Semantics concerns the conventional meaning of linguistic expressions; pragmatics concerns how people use those expressions in context and what an audience can infer. In brief, semantics asks what an utterance says by virtue of its words and structure, while pragmatics asks what a speaker or signer means by using it here. [1]
Start with the sentence
Imagine someone in a warm room says, “The window is open.” The sentence’s semantic content describes the window as open. That description can be assessed without knowing why the speaker said it. Yet a listener may infer a practical request: perhaps close the window, or notice the draft. The request is not spelled out in the sentence; it is inferred from the situation and shared expectations.
Pragmatics studies that contextual work. It considers who is speaking, what they know, what has just happened, and what action makes sense in the exchange. The same sentence can be a neutral observation, a warning, or an indirect request in different circumstances. The words stay the same while the likely communicative point shifts.
Meaning can depend on context
Pronouns and phrases such as here, now, and that need context to identify what they pick out. “Put that over here” does not specify the object or location on its own. A gesture, shared attention, or earlier conversation may supply them. The linguistic expression contributes a general instruction, while the context helps determine its particular reference.
Context can also shape how a question functions. “Can you pass the notebook?” has the grammatical form of a question about ability. In an ordinary shared setting, it often functions as a request. A person who answers “Yes” without passing it has responded to the literal ability question but may have missed the intended action. The distinction describes interpretation; it does not mean that every indirect request is understood identically.
A contrast that is easy to test
Compare “Some of the lanterns are lit” with a situation where every lantern is lit. The sentence can be true if some are lit, even when all are. A listener might infer that the speaker means not all are lit, because speakers often choose a more informative statement when it is relevant. That “not all” inference is pragmatic: it is a reasonable conclusion from the choice of wording, but it is not forced by the basic meaning of some. [1]
This illustrates why implication and entailment should not be collapsed. If a statement entails a conclusion, the conclusion follows from its meaning. A pragmatic inference can be plausible and useful yet disappear when the context changes. If a speaker says “Some are lit—in fact, all of them are,” the continuation is not contradictory, even if it revises an expectation.
A short listening exercise
Use this invented exchange. At the end of rehearsal, one person says, “The chairs are still in a circle.” Write down two readings: the direct content and a possible intended point. The direct content says something about the chairs’ arrangement. Depending on the shared task, the person may be asking others to reset them, noting that the room has not been prepared, or simply reporting what they see.
Now change one detail at a time. If the group has been asked to arrange chairs in rows, the remark may signal a problem. If someone asked where the chairs are, it may answer the question. If the speaker is reading a stage direction aloud, it may be neither a complaint nor a request. The sentence’s basic content remains similar while its contextual force changes. This is a practical way to separate semantic content from pragmatic inference.
Use the distinction without overreading
When interpreting a message, first paraphrase what the words conventionally express. Then list what the context suggests, and mark those conclusions as inferences rather than certainties. This two-column habit can prevent a listener from treating an implication as a promise or a request as a literal statement. It is equally useful when revising dialogue: decide whether a character should state a point plainly or let the scene make it inferable.
Semantics and pragmatics are connected: people use context to resolve meaning, and the available wording shapes what can be inferred. The distinction is an analytic tool, not a claim that real conversations arrive in separate layers.
