Conflict reports
Flag: --audit (or --compute).
The conflict report is the mirror’s flagship: it surfaces where what you asserted disagrees with what your own structure implies. It ships the conflict; it never auto-corrects.
A separate channel
Conflicts are not diagnostics. Diagnostics judge
a document’s form (is it well-formed?). Conflicts judge its coherence (do
your beliefs hang together?). A document can be perfectly strict-clean and still
carry a conflict — that’s the interesting case. So conflicts ride their own
channel, in an audit section, and never affect strict parsing.
"audit": {
"conflicts": [
{ "kind": "confidence-vs-status", "severity": "error",
"subjects": ["site-engineer-holds-conditions-are-fine", "conditions-are-fine"],
"message": "`site-engineer` asserts confidence 0.88 in `conditions-are-fine`, but your own structure defeats it (argument status: out)" }
]
}
Each conflict has a kind, a severity (error or warning), the subjects
it concerns, and a human-readable message. That output is
examples/pour-the-slab.thml
run with --audit, verbatim.
confidence-vs-status
It compares each authored stance’s confidence against the grounded argument status of its target. Two cases fire:
| Condition | Severity | Reading |
|---|---|---|
target is out and confidence ≥ 0.66 | error | high credence in a claim the structure defeats |
target is in and confidence ≤ 0.34 | warning | low credence in a claim that survives every attack |
(A confidence range is taken at its midpoint. Stances on targets that don’t take part in the attack graph are not compared.)
The first case is the flagship — you wrote down the objection and believed the claim anyway. The second is the inverse tell — you’re underweighting something your own evidence upholds.
It reports; it does not decide
This is worth stating plainly, because it’s the whole design. When the mirror
finds a confidence-vs-status conflict, it does not:
- lower your confidence,
- flip the argument status,
- or tell you which one is right.
Maybe the structure is incomplete (a missing rebuttal would change the status). Maybe the confidence is the honest number and the structure overstates the attack. The mirror can’t know — you do. It just makes the disagreement impossible to miss.
What it cannot see
The mirror checks a document against itself, never against the world. A
falsification pass (crates/thoughtml/tests/falsification.rs) pins the boundary,
and it is worth knowing precisely, because everything inside it is trustworthy
only if you know where it ends.
These produce an empty conflict report and confident-looking numbers:
| Error | What happens |
|---|---|
| A false premise | earth-is-flat supports maps-are-wrong is valid reasoning from nonsense. The conclusion derives ≈0.73. |
| One fact entered twice | Two ids for one observation both support a claim. Independence is assumed, so it compounds to ≈0.96. |
| Evidence never written down | Omit the two failed trials and the graph is coherent at ≈0.86. |
There is a fourth that is visible in the graph, and so is caught — but only if
you ask. a supports b and b supports a derive above 0.5 for both, from no
outside evidence at all. thoughtml check <file> --lint reports that as
TML502, off by default like every modelling lint.
The first three are not bugs waiting to be fixed. They are the limit of what any self-consistency check can reach, and they are exactly why the mirror reports rather than decides. The part it cannot do is yours: write down the evidence that cuts against you.
The bundled pour-the-slab.thml exists precisely to
demonstrate this: clean document, real conflict, no verdict.
definition-divergence
The second conflict type catches a different kind of disagreement: the same focus defined more than once with differing content.
{ "kind": "definition-divergence", "severity": "warning",
"subjects": ["launch-date"],
"message": "`launch-date` is defined more than once with differing content; all 2 definitions are kept" }
Ordinarily a repeated focus id merges
(first-wins on body / quantity / formula). But when a later mention states a
genuinely different value, ThoughtML does not drop it — every alternative is
retained on the focus’s divergent list, and this conflict points at the
disagreement. It’s the lossless-authoring tell: two agents (or two of your own
passes) wrote down incompatible versions of the same thing, and the mirror asks you
to reconcile them rather than picking one silently.
More conflict types are coming
confidence-vs-status and definition-divergence are the first two. The conflict
report is built as an extensible channel; future readings (calibration drift,
numeric inconsistency, stale beliefs) will land here as additional kinds — each
one a disagreement surfaced, never a decision made.