Section 5: Agents and Consciousness

This section asks what it means to be an agent within a relational, open territory — and what consciousness is, understood not as a mystery to be dissolved but as a structural feature of a particular kind of node in the relational fabric.

Section 5A: The Agent as Mechanism

The case for Frame A on agents and consciousness, argued from within.

The Starting Point

Frame A’s account of agents begins from its strongest ground: the extraordinary explanatory success of mechanism. Cognitive science, neuroscience, and artificial intelligence have between them produced a picture of cognition as information processing — the reception, transformation, and output of structured signals — that has generated genuine understanding and genuine technology. The brain is a physical system. Its operations are, in principle, physical operations. The agent is a sophisticated mechanism, and mechanism is a framework with a formidable track record.

This is not reductionism by assumption. It is reductionism by track record. Every time a cognitive capacity has been examined carefully — perception, memory, attention, language, emotion — it has yielded, at least partially, to mechanistic analysis. The mechanisms are complex and the analysis is incomplete. But the progress has been consistent, and the consistent direction of progress is evidence about where the complete account lies.

Dennett and the Dissolution of the Hard Problem

The most intellectually serious development of Frame A’s position on consciousness is Daniel Dennett’s heterophenomenology, developed at length in Consciousness Explained (1991) and refined in subsequent work.

Dennett’s move is precise and should be understood on its own terms before being evaluated. He does not deny that subjects have experiences. He denies that the intuitions we have about those experiences — that they are unified, that they occur at a single moment in a single location, that there is a privileged inner theatre in which they are presented — accurately track what is actually happening. The brain, on Dennett’s account, runs multiple parallel processes that continuously revise and update representations. What we call conscious experience is a model the brain constructs of these processes — a user interface, not a window onto the underlying computation. The feeling of unified, present, subjective experience is itself a representation generated by the machinery, not a direct acquaintance with a further fact.

This is Frame A pursued to its logical conclusion with full intellectual honesty. The hard problem — why there is something it is like to be a brain in a particular state — does not, on this account, identify a genuine explanandum. It identifies a persistent illusion generated by the representational machinery itself. The machinery generates reports of experience. The reports feel, from the inside, like reports of something irreducible. But the feeling is itself a product of the machinery, and once this is seen clearly, there is no residual mystery. The explanatory gap between physical process and subjective experience closes because the gap was always between physical process and a representation of physical process — not between physical process and something genuinely other.

This is a genuinely powerful argument. It is not a dismissal of consciousness. It is a careful attempt to show that the phenomenon, properly understood, does not require more than the mechanism can provide.

The Achievements and the Resources of Frame A

Beyond Dennett, Frame A has significant resources on consciousness and agency.

Functionalism — the view that mental states are defined by their functional role rather than their physical substrate — allows Frame A to take the reality of mental states seriously without invoking anything beyond the physical. Pain is whatever state plays the pain role: receiving damage signals, motivating avoidance behaviour, generating reports of suffering. The state is real. Its reality is fully captured by its functional relations. Nothing further is required.

Global workspace theory (Baars) and predictive processing frameworks (Clark, Friston) offer mechanistic accounts of how unified, flexible cognition emerges from distributed neural processes. These are not merely promissory — they generate testable predictions and have empirical support. They show that the appearance of a unified conscious agent can be explained by specific computational architectures without invoking anything outside the mechanism.

On agency specifically, compatibilist accounts of free will — from Hume through Frankfurt to contemporary philosophers — show that the freedom relevant to moral responsibility is compatible with determinism. What matters is not whether actions are undetermined but whether they flow from the agent’s own reasons, values, and deliberative processes. An agent who acts from its own settled dispositions, responsive to reasons, is free in every sense that matters for ethics — regardless of whether those dispositions are themselves physically determined. Frame A does not need indeterminism to ground responsibility. It needs only the right kind of mechanism.

Where Frame A Is Pressed

Frame A’s position is coherent and productive. It faces pressure at one point that has proved remarkably resistant.

Dennett’s dissolution of the hard problem has not convinced the majority of philosophers of mind, and the reason is not confusion or insufficient attention to the argument. It is that the argument, however carefully made, seems to leave something out. After the heterophenomenological account is complete — after the user illusion is described, the Cartesian theatre debunked, the multiple drafts explained — the question remains: why is there something it is like to undergo these processes? Not why do systems generate reports of experience, but why is the generation of those reports accompanied by anything at all?

This is Chalmers’s hard problem, and its persistence is a data point. It has survived thirty years of serious philosophical and empirical engagement without resolution. Frame A’s response is that the question is confused — that it is asking for an explanation of something that is already fully explained, and that the residual sense of mystery is itself a product of the machinery. This response may be correct. But it requires the phenomenon — the first-person fact of experience — to be less than it appears. Whether that requirement is a genuine insight or a structural limitation of Frame A is precisely what is at issue.

Section 5B: The Agent as Selection Node

The case for Frame B on agents and consciousness, argued from within.

The Starting Point

Frame B’s account of agents begins from the generative ladder. The ladder produced, at its highest currently known rung, systems that do not merely instantiate relational structure but represent it. A cognitive system models its own relations with the environment, holds multiple possible outcomes simultaneously, and participates in the resolution of genuine openness. The question is what it is to be such a system from the inside — and whether that question has an answer that the mechanism, by itself, can provide.

Frame B’s answer is that consciousness is not an addition to the mechanism, nor an illusion generated by it. It is the structural consequence of being a selection node that represents its own bifurcation points. It is what the inside of genuine openness feels like when the system is complex enough to have an inside.

Interiority as Structural Feature

At the cognitive threshold of the generative ladder, the relational fabric generates a system whose self-maintaining loop includes a model of itself in relation to its environment. This model is not merely a representation in the computational sense — a data structure that encodes information about the system’s state. It is the system’s way of being inside its own relational field. The model does not merely describe the system’s situation. It is the mode in which the system inhabits that situation.

This distinction matters. A thermostat has a representation of temperature — a physical state that correlates with temperature and triggers a response. But the thermostat is not inside its representation. It does not inhabit its model of the world. An agent, at the cognitive threshold, is inside its representation. Its model of the relational field is its way of being in that field. The difference between these two things is not a difference of degree or computational complexity. It is a difference in kind — the appearance of genuine interiority, which is the appearance of a self that is inside its own situation rather than merely responding to it.

Consciousness, on this account, is interiority: the structural feature of a selection node that is inside its own bifurcation points. This is broadly consistent with enactivist approaches to mind (Maturana and Varela, The Tree of Knowledge, 1987) and with integrated information theory (Tononi, Consciousness as Integrated Information, 2004), both of which treat consciousness as constitutively relational — as something that exists in the relation between a system and its world, not as an internal property of a substrate considered in isolation.

Genuine Openness Represented

The previous sections established that the relational territory has genuine ontological openness: regions where multiple outcomes remain simultaneously viable, where the future is not fixed by the present. Lower rungs of the generative ladder operate at bifurcation points without representing them. A chemical system at a bifurcation point resolves it without knowing it is doing so. A biological system resolves bifurcation points through the dynamics of its self-maintaining loop, still without representation.

An agent at a bifurcation point holds the openness explicitly. Multiple outcomes remain simultaneously viable within the model. The resolution requires a selection that the model participates in. This is what deliberation is — not the mechanical working out of a predetermined output, but the genuine navigation of a space where multiple outcomes remain viable until the selection is made.

Hofstadter’s analysis of strange loops is essential here. The transition from cognition to full self-aware agency is not merely the addition of more processing power. It is the appearance of a loop: a system whose representations loop back on themselves, modelling the modelling process itself. (Gödel, Escher, Bach, 1979; I Am a Strange Loop, 2007.) The strange loop is not a feature of the underlying biology. It is an emergent relational structure at the cognitive level — one that arises when self-representations achieve sufficient complexity and coherence to form a stable, self-referential pattern. The self that emerges from this loop is constituted by the loop itself. It is not a substance that happens to be self-aware. It is a relational structure whose identity just is its pattern of self-reference.

This is the mechanism by which the bifurcation process becomes reflexive — where the territory generates a node capable of representing genuine openness and participating in its resolution from the inside.

Selection Within Structure

The agent’s selections are not unconstrained. They occur within the relational field the agent inhabits — within the history that formed it, its biological constitution, its social and cultural embedding, the specific bifurcation point it faces. Frame B does not treat this as a limitation on freedom. It treats it as the condition of freedom.

A selection with no constraints is not a choice. It is noise. A genuine choice is a resolution of genuine openness within a structured relational field. The structure is what makes the choice coherent — what makes it this agent’s choice, expressing this agent’s values and history and understanding, rather than a random fluctuation in the relational fabric. The openness is what makes it real — what distinguishes it from the mechanical working out of a fixed programme.

Freedom, on this account, is neither the absence of constraint nor the exemption from the relational field. It is the capacity to participate in the resolution of real openness within a structured fabric. This is more than the compatibilist freedom of Frame A — not because it ignores the agent’s history and dispositions, but because it insists that the openness is genuine rather than a complexity that merely appears open from the inside. The difference is ontological, not merely psychological. Whether that difference matters for ethics is the subject of the next section.

The Hard Problem Revisited

Frame B does not dissolve the hard problem. It relocates it.

Dennett’s dissolution requires the first-person fact of experience to be less than it appears — a representation of a representation, a user illusion generated by machinery that has no genuine inside. The residual sense that something has been left out is, on his account, itself an illusion. Frame B takes a different position: the sense that something has been left out is correct, and what has been left out is interiority — the structural feature of a selection node that is inside its own bifurcation points.

This does not make consciousness less difficult to understand. It makes it differently difficult. The question is no longer why a physical mechanism generates the illusion of experience, but what it means for a relational structure to be inside its own situation rather than merely responding to it. This question is hard. But it is a question about a genuine feature of the relational territory rather than a question generated by confused intuitions about a non-existent phenomenon.

Frame B predicts that the hard problem will not be dissolved by better neuroscience or more sophisticated functionalism. It will be approached — though perhaps not fully resolved — by understanding what it means for a relational structure to achieve genuine interiority. That understanding is likely to require concepts that are not yet available, because the structure of interiority is genuinely novel at the cognitive rung of the generative ladder — as novel as life was at the biological rung, and as resistant to reduction to what came before.

Section 5C: Agents as Witness

What the agents and consciousness argument establishes — and what it hands to the sections that follow.

The asymmetry between the two sections is real but requires careful statement.

Frame A’s account is powerful and productive. Dennett’s dissolution of the hard problem is the most sophisticated attempt to close the explanatory gap from within Frame A’s resources, and it deserves to be taken seriously rather than dismissed. The cognitive sciences built on Frame A’s mechanistic assumptions have generated genuine understanding. Compatibilist accounts of free will are serious philosophy, not evasion. Frame A is not wrong about agents. It may be incomplete.

Frame B’s account makes a stronger claim: that consciousness is a structural feature of the relational territory — the interiority of a selection node — rather than an illusion generated by mechanism. This claim is harder to establish than Frame A’s, because interiority is precisely the thing that resists third-person description. But the difficulty of establishing it is not evidence against it. It may be evidence that the concepts required to establish it have not yet been fully developed.

What can be said at this stage is this. The hard problem has survived thirty years of serious engagement without resolution from within Frame A. This is not conclusive — hard problems survive for a long time before yielding. But the pattern of survival matters: the problem has not merely proved technically difficult. Every serious attempt to dissolve it has produced the reaction that something genuine has been left out. That reaction is itself a data point. It is what Frame B would predict: that the attempt to account for interiority from outside — from the third-person perspective of mechanism — will always leave a residue, because interiority is precisely what the third-person perspective cannot fully capture.

What this section hands to the sections that follow is twofold.

First, if agents are genuine selection nodes — if their deliberations are real navigations of genuine openness rather than complex mechanisms producing the appearance of choice — then their selections matter ontologically. The relational fabric is genuinely different depending on how the resolution goes. The branch not taken is not recoverable. This is what grounds the weight of ethical responsibility. It is not a moral postulate. It is a structural consequence of what agents are.

Second, if the self is a relational structure — constituted by its history, its embedding, its strange loop of self-reference — then the ethics that apply to it are not the ethics of an isolated substance protecting its interests. They are the ethics of a node in a relational fabric, whose identity is partly constituted by the very relations that ethics governs. What it means to act well, within this account, is not separable from what it means to be well — to inhabit one’s relational field with integrity rather than extraction. That argument is the subject of the next section.

One further point belongs here, because it shapes everything that follows. The constraints that structure an agent’s selections — its history, its attractor landscape, its specific configuration of relations — are not obstacles to genuine agency. They are its medium. A choice made by no one in particular, against no background, within no history, is not a purer form of freedom. It is not freedom at all. It is noise. What makes a forgiveness genuinely a forgiveness, what makes a choice genuinely this person’s choice, is precisely the texture of constraint within which it is made — the pull of the attractor that makes it costly, the history that makes it meaningful, the relational field that makes it matter. The constraints are not incidental to the self. They are constitutive of it. This has a consequence that Frame A, with its intrinsic individual prior to all relations, cannot easily accommodate: the relational self is not less particular than the Frame A individual. It is more particular — irreducibly this one, shaped by a history of relations that no one else has inhabited. That particularity is not a limitation to be overcome on the way to genuine agency. It is what genuine agency is.