does not acknowledge diagnosis as a task in its own right it cannot peacefully differentiate according to different types of illness. There is some differentiation according to types of patient (children, couples and organisations), but psychoanalysts view these forms of specialisation with ambivalence and as a form of selling out to external concerns. Because psychoanalysis does not formalise itself as treatment, it cannot peacefully differentiate according to different courses of action, or more or less successful forms of treatment. Thus each version which claims to do something different or better ends up being a different form of therapy, rather than a specialized form of therapy within psychoanalysis.

In the terms of Lakatos’ philosophy of science, psychoanalysis is not a program of research that changes continuously; rather due to its theoretical and technical rigidity and the continued production of new unique cases, it continues to produce seemingly incommensurable paradigms. These emerge as a caricature of Kuhn’s paradigms (Kuhn Reference Kuhn1996). Paradigms are not triggered by incompatible views of an object, which then trigger social differentiation, as Kuhn describes it. Rather paradigms follow from a social differentiation, which can then be criticised as ideology because of the lack in technical differentiation.

It is instructive to compare this to the social sciences. For example, there is also some research following Foucault that merely elaborates and confirms some aspect of Foucault’s work. The researcher points at a single case and generalises via Foucault by calling it an instance of “governmentality”. The same can be said of some research following Bourdieu, Actor-Network-Theory, systems theory or any other strong theoretical tradition. In sociology, however, this form of generalisation via school founders is supplemented by generalisation via object (education, organisations), methods (ethnomethodology, network analysis) or problem (crime, inequality). These different approaches are usually tolerated within the discipline itself, rather than excluded on the ground that they use a different mode of generalisation.