In a Word - Transference (with a dash of neuroscience)

Dr. O'Leary discusses the term Transference, and if you listen until the end, he relates it to some computational neuroscience.   Transference is a historically loaded term.  Transference is supposed to be an unconscious process, so it can not really be observed, only inferred, so this means that both the definition of transference and any instance of it in psychotherapy is dependent upon the therapist’s model and their particular way of interpreting that model.  But what exactly is this elusive but data rich unconscious process?

Please leave feedback at https://www.psydactic.com.

References and readings (when available) are posted at the end of each episode transcript, located at psydactic.buzzsprout.com. All opinions expressed in this podcast are exclusively those of the person speaking and should not be confused with the opinions of anyone else. We reserve the right to be wrong. Nothing in this podcast should be treated as individual medical advice.

Transcript

In a Word - Transference

Welcome to PsyDactic! Today is August 17, 2024 and I am Dr. O’Leary a Child and Adolescent Psychiatry Fellow in the National Capital Region.  This is a podcast about psychiatry and neuroscience.  I make this as a way for me to have an excuse to nerd out and look more deeply into topics that I love to think about.  By synthesizing it and recording it, I get to share this with you, for better or for worse.  Everything in this episode is my opinion and was curated by me with all of the biases inherent in the way I think.  It is not medical advice, and it is certainly not the opinion of any institution I work for or any person who happens to be standing next to me on the Metro.

Today I am going to discuss the term Transference, and if you listen until the end, I am going to relate it to some computational neuroscience.  Woo Hoo!  Transference is probably the most historically loaded term I have dealt with to-date.  Transference is supposed to be an unconscious process, so it can not really be observed, only inferred, so this means that both the definition of transference and any instance of it in psychotherapy is dependent upon the therapist’s model and their particular way of interpreting that model.  But what exactly is this elusive but data rich unconscious process?

For some it is very broadly defined as the way that the patient relates to anyone based on their past experiences and relationships.  If you are Carl Jung you might include the collective unconscious as a player.  In a more psychoanalytic approach, transference may be more strictly defined as, for example, the unconscious fantasies that a patient generates about their therapist.  It might also be defined as the transfer onto the therapist of an object relationship that the patient has previously developed.  The most primary object is a parental object, and this may dove-tail with more Freudian concepts like the Oedipus complex.  Transference may also be defined as what the patient assumes that the therapist thinks or feels, based on whatever their psychopathology is.

Psychoanalytic and psychodynamic therapists use, as a primary way to evaluate the patient, their inferences of the patient transference.  While other therapies recognize in general that a patient’s past can creep into the therapeutic relationship, most don’t rely on identifying this transference.  A paper called, “The Concept of Transference” by Thomas Szasz in the International Journal of Psychoanalysis reports on the many different definitions of transference available in 1963.

He quotes Karl Minninger who wrote in 1958:
“I define transference … as the unrealistic roles or identities unconsciously ascribed to a therapist by a patient in the regression of the psych-analytic treatment and the patient’s reactions to this representation derived from earlier experience.”

Szasz also quotes Ida Macalpine from 1950:
“Transference may be said to be an attempt
of the patient to revive and re-enact, in the
analytic situation and in relation to the analyst,
situations and phantasies of his childhood.
Hence transference is a regressive process.”

Szasz quotes an even earlier writer, Gover from 1939:
“As the transference develops, feelings
originally associated with parental figures are
displaced to the analyst, and the analytic
situation is reacted to as an infantile one.”

All three of these analytical definitions of transference rely on another concept called regression, and state that transference can be identified by the nature of the regression toward an infantile state that a patient experiences in psychotherapy.

Szasz also explores transference as something that may transcend the therapeutic relationship.  He explores how some analysts have proposed that transference cannot be used to account for some (or possibly even most of) the affective states that a patient experiences in therapy.

For example he quotes Finichel from 1941:
“Not everything is transference that is experienced
by a patient in the form of affects
and impulses during the course of an analytic
treatment. If the analysis appears to make no
progress, the patient has, in my opinion, the
right to be angry, and his anger need not be a
transference from childhood-or rather, we
will not succeed in demonstrating the transference
component in it.”

Aside from relying on regression, transference as a means of formulating a patient uses a Bayes Optimal approach that fits the data to a model.   Dr. O’Leary, now you are talking nonsense.  What is a Bayes optimal approach?  Let me summarize like this:  When we get new data, we can do two primary things with it, we can either change how we experience the world or we can change what we think about the world.  A mathematical way to say this is that we can change our model to fit the data or we can change the data to fit the model.

When we have tightly defined and rigid models, then the easiest thing to do is to change the data to fit the model and the easiest way to do this is to gather a lot of data and then pick out the patterns in that data that appear to fit our model.  This is a pseudoscientific process because it appears that our models are well supported by data, but we are only considering the data that supports our model, and ignoring the rest.  We are in effect changing the world to fit our model.

Maybe an easier way to conceptualize this is to imagine that you or your family just bought a used Toyota Prius.  Suddenly, you are seeing Priuses everywhere you go, and you may be tempted to think that this was not the case in the past.  What has happened is NOT that there are suddenly more Priuses on the road.  What has more likely happened is that now the way you are sampling the world has changed.  But how does this relate to psychotherapy and transference?

Let’s assume my model states that the primary transference in a therapeutic relationship between a female patient and male psychotherapist (myself in this example) is either one of sexual desire OR envy for the power that the therapist holds in the relationship.  The primary love object of my female patient is the patient's father, who she cannot have, due to society's restrictions, so she must transfer that desire onto someone else in order to have it be fulfilled.  I am an adequate substitute, though similarly elusive.  A sexual relationship between provider and patient is forbidden.  This recapitulates the relationship with her father.  The vehicle of this transference is the patient's unconscious fantasy of sexual intercourse with me, which at some point during therapy will likely percolate to the surface and ultimately come to conscious awareness.  Long before the patient realizes this, I (as the therapist) should have already inferred its existence.  At the point the patient reveals this realization, I will be confirmed that my inference of transference was true.  I will see a Prius where I might not have before.

For those of you who either already knew or who have quickly figured out that transference is not necessarily a one-way street, the complement to transference is countertransference, which is a similar phenomenon, though it is supposed to be based on the therapist's unconscious fantasies regarding the patient.  I may desire the patient and have fantasies of breaking free from society's restrictions and eloping with her.  I might, alternatively, view the patient as a daughter, and may unconsciously adjust my actions to protect and nurture her, instead of to challenge her and expose her to opportunities to gain power either in the therapeutic relationship or in the world at large.

In the past, the roots of transference were proposed to be highly deterministic and a particular type of transference had a strict relationship with the kind of disorder a patient presented with.  For example,  Freud noticed that many of his female patients, especially those determined to be hysterical, reported having been raped by family members, and because so many of them had similarly disturbing stories, he rejected the idea that any of them were actually sexually molested.  They must be inventing these tails, and therefore these sexual fantasies are the cause of their hysteria.  He later waffled on this idea, proposing that some kind of actual sexual activity may have occurred, but instead of molestation, it was masterbation.  Patients are causing their own hysterical or neurotic reactions by failure to suppress their sexual urges and giving in to fantasy.  The sexual fantasies of the patient toward the therapist serve to prove this.  He revised his ideas many times over his life, but the primacy of using transference to determine the cause of neurosis remained central to the psychoanalytic method.

In the more modern age, loosey goosey definitions of transference and countertransference are used.  In general, those without intense formal training in psychoanalysis or psychodynamics (like me) use the word transference to mean the baggage that a patient brings with them to therapy that affects how they relate to the therapist.

I like a quote I found in a paper from 2007 in the American Journal of Psychotherapy, and I think it does a better job than I can in explaining this.  This is from The role of clinical inference in psychoanalytic case formulation published in 2007 by David Wolitsky.

“In the current age of theoretical pluralism we can no longer speak of the
theory of psychoanalysis, we have increasingly accepted the notion that
when formulating a case, the clinician creates a narrative structure influenced
by a preferred theoretical orientation. This structure is an attempt to
provide a coherent, comprehensive, plausible, and hopefully accurate
account of the individual's personality and current functioning based on
his or her life history, as that history is told, lived, and retold by the patient
(and analyst) in the course of the psychoanalytic encounter. As Schafer
(1992) has stressed, there is no single, definitive, unchanging, final narrative
that can qualify as the correct understanding of the patient's psychic
life. Implicit in this view is that what we call "facts" in psychoanalysis are
not pure observations, but inferences that attribute meanings and motives
to patients' behaviors and experiences. What is observed (i.e., what is
attended to selectively), stored, and retrieved from memory to arrive at
particular clinical inferences is influenced by the nature of the patient therapist
interaction, including the therapist's countertransference reactions
and working model of the patient, which is active and changes as new
material is produced. It is inevitable that what the therapist selects for
observation, and the inferences drawn from the observations, are influenced
by theory. Long ago it was observed by Marmor (1974) that
Freudian patients have Freudian dreams and Jungian patients have Jungian
dreams. The influence of theory on clinical inference has been demonstrated
empirically (Fine & Fine, 1990).”

This is a nice way of saying that the way that psychoanalysts interpret the transference in a case is entirely arbitrary, not in the colloquial meaning of arbitrary, which to most means randomly generated, but in the sense that the arbiter (who is the therapist) determines which preconceived models to apply.  While this is a non-random process, the model was not selected from all possible models based on the data.  In this way, data does not generate the hypothesis, but is combed through for evidence that appears to fit an already assumed set of possible explanations.  Whatever convenient model fits the data is selected, whether it is a Prius or a Camry.  This is not a passive process.  The data that we get, we get because we have to some degree forced that data into existence and then forced to fit a predetermined model because the data that we collect is collected only because it reminds us of a model.  Everything else is considered unimportant random noise.

It may be tempting to say that an eclectic approach, which allows for model switching, is a method that fits the model to the data, instead of fitting the data to the model, but all it does is provide a larger number of arbitrary narratives.  Therapy assumes at the outset that the patient starts either lacking insight into their problems, or having deceived themselves as to the source of the problem.  All of the various therapeutic frameworks seek to disabuse the patient of their ignorance or self-deception, and the dynamic and analytic ones assume that the formulation determined by the therapist who is attuned to transference-countertransference is superior to that of the patient.  None of these methods derive their explanatory narrative from the data itself, but instead, find patterns in the patient’s narrative that fit the story that the therapist wants to tell back to the patient.  The transference they identify is the transference they are looking for.  When we see details that fit a familiar story, we focus on these.


To be fair, it would be extremely resource intense to generate a new model for every new patient.  We need some kinds of predetermined models to guide our work.  Psychotherapy happens to be flush with competing models.  Confirming out own suspicions is also a very satisfying and attractive way to think, and I know I think like this very frequently as a clinician, because it is convenient and helps me make decisions.  I can see that the patient is regressing into a more childlike state and relying on me as their parent to prescribe an anxiolytic, because I, too, am distressed by their anxious presentation.  It is an easy thinking trap to fall into, and in some cases has some level of validity.   Any good story that promotes adaptive behaviors will help the patient to exert more control over their lives or to release the pathological amount of control that they desperately cling to.

One of the ways that many modern physicians deal with the plethora of theoretical frameworks is by not necessarily even attempting to explain why any particular patient is anxious or depressed or obsessive, etc, but merely treating the patient because they are anxious or depressed or obsessive or whatever.  You are depressed, here is an antidepressant and a round of Cognitive Behavioral Therapy.  If you do both you’ll have about a 40 to 60% chance of feeling significantly better than by just waiting for it to go away.  It is an impersonal approach, but it is arguably efficient. There is no need for 104 weekly sessions to wait for the patient to reveal their life themes and transference.  In the more modern perspective identifying an etiology is not necessary for their treatment.  We assume a number of other kinds of models, such as a chemical imbalance or a trauma reaction or a genetic predisposition and use these as justifications for what we do.

The worst outcome of this is not that the therapist will make poorly evidenced conclusions; it is that our methods systematically blind us to all other possible interpretations of the data and also blind us to the data itself, because we only see what we are trained to see.  Classical psychoanalytic and psychodynamic approaches rely on a relatively deterministic model of human cognition and behavior where we can draw strong causal relationships through time and weave a narrative that is so satisfying, we don’t need to look anywhere else.  They are not the only ones.  Those of use preaching chemical imbalances are doing the same thing.

Modern psychodynamic approaches are less attached to a particular theoretical framework and try not to overfit the data to a single model.  However, our overall approach in any psychotherapy has a limited number of models that data can be fit to.  Transference, how we assume the patient relates to the world, is often the key to determining which of those models we chose to explain the patient. 

Before I stop blabbering today, I want to relate this to some neuroscience.

Most of our therapies are, to some degree, an expression of neuroscience as it was understood at the time.  Ideas like neurasthenia, or exhaustion of the nerves, had a great impact on Freud, and probably influenced why he identified things like masterbation as pathogenic.  So did his rudimentary understanding of evolutionary biology at the time, which supposed that humans shared with animals basic drives that are necessarily dysfunctional in psychopathology.

Until recently, much of neuroscience merely assumed that the different functions of the brain are anatomically distinct from each other, for example emotions are a function of the limbic system, anxiety particularly of the amygdala, and cognition of the cortex, etc, but the most advanced data does not support this at all.  Some have even called this a modern form of phrenology, which is a debunked form of psychological inference where doctors would claim to be able to determine personality traits of individuals by their skull anatomy.

Neuroscientists are showing more and more that degeneracy of neurological processes is the norm.  What I mean by this is that our expressions of anger, contentment, anxiety, joy, etc are not generated in the same way over time.  The brain patterns I have that supposedly are the fingerprint of happiness on an MRI are hard to reproduce, even within a single person and especially divergent among separate individuals.  What causes most complex experiential phenomena is not one thing.  It is degenerate.  It is a multitude of things that are subjectively interpreted as the same thing.

I highly encourage you to check out the Brain Science Podcast by Dr. Ginger Campbell, especially episode 135 with Dr. Lisa Barret who describes how she was forced to give up the idea that there are reliable neurological tracings for how emotions are generated in the brain.  This may be because there are an infinite number of Bayes optimal ways to produce a belief that maximizes free energy, and emotions at their most fundamental level are beliefs about the state of the world.

Holy schmagolly, Dr. O.  What is happening here?  I hope I have piqued your interest.  

Instead of doing a poor job of explaining how Bayes optimal solutions relate to psychotherapy, I am going to provide you links to two episodes of the podcast Doorknob Comments.  The first is an interview with the neurophysicist Dr. Karl Friston, who is one of the main developers and proponents of active inference.  The second is an interview with Dr. Alex Tolchinsky where he discusses the Narrative Fallacy.  Dr. Tolchinsky was also on this podcast a few episodes ago talking about the same thing.

In the future, you will hear me come back to the concept of active inference, Bayes optimization, and the free energy principle frequently.  These principles, I believe, unify neuroscience and cognitive philosophies and can even provide more fundamental ways to understand sciences like evolutionary biology by modeling the common thread in how various processes of Natural Selection work.  Bayes optimization can even help explain why patients act the way that they do without making any assumptions about their daddy issues.

Thank you for listening.  I am Dr. O’Leary, and this has been an episode of PsyDactic.





Wolitzky DL. The role of clinical inference in psychoanalytic case formulation. Am J Psychother. 2007;61(1):17-36. doi: 10.1176/appi.psychotherapy.2007.61.1.17. PMID: 17503675.

SZASZ TS. THE CONCEPT OF TRANSFERENCE. Int J Psychoanal. 1963 Oct;44:432-43. PMID: 14084096.

Macalpine, I. (1950). The development of the transference. The Psychoanalytic Quarterly, 19, 501–539.

https://brainsciencepodcast.libsyn.com/podcast/bs-199-batja-mesquita-on-cultural-origins-of-emotion

https://brainsciencepodcast.libsyn.com/podcast/bs-135-lisa-barrett-on-how-emotions-are-made 

https://brainsciencepodcast.libsyn.com/podcast/bs-132-william-uttal-on-the-limitations-of-brain-imaging

https://www.doorknobcomments.com/episodes/xe5tgw2m7xrk5fh-emldc-ts9r9-pxcec-t9m9c 

https://www.doorknobcomments.com/episodes/xe5tgw2m7xrk5fh-emldc-ts9r9-pxcec-t9m9c-zw7af