Chapter 8: The Brain as an Organ of Re-Disposedness, not an Organ of Pre-Diction

Chapter 8: The Brain as an Organ of Re-Disposedness, not an Organ of Pre-Diction

Predictive processing can be a misnomer, and potentially contributes to creating a category mistake. This category mistake is to claim that “the brain is an organ of prediction.” This chapter argues that this claim is a reification which confuses models with reality. 

A brain is not a specially-powered cognitive organ for priori prediction, but an embodied organ of posteriori re-disposition. Brains do not merit a special status of organ dualism. Brains alone do not generate action, and experience does not happen in brains. Brains (and the rest of the nervous system) are neurological substrates that attune organisms and environments via physical reorganization following reiterated historical coupling. Nervous systems are reactively plastic, governing and mediative substrates. Nervous systems act as enabling and governing constraints for holistic brain-body-environment couplings. As constraints, nervous systems enable work cycles even as these work cycles enable the continuous production of constraints (as in constraint task closure), including the metabolic turnover of the nervous system itself. As an abstracted organ, and as constituent within a whole organism, brains exist for and by means of the whole. 

Nervous systems act as governing or enabling constraints relative to the specifically situated eco-evo-devo histories and memory of the organism-environment system. The nervous system, body, its historized memory, and its co-evolving environments are dynamically co-plastic over these various time scales. In some ways the constraining nervous tissues will be enabling to open more possibilities, in other ways governing to constrict possibilities. The embodiment of such constraints relate to different neuroplastic connective partnerships, and relate morphological changes in anatomical, physiological integration and differentiation. Experience and cognition are holistic and happen with a brain, not in a brain. The nervous system is concrescent within an operational closure and does not stand alone like a module; experience and cognition are likewise not reducible nor localizable to the nervous system. 

Not only are nervous systems constraints upon behavior, but loaded behaviors are constraining upon nervous system states (Raja & Anderson, 2020). Behaviors can act as enabling constraints to nervous system processes, and nervous systems can act as governing constraints to behavioral processes. Together nervous systems and behavioral processes obtain operational constraint closure. Constraining, enabling and governing interactions between nervous systems and personal-level behaviors are causally circular, enabling and governing, dynamically co-emergent, lay paths of memory with and in each other, and are co-adaptive as nested complex system heterarchies.

Is the brain an organ of predictive processing or an organ of posteriori redisposedness? This chapter argues that the brain is best ontically described as the latter. While predictive processing and Bayesian modeling can be useful and pragmatically powerful models, models should not be confused with reality as reifications. Humans use models as tools to predict; the brain is not a tool, because it is not heteronomous to the organism. Brains do not predict to their organism, even if brains enable an organism to pre-reflectively anticipate reiterated situations in its dispositions. Such anticipation is due to posteriori redisposedness. That is, via memory assimilation due to an adaptive hysteresis effect, and co-extensive effects upon a stably reiterated environmental niche. Neural assimilation can be via processes of neural plasticity, neural reuse, and interactive differentiation-and-search of adaptive connective partnerships (Anderson, 2014). Anticipation is not due to a brain literally predicting and anticipating a world to an organism; this cognitivist conception of predictive processing can be a model, but is not the ontical case of reality. Cognitivist prediction requires the brain to be a heteronomous tool to an organism, with special priori powers of expectation, modeling, planning, and acting upon plans. This leads to homuncular regress and organ dualism; what or whom is a brain predicting to? Such cognitive activities of predicting, planning and representing emerge on a personal level (Chapter 14), but are not activities of the neural substrate itself. 

Instead of being labeled an “organ of prediction,” the brain can be considered a substrate that governs and mediates behaviors. Cognitivist “prediction” based in operations upon stored information, data, knowledge or beliefs (as in “good old fashioned action theory” or GOFAT) is replaced with a concept of embodied “posteriori re-disposedness'' of the governing substrates’ organization. This does not happen via cognitivist “sense-model-plan-act” prospection based in information processing, but via a hysteresis effect. This form of memory involves a temporally-delayed historicizing of multi-systemic plastic/reactive governing substrates, including the brain. Adaptive hysteresis effects re-dispose all these enmeshed and entangled autonomous bodies following reiterated experiential and behavioral loading. Types of autonomous bodies span metabolic, sensorimotor, intersubjective and linguistic bodies. These interdependent types of autonomous bodies are abstractions from a concrescent organism’s obligate operational closure (Chapter 10). Applications of adaptive hysteresis are varied, and this process is a substrate-flexible formal activity-structure which stabilizes memory. Adaptive hysteresis is a general activity-structure resulting from the self organization of complex systems as they coevolve with their embedded field. For example, adaptive hysteresis mediates the reattunement of habitus to loaded field behaviors, re-attunement of neural network TALoNs to loaded NRP factors, and genetic networks to loaded natural selection. This reorganization of the reactive, plastic, and governing/mediative substrates mediates a training effect, thereby enabling phenotypic reorganization, re-disposedness, re-disclosedness, training of praktognosia, training of habitus and capital, natural selection (training of a genetic network), neural network training, etc. 

The adaptive hysteresis effect is one way of mediating memory in complex adaptive systems. It involves an embodied re-attunement of phenotypic organization to past behavioral loadings; a “posteriori re-disposedness.” This temporally-delayed assimilation of phenotypic reorganization enables a re-attunement to one’s fluctuating niche. Re-attunement is contingent on the stable reiteration of similar challenges. In a niche construction, such experiences are relatively stable, reloaded, and reiterated over the organism-environment system closure’s coevolutionary history together.

Hysteresis is a general term originating in the science of ferromagnetism, and later applied across many disciplines including engineering, biology, neuroscience, and sociology. In Pierre Bourdieu’s original usage of the term, hysteresis can also involve behavioral displacement due to gross habitus-field mismatch. In Bourdieu’s examples, the gross habitus cannot adapt in time to gross field changes, and the general adaptation of genetic-assimilation has a much higher degree of inertia (habitus cannot re-assimilate gross field changes without significant delay). In this way, hysteresis and memory can be maladaptive, and can lead to inertia if system memory is saturated and/or slow to change and adapt in relation to to field changes. This is memory and hysteresis exemplified as resistance to change, the inertia of habitus assimilation to changes in field, memory as baggage, and memory as anachronism. 

Systems with highly saturated and long-lasting memories can be less adaptive in fluctuating environments. These systems are inertially static and resistant to re-organizational change, especially following loading of new experiences and fading of old experiences. For a complex adaptive system to be maximally adaptive between its habitus and field, memory must have a consonant balance between plasticity, turnover, and an ability to be lost. This balance is contingent on the dimensionalities and degrees of flux and pressures of a changing field. In the most general sense, a highly fluctuating field benefits from a highly co-plastic degree of memory, and a highly stable and unchanging field benefits from static and long lasting memory. 

This characterization of adaptive hysteresis as memory via assimilated posteriori re-disposedness of governing substrates is contrary to the cognitivist knowledge-storage and prospection of GOFAT. This develops an enactive and anti-representationalist account of memory and phenotypic reorganization. Posteriori re-disposedness of habitus (via adaptive hysteresis) is contrasted to the cognitivist “brain is an information processor organ of prediction.” Instead of being called an “organ of prediction,” the brain can be considered a reactive, plastic, governing and meditative substrate of “posteriori redisposedness” over the organism-environment system’s history. 

To build upon the difference between posteriori re-disposedness versus pre-diction is to distinguish an enactive, embodied and embedded conception of memory. This is distinct from the cognitivist and predictive-processing conceptions of memory. In this way, organism memory is not analogous to the process of “data storage to be input into rule-entailed information processing programs, outputting a prediction to be used by the tool-user of the program in order to plan action.” This is the sense-model-plan-act conception of memory as data storage and retrieval, for predictive processing. 

Memory as data storage and retrieval for predictive sense-model-plan-act algorithms is heteronomous. Data storage is memory conceived as a tool, for use by an other, not use for the computer system’s own autonomous goals (for which it has none, being heteropoietic and designed by an other). Organism memory is embodied and embedded, it is part of an autopoietic and autonomic precarious operational closure with its own instrumental norms. Organism memory is hence autonomous. Memory is instantiated processually by redisposing an organism’s governing substrates’ poised stance to act. This enables (not entails) the organism to act skillfully upon afforded opportunities-to-act within a perceptually indeterminate environment. Organism memory is mediated by the process of temporally-delayed (posteriori) assimilation of reorganizational effects, following reiterated behavioral loading with experiential challenges. This entire process is definitive for the adaptive hysteresis effect as a substrate-neutral activity-structure for memory.

Intra-, inter-, and trans- generational closures of autopoiesis and autonomy

An example of intra-organism memory and trans-generational memory can be studied in slime molds. These single-celled colonial organisms can behaviorally move during the history of the single intra-generational organism’s lifespan. The colonial closure can also reconfigure over trans-generational forms of behavioral organization. These trans generational memories enable the induction of reproductive spore forms, motile forms, feeding forms, drought forms, and others. The ecologically loaded demands span time scales beyond the life of a single cell’s closure, and thereby affect the trans-individual form of the colonial closure, as a whole. This trans-organism closure then remembers, moves and transforms in ways that supersede the memories of individuals. This is an exercise of thinking about organizational closures in flexible ways, e.g. on different abstract time scales, evolutionary transitions, orders of individuation, integration and differentiation. 

This is to flexibly think about autonomous bodies as having dimensions spanning and penetrating different time scales and types of closures. Closure may be intra-organism, sub-organism (e.g. organelles, organs, endosymbionts), trans-organism (e.g. food web ecology, reproductive closures), and epi-organism (e.g. viruses, parasites). 

Evolution is a process that spans inter-generational closures. Experiential time scales span intra-generational closures. 

Development spans intra-generational closure. Ecology spans both intra- and trans-generational closures. Sociocultural effects similarly span both intra- and trans-generational closures. 

In a processual-enactive framework, memory involves hysteresis effects from behaviors loading-upon and historicizing their plastic, reactive governing substrates. Hysteresis effects yield posteriori-attuned re-disposedness. This enables the historical attunement, praktognosia, and prereflective embodied “intuitiveness” of an agent’s disclosedness. This reorganization of an organism’s poised stance with its environment enables (not entails) skillful behavioral reorganization when facing historically stable opportunities-to-act in a perceptual indeterminate world.

Memory can enact upon different time scales and orders of evolutionarily transitioned substrate for each type of autonomous body within a concrete organism closure. As operationally defined, organism memory processes can be abstracted on evolutionary, developmental, ecological, and experiential time scales and closures. These span intra-, inter-, and trans-generational closures for an organism. 

Processes of memories are considered in the 4E sense (embodied, enactive, embedded, extensive). These processes historize reactive, plastic and governing substrates of an organism. A temporally delayed effect of posteriori re-attunement follows (i.e. the assimilation of habitus organization following experiences with a field). This effect thereby attunes the embodied organization of an organism’s habitus, poised to act with a field. 

Hysteresis effects thereby do not mediate memory via cognitivist functions of forward-planning pre-dictions, but via embodied and pragmatic posteriori-redisposition. Future planning, prospection, and cognitivist predictions are important phenotypic abilities, but these require further scaffolding. Such behaviors are exemplified in animals such as scrub jays, orangutans, and humans. Reference to prospective predictions as behaviors within the subpersonal neural substrate should be considered as metaphor, but not as literal. For example: hysteresis effects can be seen as metaphorically similar to Bayesian-based predictions via “updating priors,” but still, the temporal direction is inverted. These metaphorical models can be extremely helpful as pragmatic and heuristic tools, but to reify predictive processing models as “how neurons actually operate” is to confuse models with reality. Perhaps it is safest to remain agnostic on the verdict of whether a brain is an “organ of pre-diction” or an “organ of posteriori re-disposedness.” Even so, the temporal direction of these processes seems to be inverted in the former. Neurons cannot model, plan, and act upon the future in the same way that orangutans, jays and humans can, via their behaviorally scaffolded intelligence (note that intelligence will be operationally defined under this framework in Chapter 14). To claim that neurons can non-metaphorically predict is to attribute a personal-level capability to them. Neurons can not act on probabilistic outcomes of predictive models that they generated by an algorithm. To attribute neurons with the behavioral capabilities of prediction is a fallacy of equivocating inter-level emergent behaviors (Chapter 7).

In the processual-enactive framework, memory is not mediated by cognitivist functions of prediction (rule entailed deductions, inductions, logic), but by embodied, enacted posteriori re-disposedness. This involves behaviorally  reorganizing an organism’s poised, attuned stance with its perceptually-indeterminate environmental opportunities to act. Temporally delayed attunement between coupled organism-and-environment follows  historically reiterated behavioral loading upon a reactive, plastic, and governing embodied substrate.

Organism memories are not heteronomous data stored-and-retrieved by cognitive predictive information-processing. Organism memories are embodied following adaptive hysteresis of pragmatically poised dispositional organization. The brain does not deserve a special extra-organism power by appealing to a type of organ-dualism. 

Organism memory is not a process of “pre-diction” in the way of computationalist sense-model-plan-act paradigms. Organism memory invokes a process of “posteriori re-disposedness” for autonomous substrate embodiment.

Conclusions

The brain does not exert Bayesianism and predictive processing upon an organism’s actions as a special organ that has centralized powers over the body. Complex adaptive systems are self organizing and likewise uncontroversially involve decentralized control. Therefore, the claims that “the brain is an organ for prediction” and “the brain is an organ that predicts” are inappropriate. The brain has nothing and no-one to predict to, because the brain is not a centralized controller, the brain is not a heteronomous tool of prediction for the organism, and the brain does not have a special status of organ-dualism within the concrescent organism closure. The brain also cannot literally pre-dict, meaning “to say prior.” The temporal direction is of delayed posteriori re-disposedness of an organism’s poised dispositions in the wake of reiterated experience (i.e. the assimilation of reorganization by the habitus following reiterated experiences solicited by the field). Further, the brain cannot pre-dict nor pre-state an entailed result, instead it enables and poises a bodily disposition to act within a valenced world. 

Predictive processing and Bayesian prediction do not happen within an organism’s autonomous neural networks. Predictive and Bayesian processing are attributes of heteronomous tools that input data storage, condition priors, and output predictive models as information for tool-users to treat as plans for action.

A computer is heteronomous so its means of prediction are heteronomous. This is because it has been engineered by-and-for its coupled tool user. The computer cannot predict by-itself, to-itself, and for-itself.

The brain as an organ does not pre-dict to a severed operator because it is not heteronomous, and the brain as an organ that exists for-and-by-means of the organism does not have autonomy of its own (decoupled from the organism). Organisms modeled as complex adaptive systems do not have a central operator, but instead self-organize decentralized control. There exists no res cogitans ghost in the machine, and the brain does not have privileged status of organ dualism. Autonomous bodies instead use the process of adaptive hysteresis as one way of mediating memory. This remnant effect is assimilated in temporal delay, following reiterated experiential loading, for which the bodily substrates are reactive and plastic. This is similar to models of simulated annealing in complex adaptive systems, but specifically applied to embodied-embedded memory processes.

Heteronomous computers can pre-dict for their users, as they are heteronomous tools given final cause by their autonomous users and engineers. Brains of autonomous and autopoietic organisms cannot be said to pre-dict for the organism since brains do not have a dualistically decoupled operator to predict to. A brain is an organ within the autonomous closure of the organism, existing for-and-by-means of the organism, and the brain is not a heteronomous tool for prediction. Brains cannot be organs that predict, since there is no heteronomous recipient as the predictor, no ghost in the organ. Instead, brains (as but one of many reactive-plastic-mediative substrates embodied within the autonomous organism) assimilate reorganization of their disposition following solicitations-to-act within a field (posteriori re-disposedness via the hysteresis effect). The brain, as an embodied substrate, mediates behavior no-less-bodily than any other sensorimotor bodily substrate. The brain’s plastic assimilation of re-organizational change in the wake of reiterated experiences yields an embodied, practical, dispositional (non-cognitivist) form of memory. The brain’s posteriori-assimilation of experiential loading attunes the poised, practical behavioral organization of the organism. In this way, dispositions to act are attuned to skillfully unfold in the face of future field solicitations-to-act relative to the outcomes of past experiences. Memory processes train the organism’s embodied know-how when facing an open-ended but potentially masterable world.