Implicature systems at C2

The frontier between C1 and C2 is not vocabulary or grammar — it is implicature. A C1 speaker understands what is said; a C2 speaker understands what is meant given what is said. The gap between the two is the territory Paul Grice (1975) opened up when he proposed that conversation is governed by a Cooperative Principle and by four maxims — Quality, Quantity, Relation, Manner — and that listeners routinely compute meanings that go beyond the literal by assuming the speaker is cooperating even when she appears not to be.

Russian-speaking C1 graduates already know this in their L1 — Russian is, if anything, a more implicature-heavy language than English in some domains. The C2 task is not learning the phenomenon but learning which implicatures English conventionally licenses and which it does not, which it cancels with which markers, and which patterns Russian and English diverge on enough to cause systematic misunderstanding.

This lesson lays out the Gricean framework, then walks the major implicature types — generalized vs particularized, scalar, conversational vs conventional — and closes with the diagnostic test (cancellability) that distinguishes implicature from entailment.

The Cooperative Principle and the four maxims

Grice’s Cooperative Principle: Make your conversational contribution such as is required, at the stage at which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged.

The four maxims unpack this:

• Quality — Do not say what you believe to be false; do not say that for which you lack evidence.
• Quantity — Make your contribution as informative as required; not more informative than required.
• Relation — Be relevant.
• Manner — Be perspicuous: avoid obscurity, avoid ambiguity, be brief, be orderly.

A speaker can observe a maxim, violate it (quietly break it, often misleading), opt out (signal she is not playing — no comment), or flout it (visibly break it in a way the hearer is meant to notice). Flouting is the engine of conversational implicature: the hearer notices the apparent violation, assumes cooperation, and infers what the speaker must have meant.

Example — flouting Relation

A: Is Sarah still working at the firm?

B: Her car is in the lot every morning.

B’s answer is literally about a car, not about Sarah’s employment. Flouting Relation, B forces A to compute: if B is cooperating, this comment must be relevant; the relevance must be that car presence implies Sarah-presence implies still-employed. The implicature yes, she’s still there is computed.

A Russian speaker may take B’s answer literally and be confused. A C2 English speaker computes the implicature automatically.

Why flouting works — the inference schema

The hearer’s reasoning, formalized:

1. Speaker said P.
2. P appears to violate maxim M in this context.
3. I have no reason to think Speaker has opted out of cooperation.
4. Therefore Speaker must intend Q, where Q is consistent with cooperation and explains the apparent violation.
5. Speaker knows I will reason this way and made the contribution accordingly.

The whole schema runs in under a second in fluent listeners. The C2 task is making it run in real time in a second language — and especially recognizing which apparent violations trigger which inferences, since languages calibrate this differently.

Conversational vs conventional implicature

Conversational implicatures arise from the maxims and context. They are cancellable (you can add but… and undo them) and non-detachable (paraphrasing the sentence preserves them).

Conventional implicatures are attached to particular lexical items — but, even, yet, therefore, still — and are not cancellable and detachable.

Conventional examples

• She is poor but honest. — The but conventionally implicates a contrast (poor ≠ honest). You cannot say She is poor but honest, and there is no contrast intended without sounding incoherent.
• Even John passed. — Even conventionally implicates that John was a less-likely passer than others. Removing even removes the implicature.
• She managed to finish. — Manage conventionally implicates difficulty. She finished makes no such implicature.

C2 speakers control these because they are lexicalized — once you know the word, you carry the implicature.

Generalized vs particularized conversational implicature

Generalized conversational implicatures (GCIs) arise without special context. They are default inferences attached to particular forms.

Particularized conversational implicatures (PCIs) arise from specific context.

Generalized examples

• I broke a finger yesterday. → GCI: one of my own fingers (not someone else’s). Saying a finger rather than my finger normally still implicates own-finger ownership by Quantity.
• I walked into a house. → GCI: not my own house (otherwise I would have said my house). By Quantity, the indefinite signals non-possession.
• Some of the students passed. → GCI: not all of them passed (the scalar implicature, treated in detail below).

Particularized examples — context-driven

A: Did you finish the report?

B: I had a meeting at three.

PCI: no, I didn’t finish. B’s answer is non-relevant on its face; the hearer infers the meeting-as-obstacle explanation only because of the question that preceded it. Strip the question and I had a meeting at three generates no such inference.

Russian and English largely agree on PCIs — context-driven reasoning is universal. They diverge on GCIs, where convention does the work.

Scalar implicature — some, most, all and friends

Scalar implicature (Horn 1972) is the most-cited type of generalized implicature. The intuition: linguistic items form scales ordered by informativeness, and using a weaker item implicates that the stronger one does not hold.

Classic scales:

Asserting the lower scalar item implicates that the higher item does not apply (by Quantity: if you knew the stronger, you would have said it).

Examples

• Some of the kids passed. → implicates not all of them passed.
• It’s possible the deal will close. → implicates it’s not certain.
• She has three kids. → implicates exactly three (not four or more) in default context.
• You can have soup or salad. → implicates not both.

Cancellation

Scalar implicatures are cancellable — you can add a clause and undo:

• Some of the kids passed — in fact, all of them did.
• I have three kids — at least, three; I might be forgetting one.

The cancellability is the diagnostic test that distinguishes implicature from entailment. Some does not entail not all; it merely implicates it.

Russian-speaker trap

Russian некоторые implicates not all less strongly than English some. The English implicature is robust enough that asserting some students passed when in fact all did is misleading. A Russian speaker may produce some where all is true and be heard as misrepresenting.

The cancellability test

The diagnostic for any implicature:

1. Take the sentence.
2. Add but in fact / actually / as it turned out [contradiction].
3. If the result is coherent, you had an implicature.
4. If the result is contradictory, you had an entailment.

Examples:

• Some students passed — in fact, all of them did. Coherent → some implicates not all; entails only at least one.
• Three students passed — in fact, only two did. Contradictory → three entails at least three; the at least reading is the implicature, the not fewer than three is the entailment.
• She managed to finish — but it was easy. Awkward → manage conventionally implicates difficulty; conventional implicatures resist cancellation.

C2 reading test: when you encounter an ambiguous claim, run the test. The result tells you whether you can cancel the inference or whether you have committed to it.

Flouting in real US speech

Flouting is everywhere in casual American English. Recognizing flouts is the C2 listening skill.

Flouting Quantity — saying too little or too much

Co-worker: How was the meeting with Dave?

You: Dave is… Dave.

Quantity-flout: tautology gives zero information on its face. The hearer infers Dave was Dave again — frustrating, predictable, with affective content the bare description cannot carry.

Flouting Quality — irony

(After a presentation goes badly)

Manager: Well, that was a triumph.

Quality-flout: the speaker doesn’t believe it. The hearer computes the opposite — that was a disaster — as the intended message. Irony is the canonical Quality-flout.

Flouting Relation — apparent non-sequitur

A: Did you tell her about the layoffs?

B: Oh, look, the coffee’s ready.

Relation-flout: B’s topic shift implicates I’m not going to answer / I haven’t done it / let’s drop it.

Flouting Manner — deliberate obscurity

Parent (about child, child present): We had a… situation… at school today.

Manner-flout: the deliberate vagueness implicates that the speaker doesn’t want to be specific in front of the child.

Real US example — flouting in a Veep exchange

Selina: Was the speech good?

Amy: It was… received.

Quantity-flout (saying less than is informative): received tells Selina almost nothing about quality, which by the schema implicates the speech was bad. If Amy could have honestly said good or great, Quantity would require her to. Saying received is the lowest informative truth she can offer; the implicature not good is computed instantly.

Real US example — talk radio, flouting Quality

Host (deadpan, after a politician’s evasive answer): Well, that was a master class in saying nothing.

Quality-flout via irony: master class literally praises, prosody and context cancel the praise. Implicature: that was a particularly poor evasion. The mock-praise is more cutting than direct criticism.

Manner implicatures — the marked-vs-unmarked test

Manner generates implicatures when speakers choose a marked (longer, less common, more complex) form over an unmarked one. The principle (Levinson 2000): unmarked form for stereotypical situation; marked form for non-stereotypical.

Examples

• She made the car stop. (marked) vs She stopped the car. (unmarked). The marked form implicates an unusual stopping method — perhaps by jumping in front of it.
• He caused the door to open. (marked) vs He opened the door. (unmarked). Marked implicates indirect causation — maybe a remote-control or a triggered mechanism.
• Bill is not unattractive. (marked double negative) vs Bill is attractive. (unmarked). Marked implicates attractive but not strikingly so — the double-negative carries a hedge of doubt.

Russian-trained speakers sometimes pick the marked form (longer phrase, more morphology) because it feels more careful. In English the marked form is itself an implicature trigger. He made the report be written implicates something unusual about the writing; he wrote the report says he wrote it.

Real US example — Atlantic essay

The president didn’t say the deal was dead. He said the conditions for its survival were no longer present.

Marked phrasing of no longer present over unmarked gone or missing. The marker implicates a careful, perhaps deliberate, choice of words by the president — and signals to the reader that the columnist noticed the choice. C2 reading-for-tone depends on catching the marked-form-implicature.

Implicatures English licenses by default

C2 speakers should internalize these GCIs as English-conventional, not universal:

• Indirect requests: Can you pass the salt? implicates a request, not a question about ability. Russian Можешь передать соль? works similarly, but English is stricter — answering Yes without passing the salt reads as deliberate flout.
• Numerals: I have three kids defaults to exactly three. Russian у меня трое детей is the same; no L1 trap here, but worth verifying in test items.
• Connectives: and in narrative implicates and then (temporal order). He went home and had a beer implies sequence; reversing reads as different.
• Negation under modal: You shouldn’t have (after a gift) implicates thanks for the gift. Russian не стоило maps cleanly; safe.
• Indirect refusals: That’s an interesting idea in business contexts implicates no. Russian directness can miss this — интересная идея in Russian often means I’ll consider it.

Real US example — indirect refusal in business context

Hiring manager (email): Thanks so much for the conversation last week. We were really impressed with your background. After much deliberation, we’ve decided to move forward with another candidate whose experience more closely matches the role’s current needs. We hope you’ll keep us in mind for future opportunities.

Surface meaning: polite rejection. C2 implicatures: (1) the much deliberation is conventional padding implicating we considered you seriously; (2) more closely matches the role’s current needs implicates you weren’t a poor fit; we found a better one — preserving the rejected candidate’s face; (3) keep us in mind for future opportunities is conventional — implicates almost nothing about actual future openings. Russian speakers sometimes over-interpret the closing as a real invitation; in American business it is a face-preserving formula, not a substantive offer.

Implicatures English does NOT license that Russian might

• Bare claim = strong commitment: in Russian, a bare assertion can be a tentative guess that the hearer is meant to soften. In English, a bare assertion commits the speaker. Russian speakers expecting a Russian-style softening default produce statements that read as more confident than intended.
• Silence = continued floor: in Russian, the silent listener may still be processing. In English, silence over 2-3 seconds implicates I’m done; over to you or I disagree but won’t say. The implicature is wrong but the marker is missing.
• Negative question = sympathetic invitation: in Russian, Не хочешь ли пойти? (negative form) is warmly inviting. In English, Don’t you want to go? implicates I think you should want to but you don’t. The negative form carries presupposition.

Presupposition — what’s taken for granted

Presupposition is the close cousin of implicature: it is what a sentence takes for granted regardless of whether the sentence is true or false. Presuppositions survive negation; implicatures often don’t.

Examples — presupposition triggers

• John regrets selling his car. — presupposes John sold his car. Negate to John doesn’t regret selling his car — still presupposes the sale.
• The king of France is bald. — presupposes there is a king of France. Both the negated and non-negated sentence carry the presupposition.
• When did you stop beating your wife? — presupposes you used to beat your wife. The classic loaded-question presupposition trap.
• Even John passed. — presupposes John was less expected to pass than relevant others.

C2 listening: presupposition triggers in questions and assertions reveal what the speaker takes as common ground. Russian and English presupposition triggers overlap heavily; the trap is when a speaker introduces a presupposition the hearer has not accepted (called accommodation in pragmatics — the hearer accommodates by silently accepting). American argumentation often relies on presupposition smuggling — putting controversial claims as presupposed rather than asserted. C2 readers spot the move.

Real US example — political ad

When the senator finally addresses why the budget collapsed under his leadership, we’ll listen.

Two presuppositions smuggled: (1) the budget collapsed (factive presupposition from why) and (2) it was under his leadership (presupposed). The sentence is grammatically about listening, but the listener has accommodated two assertions without explicit assent.

AmE vs BrE in implicature

• Indirect refusals: BrE goes further than AmE. That’s quite good in BrE often implicates mediocre; in AmE quite is closer to its literal meaning. I’ll bear it in mind in BrE implicates I will not act on this; in AmE it can be sincere.
• Understatement: BrE conventionally understates (a bit of a problem = a major crisis); AmE understates less and signals scale more directly.
• Sarcasm: AmE relies on prosody and context; BrE relies on flat-affect deadpan, which Americans sometimes miss.
• Yes/no answers: BrE Yes, no, definitely is a real agreement marker; AmE Yeah, no is a turn-opener with different function. Don’t confuse them.

Functional view — the implicature toolkit

Real-time implicature reading — three drills

The C2 implicature skill is fast, not slow. Three drills to train it:

Drill 1: scalar audit

For every quantifier or modal you read or hear, ask: what stronger item could the speaker have used? Why didn’t she? Train this until it is automatic. Native speakers do it under 100 milliseconds.

Drill 2: cancellation test

For every claim you encounter, mentally append in fact, [opposite]. If the result is coherent, the claim was an implicature, not an entailment. The test sharpens your sense of what was actually committed to.

Drill 3: marked-form audit

For every longer-than-necessary phrasing — made the door open over opened the door, not unattractive over attractive, it would not be incorrect to say over one could say — ask why. The marked form is the implicature trigger; the answer tells you the speaker’s stance.

Common Russian-speaker mistakes

1. Treating some as scalar-free: Some of the team passed the review when in fact all did. The English implicature is robust; this produces misleading reports. Use all when all is true.
2. Missing irony marked by prosody: Quality-flouts in AmE rely on prosody (falling-rising on the stressed word) more than on lexical irony markers. Russian speakers trained on text miss spoken irony.
3. Taking that’s interesting as endorsement: in US business contexts that’s an interesting point often implicates no via Quantity (if she agreed, she’d say so). Read the move, not the words.
4. Failing to cancel implicatures when needed: if you mean all, say all. If you mean some, possibly all, mark it — some, in fact maybe all. Bare some commits you to not all.
5. Producing the negative-question invitation: Don’t you want some coffee? sounds reproachful in AmE. Use Would you like some coffee? or Want some coffee? in casual register.
6. Bare claims with no Quality-hedge in low-evidence territory: They’re going to fire him (Russian-direct) when you only suspect → I’d imagine / I suspect / from what I’ve heard, they’re going to fire him. Bare claim commits you to Quality you may not have.
7. Reading I’ll think about it as deliberation: in AmE business it often implicates no, but politely. In Russian context я подумаю can be sincere. The English form is more conventionalized toward refusal.

Summary

• Grice’s Cooperative Principle and four maxims (Quality, Quantity, Relation, Manner) generate conversational implicatures via flouting.
• Conversational implicatures are cancellable and non-detachable; conventional implicatures (carried by but, even, yet, manage) are not.
• Scalar implicature is the C2 workhorse: some implicates not all; weaker scalar items implicate the negation of stronger ones.
• The cancellability test (X — in fact, [contradiction]) distinguishes implicature from entailment.
• AmE and BrE diverge in indirect refusal strength: AmE understates less than BrE; bear it in mind is conventional refusal in BrE but can be sincere in AmE.
• The Russian-speaker frontier is mostly here: bare claims that should be hedged, missed indirect refusals, missed scalar implicatures, missed Quality-flouts (irony).

Next lesson: Politeness theory applied — Brown & Levinson on face, FTAs, on-record vs off-record, and the US English politeness system in detail.