↩ ExploRational · A Catalogue of Divergent Rationality
explorational · a divergent move
True, false — and the two corners a yes-or-no habit cannot see: true-in-one-respect-false-in-another, and neither — the question doesn't apply. Walk a claim around all four corners and watch it stop being one claim.
Indian logic kept four corners where the modern habit keeps two. The catuskoti — literally "four corners" — runs every proposition through the full square: true, false, both, neither. The last two are live logical positions, not decoration. "Both" is what a claim looks like when it holds under one index and fails under another; "neither" is what it looks like when the question itself misfires — a category mistake, an undefined term, the wrong unit. The extra corners are precision instruments for exactly the cases a yes-or-no reflex mangles.
The Jaina logicians made the index mandatory. Anekantavada, the many-sidedness doctrine, holds that every judgment is true from a standpoint, and its method — syadvada — prefixes each assertion with syat, "in some respect." Their parable is the blind men and the elephant: the hand at the trunk reports a snake, the hand at the ear a fan, the hand at the leg a pillar. Each report is true in its respect; the brawl starts when the respects are dropped. None of this is mysticism. It is a checklist against premature binary collapse, and the bench below makes you run it.
the signature instrument
Type a verdict, or load one. Four panes wait for it, and each is gated: a pane will not close until it names the standpoint it speaks from — the enforced form is in respect of ___: … The gate checks form, never content. Every count is structural, and the ledger that grows below is the product.
The four-cornered check
0 / 4 corners · 0 respects named
In what respect is it true?
In what respect is it false?
Where does it split — true one way, false another?
Is there a reading on which the question misfires?
pick the way the question fails, then describe the reading.
corners closed: 0 / 4
the check completes when all four corners close.
no respects named yet — the ledger fills as the corners close.
the two missing corners
A yes-or-no habit files the third and fourth corners under failure of nerve. They are findings. "The project succeeded — true in respect of the shipped features, false in respect of the budget" describes the world more exactly than either verdict alone, and it says where the argument moves next. The both corner is where most live disputes actually sit: two people shouting true and false at each other are usually holding different respects and declining to name them.
The neither corner is rarer and sharper: the reading under which the question does not arise. Ask whether a cancelled project "succeeded" and there may be no fact to report — the word has nothing left to grip. Ask whether "we are happy" and the we may be a committee wearing a person-sized predicate. Declaring a misfire is a substantive claim about the question's structure, which is why the pane makes you name the failure: category, term, unit, or presupposition. Corner four asserts something.
what to try
Run a verdict you delivered this week — a project judged, a person summarized, a policy graded. Type it exactly as you said it. Then close the TRUE pane: the gate refuses "overall" and "in general," and the moment a real standpoint goes in, feel the verdict shrink to its size.
Feel the FALSE pane resist. For a claim you believe, false-in-what-respect is the hard field — and the check will not complete without it. The bench asks only for the standpoint from which a fair observer scores your claim false. There is one; a claim false nowhere is a tautology, and tautologies need no defending.
Close all four corners, then read the respects ledger top to bottom. That column is the finding: the one claim you started with has split into three or four smaller claims, each wearing its own address — and the re-statement pane assembles them into a sentence built for cross-examination.
indexed, not dissolved
Anekantavada is a discipline of assertion. The syat prefix marks a coordinate, never a shrug: the Jaina sevenfold predication holds each indexed judgment to be definitely true from its standpoint. "In respect of schedule, we succeeded" commits you to more than "we succeeded" — it names the ground you agree to be held to, and it stands even against an opponent who owns a different respect.
So the restated claim at the bottom of the bench is more assertive than the sentence you began with. A bare verdict retreats under attack — "well, I meant broadly." An indexed claim has nowhere to retreat to and no need: every clause says where it stands and defends that ground alone. Indexing splits a claim into claims; it dissolves nothing. The ledger is the claim's true shape, finally written down.
the honest bench
Every readout is structural. Corner completion counts entries. The ledger collates your respect fields verbatim. The duplicate flag is a normalized string comparison — lower-cased, trimmed, leading articles dropped — between respects on the true side and respects on the false side; the same string on both sides is either a slip or the fault line worth arguing, and the bench flags it without deciding which. The respect gate checks form only: at least two characters, at least one letter, and off a short stop-list of index-refusers ("everything," "overall," "in general," and kin). Whether a named respect is a real distinction or a dodge is your judgment. The bench refuses exactly one thing: a respect-less answer.
the move ↔ the bench
| The move | The bench |
|---|---|
| the claim | the verdict as you delivered it — one sentence, no coordinates, before indexing. |
| a corner | one logical posture toward the claim: it holds, it fails, it splits, or the question misfires. |
| the respect | the standpoint a truth is true from — the index a yes-or-no habit silently drops. |
| the ledger | the claim's actual structure: the standpoints it was averaging over, collected verbatim. |
| the duplicate flag | the same respect on the true side and the false side — a slip, or the exact fault line worth arguing. |
| the re-statement | commitment with its coordinates shown: each clause defended from its own named ground. |
how this opening fails
risk · flattening
Run loosely, the check becomes a way to never be wrong: every criticism answered with "true in some respect," every commitment softened into a cloud of standpoints, until the claim asserts nothing and its owner cannot lose. The enforced respect-index is the antidote built into the bench. A respect that cannot be named — that stays at "in a sense," "in some way," "in general" — is not a respect, and the gate holds. A ledger of standpoints nobody could locate or contest creates fog.
risk · corner-four abuse
Corner four has a precise job: readings on which the question genuinely misfires — a category error, an undefined term, the wrong unit. Its abuse is a classic of accountability-dodging: "was the launch late?" met with a seminar on the social construction of deadlines. The tell is selectivity — questions that misfire only when their answers would cost you. The bench keeps the abuse visible by gating all four panes at once: your misfire reading must sit in the ledger beside a named respect in which the claim is, after all, true, and one in which it is false.
Walk the claim around all four corners, and make every truth show the respect it is true in.
can you use it?
RECOGNITION — Which situation calls for the check? A: two teammates shouting "the launch succeeded" and "it failed" at each other. B: a criticism you would like to soften into "true in some respect" until nobody is wrong. C: a metric dropped and you need rival explanations for it.
A. A live true-versus-false shouting match is usually two unnamed respects — the both corner's home case. B is the near-miss: flattening, the fog the respect gate exists to block.
THE NEAREST NEIGHBOR — Perspective rotation also walks a question around standpoints. What makes this move different?
The verdict. Each named respect must score the claim definitely true or false, and the corners assemble into a re-statement you can be held to. Collecting views without scoring the claim in each is rotation.
PRODUCTION — You told the review "the migration worked"; a colleague flatly disagrees. Walk it around all four corners before looking.
True in respect of data integrity; false in respect of the cutover deadline; both — that split named; neither — for the modules not yet moved, "worked" rests on a false presupposition. Yours works if every corner names a standpoint a fair observer could locate and contest, the false one included.