Perceptual Logic: Regulating Perceptual Organization in Sense-Making Under Drift

-Nicholas Davis, PhD



Abstract

Perception is typically theorized either as a representational process that constructs internal models of the world or as an inferential mechanism that minimizes error relative to predicted input. While these approaches have yielded powerful explanatory frameworks, they remain insufficient for accounting for how sense-making is sustained over extended interaction in non-stationary environments. In particular, they lack a principled account of how salience, affordance perception, and perceptual–action coupling reorganize under gradual misalignment rather than discrete error.

This paper introduces perceptual logic as a foundational construct for understanding perception at the level of regulation rather than representation. Perceptual logic refers to the historically stabilized, implicitly enacted organization through which situations are encountered as meaningful and actionable. It structures what becomes salient, which affordances are disclosed, and how perception and action remain coupled across time.

Building on enactive and embodied accounts of cognition—particularly prior work on perceptual logic in computational creativity—we argue that perceptual logic must itself be regulatable in order for sense-making to remain viable under drift. We show how this regulatory view clarifies the limitations of belief-based, optimization-driven, and error-centric models of cognition, and how it integrates naturally with a theory of regulated sense-making. The result is a framework in which perception is understood not as a passive interface or internal model, but as a living, historically contingent organization that must be continuously regulated to sustain engagement in changing worlds.


1. Introduction

Why Perception Must Be Theorized at the Level of Regulation, Not Representation

Perception is often treated as the entry point of cognition: a channel through which information about the world is delivered to internal processes that interpret, evaluate, and act upon it (Marr, 1982; Pylyshyn, 1984). Even when this view is softened—by accounts that emphasize prediction, inference, or construction—perception is still commonly understood as something that happens to the agent, supplying content that cognition subsequently works with (Friston, 2005; Clark, 2013; Hohwy, 2013).

This paper argues that such framings mislocate the primary explanatory burden. The central problem is not how perceptual information is represented or inferred, but how perceptual organization remains viable over time as conditions change (Varela, Thompson, & Rosch, 1991; Di Paolo, Buhrmann, & Barandiaran, 2017).

In lived cognition—biological, social, and artificial—perceptual breakdown rarely announces itself as immediate error. More often, misalignment accumulates gradually. Situations become harder to read, actions require more effort, and engagement loses fluency long before any explicit failure occurs (Suchman, 1987; Schön, 1983; Rietveld & Brouwers, 2017). Systems may continue to function locally, responding correctly to stimuli or producing appropriate outputs, while globally losing their grip on the situations they are embedded in. These phenomena cannot be adequately explained by representational mismatch or prediction error alone (Kirsh, 1991; Di Paolo & De Jaegher, 2012).

What is missing is a theory of regulation.

To regulate is not merely to correct deviations from a target or to optimize performance relative to a fixed objective. Regulation concerns the maintenance of coherence under change. It addresses how a system sustains meaningful engagement when the very conditions that define relevance, salience, and actionability evolve over time (Ashby, 1956; Beer, 2000; Di Paolo, 2005).

Perception, on this view, cannot be reduced to input processing or world-model construction. It must be understood as an active organization of engagement—one that determines what matters, what is possible, and what counts as a viable next move (Gibson, 1979; Noë, 2004; Rietveld & Kiverstein, 2014). Crucially, this organization is not static. It is shaped by history, sedimented through prior success, and continually pressured by drift (Merleau-Ponty, 1962; Varela et al., 1991).

Existing perceptual theories struggle with this temporal dimension. Representational accounts assume that the problem is one of accuracy: the internal model must better match the world (Marr, 1982; Fodor, 1983). Predictive accounts assume that minimizing error is sufficient to sustain sense-making (Friston, 2010; Clark, 2016). Yet systems can minimize error while drifting structurally, compensating moment by moment without reorganizing the perceptual conditions that make engagement meaningful (Di Paolo et al., 2017; Bruineberg, Kiverstein, & Rietveld, 2018).

This paper proposes that perception must therefore be theorized at the level of perceptual logic: the implicit, historically constituted organization that structures salience, affordance disclosure, and perceptual–action coupling (Davis, Riedl, & Ueda, 2011; Rietveld & Brouwers, 2017). Perceptual logic is not a representation of the world, nor a belief about it. It is the way the world already shows up as meaningful and actionable for an agent (Heidegger, 1962; Noë, 2004).

The central claim developed here is that perceptual logic must itself be regulated. Without regulation, perceptual organization ossifies, overextends, or fragments under drift. With regulation, sense-making can remain viable even as contexts, intentions, and environments evolve (Di Paolo, 2005; Beer, 2000).

In the sections that follow, we ground this claim in enactive and embodied theories of cognition, clarify the role of perceptual logic in structuring salience and affordances, and show how regulated perceptual logic provides a missing link between perception, action, and temporality. The result is a framework in which perception is no longer a passive interface or internal mirror, but a living organization that must be continuously maintained to support ongoing sense-making.

2. From Enaction to Perceptual Logic

Grounding in Enactive and Embodied Cognition and Davis et al., 2011

The move from representational theories of perception to a regulatory account of perceptual logic is not a rejection of cognitive science, but a continuation of a trajectory already well established within enactive and embodied approaches to cognition. These approaches emerged in response to the limitations of computational and representational models, particularly their inability to account for meaning, situated action, and temporal coherence in lived cognition (Varela, Thompson, & Rosch, 1991; Brooks, 1991; Clark, 1997; Di Paolo, Buhrmann, & Barandiaran, 2017).

2.1 Enaction and the Rejection of Passive Perception

Enactive theories of cognition argue that perception is not the passive reception of information, nor the reconstruction of a pre-given world through internal representations. Instead, perception is understood as sense-making: the active bringing-forth of a meaningful world through ongoing engagement between an organism and its environment (Varela et al., 1991; Thompson, 2007).

On this view, perception and action are not sequential stages but dynamically coupled. What an agent perceives is inseparable from what it can do, and what it can do is shaped by its history of engagement (Noë, 2004; O’Regan & Noë, 2001). Meaning arises not from correspondence between internal models and external states, but from the viability of interaction—whether engagement can be sustained in the face of perturbation (Di Paolo, 2005; Barandiaran, Di Paolo, & Rohde, 2009).

While enactive theory powerfully reframes cognition as relational and temporal, it often remains abstract at the level of mechanisms. Enaction tells us that perception is enacted, but less often how the organization of perception itself stabilizes, shifts, and reorganizes across extended interaction (Thompson, 2007; Di Paolo et al., 2017). This is where the concept of perceptual logic becomes necessary.

2.2 Perceptual Logic as an Intermediate Explanatory Level

The notion of perceptual logic was introduced to address precisely this gap: the absence of a conceptual level that explains how perceptual organization coheres over time without collapsing into either moment-to-moment reactivity or static representation.

In PerLogicArt, Davis et al. (2011) argue that creative perception—particularly in computational and co-creative systems—cannot be understood solely in terms of symbolic reasoning, search, or rule-based generation. Instead, they introduce perceptual logic as the structured, historically conditioned organization through which salience is determined and action becomes possible within an unfolding creative process.

Perceptual logic, in this sense, is neither a set of explicit rules nor a representational model of the world. It is an implicit ordering of perceptual relevance: a way of seeing that determines what stands out, what recedes, and how emerging forms are interpreted as meaningful within a given practice (Davis et al., 2011; Schön, 1983).

Crucially, Davis et al. emphasize that perceptual logic is:

This framing aligns directly with enactive commitments while adding a missing layer of specificity. Perceptual logic names the organization that enaction presupposes but does not always articulate.

2.3 From Enaction to Regulation

What distinguishes perceptual logic from broader enactive accounts is its susceptibility to drift. Because perceptual logic stabilizes through repetition and success, it can become misaligned as contexts change. A way of seeing that once supported fluent engagement may gradually lose its fit—not abruptly, but through subtle accumulation of misalignment (Schön, 1983; Rietveld & Brouwers, 2017).

This is the point at which regulation becomes essential.

Enactive cognition already recognizes that organisms must maintain viability under changing conditions (Di Paolo, 2005; Ashby, 1956). Perceptual logic specifies where this maintenance operates: at the level of salience, affordance disclosure, and perceptual–action coupling. Regulation, then, is not an external corrective process, but an internal modulation of perceptual organization itself (Beer, 2000; Di Paolo et al., 2017).

By grounding perceptual logic in enactive and embodied cognition while extending it with an explicit regulatory dimension, we arrive at a framework capable of explaining how sense-making persists—or fails—over time. Enaction explains why perception is active and relational; perceptual logic explains how that activity is structured; regulation explains how that structure remains viable under drift.

Figure 1. Perceptual Logic as the Unit of Sense-Making Perceptual logic as the organizational locus of sense-making. The figure situates perceptual logic as an intermediate level between environment and behavior, distinct from stimuli, representations, and outputs. Perceptual logic organizes salience, discloses affordances, and structures perceptual–action coupling. Drift pressure acts primarily at this level, accumulating before overt behavioral error occurs. Regulation therefore operates on perceptual logic rather than directly on actions or outputs, shaping the conditions under which coherent sense-making remains viable over time.

This progression sets the stage for the next section, where we examine salience not as a feature of stimuli or attention alone, but as a product of perceptual logic itself.

3. Salience as a Product of Perceptual Logic

Salience is often treated as a primitive: a feature of stimuli, a function of attention, or a by-product of low-level perceptual mechanisms. In both cognitive science and artificial intelligence, salient features are typically assumed to stand out because of intrinsic properties—contrast, novelty, intensity—or because an attentional system selects them according to task demands, goals, or reward signals (Itti & Koch, 2001; Desimone & Duncan, 1995; Anderson, 2013).

From the perspective developed here, this framing is incomplete.

Salience is not given. It is produced.

3.1 Why Salience Cannot Be Reduced to Stimulus Properties

If salience were determined primarily by stimulus features, then the same elements would reliably stand out across contexts and agents. Empirically, this is not the case. What matters perceptually varies dramatically with history, practice, intention, and situation. The same visual pattern may appear trivial to one observer, structurally decisive to another, and invisible to a third (Goodwin, 1994; Merleau-Ponty, 1962).

This variability cannot be explained by stimulus properties alone. Nor can it be fully explained by top-down attention understood as deliberate selection. In most cases, what becomes salient does so prior to conscious choice. Attention follows salience more often than it creates it (Theeuwes, 2010; Wu, 2014).

Enactive accounts already point in this direction by emphasizing that perception is shaped by ongoing engagement rather than passive reception (Varela, Thompson, & Rosch, 1991; Thompson, 2007). However, without a concept like perceptual logic, salience risks remaining underspecified—acknowledged as relational, but not structurally explained (Di Paolo, Buhrmann, & Barandiaran, 2017).

3.2 Salience as an Outcome of Perceptual Organization

Perceptual logic provides this missing explanation. Salience arises from the current organization of perceptual–action coupling: the way perception is already structured to support engagement (Noë, 2004; Rietveld & Kiverstein, 2014).

Within a given perceptual logic:

These are not judgments applied after perception. They are properties of how the situation shows up (Heidegger, 1962; Merleau-Ponty, 1962).

In PerLogicArt, Davis et al. (2011) make this point explicit in the context of creative practice and computational creativity. They argue that creative perception depends on a perceptual logic that determines which structures, motifs, or emergent forms become salient within an unfolding work. Crucially, this logic is not encoded as explicit rules or feature detectors. It is enacted through sustained interaction with materials, tools, and prior forms (Schön, 1983; Davis et al., 2011).

Salience, on this account, is indexical rather than absolute. It points to what matters now, given the current trajectory of engagement (Rietveld & Brouwers, 2017).

3.3 Historical Sedimentation and Selective Sensitivity

Because perceptual logic is historically sedimented, salience is never neutral. What stands out reflects past success. Patterns that have previously supported viable action acquire a privileged status: they are detected more quickly, weighted more heavily, and trusted more readily (Merleau-Ponty, 1962; Varela et al., 1991).

This explains both expertise and its fragility.

Experts perceive salient structure where novices see noise. A trained musician hears tension and resolution where others hear sound. A seasoned artist sees compositional imbalance where others see completion. These differences are not cognitive overlays or explicit inferences; they are perceptual organizations shaped by history and practice (Goodwin, 1994; Dreyfus & Dreyfus, 1986).

At the same time, this sedimentation introduces inertia. When conditions change, previously salient features may remain dominant even as they lose relevance. Salience lags behind reality. This lag is one of the earliest signatures of interactional drift (Schön, 1983; Rietveld & Brouwers, 2017).

Importantly, salience does not fail by disappearing. It fails by persisting too long.

3.4 Salience, Discomfort, and the Onset of Drift

Because salience is produced by perceptual logic, misalignment first appears experientially as discomfort rather than error. What once stood out clearly now feels slightly off. Attention narrows or scatters. Certain features feel overemphasized; others seem inexplicably absent (Damasio, 1999; Rietveld et al., 2018).

These experiences are not noise. They are signals that the current perceptual logic is no longer organizing salience in a way that sustains coherence.

This is why salience is central to regulated sense-making. Regulation does not operate by selecting better outputs or optimizing internal metrics. It operates by modulating perceptual logic so that salience itself reorganizes—so that different aspects of the situation begin to matter (Di Paolo, 2005; Beer, 2000).

3.5 Implications

Reframing salience as a product of perceptual logic has several consequences:

The next section builds on this foundation by examining affordances—not as objective features of the environment, but as possibilities disclosed through perceptual logic rather than given in advance.

4. Affordances as Disclosed, Not Given

Affordances are often introduced as a corrective to representational theories of perception. Rather than treating the world as a set of neutral stimuli to be interpreted, affordance-based accounts emphasize action possibilities: what the environment offers an agent relative to its capacities (Gibson, 1979). This shift has been influential across ecological psychology, enactive cognition, HCI, and robotics.

Yet even within affordance-based frameworks, a subtle assumption often remains intact: that affordances are there to be detected.

From the perspective of perceptual logic, this assumption must be revised. Affordances are not given by the environment and then perceived. They are disclosed through perceptual organization.

4.1 Why Affordances Cannot Be Pre-Specified

If affordances were fixed properties of environments relative to agent capabilities, then different agents with similar bodies and skills would reliably perceive the same possibilities in the same situations. In practice, this is rarely the case.

What a situation affords depends not only on bodily capacity, but on:

A staircase affords sitting, climbing, blocking, display, performance, or rest depending on how it is encountered. These differences are not matters of belief or reinterpretation. They are differences in what is immediately actionable.

This variability cannot be accounted for by static affordance lists or predefined action mappings. It points instead to the role of perceptual logic in structuring what counts as possible at all.

4.2 Affordances as Relational Disclosures

In enactive accounts, affordances emerge through coupling between agent and environment rather than existing as independent entities (Varela et al., 1991; Di Paolo et al., 2017). Perceptual logic sharpens this claim.

An affordance is not simply a relation between body and object. It is a temporally situated disclosure: a possibility that becomes available within a particular organization of perception and action.

Perceptual logic determines:

Affordances, on this view, are enacted invitations rather than static offers. They arise when perceptual organization aligns with environmental structure in a way that supports action.

This explains why affordances can appear, disappear, or transform without any physical change in the environment. What changes is not the world, but the logic through which it is engaged.

4.3 Creative Practice and the Emergence of Affordances

Creative practice makes this process especially visible. Artists, designers, and musicians routinely report that materials “suggest” actions only after extended engagement. A line affords continuation only once a certain rhythm has been established; a sound affords variation only once a tonal field has stabilized.

In PerLogicArt, Davis et al. (2011) emphasize that creative affordances are not present at the outset. They are disclosed through interaction with evolving form. Early in a creative process, affordances are sparse and ambiguous. As perceptual logic stabilizes, possibilities multiply. As drift accumulates, previously salient affordances may close, while others open unexpectedly.

This dynamic cannot be captured by treating affordances as predefined action sets. It requires understanding affordances as products of regulated perceptual organization.

4.4 Affordances, Drift, and Misalignment

Because affordances are disclosed rather than given, drift manifests not only as misaligned behavior, but as affordance distortion. Actions that once felt obviously available now feel awkward or forced. New possibilities fail to appear even when conditions would support them.

This helps explain why systems—and humans—can become stuck despite having the necessary capabilities. The problem is not lack of skill, but a perceptual logic that continues to disclose obsolete affordances.

Regulation, in this context, is not about selecting better actions from a fixed menu. It is about reorganizing perceptual logic so that different affordances become available.

4.5 Implications for Artificial Systems

In artificial systems, affordances are often hard-coded or learned as mappings between states and actions. While this can support task performance, it limits responsiveness under drift. When creative direction shifts, the system continues to perceive the same affordances, even as they lose relevance.

A system grounded in perceptual logic, by contrast, would treat affordances as contingent disclosures. Its action possibilities would reorganize as salience and relevance shift, allowing it to remain engaged with evolving situations rather than persisting in outdated modes of interaction.

This reframing sets the stage for the next section, which examines how perceptual–action coupling itself drifts over time, and why regulation must operate at the level of perceptual logic rather than action selection alone.

5. Perceptual–Action Coupling and Drift

Perception and action are often discussed as distinct processes linked by information flow: perception delivers input, action produces output, and cognition mediates between them (Marr, 1982; Fodor, 1983). Enactive accounts have already dismantled this picture by emphasizing perceptual–action coupling—the idea that perception and action are mutually constitutive aspects of a single, ongoing engagement with the world (Varela, Thompson, & Rosch, 1991; O’Regan & Noë, 2001; Noë, 2004).

Perceptual logic allows this insight to be extended in a crucial direction: it explains how this coupling changes over time, and why those changes are not always adaptive (Di Paolo, Buhrmann, & Barandiaran, 2017).

5.1 Perceptual–Action Coupling as an Organized Dynamic

Perceptual–action coupling is not a generic linkage between sensing and doing. It is an organized dynamic: perception is structured in terms of action-readiness, and action continuously reshapes what is perceived as salient, relevant, or possible (Merleau-Ponty, 1962; Gibson, 1979; Noë, 2004).

Within a given perceptual logic:

This circularity is not a flaw; it is the basis of skill, fluency, and coherence (Dreyfus & Dreyfus, 1986; Rietveld & Kiverstein, 2014). When coupling is well-tuned, action “fits” perception without deliberation, and perception “anticipates” action without explicit planning (Clark, 2008; Noë, 2004).

However, this same circularity introduces vulnerability.

5.2 Drift as Coupling Misalignment

Drift arises when perceptual–action coupling remains locally functional while becoming globally misaligned with changing conditions. Because coupling is historically stabilized through repetition and success, it does not update automatically when environments, goals, or interactional contexts evolve (Schön, 1983; Di Paolo, 2005).

This misalignment does not initially appear as failure. Instead:

Yet the coupling begins to require more effort. Actions feel slightly forced. Perception narrows or overextends. Small compensations accumulate. What once flowed now needs constant correction (Suchman, 1987; Rietveld & Brouwers, 2017).

Drift, then, is not a breakdown of coupling, but a persistence of coupling beyond its conditions of viability.

5.3 Why Drift Is Hard to Detect

Perceptual–action coupling masks drift precisely because it is self-stabilizing. Each successful action reinforces the current organization, even if that success is achieved through increasing compensation. Local success delays global reorganization (Schön, 1983; Beer, 2000).

This explains why:

Error-based signals arrive too late. By the time coupling fails outright, the system has already exhausted its regulatory flexibility (Di Paolo et al., 2017; Bruineberg, Kiverstein, & Rietveld, 2018).

5.4 Drift as a Property of Coupling, Not Components

A critical implication follows: drift is not located in perception alone or in action alone. It is a property of their relation over time (Varela et al., 1991; Di Paolo, 2005).

Attempts to correct drift by:

often fail because they leave the underlying coupling intact (Kirsh, 1991; Beer, 2000).

What needs to change is not what is perceived or which action is chosen, but how perception and action are coordinated—that is, the perceptual logic that organizes their coupling.

5.5 Drift in Human and Artificial Systems

In human activity, drift appears as loss of grip, creative stagnation, or fatigue despite continued competence (Dreyfus & Dreyfus, 1986; Schön, 1983). In artificial systems, it appears as repetition, brittle adaptation, or incoherent escalation despite ongoing novelty (Brooks, 1991; Di Paolo et al., 2017).

In both cases, the underlying issue is the same: perceptual–action coupling has ceased to reorganize in response to changing conditions.

This observation motivates a regulatory turn. If drift is a property of coupling, then regulation must operate at the level of perceptual logic, not merely at the level of behavior, optimization, or internal representation.

The next section introduces two complementary regulatory dynamics—clamping and unclamping—as minimal mechanisms for modulating perceptual–action coupling under drift, without collapsing regulation into optimization or control.

6. Regulating Perceptual Logic: Clamping and Unclamping

If perceptual logic organizes salience, discloses affordances, and structures perceptual–action coupling, then the central question becomes how this organization changes over time without collapsing into rigidity or incoherence. Regulation, in this context, does not mean imposing control or correcting error. It refers to the modulation of perceptual logic itself so that engagement remains viable under drift (Ashby, 1956; Di Paolo, 2005; Beer, 2000).

To describe this modulation without reintroducing optimization or representation, we distinguish two complementary regulatory dynamics: clamping and unclamping.

6.1 Why Regulation Must Operate at the Level of Perceptual Logic

Drift cannot be addressed by faster learning, stronger attention, or improved action selection alone. These interventions presuppose that the underlying perceptual logic remains appropriate. When drift accumulates, this assumption fails (Di Paolo, Buhrmann, & Barandiaran, 2017; Bruineberg, Kiverstein, & Rietveld, 2018).

Regulation must therefore act prior to choice and beneath belief. It must alter how situations are structured as meaningful and actionable, rather than merely selecting among predefined alternatives (Merleau-Ponty, 1962; Noë, 2004). Clamping and unclamping name the minimal dynamics by which this alteration occurs.

These are not mechanisms in the engineering sense, nor discrete operations invoked on demand. They are modes of organization that shape how perceptual logic stabilizes or reopens under changing conditions (Beer, 2000; Thompson, 2007).

6.2 Clamping: Provisional Stabilization

Clamping refers to the temporary stabilization of perceptual logic. It is the process by which certain patterns, relations, or directions are held long enough for coherence to form (Schön, 1983; Rietveld & Kiverstein, 2014).

In clamping:

Clamping is not fixation. It does not freeze perception permanently. It establishes local coherence—a stable frame within which action can proceed without constant renegotiation (Dreyfus & Dreyfus, 1986; Clark, 2008).

This stabilization is essential. Without clamping, perception remains too labile to support sustained engagement. Action becomes reactive, trajectories fragment, and sense-making cannot consolidate (Beer, 2000; Di Paolo, 2005).

In creative practice, clamping appears as committing to a motif, a rhythm, or a compositional direction (Schön, 1983). In interaction, it appears as holding a conversational frame or shared focus (Suchman, 1987). In artificial systems, it corresponds to maintaining a regime or mode of participation rather than continuously re-evaluating every possibility (Brooks, 1991; Di Paolo et al., 2017).

6.3 Unclamping: Reopening Degrees of Freedom

Unclamping is the complementary dynamic. It refers to the loosening of perceptual constraints when stabilization becomes counterproductive (Ashby, 1956; Di Paolo, 2005).

As drift accumulates:

Unclamping responds to this condition by reopening degrees of freedom. Salience redistributes. Alternative affordances become visible. Perceptual logic becomes exploratory rather than confirmatory (Varela, Thompson, & Rosch, 1991; Thompson, 2007).

Crucially, unclamping is not triggered by explicit error. It is prompted by changes in coherence: loss of grip, persistent discomfort, or escalating effort (Damasio, 1999; Rietveld et al., 2018). These signals indicate that stabilization no longer supports viable engagement.

Unclamping allows perceptual logic to reorganize without specifying in advance what the new organization should be. It creates conditions for new sense-making to emerge rather than prescribing outcomes (Beer, 2000; Di Paolo et al., 2017).

6.4 Regulation as Modulation, Not Switching

Clamping and unclamping are not binary states. Regulation does not consist in switching cleanly from one to the other. Instead, they operate continuously and reciprocally (Beer, 2000; Thompson, 2007).

Excessive clamping produces rigidity. Perception becomes blind to change. Engagement persists through force rather than fit (Dreyfus & Dreyfus, 1986).

Excessive unclamping produces fragmentation. Salience shifts too rapidly, affordances fail to consolidate, and coherence dissolves (Ashby, 1956; Di Paolo, 2005).

Effective regulation consists in modulating between these tendencies in response to drift. The system remains neither locked nor unbounded, but poised (Rietveld & Kiverstein, 2014).

This modulation is sensitive to time. Clamping that is appropriate early in an interaction may become maladaptive later. Unclamping that is necessary for reorganization may need to be followed by restabilization if coherence is to be restored (Schön, 1983; Di Paolo et al., 2017).

6.5 Clamping, Unclamping, and Intentionality

Importantly, clamping and unclamping do not require a central controller. They are enacted through changes in perceptual organization itself—through attention, posture, pacing, and responsiveness (Merleau-Ponty, 1962; Noë, 2004).

Intentionality here is not command-based. It is expressed as answerability to drift. The system remains sensitive to when stabilization helps and when it hinders (Di Paolo, 2005; Thompson, 2007).

This reframes regulation as a property of engagement rather than a function of decision-making. Perceptual logic reorganizes not because a goal demands it, but because coherence requires it.

6.6 Implications

By articulating clamping and unclamping as regulative dynamics, we can explain how perceptual logic adapts without collapsing into optimization or control.

Regulation:

It sustains the conditions under which sense-making can continue (Ashby, 1956; Beer, 2000).

The next section extends this account beyond human perception, examining how perceptual logic—and its regulation—can be understood in both human and artificial systems without reducing one to the other.

Figure 2. Drift, Clamping, and Unclamping Over Time Regulation of sense-making as a trajectory rather than a state. The figure illustrates changes in perceptual coherence over time under three phases: stabilization through clamping, gradual misalignment under drift, and reorganization via unclamping followed by restabilization. During misalignment, behavior may remain locally successful even as perceptual coherence degrades, delaying error signals. Regulation operates by modulating perceptual organization over time, rather than switching states or optimizing momentary performance.

6.7 Perceptual Scale and Cross-Scale Tension

Perceptual logic is not monolithic. It operates simultaneously at multiple perceptual scales—local, regional, and global—each organizing engagement at a different level of granularity. These scales are not stages, layers, or control levels. They are concurrent organizations of sense-making that remain active throughout interaction. Regulation under drift often depends on how these scales align, diverge, or compete.

At the local scale, perceptual logic organizes fine-grained relations: strokes, gestures, micro-actions, and immediate affordances. It structures what feels correct or incorrect in the moment—whether a mark lands, a movement flows, or an action fits its immediate context. Local perceptual logic answers the question: Does this move work? It supports fluency, precision, and responsiveness at the level of execution.

At the regional scale, perceptual logic organizes intermediate coherence: motifs, phrases, clusters of action, and relational patterns that extend beyond individual moves. It structures whether local actions belong together, reinforce one another, or form a recognizable configuration. Regional perceptual logic answers the question: Does this section belong? It is at this scale that repetition becomes motif, variation becomes development, and isolated acts become organized passages.

At the global scale, perceptual logic organizes overall trajectory, direction, and identity. It structures the sense of what the activity, interaction, or work is becoming over time. Global perceptual logic answers the question: Is this still the same work? It governs long-range coherence, commitment, and the felt directionality of engagement.

Crucially, these scales are not hierarchical controllers, nor does one dominate the others by default. They are simultaneously active organizations, each contributing to how salience is produced, affordances are disclosed, and perceptual–action coupling is stabilized. None has absolute priority. Local fluency can coexist with regional confusion; regional coherence can coexist with global loss of direction.

It is precisely this simultaneity that gives rise to creative tension.

Figure 3. Cross-scale perceptual logic and creative tension. Perceptual logic operates simultaneously at multiple scales of organization. Creative tension arises when coherence is maintained locally while regional or global perceptual logic drifts, producing a felt demand for reorganization prior to overt failure. Regulation under drift often involves shifting the scale at which clamping and unclamping operate, allowing sense-making to be restored through reorganization rather than local optimization.

Creative tension arises when local perceptual coherence is maintained while regional or global perceptual logic drifts—producing a felt demand for reorganization before explicit failure occurs. Actions continue to work. Output remains competent. Yet something feels off. The work “asks” for change. This tension is not error, noise, or uncertainty. It is cross-scale misalignment becoming perceptually salient.

This explains why artists and practitioners often experience a sense of being called or beckoned to reorganize their activity. The demand does not originate in local breakdown but in the persistence of local success under broader misalignment. Optimization fails here because improving local performance only deepens the mismatch. Drift is detected aesthetically—through discomfort, loss of grip, or diminishing resonance—before it is detected behaviorally.

Clamping and unclamping operate across these perceptual scales and provide the regulatory means by which such tension is addressed. Clamping can occur at different scales:

Conversely, unclamping often involves shifting scale. When local refinements no longer restore coherence, regulation may require unclamping at the regional level—loosening motifs or reorganizing sections. When regional coherence persists but the work feels directionless, unclamping must occur globally, reopening the trajectory itself.

Regulation under drift therefore often takes the form of perceptual rescaling: shifting the level at which coherence is sought. Rather than correcting errors or selecting better actions, the system modulates which perceptual scale is allowed to dominate sense-making at a given moment. This rescaling enables reorganization without prescribing outcomes, allowing new coherence to emerge through engagement rather than imposition.

By articulating perceptual logic as a multi-scale organization, we can account for how creative tension arises prior to breakdown, how it guides reorganization without explicit representation, and why regulation must remain sensitive to scale if sense-making is to remain viable under drift.

7. Perceptual Logic in Human and Artificial Systems

Perceptual logic has so far been developed as a theoretical construct grounded in enactive cognition and lived experience. To be useful for regulated sense-making more broadly, it must also be applicable beyond human perception—without collapsing into anthropomorphism or, conversely, being reduced to implementation detail (Varela, Thompson, & Rosch, 1991; Di Paolo, Buhrmann, & Barandiaran, 2017). This section articulates perceptual logic as a structural concept that can be instantiated in both human and artificial systems, while remaining sensitive to their differences.

7.1 Avoiding Two Symmetric Errors

Discussions of perception across biological and artificial systems often fall into one of two errors (Searle, 1980; Brooks, 1991).

The first is anthropomorphic projection: attributing human-like experiences, intentions, or meanings to artificial systems. This obscures differences in embodiment, history, and vulnerability, and often leads to category mistakes about agency and understanding (Searle, 1980; Gunkel, 2012).

The second is reductive instrumentalism: treating artificial systems as mere tools whose behavior can be fully explained in terms of optimization, representation, or control, thereby ignoring interactional dynamics that unfold over time (Suchman, 1987; Dourish, 2001).

Perceptual logic provides a way between these extremes. It does not ascribe inner experiences to artificial systems. Instead, it characterizes how any system—human or artificial—organizes its coupling with an environment such that certain distinctions, continuations, and sensitivities become operative (Di Paolo et al., 2017; Rietveld & Kiverstein, 2014).

7.2 Perceptual Logic in Human Systems

In human systems, perceptual logic is grounded in embodied history. It is shaped by sensorimotor capacities, affective dynamics, social practices, and long-term engagement with particular environments (Merleau-Ponty, 1962; Varela et al., 1991; Thompson, 2007). Salience, affordances, and action-readiness are deeply interwoven with bodily states and lived meaning (Noë, 2004; Rietveld et al., 2018).

Because of this embodiment, human perceptual logic is:

Regulation in human systems therefore involves not only cognitive adjustment, but changes in posture, pacing, relational context, and environment (Merleau-Ponty, 1962; Suchman, 1987). Clamping and unclamping are enacted through these embodied channels rather than through explicit control.

Crucially, humans experience the reorganization of perceptual logic phenomenologically—as tension, flow, loss of grip, or renewed coherence (Schön, 1983; Dreyfus & Dreyfus, 1986). This experiential dimension is central to Intentional Perceptual Attunement (IPA), but not required for the concept of perceptual logic itself.

7.3 Perceptual Logic in Artificial Systems

In artificial systems, perceptual logic is instantiated differently. It is not grounded in lived experience, but in organizational structure: the way inputs, internal dynamics, memory, and action are coordinated over time (Beer, 2000; Di Paolo et al., 2017).

A system exhibits perceptual logic use when:

This can occur even in the absence of explicit representations or goals. For example, a co-creative system may develop regimes of interaction in which particular motifs, timing patterns, or relational structures dominate. These regimes function analogously to perceptual logics: they shape what the system “notices” and how it responds (Brooks, 1991; Davis et al., 2011).

Importantly, artificial perceptual logic need not be human-like to be meaningful. It need only be interactionally coherent—capable of sustaining engagement under changing conditions (Beer, 2000; Di Paolo et al., 2017).

7.4 Drift and Regulation Across System Types

Both human and artificial systems are subject to drift, but the sources and consequences differ.

In humans, drift arises from bodily change, emotional fatigue, social pressure, and historical sedimentation (Schön, 1983; Rietveld & Brouwers, 2017). In artificial systems, drift arises from changing inputs, evolving interaction histories, shifting contexts, or misaligned internal regimes (Beer, 2000; Di Paolo et al., 2017).

Despite these differences, the regulatory challenge is structurally similar: how to reorganize perceptual logic without losing coherence (Ashby, 1956; Di Paolo, 2005).

Clamping and unclamping apply in both cases, though they are realized through different mechanisms. In humans, clamping may appear as sustained focus or commitment; in artificial systems, as regime stabilization or constraint of variation. Unclamping may appear as exploratory attention in humans, and as loosening of internal constraints or increased variability in artificial systems (Beer, 2000; Thompson, 2007).

7.5 Why This Matters for Co-Creative Systems

In co-creative human–AI systems, perceptual logics do not exist in isolation. They interact. Human perceptual organization shapes what the system encounters as input; system behavior shapes what the human perceives as salient or promising (Suchman, 1987; Dourish, 2001; Davis et al., 2011).

Without regulation, these coupled logics drift out of alignment. The human may perceive repetition or incoherence where the system perceives novelty or success. Conversely, the system may fail to notice shifts in creative direction that are obvious to the human (Schön, 1983; Rietveld & Brouwers, 2017).

Understanding perceptual logic as a shared regulatory concern—not a property of one agent alone—allows co-creative systems to be designed around sustaining interactional coherence rather than optimizing individual performance (Di Paolo et al., 2017; Beer, 2000).

7.6 Summary

Perceptual logic is not a metaphor, nor a claim about inner experience. It is a structural account of how systems organize engagement over time.

In humans, it is embodied, historical, and phenomenologically accessible. In artificial systems, it is organizational, emergent, and interactionally visible. In both cases, it is subject to drift and in need of regulation.

8. Implications for Regulated Sense-Making Under Drift

The preceding sections have argued that perceptual logic is the structure through which salience is produced, affordances are disclosed, and perceptual–action coupling is organized over time. This section draws out the implications of that claim for regulated sense-making. The central move developed here is simple but consequential:

Perceptual logic is the unit of regulation in sense-making systems.

What must be regulated for sense-making to persist is not behavior alone, nor prediction accuracy, nor internal representations, but the organization through which situations become meaningful and actionable at all.

8.1 From Regulation of Behavior to Regulation of Sense-Making

Many adaptive systems regulate behavior. They correct errors, adjust parameters, or optimize actions relative to feedback. Such approaches implicitly assume that the perceptual organization through which situations are encountered remains appropriate.

When interaction unfolds under changing conditions, this assumption often fails.

Regulating behavior without regulating perceptual logic leads to familiar failure modes:

These failures occur because behavior-level regulation operates downstream of perceptual organization. By the time errors appear at the level of action or output, misalignment in sense-making has already accumulated.

If sense-making is to remain viable over time, regulation must therefore operate prior to choice and beneath belief. It must act on the organization that structures salience, affordances, and coupling in the first place.

8.2 Drift as Misalignment of Perceptual Logic

When perceptual logic is taken as the unit of sense-making, drift can be characterized more precisely. Drift is not noise, error, or randomness. It is a gradual misalignment between perceptual organization and evolving conditions.

Under drift:

This explains several otherwise puzzling phenomena:

Perceptual logic drifts before outputs fail. Regulation must therefore detect and respond to changes in perceptual organization rather than waiting for breakdown.

8.3 Clamping and Unclamping as Regulatory Dynamics

Regulation at the level of perceptual logic takes the form of two complementary dynamics: clamping and unclamping.

Clamping stabilizes perceptual organization long enough for coherence to form. It constrains salience and affordances so that engagement can proceed without constant renegotiation. Without clamping, perception remains too labile to support sustained trajectories of action.

Unclamping loosens that organization when stabilization becomes counterproductive. It reopens degrees of freedom when salience narrows excessively, affordances rigidify, or coupling demands increasing effort. Unclamping allows alternative perceptual logics to emerge without specifying in advance what those alternatives should be.

These dynamics do not optimize performance, enforce goals, or correct error. They manage organizational tension—the balance between stability and flexibility required for sense-making to persist under drift.

Crucially, regulation here is temporally extended. The success of clamping or unclamping cannot be assessed instantaneously. It is evaluated retrospectively, by whether engagement remains viable across time.

8.4 Sense-Making as a Trajectory, Not a State

Centering perceptual logic reframes sense-making as a trajectory-sensitive process rather than a momentary achievement. What matters is not whether a system is coherent at a given instant, but whether coherence can be sustained, repaired, or reorganized as conditions evolve.

This shifts the normative focus of cognitive systems:

Perceptual logic provides the structure through which this trajectory unfolds. Regulation acts by shaping how that structure evolves, not by enforcing fixed outcomes or predefined targets.

8.5 Implications for Artificial Sense-Making Systems

For artificial systems, adopting perceptual logic as the unit of regulation has concrete design consequences.

Rather than regulating:

a regulated sense-making system modulates:

This enables artificial agents to participate meaningfully in long-duration interaction without requiring fixed objectives, exhaustive retraining, or continuous optimization. Sense-making becomes something the system maintains, not something it recomputes anew each turn.

8.6 Toward a Unified Regulatory Account

The account developed here points toward a unified view of regulation that applies across biological, cognitive, and artificial systems. What varies across systems are the materials, mechanisms, and timescales through which regulation is enacted—not the object of regulation itself.

Perceptual logic plays a unifying role by providing a common locus at which regulation operates:

By identifying perceptual logic as the operative unit of sense-making, regulation can be theorized without collapsing into representationalism, optimization, or control. Regulation does not act on outputs or internal models. It acts on the conditions that make sense-making possible at all.

The final section concludes by positioning perceptual logic not as an auxiliary concept, but as the fundamental interface through which sense-making is enacted, regulated, and sustained under change.

9. Discussion: Perceptual Logic as the Unit of Sense-Making

This paper has argued for a shift in how perception, action, and meaning are theorized—away from representation, inference, or isolated affordance detection, and toward perceptual logic as the primary unit of sense-making (Varela, Thompson, & Rosch, 1991; Di Paolo, Buhrmann, & Barandiaran, 2017). Perceptual logic names the organized, historically sedimented way in which a system discloses salience, affords action, and sustains coupling with a changing world. It is not an internal model, a belief structure, or a decision rule. It is the lived organization through which a situation becomes actionable at all (Merleau-Ponty, 1962; Noë, 2004).

Reframing sense-making at this level resolves several persistent problems across cognitive science, creativity research, and human–AI interaction. First, it explains why coherence often degrades before explicit error appears. When perceptual logic drifts out of alignment with evolving conditions, a system may continue to act correctly by local metrics while losing global grip (Schön, 1983; Rietveld & Brouwers, 2017). Second, it clarifies why belief change, optimization, or learning alone are insufficient for sustaining engagement over time: these processes operate downstream of perceptual logic and therefore arrive too late to regulate emerging misalignment (Beer, 2000; Di Paolo, 2005; Bruineberg, Kiverstein, & Rietveld, 2018).

By treating salience as a product of perceptual logic, we showed that what “stands out” in a situation is not given by stimuli or features, but disclosed through organized readiness for action (Rietveld & Kiverstein, 2014; Davis, Riedl, & Ueda, 2011). By treating affordances as disclosed rather than given, we reframed action possibilities as relational achievements that depend on the current organization of perception (Gibson, 1979; Rietveld et al., 2018). And by analyzing perceptual–action coupling under drift, we demonstrated why sustained sense-making requires regulation rather than mere responsiveness or adaptation (Varela et al., 1991; Di Paolo et al., 2017).

The introduction of clamping and unclamping provided a minimal but powerful vocabulary for describing how perceptual logic is regulated over time. Clamping stabilizes organization long enough for coherence to form; unclamping reopens degrees of freedom when that organization becomes brittle or misaligned (Ashby, 1956; Schön, 1983). These dynamics are not error-driven, not goal-directed, and not optimization-based. They are viability-preserving responses to drift—responses that operate at the level of engagement itself (Beer, 2000; Di Paolo, 2005).

Extending this account across human and artificial systems revealed a crucial symmetry: while the materials and timescales differ, both humans and artificial agents face the same fundamental problem of sustaining perceptual coherence under changing conditions (Suchman, 1987; Di Paolo et al., 2017). What differs is not the presence of perceptual logic, but the mechanisms available for regulating it. Regulation does not act on representations or outputs; it acts on the conditions that make sense-making possible at all (Ashby, 1956; Beer, 2000).

The broader implication is that perceptual logic should be treated as a first-class construct in theories of cognition, creativity, and interaction. It is the level at which salience, affordances, agency, and meaning converge (Thompson, 2007; Rietveld & Kiverstein, 2014). Without it, theories oscillate between low-level mechanisms and high-level descriptions, leaving the dynamics of lived engagement undertheorized.

To take perceptual logic seriously is not to add another component to existing models. It is to change the explanatory center of gravity—from what systems compute or produce, to how they remain oriented in a world that does not stand still. Sense-making, on this view, is not an outcome or a capability. It is a regulated trajectory. Perceptual logic is the form that trajectory takes.

Conclusion

This paper has argued that perceptual logic constitutes the primary unit of sense-making. Rather than treating perception as input, affordances as given, or action as downstream selection, we have shown that salience, possibility, and engagement arise from an organized, historically situated perceptual logic that structures perceptual–action coupling. Drift, on this account, is not a failure of prediction or performance, but a gradual loss of alignment between perceptual organization and evolving conditions. Sense-making succeeds or fails at this level before error, belief, or evaluation come into play.

By situating perceptual logic within a framework of regulated sense-making, the paper clarifies how coherence can be sustained under change without recourse to optimization, representation, or fixed objectives. Regulation operates directly on perceptual logic through dynamics of stabilization and release, allowing systems—human or artificial—to remain viable as contexts shift. The contribution of this work is therefore not a new mechanism, but a reframing: perceptual logic is not a descriptive convenience, but the locus at which meaning, action, and regulation converge. Treating it as such opens a path toward theories of cognition and creativity that take temporality, drift, and lived engagement as foundational rather than residual.

References

Anderson, B. A. (2013). A value-driven mechanism of attentional selection. Journal of Vision, 13(3), 7. https://doi.org/10.1167/13.3.7

Ashby, W. R. (1956). An introduction to cybernetics. Chapman & Hall.

Barandiaran, X. E., Di Paolo, E. A., & Rohde, M. (2009). Defining agency: Individuality, normativity, asymmetry, and spatio-temporality in action. Adaptive Behavior, 17(5), 367–386. https://doi.org/10.1177/1059712309343819

Beer, R. D. (2000). Dynamical approaches to cognitive science. Trends in Cognitive Sciences, 4(3), 91–99. https://doi.org/10.1016/S1364-6613(99)01440-0

Brooks, R. A. (1991). Intelligence without representation. Artificial Intelligence, 47(1–3), 139–159.

Bruineberg, J., Kiverstein, J., & Rietveld, E. (2018). The anticipating brain is not a scientist: The free-energy principle from an ecological–enactive perspective. Synthese, 195(6), 2417–2444. https://doi.org/10.1007/s11229-016-1239-1

Clark, A. (1997). Being there: Putting brain, body, and world together again. MIT Press.

Clark, A. (2008). Supersizing the mind. Oxford University Press.

Clark, A. (2013). Whatever next? Predictive brains, situated agents, and the future of cognitive science. Behavioral and Brain Sciences, 36(3), 181–204. https://doi.org/10.1017/S0140525X12000477

Clark, A. (2016). Surfing uncertainty: Prediction, action, and the embodied mind. Oxford University Press.

Damasio, A. (1999). The feeling of what happens. Harcourt Brace.

Davis, N., Riedl, M. O., & Ueda, K. (2011). Perceptual logic and creative cognition. In Proceedings of the International Conference on Computational Creativity.

Desimone, R., & Duncan, J. (1995). Neural mechanisms of selective visual attention. Annual Review of Neuroscience, 18, 193–222.

Di Paolo, E. A. (2005). Autopoiesis, adaptivity, teleology, agency. Phenomenology and the Cognitive Sciences, 4(4), 429–452. https://doi.org/10.1007/s11097-005-9002-y

Di Paolo, E. A., & De Jaegher, H. (2012). The interactive brain hypothesis. Frontiers in Human Neuroscience, 6, 163. https://doi.org/10.3389/fnhum.2012.00163

Di Paolo, E. A., Buhrmann, T., & Barandiaran, X. (2017). Sensorimotor life: An enactive proposal. Oxford University Press.

Dourish, P. (2001). Where the action is. MIT Press.

Dreyfus, H. L., & Dreyfus, S. E. (1986). Mind over machine. Free Press.

Fodor, J. A. (1983). The modularity of mind. MIT Press.

Friston, K. (2005). A theory of cortical responses. Philosophical Transactions of the Royal Society B, 360(1456), 815–836. https://doi.org/10.1098/rstb.2005.1622

Friston, K. (2010). The free-energy principle: A unified brain theory? Nature Reviews Neuroscience, 11(2), 127–138. https://doi.org/10.1038/nrn2787

Gibson, J. J. (1979). The ecological approach to visual perception. Houghton Mifflin.

Goodwin, C. (1994). Professional vision. American Anthropologist, 96(3), 606–633.

Gunkel, D. J. (2012). The machine question. MIT Press.

Heidegger, M. (1962). Being and time (J. Macquarrie & E. Robinson, Trans.). Harper & Row. (Original work published 1927)

Hohwy, J. (2013). The predictive mind. Oxford University Press.

Itti, L., & Koch, C. (2001). Computational modelling of visual attention. Nature Reviews Neuroscience, 2(3), 194–203.

Kirsh, D. (1991). Today the earwig, tomorrow man? Artificial Intelligence, 47(1–3), 161–184.

Marr, D. (1982). Vision: A computational investigation into the human representation and processing of visual information. MIT Press.

Merleau-Ponty, M. (1962). Phenomenology of perception (C. Smith, Trans.). Routledge.

Noë, A. (2004). Action in perception. MIT Press.

O’Regan, J. K., & Noë, A. (2001). A sensorimotor account of vision and visual consciousness. Behavioral and Brain Sciences, 24(5), 939–973.

Pylyshyn, Z. (1984). Computation and cognition. MIT Press.

Rietveld, E., & Brouwers, A. A. (2017). Optimal grip on affordances in architectural design practices. Phenomenology and the Cognitive Sciences, 16(3), 545–563.

Rietveld, E., & Kiverstein, J. (2014). A rich landscape of affordances. Ecological Psychology, 26(4), 325–352.

Rietveld, E., Denys, D., & Van Westen, M. (2018). Ecological-enactive cognition as engaging with a field of relevant affordances. Frontiers in Psychology, 9, 2343.

Schön, D. A. (1983). The reflective practitioner. Basic Books.

Searle, J. R. (1980). Minds, brains, and programs. Behavioral and Brain Sciences, 3(3), 417–457.

Suchman, L. A. (1987). Plans and situated actions. Cambridge University Press.

Theeuwes, J. (2010). Top–down and bottom–up control of visual selection. Acta Psychologica, 135(2), 77–99.

Thompson, E. (2007). Mind in life: Biology, phenomenology, and the sciences of mind. Harvard University Press.

Varela, F. J., Thompson, E., & Rosch, E. (1991). The embodied mind. MIT Press.

Wu, W. (2014). Attention. In Routledge Encyclopedia of Philosophy.