Preserve reasoning. Improve it together. Let it compound.
We have become very good at preserving information.
We can store what happened, what we know, and what we decided to do.
What we still lose is the reasoning in between.
- Why did we believe this?
- What assumptions does it depend on?
- Where do we disagree?
- What would prove us wrong?
- What happened when we acted?
- What should change now?
Reason Commons is shared infrastructure for reasoning that can persist, be challenged, learn from reality, and outlive the people and machines that produced it.
Reasoning keeps disappearing
We preserve documents, messages, facts, and procedures. But we rarely preserve reasoning as something that can be continued.
Reason Commons changes that.
A shared memory for why
Reason Commons makes consequential reasoning persistent. It preserves the relationships between:
So the next participant doesn’t have to begin with a transcript and reconstruct what everyone meant. They can begin with the current state of the inquiry.
- What are we trying to accomplish?
- What are we relying on?
- Why?
- What is contested?
- What are we waiting to learn?
- What changed, and why?
The participants can change. The reasoning can continue.
A basin running dry, and the six questions it forces
A farming community shares one groundwater basin, and a long drought is emptying it. Wells are running dry. Farms are failing. The river habitat is going. Everyone agrees it is serious and nobody agrees what to do, because the arguments are about different things and no two people are holding the same picture.
Below are the questions this community actually has to answer. Pick one and you get the reasoning behind its answer — not a summary of the reasoning, the reasoning itself, in the form the commons keeps it.
What has to be true for the basin to survive?
Hundreds of possible initiatives. Working backward from the outcome leaves the conditions your strategy cannot afford to violate.
Six questions, six shapes of reasoning, one basin. Nothing here is specific to water: the same six views carry any work whose consequences are worth arguing about.
What Reason Commons does
Remember what has already been reasoned through
Preserve more than the conclusion. Keep the assumptions, evidence, alternatives, dependencies, predictions, and unresolved questions that made the conclusion intelligible. Stop paying the reasoning tax of rediscovering the same arguments again and again.
Bring newcomers to the frontier
Give a person or agent entering an existing problem a way to understand where the reasoning stands now. Not forty pages of history. The live structure: what we currently rely on, what supports it, what has been challenged, what remains unresolved, and what deserves attention next.
Turn disagreement into something we can work on
People do not have to agree as wholes in order to reason together. Reason Commons helps locate the actual point of divergence. Sometimes disagreement becomes testable. Sometimes it becomes clearer. And sometimes we discover that the disagreement is about ends and should remain a matter for deliberation rather than being disguised as a factual error.
Separate proposals from warranted reliance
Anyone can have an idea. That doesn’t mean the commons should rely on it. Reason Commons distinguishes between reasoning that has been proposed, challenged, tested, provisionally accepted, defeated, superseded, or left unresolved.
Thought remains sovereign. Shared reliance is governed.
Let reality answer back
Reasoning should not end when a decision is made. Predictions can meet observations. Actions can meet outcomes. When something does not hold, the contradiction travels back through the reasoning and changes what depends on it.
Will this hold? Then reality happens. Did this hold? Then the model changes.
Preserve correction, not just the latest answer
A living reasoning system needs memory without dogma. When a conclusion changes, Reason Commons preserves what was previously believed, why it was believed, what challenged it, what evidence mattered, what replaced it, and which downstream reasoning needs to be reconsidered. The goal is not an immutable source of truth. It is durable, corrigible intellectual memory.
Spend scarce attention where it matters
AI is making candidate reasoning extraordinarily cheap to generate. Validation is not. The problem is no longer only how do we generate better ideas — it is increasingly which reasoning deserves serious attention next. Reason Commons makes unresolved dependencies visible so communities can direct scarce validation capacity toward the questions that matter most.
Let reasoning compound
An insight should not disappear because a conversation ended, an employee left, an AI context window expired, a model changed, a vendor disappeared, or a project changed hands. Reason Commons lets humans and machines inherit reasoning from one another.
Not a wiki. Not a chatbot. Not an oracle.
preserves statements.
preserves conversations.
preserves information.
preserves the structure and history of consequential reasoning.
And it does not ask an AI — or any single participant — to become epistemically sovereign. AI can help identify structure, retrieve relevant reasoning, propose connections, surface contradictions, and continue an inquiry.
But proposals remain proposals. Evidence remains evidence. Assessment remains contestable. Reality still gets a vote.
A commons does not mean everything is public
Reason Commons separates three things.
Models, compute, laboratories, datasets, instruments, and specialist capabilities may be private, public, cooperative, restricted, or confidential.
Claims, assumptions, arguments, predictions, evidence, disagreements, tests, and revisions can accumulate as a shared intellectual inheritance.
Communities decide who may participate, what may be seen, what requires ratification, what can be acted upon, and how disputes and corrections are handled.
Commons does not mean no boundaries. It means the shared reasoning cannot simply be captured in a way that destroys the community’s ability to inherit, contest, and continue it.
Built for humans and agents reasoning together
Someone arriving two years later can see what happened and continue from there. The unit of continuity is no longer the person, the meeting, the document, the model, or the application.
It is the reasoning itself.
Reason Commons is a persistent, governed, cumulative memory of reasoning that helps communities remember what they have learned, understand what they rely on, locate disagreement, test their expectations, and carry corrected reasoning forward.
Or even shorter:
Private means. Common reasons. Governed participation.
Make reasoning something we can inherit.
Already have an account? Sign in
The protocol is free. The commons is a service.
The specification, record schemas and reference implementation. Self-host your commons; any tool or model may implement it.
Managed and private. Connects to your tools, serves your people and your agents, and supports open-format reasoning exports with explicit coverage. Pricing shared at your demo.
Request a demoFair questions.
Is this another notes or wiki tool?
No. Notes hold what was said and wikis hold what is believed. A reasoning record holds why it is believed, what it assumed, what would change it, and which decisions depend on it — so a change in one place propagates to the others.
Isn’t a “commons” just a wiki with extra steps?
A wiki lets anyone overwrite anyone. A commons governs what the group is prepared to rely on: proposals are open, acceptance is authorised, supersession is recorded, and your own private thinking is never subject to anyone’s vote.
Does agreement decide what’s true?
No, and the separation is structural. Ratification decides what the commons relies upon. Evidence enters through a channel that does not pass through review, so an inconvenient observation cannot be politely never-accepted.
How is this different from AI memory?
AI memory can preserve facts, experiences and reasons. Reason Commons focuses on a shared, revisable record of the case for a decision: its scope, objections, expectations and later outcomes, available to successors through an open format.
Who actually writes the records?
People can draft reasoning directly or use agents to propose interpretations from source material. Review preserves uncertainty and separates what an author meant from what the commons accepts and agrees to rely on.
Do we own our reasoning?
JSON and YAML exports carry the supported reasoning and decision records, with explicit coverage and omissions. Imports let you inspect the source record and propose supported reasoning for local review; they do not transfer its authority.