Logical consequence

• 1. Introduction

  ›

  Logical consequence sits at the heart of every argument ever made. When someone says "therefore," they are staking a claim: that what comes next must follow from what came before. But what does "must" actually mean here? That question turns out to be one of the oldest and most contested in all of philosophy.

  Logic, as a discipline, exists largely to answer it. Every formal system ever devised, every proof written in symbols, every model built to test an argument, traces back to this single idea: can a conclusion follow from its premises in a way that is airtight, necessary, and independent of what any individual happens to believe?

  The Polish logician Alfred Tarski identified three features that any satisfying account of logical consequence must have. Teasing apart those three features, and the debates that surround them, is what this documentary is about.

• 2. Form Over Content

  ›

  Fred is Mike's brother's son. That sentence is perfectly understandable, and most people would immediately conclude that Fred is therefore Mike's nephew. But that inference is not a formal consequence. It depends entirely on what the words "brother," "son," and "nephew" happen to mean in English.

  Formal logical consequence strips meaning away. Whether one statement follows from another, on the formal account, depends only on the structure of the statements, not the contents they carry. Consider the argument form: all X are Y; all Y are Z; therefore, all X are Z. Every possible substitution into this structure yields a valid argument, regardless of whether X, Y, and Z stand for frogs, planets, or prime numbers.

  The contrast matters because it sets the boundary between logic and everything else. Material consequences, like the Fred-and-Mike example, are governed by world knowledge and the meanings of words. Formal consequences hold across all possible contents. A formal account of logical consequence needs to capture that difference precisely, which is why logicians use syntactic inference rules and symbolic notation rather than everyday language.

• 3. Two Frameworks

  ›

  Syntactic consequence asks whether a formula can be derived by applying formal proof rules step by step. The turnstile symbol used to write this relationship was first introduced by Frege in 1879, though its current standard use dates back to Rosser and Kleene in the period 1934-1935. Proof theory is the branch of logic that studies these syntactic relationships, entirely without reference to what the symbols mean.

  Semantic consequence works differently. A formula is a semantic consequence of a set of statements if and only if there is no model, no interpretation of the formal system, in which all the premises are true and the conclusion is false. Model theory is the branch that studies this side of things. The key phrase is "no model": semantic consequence is defined by the impossibility of a counterexample across all possible interpretations.

  These two approaches, proof theory and model theory, represent the main technical options available to logicians who want to give a rigorous account of what it means for one statement to follow from others. Whether the two frameworks always agree with each other is itself a deep question in mathematical logic.

• 4. The A Priori Condition

  ›

  Tarski's second requirement for an adequate account of logical consequence is that the relation must be knowable a priori: determinable without appeal to sense experience or empirical evidence. If you know that a conclusion follows logically from its premises, no new observation about the world can take that knowledge away.

  Deductively valid arguments can be recognized as valid without any recourse to experience. That is what distinguishes them from empirical generalizations, which can always be overturned by a counterexample. A logical consequence holds not merely for all observed cases but for all possible cases.

  Crucially, Tarski's framework treats the a priori property as independent of formality. Formality alone does not guarantee that empirical knowledge plays no role; the two conditions must be held separately and satisfied independently. That independence is one of the more subtle points in the philosophical analysis, and it marks the line between logical entailment and the many other kinds of reasoning humans routinely rely on.

• 5. Necessity and Possible Worlds

  ›

  Modal accounts of logical consequence trade on the idea of necessity. A conclusion is a logical consequence of a set of premises if and only if it is impossible for the premises to all be true while the conclusion is false. Equivalently, the conclusion holds in every possible world where the premises hold.

  Tarski identified a modal component as the third feature any adequate account must possess. The modal framing makes this precise: logical consequence is not merely a contingent pattern observed across many cases, but a necessary relationship that holds across all conceivable situations.

  The example that best illustrates this is simple. All frogs are green. Kermit is a frog. Therefore, Kermit is green. There is no possible world where both premises hold and the conclusion fails. Modal-formal accounts go further, combining the modal and formal criteria: it must be impossible for any argument sharing the same logical form to have true premises and a false conclusion. This combined standard is the most demanding version of what logical entailment requires.

• 6. Beyond Truth Preservation

  ›

  All of the accounts discussed so far share a common assumption: that a good inference is one that never carries you from true premises to a false conclusion. Warrant-based accounts challenge that assumption at its root.

  On the warrant-based view, the defining feature of a good inference is not truth preservation but justification preservation. A sound inference should never move you from premises you are justified in asserting to a conclusion you are not justified in asserting. This is roughly the position favored by intuitionists, who are skeptical of classical logic's treatment of truth in ways that outrun our capacity to verify claims.

  Non-monotonic consequence relations introduce a further departure from the classical picture. Standard logical consequence is monotonic: if a conclusion follows from a set of premises, it still follows when you add more premises to that set. Non-monotonic consequence allows that adding information can defeat a previously valid inference. The example is instructive: "Tweety can fly" follows from "birds can typically fly" and "Tweety is a bird"; but once you add "Tweety is a penguin" to the premise set, the conclusion no longer holds. Classical logic has no room for this kind of defeasibility, which is why non-monotonic systems were developed as a separate family of frameworks.