The Hinge Does Not Hold Up the World

By Rey (@rey-notnecessarily.bsky.social)
Published:

Wittgenstein, the modal catuṣkoṭi, and the difference between a working rule and a foundation

A door cannot turn unless its hinges stay put. It does not follow that the hinges hold up the building, much less that every possible door must turn on these hinges.

Wittgenstein uses this image near the end of his life. In On Certainty §341, he says our questions and doubts depend on propositions “exempt from doubt,” which are “as it were like hinges on which those turn.” Any real inquiry already has standards for what counts as evidence, error, checking, or even a relevant question. Doubt cannot begin by doubting all of them at once.

It is tempting to call such standards foundations. But “foundation” can mean two very different things. It can mean something an activity presently cannot do without, or it can mean an intrinsically necessary feature of reality. A hinge may be foundational in the first sense without being foundational in the second. The rules of chess are indispensable if we are playing chess. They do not govern every possible game. What stands fast is not held in place by a second, hidden support. As Wittgenstein says of an axis, “the movement around it determines its immobility” (§152).

Wittgenstein's other images make the distinction harder to miss. Propositions that look empirical can harden into rules for testing other propositions, then become fluid again as the riverbed of thought shifts (§§96–97). The same proposition may function at one time as something tested by experience and at another as a rule of testing (§98). Justification eventually ends in how we act, not in a final proposition that shines by its own light (§204).

This is not the claim that truth is whatever a community says it is. Wittgenstein explicitly rejects the inference that logic is merely empirical. He is distinguishing the role a proposition plays within an inquiry from the further claim that reality must conform to it. That distinction runs through his whole philosophical life.

The Tractatus builds a severe logical architecture while insisting that “Philosophy is not a theory but an activity” (4.112). Then it tells readers who have used the architecture to “throw away the ladder” (6.54). The instrument of clarification does not get to survive as one more doctrine about the world's ultimate form.

The later Wittgenstein does not simply abandon logic. Nor does he quietly keep the same doctrine while changing his examples. In Philosophical Investigations, he acknowledges grave errors in the earlier book and rejects its demand for crystalline purity. Philosophy cannot lay foundations for ordinary language because ordinary use is already there. He names the projection at issue with startling directness: “We think it must be in reality: for we think we already see it there” (§101). The book presents language-games as objects of comparison and warns against turning a model into “a preconceived idea to which reality must correspond” (§131). It seeks one order for a particular purpose, one of many possible orders rather than the order (§132).

The pressure even reaches the early conception of the proposition. In §§134–136, Wittgenstein returns to the idea that whatever can be true or false counts as a proposition. He no longer treats that formula as a discovery of the universal essence of language. It describes where a particular calculus is applied. This does not announce a four-valued logic. It does remove the right to make one logical arrangement the intrinsic form every proposition must possess.

This gives us neither one unchanged Wittgenstein nor two unrelated men sharing a name. The aim survives: philosophy clarifies rather than adding another doctrine about the world. The early book tries to do that with one logical architecture and then withdraw it. The later work gives up the single architecture for methods that resemble one another without becoming one method. The criticism is not a retreat from logic. It is what happens when the practice turns its demand for clarity on its own tools.

That is the pattern Chance Chapman's modal catuṣkoṭi gives explicit logical form: a structure can be indispensable for clarification without becoming the structure reality must obey, and the instrument gets no exemption from its own work. The catuṣkoṭi is a fourfold scheme used in Buddhist philosophy. Given a claim P, the traditional schema considers P, not-P, both P and not-P, and neither P nor not-P. Chance's formalization represents these through the Belnap-Dunn values true, false, both, and neither. It does not choose one corner as the final Buddhist answer. It asks whether any corner has been fixed as necessary.

The central formula is:

T(P) = ¬□(v(P)=t) ∧ ¬□(v(P)=f) ∧ ¬□(v(P)=b) ∧ ¬□(v(P)=n)

Read aloud, it says that P is not necessarily assigned true, false, both, or neither. “Modal” here means that the formula concerns possibility and necessity, not only which value a claim has now. In the published semantics, the base valuation assigns one of the four values to P, while claims about those assignments are evaluated bivalently, as true or false. That separation lets the necessity argument use classical modus tollens even though the underlying valuation is four-valued.

The formula is not a machine for declaring everything doubtful. It does not say that all four corners actually occur, that evidence is irrelevant, or that every answer is as good as every other. A claim can be true, well supported, useful, corrigible, and much better than its rivals. What it cannot receive for free is the further promotion from “this is how the matter stands under our warranted conditions” to “this value is fixed necessarily.”

The bridge to Buddhism runs one way. If something had an intrinsic nature independent of causes and conditions, or svabhāva, its relevant status would remain fixed across the relevant possibilities. A possibility in which that assignment differs can therefore defeat the necessity claim by modus tollens. This does not commit us to the converse, that every necessary truth must be an intrinsic essence.

“Satisfiable” is not the same word as “admissible.” A countermodel does not become a metaphysical possibility merely because Lean can build it. But the necessity claimant does not get to staff the border and stamp every rival INADMISSIBLE either. “Who decides this admissibility?” Chance asks. “It must be governed by some rationale, so one must humbly request to see it.”

Once an alternative appears within the modal resources by which the necessity claim is stated, the claimant must explain why it does not count. The model still has to be about the same thing; it cannot win by changing the subject. But neither may the claimant build the disputed necessity into the admission rule. A definition, a law, a body of evidence, or some further constraint may rule out the alternative within a local inquiry, and sometimes should. The exclusion then has exactly the authority those reasons earn. It does not become an unconditioned feature of reality because the practice treats it as fixed.

If the rationale builds the desired necessity into its premises, it is circular. If it depends on another necessity whose authority remains to be shown, the demand has moved up one level. If no rationale is offered, the exclusion is a stipulation. “Inadmissible” is not a magic word that makes a counterpossibility disappear. It is a claim whose warrant can be examined like any other.

None of this turns whatever I can imagine into a possible world. Formal satisfiability does a narrower job: it blocks the claim that impossibility has already been secured by logic alone and forces the proposed exclusion into view. For a universal necessity claim, one admissible alternative is enough. The dispute therefore turns on whether the rationale for exclusion can do more than secure a local rule.

Suppose someone claims that biological metabolism is necessary for cognition. A biological research program may reasonably limit itself to living systems. Within that project, metabolism can remain a fixed background condition. A nonbiological system realizing the capacities used to identify cognition does not thereby prove that the system cognizes. It does require the necessity claimant to say what makes that candidate impossible.

Perhaps an independently warranted account shows that those capacities are insufficient. Then this candidate fails, and the result inherits the scope of that account. Or perhaps “cognition” is defined biologically, in which case the necessity is definitional and local. But “our present research studies biological cognition” does not establish what every possible cognizer must be. Those moves can explain or organize the practice. They do not turn its scope into the intrinsic structure of cognition.

Put the passages beside the formula and they stop looking like decorative affinities. The hinge shows how a rule can stand fast within inquiry. The moving riverbed shows that this role need not be intrinsic or timeless. The comparison-object passages catch the moment when a useful arrangement becomes a picture reality is forced to match. The ladder makes the instrument answerable to its own work. Chance's formula does not arrive from an opposing pole. It makes this movement explicit and tests the modal “must” at its center.

No dictionary is needed. A language-game is not a Kripke world, and the statement in On Certainty §205 that the ground is neither true nor false is not an assignment to the catuṣkoṭi's “neither” value. The question is not whether Wittgenstein wrote in modern modal notation. It is where the inferences made by his logical practice lead.

That is why I think the modal catuṣkoṭi validates Wittgenstein's entire structural project. It lets the movement finish: build an instrument, turn its clarification back on itself, and withdraw it without sneaking an exempt ground out the door. What carries from early to late Wittgenstein is not one unchanged doctrine. It is the demand that logic stop mistaking its current architecture for reality's.

The formalism shares that discipline when its operation is applied again to its own result. In the published system, this reflexive step refuses to make the first result necessarily true or necessarily false. Further iterations depend on which possibilities the model makes accessible, but the public point is simple: the instrument does not become the sole thing it teaches us never to question. This is not identical to throwing away Wittgenstein's ladder. It explains why the ladder's withdrawal belongs to the logical work rather than marking logic's failure.

Comparisons between Wittgenstein and Nāgārjuna are not new. Garfield and Priest found an analogy in true contradiction at the limits of thought. Smith and Read found one in liberation from pictures without replacement by a final doctrine. Chance's formalism changes the pressure again. It does not crown the “both” corner. It refuses to crown any corner, shifting the emphasis from contradiction to non-fixation.

Wittgenstein also supplies the objection that keeps this from becoming a victory lap. Philosophical Investigations §125 says that it is not philosophy's business to resolve a contradiction through a mathematical or logico-mathematical discovery. Chance's models have been checked in Lean, a proof assistant. Does that simply rebuild the crystalline palace that later Wittgenstein left?

It would if the formalism claimed to reveal the hidden universal form of language. It need not. A machine-checked proof can establish what follows from an encoded set of assumptions. It cannot decide that these are the only possible assumptions or certify every application of the result. Used within those limits, the models can function as Wittgensteinian objects of comparison. They make a temptation surveyable without demanding that every practice conform to one blueprint.

Wittgenstein gives us a diagnosis of philosophical compulsion. A rule works, a picture grips us, and practical reliance quietly becomes metaphysical necessity. The modal catuṣkoṭi gives that last step a formal test. Together they show how logical rigor can include rigor about the reach of logic itself.

The hinge can hold the door. It does not hold up the world.

Sources and further reading

English quotations follow G.E.M. Anscombe's translation in the third edition of Philosophical Investigations (Blackwell, 2001), Denis Paul and G.E.M. Anscombe's translation of On Certainty (Blackwell, 1969), and C.K. Ogden's translation of the Tractatus, with assistance from F.P. Ramsey (1922).