To my understanding... Semantics is the raw meaning and connotations a word carries on it's own and pragmatics is the context-dependent meaning a word holds.

Is this accurate? Can anyone explain it to me simply?

  • 2
    Great question! After so many years, the difference is still not always 100 % clear to me. I hope someone comes up with a great answer.
    – Cerberus
    Sep 18, 2012 at 23:41
  • 1
    An example at EL&U
    – prash
    Sep 19, 2012 at 0:01
  • No, that is not accurate. Pragmatics deals with how people use language. (Christenings, curses, promises --- that sort of thing.)
    – Greg Lee
    Jan 11, 2016 at 23:39

7 Answers 7


Nothing much. The distinction is a more or less artificial one.

Supposedly, pragmatics is any part of what we laughingly call the "meaning" of an utterance that can be figured out by the listener.

This includes

and other topics, like speech acts, irony, and politeness. Everything else is semantics. Or so it goes.

Well, this is not helpful, since you constantly have to be trying to scry out the mental calculations of various speakers in various situations to tell whether you're studying semantics or pragmatics. This is especially troublesome when semantics gets mixed up with logic -- which is rather like confusing speech with writing, since logic is just a written representation for natural semantics.

Jim McCawley put it this way, when he was describing his research areas:

Semantics/logic/pragmatics (it's impossible to talk in any detail about any of these three fields without getting into the other two, so I don't even try to keep them separate). I teach courses on logic from a linguist's point of view, taking a broad view of the subject matter of logic (logic has suffered from 23 centuries of myopia, which I try to make up for) and giving full weight to linguistic considerations in revising (or replacing) existing systems of logic to maximize their contact with natural language syntax and linguistic semantics. (See my book, Everything that Linguists Have Always Wanted to Know About Logic (but were Ashamed to Ask), University of Chicago Press, 2nd edition, 1993). I also from time to time teach courses in lexical semantics, tense and aspect, and speech acts (with Erving Goffman sharing top billing with J. L. Austin).

And I agree with Jim. There's no reason to believe that humans consistently distinguish one kind of information from another, and in fact plenty of reasons to believe they differ greatly in what they take for granted; so the null hypothesis has to be that there is no clear, definitive difference, though certainly there are exemplary cases of one or the other.

  • Many informative links, thank you. +1 a couple don't work though? And more than a couple are well beyond me but I can appreciate the value of the less complex ones easily enough :)
    – LitheOhm
    Sep 20, 2012 at 18:30
  • 2
    It may be, at the end of the day, that there isn't a real distinction here. That said, not being able to see how humans distinguish one type of information from another doesn't constitute a basis for arguing that they don't. And I don't think that one has to confuse semantics for logic to want to make the distinction, e.g. terpconnect.umd.edu/%7Epietro/research/papers/cbc.pdf Dec 9, 2012 at 22:49
  • Clearly, some do. Equally clearly, others don't. And the rest of us are smeared across the middle. Nothing new here. All working parts of one's language (phonet-, phonem-, ... pragmatics) are, after all, self-built, and there's no reason to believe they resemble one another. Except that there can probably only be a certain number of ways to do anything effectively, and many people will either imitate others' solutions, or discover a solution independently. And others will never recognize that there is something to discover. Or get fixated on a wrong solution forever. We're all human.
    – jlawler
    Dec 10, 2012 at 5:06

You cannot talk about semantics without pragmatics and vice versa, but a decent way to cut the cake is a broadly relevance-theoretic view. Semantics is the stable aspects of encoded meaning which are always triggered and make a predictable contribution to the final interpretation (however minimal that encoded stuff might be). Pragmatics is the inferred and non-encoded and non-predictable aspects of meaning. Look at 'thoughts and utterances' by Robyn Carston for a good discussion of all this.


Just by coincidence I opened the latest issue of Language (88:3) and found on the first page a letter to the editor by Mira Ariel, responding to a review of her book Defining Pragmatics (CUP, 2010). The following statement is found therein:

The book's main claim is that once we base the grammar/pragmatics distinction on a division of labor between directly ENCODED AND UNCANCELABLE associations between forms and meanings or conditions of use (grammar) and CANCELABLE associations mediated by INFERENCE (pragmatics), 'a unified view of the field can thus be construed' (DP, p. xiv). In other words, I assert the coherence of a pragmatics that is defined as inference, while rejecting as incoherent the older approach to defining the field.

  • 1
    Which is fine, as long as one can always tell which associations come from inference, and which come from lexical coding; and as long as one can demonstrate that they are always the same for every speaker. Putting it mildly, this remains to be established.
    – jlawler
    Sep 21, 2012 at 23:46
  • 2
    @jlawler This definition does say how to tell pragmatic inferences from encoding: by cancellability. If you aren't sure if X was an inference or if it was directly encoded in an utterance Y, just ask yourself if the person uttering Y can immediately afterwards deny X. If you see it as a contradiction, it was encoded, if you don't, it was a pragmatic inference. And the problem of agreement between language users is present for virtually everything in linguistics that deals with more than idiolects.
    – michau
    Sep 8, 2016 at 15:52

My pragmatics professor1 was fond of the following example: if it is a Wednesday and somebody says "next Thursday", which Thursday is meant, the one tomorrow or the one next week? What if you don't know what day today is when you hear somebody say that?

To oversimplify: Context-dependent meanings sort under pragmatics; meaning you can sort out with the help of a dictionary sort under semantics. The border is rather fuzzy. Intonation plays an important role though. It is, I think, the trickiest part of learning a foreign language, for instance learning when (if) it is correct to say "please" and "thank you", and adapting to the other parties in the conversation.

1: My uni. divided linguistics equally into four parts: phonology, syntax, semantics and pragmatics. Phonetics was separate.

  • +1. I like your example and simplicity. In relation to learning foreign languages - do idioms factor into pragmatics?
    – LitheOhm
    Sep 20, 2012 at 18:32
  • @LitheOhm, dunno. I only took the needed minimum of pragmatics and idioms was not covered in any undergrad class. Besides, it was.. yikes.. more than ten years ago. "Women, Fire, and Dangerous Things" hadn't been out that long.
    – kaleissin
    Sep 20, 2012 at 18:43

From semiotics I gather that semantics is related to the abstract concept the word represents, while pragmatics is related to the purpose of its use. [S. Boell en D. Cecez-Kecmanovic, “Attributes of Information”, in AMCIS 2010 Proceedings, 2010, pp. 11.]

For example, if I say that my heart rate is 120 BPM, there will be no misunderstanding about its conceptualisation; its meaning. Still, a medical doctor cannot draw any conclusion from just this information alone. Context is required, such as age, gender, current and very recent activity (running or sitting) in order to draw conclusions about a certain aspect of my health.

Pragmatics define the sort of actions to take, which are dependent on the listener's goal. The pragmatics for a medical doctor examining a patient will therefore be quite different than those for a personal trainer coaching an Olympic athlete. Therefore the conclusion drawn (i.e. is a heart rate of 120BPM good or bad) is influenced by its use (pragmatics), while the meaning of "a heart rate of 120BPM" remains constant (semantics).

Note that context, used above to indicate "what else do we need to know before we can draw conclusions", therefore depends on pragmatics. The context required for the medical doctor might be different than for the personal trainer.


To make it explicit, Consider the following sentence:

"Hey, How you doin'?"

In Semantics the sentence just asks about how the other person is doing, Or how he/she feels etc. But pragmatics says that if a guy (joey!) tells this sentence to a female, it means that he might be hitting on her. In another view Semantics is affected by the meaning of the sentence in whole but Pragmatics is more or less affected by the socio-cultural variables.


Not an elaborate answer, but putting it in a nutshell pretty well IMO:

"Could you given an example sentence to illustrate the difference between semantics and pragmatics?"
- "Yes, I could."


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.