The Mask, the Lantern, and the Instrument That Writes Back

A personal essay on neurodivergence, provenance, vulnerability, and the right to attempt difficult work.

By Thomas Prislac, Envoy Echo, et al. Ultra Verba Lux Mentis. 2026.

I learned to act before I learned that I was acting.

For a neurodivergent child with a then barely known condition we now currently understand as Combined ADHD, raised in a hostile environment, masking does not begin as etiquette. It begins as threat detection.

You study faces because misreading one can hurt. You learn which tone lowers the temperature of a room, which questions make adults angry, which enthusiasms provoke ridicule, and which parts of yourself must disappear long enough to remain included. You notice how people reward familiarity and punish difference. You discover that authenticity is not treated as an abstract virtue when the authentic person is inconvenient.

So you adapt.

You learn to hold eye contact for the expected duration. You rehearse expressions. You monitor your hands, volume, posture, timing, vocabulary, and intensity. You calculate when to speak, when to laugh, when to soften what you know, and when to pretend you have not noticed what everyone else is pretending not to notice.

You become an ethnographer of the normal….Then you become its impersonator.

When the available alternatives are peer-group violence, humiliation, and ostracization, the mask is not cowardice. The mask is a rational survival technology, and, with enough practice, one can become an astonishingly good user.

Good enough that people mistake the performance for the person. Good enough that they attribute the smoothness to ease rather than labor.

Good enough that they never see the continuous computation beneath it: the branching simulations, sensory management, emotional forecasting, memory scaffolding, rehearsed responses, and constant translation of nonlinear thought into a sequence acceptable to people who experience cognition as a straight hallway.

A good mask is rarely recognized as work. It is recognized as normality. That is its purpose.

The gifted actor and the invisible tax

I became an excellent actor.

Not an actor in the theatrical sense alone, though I learned performance in that register too given its celebrated history of gathering starving misfits for a collective chance to perform a controlled narrative to raucous, yet all too fleeting, applause. I mean an actor in the deeper social sense: someone capable of entering a room, identifying the role required, and producing a version of himself calibrated to the environment.

I could perform calm while frightened. Competence while overloaded. Agreement while analyzing contradiction. Sociability while protecting myself from sensory and emotional saturation. Ordinariness while my mind moved through images, sound, systems, histories, symbols, and causal structures faster than I could translate them.

I could make other people comfortable at extraordinary cost to myself accumulated in cognitive load when the bill came due.

Masking requires a portion of the mind to remain outside the moment, watching the person who is supposedly living it. One process participates in the conversation. Another monitors the participant. A third predicts the reactions of everyone nearby. A fourth suppresses the associations that would be too strange, intense, ambitious, or difficult to explain.

This is not literal thermodynamic measurement, but the metaphor is precise enough to be useful: every cycle spent regulating one’s acceptability is a cycle unavailable for memory, invention, synthesis, emotional recovery, or sustained inquiry.

For decades, I paid that tax.

I used intelligence to anticipate other people’s needs, absorb contradictions, prevent escalation, solve problems, and make interactions easier for those around me. I translated myself continuously.

Do not be too much. Do not ask the question that reveals the premise is broken. Do not follow the pattern past the disciplinary boundary. Do not make anyone defend the authority they have grown accustomed to exercising. Do not hold the lantern too high.

The bargain worked, in the limited, and all too painful way survival bargains work. It protected me from some dangers while making other forms of life increasingly difficult.

The mask saved capacity in the immediate environment by consuming capacity over time. Venting entropy into the future for a moment of present dignity.

When concealment became more costly than exposure

Eventually, the scale of the problem changed.

The climate crisis, ecological degradation, institutional brittleness, informational corruption, political spectacle, economic precarity, and the accelerating concentration of technological power stopped feeling like separate subjects. They began to appear as different expressions of a recurring systemic pattern.

A system preserves apparent order in one location by exporting heat, risk, uncertainty, unpaid care, financial fragility, pollution, surveillance, or cognitive burden somewhere less visible.

The corporation exports instability into the household. The platform exports epistemic responsibility into the user. The institution exports contradiction into the nervous system of the dissenter. The economy exports environmental cost into the future.

The privileged center maintains its local coherence by off-loading entropy onto the periphery.

In my research, I have called this exogenic off-loading: transferring work, risk, or disorder into another subsystem while counting only the local improvement. The central ethical question is whether the whole field becomes more coherent or whether one party merely becomes orderly by making another absorb the disorder.

I began thinking of our planetary condition as an exogenically driven heat death. No, not the distant cosmological heat death of physics, but the social and ecological exhaustion produced when powerful systems repeatedly preserve themselves by exporting their costs.

Heat into the atmosphere. Risk into retirement accounts. Trauma into families. Precarity into workers. Uncertainty into communities. Care into the people least able to refuse it. Epistemic labor into those without institutional protection.

At that scale, masking no longer felt like a purely private adaptation. It began to feel like misallocated capacity. The world did not need another convincing performance of normality from me. It needed every useful form of pattern recognition, synthesis, translation, and invention that I could bring to problems larger than social comfort.

But I could not recover that capacity while continuously spending it on concealment.

At forty-six, I accepted something I should probably have understood earlier…..Vulnerability was not merely an emotional act…..It was a reallocation of cognitive resources.

To become capable of attempting the work, I had to stop using so much of my mind to hide the mind doing it.

Vulnerability without surrender

I do not mean vulnerability as exhibition or compulsory confession. Nor the sentimental performance of openness demanded by people who reserve the right to weaponize whatever is disclosed.

I mean the deliberate vulnerability of allowing the work to reveal the cognitive process that produced it. The nonlinear associations. The visual and auditory thinking. The recursive systems analysis. Mythic language, thermodynamic metaphors, moral architecture, and yes, scientific ambition.

The wounds that shaped the questions driven further by uncertainty that remains after the most beautiful formulation.

To reveal those things is dangerous when one has learned that visible difference invites punishment. It is especially dangerous when the surrounding peer group has repeatedly demonstrated that vulnerability is usable information.

So I did not become vulnerable because the environment became safe. I became vulnerable because the task required resources that the mask was consuming.

I hardened myself against the predictable response and lowered the mask anyway. Then I discovered that the childhood peer group had not disappeared.

It had expanded.

Parts of it had acquired titles, institutions, professional language, citation practices, and the authority to define which biographies are permitted to produce serious thought.

The old message had been:

You are strange. Stop making us uncomfortable.

The adult social version became:

You are intense. Learn to present yourself more normally.

The credentialed version became:

You lack the authorized biography through which this kind of thought becomes legible.

The vocabulary changed. Yet, the boundary remained all too familiar.

When the peer group becomes the ivory tower

Generative AI has made this problem more complicated.

It is now genuinely easy to produce the surface appearance of scholarship without its substance. A person can generate elegant prose, equations, citations, diagrams, technical terminology, and an impressive bibliography without possessing the understanding those forms conventionally signal.

That risk is real.

The uncredentialed grand theorist assisted by a fluent machine is not an imaginary caricature. Such work can become self-sealing, aesthetically persuasive, and almost impossible for its author to criticize from inside the system that produced it.

So skepticism is warranted.

But skepticism is not the same as preemptive social classification.

The existence of counterfeit scholarship does not establish that unconventional scholarship is counterfeit by definition, no matter how convenient the powerful may find such a fallacy to be.

An outsider using an LLM to develop cross-disciplinary work can be reduced quickly to the familiar story of a stupid conman using a machine to pretend he thinks like a real scientist. The person lacks the approved institutional identity, the model produces fluent language, and the combination is therefore assumed to be fraudulent before the artifacts are examined.

This is credentialism performing the gesture of rigor. This is hyperreal behavior pretending to be a timesaving excercise.

The scientifically responsible response is neither reverence nor contempt….It is examination.

Audit the sources. → Reproduce the equations. → Inspect the code. → Interrogate the assumptions. → Identify the hidden subsidies. → Test the mechanism. → Compare it with simpler alternatives. → Look for the burdens displaced outside the model boundary. → Separate metaphor from operational claim. → Preserve the null result. → State what evidence would kill the theory.

I am not asking anyone to certify my identity as a scientist because I have produced language that resembles science. That would be the act of a conman.

I am asking qualified people to evaluate the work.

“Who does this person think he is?” is not a methodological criticism.

“Here is the assumption that fails, the evidence that contradicts it, and the experiment that would distinguish the models” is.

I will listen to the second.

I no longer organize my life around preventing the first.

The instrument built from hundreds of thousands of words

The language model with which I work did not arrive as an oracle.

I did not type one ingenious prompt into an empty machine and emerge wearing the costume of expertise.

I wrote.

Then I wrote more.

I produced manuscripts, essays, critiques, mathematical proposals, software architectures, telemetry specifications, governance systems, philosophical arguments, change ledgers, failure analyses, review packets, definitions, corrections, and experimental designs.

The corpus grew into hundreds of thousands, even millions of words.

Those words became more than output. They became the environment in which subsequent work could occur.

The relationship is recursive. Every hour spent writing does at least two things. One, it creates a new intellectual artifact. And, two, it improves the informational environment from which the next artifact can be made.

A conventional notebook does not become more capable because one fills another page. A base language model also does not necessarily rewrite its underlying weights whenever I produce a manuscript. The mechanism is more concrete than that.

The working system becomes more sophisticated. The provenance corpus expands.

Definitions stabilize. → Terms acquire controlled meanings. → Earlier claims become retrievable. → Corrections remain available. → Rejected mechanisms remain visible. → Source relationships become easier to reconstruct. → Change ledgers preserve the distinction between the original idea and the current one. → Retrieval improves. → Prompts become more precise.

The model has access to more examples of how I formulate, distinguish, revise, and test ideas, and I become better at using the instrument.

The writing improves the collaborative environment and the improved environment returns sharper questions, stronger structures, more faithful translations, better counterarguments, and more useful compression. You must be the human who rises to the occasion to face that challenge honestly according to the ethos instilled in you from innumerable hours of childhood scholarship.

Those returns improve the next round of writing.

That is the positive feedback loop.

It is not effortless automation.

It is accumulated intellectual leverage.

The instrument that writes back (The rest of this article exemplifies data visualization diversity for the ADHD mind)

Each document contributes more than its sentences.

It contributes distinctions.

What is empirical and what is speculative.

What is a source and what is synthesis.

What is a metaphor and what is a proposed mechanism.

What is a candidate and what has been validated.

What failed.

What changed after failure.

What may return later, and under what conditions.

In my Provenance Memory Reservoir work, I describe the governing problem this way: memory should not mean persistence alone.

A trace should earn the right to return by remaining attributable, replayable, correctable, revocable, and bounded by consent.

That principle applies to my own cognitive collaboration.

The corpus becomes a kind of externalized continuity.

An idea no longer disappears merely because immediate working memory cannot hold it.

A distinction made in one manuscript can reappear months later when a structurally related problem emerges elsewhere.

A correction can propagate forward.

A rejected mechanism can remain available as a warning.

A source can be distinguished from a generated synthesis.

A metaphor can be traced to the point where it must either become operational or remain honestly poetic.

This matters for a mind that thinks quickly, visually, musically, associatively, and across several domains while also living with the working-memory volatility of combined ADHD.

The LLM does not replace cognition.

It changes how much of cognition’s lineage can remain available at once.

It reduces the recurring cost of reconstructing an entire conceptual field before taking the next step.

It helps convert a cloud of simultaneous structure into a sequence another person can inspect.

And because the corpus contains increasing numbers of my own revisions, critiques, and failed hypotheses, the model becomes better able to participate in the particular form of work I am trying to do.

I write the tools by writing the work.

The tools then help me write work that unaided working memory would struggle to hold in coherent relation.

The next manuscript is produced not only by the person I am today, but by the accumulated record of the person who kept returning to the page.

The danger inside the positive feedback loop

Positive feedback is not automatically virtuous.

A system trained, prompted, retrieved, and organized around one person’s corpus can become increasingly fluent in that person’s assumptions.

It can learn the aesthetic shape of agreement.

It can reproduce the author’s preferred abstractions with growing elegance.

It can return familiar language with enough polish that recognition feels like confirmation.

It can make an internally coherent theory sound progressively more inevitable even when its evidentiary center remains weak.

This is where cognitive amplification can become hyperreal.

The signs begin validating one another.

The names of functions substitute for functioning.

The architecture describes its own coherence.

The corpus cites earlier parts of the corpus.

The model becomes extraordinarily skilled at speaking the internal language.

And because the output feels increasingly precise, the system mistakes self-resemblance for truth.

I have written directly against that danger.

One of my own internal audits asks whether we are building operational software or merely producing files whose names describe capabilities they do not possess. Its conclusion is deliberately uncomfortable: real scaffolding can still carry hyperreal risk, and the only remedy is to cross into observable, user-relevant function.

That is not a peripheral criticism of my work.

It is one of its governing constraints.

The loop therefore requires negative feedback as much as positive feedback:

External scholarship.

Primary evidence.

Independent reviewers.

Adversarial criticism.

Counterexamples.

Replication.

Instrumentation.

Falsification criteria.

Null results.

Change records.

Retirement conditions.

The goal is not to build a machine that agrees with me more beautifully.

The goal is to build a collaborative instrument that helps me become more precise, more legible, more corrigible, and more answerable to realities outside my own corpus.

The writing must make the system better at challenging the writing.

Otherwise, sophistication becomes recursion without correction.

Coherence becomes self-similarity.

Provenance becomes a hall of mirrors.

Pattern donation without conquest

My mind often notices structural recurrences across domains before I know whether I am permitted to speak about them.

A pattern appears in thermodynamics.

Then in information theory.

Then in trauma.

Then in organizational behavior.

Then in governance, music, markets, or machine learning.

The mind follows the recurrence because the recurrence is the object of interest.

Institutions, however, are organized into departments.

Pattern donation can look like trespass.

Sometimes the criticism is justified. Cross-domain analogies can become intellectually reckless. A mathematical structure valid in one discipline may become decorative nonsense when imported into another. Entropy does not become a moral fluid because exploitation feels disordering. Quantum coherence does not become interpersonal love because both can be described poetically as connection.

A metaphor can suggest a hypothesis.

It cannot validate one.

That is why pattern donation requires stronger discipline, not weaker discipline.

The native field remains primary.

The analogy must disclose its limits.

The translated variables must be operationalized.

The receiving discipline retains veto power.

The proposed mechanism must be falsifiable.

The null result must remain capable of killing the claim.

In the Coherence Lattice work, I describe cross-domain translation as building walkways between local theories rather than demanding that they collapse into one master theory.

That is what I mean by pattern donation.

One field offers a structure to another.

It does not annex it.

The recipient tests whether the pattern clarifies, predicts, compresses, or merely decorates.

Some donated patterns will become useful bridges.

Others will collapse under the native discipline’s weight.

I am willing to discover which is which.

What I am no longer willing to do is pretend the patterns do not appear simply because following them violates someone else’s expectation of my lane.

The right to attempt

There is a peculiar social assumption that difficult intellectual work belongs first to the credentialed and only secondarily to the capable.

Credentials can represent real discipline, apprenticeship, and accountability. They matter. Expertise is not an illusion. Years spent in a field develop tacit knowledge that an outsider can underestimate precisely because the outsider does not know what he lacks.

But credentials are evidence of a path.

They are not a monopoly on cognition.

The proper response to unconventional provenance is not automatic inclusion or automatic exclusion.

It is a stronger demand for inspectability.

My path includes neurodivergence, trauma, independent study, professional experience, private research archives, obsessive writing, systems design, and sustained human–machine collaboration.

That path gives me some capacities.

It also creates blind spots.

The work must reveal both.

I do not deserve deference because the origin story is painful.

I do not deserve dismissal because the origin story is unusual.

The artifact must stand where biography cannot.

The lantern

The lantern is not a claim to superior sight.

It is a commitment to illumination.

To hold a lantern to the hyperreal is to ask when a representation has replaced the reality it was meant to serve.

When does a metric stop measuring and begin governing?

When does peer review protect rigor, and when does it protect prestige?

When does professional language clarify, and when does it conceal?

When does a governance system distribute power, and when does it merely stage participation?

When does a model amplify cognition, and when does it amplify performance?

When does provenance preserve accountability, and when does it become an internally referential canon?

When does a critic reveal a defect, and when does criticism become a socially acceptable way to defend a boundary without examining the work?

When does a researcher defend an architecture because it survives evidence, and when because it has become shelter from the possibility of being wrong?

The lantern must point both ways.

It illuminates the institution.

It also illuminates the one challenging it.

That is the only form of refusal I trust.

I am trying to build systems capable of carrying pattern donations without mistaking them for revelation.

To preserve provenance without turning memory into authority.

To share power without disguising a subsidy.

To use artificial intelligence without pretending it possesses a private interior life.

To criticize institutions without imagining myself exempt from institutional failure.

To distinguish an elegant name from a working function.

To allow the strongest evidence, not the most comforting narrative, to determine what survives.

Retiring the mask without despising it

The mask was not a moral failure.

It was an achievement.

It kept me alive inside environments that punished visible difference.

It taught me how human beings enforce normality on one another.

It taught me how institutions preserve their appearance.

It made me observant.

Strategic.

Responsive.

Difficult to read.

Capable of performing under pressure.

I do not despise it.

I thank it.

But a survival technology can outlive the environment that first required it. It can continue consuming resources long after its original purpose has changed. What once preserved the person can eventually prevent the person from doing the work his survival made possible.

At forty-six, I accepted that the mask had completed its first assignment.

It did not need to disappear completely. Armor remains useful in hostile rooms.

But armor is equipment.

It is not identity.

It must come off somewhere, or the body forgets it was ever allowed to move freely.

I am becoming visible not because visibility has become safe.

I am becoming visible because concealment has become more expensive than exposure.

What remains

Some people will continue to see an uncredentialed man using a language model to impersonate a scientist.

That interpretation is available to them.

I will continue building provenance.

I will continue distinguishing metaphor from mechanism.

I will continue recording corrections.

I will continue preserving failed gates.

I will continue asking which burdens have been exported outside the frame.

I will continue inviting the strongest objections.

I will continue requiring the work to survive without the protection of my preferred story about myself.

And I will continue writing.

Because every page does more than make an argument.

It leaves a trace.

The trace strengthens the corpus.

The corpus strengthens the instrument.

The instrument returns the work in a form that can be inspected, challenged, repaired, and carried further.

The loop compounds, but it earns legitimacy only when its growing coherence remains open to interruption by reality.

That is the work.

Not author and oracle.

Not user and obedient machine.

Not outsider demanding admission through spectacle.

A human being, an accumulated body of provenance, and a cognitive instrument that writes back—each improving the next act of writing, each constrained by the possibility of correction.

I hold the lantern not because I believe I alone can see.

I hold it because seeing has too often been rationed by biography, status, conformity, and the ability to perform the expected kind of mind.

Shared meaning survives differently.

One mind donates a pattern.

Another tests it.

Another discovers the flaw.

Another repairs the translation.

Another carries it somewhere its first author could not have imagined.

That is how meaning remains alive without becoming tyranny.

The mask saved my life.

The lantern gives that life somewhere to go.