Abstract
A writer has been careful with an early draft. The working file contains provisional language, notes about living people, and revision traces that would be misleading if read as settled prose. When an opportunity for rapid feedback appears, the writer sends that same unredacted version to a large informal group. The text has not changed, but the situation has. More people may now receive it, copies may travel by new routes, and it may be harder to correct a misunderstanding or retrieve an unintended version.
The familiar instruction to “keep the manuscript private” cannot explain this change. Manuscripts do not have one stable privacy status. They move through private drafting, collaborative editing, submission and review, public release, and longer-term archival transfer. At each point, a different part of the record can matter; a different audience may have a legitimate role; a different route can carry the material; and the consequences of a mistake may have a different shape.
This paper develops stage-sensitive manuscript exposure: the relationship among a particular manuscript asset, its creative-work stage, the people who may encounter it, an access channel, a possible consequence, and the recovery or access condition that must remain available. It translates the data-focused starting point of threat modeling and the proportional questions in public security guidance into a creative-production field guide.[1] [2] The resulting teaching instrument—the Manuscript Stage Threat Register—does not score a person’s safety or prescribe a technology. It makes a manuscript boundary crossing clear enough for a writer, editor, publisher, educator, or tool steward to choose a proportionate response or state a reasoned exception.
1. The wrong precaution in the next room
The writer who widened the audience for the early draft did not make a foolish choice simply because the work reached readers. Drafts require readers; editing is not a solitary condition forever. The problem is that a precaution suited to one setting was treated as though it answered every later setting. The early working draft raised one question: what should remain inside a small writing boundary while the work is incomplete? The broader review copy raised another: who is receiving this particular version, by what route, with what context, and what can happen if the group or channel changes?
A manuscript repeatedly enters new rooms. A note becomes a co-editor’s working copy. A marked-up file enters a submission workflow. A selected version becomes public. A cluster of drafts and correspondence becomes material for a future repository. The word privacy may appear in all four conversations, but its practical meaning changes with the audience, purpose, copies, and paths of access.
The Electronic Frontier Foundation offers a discipline well suited to this problem. Its security-planning guide begins by observing that protecting all information from every conceivable danger, all the time, is neither practical nor sustainable. It asks a person to identify what needs care, who or what creates the relevant concern, the consequences of failure, likelihood, practical burden, and allies.[2] These are not answers to be applied mechanically. They are questions that prevent a vague request for “security” from becoming an empty command.
The useful question is not “Is this manuscript private?” It is “What is exposed at this stage, who may encounter it through which route, what consequence matters, and what legitimate activity must still be possible?”
That formulation changes the work. A private early draft containing tentative names does not pose the same decision as a review copy inside a defined editorial process. A public essay is not equivalent to the correspondence, deleted passages, or contributor information that may surround it. A deposit intended to help future researchers is not simply an old version of the writer’s working environment. The manuscript has crossed a boundary; the model should cross one too.
2. Begin with the asset, not a generic instruction
A filename can make a manuscript appear to be one object. A working record is often a collection of different things. It may contain readable prose alongside notes, comments, names, alternate versions, permissions information, correspondence, images, metadata, and editorial decisions. Any one of those components might matter for reasons that do not attach to the others. A final public text does not automatically turn every adjacent artifact into public material. Conversely, a working draft can be shared responsibly when its intended audience and condition are clear.
NIST’s Guide to Data-Centric System Threat Modeling describes threat modeling as a form of risk assessment that considers attack and defense aspects of a particular logical entity, and it focuses attention on selected data within a system.[1] That document is an initial public draft designed for organizations and systems, not an author’s private writing practice. Its value here is deliberately narrow: start by identifying the particular asset that matters rather than reaching first for a stock list of safeguards.
The Manuscript Stage Threat Register borrows that discipline of specificity. It does not implement NIST’s method, establish compliance, or transform a writer’s workflow into an organizational security program. It instead asks whether the reader is referring to final prose, a comment history, an editorial exchange, a list of names, a version selected for review, or an archive package whose access has not yet been decided. Such distinctions are not bureaucratic excess. They are the difference between a condition a reader can examine and a label—“the file,” “private,” “secure”—that hides the decision.
The following frame defines stage-sensitive manuscript exposure.
| Field | Reader question | Why it matters |
|---|---|---|
| Stage | Where is this material in the life of the work? | A private draft, shared edit, formal review, release, and archival transfer organize access in different ways. |
| Specific asset | What exactly needs attention in this situation? | The prose, a set of names, a comment trail, a version history, and a rights record can carry different implications. |
| Audience or plausible actor | Who needs access, may encounter the material, or might receive it in error? | A named collaborator, an undefined recipient, a reviewer, and a future archive user are not the same relationship. |
| Access channel | By what route can the material be handed over, copied, lost, redirected, or made available? | A direct handoff, shared workspace, submission process, public page, and repository arrangement produce different conditions. |
| Consequence | What changes if exposure, loss, misuse, or confusion occurs? | The effect might be delay, loss of editorial control, damage to trust, unwanted identification, or impaired recovery. |
| Existing condition | What permissions, restrictions, copies, roles, or records already exist? | A proportionate response begins with actual circumstances rather than a fantasy of total control. |
| Recovery or exception | What access, collaboration, correction, preservation, or remedy must remain possible? | A manuscript is made to move. A workable model does not protect it by making its required work impossible. |
The final field prevents a central mistake. Security planning is not only the denial of access. EFF’s guide identifies allies and asks readers to weigh practical burden alongside consequences and likelihood; it emphasizes that privacy, cost, convenience, and available support shape what can be maintained.[2] Writers need editors. Editors need permissions. Repositories need a way to describe and provide appropriate access. A model that reads all other people as threats and all copies as failure cannot explain creative work.
3. One manuscript, five changing conditions
To see the model in use, follow a constructed manuscript, River Names. It contains an essay in progress, research notes, a few identities that may later be removed, and editorial conversations. This is not a real author’s record or a disguised case study. It is a fixed teaching scenario that lets the manuscript’s surrounding conditions vary.
3.1 Private draft: working material before circulation
At the private-draft stage, River Names may include unresolved prose, source leads, personal observations, names that need further consideration, and earlier phrasing that will not survive. The asset needing attention might be the whole working version, but it might instead be a short note that identifies someone. The audience could be one writer, with other people having only ordinary physical or digital proximity to the working environment. The channel might be a working copy, a copy that synchronizes elsewhere, a printout, or an attachment sent in error.
The paper does not ask a reader to regard every early draft as secret. Some writers intentionally share unfinished work; some work in classrooms, workshops, teams, or shared environments. The stage instead calls for a modest act of naming. Is this copy for the writer alone? Is it ready for a particular reader? Does it include material that should not travel with the prose selected for feedback? If the version is lost or misdirected, what would be difficult to repair—time, trust, context, a person’s identity, or the ability to distinguish working language from a finished statement?
A register entry at this point may reveal nothing more dramatic than a need to identify a shareable version before a request arrives. It may show that a particular loss would be inconvenient but recoverable. It may surface that the most consequential material is not the draft’s argument at all, but the notes embedded around it. The gain is not a technical prescription. It is a clearer account of the intended boundary.
3.2 Collaborative edit: access becomes part of the work
Once River Names is being edited with others, the text may be more deliberately shareable while other assets become newly important. Comments can reveal who gave a particular critique. A version history can expose discarded decisions. A co-author may have a legitimate need to see material that a larger workshop does not need. A request for feedback from two named editors is not the same condition as an invitation sent to an undefined group.
This stage is where a threat model can become unhelpful if it treats collaborators as hazards rather than participants. Collaborators can be the allies who make revision possible. Yet collaboration still has a shape: which people need which version, whether comments are intended to travel, whether an exported copy preserves the same boundary as a shared working space, and what happens if new readers are added.
The practical question is therefore not “How can sharing be eliminated?” It is “How can the manuscript reach the readers needed for this editorial task without silently expanding the audience for every contextual artifact?” In one situation, the answer might involve distinguishing a review copy from a working copy. In another, it may require a stated feedback scope or an agreed path for reporting a mistaken handoff. The paper does not prescribe any of these as a universal workflow. It teaches the categories of decision that a broad word such as share can hide.
3.3 Submission and review: confidentiality is a process condition
Submission and review make roles, permissions, and channels unusually visible. The relevant material can include the manuscript itself, author information, reviewer comments, editorial correspondence, supplementary records, and an understanding of the review model. A reviewer may need to evaluate a text without authority to redistribute it. An editor may need a decision record without a reason to reveal a reviewer’s identity. The manuscript continues to circulate, but within a workflow whose participation has a stated purpose.
Research publishing provides a bounded illustration. Springer Nature’s editorial policy says that correspondence among editors, authors, and peer reviewers should be treated as confidential by default.[3] The Committee on Publication Ethics states that peer-review responsibilities vary with the model in use and instructs reviewers to respect confidentiality, avoid using review information to advantage or disadvantage others, and obtain permission before involving someone else in reviewing a manuscript.[4]
These sources establish norms for scholarly publishing. They are not offered as universal legal, contractual, or ethical rules for literary, classroom, trade, or independent editorial practices. Their teaching value is that they show confidentiality as a workflow condition rather than a mood attached to a document. If a review model, participant list, or route changes, the exposure analysis may need to change as well. A closed editorial assessment, an anonymous review, an open report, and a public post-publication discussion do not make the same promises.
A stage register makes that difference visible before convenience decides it. It asks whether the people involved have defined roles, what channel the stated process permits, what the manuscript carries alongside its prose, and what response is available if a role or route changes. It does not ask a writer to imitate scholarly peer review where it does not belong.
3.4 Public release: the public work has a perimeter
Publication changes the purpose of access. The selected text may now be intended for public reading, quotation, retrieval, discussion, and retention. In this stage, circulation is not a defect to be prevented. It is the object of the work. But the public status of a defined release does not automatically extend to every artifact created while the work was made.
Private correspondence, unreleased versions, contributor contact information, permissions documents, editorial deliberations, and notes may remain subject to a different decision. Springer Nature’s policies, for example, treat confidential editorial correspondence as distinct from published material and require consent for the publication of identifiable personal data.[3] This paper does not treat that publisher policy as a legal analysis of privacy or consent. It is a concrete reminder that release is not one undivided event.
At this point, a reader can ask what is actually being released; what contextual material will accompany it; whether names, images, excerpts, and acknowledgments have the status intended; and who can correct or clarify a boundary if it has been misunderstood. The appropriate action may be a particular public version, a change to contextual material, a delayed release, an explanation, or another arrangement shaped by the work’s setting. The register has not been designed to make that choice for the reader. It prevents the phrase “the work is public now” from ending the analysis too early.
3.5 Archival transfer: access for people not yet known
Archival transfer introduces a future audience into the model. The materials may include published work, drafts, correspondence, images, metadata, and versions that help explain how a text took shape. Repository staff, future researchers, and other users may be relevant, even though their individual identities cannot be known at transfer. The access route can include deposit, processing, catalogue description, access copies, reading-room arrangements, and discovery systems.
The National Archives of the United Kingdom supplies a useful vocabulary for this situation. Its digital-preservation access workflow instructs custodians to identify intellectual-property and other access restrictions, including sensitive and personal information. It refers to terms such as closed, open, conditional access, and sensitive, and it discusses redaction, access copies, discoverable catalogues, and controlled onsite access.[5]
That vocabulary avoids a weak binary: make everything available or hide everything forever. It also marks an essential boundary in this paper. Archival access controls govern future use of retained material. They do not substitute for deciding what a current draft should reveal during collaboration, and they cannot undo a version already distributed beyond its intended setting. Likewise, a carefully restricted transfer need not require the removal of every contextual record that could help a future user understand the work.
An archival-stage register asks what needs an access condition; what should remain discoverable; whether an access copy or redaction has a role; how a restriction can be revisited; and what context would be unnecessarily lost by a blunt removal. The focus is not “secrecy after publication.” It is a deliberate relation between preservation and access.
4. The Manuscript Stage Threat Register
The Manuscript Stage Threat Register is the paper’s original teaching instrument. It is intended for a specific handoff, release, review process, or archival transfer. It is not a numerical risk score; it cannot certify a file as safe; and it does not claim to predict every concern. A completed row may reveal a sensible safeguard, a reasoned decision to accept a small risk, a need for clarification, or a requirement to consult a person or institution with the relevant responsibility.
The register is arranged in two passes so that it can be read and completed at ordinary page scale. The first table establishes what is moving and through which relationship. The second records why the boundary matters and what must remain possible. A reader may complete both rows for one stage before moving to the next; the paired tables are one instrument, not two separate assessments.
| Stage | Specific asset | Audience / plausible actor | Access channel |
|---|---|---|---|
| Private draft | Working prose, notes, and tentative names | Writer; people with ordinary proximity to the work | Working copy, synchronized copy, printout |
| Collaborative edit | Draft plus comments and version history | Named collaborators; an unintended recipient | Shared workspace, export, attachment, link |
| Submission/review | Manuscript, editorial correspondence, review materials | Editors and assigned reviewers | Defined venue workflow |
| Public release | Released text and surrounding nonpublic material | Public readers, media, future users | Publication page, copies, excerpts |
| Archival transfer | Versions, correspondence, metadata, associated files | Repository staff and future users under access rules | Deposit, processing, description, access copy |
| Stage | Consequence if exposure, loss, misuse, or confusion occurs | Proportionate safeguard or justified exception | Residual concern and recovery route |
|---|---|---|---|
| Private draft | Unfinished or identifying material is mistaken for release-ready text | State the intended working boundary and separate any shareable version | Can the status be restored or explained if copies diverge? |
| Collaborative edit | Comments, identities, or earlier states travel beyond the editorial task | Match access scope to defined roles; state a reasoned exception for wider review | Can participants recognize what is shareable and report an error? |
| Submission/review | A confidentiality expectation or permission condition is breached | Follow the stated process and obtain required permission for altered participation | What should occur if the model, personnel, or route changes? |
| Public release | Public status is assumed to include material not selected for release | Separate the release decision from unrelated records and context | Who can correct an unintended disclosure or clarify the intended scope? |
| Archival transfer | Sensitive material is opened without a condition or valuable context is needlessly unavailable | Record an appropriate access condition, redaction, or access-copy rationale | How will restrictions be reviewed without erasing preservation value? |
The register is most useful when used in short passes. First, identify the stage. Second, narrow the asset: “comments that identify readers” is more actionable than “my manuscript.” Third, identify a realistic actor and channel without assuming hostile intent. Fourth, state the consequence in ordinary language. Fifth, record the least burdensome workable response or the reason access must be allowed. Finally, ask what can happen after a mistake: clarification, restoration, correction, reporting to the responsible party, or a later reconsideration of an access condition.
The last prompt makes recovery part of the teaching model. Creative work is conducted by people across time. Readers are added, drafts are revised, access conditions are reconsidered, and errors occur. A file that cannot reach an editor cannot be edited. An archive with no way to revisit a condition cannot adapt. A collaborator who cannot tell whether a comment may circulate cannot reliably preserve the boundary. Recovery is not an afterthought to prevention; it is one of the conditions that makes a manuscript workflow usable.
5. The counterpoint: restriction can also damage the work
A stage register would fail if it became a catalogue of reasons to restrict every route. Manuscripts may require feedback, redundant care, accessibility, preservation, catalogue description, correction, publication, and future interpretation. Each of these functions depends on someone being able to encounter a defined portion of the material under an appropriate condition.
EFF’s guidance explicitly rejects the fantasy of a perfect choice shared by everyone; it frames security planning as a judgment about consequences, likelihood, privacy, cost, convenience, and what a person can actually carry out.[2] The National Archives workflow likewise presents more possibilities than a permanent open-or-closed decision. It identifies closed, open, conditional, and sensitive conditions; it allows for redaction and access copies; and it recognizes that online discovery can coexist with controlled or onsite viewing.[5]
The sources work in different domains, but together they support a teaching principle: the goal is governed circulation, not either total secrecy or indiscriminate openness. A writer, editor, or custodian should be able to say what has a condition and why; who receives a copy and for what task; what consequence is being addressed; and what possibility of correction, access, or future use remains.
This principle also keeps the word threat in proportion. A plausible actor in the register need not be malicious. It may be an unintended recipient, a group that has changed without a new agreement, a reviewer who must seek permission before bringing in help, or a future user whose access terms are still unsettled. A channel can be an ordinary export, a submission workflow, or a repository catalogue. The register asks what makes access possible and what social agreement makes that access legitimate. It is not a theory that every manuscript is under attack.
6. A proportional rule for boundary crossings
The paper’s reader action is deliberately compact. Before a manuscript moves into a new relationship, the reader can follow one sequence:
Name the stage. Identify the specific asset. Identify the actual role and channel. Name the consequence worth avoiding. Record the smallest workable safeguard and the collaboration, recovery, or access exception it must not destroy.
This rule cannot guarantee a good outcome. It cannot decide whether a particular person is trustworthy, whether an institutional policy applies, whether a technology is suitable, or whether a legal duty exists. NIST’s Privacy Framework describes itself as a tool for improving individuals’ privacy through enterprise risk management, while NIST’s data-centric threat-modeling guidance addresses organizational system and data contexts; neither source produces an individual manuscript-security verdict.[1] [6]
The Register is therefore a teaching device rather than a compliance framework. Its contribution is to make one familiar error more difficult to overlook. The writer in the opening scene did not choose an irrational precaution. The writer took care designed for one room into another room with different participants, thresholds, and obligations. Manuscript privacy is not the search for a permanent lock. It is the discipline of revisiting a changing relationship when the work crosses a boundary.
Limitations
This is a conceptual teaching manuscript, not a security audit, legal opinion, technical guide, or publisher policy. It does not guarantee confidentiality; prescribe products, technical configurations, encryption approaches, or individual security plans; or assess a reader’s personal risk. The River Names scenario is constructed and is not evidence about a real author, editor, archive, or workflow. NIST SP 800-154 is an initial public draft for data-centric system threat modeling, so the paper uses only its bounded-asset premise rather than claiming adoption of a NIST method.[1] EFF’s resource is public security education rather than an individualized assessment.[2] Springer Nature and COPE supply research-publishing and peer-review examples, not rules that automatically govern all literary, educational, professional, or independent workflows.[3] [4] The National Archives (UK) guidance concerns archival access and does not replace an earlier drafting or collaboration decision.[5] The Manuscript Stage Threat Register is Wexley’s original instrument. It should help a reader formulate a focused question and, where a situation warrants it, identify the appropriate institutional, legal, archival, accessibility, editorial, or security expertise to consult.
References
[1] National Institute of Standards and Technology, “SP 800-154: Guide to Data-Centric System Threat Modeling,” initial public draft, 2016.
[2] Electronic Frontier Foundation, “Your Security Plan,” last reviewed 2023.
[3] Springer Nature, “Editorial Policies.”
[4] Committee on Publication Ethics, “Ethical Guidelines for Peer Reviewers,” Version 2, 2017.
[5] The National Archives (UK), “Digital Preservation Workflows: Access.”
[6] National Institute of Standards and Technology, “Privacy Framework.”