Cheap Talk¶
Costless, non-binding communication whose credibility rests entirely on incentive alignment: as the gap between sender and receiver interests widens, the information a costless message can carry shrinks through ever-coarser partitions down to meaningless babble.
Core Idea¶
Cheap talk is game-theoretic communication in which messages are costless, non-binding, and carry no direct payoff — the sender can say anything without penalty or commitment. Crawford and Sobel (1982) show its behaviour turns on a bias parameter b, the gap between the sender's and receiver's preferred actions: at b=0 full transmission is credible, at large b only babbling survives, and intermediate b yields a finite interval partition that coarsens as b grows. It is the natural foil for costly signaling.
How would you explain it like I'm…
Free Words
Talk That Costs Nothing
Costless Strategic Communication
Scope of Application¶
Cheap talk lives across the signaling subfields of game theory and disciplines modelling strategic interaction among agents with possibly-misaligned interests; its reach is co-extensive with strategic communication, since it bites only where a sender chooses what to say against a divergent interest.
- Game theory and mechanism design — the home turf; the cheap-talk-versus-costly-signaling distinction and persuasion games.
- Political economy — campaign promises and policy announcements, credible only as far as interests align.
- Corporate finance and disclosure — management guidance whose content tracks the management-investor gap.
- Organizational behavior — unaccountable peer recommendations and "expert" opinions.
- Multi-agent AI and protocol design — messages between agents lacking a commitment device.
Clarity¶
Naming cheap talk separates two sources of message credibility intuition conflates: credibility from a sender bearing a cost, and credibility from nothing but aligned interests. Isolating the costless pole reframes the credibility problem as structural, not psychological — not whether the speaker is sincere but whether the incentive geometry lets the words carry information at all. It sharpens informativeness against honesty-as-character: bandwidth is capped by the incentive gap alone.
Manages Complexity¶
"When should I believe what someone says?" untamed has as many answers as speakers. Cheap talk compresses the space of costless messages onto one scalar — the bias parameter b — from which the whole regime follows: full transmission, babbling, or a partition coarsening with the gap. A clean two-parameter front end sorts any communication first: is anything costly or binding the speaker, and how large is the incentive gap? The analyst reads believable bandwidth off that small structure instead of judging each utterance.
Abstract Reasoning¶
The concept licenses diagnostic reasoning — classifying a message as cheap talk or not, then reading its credible bandwidth off the incentive gap. It supports interventionist reasoning that widens a channel by aligning interests or adding a cost or commitment, boundary-drawing that separates cheap talk from costly signaling and from Bayesian persuasion, and predictive reasoning that a misaligned speaker delivers coarse or no information while a stake widens the channel by a forecastable amount.
Knowledge Transfer¶
Within game theory and strategic communication cheap talk transfers intact as mechanism wherever its preconditions hold — a privately-informed sender, an acting receiver, a costless channel, and possibly-divergent payoffs — across political economy, corporate finance, organizational behavior, and multi-agent AI, each a literal instance of the same game. Beyond agent-with-interests settings the precondition fails outright. Loose uses ("just cheap talk") are analogy. The general credibility-from-alignment lesson travels via Signaling, incentive_alignment, and Credible Commitment.
Relationships to Other Abstractions¶
Current abstraction Cheap Talk Domain-specific
Parents (1) — outgoing typed relations
-
Cheap Talk includes Bayesian Nash Equilibrium Domain-specific
Cheap Talk contains a Bayesian equilibrium linking private sender states, message partitions, receiver posteriors, and receiver best responses.
Hierarchy paths (11) — routes to 8 parentless roots
- Cheap Talk → includes Bayesian Nash Equilibrium → is a kind of Nash Equilibrium → is a kind of, typical Equilibrium → is a kind of, typical Fixed Point
- Cheap Talk → includes Bayesian Nash Equilibrium → presupposes Information Asymmetry → is a kind of Asymmetry
- Cheap Talk → includes Bayesian Nash Equilibrium → is a kind of Nash Equilibrium → is a kind of Fixed Point
- Cheap Talk → includes Bayesian Nash Equilibrium → includes Bayesian Updating → is a kind of Inductive Reasoning
- Cheap Talk → includes Bayesian Nash Equilibrium → is a kind of Nash Equilibrium → includes Game-Theoretic Strategy → is a decomposition of Function (Mapping)
- Cheap Talk → includes Bayesian Nash Equilibrium → presupposes Common Knowledge → includes Hierarchy → is a kind of Network
- Cheap Talk → includes Bayesian Nash Equilibrium → presupposes Common Knowledge → includes Hierarchy → is a kind of Order → presupposes Comparison
- Cheap Talk → includes Bayesian Nash Equilibrium → presupposes Common Knowledge → includes Hierarchy → is a kind of Order → presupposes Relation
- Cheap Talk → includes Bayesian Nash Equilibrium → includes Bayesian Updating → presupposes Probability → is a kind of Measure → presupposes Set and Membership
- Cheap Talk → includes Bayesian Nash Equilibrium → presupposes Common Knowledge → includes Hierarchy → is a kind of Order → presupposes Set and Membership
- Cheap Talk → includes Bayesian Nash Equilibrium → includes Bayesian Updating → presupposes, typical Conditional Probability → is a kind of Probability → is a kind of Measure → presupposes Set and Membership
Neighborhood in Abstraction Space¶
Cheap Talk sits in a moderately populated region (51st percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.
Family — Persuasion & Interpersonal Communication (17 abstractions)
Nearest neighbors
- Bayesian Persuasion — 0.90
- Global Games — 0.86
- Data Extraction Through Prompting — 0.85
- Monty Hall problem — 0.85
- Ostrich Effect — 0.84
Computed from structural-signature embeddings · 2026-10-25