Chapter 12: Attention and Perception: A Positive Feedback Loop

Chapter 12: Attention and Perception: A Positive Feedback Loop

Memory processes temporalize perceptual processes. Two memory-stabilizing processes include adaptive hysteresis and feedback loops. Chapter 11 emphasized the role of adaptive hysteresis in generating the temporal and embodied, dispositional dimensions of perceptual processes. The present chapter emphasizes the role of positive feedback loops for temporalizing and stabilizing memory and perceptual processes. Feedback loops may organize on different scales of space and time, including the ecological scale and the individual scale. 

Ecological Positive Feedback Loops

This chapter discusses the positive feedback loop between perception and attention. To preface this discussion, another positive feedback loop on a slower timescale can be drawn between the organism and environment closure. 

Processual-enactive memory differs from memory conceived under substance ontology. The adaptive hysteresis effect and feedback loops enable memory on the scale of the holistic organism-environment system closure. These processes differ from memory conceived as data-storage-and-retrieval in a substance, e.g. the hard drive of a computer, or the hippocampus treated like a writeable substance. Organism-environment memory is not statically written and stored like data-engrams in the hard drives of the brain. Organism memory differs from hard drive memory and does not function via heteronomous data storage-and-retrieval as properties of a static substance, to be fed into neuronal information processing algorithms (Chapter 8).

The environment modulates the organism as the organism transduces its environment. In this way, ecological-time scale processes of memory can be seen as co-attunements between the organism’s autonomous bodily substrates (governing for its behavioral dispositions) and the changes left upon the lived-in environment. Organisms and environments continuously change and coevolve with each other to each other’s constraints and pressures as they reiterate couplings. This is relevant for perceptual development. The sensitivity and reactivity to frequented environmental patterns (and opportunities to act) becomes increasingly ingrained within an organism’s prereflective habitus over its reiterated experiences. In a circular way, the lived-in constructed niche becomes increasingly shaped in ways that are readily signified by its inhabitants. Through this circular concrescence, the organism-environment closure obtains co-attunement between the signifiers’ habitus and their signified field. 

On an ecological time scale: organisms shape the accumulation of opportunities-to-act within the environment. In circular return, environments shape an organism's capabilities to act via reiterating common developmental experiences and selective pressures. This ecological-scale positive feedback loop sets the stage to enable attunement on a more rapid sensorimotor timescale. In this way, the sensorimotor-time scale positive feedback loop between attention-and-perception is thrown into a niche that was already laid by the generational paths of phylotypic progenitors. The perception-and-attention loop has a sensitivity that is already highly attunable to the attractors, disturbances, figures and grounds that relate an organism-environment system together, having shared histories. 

Hubert Dreyfus on Merleau-Ponty’s Phenomenology of Perception 

Attention has a genetic role for the phenomenal field, beyond attention’s role as a spotlight. Attention is not a mechanism to reveal a pre-given world; instead, attention is genetic for the perceptual field. Attention uncovers perception as perception enhances perception. This circular positive feedback contributes to enacting a figure-ground enriched phenomenal field. 

As attention moves around, it is an actively genetic spotlight, not just a representationalist/revealing spotlight. 

To clarify some deconstructed word usage: 

To “Signify” involves a dialectic co-generation for signal-noise. 

To “Figure out:” involves dialectic co-generation for figure-ground 

According to Hubert Dreyfus in his archived lectures of Merleau-Ponty, attention has the positive role of “opening up the conceptual world, where everything  is seen as determinate figures, determinate objects with properties.”

The next step in the dialectic tension between discriminating figure and ground is the genetic emergence of a phenomenal field and behavioral field. This new synthesis involves a new tension between signification and opportunities to act. 

Attention opens up to perception as perception enhances attention, as a reinforcing genetic dialectic between the two. This feedback loop is stabilizing for the persistence of a phenomenal field, differentiated into figures and ground, attractors and aversions, announcements against silence. The figuring-out of disturbances and affordances are not background-independent from the ground, but instead continuous in their genetic self organization; figures are continuous with the background even as they stand out from the background. The field is holistic in this tensioned discrimination, and tensioned figures and grounds remain pragmatically determinable but objectively indeterminate. The field is not constructed via pure dichotomies cut with digital and determinate boundaries, but enacted from continuous tensions. Likewise, figures do not have true constitutive background-independence from the ground of a field.

A persistent enactive process of objectification is synthesized via this genetic role of attention and perception in a self-stabilizing positive feedback loop. This is distinct from reification, occurring when propositional stances cross the line into conclusions of perceptual realism, e.g. the constancy hypothesis or the perceptual core hypothesis that Merleau-Ponty and Dreyfus contradict. 

Objectification without reification is stabilized via the positive feedback-reinforcement and dialectic between attention as genetic for a field (“signifying,” “figuring-out”), and attention spotlighting a field. However, the field is not to be reified as observer-independent nor pregiven. The field must be processually enacted. Stabilized objectifications likewise must be processually enacted, at least until they synthesize a consequent process following dialectic interaction with afforded opportunities to act. The field itself is processually and continuously co-constructed/enacted with the stabilized figures and grounds, both having non-dichotomous background dependency. The two aspects of perception and attention positively reinforce each other’s feedback, stabilizing determinably tensioned (but indeterminately bounded) significations from insignificance. During self-stabilizing and self-reinforcing processes of objectification, the field is treated as if it was “already there” in its remanence. However, if the field is reified, then the objectified properties of a field are propositionally concluded as being “already there” prior to enactment. That is, determinate properties of a reified field are taken to be empirically realizable and objectively idealiziable. Under reification, the objectified field is a separate background independent from its subjects, and subjects can gain representational access to the determinate objects in the field. 

Distinguishing Process Ontology from Subjective Idealism 

The process of enactive objectification by-and-for an organism is not equivocal to reification, nor is it equivocal to an idealist stance of solipsistic creation of a perceptual world ex-nihilo. Reification involves a propositional attitude resultant from the metaphysics of substance ontology combined with a philosophical stance of objective realism, as previously outlined. Solipsism is a propositional attitude following certain metaphysical interpretations of subjective idealism, that reality is indistinguishable from perception. Processual enaction rejects both substance ontology and subjective idealism. Processes of enacted objectifications are not to be confused nor conflated as metaphysically idealist: objectifications are abstracted from physical patterns, but it strongly does not follow that physical patterns of reality are indistinguishable from these abstractions. Process ontology claims ontically real processes and ordered patterns of flowing matter-energy. Determinable bounds and phenomenal appearances are stabilized contingent to vital norms and pragmatic epistemics. Processes are ontically real, but their determinable bounds are abstracted. This does not mean that ontically real processes are indistinguishable from an organism’s abstractions and know-how when perceiving with them.

Process ontology is a stance of ontically real physical processes and negentropically organizing patterns. Ontically real flowing processes can be temporally cross-sectioned as stable things due to an organism’s action of signification (and embodied know-how with such ordered physical patterns), relative to vital norms and history. 

Objectification as an enacted biosemiotic process develops through the genetic positive feedback loops between attention and perception, leading to a stable figure or signal that can be pragmatically interacted with. The stabilized figures and signals are treated as having been already there in their persistence, but objectifications are not to be propositionally reified as actually/really “already there.” Objectifications have virtual and pragmatic persistence via self-stabilizing positive feedback loops between attention and perception. Substantial objects are not real stabilities with determinate properties as claimed under substance ontology, and as claimed to be really discoverable via objective realism.

Positive Feedback Loops and Circularly Concrescent Becomings 

Is precedent consciousness constituting for perception? Or is consciousness co-constituted with perception in co-precedence? The former is Intellectualist and representationalist, the latter is genetic and enactive. Consciousness and perception are co-concretizing in situ, and the one cannot be reductively abstracted and taken to stand viably in isolation. Perception and attention are circularly inextricable in their co-becoming positive feedback loop together. The same applies for meta-attention and awareness, as self-stabilizing in a positive feedback loop with pre-reflective attention. 

Experience is self-organizing and self-stabilizing, and experience is genetic for itself. Experience is embodied and embedded, requiring continuous and active precedents of autopoiesis and autonomy. In other words, life and mind are constitutively continuous in self-organizing and self-stabilizing positive feedback loops, and cannot stand viably alone without collapse to an equilibrium state (i.e. death). This is the experience of an “organ-ism” as a literal “doctrine of instrumentality” enacting its own vital normativity via autonomy (from “organon” meaning instrumental, and “ism” meaning doctrine; “auto” meaning self, and “nomos” meaning law/legislating). 

To clarify some word usage: “constitution” to Merleau-Ponty is not synonymous with “constructivism” (as in the philosophical movement), and his usage of constitution remains ambiguous. The word usage of “constitution” could be interpreted in two ways. “Constitution” under a representationalist and intellectualist usage relates to the “constituting consciousness,” i.e. a precedent and stand-alone consciousness that constitutes its perception. “Constitution” under a processual-enactive usage relates to the “co-constitutive consciousness,” as concrete and genetic for perception. This usage of constitution refers to a concrete organization of figure-ground differentiations, not the constitution of perceptual figures from aggregate sense data as partes extra partes. Pre-reflective consciousness and perception are holistically cogenerative, self-reinforcing, and self-stabilizing in a positive feedback loop. One can be heuristically abstracted from the other, but an actual constitutive separation or dichotomy would lead to decoherence. Hence, the enactive word usage of constitution in this text contradicts and rejects the cognitivist-representationalist usage. Thus, constitution refers to a genetic, concrete organization of wholes from which interdependent parts are abstractions that exist for-and-by-means of the whole. The whole organizes and differentiates while still obtaining integration of interdependent “parts.” Constitution is not the aggregate assembly of independent, atomistic sense data into a whole.

The Antinomy Between Representing a World and Constructing a World to Represent, under Substance Ontology

Intellectualism claims that a stand-alone consciousness represents external objects to itself in order to construct perception. In a related way, empiricism claims that a stand-alone consciousness represents and predicts the external world in order to construct knowledge. 

However, both stances take consciousness as the already precedent “constituting consciousness.” These stances are empty of genetic power. Stand-alone consciousness is taken to already exist as the stable functional property of a stable substance, separate from the stable external world of objective substances with properties. Consciousness does not and cannot construct a world, it has to represent the word. This is a constituting consciousness that constructs perception to represent objective, substantial reality. This is not a co-constitutive consciousness-perception system that constructs stable abstractions from a processually flowing and indeterminate world. 

To intellectualism, consciousness is already existing as precedent, and represents a determinate external world. To empiricism, consciousness already exists as precedent, and predicts or deduces a determinate external world. Both frameworks belie an assumption of substance ontology, and the validity of representational content is relative to the truth and accuracy conditions of an objective determinate reality of stable and ontically real substances-with-properties. The representational or deductive/predictive success of contentful knowledge of things is relative to the truth and accuracy conditions inherent to objectively real substances and their properties. 

Either way, both frameworks must face the intractable symbol grounding problem and related hard problem of content (per Hutto-Myin). Both empiricism (e.g. predictive processing frameworks) and representationalism are empty of generative/genetic power for phenomenal experience, as consciousness is constitutive for perception, but not for the enactment of an abstracted world. Stand-alone consciousness constructs perception as a function to represent a substantial reality having true, objective properties. This is contrary to the processual-enactive framework, in which perception and consciousness are co-constitutive, and genetic for phenomenal experiences. Abstractions are enacted as an embodied know-how with ontically real but indeterminate physical processes, and are stabilized contingent to embodiment with vital and pragmatic norms. 

The antinomy of representationalism and empiricism is that consciousness constructs perception with the function of representing a world, but representation can not construct the world it needs to represent. Under substance ontology, this would be taken as a commitment to subjective idealism. These frameworks are impotent to construct or enact their worlds, the world already exists separate from consciousness and has to be predicted or deduced. 

To prevent this antinomy, the alternative solution is to adopt a framework of processual-enaction. 1. Consciousness does not construct a function of perception. Instead, consciousness and perception processually co-arise. They self-organize and self-stabilize in a positive feedback loop, generated continuously with the organizational processes of autopoiesis and autonomy for a living organism. Mind is continuous with life in enaction. 2. The function of perception is not to represent a determinate objective world with true properties. Instead, perception abstracts stable patterns from diachronically indeterminate (but ordered) physical patterns, and relates to them with contingencies of vital and pragmatic norms. Consciousness and perception are co-arising, and together they are genetic for an enactively objectified world. Stand-alone consciousness does not construct perception which is then left to represent, but not construct, a determinate world’s properties. Overall, adoption of process ontology combined with enaction wholly prevents the intractable antinomy caused by substance ontology and representationalism, empiricism or predictive processing. Processual-enaction obviates the need to solve artificial problems of symbol grounding, the hard problem of content, and perceptual indeterminism. Perception is abstractly determinable but always indeterminate because the world is ontically made of indeterminate processes that have to be pragmatically used and signified, but not “accurately represented.” 

The Antinomy of Perceptual Indeterminacy and Direct Perception under Substance Ontology

Perceptual indeterminism is not a problem for direct perception under processual-enaction. However, the notion of perceptual indeterminacy remains an intractable antinomy under the framework of direct perception plus substance ontology. The former framework (direct perception plus process ontology and enaction) conceives of perception as constitutive for a world, contingent to vital and pragmatic norms. Thus, indeterminacy is assumed and determinacy is a product of contingent abstractions and enactive objectifications. The latter framework (of direct perception plus substance ontology) struggles to see how direct perception of a substantially real world can ever be indeterminate. If perception as a function to directly perceive a substantially real world with determinate properties, problems like illusion and perception indeterminacy defy explanation. Again, adopting a framework of process ontology and enaction wholly prevents the artificial antinomy of direct perception and indeterminacy. This is because perception is taken to be genetic, not determinist. 

For empiricism and representationalism, both consciousness and the world are taken to have already been determined and generated, and the two are left to merely interface, but remain separate and external to each other. These stances are in contrast to the tensioned dialectic which is constitutive and enactive, and not to be resolved (i.e. remain in non equilibrium) as long as life persists in its non equilibrium steady states. 

Returning to the Positive Feedback Loop between Genetic Attention and Perception

How you take something to look changes how you take it to be, and how you take it to be changes how it looks. The activity of enactive objectification has positive feedback upon signifying the perceptual field, and the perceptual field is generative for objectification— therefore there is a positive feedback loop emerging in the diachronic process of enactive objectification. 

Reification would be a subsequent wrongheaded propositional step. I.e. Reification proposes that objectifications are real ontical constancies, instead of transiently and relationally enacted objectifications. Objectifications (v) are concretized holistically during enactive objectifying; objects (n) are not uncovered via abstracting determinate/realist constancies. Note that the word usage of “abstracting” applied to objects as nouns would refer to intellectualist representation, or empiricist prediction/deduction.

To enact a phenomenal field: The being-in-time has to both signify (v) and figure-out (v) signal from noise and figure from ground, and interact relationally to it. This process is enabled as attention-with-perception diachronically organize positive feedback loops that are constitutive for the phenomenal field. The mistake is to think the actively objectifying (v) feedback loops are instead static realist objects (n). Objectifying (v) patterns are emergent; realist objects (n) are not uncovered.

Conclusions 

This chapter ends with a brief recap, and applies several instructive quotations by Hubert Dreyfus (from his archived lectures on Merleau-Ponty). 

“Perception awakes attention, and attention develops and enriches perception.” 

Here, Dreyfus draws a distinction between what he calls primary attention (genetic attention plus spotlight attention) and secondary attention (pure attentional spotlight). He notes that intellectualism and empiricism only recognize the secondary type. Secondary attention only involves the mechanistic and intellectualist role of the attentional spotlight, and this type of attention refers to the representationalist role of perception under substance ontology. Primary attention refers to the genetic positive feedback loop between perception and attention. 

“We have to get the right body set to unlock what is calling us.” This quotation from Dreyfus alludes to the role of continuous attunement between the organism and environment as a concrescent system, as the organism perceives with its environment and affords opportunities to act. This quotation also alludes to the role of attunement in enabling an Uexküllian search-image for perception, and the role of attunement in enabling Gibsonian affordances as “announcements” and “invitations.” 

“I don’t just copy what’s out there, and I’m not creating it ex nihilo, either.” This quotation is to emphasize that enaction or genetic phenomenology must not be considered as organizing a world ex nihilo per subjective idealism, and must not be considered as representing nor predicting a world per objective realism under substance ontology.  

To pull these ideas together, it is helpful to show how the phenomenal field’s differentiated figures and grounds are not background independent. Instead, figures and ground are continuously background-dependent to their co-constituting field. In this way, figures, ground and field differentiate themselves holistically together as concrescence, under an organism-environment system’s operational closure. Note that the notion of “background dependency” is a topic originally used in the field theories of physics, but this is also a helpful concept when reapplied to thinking about phenomenal fields in philosophy. 

Overall, the condition of “background independence” describes secondary (representationalist) attention, transcendence, reified objectivism, realism, intellectualism and empiricism. 

Primary (genetic) attention, ecstasis, etc involve a lack of background independence. The phenomenal field itself must be co-constructed continuously with the figure-ground/signal-noise differentiation, therein. The background and its elaborations are continuous in their holistically concretizing enaction, and the figures/signals are not independent from the “background” phenomenal field. Figures/signals can not be abstracted as reified objects (n) apart from the field. This is because figures, ground and the field itself are concretized and differentiating holistically together, integrated under organism-environment system closure. Field, figures, and ground are continuously enacted processes, differentiated and integrated under concrescent closure. Likewise, figures are not background-independent “real, determinate objects” apart from their ground. Figures, ground and field are tensioned and continuous, but not dichotomous and reified independent of each other. 

Merleau-Ponty refutes stances of empiricism, objective realism, intellectualism and cognitivism. Never fully determined, the lebenswelt is stable but always becoming; the life world is always undetermined but determinable transiently and relatively to vital and pragmatic norms. Full determinism per substance ontology leads to antinomies and false dichotomies, impotent of genetic process. Instead, perception and phenomenal fields enact from indeterminate, unresolved and creatively tensioned dialectics, leading to processual becomings as status nascendi.