Lab · Research preview

Do not trust the conclusion. Reconstruct the authority.

The Lab is a plan for making derivations inspectable. Nothing described on this page has been built.

NONCANONICAL RESEARCH

Research preview. No software exists yet, and nothing here is accepted doctrine. The derivation system is an instrument for representing and testing derivations under the theory. It is not the theory, and its output will have authority effect none.

Three things to keep apart

What is accepted is the theory: its sources, its grammar and its chain, as set out under Canon. The project has also accepted, as a research roadmap and nothing more, the idea that this architecture might be formalized as a graph of derivations and a ledger of normative events.

What is current research is the groundwork for that formalization, listed under Research. No schema has been fixed, deliberately.

What is planned is everything below. The sequence is a plan, drawn dotted because none of it exists.

  1. Settle what the record may represent

    Before any schema is fixed, several questions have to be answered so that the software’s structure does not decide them: the minimality hypotheses, how organizations hold positions, and the exact difference between a normative source, an empirical fact, a piece of provenance evidence, an inference rule and a technical record.

  2. Write the formal specification

    A specification able to represent prior positions, sources, mechanisms, trigger facts, the resulting relations, scope ceilings, residual liberty, and each downstream stage, along with contradictions, unresolved dependencies and the point where a derivation terminates.

  3. Reconstruct the theory itself

    The first test is whether the formal system can rebuild derivations the theory already accepts: three foundational sources, one violation-to-liability derivation, one case where liability does not establish jurisdiction, and one rule with a real scope ceiling. A chain that cannot be rebuilt is classified, not quietly repaired.

  4. Audit an external framework

    A pilot reconstruction of a framework from outside the project, to see whether its propositions survive the same source-first test. External material enters with authority effect none, and passing a reconstructability test would not make it normatively valid.

  5. Build the runtime

    An engine over the validated graph that tests whether a claimed conclusion is derivable from a chosen source profile and proven facts. It must reproduce the earlier audits without substituting a hidden source.

  6. Open the Lab

    Public inspection of real derivations: source ancestry, trigger facts, the relations changed, scope ceilings, residual liberty, and where a derivation stops.

What it is meant to do

Given a claimed conclusion, a set of accepted sources and a set of proven facts, the system is meant to show whether the conclusion can be derived and, if not, exactly where the derivation stops. The intended answers are a short fixed list:

  • Derived
  • Not derived
  • Partially derived
  • Underdetermined
  • Open
  • Requires a separate source

The intended architecture runs from canonical sources to a formal graph, through deterministic checks, to a derivation record, with an explanatory layer on top. It does not run from a language model's intuition to an answer. A model may help normalize facts, propose candidate derivations, explain a result and ask adversarial questions. It may not invent a missing normative source.

What it cannot do

A record represents a derivation. It does not become a source because it has been recorded, signed, hashed, made immutable, validated by a machine or produced by software. If the theory's own derivation of a rule cannot be reconstructed, the failure is reported and sent back to the project's review process. The technical layer does not patch the canon.

The Lab will not be an oracle of objective morality. It will report what follows from a stated set of sources, and those sources remain open to dispute. It will not be a substitute for legal advice in any jurisdiction, and it will not describe what existing law requires.

How this system is built

The Lab is intended to operate over a controlled derivational structure: a canon whose status is explicit, with dependencies that are written down and changes that are separately reviewed. The runtime will not create normative authority. It will test whether a claimed conclusion is supported by the sources and dependencies supplied to it.

How Default Liberty is built