“That’s What You Say”: Reply to Levine and Stern

A joke I heard in graduate school: A young philosopher is working deep into the night. Feverish and sleep deprived, he struggles to craft a solution to the great philosophical dilemmas posed by sages from the past. Exhausted by his efforts, he falls into a deep sleep on his bed and begins to dream. In his dream, all the great philosophers—Aristotle, Plato, Kant, Descartes, Hume, Nietzsche, de Beauvoir, and others—come before him, one by one, to present their arguments, and one by one, the young philosopher refutes them all. To each, he offers a devastating critique, so powerful that each of them disappears in a puff of smoke. The young man awakens briefly, and, recalling the dream, he stumbles to his desk to record his shattering critique while he can still remember it. He manages to scribble it down before collapsing again into sleep. When he awakens fully the next morning, he remembers the dream but not his retort, and is relieved when he recalls that he wrote it down. Rushing to his desk he sees, written in his own wobbly hand, the following: “That’s what you say.”

We psychoanalysts often find ourselves in a similar position. We read, we write, we argue over differing theoretical positions, but in the end, our replies devolve into something like the young philosopher’s retort: “Well, that’s what you say.” We have little alternative, as we have few consensually agreed upon criteria by which we could validate a construct, or confirm one theory and disconfirm another. Levine, citing Bion, argues that the only criterion for evaluating a psychoanalytic theory should be its clinical usefulness. Sounds reasonable, except that there is virtually no evidence that any one psychoanalytic theory is more clinically “useful” than any other, to say nothing of comparing any psychoanalytic theory with nonanalytic theories such as cognitive behavioral therapy, dialectical behavior therapy, eye movement desensitization and reprocessing, or any of the array of therapies that have arisen in more recent years. It’s not even clear that there would be consensus on how to define “clinical usefulness.”

How do we escape this state of affairs? Within psychoanalysis itself, we have a panoply of mental constructs, processes, and theories, amassed since the very beginnings of our science. Some are surely zombie concepts, never to be dispatched and buried, even when they have been seriously questioned or superseded by more contemporary findings, constructs such as primary and secondary processes (but see Litowitz, 2007, and Erreich, 1994), regression (but see Inderbitzen & Levy, 2000, and Rizzolo, 2019), psychic energy, the death drive, the topographic model, and others. Philosophers of science speak of pitting theories against each other to determine which should be retained and which discarded, but psychoanalytic constructs and theories don’t generally get discarded. Rather, our theoretical store has become bloated by the accretion of new constructs, new theories, as they arise in our history (e.g., alpha and beta elements, unformulated and unrepresented experience, figurability, field theories). Are these novel proposals? Or are they merely notational variants of already extant constructs and theories? We often don’t bother with these niceties.

There have been some changes, mostly brought on by external pressures that emerge either from society or from outside our discipline. For example, social changes have led us to dispense with penis envy, to rethink homosexuality, and to reckon with transgender identities. Pressure from neighboring disciplines, such as developmental psychology, has led most of us to no longer think, as Freud did, of infants as passive beings engaged in hallucinatory wish fulfillment. Similarly, Mahler’s notion of the autistic infant and separation-individuation appears to have been eclipsed by a focus on the vicissitudes of attachment. As with changes occasioned by societal pressures, there was significant resistance to allowing these findings into our literature (e.g., Psychoanalytic Study of the Child, 1960, Volume 15, Issue 1).

Since the 1960s, cognitive developmental psychology has used ingenious modes of observation and experimentation to investigate the development of mind, a potential boon to psychoanalytic thinking about the same topic. I have proposed that one way out of the “that’s what you say” conundrum is via relevant data from other disciplines that may help us adjudicate the validity of psychoanalytic constructs and theories.

Any discussion regarding the disagreements among Levine, Stern, and myself needs to be situated in the answer to a critical question: Do you, dear reader, believe that there should be consilience between psychoanalytic thinking and findings in related disciplines, or do you believe that we should continue to remain ensconced in our splendid isolation almost 125 years after the beginnings of our science? If the latter, this commentary is not for you; burn before reading. The context for my discussion of Stern and Levine, articulated in other writings (Erreich, 1994, 1999, 2003, 2015, 2017, 2018), will be the assumption that psychoanalysis, like every other scholarly discipline, should be in dialogue with neighboring disciplines, that it cannot thrive if it continues to ignore what others have discovered about human mentation and development.

Levine: Plausibility and Pragmatism

I am grateful for Levine’s clarity with respect to his two-track model; he has opted for the creative benefit of developing our theories in “splendid isolation” from other disciplines:

I have not been concerned with how this expansion [Levine’s theory] fits with the facts, findings, or theories of related disciplines. The problem I have sought to address has been and continues to be how to understand and respond to my next patient. (p. 54).

And

I have never sought congruence with the findings of these or other extraclinical, extra-analytic fields. My only concern has been the clinical usefulness of the model I was proposing. . . . I proposed that we limit the term representation in psychoanalytic discourse to a Freud-based psychoanalytic meaning, in the hope of defining and restricting its usage to being a specifically psychoanalytic term of art. (p. 54).

Why this narrow lineage? As I have noted, not only do mental representations stand for propositional knowledge (involving symbol systems such as language or mathematical formulas) and analogic knowledge (such as imagery, sound, taste, or smell), but they also carry affective valence. Sigel (1999), a cognitive psychologist, notes that mental representations carry a strong affective valence; they

can induce excitement, fear, discomfort, exhilaration, and an assortment of other emotions and motivations. . . . Conceptions of representations that fail to acknowledge the “hot” nature of representations in use are incomplete. Representations are therefore explicitly recognized . . . to possess affective and motivational characteristics. (p. 25)

What more could a psychoanalytic theory of mental representation long for?

In my presentation, I wholeheartedly agree with Levine’s notion that plausibility is an important criterion for psychoanalytic hypotheses. Plausibility depends on some hypothesis being reasonable or probable given other known facts, but it is an abuse of this principle to simply posit entities without any regard for what is already known, in this case, about the human capacity for mental representation. Levine insists that

Erreich’s demonstration of the competence and representational capacity of the infant’s psyche is not at all an argument against there also being unrepresented forces present. As we shall see, I think it is pretty clear that unrepresented force without meaning is how Freud thought of the id and probably the drives in their originary state. (p. 52)

However, Levine does not cite any evidence to support the plausibility of the existence of such forces, and Freud’s authority does not abrogate the need to offer such evidence. Despite my fulsome catalogue of precocious infant abilities, Levine writes,

For Bion, Green, and others, the idea of primordial mind refers to a part or level of organization that is inherent in each of us from before birth and remains so throughout life. It is not to be equated with the infant’s psyche. (p. 52, fn. 6).

Is the notion of a “primordial mind” in infancy a plausible hypothesis? What can Levine offer in response to the findings I have enumerated regarding the infant’s significant innate perceptual, representational, and memorial capacities that plausibly suggests that subjective experience is represented and memorialized both pre- and postnatally?

Levine (p. 49) repeatedly cites pragmatism, efficacy, or “expediency” in the clinical situation, as the criterion for importing into psychoanalysis any ideas from without. Opting for what “works” in the clinical situation is a very shaky foundation for any theory, as there is scant agreement among psychoanalysts about what “works” in psychoanalysis, or even what “works” might mean. The ambiguities and multiple determinants in clinical material and the underspecified nature of our hypotheses lead to abundant disagreement about any clinical situation. A recent example comes to mind: During a clinical presentation, a dark-skinned patient accuses his light-skinned analyst of racism in response to an intervention. Some participants see this as the patient’s straightforward assessment of the analyst’s racism; others see the epithet as merely the latest manifestation of the patient’s negative transference. Apart from what sort of intervention might “work” best in such a situation, there is no consensus regarding even the nature of the problem.

One consequence of remaining isolated from other disciplines is that we wind up having some nonsensical debates. For example, Levine writes, “Green ([1997/2023]), probably here influenced by Bion ([1962a], 1970), argues in favor of metapsychology” (p. 55). This issue of whether psychoanalysis needs a metapsychology is not a sensible question. In the larger world of scholarship, it is generally understood that every psychology is embedded within a larger metapsychology that specifies fundamental assumptions that underly that brand of psychology. A metapsychology stipulates a set of principles that cannot as yet be verified empirically and so must be assumed, for example, the nature of the relation between mind and body. Thus, Levine’s comment about “the absolute clinical necessity of metapsychology” (p. 56) isn’t incorrect; rather, it states the obvious: that for every psychology, its metapsychology is a logical necessity. The metapsychology within which Freud’s and Klein’s clinical theories are embedded is entirely different from that of cognitive behavioral therapy or attachment theory; each depends on a different model of mind and development and different assumptions regarding the mind-body relationship. Throughout his essay, Levine, with his citations of Bion and Green, appears to assume a particular metapsychology, one in which the traffic between mind and body moves easily and confidently. Psychic energies, drives, and somatic forces and discharge can be bound, contained, and linked to ideational meanings, and when they are not, they remain unrepresented.

Finally, a logical question arises: What is the ontological status of “unrepresented experience”? As we know, somatic tensions, forces, and sensations occur not in the body but in the mind. Bodily sensations, like all other subjective experiences, must be somehow represented in mind in order to be experienced; that is, there is no subjective experience of bodily sensation that is not mediated by mental representation. So where do these elusive somatic tensions, forces, and sensations reside if not in the mind?

Stern: A Theory of Consciousness

In his commentary, Stern writes that he is interested in developing a theory of consciousness from the perspective of the hermeneutic tradition and phenomenological psychology. Stern protests that he has not addressed the sources of severe psychopathology or questions of mental representation in infancy.

I was puzzled by Stern’s insistence that his theory of unformulated experience is unrelated to psychopathology; in this assertion, he seems to want to differentiate his thinking from Levine’s on the relationship between unrepresented experience and psychopathology. Yet in 2004, Stern presented a clinical vignette intended to illustrate that very relationship: “Enacted experience, and thus dissociated states as well, cannot be symbolized and therefore do not exist in any other explicit form than enactment itself. Enacted experience is unformulated experience” (p. 213). (Recall that I’ve argued that enactments are in fact represented in the indexical register, per Pierce’s taxonomy.)

Stern correctly notes my mistaken attribution to him of comments by Green that Stern (2015b) cites in a review of Levine, Reed, and Scarfone (2013):

In this frame of reference, an important aspect of psychopathology, especially more severe types, becomes the incapacity to create representations and link them with drive, leaving the mind to function somehow with voids and absences. The crux of clinical work in these cases, and in those parts of less severe cases in which absences and voids nevertheless figure, becomes the creation of representations that were literally ‘not there’ before [italics added]. (p. 493)

But notice the similarity between Stern’s description of Green’s model above and Stern’s own words regarding unformulated experience as recently as 2020:

Some formulations must simply not be made. These “not-me’s” habitually go unformed; they remain unsymbolized. They have no shaped or structural presence in the mind. They are like the white spaces in a painting: they have no actual presence, and yet their absence gives shape to other presences. These white spaces—unformulated experience that, in an ongoing way, is continuously denied formulation for unconscious defensive reasons—are what I describe as “dissociated experience” [italics added]. (Stern, 2020, p. 914)

A careful reading of Stern’s (2015b) review of Levine et al.’s (2013) book in particular demonstrates Stern’s view of the close connection between unrepresented states and unformulated experience, and his approval of Levine et al.’s ideas regarding psychopathology and the development of mental representations. ¹

Perhaps I may be forgiven for my misattribution of Green’s quote to Stern; like those International Journal of Psychoanalysis editors who also noticed the close relation of Stern’s work to the thinking in Levine et al. (2013) and thus asked him to review their book, I too was impressed by the similarities in these two models. Given his approval and agreement with so much of Levine’s thinking, Stern’s comment below seems rather disingenuous:

I am incredulous at the suggestion, made in more than one place in the [Erreich] article, that, “The ‘infantile,’ that is, the infant’s ‘unrepresented’ or ‘unformulated’ experience, has come to be viewed as a primary pathogenic factor in Kleinian, French, and relational theories.” I will not comment on whether that is a reasonable characterization of Kleinian and French psychoanalysis, but it is certainly not supportable as a characterization of relational psychoanalysis, as written by me or anyone else (p. 63).

I did not originally include a consideration of Stern’s 2019 book, The Infinity of the Unsaid, because its topic was not pertinent to my argument. However, even as Stern’s interests have pivoted toward attempts to formulate a theory of ordinary consciousness (Stern, 2019), he still seems to feel a strong commonality between his thinking and Levine’s: “the subject of unrepresented states is itself, independently, a highly significant part of the scholarly and clinical context for the theory of unformulated experience” (p. 27).

Stern is also correct that the shorthand in my abstract collapses his work and Levine’s with respect to the provenance of unrepresented, unformulated experience. However, as I repeatedly stated, “it is not always clear whether the provenance of the unrepresented and unformulated experiences they [Levine and Stern] refer to is the infant’s alone, or whether such states can also arise from adult experiences” (p. 9). In his discussion, Stern clarifies:

It [his model] does not have a developmental thrust at all. It is, instead, an attempt to say something about how ordinary knowing and experiencing occurs, moment to moment. . . . that all conscious experience begins in an unformulated state [italics added].

He further writes, “I set out to characterize all conscious experience [italics added]” (p. 61).

In The Infinity of the Unsaid, Stern (2019) lays out his change in focus to the formulation of nonverbal experience, of procedural knowledge, in addition to verbal experience. I would argue that it behooves him to consider the developmental issues raised by this more recent formulation of his model. After all, infants, despite their innate store of rich metalinguistic knowledge, have limited expressive language for a year or so after birth; that is, they are nonverbal, and they demonstrate the ability to represent procedural knowledge (Erreich, 2015). How does his model apply to them? Stern himself begs this question: “Unformulated experience . . . is constantly and expectably being created and then going through the process of formulation: in childhood [italics added], in adulthood, anytime in human living at all” (p. 64).

Stern currently views his theory of unformulated experience as “a way of grasping the way ordinary knowing takes place” (p. 64). The larger scholarly world does not look to psychoanalysis for models of unconsciousness or consciousness for obvious reasons: We have little empirical evidence to offer. It is not provocative to insist that any proposal regarding nomothetic principles of human mentation (i.e., a general theory of consciousness) should consider what is already generally accepted knowledge in this domain. Perhaps in the early 1960s it was considered acceptable for Bion (1962b) to conjure a theory of thinking that made no reference to what was even then known about human mentation, but one would hope that time had passed, and that we could by now offer more than a “that’s what I say” argument on this very important topic.

Stern tells us that he relies on Fingarette’s (1963) thinking on self-deception and William James’s (1890) work on the stream of consciousness, as well as others from the hermeneutic tradition and phenomenological philosophy. I have no brief to make against these views, ² but to my mind, they do not substitute for more recent sources, especially those anchored in a different tradition, one that respects empirical findings regarding human mentation and unconscious and conscious states in particular. A Google search for the word consciousness provides an almost overwhelming abundance of material on this topic, largely from neuroscientists and philosophers of mind (many well versed in cognitive psychology); the bulk of this material is ignored by Stern. ³ I would argue that such thinkers offer more plausible views of consciousness than the musings of writers from the phenomenological and hermeneutic traditions and that psychoanalysis is more productively viewed as a branch of psychology than a branch of philosophy. In this respect I agree with Coelho’s (2020) review of Stern’s (2019) book:

Usually, the more an author has difficulty recognizing the value and role of Freudian and post-Freudian metapsychologies (Klein, Lacan, Ego Psychology, and the British Independent Group), the more they tend to resort to importing a philosophical resource to replace established psychoanalytic theorizing. Thus, a theoretical-clinical hybrid is created in which philosophy or a certain philosophical idea emerges as the guarantor of a psychoanalytic practice. . . . In such situations, in my opinion, psychoanalysis loses as a theory and as a clinical field of work. (p. 412)

I am also in agreement with Coelho’s view that the phenomenological and hermeneutic traditions tend to undermine the critical value of unconscious mentation:

Emphasis on the ideas of “spontaneity” and “authenticity” on one side, and “use” on the other, marks a certain model of thinking that puts Stern’s work, even if he does not want it, closer to a psychology of consciousness than to a psychoanalysis centered on unconscious conflict. In this sense too, the rapprochement with phenomenological philosophy, Gadamer’s hermeneutic tradition and with a certain linguistic position also brings his psychoanalytic vision closer to the philosophies of consciousness. (p. 413)

Perhaps the most difficult challenge for Stern’s model of unformulated experience is the same charge I leveled against Levine: What is the ontological status of unformulated experience? With Levine, I emphasized the fact that all somatic sensation is registered or “represented” in the mind, not in the body, giving lie to the notion that somatic experience is unrepresented. Stern (2020) asserts that unformulated thoughts are not only not symbolized (not verbalized in Peirce’s taxonomy), but they have no “structural presence in the mind” (p. 914). What then is the nature of these entities? What is their ontological status? I have argued that registration in the mind, conscious or unconscious, is a necessary and sufficient condition for subjective experience. It seems unsupportable to assert that unformulated experience is subjective experience which remains untethered to anything mental. My formulation actually supports a cornerstone of Stern’s current proposal regarding his expansion of unformulated experience to include the nonverbal as well as the verbal. The verbal and nonverbal are united by the fact that both are represented in the mind, conscious or unconscious.

Finally, I am very much in agreement with Stern regarding the notion that a theory of unconscious processes must be “sensitive to context, especially the interpersonal context” (p. 67), which, he correctly claims, has not usually been included in what has generally been considered “intrapsychic.” In 2003, I proposed that we redefine unconscious fantasy as the intersection of (a) endogenous wishes (fantasy), (b) veridical perception of exogenous events (reality), and (c) naive misattribution or misinterpretation of reality (due to children’s limited knowledge about the world). Note that veridical perception of reality includes family and cultural biases, allowing for a principled conceptualization of how family, cultural, and societal values and prejudices become part of our intrapsychic realty in the form of unconscious fantasies. ⁴ That is, one is not obligated to choose between fantasy and reality, because our mental representations (unconscious fantasies) are composed of both.

Concluding Remarks

Any psychoanalytic theory that claims our attention should offer a model of the mind, a model of development, and a model of technique that devolves from those two. Psychoanalysis is a developmental discipline; any theory that ignores what is known regarding the development of human mentation can only remain deficient. The fact that Stern does not articulate a developmental perspective does not absolve him of the requirement to do so, especially if he’s writing about normative rather than pathological states. If we take on authority or on faith what Freud, Bion, Green, and others have written, we are in the realm of hallucinating babies and refrigerator mothers.

I will not restate my previous arguments regarding the value of extraclinical data to confirm or render more plausible our hypotheses and theories, but they apply to both Stern and Levine. The robust evidence that very young infants and even fetuses can represent subjective experience, especially highly traumatic experience, seems to plausibly argue against the existence of any kind of unrepresented experience. In fact, terms such as unrepresented experience and unformulated experience are oxymoronic; there can be no subjective experience without representation in the mind.