The Three Levels of Truth

Pragmatism, correspondence, and coherence as a single hierarchy

Put three questions to a ship’s navigator about her chart. Why do you care whether it is accurate? Because I want to arrive safely. What would make it accurate? Its matching the actual coastline, the actual depths, the actual hazards. How do you know it does? Because it agrees with my other instruments, with the charts of other navigators, with what I saw the last time I sailed these waters — and nothing I encounter contradicts it.

Three questions, three answers — and each of the three classical theories of truth has seized one answer and mistaken it for the whole. Pragmatism heard “because I want to arrive safely” and concluded that truth is usefulness: a belief is true if it enables successful interaction with reality and reliably fulfills goals. Correspondence theory heard “matching the actual coastline” and concluded that truth is accurate representation: a statement is true if and only if it depicts the way the world actually is — a map-to-territory relationship. Coherence theory heard “it agrees with everything else” and concluded that truth is internal harmony: logical consistency and explanatory power within a system of beliefs, since contradictory beliefs cannot simultaneously be true.

Philosophers have spent more than a century treating these as rival accounts, and each rivalry generates a familiar stock of objections. Correspondence is intuitive but faces the obvious embarrassment that we can never step outside our perceptions to check a belief against unmediated reality. Coherence risks circularity and solipsism — a perfectly consistent fantasy is still a fantasy. Pragmatism invites relativism: if truth is whatever works, then whatever works for you is true for you, and the word has stopped doing its job.

I propose that much of the rivalry comes from answering different questions. Pragmatic considerations explain why agents value accurate inquiry and which distinctions matter for a task. Correspondence states a realist success condition for empirical claims. Coherence is one family of methods for assessing claims within a wider evidential network. This is a useful division of labor, not a proof that every theory of truth reduces to one strict hierarchy.

Why We Want Truth

Start at the top: the pragmatic imperative. Truth-seeking is driven by pragmatism. We do not chase true beliefs because they are intrinsically decorative; we chase them because correct beliefs about reality lead directly to greater agency — more reliable prediction, more effective decisions, more successful action in the world. An organism indifferent to truth walks off cliffs. The fundamental reason humans value truth is that it pays.

Pragmatism alone cannot define truth. Utility divorced from accurate representation is often fragile, although idealized and even false models can work reliably within restricted domains. Pragmatic success supplies evidence that a model preserves relevant structure; it does not make usefulness identical to truth.

What Makes Beliefs Useful

Correspondence anchors pragmatism. If true beliefs did not reflect external reality, their usefulness would reduce to pure chance: any pragmatic success would be accidental and short-lived. Engineers, doctors, scientists, and navigators all rely on correspondence explicitly — their effectiveness depends on accurate maps of reality, in the literal case and the figurative one alike.

The navigator makes the dependence vivid. Her pragmatic goal is safe arrival. That goal requires a chart that preserves the geographic structure relevant to navigation. The chart need not reproduce the territory in full; its usefulness depends on selective, sufficiently accurate correspondence at the task’s grain.

Yet correspondence is notoriously elusive. We never apprehend reality in unmediated form — our perceptions, measurements, and interpretations are all indirect, all filtered through models of one kind or another. No one has ever held a statement up against the world and observed the match directly. This is the standard objection to correspondence theory, and it is a fair one — as an objection to correspondence as a method. As an account of what truth is, correspondence survives it untouched. The objection merely forces the question down a level: if we cannot inspect correspondence directly, how can we reliably infer it?

How We Check

Through multiple, mutually correcting methods: observation, measurement, intervention, prediction, replication, and coherence with other well-supported claims. Because correspondence is not inspected from outside every representation, it is inferred through the performance of this network. Coherence is indispensable within that network, but it is not the sole empirical test.

Science exemplifies this. No scientific theory has ever been confirmed by someone observing its correspondence to reality; theories gain credibility by integrating vast, diverse bodies of evidence into coherent frameworks. A theory is accepted when it minimizes contradictions, explains observations comprehensively, and meshes with established knowledge. And when coherence fails — when anomalies, contradictions, and unexplained phenomena accumulate — the theory’s presumed correspondence is what comes under suspicion. Incoherence is how misrepresentation announces itself.

The crucial point is that coherence does not replace correspondence; it helps operationalize and audit claims about it. Internal consistency alone cannot distinguish a coherent fiction from a successful empirical model. Sacred Coherence will defend coherence as an owned epistemic commitment while preserving its dependence on observation and correction.

The Hierarchy at Work

Medicine displays all three levels operating at once.

The pragmatic goal is healing: curing illness, saving lives, promoting health. That is why medical knowledge is pursued at all.

The correspondence foundation is physiology: medical effectiveness depends on theories that accurately reflect human biology, disease mechanisms, and biochemical reality. A treatment that works does so because the model behind it matches the body it intervenes in.

The assessment layer is the entire evidential apparatus: clinical trials, statistical analyses, replication, peer review, mechanistic models, and consistency across populations and contexts. Together they support or weaken an inference of correspondence; no one component is sufficient by itself.

Arranged this way, the hierarchy dissolves the classical objections rather than answering them one by one. Correspondence keeps its foundational importance without anyone needing impossible unmediated access to reality — the access problem is handled a level down. Coherence escapes its circularity because it is no longer self-supporting: it is disciplined from above by pragmatic goals and pointed outward at correspondence, so a merely internal harmony that keeps failing in practice gets overruled. And pragmatism loses its vulnerability to relativism because “what works” is no longer free-floating: it is anchored in a correspondence relationship, validated through coherence, that does not care what works for you.

Truth on Conditions

This hierarchy is not a standalone doctrine bolted onto Conditionalism; it is Conditionalism’s account of truth seen from the inside. If all truth is conditional — if every truth claim holds only relative to a lattice of background assumptions, interpretive frameworks, and standards of evaluation — then the three levels are exactly what you should expect to find when you unpack the conditions of any working truth claim.

Pragmatism supplies conditions of purpose: goals fix relevant precision, grain, and acceptable idealization. They do not make an inaccurate claim true merely because it is useful. Correspondence supplies conditions of reference: the target and structure to which the claim is answerable. Coherence supplies part of the conditions of assessment, alongside observation and intervention, and is itself relative to stated rules and background theories. There is no view from nowhere at any level, but there are better and worse contacts with the target.

The division of labor and Conditionalism support each other without settling every debate about truth. A statement evaluated under explicit conditions, answerable to its target, and supported by converging evidence earns confidence without requiring a view from nowhere. The three-level framework is this book’s audit of that process, not a definition imposed on every use of “true.”