How it works¶
The Legible Protocol makes it possible to record what each person formulated, the criteria and choices they took responsibility for, and relationships with other work. This information can accompany a deliverable and help its recipient understand how it was produced, assess its reasoning or develop another solution.
The protocol organizes declarations about work, while documents, code, analyses and other materials can remain as referenced evidence. Each participant describes what they take responsibility for and can record relationships with other work. Recipients of these records can examine formulations and declared dependencies even without having followed the entire production process.
How can contributions and their relationships be recorded?¶
Legible defines three fundamental declaration types with different functions. They allow a formulation to be recorded, the work it drew on to be declared and a consequence received to be acknowledged afterward. Creating one type of record does not require the others to be produced too.
Appoint — Declare what you formulated.¶
An Appoint attributes a formulation to its declarant. That formulation can be a problem, hypothesis or solution, even if it has not yet been selected or executed.
The detailed work, including work carried out with AI assistance, can be referenced as evidence. Formulating the problem can itself be the contribution, without requiring a completed solution.
Trace — Declare what a piece of work or decision drew on.¶
A Trace declares which other work a formulation, solution or decision drew on. Its author identifies the related material and explains the dependency they assert. This relationship allows the use of a contribution to be followed without attributing the origins of everything incorporated to the new work.
A decision document can also be related to the work that informed it. The Trace makes that declared dependency examinable, while approval of resources and authority to decide remain within the corresponding organizational processes.
Nod — Recognize a consequence received.¶
A Nod records acknowledgment of a consequence received from an earlier contribution. The declaration is made by whoever experienced the benefit and describes what they acknowledge, with relevant references.
Beyond documenting what happened, that acknowledgment adds information other participants can consider in their own decisions. Knowing who acknowledged a contribution, under what circumstances and for which consequence can help people assess where to collaborate, what work to develop or which activities to support.
Recognition remains situated: it describes the experience of its issuer, and each reader assesses its relevance according to the evidence and criteria they consider pertinent. This act comes after the consequence and is distinct from the earlier decision to adopt, fund or implement a proposal.
Each declaration has its own author.¶
Each participant takes responsibility for the declaration they make. Formulating a problem does not mean predicting who will benefit from it; declaring a dependency does not speak for the authors of the related work. This distinction allows the content of each assertion and who took responsibility for it to be examined separately.
A relationship between pieces of work can also be declared by someone who did not participate in producing them. That person can create a Trace, explain the relationship they identified and present the evidence they relied on. The record identifies who made the assertion without automatically attributing it to the authors of the related work.
For a Nod, the declaration belongs to whoever received the consequence. Someone can recognize how particular work benefited their activity, but observing a benefit received by another person is not the same as receiving it.
Formal records are digitally signed. A signature allows verification of their integrity and the key that assumed each declaration, but does not by itself prove the truth, originality or merit of what was asserted.
Readers remain responsible for examining evidence and forming their judgment. The protocol also does not produce an official score to measure how much each person contributed.
Starting with a draft¶
You can prepare these accounts as unsigned drafts. The first step is to check whether the document corresponds to what each participant formulated or took responsibility for and preserves the relevant uncertainties and relationships.
In review, preserve the origins of formulations and describe separately the choices of those who used them. A question does not become a conclusion; adopting a proposal does not mean having originated it. If AI suggests an alternative, the description should distinguish the suggestion received from the human decision to adopt it.
A draft can remain informal throughout its use. If you choose to formalize it, you prepare a new signed declaration that can reference the draft as evidence. Keeping that record privately and sharing it are separate decisions.
Prepare your first draft · Understand records and verifiability
A rejected proposal can help shape another¶
Preserving a formulation and deciding to use it are different operations. A proposal that does not justify investment may contain a well-identified problem, an assessment criterion or an analysis useful in another context. Deciding not to execute that solution does not remove the relevance of those parts to other readings.
When the formulation remains accessible, another participant can return to it, assess its reasoning and develop an alternative. The new work can declare what it used from earlier work, preserving both the origins of formulations and what its own author added.
A subsequent adoption decision can also be related to the contributions that informed it. The record references the decision document and the work considered, while authorization to commit resources remains documented in the organization's processes. The dependency declaration makes the relationship asserted by its author examinable; it does not replace verification of the corresponding authorizations and conditions.
Wouldn't good documentation already make this possible?¶
Yes. Well-maintained documentation can also preserve problems, criteria and reasoning for future work. Legible adds common conventions so participants can take responsibility for their own declarations, record dependencies and transport and verify them across contexts.
Legible lets each participant adopt a declaration through a verifiable digital signature. The signature is sovereign in the sense that it can be made with a key under the participant's control, without prior approval from a platform or the institution receiving the work. The signed declaration can be stored privately, shared and verified in other environments.
A person can associate that key with the identity by which they are already known. Readers who know that association can also consider the author's professional history and relationships of trust when examining the declaration, even if it is their first Legible record. The key allows the signature to be verified; its connection to the person is known and examined through other means.
The common format preserves these declarations and the relationships between contributions, even when they arise at different times, in different tools or in different organizations.
Evidence can remain in the tools where the work was produced. A record can reference an analysis, code repository, test or decision document. That reference preserves the distinction between the declaration taken responsibility for and the material that allows it to be examined.
Recipients in another environment can examine those relationships without using the same application. Access to evidence continues to depend on its permissions: referencing an internal report does not make its contents public.
Result, history and genealogy¶
While the result presents the work delivered, the history retains a conversation that mixes questions, attempts, instructions and suggestions. Without an account of what each participant chooses to take responsibility for, understanding a decision may require the reader to make that selection again.
A declared cognitive genealogy describes contributions participants choose to declare: a question that changed the analysis, an adopted criterion, a correction or a relationship to other work. It points to relevant evidence, including the result and its versions.
It can begin informally and remain that way. Formalization means signing a specific declaration, without turning this selection into a complete reconstruction of thought. The work and its records remain distinct; a reference may point to evidence that remains restricted.
What each person's history adds¶
Different people can take responsibility for a similar formulation with different experiences and relationships. That history may be known through their profession, previous work or direct relationships, even before their first Legible record.
Someone who knows the author can consider this context when examining the declaration. Records help relate what the author assumes to declared work and consequences, but they do not produce an official reputation or reconstruct the person's entire history.
A question, incomplete work or a mistaken hypothesis can also be described. Recording participation allows it to be examined; it does not make the result correct or require an extensive history before someone can start.
Git, PROV and notes with AI¶
Legible relates to resources that already preserve work and context: Git allows contributions to develop before integration into a main repository; the W3C PROV model describes provenance, including human participation; and good notes with AI can retain criteria and decisions. The protocol's specific question is how to exchange declarations of contribution, relationship and consequence between participants in different contexts.
| Resource | What it offers | How it can relate to Legible |
|---|---|---|
| Git | Versions, branches, distribution and integrity through content addressing; commits can be signed | Can store Legible records, whose semantics need not correspond to commits |
| W3C PROV | A general model of entities, activities, agents and provenance relationships | Could support a mapping of declarations; that correspondence would need specification and is not yet implemented |
| Notes with AI | Context, rationale and retrieval according to the organization of the notes and the tool's capabilities | Can remain in use; Legible adds common conventions for interchange, relationships and formalization |
The intended difference should emerge in practice: less effort to understand and continue contributions across contexts, with origins and relationships preserved. This must outweigh the additional work of preparing and reviewing records. A fair comparison uses the same people, AI, information and channels, including good, well-maintained notes.
References: Git fundamentals, distributed workflows and the W3C PROV primer.
From documentation to coordination¶
A preserved formulation can inform decisions different from those that motivated its creation. Its relevance also depends on who examines it, the question guiding the reading and the knowledge to which it can be related.
The same set of records can therefore support different readings. Developing a solution, investigating its dependencies and assessing its risks are questions that can lead to different parts of the collection being considered. Each participant can consult accessible material with their own tools and criteria without having to adopt a single interpretation of the work.
For this possibility to be effective, references need to allow relevant materials to be located, and their access conditions must be clear. A formulation not selected for a particular use can remain available for other readings, provided someone preserves it and keeps it accessible under the corresponding conditions.
When choosing to formalize, each participant can sign their own declaration and preserve relationships with the work used. This allows the content taken responsibility for and its attribution to the key to be verified, while the author's known identity and history, the evidence and the conditions of each activity help readers form their judgments. Those who decide to collaborate can do so with a clearer understanding of what others assert and what still needs investigation.
AI tools can help locate relevant contributions, compare criteria and prepare questions from accessible material. The intended gain depends on that assistance reducing the effort needed to understand and develop others' work enough to offset the cost of preparing and reviewing records.
The exchange can then produce knowledge that was not yet documented: a question can reveal an unknown constraint, a test can change a criterion and reading a proposal can lead to another problem being formulated. Each participant decides what to add and take responsibility for. This work can be developed before an investment decision and remain accessible for other uses under the conditions in which it is preserved and shared.
For contributors, preserving that participation and being able to follow declared uses and consequences also informs their choice to keep collaborating. Operating in Open Legibility explains how to organize this practice and which economic conditions can sustain it.
Next steps¶
Prepare a draft · Share and develop · Participate in a space · Create a space
Specification · Glossary · Articles and books
Usage examples¶
Follow a service provider delivering in cycles, an R&D team preserving observations before funding tests and a team preparing decisions before meetings. The examples area shows what each person declares and how they use the records to choose the next step.
The economic coordination guide develops these cases: contributions before demand, access to results and recognition chosen after the consequence. It is an initial presentation of possibilities and hypotheses to examine in practice.