Abstract
“Please return a marked-up, accessible proof by Friday.” The request sounds procedural. It is also a material demand. It presumes a usable input path, an adequate display, compatible software or a workable review route, enough storage, a power source, connectivity, time in a safe place, a way to save the result, and a device that will still function when the task must be done. For some contributors it additionally presumes screen-reader compatibility, alternative input, magnification, or other assistive technology. The issue is not merely whether someone is online.
This paper introduces the capability floor: the minimum, ongoing constellation of material conditions required to complete a specified publishing task reliably and accessibly. The construct shifts the conversation from device ownership and sticker price to task fit, continuity, repair, time, shared infrastructure, and accessibility paths. It does not name a universal minimum specification or recommend a product. It asks a more exact question: what must remain available for a person to participate in a particular editorial or publishing workflow without accepting disproportionate risk?
The paper develops the Capability Floor Ledger, a task-based audit that records requirements, available routes, continuity exposures, workarounds, responsibility, and residual exclusion. Current U.S. evidence shows that connection, computer access, smartphone reliance, and multi-device access remain unequally distributed; global digital-inclusion guidance similarly treats devices, affordability, skills, quality, availability, and security as interdependent.[1] [2] [3] [8] Public libraries, lending programs, and digital-navigation services can lower important barriers, but they do not make continuity, privacy, opening hours, accommodation, or deadline risk disappear.[6]
1. The editorial request that assumes a room of equipment
A small press sends a contributor a proof. The contributor is asked to read the text, identify any structural problem, respond to tracked changes, add a comment, save the revised file, and return it by a fixed deadline. The press also asks that the proof be checked for an accessible reading path. No member of the editorial team intends the request to exclude anyone. Yet the request contains an unspoken room of equipment.
It assumes that the contributor has, or can reliably reach, a suitable device. It assumes a usable display and input route. It assumes an application that can open the file, expose the relevant markup, permit a response, and save or export a returnable version. It assumes a battery, electricity, storage, network access at the necessary moment, and enough uninterrupted time to work. It may assume a quiet or private place. It may assume that assistive hardware or software works with the document and the chosen application. It assumes that if the device breaks, an alternative exists before Friday.
The scenario is constructed, not a claim about a particular software product or an observed contributor. Its purpose is to make visible a problem that device counts hide. A person may own a smartphone, have used the internet recently, and still lack a reliable path through this particular task. A person may have access to a community computer yet be unable to use it outside business hours, hold confidential editorial material there safely, install a needed assistive tool, or retain a private working copy. Another person may own a laptop but face a failed battery, an inaccessible interface, missing storage, unsupported software, or a repair barrier at the moment of need.
Digital participation is not a possession. It is a continuing capability to complete a defined action under actual conditions.
That distinction matters because publishing increasingly coordinates work through digital documents, portals, files, devices, and services. When an organization describes participation as “online,” it may confuse several conditions: a connection exists, a device exists, a suitable device exists, the device is available to this person, the task can be completed on it, the result can be saved and returned, and the arrangement will survive until the deadline. The capability floor provides a way to separate these conditions.
2. Access is not a yes-or-no condition
U.S. data illustrates why a simple access category is inadequate. The Census Bureau reported that 95% of U.S. households had at least one type of computer and 90% had a broadband subscription in 2021; smartphones were the most common reported device, while desktop/laptop ownership was lower.[3] At the aggregate level, these figures can suggest that a device problem has largely been solved. They cannot answer whether an individual has private use of an appropriate device, whether it supports a particular task, or whether it will remain usable.
The distributions inside the aggregate matter. Pew Research Center’s 2021 U.S. adult survey found that 41% of adults in households below $30,000 did not have a desktop or laptop and 43% did not have home broadband; 27% were smartphone-only internet users.[1] NTIA’s 2023 Internet Use Survey summary similarly distinguishes any connection from both fixed-and-mobile household connection and desktop/laptop/tablet use. It reports that lower-income people remained less likely to have both fixed and mobile connections and that device use and smartphone-only reliance varied across groups.[2]
These figures are not a publishing-access study. They are U.S.-specific snapshots, and they do not determine what any one person can do. Their role is narrower: they show why “internet user” and “task-capable producer” should not be treated as the same category.
The International Telecommunication Union places that distinction in a broader context. Its Global Connectivity Report 2025 frames universal and meaningful connectivity through six interdependent dimensions: quality, availability, affordability, devices, skills, and security.[8] That framing does not let the paper generalize U.S. percentages worldwide. It establishes a scope principle: meaningful participation is multi-dimensional, and a device is one component within a wider infrastructure of use.
| Access statement | What it establishes | What it leaves unanswered |
|---|---|---|
| “This person has internet.” | Some connection route exists. | Device suitability, privacy, bandwidth quality, timing, file compatibility, power, and task completion. |
| “This household has a computer.” | At least one computer is present at the household level. | Individual availability, condition, accessibility, software, storage, repair path, and deadline access. |
| “This person owns a smartphone.” | A portable connected device may be available. | Whether the editorial task is usable, accurate, accessible, private, or sustainable on that device. |
| “There is a library nearby.” | Some shared infrastructure may exist. | Opening hours, travel, reservations, wait lists, confidentiality, assistive setup, file retention, and ongoing availability. |
| “The tool is accessible.” | A relevant accessibility claim or standard may apply. | Whether it supports this person’s assistive technology, task, configuration, training, and workflow. |
The point is not to deny the value of any of these conditions. A smartphone can make communication, capture, reading, and some forms of composition possible where no computer is available. A library can provide public computers, lending, expertise, and a place to work. An accessibility standard can establish meaningful obligations. But each provides a part of a pathway, not a guarantee of full participation in every publishing activity.
3. The capability floor
The capability floor is a paper-defined teaching construct. It names the minimum ongoing conditions needed for a defined task to be completed reliably and accessibly. The floor is not a price threshold and not a universal technical specification. It changes with the task, the person, the workflow, the time constraint, the available shared infrastructure, and the consequences of failure.
For the constructed proof-review task, the floor might contain the following layers.
| Layer | Teaching question | Possible requirement | Why it is not reducible to purchase price |
|---|---|---|---|
| Input and output | Can the person read, navigate, comment, and verify changes? | Keyboard, touch, mouse, switch, screen reader, display, magnification, or other appropriate route. | The required route is person- and task-specific. |
| File and software fit | Can the document open, reveal relevant content, and return in a usable form? | Compatible editor/viewer, markup path, export/save option. | A cheaper device can be adequate for one file and inadequate for another. |
| Storage and continuity | Can work be saved, recovered, and preserved through the task? | Enough local/available storage, a dependable save path, version awareness. | Capacity depends on file, workflow, and recovery route. |
| Power and connectivity | Can the work be done and sent at the needed time? | Battery/charging, reliable connection or suitable offline route. | A device can be technically capable but unavailable during outage or travel. |
| Maintenance | What happens if the equipment fails or support ends? | Repair, parts, documentation, software support, replacement or loan route. | Initial ownership says little about useful life. |
| Time, privacy, and place | Can the person complete the task without disproportionate exposure or delay? | Private session, sufficient duration, quiet/supportive environment. | Shared access has social and temporal conditions. |
| Accessibility path | Does the system support the individual’s functional route? | Compatible hardware/software, accommodation, configuration, support. | A generic “computer” category hides alternate inputs and outputs. |
The floor is deliberately relational. A fully maintained personal workstation may meet several layers for one user. A phone plus a library workstation plus a borrowed keyboard plus a hotspot may meet them for another. A high-cost laptop may still fail if an essential assistive path is absent or the account cannot open the required document. The construct directs attention to capability under conditions, not prestige or ownership.
4. Three environments, one editorial task
The same proof-review task can be imagined across three constructed environments. These are not performance tests or advice about particular devices. They are teaching scenes that make different constraints visible.
Environment A: Smartphone-only access
The contributor has a smartphone and mobile data. The phone can receive the proof, read the visible text, and send an email. It may support quick comments, reading, or urgent communication. But a lengthy structural review can introduce effort or precision burdens; the markup may not appear in the same way; the contributor may lack a preferred input method, suitable display condition, robust file storage, or an independent return path. The task may still be completed. The ledger’s role is to ask what burden, time, accuracy, privacy, and continuity cost the workaround creates rather than calling the arrangement automatically adequate or automatically impossible.
Pew’s and NTIA’s evidence matters here only as context. Smartphone-only use is a reported condition for part of the U.S. population, particularly among some lower-income groups—not proof that a phone cannot support meaningful writing.[1] [2] The paper’s task-specific question is more useful: which exact operations become uncertain or costly in this workflow, and who must absorb that cost?
Environment B: Shared community access
The contributor has access to a public library. The Public Library Association’s 2023 weighted survey shows that public libraries provide a substantial layer of shared technology infrastructure: public computers, laptop and hotspot lending, digital-literacy training, digital-navigator programs, and—in some places—digital media production equipment.[6] The report states that 46.9% of public libraries circulated hotspots and 24.8% circulated laptops for off-site use, while 95.3% reported at least one kind of digital-literacy training.[6]
That is a powerful counterpoint to individual-device fatalism. Shared infrastructure can expand capability without requiring every contributor to purchase and maintain an identical private system. It can also provide the human support that turns a device into a usable tool.
But shared provision has conditions. The PLA report notes costs, wait lists, lost or damaged devices, locality differences, and uneven availability of specialized resources.[6] A library session may not provide a quiet space for confidential review, enough unbroken time, a custom accessibility setup, or access at the exact moment a deadline arrives. The lesson is not that community access is secondary. It is that collective provision changes the floor’s composition; it does not erase the floor.
Environment C: Maintained personal workstation
The contributor has a computer suited to the task, a compatible application path, appropriate input/output, power and storage, and a workable save/return routine. This arrangement may provide privacy, immediate access, continuity, and a more stable opportunity for sustained work. It can also fail. A battery may degrade. An operating system or application may stop receiving support. A component may fail. A device may be too expensive or difficult to repair. The ownership condition therefore needs a maintenance question.
PIRG’s 2026 repairability report is not a neutral product benchmark and should not be used to recommend brands. It is useful for a narrower reason: it identifies the practical dimensions through which a device’s usable life can depend on disassembly, parts, tools, repair documentation, software-support duration, and a support ecosystem.[4] The capability floor therefore includes continuity cost—the risk that a usable task path disappears before the work is completed or repeated.
5. The Capability Floor Ledger
The following instrument can be used before an editorial process, in an educator’s assignment design, during a community-access planning discussion, or in product requirements review. It does not assign a universal pass/fail result. It asks the responsible organization to name the conditions it has assumed.
| Ledger field | Question | Illustrative entry for the proof-review task | Responsible party to identify |
|---|---|---|---|
| Defined task | What action must be completed, for whom, and by when? | Review a tracked proof, leave structural comments, save and return a marked version by Friday. | Publisher/editor and contributor. |
| Functional requirement | What input, output, file, application, storage, power, and connection conditions are required? | Readable markup, usable commenting path, version save, return route, suitable input/output. | Workflow designer/product steward. |
| Available route | Which personal, shared, loaned, or assistive routes can meet the condition? | Personal device; library workstation; loaned laptop; alternate input with compatible software. | Contributor with support organization. |
| Continuity exposure | What could interrupt the route? | Device damage, dead battery, closed library, wait list, software expiry, lost file, unavailable accommodation. | Organization that sets timeline and service conditions. |
| Workaround and burden | What alternative exists and what does it cost in time, privacy, travel, money, or error risk? | Travel to library, borrow a device, convert file, request extra time, use a different review path. | Shared: organization should not silently shift all burden to contributor. |
| Residual exclusion | What remains inaccessible, unreliable, unsafe, or disproportionately costly? | No screen-reader-compatible markup; no private room; insufficient loan duration; unstable connection. | Organization must name and address or accommodate it. |
The ledger changes the question a publisher asks. Instead of asking “Can you open the attachment?” it can ask: “What review path would allow you to complete this task appropriately, and what will we provide if the ordinary route is not suitable?” Instead of treating deadline extension as a charitable exception, it can identify whether the original workflow was designed around an unexamined material floor.
The instrument also keeps the scope honest. It does not show that every barrier can be solved by technology procurement. It can reveal that an organization needs a different file format, a longer response window, a phone call, a lower-bandwidth route, access to a loan device, an accessible review method, training, compensation for a necessary cost, or a process that does not treat urgent digital completion as the only legitimate participation mode.
6. Repair, replacement, and continuity
A device’s initial purchase cost is only part of its cost of capability. The relevant working life includes what happens when a battery declines, a charger fails, storage fills, a component breaks, software support ends, an account requires a new version, or a device is lost. A person who can complete today’s proof review but cannot repair or replace the device next month has a fragile capability floor.
The repairability dimensions outlined in PIRG’s 2026 report make this fragility legible: ease of disassembly, availability of parts, use of standard or proprietary tools, availability of repair documentation, and length of software support affect whether a device can continue to function.[4] These dimensions should be treated as one advocacy-oriented account of repairability, not a universal scoring system. Their teaching value is that they move the conversation from ownership to continuity.
This matters for authorship because publishing work has rhythms. A contributor might need only occasional device access to read a message. But a revision process, design pass, grant application, book upload, accessibility audit, or proof deadline can require sustained availability at a specific moment. When a platform or organization assumes uninterrupted device continuity, it can turn a repair policy, service life, or household replacement decision into a hidden editorial condition.
7. Accessibility is a material path
Accessibility cannot be represented by a general category called “device.” Section 508 guidance requires U.S. federal agencies to provide employees and members of the public with access to electronic and information technology comparable to that available to others, and it connects accessibility to procurement and ICT standards.[5] The policy’s scope is federal; it does not dictate a writer’s device or certify a publishing workflow. Its relevance is conceptual: access must be considered in relation to actual functional participation.
The Assistive Technology Industry Association describes assistive technology as equipment, software, and systems that improve or maintain functional capability, with examples including specialized keyboards, switches, pointing devices, screen readers, and eye-gaze or head trackers.[7] This source is a professional resource, not an individualized assessment. The paper does not infer that a particular person needs any particular technology. It uses the variety of examples to show why an assumption such as “a laptop is available” may fail to describe the path through which a contributor reads, writes, points, edits, or communicates.
A capability floor is therefore not inclusive when it specifies one preferred setup. It is inclusive when it asks whether a workflow can be completed through an appropriate accessible path, and whether the organization has designed enough flexibility, time, documentation, and human support to make that path real.
8. Shared infrastructure is not a footnote
The public-library counterpoint changes the paper’s politics. If access were only an individual purchase decision, the solution would appear to be acquiring more devices. The PLA survey shows a more complex landscape: libraries provide public computers, Wi-Fi, hotspots, loans, digital-literacy training, digital navigators, and, in some communities, media-production equipment.[6] These resources make authorship and participation less dependent on private capital alone.
They also make visible a collective responsibility. A library’s public workstation can be a form of publishing infrastructure. A loaned laptop can turn an excluded deadline into a possible task. A digital navigator can help identify a barrier that a device donation alone cannot solve. Training can make a tool usable rather than merely present.
Yet collective infrastructure should not be romanticized. Services depend on budgets, staff, schedules, eligibility, queues, geographic proximity, device condition, local policy, and the capacity to support confidential or specialized work. A public computer may be crucial for a job application and still be unsuitable for a long private edit involving sensitive correspondence. A hotspot loan may solve one connection problem while leaving display, input, software, and repair gaps untouched.
The Capability Floor Ledger therefore asks organizations to name a plural set of routes. Personal ownership, shared infrastructure, accommodation, lending, support, and workflow flexibility can all matter. The appropriate arrangement changes with the task. The ethical failure is not that every participant has a different path; it is that the path is assumed rather than designed.
9. Design implications and boundaries
The material floor of authorship produces four practical questions.
First, publishers and editors can examine every required digital action before imposing it. If a proof requires markup, which review paths work? If a deadline is fixed, what happens when a device, power source, or accessibility route fails? If a file is confidential, which community or loaned routes are realistic? The ledger makes these assumptions discussable.
Second, educators and funders can distinguish a one-time equipment intervention from ongoing capability. A device program may need repair, replacement, training, accessible technology, data, support, and an appropriate place to work. This does not devalue hardware provision. It makes the provision more truthful.
Third, product stewards can avoid using “mobile-friendly” or “works anywhere” as substitute claims for task fit. An interface should make its alternate input/output paths, offline conditions, file behavior, compatibility requirements, and recovery options intelligible. This paper does not test a named application. It identifies questions a responsible design review should answer.
Fourth, writers and community organizations can use the ledger to identify where a material arrangement is fragile without accepting a shopping list as the only answer. A shared computer, a different file path, an earlier hand-off, a loan program, a repair referral, an accessible document alternative, or a negotiated deadline can all be meaningful interventions.
The hardware floor is not a line dividing “real” writers from others. It is the material threshold a workflow quietly imposes—and therefore a threshold that can be inspected, altered, shared, or accommodated.
This paper does not recommend a device, set an individual budget, make consumer-purchase claims, establish a universal hardware minimum, diagnose a person’s accessibility needs, or claim that hardware is the only source of publishing inequality. Its conclusion is narrower. A serious account of access must ask what a person needs to complete a defined creative task, how that capability continues across time, what shared routes exist, and who bears the cost when the assumed route fails.
Limitations
This is a conceptual teaching manuscript, not a device benchmark, price guide, market study, accessibility evaluation, or global prevalence report. The proof-review scenario is constructed to expose dependencies; it makes no claim about the performance of a particular device, application, or contributor. The numerical evidence is chiefly U.S.-specific and time-bounded. It illustrates differentiated access, not universal conditions or individual capability. The ITU source is used only to establish a multi-dimensional global scope frame, because the full underlying report was not inspected. The PIRG repairability report is an advocacy/consumer methodology source and is used for repairability dimensions rather than unqualified brand conclusions. Section 508 concerns U.S. federal ICT obligations, and the assistive-technology source is informational rather than an individualized assessment. Finally, community infrastructure is diverse; public-library figures show service patterns, not a promise that any person can obtain a particular resource when needed.
References
[1] Pew Research Center, “Digital divide persists even as Americans with lower incomes make gains in tech adoption,” 2021.
[2] U.S. National Telecommunications and Information Administration, “New NTIA Data Show 13 Million More Internet Users in the U.S. in 2023 than 2021,” 2024.
[3] U.S. Census Bureau, “Computer and Internet Use in the United States: 2021,” 2024.
[4] U.S. PIRG Education Fund, “Failing the Fix 2026.”
[5] Section508.gov, “IT Accessibility Laws and Policies.”
[6] Public Library Association, 2023 Public Library Technology Survey: Summary Report, 2024.
[7] Assistive Technology Industry Association, “What is AT?.”
[8] International Telecommunication Union, “Measuring Digital Development: Global Connectivity Report 2025.”