Pragmatics of Persuasion L15 Glazer and Rubinstein (TE 2006) A study in the pragmatics of persuasion : a game theoretic approach. Setup • Today - Sender choses which verified fact to reveal - Sender has limited capacity to verify facts - Receiver choses their interpretation • Optimal interpretation rule depends on S incentives • Pragmatics: A field of linguistics - Interpretation of utterances depends on the context - Cooperative principle (Grice 1989) requires aligned preferences - Noncooperative approach Benz at al. (2006) A persuasion game • Finite state space • Action space • Sender always prefers • Acceptance and rejection region • (Arbitrary) type dependent message structure • Persuasion rule • Rule • Rule is deterministic if is finite if Optimal persuasion rule • For probability of acceptance • Let • Optimal mechanism solves • Relative to Glazer and Rubinstein (2006): - S controls which verified facts are revealed - Random decision rule - Abstract (possibly infinite) type dependent message space Milgrom’s message structure • Assume - Cheap talk with type independent preferences - Trivial optimal rule • Assume - Variant of Milgrom’s persuasion game - In any PBN equilibrium unraveling of information - Trivial optimal persuasion rule: • Problem interesting when high types cannot verify that they are high • Example: Vectoric message structure • Let Example • Acceptance region • Vectoric message structure, capacity two aspects • Rule 1: Accept if any two aspects are verified • Rule 2: Accept if any to neighboring aspects are verified 3 sets of results 1. Randomization is not needed 2. Optimal rule given by a solution to linear optimization problem 3. Credibility (ex post optimality) 4. Side product: which mechanism is better (GR 2004) or (GR 2006)? Lemma 1 L: There exist an optimal persuasion rule Proof: Claim 1: For any there exist finite. that is finite such that Randomization is not needed P1: There exist an optimal persuasion rule that is deterministic. • For any type probability of a mistake is • Implications (vectoric message structure) - Deterministic mechanism (GR 2004) equivalent to deterministic rule (GR 2006) - Optimal mechanism (GR 2004) weakly dominates optimal rule (GR 2006) Proof Let be finite optimal mechanism with the smallest number of noninteger values. Suppose Generalization of • L with Vectoric message structure • For abstract message structure Characterization L: Fix • Let . Sum of errors on any satisfies be a solution to a linear programming problem P: There exists optimal deterministic mechanism P: Any optimal mechanism is credible inducing Conclusions • Ability to verify facts improves information transmission • Skepticism and selective reporting • Unraveling of information from the top (frictionless verifiability) • Frictions: - Uncertainty about verifiability - Costly verifiability - Capacity constraints
© Copyright 2026 Paperzz