I have provided an example that hopefully highlights what I am trying to articulate.

Person A believes that Something can be bad or good, but not both (XOR)

Person B believes that Something can be bad and/or good (OR)

If X is deemed bad, and subsequently deemed good by the same source, Person A believes that the source has stated a contradiction, whereas Person B believes that it is simultaneously good and bad (albeit in different respects, or bad at t=0 and good at t=1, et cetera)

Because the way Person A and Person B interpret a statement differs, is the way they differ described by Pragmatics as opposed to the other fields of linguistics?

Below is the previous abandoned example.

Statement(S): Person A is bad, and Person A is good.

Evaluation System 1 (ES1) has axiom: "All subject complement's describe identical respects of Subject. (Within a single sentence)"

Evaluation System 2 (ES2) has axiom: " All subject complement's do not necessarily describe identical respects of a Subject. (within a single sentence)"

Inputting S into ES1, yields - terms "bad" and "good" describe identical respects of Person A, thus the statement is false.

Inputting S into ES2, yields - terms "bad" and "good" do not necessarily describe identical respects of Person A, thus is permissible, not false (unless Person A is further qualified, in other clauses or sentences.)

ES1 and ES2 differ in axioms used, and thus evaluate the statement differently, are then evaluation axioms to be considered as relating to Pragmatics?

(Apologies if I, in creating the question, used vocabulary in an inappropriate manner relative to their use in other well-established fields of linguistics.)

  • Are you asking whether pragmatics deals with the question of utterances having different semantic interpretations, depending on the condext? May 25, 2020 at 7:20
  • If the axioms that one uses to interpret something is context then yes?
    – TomDot Com
    May 25, 2020 at 7:27

1 Answer 1


This is likely in the domain of semantics rather than pragmatics because it relates to the truth conditions of the words rather than the evaluation of statements in context. In your updated example, the difference is that speakers A and B have different definitions of "good". In the linguistic sense, in order to evaluate the phrase "X is good", we interpret the word "good" as denoting a function, good(X,w), which takes the subject X and an evaluation context w and returns a truth value.

Lexical semantics deals with what goes on inside that function, and the answer to your question is that the difference is speakers A and B define the function good() differently. Speakers A and B both evaluate the same utterance and the same state of affairs, yet they return different truth values. The only possible conclusion is that they have different lexical semantics for good() because the rest of the function is held constant.

The rational speech act (RSA) model is a good framework to exemplify this (see Goodman & Frank, 2016). RSA defines a probabilistic model of world states given an utterance (from a set of possible alternative utterances). In this model, the semantic meaning of good() is operationalized as an explicit mapping from inputs (utterances and world states) to truth values. It would be at this level where the differences you ask about would be implemented. The definition of good() for Speaker A would return false for all utterances of "X is good", but speaker B would define good() in such a way that the utterance "X is good" returns true in at least one world state. Given these semantic differences, the model would predict different pragmatic inferences, but the main difference is the lexical semantics, not the pragmatic reasoning.

Getting much beyond that becomes a philosophical rather than linguistic question. "X is good" has no meaning outside of the way listeners interpret it, and evaluating sentences against metaphysical properties rather than speaker beliefs quickly leads to problems.

A useful but oblique example is gradable adjectives (see Kennedy & McNally 2005). Consider "X is tall". We could define tall() as simply evaluating whether X has the property "tall". However we can say sentences like "John is tall, but John is not tall for a basketball player." The same entity is tall in one context and not tall in another, meaning that the entity is both in and not in the set of tall things which is a logical contradiction. From this contradiction we can conclude that the attribute denoted by "tall" is not a property of the entity at all. A similar argument exists for "good" in the way you've been using it.

Edited to address comments.

  • They're not necessarily using different definitions of "good", but they're applying goodness to different respects. "respects" might not be the best word, maybe "aspects" or "dimensions", which describe properties of the object. You could infer that Person A is using a mono-dimensional evaluation schema or something like that, and so goodness of any one quality will be negated, and contradicated by the badness of any other, of identical or differing aspect.
    – TomDot Com
    May 27, 2020 at 17:06
  • Thus, it's possible that if X commits a good act, donating to charity, it is conferred the quality of goodness onto itself,(thus X is good), but if subsequently it commits a bad act, kicking a dog, the quality of badness is conferred onto it,(thus X is bad) and Person A finds the statement "X is good, X is bad" contradictory because although the aspects which good and bad apply to differ, because they confer goodness and badness upon X, it is a contradiction. Hopefully i'm articulating this clearly.
    – TomDot Com
    May 27, 2020 at 17:07
  • Whereas for Person B, they use a multi-dimensional evaluation schema, meaning the dimensions that good and bad describe, if not identical, can confer qualities of "goodness" and "badness" onto X, without the statement necessarily ( or universally) being considered a contradiction. Thus "X is good, X is bad" is not necessarily false (only when the respects that good and bad apply to are identical, would it be false).
    – TomDot Com
    May 27, 2020 at 17:19
  • Apologies for the lengthy response
    – TomDot Com
    May 27, 2020 at 17:19

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.