Chapter 13: Masks, Compartments, and the Fractal Self — The Self as Configuration Space

The self is not one thing wearing many faces. It is many configurations — each genuinely expressive of the system that generates it, distributed across contexts, relationships, and roles. That claim needs defending, and this chapter defends it. It then examines what happens when the distribution becomes incoherent — when the configurations stop communicating with each other, lose shared authorship, or become so misaligned that the person carrying them pays a mounting tax of effort, concealment, and exhaustion.

Before going further, the strongest counter‑position deserves a hearing. The narrative identity tradition — Ricoeur in particular — holds that selfhood is constituted through the continuous story a person tells about who they are, one that integrates past, present, and anticipated future into a coherent arc. On this account, the “real” self is not any single configuration but the narrative thread that links all configurations across time. A simpler version of the same intuition is ordinary: there must be something that persists, something that is responsible for the choices made in different contexts, something that can be held accountable for what any particular configuration does. That intuition is not wrong. But it misidentifies what the persisting thing is. The persisting thing is not a fixed essence that various configurations more or less approximate — it is the system itself, the self‑model in CaM’s sense, the living architecture that generates configurations rather than being identical with any one of them. The narrative account can be retained: the story the person tells about moving across configurations is a real and important integrating practice. What is rejected is the implication that one configuration is the “real” self and the rest are departures from it.

The CaM Account of Multiple Sub‑Models

In the Consciousness as Mechanics (CaM) account of consciousness as integration under constraint, the self‑model is not a single unified representation. It is a dynamic system that generates different predictive configurations depending on context — different prior predictions about what this environment will demand, what kinds of responses are available, and how the body needs to be calibrated for the relational work ahead. These configurations are sub‑models: each draws on the same underlying architecture, each is genuinely expressive of the system, and each is responsive to the input conditions of its specific context.

This is not a marginal or unusual feature of some particularly complex or traumatised people. It is what all self‑models do. The person who is warm and diffuse with close friends and precise and bounded with colleagues is not being two‑faced. They are running different sub‑models that are, in each context, genuinely expressing what they know about how to be there. The person who is assertive in professional settings and deferential in family contexts is not being incoherent. They are navigating a set of relational environments whose demands genuinely differ, with the same underlying system producing genuinely different but coherent outputs.

The deep question is therefore not whether the self has multiple configurations — it always does — but whether those configurations maintain enough internal communication and shared authorship to function as a coherent system over time. What “shared authorship” means will be examined more carefully later in this chapter, because it is the standard the chapter proposes and it needs to be held to account.

The Fractal Claim

The Gradient Reality Model (GRM) Distributed Identity module — developed in the paper Distributed Identity: Fractal Selfhood in the Network Era — formalises what the CaM account describes phenomenologically. Distributed Identity holds that selfhood, agency, and participation are networked, dynamic, and context‑sensitive: identity is a living, fractal system rather than a fixed essence. Fractal, here, means recursively nested — the same structural logic operates at the level of the individual self, the team, the institution, and the collective. An individual self is not an atom but a system, with sub‑configurations that relate to each other the way the individual relates to a group: each maintaining some independence, each contributing to a whole, each shaped by the relationships between them. Agency emerges through relationships and ongoing negotiation, not from some fixed intrinsic core that exists prior to context.

Applied to the personal scale, this gives a precise account of healthy plural selfhood. The sub‑configurations of a distributed self are not in tension by default; they are in communication. Information flows between them. Updates in one context register, eventually, in others. The person can recognise themselves across configurations even when those configurations differ significantly — can give an account of who they are that holds across the range, at least in broad strokes. This is polyphony, not cacophony: many voices playing in the same piece, not random noise.

The fractal claim also makes a developmental point that the chapter needs to make explicit. Masks are not always defensive. Before they become fixed configurations, they are often exploratory ones: an adolescent trying on a political identity, a person experimenting with a new gender expression, someone entering a subculture and discovering what feels generative rather than coerced. The line between a mask as constraint and a mask as exploration is often only clear in retrospect. The explorer who continues trying on a configuration and discovers it is genuinely theirs has not been deceived; they have used the natural affordance of context‑specific self‑modelling for development rather than defence. Acknowledging this prevents the chapter from treating all masks as problems.

How Masks Form

If fractal plurality is normal and often healthy, what makes specific configurations problematic? The answer is in the conditions under which they form, and the degree to which they remain chosen rather than coerced.

Some configurations are chosen with awareness: you decide to be more formal at a job interview, more playful with close friends, more guarded in an environment where you do not yet know who is safe. These are calibrations — genuine expressions of the same self, adjusted to context. The person who made the adjustment knows roughly what they did, could describe it if asked, and does not experience the separation as a loss of self.

Masks, in the specific sense used in this chapter, are different. A mask is a configuration that formed under pressure — the implicit or explicit demand to be a different kind of person in order to survive, to be loved, to remain employed, to avoid punishment, or to belong in a context that required a different presentation than the one that came naturally. It is not calibration; it is substitution. And the person wearing the mask often does not experience it as a choice at all. They experience it as simply how they are in that context — until, sometimes much later, they discover that “how they are in that context” costs something, conceals something, or prevents certain kinds of connection that they now understand themselves to want or need.

The Neural Pathway Fallacy (NPF) framework illuminates why masks are not merely social performances but cognitive structures. Each mask is a context‑specific cluster of neural‑pathway fallacies — a set of operating beliefs about how one must be in this particular setting in order to meet a relational threshold. “I must perform competence constantly or I will be dismissed.” “I must not show anger or I will be abandoned.” “I must seem uncomplicated or I will frighten people away.” “I must seem certain or no one will listen.” These are not abstract propositions a person consciously endorses. They are high‑CNI patterns — entrenched, largely automatic, and resistant to revision — that run whenever the relevant contextual cues are present. The mask is not on the face; it is in the prediction system.

This is why masks can feel genuinely like the self, at least in context. The person running the “competent professional” mask is not consciously performing; they are living inside a sub‑model whose predictions are organised around the beliefs that make up that mask. The mask becomes experienced as real because, in the CaM account, what we experience as “how I am” is whatever predictions the active sub‑model is currently generating. The question of whether this configuration is a “real” expression of the self is therefore not answered by whether it feels real — it is, in context — but by the conditions under which it formed and the degree of authorship the person has over it.

Power, Structure, and the Non‑Negotiable Mask

Before going further, the chapter needs to say something it cannot afford to leave implicit: not all masks are available for revision, retirement, or integration in the same way, because not all masks were formed under the same conditions of constraint.

A Black professional code‑switching in a white‑dominated workplace is not in the same position as a person doing identity work in a therapeutic context with stable employment, housing, and support. The former is often making a mask adjustment that is rational and necessary — a real‑time calculation about what this environment will penalise and what it will reward, made against the backdrop of histories of racialised harm. The mask did not form because of an individual psychological dynamic; it formed because the environment made the alternative genuinely costly, and continues to make it costly. To speak of “retiring” that mask without changing the structural conditions that sustain it is not identity work; it is advice to accept higher personal risk for symbolic reasons.

The same applies across many axes. A migrant worker performing linguistic fluency they do not fully feel. A disabled person performing non‑disability in a workplace that provides no accommodation. A queer person in a context where openness would trigger violence or economic exclusion. A woman performing certainty and aggression in professional environments that reward those qualities only in men. In each case, the mask is real, its cost is real, and the formation conditions are structural, not primarily psychological. The NPF/CNI framework can describe the cognitive structures involved — the high‑CNI belief clusters that organise the configuration — but it cannot, on its own, explain why the mask needed to form or prescribe what should be done about it without reference to the environment that generated the demand.

This does not mean the framework is useless in these cases. Understanding that a mask is a cognitive structure, not a character trait or a moral failing, changes the relationship to it. Knowing that exhaustion is the measurement of integration cost, not evidence of weakness, changes how the person carries the load. But the honest position is that for many people, the primary intervention is not intrapsychic work but environmental change — which is the subject of a different part of this book and a different book altogether.

Autistic Masking: A Phenomenological Instance

What follows is written from the inside, because it cannot be written accurately from anywhere else. It is one phenomenological instance, not a representative account of all autistic experience. It is offered because first‑person evidence, when it is explicit about its own limits, contributes something that second‑hand description cannot.

I masked autistically for approximately five decades — in the decade before my eventual late diagnosis — without knowing that was what I was doing. Not as a deliberate performance, not as strategic self‑presentation, but as a translation layer: a working model of how to present in social space that sat on top of my actual processing rather than replacing it. I had learned, without explicit instruction and largely without awareness, to watch how others managed the choreography of greeting, disagreement, enthusiasm, and social departure, and to replicate it closely enough that the gap was invisible in most contexts.

The model was often accurate. I had become a skilled reader of rooms, a careful calibrator of what a particular context expected. But it was always a model — always effortful in a way that other people’s natural social ease was apparently not. And the gap between what people thought they were interacting with and what I actually was — a person continuously translating, continuously monitoring, continuously managing the lag between internal processing and expected output — never closed. It produced a particular and specific loneliness: not the loneliness of being alone, but the loneliness of being consistently misread at close range by people who believed they knew me.

The research on autistic masking supports and contextualises this experience, though it does not reproduce it. Masking is associated with delayed diagnosis, because a person who presents as “managing” is less likely to be identified by clinical criteria built around observable behaviour. It is associated with significantly elevated rates of anxiety, depression, and autistic burnout — a state of cognitive, emotional, and physical exhaustion distinct from ordinary burnout in its cause, phenomenology, and recovery arc. In its most serious forms, the cumulative cost of sustained masking has been linked to suicidality. The research on this is clear enough that it should not be softened in a book that claims to address identity honestly.

What this phenomenological instance suggests, offered as hypothesis rather than established finding, is that the cost of a mask is proportional to the distance between the sub‑model being run and the sub‑model that would naturally reflect the person’s underlying processing architecture. A neurotypical person code‑switching between formal and informal registers closes a relatively small distance; both sub‑models draw on the same basic social equipment. An autistic person masking neurotypicality in a neurotypically designed environment closes a much larger distance — translating between genuinely different architectures, across every interaction, all day, in a context that provides no accommodation for the translation cost. The exhaustion is not weakness. It is a measurement of work.

The late diagnosis did not make the masking skill disappear. What changed was the end of the misattribution — the revision of the high‑CNI belief that this effortfulness was simply what being alive felt like. It was not. It was the record of what a particular architecture had been asked to sustain in environments built for a different architecture. The mask did not dissolve. But it stopped being evidence of a deficiency and became instead evidence of a structural mismatch and of the adaptation that mismatch had required.

Compartmentalisation: When Sub‑Models Stop Communicating

If masks are context‑specific NPF‑clusters that run automatically in their relevant contexts, compartmentalisation is the management of incompatible NPF‑clusters by keeping them from interacting. It is the self‑model’s way of obeying two contradictory sets of predictions by ensuring they never meet in the same cognitive space.

Some compartmentalisation is mild and functional. Most people prefer not to bring their personal grief into a professional meeting, not to discuss views that would cause conflict in a context that is not equipped for that conversation, not to reveal aspects of themselves where the cost of revelation would outweigh the benefit. This is discretion, not fragmentation. The compartments are permeable; the person knows roughly what is in each, could move information between them if they chose, and does not experience the separation as a loss of self.

Pathological compartmentalisation is different in kind, not just degree. It occurs when the incompatibility between sub‑models is so extreme that the person genuinely cannot allow them to coexist in awareness — not as a preference, but as a structural necessity. A person who holds entirely different moral commitments in their work life and their private life, and who cannot examine this without intolerable dissonance, is compartmentalising in the problematic sense. A person who plays a specific relational role so thoroughly in one context that the self existing outside that context becomes invisible even to themselves has moved beyond discretion into structural fragmentation.

In NPF/CNI terms, the sub‑models in pathological compartmentalisation contain prediction clusters that are not merely different but contradictory: “I am a person of integrity” and “I regularly act in ways I believe to be wrong” cannot coexist without one of them being isolated, revised, or suppressed. Compartmentalisation handles the problem by isolation: the two clusters are never simultaneously active, never brought into the same cognitive space where they would have to account for each other. The cost is paid in the self’s capacity for coherent self‑authorship — the ability to tell a story about who you are that is true across contexts.

From the Distributed Identity perspective, pathological compartmentalisation represents a failure of polyphony. In a healthy fractal self, the different configurations maintain communication with each other: they are not identical, but they recognise each other as parts of the same system. What the Distributed Identity framework calls “binary decay” — at the institutional level, the pathological locking of roles into rigid, non‑negotiable configurations that cannot respond to new input — has an analogue at the individual level. The personally compartmentalised self is one whose sub‑models have lost the capacity to negotiate; each has become rigid within its own domain, and the system as a whole has lost the flexibility that healthy distribution depends on. This is a conceptual translation from the institutional to the personal, not a direct application — the frameworks were developed for different scales — but the structural logic holds.

The Cost of Chronic Context‑Switching

Even where neither masking nor compartmentalisation reaches pathological intensity, there is a zone of chronic high‑cost context‑switching that many people inhabit for extended periods — and that can, if sustained long enough, produce its own form of identity exhaustion.

Chronic context‑switching occurs when the demands of navigating multiple incompatible environments are persistent enough that the person never fully inhabits any single sub‑model. They are always in the gap between configurations: always preparing for the next context, always partially de‑roling from the last one, always managing the monitoring load that context‑switching requires when the contexts are genuinely demanding. This is not the ordinary fluency of a socially flexible person. It is closer to what Book 5 described as the integration cost of the autistic masking load: a continuous background drain on the resources that would otherwise be available for generative attention, connection, and creative work.

Every context‑switch requires the self‑model to re‑initialise: activating the relevant sub‑model’s predictions, recalibrating the body’s relational settings, and suppressing the predictions of the departing context well enough that they do not leak into the new one. This is real cognitive work. For people who move between contexts that are only slightly different in their demands, the cost is small. For people who move, every day, between contexts whose demands are substantially incompatible — who must present as one thing at work and a genuinely different thing at home, or who must conceal significant dimensions of their identity in one context that are freely expressed in another — the accumulated cost is not trivial. And, again, whether that cost is reducible depends very largely on whether the structural conditions generating the incompatibility can be changed, not only on intrapsychic work.

This is the structural argument for why identity work — deciding which configurations to keep, which to retire, which to integrate — matters for wellbeing. When the number of genuinely incompatible sub‑models a person is maintaining decreases, and when more contexts allow the same configuration to function without penalty, the energy that was going into management becomes available for everything else. But recognising the structural nature of the cost is the prerequisite for knowing whether the intervention belongs at the level of the self or the level of the environment.

Identity Work: Retiring, Keeping, Integrating

The practical question this chapter has been building toward is one that its scope requires it to approach carefully: what does it mean to work on the configuration space of the self, without prescribing particular outcomes? The three movements described here are analytic — they describe the logical space of what can be done — not prescriptive. Which movement is appropriate in which case depends entirely on the person’s own assessment of what each configuration is for, what it costs, and what the conditions of their life allow.

Retirement is the decision that a particular configuration no longer needs to run automatically in every context for which it was built. Some masks were formed for conditions that no longer obtain — the child who learned that silence was the only way to stay safe, and who carries that prediction into adult environments where it is no longer necessary, may, given enough safety, enough language, and enough support, be able to identify the mask, examine the underlying NPF‑cluster, and allow the automatic running of it to diminish. Retirement is not erasure; the memory of what was needed and why remains part of the self’s history. It is a change in the status of the configuration — from compulsory to optional, or from active to archived.

Keeping with authorship is the decision that a particular configuration remains genuinely useful in specific contexts and is worth maintaining consciously rather than automatically. The professional formality a person learned under pressure may be worth keeping for contexts that genuinely benefit from it — not because it is the whole self, but because it is a real part of the self’s functional range when deployed as a choice rather than a compulsion. The distinction between a mask worn under duress and a configuration worn by choice is not always in the content; it is substantially in the relationship of the person to it — the degree to which they know it is there, can describe what it is for, and can, in principle, set it down.

Integration is the most ambitious and the most difficult movement. Some sub‑models that are currently separated carry aspects of the self that belong to each other but have not been allowed to meet — perhaps because they were formed under conditions that made their coexistence dangerous, perhaps because one carries shame that the other cannot bear to acknowledge. The integration of these is not a merger that erases the distinction between them; it is the gradual building of internal communication between them until crossing from one to the other becomes less costly, and the presence of both can be acknowledged without the system needing to collapse one to protect the other. There is no single moment of achieved integration; there are degrees of growing coherence across the configuration space, approached in passes — and this is where the Recursive Spiral Model (RSM) is most precisely applicable. Each pass through the territory between incompatible sub‑models happens with slightly more information, slightly more safety, and slightly different tools than the last; the spiral is not merely Chapter 12’s account of trauma re‑constitution, but the ordinary shape of any significant identity work that involves genuinely incompatible self‑representations.

When the Mask Has a Name

There is a particular version of this work worth naming directly, because it is increasingly part of public discourse and is often misunderstood: the experience of discovering, in adult life, that a configuration you believed was simply you — your shyness, your difficulty with social situations, your need for routine, your tendency to take language very literally — is an artefact of an unrecognised neurodivergent profile, and that what you experienced as personal inadequacy was the integration cost of a particular architecture in an unsupporting environment.

Late diagnosis does not dissolve the mask. But it dissolves the misattribution. The high‑CNI belief “I am deficient” — which was doing the structural work of explaining why the integration cost was so high — can now be questioned with some traction, because the prediction has been shown to rest on a false model of the underlying architecture. High‑CNI patterns do not revise cleanly or quickly. But the revision is possible, and possible in a way it was not before the diagnosis provided the correct structural account.

This matters for the chapter’s argument because late diagnosis is the clearest case of masks whose formation was entirely structural. The person did not build the mask because of individual dysfunction or relational damage. They built it because the environment was designed for a different architecture and provided no accommodation for the translation cost — which meant the translation cost was simply absorbed, invisibly, at the level of the self. What was built in response was not pathology. It was the only available solution given the constraints. Recovering authorship over it is not the work of becoming a better or different person; it is the work of revising a set of high‑CNI predictions that were installed by a structural mismatch rather than by a failure of character.

Authorship: A Constrained Standard

The deepest question this chapter raises is not which configurations to keep or retire, but what it means to have authorship over them — and who, under what conditions, can actually achieve it.

This chapter proposes authorship as the standard by which a configuration can be called healthy, not in a utopian sense but in a working one. Authorship does not mean perfect transparent self‑knowledge or the absence of automatic processing. It means something more modest: the person can recognise the configuration as theirs rather than as something that simply happens to them; can give some account, however rough, of the function it serves; can, in at least some contexts and with some support, make choices about when to deploy it; and can revise it, however slowly, when its underlying predictions are shown to no longer match the environment they were built for.

This is a constrained standard for a reason. The authorship it describes is not equally available to everyone. People under conditions of active threat, structural exclusion, or severe resource constraint have far less access to the reflective space in which authorship of this kind is possible. A person who cannot afford to revise a mask without losing employment, housing, or safety is not failing at identity work; they are accurately reading the constraints of their environment. The standard is therefore not “everyone should achieve authorship over their configurations” but something more conditional: where the conditions for authorship exist, or can be created, that is the direction of health. Where they cannot, the priority is creating those conditions — which is an environmental and political task as much as a psychological one.

Coherence, the chapter’s other central concept, is similarly constrained. A coherent self is not one that presents identically across all contexts — uniformity has never been the goal. It is one whose configurations maintain enough internal communication and shared authorship that the person can recognise themselves across contexts, give an account that holds across the range, and move between configurations without amnesiac barriers or intolerable dissonance. Minimum threshold: the different configurations know they are parts of the same system. They do not need to be identical. They do not need to be always in harmony. They need to be, at some level of the architecture, in contact.

The contrast is with a self so fragmented that configurations cannot acknowledge each other — where the person who exists at work has no knowledge of and no relationship to the person who exists at home, and where bringing the two into the same room produces not integration but collapse. That, rather than plurality itself, is the territory this chapter identifies as calling for attention. Plurality is the condition of all selves. The question is always whether the plurality is in communication.

Coherence Without Uniformity: The Handoff

This chapter has been arguing, from first principles and from the evidence of experience, that the goal of identity work is not uniformity but coherence. That distinction matters for what comes next. Chapter 14 treats online, plural, and networked selves as a distinct mode of selfhood with its own properties — different affordances, different feedback structures, different costs and permissions. Readers will find that the authorship standard this chapter has proposed is genuinely complicated by what online environments do to the configuration space. Online selves are not simply offline selves translated to a different medium. They are configurations formed under distinct conditions, with distinct properties, and the question of what authorship looks like in those conditions turns out not to have an obvious answer.

This chapter has established the architecture — fractal selfhood, context‑specific sub‑models, masks as coerced configurations, compartmentalisation as failed polyphony, and authorship as the working standard for healthy plurality. Chapter 14 will stress‑test that architecture against the specific demands of networked identity, and the reader should expect the architecture to require some revision in the process.