↩ ExploRational · A Catalogue of Divergent Rationality
explorational · a divergent move
Around 1121, Abelard filed 158 questions on which the Church Fathers flatly disagreed — Sic et Non, yes and no — and declined to resolve them. The conflict was the syllabus. Assemble the respected authorities that contradict each other, file them side by side, and treat the standing contradiction as information.
Classical logic has a famous house rule: from a contradiction, anything follows. Admit one true contradiction and every claim becomes provable, so the appearance of contradiction is treated as an alarm — something, somewhere, must be discarded before reasoning can resume. The rule is correct inside a formal system, and it trains a reflex that follows you outside one: when two things you trust disagree, kill one quickly. The reflex is the target of this move. In life, unlike in the calculus, a contradiction between two competent witnesses is almost never vacuous. It is a sensor reading.
What does a standing contradiction indicate? Usually one of three things, each valuable and each destroyed by premature resolution. Sometimes two senses are sharing one word, and the contradiction is a term asking to be split — the neighbouring move asking “in what sense?” finishes that case. Sometimes two frames are crossing: each authority is right inside its own system, and the clash marks the exact seam where the systems meet — the most informative location in either. And sometimes the contradiction is load-bearing: “ship fast” and “never break” are not a confusion to be tidied but a tension the organization is built to hold, and every working design is a particular way of holding it. F. Scott Fitzgerald's line — “the test of a first-rate intelligence is the ability to hold two opposed ideas in the mind at the same time, and still retain the ability to function” — describes the capacity; the ledger below is its filing system.
David Chapman's meta-rational reading goes one step further: the demand that everything eventually reconcile is itself a choice, made inside a particular picture of what knowledge must be. Outside that picture, a well-documented, long-standing, twice-respected contradiction can be the most honest summary available of a nebulous domain — and paraconsistent logicians have even built calculi where a contradiction sits quarantined without poisoning the rest. You need not go that far to use the move. Keep the ledger; visit it; resolve entries only when a resolution actually arrives.
the signature instrument
Six standing contradictions. For each, two readings: the reflex (kill one side) and the collector's (what the standing tension indicates). The reveal names which of the three kinds each entry is — split terms, crossed frames, or load-bearing — and what keeping it open buys.
The contradiction ledger
qualitative — six entries, three kinds; the classification is the exercise
Open an entry. The reflex reading and the collector's reading will differ; the difference is the move.
what to try
1 · Start the ledger. One page, two columns, Abelard's format: claims you substantially trust that contradict each other. Rules: both sides must be positions you or people you respect actually hold — no strawmen — and no entry may be resolved on the day it is filed.
2 · Classify before you resolve. For each entry ask in order: split terms? (then index the senses and the entry may close); crossed frames? (then name both systems and mark the seam — the entry stays, now as a map); load-bearing? (then ask what design currently holds the tension, and what it costs). Only unclassifiable entries are candidates for someone actually being wrong.
3 · Read the ledger under pressure. Before a big decision, reread your entries. A decision that silently assumes one side of a filed contradiction is a decision with a known unpaid debt; either pay it in argument or record that you are borrowing.
the mapping
| in the instrument | at your desk |
|---|---|
| an entry | two trusted claims in genuine conflict, filed verbatim, sources named |
| the reflex | the trained urge to discard one side before thinking resumes |
| split terms | one word, two senses — the entry dissolves under indexing |
| crossed frames | two systems, both right at home — the entry becomes a map of the seam |
| load-bearing | a tension the enterprise is built to hold — the entry becomes a design constraint |
where it tears
risk · flattening
The ledger has closings: plenty of filed conflicts end the old-fashioned way — one side was wrong, and the evidence eventually says so. A collector who never closes an entry has converted a method into a museum; Abelard's students were expected to argue their way through the filing, and some questions did get answered.
risk · the comfortable both
Load-bearing tensions still require a setting of the dial, today, for this release, this child, this budget. Naming the polarity does not excuse you from choosing a point on it. The ledger informs the choice; it is not a substitute for one.
risk · quarantine failure
The paraconsistent lesson cuts both ways: the calculi work because the contradiction is contained. A team that holds “ship fast / never break” openly must still decide which claims inherit from which pole, or the tension leaks into every estimate. Keep the entry; police its blast radius.
File the conflict, classify it before resolving it, and let the ledger — never the reflex — decide which contradictions were information.
can you use it?
RECOGNITION — Two senior engineers you trust say, respectively, “microservices saved us” and “microservices ruined us.” The move is: A: decide who is more senior. B: file both verbatim, then ask what each team's context made true. C: conclude the topic is subjective.
B. The contradiction is filed, then classified — almost certainly crossed frames: two organizations, two truths, and the seam (team size? domain boundaries?) is the finding. A resolves by status; C flattens by reflex.
THE NEAREST NEIGHBOR — Arguing both sides also holds two positions. What separates collecting the contradictions from it?
Duration and object. Arguing both sides is an exercise you perform once, on one question, to stretch your own reasoning. The ledger is an archive of other people's standing conflicts, kept open across months — the value accrues from what stays unresolved, not from your performance on either side.
PRODUCTION — File one entry tonight: two claims you actually rely on that contradict each other, with sources. Classify it.
“Premature optimization is the root of all evil” (Knuth, as usually quoted) vs. “performance must be designed in, not added later” (every architecture review). Classification: split terms — “optimization” means micro-tuning in one and structural budgets in the other; the entry mostly dissolves under indexing, and what survives is a genuine tension about when structure freezes. Yours works if both sides are things you cite approvingly, and the classification changed what you'd do next.