Originally written in French. Translated by AI — the meaning has been preserved, not the prose.
Main idea
A committee session is often judged on its ability to produce a consensus. The criterion is misleading: a complex organization needs its spaces for discussion in order to collect objections, weak signals and field constraints, not to manufacture a sentence everyone accepts.
What a committee can do, and that no other body does as well, is delimit. Separate disagreements of fact, of priority, of vocabulary and of risk. Say what cannot be reconciled. Formulate, for each possible direction, what it forces you not to do, to do later, to do less well, or not to satisfy entirely.
A committee that works this way prepares the loss instead of dissolving it: it doesn't necessarily decide, but it makes the decision possible elsewhere, and costly to dodge.
Layer added by "The Tools of Organizational Coherence" (2026-09-15). The criterion doesn't stop at formal bodies. A project glossary, a one-pager, a PR/FAQ or a hierarchy of metrics are judged in exactly the same way: they don't produce agreement, they reveal more clearly where it stops. What distinguishes a body from an artifact is therefore not their function but their reach in time — the session delimits once, in front of those who were there; the written object carries the delimitation between sessions and after the participants have gone.
Why it matters
This provides an end-of-session criterion that doesn't depend on the participants' comfort: you don't ask "do we agree?" but "where does our agreement stop, and what is the nature of what remains?".
It also rehabilitates sessions that end without a conclusion: a meeting that has delimited the disagreement has produced something, even if the minutes contain no decision.
Nuances and limits
Delimiting doesn't replace deciding. A committee that excels at qualifying disagreements and never escalates them to whoever can decide becomes a machine for endless deliberation.
And some disagreements don't classify cleanly: a conflict of priority often presents itself as a disagreement of fact, and the qualification itself is already an arbitration.
Open questions
- How do you recognize, at the moment a disagreement is qualified as a "vocabulary disagreement", that it isn't a conflict of priority in disguise?