⁂ Concepts · Knowledge · No. 56
Coherence

The books balance perfectly, and the sale never happened.

Coherence

Truth or justification understood as fit within a system, rather than as contact with something outside it.

Theme
Knowledge, truth and reason
Tradition
Western
Term fixed
1807

In plain terms

You cannot step outside your beliefs to compare them with reality. Every check you make is another belief. That observation is what coherence theories start from, and it is a good one.

So the proposal is that a belief is justified, or even true, by how well it fits with everything else you hold - not by touching something outside the system, since nothing does.

The obvious objection is equally good. A well-built fiction is entirely coherent. A conspiracy theory is coherent by construction, because every apparent counterexample is absorbed as further evidence of the conspiracy. If coherence were sufficient, those would be true, and they are not - so something else has to be doing work, and saying what has proved difficult.

As a philosopher would put it

As a theory of truth, coherence holds that a proposition is true in virtue of fitting a specified system of beliefs. Bradley's version makes truth a matter of degree and reserves full truth for the whole, which few have found usable. As a theory of justification it is far more widely held: BonJour's version has beliefs justified by their place in a mutually supporting system, with no foundational layer.

The isolation objection is the standing problem. Coherence is an internal property, so a system can be perfectly coherent and wholly disconnected from the world. Responses add observational constraints - beliefs that arrive spontaneously and carry some weight before being integrated - which critics say smuggles foundationalism back in.

Quine's web of belief is the influential modern shape: no belief is immune from revision and none is answerable to experience individually, since theories face the tribunal of experience only as a body. That is coherentist in structure while keeping the whole answerable at its edges, which is roughly where most epistemologists now sit without calling themselves coherentists.

Where you meet it

Double-entry bookkeeping is a coherence system. Every transaction is recorded twice, debits must equal credits, and the trial balance is a check that the system is internally consistent. It has been the foundation of accounting since the fifteenth century and it works.

It cannot detect a transaction that never happened. If a fictitious sale is entered as a debit to receivables and a credit to revenue, the books balance perfectly. The system will report no error, because there is no internal error - the inconsistency is between the ledger and the world, and the ledger has no access to that.

This is why audits exist and why the procedures that matter are the ones that leave the books: confirming a balance with the bank, counting the inventory, writing to a customer to ask whether they owe what the ledger says. Every one of those is an attempt to touch something outside the system, and the reason the profession requires them is that centuries of experience established that coherence alone is not enough.

What is actually disputed

Whether coherence is sufficient, or even necessary.

As a theory of truth it has few current defenders, because the isolation objection is very hard to answer and because more than one coherent system is possible with no way to choose between them. As a theory of justification it remains a serious contender, since the regress problem gives foundationalism trouble of its own.

The live question is whether any coherentist account can accommodate observation without becoming foundationalist. Adding privileged input beliefs answers the isolation worry and concedes the structure; refusing to add them keeps the theory pure and leaves it unable to explain why a system should track anything. Quine's tribunal metaphor is a way of having it both ways, and whether it is a solution or a description of the problem is not settled.

Commonly confused with

  • coherence in physics An unrelated technical sense about waves maintaining phase, sharing only the word.

Defined against

  • Correspondence The two rival accounts of what truth is, and the case against each is stronger than the case for either

Philosophers

Works

Schools

Related concepts

Correspondence c. 1270 · Truth c. 330 BC · Justification 1641 · Non-contradiction c. 330 BC · Knowledge c. 369 BC

The idea that truth is the whole is Hegel's; the explicit coherence theory belongs to the British idealists - Bradley, Bosanquet, Joachim - around the turn of the twentieth century.