Relevance Theory (Sperber & Wilson, 1986)

Human cognition is tuned to maximize relevance, and every act of communication carries a built-in guarantee of its own optimal relevance. It recasts Grice's whole cooperative apparatus as a single principle: hearers automatically pursue the interpretation that yields the most reward for the least work.

The trade-off (a relationship, not a numeric formula):

relevance  ∝  cognitive effect / processing effort

• cognitive effect — how much the utterance changes the hearer's model of the world (new inference, strengthened/revised belief). More effect → more relevant.
• processing effort — attention, decoding, and inference cost to extract that effect. More effort → less relevant. Comprehension stops at the first interpretation that satisfies the presumption of optimal relevance: enough effect to be worth the effort, and no needless effort for the effect.

Transfer. This is the exact objective function a reader's attention runs on the feed — and it is the mechanism under processing-fluency and slop-is-a-context-deficit. Slop fails on both terms at once: near-zero cognitive effect (says nothing that revises the reader's model) at nonzero effort, so relevance collapses and they scroll. To be read, maximize the ratio, not either term alone: raise effect (say the non-obvious, load-bearing thing) while cutting effort (front-load the payoff, strip preamble). A dense but well-ordered claim can out-perform an "easy" empty one, because effort is only the denominator.

Kin: grices-maxims · speech-act-theory · models · processing-fluency · slop-is-a-context-deficit