Definition

An account of what something is or what a word means - and the question of what makes one correct.

Theme: Logic, language and interpretation
Tradition: Cross-tradition
Term: ὅρος · horos
Term fixed: c. 330 BC

In plain terms

Asking for a definition sounds like asking for a fact. Often it is asking for a decision.

When a word already has a settled use, a definition reports it and can be wrong. When a word is being extended into new territory - a new technology, a new legal category, a case nobody anticipated - there is nothing to report. Somebody has to decide, and the decision then has consequences that look like discoveries.

And the classical model, in which a definition gives conditions that are individually necessary and jointly sufficient, almost never works outside mathematics. Attempts to define ordinary words that way have failed so consistently that the failure is itself a finding: most concepts do not have that shape.

As a philosopher would put it

Aristotle distinguishes real definitions, which state the essence of a thing by genus and differentia, from merely nominal accounts. Locke's version separates real essence, unknown to us, from nominal essence, the abstract idea we actually sort by - which relocates definition from metaphysics to classification.

The classical model requires necessary and sufficient conditions, and Wittgenstein's family-resemblance argument is the standard objection: look at what we call games and you find overlapping similarities with no common feature, so the demand for a definition presupposes a structure the concept does not have.

The technical distinctions still matter. Stipulative definitions introduce a term and cannot be wrong, only useful or not. Lexical ones report usage and can be. Persuasive definitions smuggle an evaluation into what looks like a report - defining democracy so that only one arrangement qualifies. And precising definitions sharpen a vague term for a purpose, which is what every legal and scientific reform of a boundary actually is.

Where you meet it

On 24 August 2006 the International Astronomical Union voted on what a planet is. The immediate cause was Eris, found the previous year and more massive than Pluto: either the solar system had a tenth planet, or Pluto was not one.

The adopted definition has three conditions - orbits the sun, is massive enough to be rounded by its own gravity, and has cleared the neighbourhood around its orbit. Pluto meets the first two. The third excludes it, and it was reclassified as a dwarf planet.

The vote was contested at the time and remains so; some planetary scientists argue that clearing the neighbourhood is a dynamical criterion imported into what should be an intrinsic classification, and would count Pluto and several moons as planets. Nothing about Pluto changed in 2006. What changed was a boundary, drawn by a show of hands among the astronomers who happened to be in the room on the last day of the assembly - and the resulting fact is now taught in schools as though it were an observation.

What is actually disputed

Whether definitions report or stipulate.

On the reporting view a definition can be checked against usage and got wrong, which is how dictionaries work. On the stipulating view, at the boundary that is where all the interesting cases sit, there is nothing to report and the definition creates the fact. Most real cases are mixed, which is what makes them contentious: the participants argue as though they were discovering something while doing something closer to legislating.

Separately, the classical model is in poor health. Family resemblance, prototype effects and the persistent failure of philosophical analysis all suggest that most concepts lack necessary and sufficient conditions. Defenders reply that this confuses how concepts are used with what they are, and that a concept can have a definition nobody can state - which is a defensible position and an unfalsifiable one.

Commonly confused with

Essence c. 330 BC An essence is what a thing is. A definition is an account of it, and on the nominal view there may be no essence for the definition to report.

Defined against

Concept A definition purports to state a concept, and the evidence that most concepts have no statable definition is what puts the pair in tension

Philosophers

Aristotle Real definition by genus and differentia
John Locke Real and nominal essence, and the limits of definition
Ludwig Wittgenstein Family resemblance, and the failure of the classical model
Rudolf Carnap Explication - replacing a vague notion with a precise successor
Socrates The demand for definitions that his interlocutors could never meet

Works

Metaphysics The classical account
An Essay Concerning Human Understanding Nominal essence and its consequences
Philosophical Investigations Family resemblance
The Logical Syntax of Language Definition within a chosen framework
Republic The Socratic demand, in its most sustained form

Schools

Peripatetic school (Aristotelianism) Where real definition is the target
British Empiricism Where it becomes classification
Ordinary Language Philosophy Where the classical model is abandoned
Vienna Circle Where explication replaces analysis
Nyaya Where definitions are tested against over- and under-extension

Related concepts

Concept 1781 · Category c. 330 BC · Essence c. 330 BC · Meaning 1892 · Sense 1892

Aristotle's real definition states the essence by genus and differentia; the modern distinction between real definitions of things and nominal definitions of words is Locke's.