Abstract
Jouissance refers to an excess enjoyment beyond (yet tied to) speech and representation. From the perspective of some Lacanian analysts, jouissance is precisely what testifies against any relationship to the brain—jouissance “slips” out of cognition. On the contrary, it is argued here that jouissance has a central place in contemporary neuropsychoanalysis. In part 1 of this series the metapsychology of jouissance was presented in relation to the real and symbolic registers. Here, in part 2, Mark Solms’s neuropsychoanalytic model of Karl Friston’s free energy principle is summarized. In this model, “predictions” aim to resolve prediction errors—most notably, those signaled by affective consciousness. “Surplus prediction error”—prediction error that arises at the point where the predictive model fails—is proposed to be a neural correlate of jouissance. This limit within prediction is analogous to the real as a structural negativity within the symbolic.
As an extensive discussion of bio-reductionism is beyond my scope here, I will make only a few brief remarks on the matter. 1 In general, psychoanalytic concepts (such as jouissance) are not reducible to neural structures or functions. In neuropsychoanalytic “dynamic localization,” psychic functions are not housed in neural structures. Rather, they emerge between constellations of activity in various neural centers (Kaplan-Solms and Solms 2002). Individual parts are thereby necessary but not sufficient for a given psychological function (Dall’Aglio 2019). Moreover, a neurobiological basis for a psychic function does not explain that function—it merely describes its properties in another register. Psychic functions must be explained in psychological terms (Solms 2015a), and often psychoanalytic concepts shed light on neurobiological functions.
This, however, does not mean that knowledge flows exclusively from psychoanalysis to neuroscience, as Blass and Carmeli (2007) would have it. Knowledge from one field can enrich the other, by providing converging evidence, by challenging contradictory ideas, or by suggesting different relationships among concepts (Dall’Aglio 2020, 2021). Neuropsychoanalysis is a dialogue, not a reductionistic translation (Solms 2015a). As Verhaeghe (1999) notes, Lacan’s registers (real, imaginary, and symbolic) are conceptual tools with which one may dissect phenomena (e.g., an analysand’s speech, psychoanalytic theory, history, science). Drawing relationships between neuroscience and Lacanian registers is an instance of such conceptual knifework.
A Naturally Unnatural Brain
Moreover, neuroscience has increasingly “de-naturalized” the brain, finding that the brain is irreducible to its natural pre-givens. For example, neuroplasticity shows that brain structure changes based on individual experience (Ansermet and Magistretti 2007). The brain is “naturally” (i.e., genetically) programmed to be open to “unnatural” (i.e., not genetically programmed) influences. Epigenetics—systems that control the extent of gene expression—extend this logic beyond ontogeny. Environmental experiences may be encoded in epigenetic effects on the individual and even the individual’s offspring. The social world can thus warp the genetic guidelines for the brain—the latter is unthinkable without the environmental histories that shape it.
Ansermet and Magistretti (2007), in a Lacanian vein, emphasize how neuroplasticity opens a “gap” between the brain and its environment. Insofar as neural traces left by experience are subject to continual modification based on visceral signals and past memories, the “trace of experience” is not faithful to that experience. In other words, the structure of the brain—its developed neural networks—bears the mark of social history without being reducible to that history. The brain, while shaped by its experience, is not a mere reflection of that experience. Thus, the object we call the brain necessarily contains much that is beyond the brain. As Ansermet and Magistretti put it, the brain seems “genetically determined not to be genetically determined” (p. 8).
These neuroscientific findings resonate with Freud’s “infantile helplessness”—developed by Lacan as the “body-in-pieces”—which describes the premature state of the newborn (Johnston 2013a). To put it simply, the organism as an organized unity does not exist from birth. For Freud (1914), the unity of the ego is not pre-given; it must be constituted by a turn to the social world (see part 1). This psychoanalytic “subject-to-be” resonates with the “premature” (i.e., genetically indeterminate) brain whose constitution does not exist without the mark of the social environment (Johnston 2019).
I hope this openness on the side of neuroscience will assuage antireductionist objections. 2 Identifying a neural correlate of a psychoanalytic concept does not replace that concept—it opens the door for interdisciplinary suggestions and new ideas. I will now turn to one such bridge: jouissance in the neuropsychoanalytic model of predictive coding.
Of Predictions and Prediction Errors: Karl Friston’s Free Energy Principle
According to the free energy principle (Friston 2010), the brain operates like a Bayesian inference machine, generating probabilistic “predictions” about experiences. These predictions are compared to sensory feedback, and any resulting discrepancy is called “prediction error,” “surprise,” or “free energy.” 3 In this model, the brain is not a passive recipient of sensory impressions. Rather, it actively explains its experiences by constructing a “predictive model.” “Prediction” is operationalized as a probabilistic inference of the cause of a sensory input, either from within the body (i.e., interoception) or from the external world (i.e., exteroception). Prediction error is the unexplained or unexpected part of experience, a measure of entropy in a system (Friston 2010).
Self-organizing systems (such as the brain) try to maximize evidence for their own existence (i.e., their predictive model)—that is, they optimize their predictions, minimizing prediction error (Connolly and van Deventer 2017). There are two broad ways to minimize prediction error: perception 4 (i.e., updating the predictive model) and action (i.e., selecting inputs that support the model). What matters is not the accuracy of the model per se, but the minimization of prediction errors (Friston 2012).
The free energy principle can be applied from basic biological systems in the body (e.g., cells) through higher psychological functions (Connolly and van Deventer 2017) such as perception, action, and attention (Parr and Friston 2018). Organized in a hierarchical fashion, unresolved prediction errors from lower levels drive changes in higher-level predictions. Top-down predictions then feed back down to the lower levels to “explain away” the ascending prediction error. However, not all prediction errors are equal—not all errors are significant enough to reach higher levels. “Precision” marks the salience of a given prediction error. Prediction errors with higher salience are more likely to drive the updating of predictions. Precision may be registered by different neurotransmitter systems (e.g., dopamine, serotonin) in different domains, such as cognition, perception, and interoception (Parr and Friston 2018).
In sum, the brain constructs a hierarchical predictive model that explains inputs. Inputs are in themselves prediction errors (i.e., unexplained quantities). Predictions decrease prediction errors by providing probabilistic explanations of their causes. Prediction therefore falls on the side of what is explained and understood. Prediction error falls on the side of what is unexplained. “Precision,” or salience, 5 marks which prediction errors are significant and require changes (through perception or action) in the predictive model. For Friston (2010), it is a biological imperative to minimize prediction error and maximize the efficacy of the predictive model.
Mark Solms’s Neuropsychoanalytic Model of the Free Energy Principle
The Conscious Id and the Unconscious Ego
Friston’s free energy principle has been vigorously incorporated into Mark Solms’s neuropsychoanalytic metapsychology. This model (see Figure 1) begins with the dynamic localization of the id to the upper brainstem and limbic structures, and the ego to the neocortex (Solms 2013). Sensitivity to the internal milieu of the body and to the pressure of somatic demands are functions of the id, whereas exteroception and cognition are functions of the ego (Freud 1923). These functions correspond to the upper brainstem (and associated limbic areas) and the neocortex, respectively.
Medial and lateral views of the brain (top) and Freud’s topographic and structural models (bottom)
As Solms points out, the structures associated with id functions are intrinsic to consciousness. If one damages upper brainstem structures, one significantly impairs—or even destroys—consciousness. On the other hand, one can damage swaths of the neocortex and, though the quality of consciousness changes, the subjective feeling of being conscious remains. Hydranencephalic children, who are born without a neocortex but intact limbic and upper brainstem structures, are striking examples (Merker 2007). These children have no functional cortex (i.e., no representational or cognitive systems), but they are fully alive with affective consciousness. They show clear emotional responses to experiences with others. Animal studies (Moruzzi and Magoun 1949) and human studies (Penfield and Jasper 1954) have supported the central role of the upper brainstem and limbic system in generating consciousness. The cortex (ego) cannot be conscious without the brainstem (id), yet the brainstem can be affectively conscious without the cortex (Solms 2013). 6
Affective Consciousness, the Preconscious, and the Unconscious
This reverses the traditional Freudian equation of the id with the nucleus of the unconscious (Freud 1923). Rather, it appears that the id is the fount of consciousness—specifically, affective consciousness 7 (Panksepp 1998). Affective consciousness is nonrepresentational in itself—it is the raw feeling state of being. Affect signals unmet needs, either within the body (e.g., hunger) or socioemotional needs within the brain (e.g., attachment). There are seven of these emotional needs or instincts: SEEKING, LUST, CARE, PLAY, PANIC (attachment needs), RAGE, and FEAR (Panksepp 1998). 8 When these needs are unmet, they generate unpleasure. Meeting them generates pleasure. In other words, these affective systems follow the homeostatic logic of the pleasure principle (Solms and Turnbull 2002).
Although these emotional instincts come with innate behavioral responses (e.g., PANIC-protest drives the organism to SEEK its caregiver), these action plans are only “rough and ready.” In other words, they are insufficient to deal with the unpredictable complexities of life. The child must learn how to meet its needs. This is reflected in the “holes” inherent in the structure of these affective systems—neuroplasticity allows experience-dependent learning of what to FEAR, what to attack (RAGE), and so on (Dall’Aglio 2019; Solms and Turnbull 2002).
Learning to meet these needs is the task of the ego. The ego (neocortex) is in itself preconscious, becoming conscious only when aroused by affective consciousness ascending from the brainstem (Solms 2013). Solms (2017) localizes the unconscious to the automatized action plans (i.e., behavioral responses) in the basal ganglia and cerebellum. In contrast to the declarative, representational processes of the neocortex, automatized plans do not permit reflective thought—they operate outside awareness as nondeclarative (i.e., nonrepresentational) memory systems. They operate and repeat independently of self-reflective cognitive control. The ego strives for automatization of action plans that meet our (id-based) socioemotional needs in the outside world.
Dynamic Localization of the Free Energy Principle
At this level of analysis, one can dynamically localize free energy to the activity of subcortical emotional systems, which are functions of the id. Specifically, affective consciousness signals the precision of prediction error, salient prediction errors that demand the predictive work of the cortex (Parr and Friston 2018; Solms 2019). Panksepp’s emotional needs are potent sources of prediction error because of their high position in the brain’s need-hierarchy (Connolly 2018). The cortical ego serves the secondary process of binding the ascending prediction error and forming predictions to meet these needs (Kaplan-Solms and Solms 2002; Solms 2019). For example, the medial prefrontal cortex (a neocortical area) is involved in the suppression of activity of subcortical affective circuits. It is argued that the cortical ego’s top-down suppression of subcortical id-affects represents the predictive resolution 9 of prediction errors (Carhart-Harris and Friston 2010). By learning how to meet these needs, the ego can better reduce the free energy of the id. Table 1 summarizes Solms’s incorporation of the free energy principle.
Comparison of concepts in computational neuroscience (free energy principle), brain structure/function, and Freudian theory
In sum, the cortical ego generates a predictive model that aims to satisfy socioemotional needs (i.e., resolve prediction error). Affects generated in the id are felt as prediction error and are surprising. Surprise here does not refer only to surprise as a “sensory affect” (Panksepp 1998)—it is “surprise” as what is not predicted, what is unexplained. Optimization of predictions increases evidence for the ego as agent by minimizing free energy associated with the external world and the internal body.
As mentioned above, the ego has different kinds of prediction at its disposal, the two major classes being declarative and nondeclarative. Declarative predictions are representational, including perception, mental imagery, thinking, episodic and semantic memory, and so on. They are accessible to the self-reflective ego, as their representational content can be brought to mind and thought through (i.e., working memory). By contrast, nondeclarative predictions (e.g., procedural action plans and emotional memories like fear-conditioning) are outside self-reflective control and operate in an automatized fashion. Nondeclarative predictions are nonrepresentational, outside of declarative thought and devoid of semantic meaning. They have high precision relative to declarative predictions and are therefore difficult to change. Correspondingly, declarative predictions are more malleable (i.e., have greater neuroplasticity and lower precision) compared to nondeclarative predictions (Solms 2019).
Just as prediction error drives the updating of the predictive model, affective consciousness drives changes in the cortical ego. This is how Solms (2014) interprets Freud’s statement that “consciousness arises instead of a memory-trace” (Freud 1920, p. 25). If affective consciousness is salient prediction error, then the activity of affective consciousness drives changes in memory-traces (i.e., predictions), a process called “reconsolidation.” As Solms notes, Freud’s statement can be interpreted literally. The arousal of affective consciousness literally dissolves the memory-traces of the prediction associated with the error (Solms 2015b). Affective consciousness arousal dissolves memory-traces, allowing new traces to form. In other words, prediction error allows the updating of predictions by removing old predictions that were insufficient to resolve prediction errors. Neuroscientific research in the field of reconsolidation points to the necessity of surprise (i.e., prediction error) in this process (Beckers and Kindt 2017; Schroyens, Beckers, and Kindt 2017).
Repression: Premature Automatization
Against criticism that this cognitive unconscious (i.e., procedural habits) has nothing to do with the psychoanalytic “dynamic unconscious” (i.e., the repressed), Solms (2017) posits “premature automatization.” He points out that self-reflective consciousness (i.e., working memory, the capacity to actively think and hold items in mind) is a mental space for prediction formation and problem solving. The infant, however, is faced with many problems it cannot solve, especially given its many instinctual needs and limited working memory capacity. Solms proposes that the infant, rather than sit endlessly trying to solve insoluble problems, “cuts its losses” and automatizes predictions that do not work. For Solms, the repressed corresponds to these “prematurely” automatized action plans. Because these solutions do not satisfy emotional needs, they produce prediction error (i.e., affect) when they are repeated. Patients suffer from these affects, the consequences of their prematurely automatized (nondeclarative, unconscious) predictions that fail to meet their emotional needs.
Jouissance Is Surplus Prediction Error
I propose to dynamically localize jouissance to the surplus prediction error in neural functions, 10 most notably the salience indicated by affective consciousness (see Table 2). Affective consciousness has properties of the Lacanian real (Dall’Aglio 2019). It is nonrepresentational in itself. That is, when viewed in its raw form, unmodulated by neocortical cognitive and self-reflective consciousness (e.g., hydranencephaly), affective consciousness signals raw emotional vicissitudes with no representational binding. The prediction errors of affective consciousness are surprising insofar as they fall outside of representation and action (i.e., predictions)—that is, they are not expected or explained. And yet, affective consciousness is at the core of all consciousness and motivates predictive work—just as the real, as the extimate center, motivates signification. As a foreign tension from the perspective of the predictive model, affective consciousness is a “fount of jouissance.” It is an excess outside of, yet extimately riveted to, the predictive agent.
Comparison of Mark Solms’s neuropsychoanalytic model of the free energy principle and the Lacanian metapsychology of the real of jouissance
Importantly, jouissance is not simply equivalent to all affect (see part 1). Prediction error that arises in the normal flux of homeostasis (e.g., Mommy left, I feel PANIC, but I cried and got her back, so the PANIC is gone) is not jouissance. Rather, jouissance refers to the surplus or residual prediction error that arises beyond predictive capacities. These prediction errors have a particularly salient and intense quality to them—they arise at the failure of the predictive model to achieve homeostasis.
To be clear, I am not reducing jouissance to surplus prediction error. Rather, when the brain is viewed at the level of neural centers, residual prediction error (i.e., excess affective consciousness) may be considered a neural correlate of jouissance. I believe this level of analysis (e.g., instead of ion-channels) is most suitable for psychoanalytic meta-neuropsychology (see Dall’Aglio 2019).
A Neuropsychoanalytic Lens on Lacanian Theory
Viewing jouissance as surplus prediction error (i.e., excess affect) gives me an opportunity to suggest something Lacanian theory might gain from dialogue with neuropsychoanalysis. 11 Panksepp’s seven emotional instincts (1998) are well established as universally present across mammals. However, Lacanians sometimes dismiss a plethora of emotions (and instincts), instead privileging anxiety (e.g., Fink 2019) and drive. Neuropsychoanalysis would challenge this position. First, there are at least two types of anxiety: PANIC and FEAR. Second, and more important, all seven instincts are equally significant within the brain. Indeed, the lack of focus on emotions is a general criticism of Lacanian theory (Bernardi and de Bernardi 2019).
I do not suggest that Lacanians simply submit and equalize all affects—clinical work remains the final “court of appeals” (Solms 2013). However, I propose that Lacanians consider the following. Recall my discussion in part 1 on instincts being “hooked” into the logic of excess. Affect that arises in this circuit deserves the name jouissance—it is a surplus enjoyment beyond the subject’s homeostatic capacities.
As I argued in part 1, a Lacanian neuropsychoanalysis would suggest that Panksepp’s instincts are vulnerable to such hooking. Premature automatization is an excellent example of this susceptibility. Here instinctual aims have been directed toward something that repeatedly ruptures homeostasis, producing surplus affect (i.e., jouissance). Instinct has drifted beyond the logic of homeostatic need and entered the circuit of the drive. The Lacanian emphasis on anxiety might thus be formalized as a focus on “surplus prediction error” that could emerge with any of Panksepp’s instincts 12 (e.g., surplus RAGE, surplus PLAY, surplus CARE, surplus FEAR).
One might reply that any excessive affect takes on an anxiogenic quality. This may be the case, and it would have implications for how neuropsychoanalysis conceives these instincts when they transgress their homeostatic limits. For instance, surplus CARE might no longer appear as CARE in a straightforward fashion; it may be “misfelt” (Johnston 2013b). However, it would be equally fruitful for Lacanians to consider what new theoretical work could emerge by considering jouissance as surplus prediction error along these many axes.
Are Predictions Signifiers?
If surplus prediction error corresponds to jouissance, are predictions then signifiers? If predictions are understood as neural traces (Solms 2015b), one might be keen to reach this conclusion (see, e.g., Ansermet and Magistretti 2007). Further, predictions come from the social world (Holmes and Nolte 2019). Although we are born with rudimentary predictions, the bulk of our predictive model must be developed via social interaction 13 (the “de-naturalized” brain). Even at the most basic level of predicting our interoceptive signals, our predictions are largely shaped by our primary caregivers (Fotopoulou and Tsakiris 2017).
The picture is more complicated, however, particularly at the level of neural centers and broader brain regions. One should distinguish between traces (i.e., predictions) at different levels of the brain (Dall’Aglio 2019). It is beyond my scope here to discuss the role of different types of predictions 14 for Lacanian theory. Here it is most essential to distinguish between nondeclarative (i.e., nonrepresentational) and declarative (i.e., representational) predictions. For now it suffices to put prediction on the side of the symbolic and surplus prediction error on the side of jouissance, the real.
Insights From The Motoric Unconscious
Jouissance, Dopamine, and Motor Tension
To develop this link between surplus prediction error and jouissance, it will be fruitful to engage with another neuropsychoanalytic model of jouissance, one proposed by Ariane Bazan and Sandrine Detandt. This venture will facilitate a more detailed sketching of the relationship between the real and the symbolic in the brain (Dall’Aglio 2019).
Bazan and Detandt (2013) localize jouissance to the mesolimbic mesocortical dopaminergic SEEKING system. Importantly, the SEEKING system is objectless. It is a motivational system promoting engagement with the world, thriving on excitation and uncertainty (Solms 2012). Solms discusses this paradoxical instinct by describing its innate prediction: “engagement with a source of uncertainty provides maximal opportunities to resolve that uncertainty” (Solms 2019, n. 41).
As Bazan and Detandt (2015) note, SEEKING is active with respect to surprising stimuli, whether the valence is positive or negative. They go a step further to emphasize arousal over valence; the experience of pleasure and unpleasure is secondary to a more rudimentary, structurally excessive (i.e., traumatic, unbound) somatic arousal. 15 Phasic dopamine spikes triggered by surprise in the SEEKING system are highlighted as quintessential instances of this arousal.
These dopamine spikes “mark” the contingent (and historical) instance of an unexpected, surprising encounter (Van de Vijver, Bazan, and Detandt 2017). Because there are no intrinsic action plans (i.e., predictions) associated with SEEKING activity besides the generalized “engage with uncertainty,” this mark is necessary to associate learned motor plans (and, later, more complex responses) to arousal. These marked motor plans are granted “incentive sensitization,” the value given to motor and representational traces associated with the dopamine spike (Robinson and Berridge 1993). Sensitization of these motor plans fills them with tension tending toward discharge (i.e., action execution) whose very excitation is rewarding. Their execution is deeply associated SEEKING arousal due to their capacity to “bind” the tension. Indeed, they are repeated whenever a similar need-arousal takes place (Bazan and Detandt 2013), much like the automatized action plans discussed above.
For Bazan (2012), these action plans always retain some baseline reserve of tension because not all actions are completely executed (e.g., motor incoordination, cortical inhibition). From the inhibition of motor plans, more complex motor representations and mental imagery arise (Jeannerod 1994). Thus, Bazan (2011) gives the signifier a neural basis as a motor articulatory pattern, 16 emphasizing (and shedding light on) the materiality of the signifier. Bazan and Detandt (2013) dynamically localize jouissance to this motor tension in these “marked” action programs (i.e., signifiers) associated with SEEKING arousal. 17
In my view, the SEEKING excitation bound to these marked action patterns corresponds with surplus prediction error—it is a surplus excitation, not tension-reduction. By localizing jouissance to surplus prediction error (and not just SEEKING), I suggest extending the dopaminergic marking dynamics described by Bazan and colleagues to any excess prediction error (e.g., other instincts, interoceptive events). Thus, mechanisms of neural sensitization (beyond the dopaminergic) may underlie jouissance or different features of jouissance (e.g., the “kindling hypothesis” [Dimitriadis 2017] or “central sensitization” in psychosomatic disorders [Woolf 2011]). This maintains a broader view of the real within the brain (Dall’Aglio 2019), fitting with neuropsychoanalytic “dynamic localization” (Kaplan-Solms and Solms 2002). Indeed, jouissance characterizes the entire body—for Lacan (1972–1973), the body is a body of jouissance. Because of its role in SEEKING, however, dopamine activity may underlie the excitatory enjoyment of jouissance. As Bazan and Detandt (2013) note, the dopaminergic marking mechanism is considered one suitable neural substrate of jouissance, not its exclusive substrate. Future neuropsychoanalytic research should inquire whether such dynamics (i.e., jouissance bound to motor patterns) require an intersection with SEEKING or whether they can occur without SEEKING. 18
The Disjuncture between Real and Symbolic
By highlighting jouissance at the interface between arousal (i.e., SEEKING) and action plans (i.e., motor predictions), Bazan and Detandt (2013, 2015) focus on the disjuncture between internal bodily arousal and action that engages with the external world. I propose that we read this as a site of deadlock (i.e., impossibility) between the real and the symbolic in the brain (Dall’Aglio 2019). Let us look at this disjuncture more closely.
Bazan and Detandt (2013) describe two “axes” of jouissance: (1) bodily arousal; (2) historical, contingently marked action plans filled with tension. 19 In our context here, bodily arousal refers to salient affective consciousness, and historically marked action plans refer to predictions (nondeclarative action plans) that bind ascending affective consciousness. This helps us distinguish between two dimensions of jouissance outlined in part 1: traumatic, unbound jouissance (das Ding), and bound jouissance.
On the historical (symbolic) side, the marked, tension-filled action plans correspond to the “sticking” of jouissance to signifiers (Zupančič 2017). Any act (and consequent binding) is better than remaining helpless before the pressure of the drive (Bazan and Detandt 2015; Solms 2017). These marked motor programs are thus nondeclarative signifiers that bind jouissance—they are “jouissance-filled,” engendering their own excitatory enjoyment (i.e., incentive sensitization), which drives continuous pressure for repetition. This is the neuropsychoanalytic perspective of jouissance being intertwined with signifiers. Insofar as there is some original, raw encounter with the arousal of affective consciousness (prior to any marked motor plans), such an encounter might approximate das Ding, the traumatic real that demands metabolization of jouissance via the symbolic.
One can also discern the two types of repetition (see part 1) at this disjuncture between SEEKING and motor traces. Insofar as the motor traces repeat in order to reproduce an historical (i.e., symbolic) event (e.g., a previously executed action plan), there is repetition as commonly understood—the mechanical repetition of this or that past trace that is jouissance-filled. On the other hand, as a dynamic interaction—that is, the repetition of the disjuncture itself between the jouissance-filled motor traces (i.e., historical, symbolic) and the arousal of prediction error (i.e., SEEKING, real)—there may be repetition that demands the new. Insofar as the dynamic “marking” of the surprising encounter is repeated, such a repetition motivates the creation of new predictions. Importantly, there is enjoyment in both cases: one concerns a repetitive surplus (i.e., jouissance-filled traces), whereas the other confronts a more radical excess.
These two axes—jouissance-filled motor traces and traumatic, unbound jouissance—and their dynamic interaction elaborate a crucial dynamic localization of the disjuncture between the real and the symbolic in the brain, the negativity of symbolic failure that demands the new. With traumatic jouissance (i.e., das Ding) on the side of the encounter with affective consciousness and bound jouissance on the side of nondeclarative motor plans, one has highlighted a neural parallel where jouissance (i.e., surplus prediction error, SEEKING) is linked to the symbolic (i.e., marked prediction). In addition, one has delineated the traumatic real and the primordial attempt of the symbolic to bind the tension. Moreover, the operation of these dynamics in response to surprise and excess excitation points to the problematic of neural systems drifting away from homeostatic need and instead grappling with the problem of the drive.
Why does prediction fail to reach homeostasis? Situating the real within the brain
It is notable that the dynamics described above operate at the point of predictive failure—that is, the point of surprising surplus, where there is no prediction that a priori is prepared to resolve the prediction error. There is an underlying question here: Why do humans not achieve homeostasis? Why does our predictive system fail to eliminate prediction error? Why are we left with this surplus prediction error that I call jouissance? In other words, how should one situate negativity (i.e., the real) in the brain, insofar as negativity is the name for the point of failure of symbolic, predictive homeostasis?
Lenses of Disjuncture
One approach is to reflect, quite simply, that life is complicated. The external world is always changing, as are our visceral states. Because the brain selectively samples from the external world and the internal body via its sense organs, its view of the complex environment is incomplete. Prediction can never fully keep up with the changing environment, always leaving some prediction error.
Via Freud, however, the problem runs deeper. It is not civilization (i.e., the external world) that causes our discontents, but rather the inevitable conflict of the drives (Johnston 2005). In fact, these intrapsychic conflicts reverberate in civilization. Moreover, as I argue in part 1, drive itself has a self-defeating nature: “Sometimes one seems to perceive that it is not only the pressure of civilization but something in the nature of the [sexual] function itself which denies us full satisfaction and urges us along other paths” (Freud 1930, p. 105).
Solms (2020) posits certain “structural antagonisms,” though he does not use the term. For example, he highlights intraneural conflicts, such as the inevitable conflict between Panksepp’s instincts. For example, attachment needs demand secure safety, yet SEEKING demands novelty. Whereas the predictive model aims at eliminating prediction error, SEEKING seeks the increase of prediction error. Even if the innate prediction of SEEKING is to engage with uncertainty to reduce uncertainty, the fact that one SEEKs uncertainty is at odds with the homeostatic goal of prediction.
Another Solmsian disjuncture is the contrast between declarative and nondeclarative predictions (Solms 2018). Nondeclarative predictions are by definition inaccessible to declarative mechanisms. Some part of the predictive system is necessarily alien to the rest of prediction. The inevitability of premature automatization—which, for Solms, forever leaves homeostasis at risk for imbalance—leans toward the “structural” nature of predictive imperfection.
As I have noted in discussing SEEKING and marked motor traces, there is also a disjuncture between affective consciousness and the motor traces called on to bind the tension. In need-satisfaction, the motor plan is tagged with incentive sensitization, becoming jouissance-filled and engendering surplus excitation. This is a possible mechanism by which instincts are hooked into SEEKING excess (see part 1), diverging from homeostatic goals and aiming at the repetition of motor traces.
Moreover, as Bazan and Detandt (2013) argue, this excess (i.e., dopamine spike, motor incentive sensitization) is necessary due to the lack of attunement between vegetative, visceral systems (i.e., those generating bodily needs; the “invertebrate body”) and the musculoskeletal motor system (i.e., the “vertebrate body”) that must act in the world to satisfy those needs. In a Lacanian vein, they emphasize that there is no predetermined guarantee or knowledge of which particular action will meet the need. Thus, the surprising enjoyment (i.e., dopamine spike) is necessary to associate the adequate action with need-satisfaction. Importantly, this disjuncture is not unique to humans—it is likely present in all vertebrates. 20 In a tone resonating with the “coincidence” of negativity and representation (see part 1), Bazan (2011) speaks of the failure of the motor system to completely eliminate the tension—a “shortfall of action”—which is at the basis of representation. Disjuncture has a productive aspect.
Johnston (2019) suggests that this bodily disharmony is most pronounced in the human brain, as evident in the disharmonious relationships among neural systems. Developing insights from neuroscientists like David Linden and Antonio Damasio, Johnston (2013a) emphasizes how evolutionarily ancient brain systems associated with affective consciousness (e.g., the brainstem) sit together in an uncomfortable “kludge” with evolutionarily recent neocortical (cognitive) systems (Linden 2008). 21 Rather than a unified system called “the Brain,” disparate brain systems developed separately, along different lines.
The “brain” is thus composed of several layers of “sedimentation,” with the thalamus—to use Damasio’s terms (2010)—as a “marriage broker” between an “odd couple”: the primitive, nondeclarative brainstem and the representational neocortex (Johnston 2019). Despite their radically opposed functions, these systems are forced to interact to facilitate survival. Thus, the brain’s “natural” (i.e., structural) state is one of incongruity. In Lacanian fashion, Johnston insists “there is no intracerebral relationship” (Johnston 2013a, p. 59). In other words, a formula for intracerebral unity and harmony among brain systems does not exist. There is a “natural” disjuncture (Dall’Aglio 2019).
Moreover, this “natural” incongruity is not simply a need-deficiency remedied in biological terms. As discussed at the beginning of this paper, the premature helplessness of the infant—mirrored in the “kludge-y” and genetically open nature of the brain—prompts the turn to a “denaturalizing” symbolic-imaginary reality (i.e., speech, images, etc.). In this way, the structural negativity of the brain (here manifest as the disharmonious relationships among brain systems) has a causal role in the turn to the de-naturalizing social world (Johnston 2019).
All these approaches (i.e., conflict between instincts; disjuncture between declarative and nondeclarative; disjuncture between predictions and affective consciousness) are valid neuropsychoanalytic conceptions of structural negativity (i.e., the real) within the brain (Dall’Aglio 2019). With the present integration of predictive coding with Lacanian theory, I venture to suggest another lens of disjuncture.
The Originally Lost Prediction
I have argued that surplus prediction error corresponds to jouissance and that predictions correspond to signifiers. 22 Recall, from part 1, that the “originally lost signifier” is one way to name the structural negativity (i.e., the real) within the symbolic (Zupančič 2017). In Lacanian fashion, one might invent the phrase “originally lost prediction” to designate the negativity within prediction.
Allow me to be precise. This formulation (the “originally lost prediction”) serves not only to map neuroscience in Lacanian terms. 23 It also allows us to link the disjunctive nature of the brain to the enjoyment of surplus prediction error (i.e., jouissance). Zupančič’s originally lost signifier (2017) designates the real firmly within the symbolic (see part 1). The signifiers that demarcate the contours of this negativity are those to which jouissance (arising at the negativity) sticks. To put the matter simply, negativity is the site of enjoyment.
In terms of predictive coding, the originally lost prediction is the (real) point within the (symbolic) predictive model that is the limit of prediction, the point where prediction fails. The failure to achieve homeostasis engenders surplus prediction error. Therefore, the originally lost prediction is the point of surprise, unexpectedness, and the unknown within prediction. Such uncertainty is precisely what excites SEEKING, which, as Bazan and colleagues have shown, “sticks” to the motor traces on the precipice of the predictive model (i.e., incentive sensitization). The predictions that fail to overcome this disjuncture are those which are jouissance-filled.
The “originally lost prediction” is a way of emphasizing the structural (“original”) disjuncture within prediction where uncertainty (“lost prediction”) may be marked. This contrasts with a view of two opposing entities with a gap between them (Johnston 2005)—the gap is within prediction. For example, consider how the free energy principle does not simply oppose free energy to prediction. This is observable in the predictive model’s calculation of expected free energy (Fotopoulou 2013; Parr and Friston 2018). Recall that a prediction is a probabilistic inference of the cause of a sensory input. It is probabilistic, not absolute (even if the self-conscious ego feels sure of its reality). While the expected prediction error and the encountered prediction error may differ, this computational operation points to the fact that prediction is not an absolute resolution of prediction error. What the brain takes to be its reality (i.e., its predictive model of the world) is itself missing something, even though this “reality” may be experienced as a phenomenological whole. There is an uncertainty within prediction.
I propose that the expected free energy calculation is not simply a cold expectation of error. It is meaningful, betraying a limit within prediction. Indeed, the incompleteness (i.e., negativity, the real) within prediction (perhaps indicated by expected prediction error) is the necessary structure of prediction that allows surplus prediction error to drive changes in the predictive model. Negativity is the well of creativity and the potential for predictions to slip from their “natural” pre-givens.
It is for this reason that I characterize the failed predictions—those at the precipice of the predictive model, beyond which is the chasm of the real—as jouissance-filled. Even though they “predict” affective consciousness, they do not totally remove prediction error; they are doomed to endlessly repeat. They are the signifiers that mark the contours of the real within the symbolic. Moreover, they themselves bind and engender jouissance, the incentive sensitization that grants them an exciting, rewarding quality beyond their homeostatic aim.
Thus, the predictions that demarcate the failure (i.e., the originally lost prediction) within prediction are paradoxically the ones that are enjoyed. In other words, the particular way we fail—the motor plans that mark the point of unpredicted surprise—is our particular way of deriving jouissance.
I make this Lacanian reading in the spirit of neuropsychoanalytic dialogue. It is a suggestion intended to stimulate discussion and engagement. To illustrate precisely how the negativity within prediction is the site of enjoyment, I will discuss the child who famously uttered “ooo-aaa,” fort-da.
Beyond The Free Energy Principle
On the paradox of a death drive that violates the pleasure principle, Freud (1920) discusses a game invented by a child (presumably his grandson). He describes the child, age one and a half, as “not at all precocious” but on “good terms” with his parents and well behaved; “he never cried when his mother left him for a few hours,” but he was “greatly attached to his mother” (p. 14). The child had a “disturbing habit” of throwing away small objects while saying “ooo,” which Freud interprets as fort [“gone”] (p. 14). This expression was accompanied by “interest and satisfaction,” although Freud attributes “greater pleasure” to a second act whereby the child sometimes pulled the toy back, uttering da [“there”] (pp. 14, 15).
One might suppose that the child has a secure attachment, given his bond with the mother and his good behavior on her disappearance. Nevertheless, separation from the mother would still arouse PANIC and the corresponding prediction to SEEK the lost mother. One might suggest that the child had learned a “secure attachment” prediction: “Mommy will return, I just need to wait.” But this would still arouse some quota of tension associated with the emergence of the gap.
Yet the child does not just sit there with the tension, nor does he cry. Instead he invents a game. Whereas Freud speculates an attempt at mastery over the mother’s disappearance, Lacan (1964) uses this scenario to illustrate the emergence of objet a (i.e., the toys thrown around). Recall that objet a is the materialization of negativity, an excessive presence in the representational world. It is the object of the drive, the ungraspable excess whose repetitive use allows the subject to derive jouissance (see part 1).
When faced with the traumatic disappearance of the mother—the confrontation with negativity—the child invents a game by playing with objects, metabolizing the drive arousal at the site of the gap. Freud’s diction is fortuitous: although there was “greater pleasure” attached to the da, Freud recognizes the “interest and satisfaction taken in” the fort (pp. 15, 14). While da might stand in for the return of the mother, which would reduce PANIC-tension, fort marks the “interest and satisfaction,” an excitatory enjoyment—that is, jouissance. Fort marks a contour of the real: “gone” is a signifier that stands in for what is absent. That the fort (i.e., throwing toys away) was repeated more than the da (i.e., their return) highlights the independence of jouissance from tension-reduction (Lacan 1964).
Here we have an aberration of PANIC. The predictions concerning fort do not aim at the mother’s return or a substitute attachment. Rather than tolerate the surplus jouissance of PANIC, the child puts this prediction error to use. He invents a game which, if we are to agree with Freud that the object stands in for the mother, might increase PANIC prediction error. In this way, PANIC has deviated from its “natural predictive course” and has been hooked into the logic of drive by targeting objet a. Chiefly, rather than this aberration being cause for discontent, the child gets his enjoyment—he does not cry and takes great joy from the game. In other words, the child found a way to enjoy his jouissance (Fink 2011).
This example illustrates how the real—as the failure of homeostasis (i.e., failure to reunite with the mother) and as cause for the slippage of prediction (i.e., the child’s focusing instead on the objet)—is the site of both creativity and jouissance. One must create when facing the surplus prediction error at the precipice of one’s predictions. It is the child’s creation of a game that allows him to overcome the negativity of das Ding and instead enjoy the use of objet a.
One might go a step further and suggest a hooking of PANIC into PLAY—the child PLAYs a game of fort-da. In this sense, PLAY is also hooked into the logic of the drive, deviating from its “natural” social prediction (i.e., cooperative give-and-take with others, mastery and submission). By engaging with the prediction error at the site of negativity (with all the slippages of instinctual predictions it entails), the child invents a way to enjoy his jouissance. Neuroscientific evidence is not foreign to this formulation, for “PLAY celebrates the joy of surprise” via its neuro-functional intersection with SEEKING circuitry (Kellman and Radwan 2019).
A Time For Understanding: Jouissance Is Surplus Prediction Error
I have argued that surplus prediction error corresponds to jouissance, and that predictions fall on the side of the symbolic. Specifically, I have focused on the surplus prediction errors signaled by affective consciousness and the primordial binding of this arousal by nondeclarative predictions (i.e., action plans) in the (prematurely) automatized unconscious. Jouissance is an irremovable, nonrepresentational excess within the symbolic, just as surplus prediction error is an irremovable excess within the brain’s predictive model. Surplus prediction error is what is left unexplained and unpredicted, yet simultaneously motivates the entire predictive machinery. As affective consciousness, surplus prediction error is the extimate, foreign core of the subject—a fount of jouissance.
There are several ways to formulate the real within the brain. Based on the link between jouissance and predictive coding, I have suggested a novel formulation: the originally lost prediction. This formulation emphasizes the extimate nature of predictive failure (the real within the symbolic), casting light on the capacity for predictions to slip, drift, and be subjected to aberration. Crucially, the signifiers (i.e., predictions) that mark the contours of the real (i.e., predictive failure) are those that are jouissance-filled. In other words, the particular way we fail is our mode of enjoyment.
This Lacanian neuropsychoanalytic model makes sex (i.e., the real, jouissance) central to the neuropsychoanalytic model of the mind. In part 1 of this three-part series, I detailed the metapsychology of jouissance. Here, in part 2, I have laid out my arguments for why jouissance is surplus prediction error and, correspondingly, why there is an originally lost prediction (i.e., a structural negativity within the brain). In part 3 I will suggest how Lacanian clinical technique can shed light on what we are to do with prediction error, if eliminating it is impossible.
Footnotes
Department of Psychology, McAnulty College and Graduate School of Liberal Arts, Duquesne University.
Submitted for publication January 2, 2020.
1
For more extensive discussions defending Lacanian neuropsychoanalysis from criticisms of bio-reductionism, see Dall’Aglio (2019, 2020) and 
.
2
For a more general discussion of the problem of associating the psychoanalytic mind with the brain, see Blass and Carmeli (2007) and .
3
4
One should not understand prediction (especially perception) as the simple absorption of sensory input. Prediction requires the active engagement of the brain in anticipating its experience and adjusting its internal models. Such mechanisms involve motoric brain functions (e.g., cognitive control). Thus, motor logic applies to both action and perception. The relevance of this point will be seen when I come to discuss the motoric unconscious.
5
I will use the terms precision and salience interchangeably. Precision is also applied to predictions. When so applied, it refers to confidence in the prediction—that is, how likely it is to explain the ascending prediction error. More precise predictions more efficaciously explain inputs. When applied to prediction errors, precision (better understood now as salience) refers to the importance of the prediction error. These two applications converge in the following way: the violation of precise (i.e., confident) predictions arouses salient prediction errors.
6
One might argue that this localization suggests a “housing” in neural structures, as opposed to a dynamic between structures. See for an extended discussion of this issue. I suggest that one should read Solms’s localizations as dynamic localizations to the multiple systems within the cortex or brainstem/limbic system, both of which are considerably heterogeneous.
7
I am focusing on Panksepp’s affective neuroscience because it is the most prevalent in contemporary neuropsychoanalysis. One should note, however, that subcortical affective consciousness is not agreed-upon in the neurosciences. , for example, argue that Panksepp’s instincts are nonconscious and that the “emotional” quality derives from cortical interpretations of this subcortical nonconscious arousal.
8
The all-capitals style is Panksepp’s convention for a formal taxonomy of these neural circuits.
9
10
It is important to remember that this does not exhaust our understanding of jouissance. The concept may still be applied at other levels of analysis, such as race (Miller 2017), sex (Zupančič 2017), and politics ().
12
Insofar as all instincts are warped by this drift toward drive (and, as elaborated below, necessarily involve some excess tension on the motor axis), one might make a Lacanian neuropsychoanalytic pronouncement: “The Instinct does not exist.”
14
One reason the present discussion of predictions as signifiers is insufficient is that it does not differentiate nondeclarative predictions and prediction error, both of which have elements of the real (see Dall’Aglio 2019); nor does it distinguish imaginary and symbolic predictions. For example, language (a system of predictions) involves motor articulation (symbolic) and meaning (imaginary). Yet Lacan also speaks of the “real” aspect of the signifier (i.e., the “letter”). Capturing these different dimensions of prediction would require a more complex dissection of the Lacanian registers, including how the real “refracts” in all three registers: imaginary real (e.g., phenomenal experience of emptiness and nothingness), symbolic real (i.e., purely empty signifiers and concepts; mathematical formulae), and real real (i.e., absolute negativity; Johnston’s corporeal negativity) (Johnston 2019; ).
15
I claim that this arousal is still affective, but in the sense of the unbound jouissance of das Ding, a “primary affect” that operates beyond the symbolic pleasure principle (; see part 1). This is one way to approach the difference between Panksepp and LeDoux. Panksepp’s raw affective consciousness is neither straightforwardly emotional nor simply nonconscious—it is a traumatic, unbound jouissance. Cortical interpretation (prediction) tames this arousal into standard emotion.
16
17
Importantly, by localizing jouissance to incentive-sensitized motor programs, any instinctual action is necessarily marked by this excess. While this may elide any difference between drive and instinct (insofar as the motor end is necessarily marked by jouissance), it would then be important to emphasize Lacan’s distinction (1972–1973) between phallic jouissance (i.e., enjoyment within the law of phallic signification) and other jouissance (i.e., enjoyment not-all within phallic signification). This topic is beyond the scope of this three-paper series.
18
See for a case of bilateral globus pallidus lesions resulting in a loss of SEEKING engagement but then, after engagement with other instinctual systems, a reengagement with the world. In such cases, we should explore whether there has been neuroplastic recovery within SEEKING or whether alternative systems are capable of such redirection.
19
20
Thus, animals may be posited to have some sort of jouissance—another point where neuropsychoanalysis can pose a challenge to Lacanian theory.
23
Although bridging neuroscience with notoriously antinaturalist Lacanian theory is no laughing matter.