Soren R. Ekstrom
Watertown, Massachusetts, USA
New York Association for Analytical Psychology
Due to discoveries in brain research over the last twenty to thirty years many of the formulations about unconscious mental life presented by Freud and Jung now seem validated by neurophysiological findings. (Solms & Turnbull, 2002) This development is in many ways encouraging for those of us in the helping professions. However, the new findings also challenge some of our established theories. What Henri Ellenberger (1970) calls “the discovery of the unconscious”, the idea that the human psyche has layers of unconscious components, rests on two components: inductions from patients in psychotherapy and the observations by analysts as participant observers. It does not rest on specific findings about how the brain participates in mental life.
Both Freud and Jung assumed that these neural manifestations of the unconscious could be inferred. In fact, Freud’s career began with his ambitious Project for a Scientific Psychology (1895), in which he attempted to establish a neurological explanation of psychology. (Sulloway, 1979) Most of his later understanding of the unconscious can be traced to this abandoned and much later discovered opus. Jung, on his part, made clear that the unconscious could only be deduced introspectively and proposed hypothetically and he suggested that a laboratory and objective methods would be needed to establish how things really are in an unconscious condition. (Jung, 1935)
The new definitions of how the mind works thus focus specifically on memory and learning and are based on findings that trillions of synapses, which make the brain’s cells and neural networks communicate, are modified by experience. This distinct characteristic, often called synaptic plasticity, may now be viewed as “innately determined”. (LeDoux, 2002, p. 9)
These findings also establish that much of what occurs in the human brain is beyond the immediate awareness of the person. In fact, the research in the various branches of cognitive science has led to what Campbell Perry and Jean-Roch Laurence (1984) describe as a rediscovery among researchers of the unconscious.
No such sense of rediscovery can be detected in the psychoanalytic community in regards to memory. Hardly mentioned by Freud and Jung, memory continues to be a neglected function, especially as it concerns the analyst. However, there are two areas in particular where findings in cognitive science offer ways to more fully understand the analyst’s contribution to successful treatments. The first has to do with research on the role narratives play in memory. This research, conducted in the fields of computational modeling and artificial intelligence, offers a strikingly different approach to how the understanding of others is developed and maintained. (Lakoff & Johnson, 1999)
The second area of research may give us a fresh and different way to understand countertransference reactions, an aspect of treatment which recent contributors to psychoanalysis have found more and more significant. (Gill, 1992; Ulanov, 1992) The notion of induced countertransference especially warrants a closer scrutiny in the light of the findings in cognitive science about problems with retrieval in states of high concentration and absorption. (Spiegel, 1995)
I have described some of this research in a previous paper, so here I will take the opportunity to further develop two models previously suggested. (Ekstrom, 2002) First, however, we need to define what we mean by “memory” and summarize what today is known about how it functions. We also need to take a look at the established approaches to the analyst’s contributions to the treatment process in most of the schools of analysis.
To Joseph LeDoux (2002), of New York University’s Center for Neural Sciences, the discovery of synaptic plasticity means that many psychological and behavioral functions are mediated by cells joined by synapses and working together. A fairly universal mechanism is, in other words, responsible for how most memory is formed. (LeDoux, 1998) The highly dynamic process of synapses being changed by experience can also be said to be responsible for our sense of self. Who we are, our individual self, is thus to LeDoux the result of what he calls “the particular patterns of synaptic connections in an individual’s brain and the information encoded by these connections”. (2002, p. 3)
LeDoux concurs, however, that much of how what happens to our experiences occurs without our explicit awareness. What makes us the individuals we are belongs mainly in this latter category. Distinction thus has to be made between two aspects of the self. Only the first, explicit aspect, involves self-awareness. (2002, p. 27) By contrast, the implicit self is “all other aspects of who we are that are not immediately available to consciousness, either because they are by nature inaccessible, or because they are accessible but not being accessed at the moment” (p. 27-28).
In practical terms, then, findings about synaptic change restate depth psychology in terms of how conscious and unconscious memory is being formed. Since the production of synaptic connections can be viewed as part of the process of encoding information into memory, what we learn from our experiences is stored for future references. But how well this storage is organized and retained depends on a variety of factors, not the least emotion or affect. As attention and arousal, emotions must be viewed as playing a central role in how certain experiences become dissociated and hidden, as it where, or reinforced and available for long-term use. (Wilkinson, 2004)
Memory, then, can be defined as “the way the brain is affected by experience and then subsequently alters its future responses”. (Siegel, 1999, p. 24) A person’s experiential history, his or her learning, is thus reflected in the structure of his or her brain.
Cognitive researchers have also shed much light on the fragile nature of memory, especially when it comes to being able to recall exact details and sequence of events. (Schacter 1996) On the basis of this research, Daniel Schacter of Harvard University categorically rejects the idea of distinct entities, whole chunks of photographic or cinematic memory. He writes: “Rather, information available in the present retrieval environment combines with stored information to yield an emergent pattern of activity that we experience as ‘memory’”. (Schacter, 1995, p. 24) In neurophysiological terms, this means that different parts of our memories are stored in different parts of the brain. A memory is always a recreation, a piecing together from many sources.
Especially when it comes to detecting and defending against threat, arousal has been found to be a critical factor, since it tends to monopolize brain resources, thus make the further processing and consolidation limited if not impossible. The learning and memory structures from such states reflect, among other things, the negative impact of childhood trauma. (LeDoux, 2002, p. 322)
But emotion may also have another and unexpected effect when it comes to the follow-up or consolidation of memory. While arousal may disturb attention at the moment of an initial experience, it seems to strengthen the ability to make experiences memorable. (McGaugh, 1995, p. 265-266)1
The most appropriate understanding of memory, then, is that each act of remembering creates new memories of old experiences. (Schank, 1990) What we are dealing with, in other words, is not only how a certain experience was encoded and stored, but the particular circumstances present in the actual situation when something is remembered. Some of this process may occur consciously, but, again, much of it happens without our awareness or involving our intent.
In this perspective, the analyst’s memory will have to be regarded as dynamic and evolving. How an analyst thinks, feels, and perceives, both in the sessions, and after, is determined by his or her memory. It is what aids the processing of new information specific to each therapeutic encounter. In short, analysts remember in order to understand.
But the analyst’s memory is not a merely passive process. It functions in a relational context and it develops from doing, that is, from what analysts do: listening, being reminded, and storing, on the one hand, and recalling, processing, and interpreting on the other. And since its focus is on meaning and beliefs – and psychotherapy or analysis is about meaning-making and the formation of structures which allow integration of otherwise fragmented memories – the analyst’s memory is a critical ingredient in the outcome of any psychotherapeutic treatment.
The difference between processing that goes on in a session, with a patient, and when later reflecting over it, points to the use of a process that involves several steps. What happens in sessions is intuitive and based on attunement to the patient’s psyche as well as to one’s own internal responses. The thinking after sessions is more extensive and often restores a sense of personal boundaries: what belongs to self can be reclaimed, something the intensely emotional involvement in sessions does not permit. (Levine, 1994)
1 LeDoux (1998) explains this phenomenon as being caused by the difference between two memory networks which need to work in tandem in order for long-term memory to be established. The first of these networks is located in the amygdala and is mainly responsible for reactions to threat. The other is located in the area of the hippocampus. Especially the hippocampal memory has been shown to be vulnerable to high levels of stress, although moderate levels of stress also seem to enhance memory encoding. Without hippocampal participation, no explicit memory will be formed (p. 240-246).
This fact has been described by Joseph Cambray (2001) in connection with an experiment of recording and reviewing all sessions for the last day before a vacation. Cambray found considerable difference between how he experienced certain interactions while they took place and when he later reflected on them. In particular his personal involvement, as it played itself out in all the sessions of the day of his experiment, came into clear focus only afterwards as a type of day residue. However, as Cambray carefully points out, upon reflection, when his countertransferences could be acknowledged as metacommunication presaging new awareness, distinctly different and relevant details surfaced. (Cambray, 2001, p. 292)
The importance of reviewing was succinctly articulated already in 1929 by C.G. Jung. He focuses on certain experiences that today seem emblematic to the analytic profession, the often unconscious overrides of theory and the risk of psychological blind spots when avoiding serious reflection and modification in the face of therapeutic failures. (Jung, 1929)
In my own experience, memory plays a critical role in how well I am able to help someone. How I perform in a given treatment appears to depend on how I am able to update my memory and integrate the patient information to a much larger extent than I was taught to believe in my training.
However, my memory of each patient’s process is based on both conscious and unconscious knowledge and relies on a type of recollection that is specific to what I do as an analyst. Although some of my personal experiences outside practicing sometimes enter into my mind in the beginning of a treatment, most of the memory material seems to come from previous treatments and, to a lesser extent, case presentations and other psychoanalytic literature. And once a therapeutic alliance has been established with a given patient, most memory associations appear to be connected to what I have experienced with the same individual patient.
My experience also tells me that my memory is aided by what I do after a session. Processing during a session, immediately after it, and when reviewing whole treatments, differs considerably and seems to call upon different functions, different ways of using my memory. As I explore the nature of the analyst’s memory more in detail, this difference will guide much of what I will review in the research literature and findings in the cognitive sciences may shed some further light on this phenomenon.
Cognitive research now appears to confirm that we use narratives to remember. Even though the exact neurophysiological pathways remain uncertain, several researchers conclude that narratives serve as memory structures. (Bruner, 1990; Siegel, 1999) Since memory is generally divided into three kinds in reference to its anatomical pathways – explicit or declarative (episodic and semantic), emotional, and implicit or procedural – it is assumed that narratives operate partially along pathways that are implicit or unconscious and partially along pathways that are explicit or conscious. (Eichenbaum & Bodkin, 2000) Siegel (1999) concludes that, especially in autobiographical narratives, the unconscious aspects can be traced to mental models or summations of repeated experiences, which are retained in implicit memory.
The emergence of narratives can be linked to the acquisition of language. By age four children appear ready to use autobiographic narratives to string together memories of many events. As perhaps the most significant cognitive achievements in childhood, verbal sharing of an experience with others now serves as a reinstatement and memory can be preserved. (Nelson, 2000) As a result, skills for social participation develop, as well as the necessary means for self-reflection.
As Jeremy Holmes (2001) shows, narratives may also reflect enduring attachment styles “by the ways in which people talk about their lives, their past and in particular their relationships and associated mental pain”. (p. 7) Attachment failures in early childhood affect what he and other developmental psychologists call narrative competence.1 Holmes defines such competence as “consciousness about our own mental life” and suggests that it translates into “a psychological equivalent of immunological competence”. (p. 4) This narrative ability, also has broad ramification for the mental health of the person. In finding such competence through psychotherapy, persons with previous attachment failures may in fact build the necessary new structures for overcoming these failures. (Holmes, 2001; Siegel, 1999)
1 A general understanding of narrative competence was first pursued by scholars in the field of literary theory or narratology. Jonathan Culler of Cornell University writes: “Children very early develop what one might call a basic narrative competence: demanding stories, they know when you are trying to cheat by stopping before reaching the end. So the first question for the theory of narrative might be, what do we implicitly know about the basic shape of stories that enables us to distinguish between a story that ends ‘properly’ and one that doesn’t, where things are left hanging?” (Culler, 1997, p. 84).
Roger Schank (1990), a cognitive psychologist and director of the Institute for Learning Sciences at Northwestern University, points out that by creating stories about our experiences, we condense them into “a story-size chunk that can be told in a reasonable amount of time”. (p. 115) In so doing, the original experiences become coherent. And by making intelligent use of indices to our stories, more complete storage is possible than if relying on a simple, event-based kind. The more complex the data and decisions, the more we have to make new stories or update old ones.
Schank (1999) views two phenomena in particular as involved in the activation of stories: the occurrence of something unexpected and being reminded. In the first instance, our mind takes notice when the expected did not occur. Without consciously having to decide, a new index will account for the circumstance we experienced and add it to the story for future use. The second instance, being reminded, has to do with how we respond when we hear someone else’s story. Our mind searches for a story of our own that may remind us of what we just heard. However, since no two stories are exactly alike, retrieval must take place by searching for stories with similar features. Listening to others, a match is assumed when we feel reminded of a story of our own.
But how do we know that our story is the same or similar to the one we heard? To Schank (1999), this is where beliefs come in. If we can construct a belief to go with the story we just heard, most likely we will find a story of our own that relates to this belief. We have connected to a story in our memory and we have done so by juxtaposing another person’s beliefs, made evident by his or her statements or actions, with our own. In the process, an index for future use is created. For instance, if person X is telling a story which seems to proclaim a belief in the blessings of free competition, we are reminded of a story that illustrates our own belief about free competition, a story that tells us what our experiences of it have come to mean.
We may safely assume that analytic interpretations and what traditionally has been called technique also serve this indexing function. Through training and experience, analysts seem especially focused on their creation. In interacting with patients presenting a wide array of traits and problems, new interpretative possibilities are continually being discovered. The understanding of narratives and indexing may in fact explain how this type of analytic memory develops.
The findings of Schank’s cognitive research thus have obvious application to what analysts do, how they contribute to each psychotherapeutic relationship, and how they train their memory to perceive, process, and communicate with patients. Some of Schank’s findings have broad application to all those in human services, others seem important in understanding analysts in particular, especially since most previous attempts leave much to be desired.
The particular expertise of analysts comes from learning stories and indices which can be used when relating to stories of others, stories about their pain and their desires, their strengths and their psychological deficit. By being reminded and extracting indices to a learned storage of stories, the benefit of developing an intricate system of indices to these stories is, at the same time, being reaped. With each treatment, certain unexpected data will surface and with it, new indexing.
But this expertise is also about how to provide the environment for new autobiography to be created, new stories to be told and remembered. For this to occur, analysts need to remember more explicitly what their patients related from session to session. Accordingly, the first priority for analysts is being attentive and interested so that as much of the relevant information as possible can be retained. This information, in turn, needs to be encoded in specific semantic memory. Siegel (1999) suggests that permanent explicit memory happens through a process called “cortical consolidation” and requires “a nonconscious activation or rehearsal process”. (p. 37) Schank (1999) points out that when we update our knowledge base, our explicit memory, only certain information is retained while previous incidents of a similar nature may be forgotten. This phenomenon points to a need to review and reflect in order to remember explicit aspects of the experiences with each patient. It also appears to explain why analysts have a hard time remembering what happened in the session before the last and the sequence leading up to it.
What is remembered form session to session is, at the same time, dependent on the stories that get activated. (Schank, 1999) Every new patient activates several stories and forces the creation of more search indices. The search for fitting stories is particularly strong in the initial phase of each treatment, but as soon as a useful set has been retrieved, the analyst can also begin the process of accumulating explicit, event-based memory. Before the activation of particular stories, the ability to relate, retain, and communicate will remain shallow.
The capacity for this type of listening and attention begins in the personal analysis, with the exploration of the analyst’s own life and the stories created about his or her life. As Mahoney (2001) suggests, particular paradigms allow analysts-in-training to create stories later to be used when listening to patients. With these as a base, further stories and indices give rise to yet other sets. In working with training supervisors yet others are added.
These narratives, initially tied to an autobiographical narrative, provide the necessary plot line, a narrative template, which organizes the analyst’s memory. An amalgamation of favorite theories, case presentations, technique formulations, and descriptions by the pioneers of analysis eventually occurs. What we call theory is in fact a personal compilation from all these sources and becomes a reusable template for what to expect. Schank (1990, pp. 147-170) calls such compilations story skeletons. They are quite general, but they do provide ways to validate what analysts do.
Applying an understanding of narrative structures to the psychotherapy process leads to a radically new understanding also of the patient’s experiences. The task of giving a coherent, livable, and adequate story about one’s own life is quite complex and as Jerome Bruner (2002) of Harvard University points out, inevitably involves pleasing an audience. He writes:
Telling others about oneself is, then, no simple matter. It depends on what we think they ought to be like – or what selves in general ought to be like. Nor do our calculations end when we come to telling ourselves about ourselves. Our self-directed self-making narratives early come to express what we think others expect us to be like. Without much awareness of it, we develop a decorum for telling ourselves about ourselves: how to be frank with ourselves, how not to offend others. (Bruner, 2002, p. 66)
Patients’ initial stories no doubt express this wish to please and follow decorum. Not only is there desire to be interesting to the analyst, still very much a stranger, and have his or her attention but the stories being activated often reflect the way a child wishes to please its parent, since the first stories learned date back to when story telling was first learned. Stories meant to prove goodness and loyalty or stories demanding attention by being vehicles for tales of failure, rejection, and despair seem to belong to this category.
In the long run, this is not the kind of story the patient favors and finds meaningful. (Ulanov, 1982) But since humans to a great extent rely on knowledge and belief already learned, old stories are easier to tell than to create new ones. As Schank (1999) points out, creating new stories involves more of an effort and only by attempting to tell someone else about new and recent experiences will a new story be memorized. For this to occur, the patient must have concluded that the old stories no longer offer a good enough fit. He or she must also have concluded that there are important new experiences to convey. So even if the patient’s first set of stories are met with appreciation by the analyst, an eagerness, or at least a willingness, to tell new stories is a necessary condition for psychological change to occur.
In this perspective, successful therapy depends on the creation of a cohesive new story, a story which can span over a longer time and encompass previously unintegrated memory. The inability to remember and make connections, means that the sense of self never becomes embodied. (Covington, 1995) So inevitably the patient is placed in the uncomfortable role of having to create a new story from many old ones, previously seen as unrelated or non-existent. The sense of integration and meaning comes from the placement of the characters or actors in the patient’s autobiography within an overall configuration, the “plot”. (Bruner, 1990)
In order to relate fully to a given patient, analysts rely on being able to respond to the particulars of his or her moods and feeling. This permeable presence and openness to induction has increasingly become an important concern in the analytic literature. The trend is to view the analyst’s contribution to the analytic process in terms of a broadly understood use of countertransference reactions (Mitchell, 1988). Analysts, according to this view, use their involuntary reactions to a patient for more in-depth information, since these reactions, upon further reflection, can be traced back to the patient’s unconscious. (Winnicott 1960; Levine, 1994)
This was not exactly how the pioneers in psychoanalysis used the term. Freud, for instance, regarded countertransference as something the analyst needed to recognize and overcome, but it soon became apparent that distinguishing between what happens in the analyst’s mind and what happens in the patient’s was difficult if not impossible. (Fine, 1979) However, traditional psychoanalysis offered few other theoretical constructs for the unconscious influence that the patient’s state of mind has on the analyst. Thus the term “induced counter- transference” soon became the common approach. (Winnicott, 1960; Mitchell, 1988)
Without taking the analyst’s memory into account, this way of describing mutual influence seems quite awkward and as Morris Eagle (2001) points out, tends to create yet another layer of technical recommendations. Instead of a one-sided emphasis on being able to detect and understand the patient’s transference, analysts are now the expert judges of what their own reactions mean for the patient.
The lack of clear formulations about mutual influence has been traced to the early experiences by Freud and Jung with hypnosis and the prohibitions both men left about undue suggestions. (Shamdasani, 2001) However, since the 1980s, many psychoanalysts, in embracing an understanding of transference and countertransference as intersubjective phenomena, have moved closer to an acknowledgement of mutual influence as an inseparable part of psychotherapeutic treatments. The problem with the new understanding is that it tends to ignore the analyst’s contribution to the dyadic experience. In Eagle’s (2001) words, “we assume that all feelings and thoughts that emerge in your experience necessarily, and in any simple, uncomplicated way, reflect what is going on in the patient’s inner world”. (p. 36) This may be just another way to regard everything in the treatment as reflecting something about the patient.
A different approach to induction and dyadic dynamics can be found in the formulations of Thomas Ogden (1997). He suggests that what he calls reverie is the important ingredient in the analyst’s processing of patient-induced feelings. Referring to a statement of Freud’s (1913), Ogden claims that “a necessary condition for the conduct of analysis” is that both analyst and analysand “gain access to a state of reverie” (p. 114), thus advocating for mutual states of altered consciousness. Implicit in his argument is that analysts enter trance-like states, a type of auto-suggestive reverie which heightens the experience of dyadic interaction while screening out its larger context. In so doing, he or she would also enable the patient to accept reveries as the means towards integration.
As David Spiegel (1995) of Stanford University Medical School notices, one of the characteristics of such states is an unusual degree of absorption, what he terms “an immersion in a central experience at the expense of contextual orientation”. (p. 130) Such a state is to Spiegel part of hypnosis as “aroused, attentive, focused concentration with relative constriction of peripheral awareness”. (p. 129)
Spiegel (1995) differentiates between three different components or hypnotic states: absorption, dissociation, and suggestibility. He argues that these components have specific corresponding effects on the three main components of memory processing: encoding, storage, and retrieval. Absorption, as a narrowing of the focus of attention, causes problems with encoding, dissociation with storage, and suggestibility with the process of retrieval.
Allan Hobson (2001), of the Neurophysiology Laboratory of Massachusetts Mental Health Center and a sleep researcher, compares hypnotic states with REM sleep. He writes:
Because recent memory is disenabled in both states, it is not surprising that orientation to time, place, and person is impaired in hypnotic trance and dreaming. That this process is considerably more floridly deranged in dreaming may possibly be due to the more extreme changes in neuromodulatory balance that occur in REM sleep. (Hobson, 2001, p. 101)
To Hobson, the similarities between dreaming, psychosis, and hypnotic states indicate an analogy between these altered states, although the brain mechanism may be quite different in each. What they do have in common is problems with memory encoding and consolidation. Altered states thus affect implicit memory structures while more or less circumventing the common pathways for episodic and semantic memory. (Knox, 2003)
This is probably what Odgen (1997) had in mind when stating that the analyst “renders his own unconscious receptive to the unconscious of the analysand”. (p. 113) To him, analysts appear to enter into one of these altered states when processing patient-induced feelings and we can conclude that these reveries are associated with a high degree of hypnotic absorption. Such semi-trances carry encoding liabilities, however. The high absorption of what transpired in a session will necessitate further processing afterwards in order that the many intense impressions from the encounters may find proper and enduring memory storage.
Based on cognitive research on memory, I have outlined two perspectives on the analyst’s contribution to psychotherapy treatments. The first has to do with the role of narrative structures in general memory and how these findings translate into a more detailed understanding of how analysts use their memory. I am suggesting that competence as an analyst depends on the capacity to train a certain type of memory. This type of memory is based on narratives which are updated and expanded via indexing. (Schank, 1999) When considering the indexing analysts learn, we seem to be dealing with narratives that are interpretive, perhaps constituting a particular kind of interpretative memory.
The second perspective presents a cognitive understanding of what in the more current psychoanalytic literature is being called induced countertransference. Since processing in sessions involves periods of intense attunement to the patient’s moods and feelings, analysts appear to enter into certain hypnotic states or reveries, when they are particularly open to induction. (Ogden, 1997) Such attunement allows for a high degree of absorption in the patient’s experiences but carries problems with encoding and consolidation.
When periods of reverie are considered, induced countertransference may best be understood as a type of learning that relies on implicit memory. However, as all implicit learning, such experiences only become part of long-term and explicit memory with further processing. The analyst’s memory must thus be based on several stages of learning. One such stage occurs during sessions, another immediately after them, and a third when reviewing whole treatments. Attunement to the patient’s moods and feelings is only the initial elements in how analysts remember. In fact, without some type of further processing of experiences in sessions, the analyst’s recall and understanding will be inadequate.
In examining the two aspects of the analyst’s memory, narratives and reveries, I also touch upon areas where new research conflicts with or alters basic conceptions in depth psychology. From the beginning, as the name indicates, depth psychology hypothesized the existence of unconscious influences on human behavior and these influences were viewed as part of a larger totality, the psyche, with common roots in all humans. (Ellenberger, 1970) The model is one of deep-seated drives or complexes, by nature more or less resistant to consciousness, but nevertheless affecting the person’s experience and behavior in many indirect ways. It describes the human mind as struggling with its past, with its darker impulses, and with needs to restore a sense of self due to fragmentation and dissociation.
Cognitive science and the research on memory introduce a different picture. In this model, the creation and alteration of memory, even when it is going on unconsciously, happens as experiences are processed via different pathways of the brain, different neural networks. In this model, the mind is no longer split or layered but dependent on several regions of the brain functioning together. (LeDoux, 1998; Damasio, 1994) Any memory or perception is an assembly from many sources and the human mind is viewed as an organism specifically geared towards remembering in order for the person – and the species – better to be able to survive. (Siegel, 1999)
The findings of the various branches of cognitive science will no doubt change analytic theory, a process that already appears to have begun. (Cozolino, 2002; Solms & Turnbull, 2002) However, cognitive science is not a clinical theory. In spite of being a burgeoning field of research, it has shown scant interest in psychotherapy. Hopefully, this will change as there are several phenomena in clinical practice worthy of cognitive research. The analyst’s memory, and the kind of learning it represents, ought to be one of them. Not only does it appear to be a form of expert memory central to the analyst’s contribution to effective treatments, how analysts remember may also give important clues to how understanding of others actually comes about.