Meaning in context: what follows, and what is only suggested
Understanding a report requires more than deciding whether it sounds natural. Separate what its words require, what the situation establishes and what a cooperative speaker normally suggests.
Course topic: Semantics and Pragmatics
Study route and exercise: University Reading List editorial project. How we curate · Sources checked:
On this page
Scope and responsibility
The cases use stipulated English readings and a fixed four-lamp domain. They illustrate semantic tests, not universal judgments about every language, dialect or conversational setting.
Choose your starting point
Start free with explicit tests
Kroeger connects descriptive meaning analysis to logical relations and pragmatic diagnostics.
Investigate word meaning
Cruse is useful when sense relations, polysemy and lexical distinctions drive the question.
Deepen contextual analysis
Levinson supplies a sustained pragmatics route after introductory linguistic concepts.
Compare the three books
Analyzing Meaning: An Introduction to Semantics and Pragmatics
Paul Kroeger · Edition: 3 · 2022
Free analytical foundation
- Choose it when
- You need worked distinctions between entailment, context and implicature.
- Use another route when
- You need a dedicated lexical reference or an advanced formal-semantics sequence alone.
- Selected reading
- Read the logical-relations discussion and chapter 8 on Gricean implicature, especially its diagnostic sections.
- Limits and trade-offs
- Mostly descriptive or semi-formal does not mean judgment-free: analyses still need explicit counterexamples.
Try before choosing
Write a world where “some lamps are lit” and “all lamps are lit” are both true under the supplied definitions.
Check your work: All four lit is a witness; it defeats the claim that “some” entails “not all”.
Meaning in Language: An Introduction to Semantics and Pragmatics
Alan Cruse · Edition: 3 · 2010
Lexical and contextual meaning
- Choose it when
- You need to explain a word’s senses and relations before analysing an utterance.
- Use another route when
- You only need a quick introduction to conversational structure.
- Selected reading
- Use the named discussions of contextual variability, sense relations and conversational implicature in the selected copy.
- Limits and trade-offs
- A lexical sense distinction must be argued; different contexts do not automatically create different dictionary senses.
Try before choosing
Contrast a change in the word’s stipulated meaning with a change in what a reporter knows.
Check your work: The lamp follow-up changes access to evidence, not the definition of “some”.
Pragmatics
Stephen C. Levinson · 1983
Sustained pragmatics
- Choose it when
- You can distinguish a proposition from an utterance and want deixis, implicature, presupposition and speech acts together.
- Use another route when
- You need a gentle first introduction to grammar or lexical semantics.
- Selected reading
- Begin with the named implicature discussion, then compare presupposition and deixis; follow your edition’s headings.
- Limits and trade-offs
- A 1983 foundational text is not a survey of all later experimental or computational work.
Try before choosing
Explain why a cancellable “not all” suggestion is unlike a contradiction between “some” and “none”.
Check your work: Give a satisfying model for the former and prove no satisfying model for the latter in the fixed domain.
Check your preparation
What is an entailment test?
Seek a case in which the first statement is true and the alleged consequence false. One valid countermodel defeats the entailment claim.
Does a false inference change the literal meaning?
Not necessarily. A defeasible inference may fail because speaker knowledge or conversational expectations fail. Keep those premises visible.
Can the domain change unnoticed?
No. Here every occurrence ranges over the same four lamps; “all” cannot quietly switch to a visible subset.
A six-step study plan
Steps completed: 0 / 6
These are your own study marks, not an assessment. They are not sent to this site.
-
Fix vocabulary and domain
Selected reading: Kroeger: logical relations.
What to do: Write definitions of some, all and none over A–D.
Check your work: Some means at least one, not exactly one or one-but-not-all.
-
Separate two evidence tables
Selected reading: Cruse: context and meaning.
What to do: Make columns for actual lamp states and the reporter’s observations.
Check your work: Knowing A is lit does not identify the other three states.
-
Test an entailment
Selected reading: Kroeger: entailment and chapter 8.
What to do: Enumerate an all-on countermodel to some⇒not-all.
Check your work: Keep the same domain and lexical definition.
-
Derive the suggestion
Selected reading: Levinson: implicature.
What to do: List the knowledge and informativeness assumptions supporting a not-all inference.
Check your work: The inference has premises beyond the sentence’s truth conditions.
-
Try two continuations
Selected reading: Kroeger: cancellation diagnostics.
What to do: Compare “indeed all four” with “none of the four”.
Check your work: Semantic consistency is separated from truth against a particular state table.
-
Revise under restricted knowledge
Selected reading: All three routes.
What to do: Write the main and follow-up conclusions in parallel.
Check your work: State exactly which truth conditions remain and which inference loses support.
Worked exercise
A truthful report can suggest more than it entails
Four lamps A–D have complete states: A on, B off, C on, D off. “Some lamps are lit” means at least one of these four; “all” means all four and “none” means zero. A reporter inspected all four, reports accurately, and is assumed to choose the stronger “all” when it is warranted and relevant. She says, “Some lamps are lit.” Assess literal truth, the suggestion “not all”, and two proposed continuations: “indeed all four” and “none of the four”.
Assumptions
The actual state table and the reporter’s access are supplied, not inferred. Informativeness and competence are defeasible conversational premises, not part of the stipulated lexical meaning. Domain A–D never changes.
- Two lamps are on: some is true, all is false and not-all is true from the table. Under the supplied full-knowledge/informativeness premises, choosing some also normally supports a not-all implicature.
- Some does not entail not-all: the alternative configuration on/on/on/on satisfies some and all together. This countermodel tests word meaning, not the truth of the actual table.
- “Some; indeed all four” has a consistent literal interpretation and cancels the usual suggestion, but is false against the actual on/off/on/off table. “Some; none of the four” is inconsistent in every configuration under the fixed definitions.
Answer and interpretation
The actual report is literally true and its not-all suggestion is supported here. That suggestion remains distinct from entailment and from the independent table evidence. Cancellation can be semantically possible without being factually correct in this particular world.
Test the result
Do not use the actual two-on table alone as a proof of entailment. Test all-on as a separate possible configuration, and never rescue “none” by changing its domain.
Change an assumption
The reporter sees only one lamp
Change the actual table to all four on. The reporter sees only A, knows it is on, and has no information about B–D. She says, “Some lamps are lit; for all I know, all four are.” All definitions and the four-lamp domain stay fixed.
- Some remains true in the actual all-on configuration. Her knowledge of A warrants at least one; it does not warrant her asserting all, even though all happens to be true.
- The full-knowledge premise for the earlier not-all inference is absent. Her explicit continuation cancels that suggestion and accurately describes her stipulated uncertainty.
- Some and all are compatible. Replacing the continuation with “none of the four” would still contradict some; limited knowledge does not change that semantic relation.
Answer and interpretation
The report can be truthful without suggesting a known failure of all. Actual truth, justified assertion and implicature come apart: all is true, the reporter cannot establish it, and she does not claim not-all.
Test the result
No domain restriction or new lexical sense was introduced. The changed access and explicit continuation explain the change in inference.
Common mistakes
Naturalness as entailment
A common suggestion can be very natural and still cancellable. Supply a countermodel rather than a confidence rating.
Speaker truth becomes omniscience
A true statement does not imply the speaker knows every relevant state.
Cancellation becomes factual correction
A consistent continuation may still conflict with the supplied facts; check both levels.
Before moving on
- Give a countermodel to a false entailment claim with domain fixed.
- Separate actual facts, speaker knowledge and conversational premises.
- Explain why cancellation differs from contradiction in both packets.
Sources and edition checks
Sources checked:
- Kroeger: official open third edition
The open third edition is available from Language Science Press under CC BY-SA 4.0, with individual credits to retain. Prior introductory linguistics is useful.
- Cruse: Oxford selected edition
Cruse’s route gives substantial attention to lexical meaning alongside pragmatics; match named topics to the selected third edition.
- Levinson: Cambridge Pragmatics
Cambridge identifies Levinson’s 1983 text and its deixis, implicature, presupposition, speech-act and conversation scope.
- Kroeger: publisher source chapter on implicature
The publisher’s chapter 8 source distinguishes conversational inference and semantic consequence, including cancellation diagnostics. The lamp packets are original.
These are proposed study activities, not official course assignments or a claim that every book has been read cover to cover.