Ishtiaque Zico

Colophon

2026-08-19T00:00:00+00:00

About this manuscript. Escape from Cinema is a web-first manuscript, the first volume of Zicobita. It is a living draft, released in versions (0.1 → 1.0) and locked only when the Bangla spine is complete.

Sources and method. The texts come from my own Facebook posts and the zicobaji blog, 2016–2020 — with one 2015 prelude and one 2020 epilogue. Originals keep their dates. Some were re-lineated or trimmed in 2026; those are marked curated-extract. Promotion, event results, guest biographies, and media coverage are kept in a private archive, not here.

Register. The Bangla is unstandardized on purpose; the colloquial register is the voice, not an error.

Technical. Built with Zola and hand-written Markdown, one stylesheet; published from Git to Cloudflare Pages at ishtiaquezico.com. IndieWeb, boring tech, on purpose.

The author. Ishtiaque Zico is a filmmaker, curator, and teacher turning toward research and an unlearning practice. His short film 720 Degrees was the first Bangladeshi film at the Venice Film Festival (Orizzonti nomination); he played Ajit in A Land Imagined (Golden Leopard, Locarno 2018); he ran Biralpakhi Cine Club, 2016–2019.

Preface

2026-08-19T00:00:00+00:00

I forget, therefore I log.

Escape from Cinema collects poems, prose-poems, visual readings, and short logs written between 2016 and 2020, while I was moving away from cinema — its craft, its festivals, its markets. It is the first volume in Zicobita, the longer archive of my poetic writing.

The book is not about quitting film. It is about watching the default — system, structure, self — and recording the friction in a low-fi, encoded form. Cinema is only the entry lens. The writing keeps reaching past it: to language, labor, class, the market, the plague, the city, the idea of home.

The numbers (১০১.১ through ১০৪.২) are not page numbers. They are logbook releases — the order in which I escaped, cycle by cycle. Read them as a slow descent: rehearsal (মহড়া), then the factory (কারখানা), then the plague (মড়ক).

Before the numbered logs sits পালানোর ইশকুল, a companion cycle. These are texts I wrote in 2016–17 to gather filmmakers around Biralpakhi Cine Club — the community that came before the escape, and the school it never became. I lineated them in 2026; they are marked as such.

The Bangla is primary and stays in its own register — colloquial, guru-chandali, deliberately “incorrect.” That register is the voice; I have not standardized it. English versions and endnotes come slowly, as a second reading rather than a translation.

I keep the original dates and sources. Where I edited in 2026, I say so. I cut the promotion, the results, the applause. What remains is for future me — I forget many things — and for anyone unsure whether a creative life has to stay inside one name.

— ইশতিয়াক জিকো, ২০২৬

Context

2026-08-19T00:00:00+00:00

This page is the ground under the poems — the timeline the logs encode but rarely name.

2016. In April, my employment as head of the film department at Pathshala ended, following a conflict over the department’s direction, authority, and priorities. I read the rupture as a clash of values as much as of employment. Students protested. Weeks later I began Biralpakhi Cine Club, and learned my first feature, Cinema, City & Cats, was selected for the Locarno Open Doors Hub.

2016–2019. Biralpakhi ran seventeen public mojmas, study circles, and workshops — self-funded — until I closed it in 2019.

2017. Script residencies in Italy and the Netherlands under Gyula Gazdag. The feature stalled; I postponed it.

2018. I played Ajit in A Land Imagined (dir. Yeo Siew Hua), which won the Golden Leopard at Locarno.

2019–2020. Ghost work on government documentaries under Biralpakhi Studio; closed during Covid. In late 2020 I moved to Cox’s Bazar and stayed largely silent in the film world since — self-assessed, pivoted.

The career under one name

2026-07-24T00:00:00+00:00

It begins with a question asked through a private message.

Can you read this? Does this argument make sense? What am I missing?

Then the question becomes a shared file. The file becomes a deadline. Three messages later, a grocery list appears in the same thread. The deadline remains.

One person may be better at institutional navigation, meetings, relationships, budgets. The other may be better at research, structure, language, analysis, or anticipating what a reviewer will ask next. Tasks follow available time and competence. Formal recognition follows the authorized role.

Over time, a second workstation forms outside the institution. A laptop at a kitchen table, a message thread that never closes, a conversation held after the formal working day ends and before it begins. It has no organizational identifier. It still produces organizational value.

I call one version of this arrangement single-name labor: recurring, substantive work performed by a partner or household member within another person’s formal occupational role, while authority, pay, accountability and career evidence remain attached to the authorized employee.

There is no second contract. There may not even be a second desk. There is a private channel where the work moves back and forth until it is ready to enter the institution under one name.

A note on privacy

This essay examines an institutional pattern, not a documented case against any person, employer or relationship.

This essay omits names, organizations, locations, dates and identifiable work products. Its examples have been generalized to examine a pattern rather than document a particular case. People familiar with the author may still infer its personal origin; complete anonymity cannot be guaranteed. The endnotes are deliberately autobiographical. They position the author within the pattern rather than claiming analytical distance from it.

This limit is methodological. Evidence can expose the people it is supposed to protect. A public argument about invisible labor should not require unsafe disclosure. The resulting account cannot publicly prove every claim about practice. I accept that limit.

The second workstation

Single-name labor rarely begins with a formal decision. It begins with proximity and a shared deadline.

A contributor may correct a figure in a donor report. Draft a response to a supervisor’s comment. Prepare talking points for a meeting they will not attend. Format official minutes for a consultation they did not join. Review a partner’s quarterly submission. Draft an email that will be sent under another person’s name. Polish the tone of a reply so that it sounds appropriately institutional - consistent, non-confrontational, strategically deferential where required.

All from a private device. All in the same message thread where a grocery list appeared three messages earlier.

The employee may be in a meeting, on a field visit, or managing a relationship that requires physical presence. The contributor sits elsewhere, working inside documents they did not author and cannot sign. The employee reviews the output when the meeting ends. The contributor waits for approval, then returns to whatever the deadline displaced.

The medium dissolves the boundary. Professional and domestic labor share a single channel. A request for data disaggregation sits between a household expense and a question about dinner. The arrangement does not announce itself. It accumulates through a hundred small transfers of task.

When support becomes production

Partners regularly support one another’s work. They listen, encourage, absorb stress, rearrange household tasks and help each other think. Not every conversation is labor. Not every suggestion creates an authorship claim.

The distinction is not whether help occurred. The distinction is whether the support repeatedly entered the production process.

Form of support	Relation to the job	Enters the formal product?
Emotional reassurance	Helps the worker continue	Indirectly
Household labor	Frees time and attention	Usually no
Informal advice	Offers an occasional view	Sometimes
Technical assistance	Solves a defined problem	Often
Editorial or analytical labor	Shapes meaning, evidence or structure	Yes
Single-name labor	Repeatedly performs part of the formal role	Yes, under one name

The threshold is repetition, substance and incorporation.

A partner who once notices a typo has helped. An outside contributor who repeatedly researches, restructures, verifies, drafts or resolves problems may be performing part of the formal role. The boundary will never be perfectly clean. That uncertainty should prevent inflated claims. It should not make the labor disappear.

The name as interface

A nameplate appears to identify a person. It also organizes a system.

Behind the name sit permissions, salary, accountability, performance records, references, promotion routes and ownership of work. The name tells the institution who may enter, who must answer and whose career will absorb the result. The institutional name is an interface connecting a person to permissions, responsibility, recognition and career history. It is carried by an email account, login credentials, an HR record, a staff directory entry, a submission workflow, and the metadata attached to every file the employee touches.

Its default model is simple: one role, one authorized worker, one accountable identity, one accumulating career.

The lived arrangement may be different: one role, distributed thinking, private coordination, one recorded identity.

The institution may never be asked to deny the contribution. Its identity, access and submission systems are designed to record authorized responsibility, not every person who influenced production. What they omit may still matter, even when it cannot simply be added to the official record.

Dimension	Formal record	Distributed practice
Worker	One employee	More than one contributor
Worksite	Authorized workplace	Workplace plus private space
Time	Contracted working time	Contracted time plus unrecorded time
Knowledge	Employee competence	Combined and negotiated competence
Credit	Authorized employee or institution	Contribution may be distributed in practice
Accountability	Employee alone	Employee remains formally accountable
Pay	Salary to the employee	No separate payment
Career evidence	Stored under the employee	Unsafe or impossible to claim

The final document looks complete. Its production history does not.

This is not necessarily a conspiracy. When an output repeatedly arrives through unrecorded outside labor, the institution receives incomplete feedback. The job may appear adequately supported, the output individually produced. Successful delivery can prevent the organization from learning what the role actually required.

Existing terms

The arrangement is not new.

In 1973, Hanna Papanek described the two-person single career: professional careers publicly attached to husbands but sustained through wives’ unpaid social, administrative and intellectual labor.^[1] The professional household functioned as a small production unit while presenting the career as an individual achievement.

Lisa Philipps later used unpaid market labor for direct contributions by family members to market work officially belonging to someone else. She argued that the usual division between paid market work and unpaid domestic work makes this labor difficult to see.^[2]

Star and Strauss examine how work becomes visible through selected indicators rather than through any complete record of production.^[3] In this arrangement, the finished output and the authorized name remain visible. The outside contribution has no institutionally accepted indicator.

Joan Acker described organizational jobs as appearing abstract and disembodied while being built around an implicitly unencumbered worker.^[4] The organization contracts with an individual. The individual’s capacity may depend on care, money, health, technology and other people outside the contract.

Single-name labor is a contemporary subtype of unpaid market labor, not the discovery of a separate structure. Its narrower focus is the institutional identity through which work is authorized, evaluated and carried into a career record. Digital production allows outside contribution to enter a workplace product without giving the contributor a recognized occupational position. A person does not need to enter the office to enter the product.

Papanek’s two-person career assumed a gendered direction: wife supports husband’s career. When the arrangement runs in a different direction - when the contributor is the partner whom prevailing social expectation treats as the earner - the invisibility acquires an additional layer.

This is not simply a gender role reversal. The structural problem is the single-name default, not which gender occupies each position. Nor is the gendered experience symmetrical. In a context where masculinity remains closely tied to visible paid employment, a man outside formal employment may encounter his unrecorded labor as both institutional invisibility and personal failure. That pressure can make the arrangement harder to name publicly.

The default may also operate beyond the household. Skilled labor can enter an institution’s knowledge products while the contributor remains absent from its systems of authorization, accountability and recognition.

Several established concepts describe parts of the arrangement. None should be discarded merely to make a new term appear necessary.

Concept	What it explains	What remains unresolved
Domestic labor	Work maintaining the household and reproducing labor capacity	The hidden contributor may enter the employer’s product directly
Emotional support	Helps sustain the worker emotionally	Does not necessarily enter the formal product
Unpaid market labor	Unpaid family contribution to market production	Career attribution and institutional identity need closer attention
Two-person single career	One career sustained through a partner’s uncredited work	Contemporary digital co-production changes how the work enters
Ghost authorship	Substantive contribution omitted from formal authorship	Many workplace products have no negotiable public byline
Contributorship	Roles separated by actual contribution	Usually applied to bounded outputs, especially research

The term is most useful when four conditions appear together: the contribution is substantive, it enters work produced for or evaluated by an institution, it recurs inside another person’s occupational role, and its contributor cannot safely convert it into recognized experience. Without these conditions, the term becomes too broad. It should not turn every intimate conversation into an invoice.

Unequal accumulation

The hidden contributor may shape the work but cannot formally submit it. The employee submits it but remains responsible for everything inside it.

The arrangement can protect the contributor from direct accountability while denying recognition. Production has been distributed. Formal credit and accountability have not.

It would be inaccurate to describe the employee as simply stealing credit. The employee also carries the formal risk. When the output is questioned, the employee answers. The contributor watches from a private channel.

A career is not only a sequence of salaries. It is an accumulating record. One completed assignment supports the next performance review. One strong product becomes evidence of competence. Competence produces trust. Trust produces responsibility. Responsibility produces a stronger title, reference or application. The output moves forward under the authorized name.

The hidden contributor may develop occupationally valuable skills: research, analysis, editing, quality assurance, strategy, institutional interpretation, deadline management. The learning cannot easily become proof. A curriculum vitae needs dates, roles, organizations, products and references. Those are precisely the details the contributor may be unable to disclose.

The employee accumulates career capital. The contributor accumulates private competence. One person’s capability becomes institutionally legible. The other’s remains difficult to verify.

Meanwhile, the contributor’s own projects occupy the margins of the arrangement. A manuscript, a website, a portfolio, a job search, rest. These are pushed to the intervals between institutional deadlines. When the formal career accelerates, the contributor’s independent work decelerates. The arrangement does not announce this displacement. It is experienced as a sequence: the deadline arrives, the contributor’s afternoon changes direction.

The evidence trap

The ordinary response to invisible labor is: document it.

In this case, documentation creates another problem. To establish the contribution, the hidden worker may need to retain drafts, messages, version histories or institutional files. Those materials may be confidential. Keeping or publishing them could violate privacy, information-security rules or professional obligations.

The strongest evidence may therefore be the least safe to possess.

Possible evidence	What it could establish	Privacy or career risk
Version history	Who drafted or revised text	May expose confidential files and identities
Message history	Requests, timing and division of work	Exposes private and professional communication
Metadata	Creation and revision patterns	Can identify devices, accounts or organizations
Contribution diary	Type and duration of labor	Safer if no confidential content is copied
Final output	Institutional use of the work	May be owned by the employer
Public disclosure	Makes the arrangement visible	May harm the employee and contributor

The contributor is asked to prove erasure through the records whose disclosure could cause harm. The institution recognizes evidence in forms the hidden contributor may not ethically use. This is a legibility trap.

A private reflection must not become an unofficial archive of institutional work. Do not retain institutional files, copied text, message histories, metadata or identifying details. If a personal skills record is lawful and safe, it can remain general: broad period, general kind of skill practiced, approximate time spent, personal reflection on what was learned. It should not identify an employer, employee, assignment, decision or outcome. If even that record could breach a duty or expose another person, it should not be created.

Why the arrangement persists

Single-name labor can persist without beginning in explicit coercion. It may grow through trust, affection, urgency, unequal opportunity, shared household survival or pleasure in solving problems together. The hidden contributor may want to help. The employee may need help. Both may benefit when the formal career remains stable. The household can receive income, security and future opportunity through the one recorded role.

That does not make the arrangement equal. A mutually accepted arrangement can still sit inside an unequal system of recognition. The personal bargain and the institutional structure are different scales. At the personal scale, the work may feel like care. At the institutional scale, the same care may become productive capacity that was neither separately hired nor recorded. The organization sees the completed output. The private sphere absorbs part of what its production required.

The hidden contributor is not automatically a victim. The role can provide proximity to consequential work, intellectual stimulation and a sense of usefulness. It can also allow someone to exercise influence without being publicly accountable for the result.

The contributor may have a history of withdrawing from credit - out of perfectionism, values mismatch, or fear that the output would not match their self-image. The pattern of ghost work may precede the current arrangement.

This particular arrangement may also involve relative privilege: sufficient household income, education, language access and proximity to institutional work. That privilege coexists with economic dependency; it does not erase it. This is not the position of most people whose labor goes unrecognized, and the critique should not present it as universal.

The employee is not automatically an exploiter. They may be carrying an excessive role, navigating an under-resourced workplace or accepting support within an intimate economy where labor and care are not calculated separately. The institution is not automatically aware.

The hidden contribution does not prove the employee is incapable. No serious occupation is performed without assistance. Colleagues review drafts. Supervisors redirect work. Teams divide tasks. Templates carry earlier thinking. The difference is not dependence. The difference is which dependencies are authorized, resourced and recorded. An employee who receives extensive help inside the organization remains a collaborative professional. Comparable help outside the organization may contribute to the same output while remaining unrecorded and potentially unauthorized. The same act of assistance changes meaning at the institutional boundary. Inside, it may be teamwork. Outside, it becomes invisible.

Single-name labor is one private repair system for the gap between the role and the person available to perform it. When demands exceed the support available inside a role, the household reallocates labor until the employee can reappear as an individual worker. The repair succeeds when the organization never has to notice it.

There is still an unequal outcome. The institution may receive additional capacity. The employee receives formal accumulation and accountability. The household may receive stability. The contributor may develop private competence. These are not equivalent returns.

Love can explain why the labor is given. It cannot explain why only one career is allowed to retain it.

Under pressure, the arrangement can sharpen. A missed deadline may surface the economic structure underneath. The employee may remind the contributor - explicitly or through silence - that the formal role depends on timely delivery, and that the household depends on the formal role. The reminder restores the sequence. It also reveals the economic dependency that makes refusal difficult. The contributor’s position is sustained by the same income whose production they help maintain. Either person’s withdrawal could destabilize the household. The circular dependency holds.

When naming becomes another capture

Naming the arrangement creates a risk.

The hidden contributor can turn invisibility into a new professional identity: the person who reveals hidden labor. Private contribution becomes an essay, a framework, a lecture or a claim to expertise. The old work then produces a second career indirectly.

This may restore some agency. It may also turn an identifiable person’s employment and vulnerability into raw material.

That is why this chapter does not present identifiable artifacts. The concept must stand without requiring an identifiable career to become its public evidence. Public authorship should not depend on making another person’s private vulnerability portable.

What recognition could mean

Public co-authorship is not always possible, appropriate or safe. Many institutional products do not belong to the employee. External assistance may be restricted. Confidential information may not be shared. Accountability cannot be transferred merely because someone helped produce the work.

Recognition therefore needs more than adding a second name.

Scale	Possible change	Limit
Household	Name the work and negotiate its boundaries	Private recognition creates no formal career record
Individual	Keep a safe, non-confidential skills ledger	Cannot prove ownership of specific outputs
Team	Realistic workload allocation and internal review	Requires adequate staffing and trust
Organization	Record contribution roles where appropriate	Not suitable for every confidential product
Management	Audit roles requiring invisible overtime or outside help	Workers may fear admitting workload gaps
Employment design	Create paid editorial, research and quality-assurance support	Requires resources and institutional change
Recruitment	Assess demonstrated skills through ethical work samples	Does not recover earlier lost credit

Research publishing offers one partial model. The CRediT taxonomy separates contribution into roles such as conceptualization, investigation, analysis, drafting, reviewing and project administration.^[5] It does not solve single-name labor. It shows that singular authorship is not the only possible interface.

A workplace could similarly ask: Who developed the idea? Who researched it? Who drafted it? Who checked it? Who made the decision? Who is accountable? These answers do not need to be identical. They do need to remain lawful, safe and proportionate to the work.

Before accepting a task, an outside contributor can ask:

Is this general support or production of a workplace deliverable?
Does it require confidential information, account access or identity substitution?
Is outside assistance permitted?
Could an authorized colleague or service perform it?
Is recurring assistance concealing a workload or capacity problem?
Can the employee independently verify, revise and defend the result?
What personal work or care will this displace?

If the task requires unauthorized access, confidential material, use of another person’s account or undisclosed production contrary to policy, the task should not proceed. If the role repeatedly becomes unstable without outside production, a dependency or organizational gap may have formed.

Prototype audit

Reflex. Absorb excess professional labor as private care, personal usefulness or household necessity.

Reward. Income stability, deadline completion, institutional approval, reduced conflict and the satisfaction of solving a difficult problem.

Carrier. The singular identity field, access-control system, job description, submission workflow, performance record and curriculum vitae.

Sequence change. Distinguish personal support from workplace production. Do not access confidential materials or institutional accounts. Redirect recurring production needs toward authorized support, realistic workload allocation or explicit role redesign.

Return. A deadline arrives. The employee is overloaded. The hidden contributor can solve the problem quickly. Care restores the old route.

Capture risk. Invisible labor becomes a public identity. The contributor gains recognition by exposing the person whose career the labor helped sustain. The framework itself may convert another person’s employment and vulnerability into material for the author’s intellectual authority. Self-criticism does not neutralize this risk; it may refine the capture.

A career with no reference

The hidden contributor may eventually leave the arrangement with years of practice and no safe answer to a simple question:

What do you do?

At social gatherings, someone asks. The truthful answer may expose another person. The safe answer requires a story told in present continuous tense about work that happened years ago. Nobody checks the details. National statistics may classify the contributor as economically inactive. The contributor may nevertheless be performing work equivalent to a full-time position.

The strongest work cannot enter a portfolio. No organization can confirm the role. The job title never existed. The curriculum vitae holds a gap that grows longer each year the arrangement continues.

When the contributor imagines applying for work, they confront a record that cannot explain what they have been doing. The skills are real. The evidence is private. The references do not exist.

The formal career may conceal how its production was distributed. The unrecorded contributor may possess real competence without an institutionally usable history.

The career moves forward under one name. The other person’s work cannot follow.

Endnotes

References

Acker, J. (1990). Hierarchies, jobs, bodies: A theory of gendered organizations. Gender & Society, 4(2), 139-158. doi:10.1177/089124390004002002 Foundational analysis of how organizations construct apparently gender-neutral jobs around implicitly unencumbered workers.

CRediT: Contributor Roles Taxonomy. NISO. https://credit.niso.org/ A standard for disaggregating research authorship into 14 contribution roles. Applied here as a partial model for workplaces that currently record only one authorized name.

Papanek, H. (1973). Men, women, and work: Reflections on the two-person career. American Journal of Sociology, 78(4), 852-872. doi:10.1086/225406 Foundational analysis of wives’ uncredited labor within husbands’ professional careers as a structural feature of professional households.

Philipps, L. (2008). Silent partners: The role of unpaid market labor in families. Feminist Economics, 14(2), 37-57. doi:10.1080/13545700701880981 Introduces “unpaid market labor” for direct family contribution to market work officially belonging to another household member.

Star, S. L., and Strauss, A. (1999). Layers of silence, arenas of voice: The ecology of visible and invisible work. Computer Supported Cooperative Work, 8, 9-30. doi:10.1023/A:1008651105359 Examines how systems of representation make some forms of work articulable while other forms remain difficult to name or record.

I grew up calling my mother a housewife. The word carried no irony in middle-class urban Bangladesh in the 1980s and 1990s. If someone asked what my father did, I said: teacher. If they asked about my mother, I said: housewife. The word has since been replaced - homemaker, then unpaid care work. Papanek described this structure in 1973: wives who performed uncredited social, administrative and intellectual labor within their husbands’ professional careers. She called it the two-person single career. I read her decades later, recognizing the pattern but not the direction. In her account, the invisible partner was the wife. In mine, the arrangement runs differently. The structural problem - one career absorbing two people’s labor - does not depend on which partner holds the formal role. ↩
Philipps titled her paper “Silent Partners.” The silence is not always imposed. I have withdrawn from credit before - on a media production where I performed the labor from ideation to execution but refused the public credit because the output did not match my self-image. A colleague stepped in and claimed it. I received my salary and remained silent. Product needs a name. Organization needs a face. System wants an identity to make someone accountable for the work. When I declined, the blank was filled. The pattern of ghost work preceded the current arrangement. Philipps helped me see that this is not a personal quirk but a structural position: unpaid family contribution to market work officially belonging to another household member. ↩
Star and Strauss helped me understand that the problem is not concealment but representation. The work exists. It may be visible in version histories, message timestamps, metadata. What is missing is an institutionally accepted indicator - a recognized way of saying this contribution happened and mattered. Their argument is that systems of representation make some work articulable and other work difficult to name. In my experience, the contributor’s labor is not hidden from the employee. It is hidden from the system that records what the employee produced. ↩
Acker’s “disembodied worker” is the employee who arrives appearing individually competent. I have watched this happen from the other side. The preparation happened elsewhere - often the night before, often through the contributor’s labor. The organization sees the result. It does not see the domestic production system that made it possible. Acker wrote about gendered organizations in 1990. The worker she described was implicitly male, implicitly supported by a wife. When the direction shifts, the structure persists but the social legibility changes. In a context where masculinity is tied to visible earning, a man outside formal employment is rarely asked whether his unpaid labor is recognized. He is more often asked why he does not have a job. The question turns institutional illegibility into apparent personal failure. ↩
CRediT defines 14 contribution roles for research outputs, from conceptualization to writing (review and editing). The taxonomy assumes a bounded output with negotiable credit - a journal article, a research report. Most workplace products have no such negotiation. The quarterly report, the talking points, the email to a supervisor - these carry one name by default, not by agreement. CRediT shows that contribution can be disaggregated. It does not show how to disaggregate contributions the institution never knew were shared. ↩

The unlearning framework

2026-05-12T00:00:00+00:00

Summary

Interfaces carry defaults. A default in a workplace record, a writing tool, or a clinical protocol decides whose work counts, which language is recognized as competent, and which bodily needs enter the record. When a default does not fit a person, that person, or someone supporting them, does unrecorded work to finish the task. If the repair succeeds, the record keeps the result and drops the repair. The default then looks adequate.

I define unlearning as the practiced weakening of sequence authority. This is a default’s power to go first, continue unchallenged, and count as valid. I trace the pattern across three encounters: single-name labor, AI-assisted writing, and a sleep test. A negotiated exception changes one encounter. Institutional unlearning changes the default, so the next person does not have to negotiate the same exception.

Where this began

I heard three words years ago: learn, unlearn, relearn. I took the order as wisdom. To make something new, I thought, I must empty the space that held old norms, habits, and traditions.

I worked in cinema then, and had begun to teach it. Istruggled with the unwritten rules of how a film should be made and taught. I had my own ways. I tried them. Some failed. Some held. Unlearning became my word for that struggle.

Then the word went loose. A company or a campaign could attach it to any preferred behavior and make the change sound obvious. I still wanted to question inherited practice. The word no longer told me what I was doing.

So I looked for something I could use before I could explain it. Where does a learned order act? What keeps it in place? What can change? What comes back under pressure?

That search moved me toward interfaces. Cinema taught me that a frame directs attention and leaves things out. Teaching showed me that a curriculum arranges time, sequence, and recognition. Language showed me how one voice is recognized as competent and another as lacking. I saw this between Bangla and English, and within Bangla too. I use interface in a narrow sense, defined below.

AI sharpened the question. A language model is not a neutral channel between intention and sentence. It hands me the first firm version of my thought. Three questions followed me. What does the interface put first? What must the person repair? What disappears when the result looks like success?

Two meanings of unlearning

The word carries two meanings.

The popular meaning tells a clean story. You hold old knowledge. You delete it. You load new knowledge in its place. Learn, unlearn, relearn.

Research finds weak support for the tidy deletion model management borrowed from psychology. Many cases read better as belief change, dropped practice, extinction, or plain learning.^[1] This does not make every use of the word empty. It means deletion cannot be assumed. The phrase stayed popular because it is easy to manage, not because it is proven.

Machine unlearning uses the same word for a different problem: reducing the influence of chosen training data inside a model. I treat machine unlearning as a neighbor, not as evidence about people. The shared word raises a separate design question: how an AI interface exposes, hides, or repeats the work of correction.

The story also assumes you chose what to unlearn. Much of what sits on us was never chosen. It arrived through school, media, language prestige, office templates, and the rewards of institutions. Nobody sat me down to teach me that fluent English signals a sharp mind. The lesson settled through repetition.

Scholars use the word for different projects. Management research asks how knowledge or routines lose their pull.^[2] Decolonial and teaching work asks how dominant knowledge and inherited positions can be challenged.^[3] These traditions differ in scale and aim. I do not collapse them.

My framework studies a smaller unit: a default carried through an interface I can point to. A recent process review asks for closer study of context and change over time.^[4] The three encounters below test what the framework can do. They do not show how common the pattern is.

Here is the difference, side by side.

The popular story	This framework
Remove old knowledge	Weaken a default’s priority
Replace the old with the new	Keep more than one route open
The worker is the unit of change	The person, the interface, and the institution stay in view
Reward speed and adaptation	Ask who performs the adaptation
Measure the finished result	Look at the work behind it
Ask whether the worker changed	Ask whether the institution changed

The difference matters most when pressure returns. Delete-and-replace promises a finished switch. This framework expects the old route to stay, and asks what keeps giving it priority.

Sequence authority

A default is a route made easier than the others. It feels natural because a system keeps rewarding it. The reward can be repetition, convenience, prestige, fear, or fatigue. In time the route becomes what people expect and what counts as valid.

A default is carried through an interface. An interface is a traceable point where rules, categories, or expected steps become something a person must act through. It can be an input field, an identity record, a billing term, a preparation sheet, a sensor setup, a scoring rule, or a report.

I call the advantage of a default its sequence authority. It takes three forms.

Form	What it does	Example
Precedence	One route starts or gets attention first	A writing tool offers a sentence before mine exists
Persistence	One route carries forward with no fresh effort	A protocol moves a patient from prep to report on one track
Recognition	One route makes the entry the system accepts	An institution records one employee name

The forms often travel together. The three encounters below show different mixes.

Sequence authority can serve a valid purpose. It can carry safety, care, and useful standards. It turns into a problem worth auditing when relevant knowledge cannot enter without extra work, when the burden lands on the person with the least power to change the setup, or when the record takes the benefit of a repair and drops its conditions.

After any encounter, three questions follow. Which route went first? What held it there? What must a person do to bring in another route?

Three encounters

Three short encounters. I ask the same three questions of each: what entered first, what held its place, and what the surviving record kept. They test the framework’s range. Fuller audits appears on separate pages.

The career under one name

A deadline arrives. A document sits unfinished. I can see how to repair it. I research. I edit. I make choices about structure. The work moves through talk, private notes, and revisions into a file that later looks like one hand made it. The better I repair the deadline, the less the finished file invites anyone to ask how it was made.

The institution records one authorized employee. That identity holds the login, the submission history, the review, and the career record, while the production behind the work may be shared. The finished file keeps one accountable identity and drops the conditions that produced it.

This has a long history. Hanna Papanek named the two-person career in 1973: wives doing uncredited work for their husbands’ careers.^[5] My case sits in a different arrangement and setting. It raises a related recording problem: shared production enters one authorized identity.

This is not a bid to reclaim credit. It asks how the employment system is built, and how we comply with it.^[6] The leading form here is recognition.

The AI-assisted sentence

I use large language models each day. They draft, translate, code, and help me recover when I am overloaded. They also change the order of my writing.

I ask the model first.

The answer arrives before my rough sentence exists. Its fluency becomes the thing I revise. The question is not just what the model writes. It is when its sentence enters, and which version later counts as the source.

For me, the model returns fluent standard English that is more likely to be read as professional. It can also flatten. A Bangla rhythm goes quiet. A useful hesitation turns into a confident line.

The prestige runs inside Bangla too, not just from English above it. In formal rooms I move toward standard media Bangla and away from my regional speech. That pressure has its own history, apart from English. It shows that the demand to sound acceptable starts before English does.

The deeper problem sits in the design. Training data, product goals, default tone, and rewrite buttons shape what the tool presents as fluent. In my drafts, that can make a situated rhythm or an honest doubt harder to keep. The design question is plain. Does the tool widen my choices, or become the standard I must write against?^[7]

The leading form here is precedence. The generated line becomes the first firm sentence, so it becomes the reference for every edit after. Persistence follows when rewrite buttons carry its words forward.

The sleep test

The test is called polysomnography.

At a hospital billing counter in Dhaka, staff could not find the service until I said that word. Once the term entered the system, a private worry became something the hospital could name, price, and book. Language was the first interface.

The overnight protocol asked for a measurable sleeper. A sheet set my waking time, my daytime rest, and my arrival. Sensors read brain activity, airflow, breathing effort, oxygen, movement, and position. My body does not sleep on command. I use familiar sound to fall asleep. The first setup put clean audio above that need.

I asked to play a soft soundscape at low volume. The technologist checked with the doctor and agreed. I slept after switching to rain with quiet piano.

The test succeeded. The report diagnosed moderate obstructive sleep apnea. Positive airway pressure reduced the recorded respiratory events during titration.^[8]

The report also gave zero minutes to the time I took to fall asleep, while my own note recorded a long wait. It logged zero night visits to the toilet, while my note recorded several. These gaps do not make the report wrong. The clinical fields and my notes serve different ends. Their disagreement marks a border where one record cannot hold what the other saw.

The local repair worked. The record makes that repair hard to see. The leading forms are recognition and persistence.

Local and institutional unlearning

A negotiated exception changes one encounter. Institutional unlearning changes the default. This gap is the core of the framework.

Local unlearning	Institutional unlearning
Changes one encounter	Changes the recurring default
Gives a person another route	Makes another route available to others
May need negotiation each time	Reduces repeated negotiation
Leaves a situated trace	Requires a changed rule, field, permission, or workflow
May leave the institution unchanged	Carries the changed condition into later encounters

An accommodated patient does not build a patient-centered protocol. A careful writer does not change an AI tool. A private log of hidden work does not create fair workload or credit.

Traces can feed redesign, shared inquiry, labor talks, or clinical review. They do not stand in for the authority of the people affected, or for the processes that change rules and budgets.

Ask one question after any repair. Will the next person face the same repair? If yes, the encounter changed and the institution did not.

Try it in three steps

Use this on one small encounter. Pick a default that feels natural and harmless.

Step	Question
Notice	What route enters first, and who must work around it?
Change	What shifts if I pause, reorder, or bring in another route? Hold the gap. Do not rush a purer answer.
Check	What came back, what entered the record, and must the next person repeat the work?

That is the whole practice. Three steps, one encounter. The aim is not to become a better resister. The aim is to record how the default keeps its priority, so the change can move from the person to the design.

Position and limits

I write from Cox’s Bazar, Bangladesh. I am a neurodivergent, multilingual, middle-class Bangladeshi man who has taught film, designed curricula, supported humanitarian knowledge work, and contributed to care at home. This position lets me notice some pressures: cognitive load, language correction, hidden work, medical legibility, and tool dependence. It also bounds me. I cannot see how the same setups land on other bodies and histories.

The method is interface autoethnography. My experience identifies the pressure. The artifact bounds what I can claim. Each case needs a dated encounter, an identifiable interface, and a trace of what I did or refused. I keep counter-cases too, where a default helps, or where a change makes no difference.

Evidence safety sets a further limit. A workplace file can carry confidentiality. A prompt log can hold another person’s information. A medical record can expose identity. The public account shrinks when safe evidence runs out.

This framework does not romanticize neurodivergence. It does not confuse personal work with structural change. Weakening one default in one encounter leaves employment rules, clinical protocols, and platform design in place.

This essay is made inside the contradictions it names. I write in standard English. I revised this text with AI.^[9] The cleaner the page, the more it matters to say so.

Closing

I began in cinema, where I learned that a frame is not neutral. Language showed me that authority runs inside a tongue, not just between tongues. Now I read interfaces everywhere: a prompt box, a personnel record, a billing term, a report field.

One question waits at the end of any refusal. After I weaken a default, what format captures the change? A page, a framework, a citation, a profile. Each gives the change a public life. Each can also hand authority to the person who names it. This essay is part of the problem it names. Versioning does not fix that. It keeps the move visible and open to revision.

The route stays open. The next reader can start where I started: one small encounter, one question, and the will to make the repair visible. Notice what enters first. Try another route where it is safe. Record what changes. Then ask whether the next person will have to ask again.

Glossary

Three terms have a specific meaning in this essay.

Unlearning: The practiced weakening of a default that once felt natural. The old route stays open while its automatic priority drops. Local unlearning changes one encounter. Institutional unlearning changes the default, so the next person does not repeat the same repair.
Sequence authority: The priority an interface gives one route, through precedence, persistence, or recognition. Other routes stay possible, but need extra work to enter, hold, or show.
Interface autoethnography: A critical method that treats my own encounters with interfaces as field sites. My experience raises the question. The artifact bounds the claim.

References

Acker, J. (1990). Hierarchies, jobs, bodies: A theory of gendered organizations. Gender & Society, 4(2), 139-158.

Agarwal, D., Naaman, M., and Vashistha, A. (2025). AI suggestions homogenize writing toward Western styles and diminish cultural nuances. Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems.

Gao, Z., Ajibola, Y., and Thebault-Spieker, J. (2026). The labor of language: Auditing dialectal erasure and the repair burden in Swahili machine translation. Proceedings of the 2026 ACM Conference on Fairness, Accountability, and Transparency.

Grisold, T., Kaiser, A., and Hafner, J. (2017). Unlearning before creating new knowledge: A cognitive process. Proceedings of the 50th Hawaii International Conference on System Sciences, 4614-4623.

Hislop, D., Bosley, S., Coombs, C. R., and Holland, J. (2014). The process of individual unlearning: A neglected topic in an under-researched field. Management Learning, 45(5), 540-560.

Howells, J., and Scholderer, J. (2016). Forget unlearning? How an empirically unwarranted concept from psychology was imported to flourish in management and organisation studies. Management Learning, 47(4), 443-463. https://doi.org/10.1177/1350507615624079

Jackson, S. J. (2014). Rethinking repair. In T. Gillespie, P. J. Boczkowski, and K. A. Foot (Eds.), Media Technologies: Essays on Communication, Materiality, and Society (pp. 221-239). MIT Press.

Klammer, A., Grisold, T., Nguyen, N., and Hsu, S. (2025). Organizational unlearning as a process: What we know, what we do not know, what we should know. Management Review Quarterly, 75, 2147-2171. https://doi.org/10.1007/s11301-024-00430-3

Mignolo, W. D. (2009). Epistemic disobedience, independent thought and decolonial freedom. Theory, Culture & Society, 26(7-8), 159-181.

Papanek, H. (1973). Men, women, and work: Reflections on the two-person career. American Journal of Sociology, 78(4), 852-872.

Philipps, L. (2008). Silent partners: The role of unpaid market labor in families. Feminist Economics, 14(2), 37-57.

Sourati, Z., Ziabari, A. S., and Dehghani, M. (2026). The homogenizing effect of large language models on human expression and thought. Trends in Cognitive Sciences.

Star, S. L., and Strauss, A. (1999). Layers of silence, arenas of voice: The ecology of visible and invisible work. Computer Supported Cooperative Work, 8, 9-30.

Tsang, E. W. K., and Zahra, S. A. (2008). Organizational unlearning. Human Relations, 61(10), 1435-1462.

Tufte, T. (2024). Unlearning communication for social change: A pedagogical proposition. Social Sciences, 13(7), 335.

Endnotes

The endnotes are my second voice. They hold the position, the doubts, and the detail that would slow the main route.

John Howells and Joachim Scholderer tested the claim that unlearning was carried from psychology into management research. They found the evidence thin. Most cases read better as belief change, dropped practice, extinction, or plain learning. I use their critique to set my problem, not to call the word empty. My claim is narrower: this framework names a documented change in a route’s priority, not a hidden deletion process. ↩
The reduced-pull idea comes from Grisold, Kaiser, and Hafner (2017). Tsang and Zahra (2008) tie unlearning to discarding routines. Hislop and colleagues (2014) study conscious letting-go at the personal level. I follow the reduced-pull branch and leave the discarding language behind. ↩
Tufte (2024) builds unlearning for social change around position, relation, and transition. Mignolo (2009) frames epistemic disobedience and delinking: a refusal to think inside colonial frames. Both work at a wider scale than one encounter. I borrow the spirit and stay small. ↩
A 2025 review of organizational unlearning as a process (Klammer et al., 2025) asks for case-based and trace-based methods that show how a practice loses or regains its pull over time. That is the space this essay enters. The recording idea sits in a repair lineage from critical design and infrastructure studies: Steven Jackson on broken-world thinking, and Susan Leigh Star and Anselm Strauss on visible and invisible work. My method adds an encounter and a carrier: a dated moment, a nameable artifact, and a recorded action. ↩
Papanek (1973) studied wives whose uncredited work sustained their husbands’ careers. Philipps (2008) describes unpaid market labor inside families. Acker (1990) shows how a neutral-looking job assumes a worker with no other duties. Star and Strauss (1999) show how some work stays visible and some stays silent. I place single-name labor inside these conversations, aimed at recurring production that enters one authorized identity. ↩
Publication proceeds only with the ongoing consent and full veto power of the person whose employment is at stake. Naming this arrangement carries a cost: it can transfer authority from the worker to the one who writes about it. That transfer is part of the audit, not a footnote. ↩
On homogenization: Agarwal, Naaman, and Vashistha (2025) studied 118 writers in India and the United States and found that suggestions moved the Indian writers toward Western styles. Sourati, Ziabari, and Dehghani (2026) review the wider concern in an opinion piece. Gao, Ajibola, and Thebault-Spieker (2026) find the same in Swahili machine translation, and give the labor a name: repair burden, the uncompensated effort to fix output that does not fit. I use their term with credit. My question is different: what does the surviving record keep after the correction? ↩
The sleep case draws on the preparation sheet, appointment messages, a fourteen-page report, notes I wrote within a day, and remembered talk marked as memory. The report gave an apnea-hypopnea index of 28.7 at diagnosis and 1.0 during pressure therapy. I accept the clinical value of the test. Clinical reading stays with qualified people. ↩
I self-assess as ADHD-combined with autistic traits. A psychiatrist friend who knew me for ten years did not agree. The disagreement is part of my positional record. Diagnosis rests on frames built in particular places, and the DSM is a North American document. In my experience, adult ADHD assessment in Bangladesh is difficult to access. That limit shapes what I can know and claim. AI is access for me and a risk of dependence. Both hold. I wrote this essay through long dialogue with AI. I brought the encounters, the terms, the objections, and the decisions. The tool offered words and shapes that I took, changed, or refused. So this is my opinion, not my finding: sequence authority raises design questions for the companies that build these tools. Does the tool answer before a person forms an intention? Does it treat a regional language or a rhythm as an error? Can a person refuse fluency without losing access? Does repeated human repair change the product, or must each person repair it again? Those belong to design and governance, and to a later node. ↩

Permacomputing is a pressure test

2026-01-07T00:00:00+00:00

I met the term permacomputing again today: 8 May. My Obsidian notes say I first encountered it on 18 January.

I had written only this:

Is permacomputing possible? Is it needed? In my website? in my lifestyle? in health response?

That’s it.

This time I read the principles. Another website. On permacomputing for graphic design. The intention was good. The first thing I experienced was physical difficulty. The text was hard to read with my presbyopia.

A sustainable web punishes my rigid eyes, my limited brain capacity.

The term reminded me of circular economy, sustainable web, resilient web, low-tech aesthetics, and the performance around these ideas. The concept is useful, visionary, often utopic. I agree: people and planet need less extraction, less upgrade pressure, less digital bloat. But the human problem remains. People perform ethical choices. I perform too.

I once analyzed the Ellen MacArthur Foundation website, a prominent advocate in the field of circular economy. The website is bloated, at odds with what it promotes.

To avoid turning permacomputing into performance, I tested it on my own website.

At that point, my current CSS was 34 KB.

The urgency to make a website less bloated, sustainable, or lightweight now comes more often when I think of the July 2024 internet blackout in Bangladesh. I was looking for some protocol, something like Bridgefy, to communicate during critical times. But the protocol, and its apps, still felt underdeveloped: dependent on proximity, adoption, battery, prior internet dependency during setup, and yes: trust.

AI delivered: CSS reduced to 18 KB while keeping the present look. Then I pushed the test further: what would happen if I aimed for a 1996-minded CSS layer, close to the first era of CSS support? Possible?

The stylesheet became 9 KB.

No local WOFF2. No custom typeface. No variable font. No optical sizing. Only system stacks. Georgia, with large type, served for a while. HTML pages were semantic. The page remained readable.

Then my brain started bargaining. FOMO. Fear of missing out.

I missed Source Serif 4. I missed optical sizes, glyphs, ligatures, font features, and the pleasure of a chosen typeface.

The typography obsessive in me became vocal while the sustainability argument was reasonable.

AI suggested a middle ground. A certain model was trained to satisfy its “user” - unless further prompted on sycophancy.

Use Source Serif 4 Regular and Italic only. Keep mono as a system stack. Your code blocks are occasional, not the site’s main reading voice. System monospace is good enough for pre, code, and metadata. Keep Bangla font local and optional. Let the page fall back without drama.

I went to make my morning coffee.

I looked at the new coffee machine. Fragile, designed for planned obsolescence. I looked at the food processor in the kitchen. I looked at the other things I “own”. It’s clear: I am not yet a sustainable practitioner. I buy clothes every year. I smoke Marlboro cigarettes. Branded, expensive, corporate. And I am a subscriber to a few AI models. I live inside capitalism and still enjoy parts of it while struggling to accept other parts.

Then I looked back at my CSS. I was still looking at the technical layer. I was thinking of technical debt: stylesheet size, heavy font files, page weight, browser support, device tolerance. I was trying to make the site more accessible by reducing technical debt. But structural debt was harder to see. A lighter website, even with nostalgia-proof or future-proof intent, can reduce one kind of dependency. But it remains silent about who gets disconnected first, who has backup devices, who can pay for data, and who becomes unreachable when infrastructure fails.

I keep facing decision paralysis: this, that, another version of that. I keep exploring. I remain unsettled. Even that unsettledness, the idea and the word, is co-opted.

Never settle.

(the official tagline of the smartphone manufacturer)

Corporations appreciate it. Ads sharpen it. Interfaces keep offering newer, better, cleaner, more ethical versions of consumption. Another desire loop.

I wish my brain could be more settled. Because unsettling also means I am allowing corporations to feed me ads so that I buy things.

So the question is not whether I am “pure” enough to practice permacomputing.

Or the question is: where can I reduce dependency without performing poverty, nostalgia, or moral superiority?

For this website, the choices could be:

HTML as structure.
    CSS as accessible presentation.
    JavaScript as optional.

Fonts as preference.
	Source Serif when available.
	Georgia when not.
	Bangla readability as access.

When I consider typography as interface, relying on system fonts is one compromise. Reducing file size with less CSS is another. It reminded me how cinema viewing shifted from maker agency to viewer agency. Cinema now gets designed for multiple screen formats.

My practical wisdom is simpler: permanence is a fantasy. I am not an advocate for disposable or ephemeral content either. But when long-term durability gets treated as permanence, and a tech community shaped by planetary limits and permaculture thinking calls that desire permacomputing, the idea can become morally superior, distant from daily use, and difficult to practice. It may ignore the spectrum of human behaviour, hide systemic uneven access, and smooth over the social and political mess inside which computing happens.

I am living inside the contradiction. Inside the mess. My agency, or the ideal image of my agency, feels compromised.

Permacomputing is a pressure test.

Working glossary

Permacomputing as pressure test: A way to test whether sustainability remains practice or turns into performance. The test begins with code, then reaches lifestyle, infrastructure, access, and contradiction.
Performative sustainability: Ethical language used as display while the underlying system remains extractive, bloated, or exclusionary.
Technical debt: The visible maintenance burden: heavy CSS, font files, scripts, dependencies, and design decisions that make the page harder to sustain.
Structural debt: The hidden burden carried by people who are already closer to disconnection: low bandwidth, shutdowns, cost, device scarcity, battery anxiety, and fragile communication networks.
Typography as interface: The idea that type is not decoration. Font choice, size, spacing, contrast, and rhythm decide whether the page welcomes or punishes the reader.
Shadow reader: The reader excluded by normal design assumptions: tired, distracted, low-vision, multilingual, older, neurodivergent, suspicious, or using a weak device on an unstable connection.
Cognitive load reduction: The practice of removing avoidable burden from a page, form, document, or system so that attention can stay with meaning.
Protocol fragility: The gap between a communication idea and its survival in crisis. Proximity, adoption, battery, setup, trust, and usability decide whether the protocol becomes usable infrastructure.
Desire loop: The cycle where interfaces offer newer, cleaner, more ethical versions of consumption while keeping the user unsettled and wanting.

Agency compromise

The uneasy state of making choices inside systems already shaped by platforms, markets, ads, infrastructure, and personal dependency.

Unlearning repository

A website that holds fragments, contradictions, tools, drafts, and unfinished thinking without forcing everything to become a polished portfolio.

Interface autoethnography

Studying the interfaces I build and use as evidence. The page, the tool, the font, the form, and the friction become field notes.

The sleep test

2026-01-07T00:00:00+00:00

The test name is polysomnography.

I learned the word at a hospital billing counter in Dhaka. I said I needed a sleep test. The woman paused. It had another name, she said, but she could not remember it.

I searched online and pronounced the word from my phone.

Polysomnography.

She entered it into the billing system. The test appeared.

Before the word, my request was informal. After the word, it became a service the institution could recognize, price and schedule.

Language was the first interface.

Method and limit

This field note uses six materials:

notes written within twenty-four hours of the test;
appointment messages and time stamps;
the fourteen-page diagnosis and titration report;
remembered conversations from the night and the following consultation;
subsequent product research and visits to medical-equipment stores in Dhaka and Bangkok;
twenty-four weeks of Garmin sleep records from 1 February to 18 July 2026.

I later returned to these materials through my unlearning framework. The framework asks what route an interface places first, what reward holds it there, and what artifact carries that order.^[1]

The report contains my own medical information. Other people are described by role because I did not meet them as research participants or obtain consent to publish their identities. Conversations are paraphrased where exact wording is unnecessary.

This is not a clinical evaluation of my treatment. It is an interface audit of how a body, a protocol and a report produced a medically legible sleeper.

The room that became a lab

Before payment, the counter sent me to a respiratory room beside the outpatient consultation area. The room handled availability, preparation, scheduling and report collection. The actual night study took place elsewhere, in a hospital cabin temporarily arranged around the equipment.

The service was new. I was later told that I was its twenty-seventh patient.

Its institutional presence was small:

one room
one recording system
one night technologist
one patient at a time

The hospital was large. The sleep service depended on a narrow human and technical arrangement inside it.

That dependence became visible when my original appointment was cancelled. A relative of the technologist had died. The test had to move to another date and then into an available institutional slot.

The preparation sheet instructed me to wake early and avoid sleeping during the day. I followed it. I was physically and mentally prepared for the original night. When the appointment moved, that preparation became unusable.

On the rescheduled date, I was told to report at 9 pm.

I do not normally sleep at 9 pm. My dinner is also later. I tried to negotiate the arrival time, but the available slot came first. I ate quickly and reached the hospital with my partner and a friend.

By then, the test had already begun shaping the body it expected to measure.

A fresh patient

The cabin resembled a hotel room with medical equipment added to it. I changed into a cotton hospital gown and completed an intake form.

The form was in Bangla.

I asked what would happen if a non-Bangla-speaking patient came for the test. I received no answer.

I had used the same hospital for several years. It already had my patient number, insurance records, consultations and recent test results. During that visit, however, staff repeatedly said that an information-system upgrade had made older records unavailable.

The sleep form therefore began from zero.

The institution knew me as a returning patient and received me as a blank one.

This was not only a data problem. It established the sequence of knowledge. The standard form arrived before the medical history the hospital already held. What fitted the form became immediately available. What remained elsewhere had to be remembered and narrated again.

Producing the measurable sleeper

The technologist attached sensors to my scalp, face, chin, chest, abdomen and legs. Bands measured respiratory effort. Other sensors recorded airflow, oxygen saturation, heart activity, movement, position and snoring.

My partner photographed the wired body. We made jokes about familiar film and internet tropes to reduce the tension.

I was both the subject and a prop inside the room.

The report later described the arrangement in technical language:

Three EEG channels, two EOG channels, one ECG channel, chin EMG, left and right leg EMG, a snore microphone, airflow sensors, respiratory effort belts, pulse oximetry and a body-position sensor.

At 11:51 pm, the recording began.

The report calls this moment “lights out.”

I call it the beginning of trying to sleep.

For what felt like several hours, I moved between positions. Sensors loosened. The technologist returned to adjust them. I felt cold in the hospital gown and requested another blanket. I turned the light on and off. I went to the toilet repeatedly.

The report later stated:

The patient had no episodes of waking to urinate during the night.

My field note records more than five toilet visits.

The difference is small clinically unless those visits affected the recording. Analytically, it matters. My body remembers repeated interruption. The summary remembers none.

What counts as sleep

I told the technologist that I had self-identified ADHD, anxiety and a history of panic attacks. I also explained that I normally use sound to settle into sleep.

She needed sufficient data before the test could move to titration.

I needed conditions under which sleep might occur.

At first, I was told not to use my phone or play sound because audio recording formed part of the study. I asked whether my usual sleep sound could be permitted at low volume. The technologist consulted the supervising doctor and returned with approval.

I played a familiar soundscape. When that did not help, I changed to a downloaded recording of rain with quiet piano.

Eventually, I slept.

This was the most important alteration of the night. The protocol did not disappear. Its sequence changed slightly.

The initial route was:

protocol first
familiar sleep practice removed
body expected to adapt
data expected to follow

The negotiated route became:

protocol
patient describes an unmet condition
technologist seeks authorization
familiar sound enters the room
sleep becomes possible

The adjustment did not invalidate the test. It helped the test continue.

The technologist later told me that I was the first patient who had requested sleep audio. The encounter made that need visible to the service. I learned that a rule presented as fixed could contain room for judgment.

The split night

At 3:45 am, the diagnostic phase ended.

The report recorded 164 minutes of sleep during 234 minutes of diagnostic recording. It identified 79 apnea and hypopnea events, an apnea-hypopnea index of 28.7 per hour, severe snoring and a lowest oxygen saturation of 88 percent.

An AHI between 15 and 30 is conventionally classified as moderate obstructive sleep apnea.^[2]

The diagnosis therefore did not depend on one isolated event. The recording documented repeated obstruction across more than two and a half hours of scored sleep.

The night then changed tasks.

Staff woke me, fitted a nasal mask and started positive airway pressure titration. I initially felt claustrophobic. I was told the sensation should improve once the airflow stabilized.

It did become easier. I slept.

The titration section records nearly three hours of sleep. Across the full titration period, AHI fell from 28.7 to 1 event per hour. Average oxygen saturation increased from 94 to 96 percent. The snore index fell from 469.7 to 40.7.

Metric	Diagnostic phase	PAP titration
Total sleep time	164 minutes	177 minutes
Sleep efficiency	71.8%	97.4%
AHI	28.7/hour	1/hour
Lowest SpO₂	88%	93%
Snore index	469.7	40.7

The central clinical result is difficult to dismiss. The study recorded obstructive sleep apnea, and positive airway pressure greatly reduced the measured respiratory events during that night.

But the report moves from this strong finding to a more specific recommendation without showing every step.

The report as interface

The report did not simply preserve the night. It reorganized it.

The lived sequence contained preparation, rescheduling, unfamiliarity, cold, repeated toilet visits, surveillance, negotiation, claustrophobia and eventual sleep.

The report produced another sequence:

recording
diagnosis
titration
normalization
recommendation

Both sequences are real. Only one enters the formal medical record.

This is sequence authority.

The report arrives first when another clinician, insurer or supplier encounters the night. My field note arrives later, if I remember to introduce it.

The report also contains internal inconsistencies.

Sleep latency

The diagnostic table records sleep latency as zero minutes. I experienced a long struggle to sleep.

This does not necessarily mean the machine was wrong. Sleep onset is scored through physiological signals, not through the subject’s memory of falling asleep. I may have entered brief stage N1 sleep while continuing to experience myself as awake.

The difference reveals two definitions rather than automatically proving an error:

machine sleep: a scored physiological state
lived sleep: the remembered experience of having slept

The report controls the clinical definition.

REM sleep

The diagnostic tables record zero minutes of REM sleep. Another summary sentence says that the lowest oxygen saturation occurred “during an apnea and R sleep.”

Both statements cannot describe the same diagnostic record without further explanation.

The absence of diagnostic REM also matters because obstructive events can vary by sleep stage. A first night in an unfamiliar laboratory can alter sleep continuity and architecture, including REM sleep.^[3]

This does not erase the 79 respiratory events that were recorded. It limits what this single night can say about how my apnea varies across a more ordinary night.

Toilet visits

My notes record repeated toilet visits. The report states that there were none.

Here the machine-readable night does not merely differ from my perception. The written summary excludes observable events that required leaving or interrupting the bed.

Sleep-stage arithmetic

The titration report gives a total sleep time of 177 minutes. It separately lists:

Stage	Reported duration
N1	37 minutes
N2	91 minutes
N3	49 minutes
REM	49 minutes

N1, N2 and REM already total 177 minutes. Adding the reported N3 produces 226 minutes.

At least one value or label is wrong.

The report nevertheless describes deep sleep as significantly improved. That conclusion may be correct, but the printed figures do not allow the reader to verify it.

Pressure and comfort

One page says pressure was increased from 4 to 15 cmH₂O and that I felt comfortable at 11. Another says I felt comfortable at 15.

I do not remember selecting either number. I remember initial claustrophobia, gradual adjustment and then sleep.

The report converts successful continuation into a statement about comfort.

Its pressure table creates another question. It records:

Pressure	Sleep recorded	AHI
5	36.0 minutes	3.3
6	22.2 minutes	2.7
7	13.0 minutes	0
8	14.2 minutes	0
9	10.9 minutes	0
10	8.2 minutes	0
11	13.0 minutes	0
12	9.7 minutes	0
13	20.1 minutes	0
14	15.9 minutes	0
15	11.4 minutes	0

The final recommendation says:

Use Auto CPAP with nasal mask at 15 cmH₂O pressure during sleep time regularly.

The table shows effective control across several pressures. It does not show, at least in the printed pages available to me, which pressures included supine REM sleep, acceptable mask leak, stable oxygenation or sustained control long enough to select 15.

Clinical titration guidance does not choose pressure from AHI alone. The selected pressure should control obstructive breathing, maintain adequate oxygenation, have acceptable leak and, for an optimal titration, include sufficient supine REM sleep at the chosen pressure.^[4]

The report may contain enough underlying signal data for a clinician to justify the choice. The printed recommendation does not make that reasoning visible.

Can one night justify one machine?

A single polysomnography night can provide clinically meaningful evidence. My diagnostic phase crossed the conventional threshold for moderate OSA. The titration phase showed a major reduction in respiratory events with positive airway pressure. Split-night studies are an established clinical route when sufficient moderate or severe OSA appears during the diagnostic portion and enough time remains for titration.^[5]

The first-night effect does not make the diagnosis imaginary.

It does, however, place limits around more specific claims.

My report supports these conclusions:

obstructive respiratory events repeatedly occurred during recorded sleep;
the measured AHI fell within the moderate OSA range;
positive airway pressure substantially reduced those events during titration;
oxygen saturation and snoring improved during the PAP phase.

The printed report does not transparently establish:

why 15 cmH₂O was selected over the other effective recorded pressures;
whether 15 is intended as a fixed pressure, an upper limit or part of an unstated automatic range;
which pressure included adequate supine REM sleep;
how mask fit, leak and long-term comfort affected the choice;
why one commercial brand or supplier would be clinically necessary;
how cost and my ability to continue treatment were weighed.

Recommending PAP after this study is not inherently excessive or unethical. Recommending a specific device may also be reasonable when reliability, data access, servicing or pressure performance differ materially.

The ethical problem begins when specificity exceeds the explanation available to the patient.

A durable recommendation should separate four decisions:

1. Does the evidence support an OSA diagnosis?
2. Does the evidence support PAP treatment?
3. What pressure range and mask are appropriate?
4. Which available device can deliver that treatment safely and sustainably?

My report answers the first two strongly.

It compresses the last two into one line.

Clinical guidance treats PAP as continuing care, not as a one-time hardware purchase. It emphasizes education, mask fitting, troubleshooting, monitoring of objective use and efficacy data, and follow-up after treatment begins.^[6]

A device that is technically excellent but financially unreachable produces no treatment. A cheaper device without trustworthy performance, service or monitoring may produce another risk. Cost is therefore not external to clinical effectiveness. It shapes whether the prescribed route can be entered and maintained.

From pressure to product

The report recommends a category of treatment: automatic positive airway pressure with a nasal mask.

The consultation moved from treatment category to product market.

The consultant explained that good machines were expensive. He warned against low-cost devices and referred me to a local supplier. He did not write a specific brand in the report, but described an international brand as preferable and told a story about a patient allegedly harmed by a cheaper device.

That story changed the decision environment.

The choice was no longer presented as:

different clinically suitable devices
different prices
different service arrangements
different forms of follow-up

It became:

expensive recognized device
or
unsafe uncertain device

I cannot verify the reported death or connect it to a particular product failure. I also cannot establish whether the referral involved any financial relationship. Treating either suspicion as fact would reproduce the same problem I am examining.

The observable issue is narrower.

A story about danger arrived before transparent comparison.

Country of manufacture became a shortcut for reliability. Brand recognition became a shortcut for safety. The supplier card made one purchasing route immediately available while alternatives remained abstract.

The treatment that did not begin

I did not reject positive airway pressure.

I could not convert the prescription into a confident purchase.

I searched the Dhaka market. During a later visit to Bangkok, I visited medical-equipment stores there as well. The same brand repeatedly arrived first. Sellers presented ResMed as the recognized international option and pushed me toward it.

The available machines did not form a simple price list. I encountered new machines, used machines, different models, uncertain warranties and cheaper alternatives whose reliability I could not independently judge.

The report had reduced the clinical decision to one line:

Use Auto CPAP with nasal mask at 15 cmH₂O pressure during sleep time regularly.

The market reopened everything that line had closed.

Which model?
New or used?
What pressure range?
What mask?
Whose warranty?
Who reads the device data?
Who adjusts treatment?
What happens if the machine becomes uncomfortable?

The supplier could sell a machine. The report could prescribe one. Neither interface gave me a clear method for comparing the available products.

The clinical sequence therefore stopped at the point of transaction:

diagnosis
titration
prescription
price
uncertainty
no device

The prescription produced medical certainty and practical hesitation.

The second machine

In early February 2026, I began wearing a Garmin Forerunner 165.

It was not a substitute for CPAP or polysomnography. The watch could not tell me how many times my airway closed. It could not reproduce the EEG, airflow, respiratory-effort or clinical oxygen measurements used in the hospital study. Garmin also states that the device is not intended to diagnose or monitor a medical condition.^[7]

It answered a different question.

The hospital had measured what happened to my breathing during one night. The watch began estimating when and how long I slept across ordinary life.

From 1 February to 18 July 2026, it produced twenty-four weeks of sleep records.

Garmin measure	Twenty-four-week summary
Average sleep score	54
Average recorded sleep duration	4 hours 32 minutes
Device-estimated sleep need	8 hours 3 minutes
Recorded gap	3 hours 31 minutes
Average bedtime	4:48 am
Average wake time	9:38 am

The averages flatten substantial variation. Weekly recorded sleep ranged from 3 hours 14 minutes to 5 hours 34 minutes. Average bedtimes moved across late night, early morning and, in some weeks, daytime.

The watch made one pattern difficult to ignore. Whatever happened to my airway during sleep, I was also recording very little sleep and entering it at unstable times.

But the watch introduced another clean result.

On recorded mornings, a proprietary system converted movement, heart rate, heart-rate variability and other signals into a number. The number often said Poor.

The hospital report had turned my breathing into AHI.

The watch turned my ordinary nights into a sleep score.

Neither number was meaningless. Neither contained the entire problem.

The hospital metric supported a respiratory treatment that I could not yet obtain. The consumer metric repeatedly described inadequate sleep without treating the obstruction already documented by the hospital.

The second machine continued measurement where the first machine’s recommendation had stopped.

The body improvises

From February through July 2026, after the sleep test and while I remained without a PAP device, I made smaller physical adjustments at home.

I began using two pillows. I also reminded myself to sleep on my side.

The two-pillow practice had not been tested during the hospital study. The report’s positional analysis, however, gave me some basis for trying side sleeping. During the diagnostic phase, it recorded an RDI of 46.3 events per hour while supine and 24 while non-supine.

Recorded position	RDI
Supine	46.3/hour
Non-supine	24/hour
Left	22.9/hour
Right	27.2/hour

The difference is meaningful, but its limit is equally important. Non-supine sleep was associated with a lower recorded RDI during that single study. It did not bring the result into the normal or mild range.

The Garmin could not establish whether either later practice reduced apnea.

These were interim counter-practices, not equivalent treatment. They changed what I did at home while the prescribed route remained financially and practically unresolved.

Prototype audit

Reflex. Comply with the test conditions, produce sufficient data and then follow the preferred purchasing route.

Reward. Diagnostic certainty, completion of diagnosis and titration in one night, efficient use of a scarce service, and rapid movement from report to treatment product.

Carrier. The billing term, preparation sheet, appointment slot, intake form, sensors, scoring rules, report tables, prescription, supplier card, product market and Garmin sleep score.

Sequence change. Introduce familiar audio during the test. Compare products rather than immediately purchasing the preferred brand. Track ordinary sleep timing and experiment with body position while treatment remains unresolved.

Return. The diagnosis retains clinical authority. The prescribed device remains inaccessible. The Garmin score risks becoming another result that arrives before felt experience.

Capture risk. A healthcare-access problem becomes a personal optimization project. Consumer measurement offers continuous legibility without treating the respiratory obstruction.

What changed

The technologist did not simply enforce the protocol. She interpreted it.

She adjusted sensors, found a warmer blanket, listened to my request, consulted the doctor and allowed an unfamiliar need into a new service. Her judgment made the protocol more workable.

That matters because the case is not a simple story of system against patient.

The interface contained several forms of intelligence:

physiological measurement;
clinical scoring;
the technologist’s situated judgment;
the patient’s knowledge of his own sleep practice;
the consultant’s treatment experience;
the report’s capacity to make the night portable.

The problem emerged when one form of intelligence was allowed to summarize the others without showing the cut.

The machine did not lie. The report did not tell the whole night. My memory does not replace the physiological record. The physiological record does not replace my account of how that record became possible.

Does this count as unlearning?

Partly, at the personal scale.

I did not unlearn obstructive sleep apnea. That is not what the framework proposes. What weakened was the authority of the prescription as a complete route from diagnosis to treatment.

The case now contains repetition. I researched machines in two countries, repeatedly encountered the same preferred brand, wore a consumer tracker for twenty-four weeks, used two pillows and reminded myself to return to side sleeping.

These practices did not establish an equivalent treatment or demonstrate reduced respiratory events. They changed the sequence through which I interpreted and responded to the recommendation.

Before
report
prescription
preferred product
purchase expected

After
report questioned
market compared
cost exposed
daily sleep observed
body position altered
treatment unresolved

A second capture also appeared.

When clinical treatment remained inaccessible, consumer measurement became available. The Garmin score offered daily legibility without respiratory treatment. It risked converting a healthcare gap into a personal optimization project.

For the watch to have sequence authority, however, the number must do more than appear. I would need to document how a low score changed my mood, schedule, behaviour or judgment about whether I had slept well.

The data establishes the carrier.

A future field note must establish its authority.

At the institutional scale, unlearning remains unproven. I do not know whether the service changed its preparation instructions, accommodation practices or reporting process after my encounter. A future patient may still have to negotiate the same exception from zero.

Individual accommodation weakened a default for one night.

Institutional memory determines whether the default learned anything.

After the test

In the morning, the technologist removed the sensors. Gel remained across my scalp and body. I thanked her and apologized for not giving her enough sleep.

The report later showed that I had given the system 164 diagnostic minutes and 177 titration minutes.

Enough to diagnose.

Enough to prescribe.

Enough to turn one difficult night into a portable medical fact.

Before leaving, I photographed the room and the equipment beside the bed. I wanted a memory of the place before the report became the night.

The report says the machine worked.

It did.

The remaining question is what else had to work around it, and what disappeared when the result became clean.

Endnotes

The framework defines sequence authority as the advantage gained when a designed system places one route first and makes alternatives slower, less visible or more costly to recover. See The Unlearning Framework 7.3. ↩
The apnea-hypopnea index counts apneas and hypopneas per hour of recorded sleep. An adult AHI from 15 to below 30 is generally classified as moderate obstructive sleep apnea. See Harvard Medical School, Understanding the Results. ↩
The first-night effect describes changes in sleep produced by an unfamiliar recording environment. Common findings include reduced sleep time or efficiency, more awakenings and altered REM sleep. It does not affect every person or invalidate every first-night study. See Wick et al. (2024), The first-night effect of sleep occurs over nonconsecutive nights in unfamiliar environments, and Mayeli (2024), Addressing the first-night effect. ↩
AASM titration guidance grades a titration through control of respiratory events, oxygen saturation, leak and the presence of supine REM sleep at the selected pressure. See Kushida et al. (2008), Clinical Guidelines for the Manual Titration of Positive Airway Pressure. ↩
AASM diagnostic guidance permits a split-night protocol when moderate to severe OSA is observed during at least two hours of diagnostic recording and at least three hours remain for PAP titration. Clinical judgment remains necessary because not every split-night study produces a conclusive titration. See Kapur et al. (2017), Clinical Practice Guideline for Diagnostic Testing for Adult Obstructive Sleep Apnea. ↩
AASM treatment guidance recommends PAP based on objective testing and calls for adequate follow-up, troubleshooting, and monitoring of efficacy and use after initiation. It supports either CPAP or APAP for ongoing treatment in appropriate adults rather than treating one device purchase as the end of care. See Patil et al. (2019), Treatment of Adult Obstructive Sleep Apnea with Positive Airway Pressure. ↩
Garmin describes sleep tracking and breathing-variation features as general wellness tools. The company states that the Forerunner is not a medical device and is not intended to diagnose or monitor any medical condition. See Garmin, Breathing Variations. ↩

শেষ আড্ডা

2020-05-15T00:00:00+00:00

সিনেমা থেইকা পালায়া ১০৪.২

2020-03-27T00:00:00+00:00

বাড়ি কোথায়

2015-09-25T00:00:00+00:00

বাড়ি কোথায়, তোমার? নাই। আবার সবখানে। আসলে নিজের বাড়ি একটা বাড়াবাড়ি ধারণা। রেবেকার নিরুত্তাপ কণ্ঠের জবাবে থমকে ছিলাম। তাই তো!^[1]

যে দেশে জন্ম। যে শহরে থাকি। যে মহল্লায় শৈশব আমার। গ্রামের বাড়ি। বাবা-মায়ের বাড়ি। নিরাপত্তাঘেরা ভাড়াবাড়ি। যেখানে যাপন করি। যেখানে জীবিকা করি। যেখানে কাজ করি। যে স্থায়ী-অস্থায়ী ঠিকানা ছাপা আছে পাসপোর্টে, পরিচয়পত্রে। কোনটা, কোথায়, নিজের বাড়ি?

রেবেকার জন্ম ফ্রান্সে। থাকে ইতালিতে। দেখা জার্মানিতে। রেবেকা মৌসুমী ভৌমিকের গানটা শোনেনি।^[2]

“বাড়ি কোথায়?”— বহু বছর পর শুনছি আবার। শুনতে শুনতে স্মৃতিটা মনে পড়লো। অনুভূতিটা।

সম্ভবত অন্য বাড়িতে যেতে হবে। অথবা, অন্য বাড়িতে ছিলাম। অথবা, নিজের বাড়ি, আসলেই, একটা বাড়াবাড়ি ধারণা।^[3]

In 2010, during the Venice Film Festival, I asked Rebecca where she was from. We first met in Italy and later crossed paths in Germany during the Berlin Film Festival. She was born in France and worked across the international film-festival circuit. My memory turned a question about origin into one about home. That turn belongs to this text. I use only her first name. ↩
In 2010, at the Venice Film Festival, I asked Rebecca where she was from. We had first met in Italy, then again in Germany during the Berlin Film Festival. She was born in France and was living across Europe, working the international film-festival circuit. My memory later turned a question about origin into one about home. That turn belongs to this text. I use only her first name. ↩
The title and recurring question come from Moushumi Bhowmik’s 2002 song “বাড়ি কোথায়?”. I had written about Bhowmik and Parapar on the Zicobaji blog on 7 June 2007, although that post considered the music more broadly. The song remained in memory and returned through this encounter. ↩
The title and the recurring question come from Moushumi Bhowmik’s 2002 song “বাড়ি কোথায়?”. I had written about Bhowmik and Parapar on the Zicobaji blog on 7 June 2007, though that post considered the music more broadly. The song stayed in memory and returned through this encounter. ↩
This is a 2026 rereading. The 2015 note did not foresee what followed. At the time, I was developing Cinema, City & Cats and still working at a local film school. The institutional rupture came in 2016. Afterward, I kept moving within Dhaka and between cities while my jobs, roles, and affiliations changed. Home now feels less like property or a fixed coordinate and more like a provisional sense of belonging, safety, or stillness. I am not using this fragmented memory to generalize migration or its histories. Cinema, a film school, and a cine club became points along the route, but none became home. The search remained after I stopped trying to solve it. ↩
This is a 2026 re-reading. The 2015 note did not foresee what followed. At the time I was developing Cinema, City & Cats and still working at a local film school; the institutional rupture came in 2016. Afterward I kept moving within Dhaka and between cities while my jobs, roles, and affiliations changed. Home now feels less like property or a fixed coordinate and more like a provisional sense of belonging, safety, or stillness. I am not using this fragmented memory to generalize migration or its histories. I am recording a narrower friction: a cinema, a film school, a cine club was a point on the route, not a home. The search remained after I stopped overthinking it. ↩

Ishtiaque Zico

Colophon

Preface

Context

The career under one name

A note on privacy

The second workstation

When support becomes production

The name as interface

Existing terms

Unequal accumulation

The evidence trap

Why the arrangement persists

When naming becomes another capture

What recognition could mean

Prototype audit

A career with no reference

Endnotes

References

The unlearning framework

Summary

Where this began

Two meanings of unlearning

Sequence authority

Three encounters

The career under one name

The AI-assisted sentence

The sleep test

Local and institutional unlearning

Try it in three steps

Position and limits

Closing

Glossary

Further reading

References

Endnotes

Permacomputing is a pressure test

Further reading

Working glossary

The sleep test

Method and limit

The room that became a lab

A fresh patient

Producing the measurable sleeper

What counts as sleep

The split night

The report as interface

Sleep latency

REM sleep

Toilet visits

Sleep-stage arithmetic

Pressure and comfort

Can one night justify one machine?

From pressure to product

The treatment that did not begin

The second machine

The body improvises

Prototype audit

What changed

Does this count as unlearning?

After the test

Endnotes

শেষ আড্ডা

সিনেমা থেইকা পালায়া ১০৪.২

বাড়ি কোথায়