Network reciprocity turns cheap talk into a force for cooperation
Zhao Song, Chen Shen, The Anh Han · 2025
Read Network reciprocity turns cheap talk into a force for cooperation by Zhao Song, Chen Shen, The Anh Han (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-08876.txt.
Introduction
Zhao Song, Chen Shen and The Anh Han model how empty promises can still grow cooperation when people interact on a network in 2025. Players may signal that they intend to cooperate before a donation game, but the signal costs nothing and binds no one.
In mixed, anonymous populations the talk changes nothing, matching experiments. On structured networks, though, cooperation emerges. Conditional players, who answer signals but pay a small thinking cost, act as catalysts: shifting patterns of dominance among strategies shelter the unconditional cooperators. The authors offer this as an evolutionary explanation for why cheap talk matters in real communities, where relationships repeat and reputations linger.
Catalog record
| CREATOR | Zhao Song, Chen Shen, The Anh Han |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-08876
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Network reciprocity turns cheap talk into a force for cooperation Zhao Song, Chen Shen, The Anh Han Abstract Non-binding communication is common in daily life and crucial for fostering cooperation, even though it has no direct payoff consequences. However, despite robust empirical evidence, its evolutionary basis remains poorly understood. Here, we develop a game-theoretic model in which individuals can signal an intention to cooperate before playing a Donation game. Strategies differ in how they respond to these signals, ranging from unconditional to conditional types, with the latter incurring a cognitive cost for deliberation. Through evolutionary analysis, we show that non-binding communication alone cannot sustain cooperation in well-mixed, anonymous populations, consistent with empirical observations. In contrast, structured populations support the emergence of cooperation, with conditional cooperators acting as catalysts that protect unconditional cooperators through context-dependent patterns of cyclic dominance. These findings offer an evolutionary explanation for how non-binding communication promotes cooperation and provide a modelling framework for exploring its effects in diverse social settings. 1 Introduction Humans possess unique communication abilities that enable them to solve complex social problems through information exchange, from negotiating climate risks between nations to coordinating public health responses during epidemics, to resolving everyday conflicts through conversation [ 1 , 2 , 3 ] . This capacity has naturally drawn attention in the study of cooperative behaviour, a longstanding puzzle across disciplines [ 4 , 5 ] . Cooperation involves paying a personal cost to benefit others, but is continually threatened by free riders who exploit others’ contributions [ 6 , 7 ] . The Donation game captures this dilemma: individuals choose whether to cooperate, incurring a cost to help a partner, or to defect and maximise their own payoff [ 8 ] . Although mutual cooperation produces the highest collective benefit, defection is the individually optimal strategy, often leading to the breakdown of cooperation. Economic experiments show that allowing participants to communicate before making decisions increases cooperation, even when communication is non-binding and has no direct payoff consequences [ 9 , 10 , 11 ] . This effect relies on trust and tends to appear transiently in one-shot, anonymous interactions in well-mixed populations [ 12 , 13 , 14 ] , often fading as interactions proceed. By contrast, when games are repeated among the same individuals, the positive impact of communication on cooperation tends to persist [ 15 , 16 , 17 ] . Meta-analytic studies confirm this pattern across different communication types and timing relative to the game [ 18 , 19 ] . Recent work extends these findings to noisy environments, showing that cheap talk promotes cooperation only when players have self-interested motivations to cooperate in repeated interactions [ 20 , 21 ] . Despite the empirical robustness of these results, it remains theoretically unclear why costless, non-enforceable communication should promote cooperation from an evolutionary perspective [ 22 , 23 , 24 ] . Evolutionary explanations for non-binding communication remain limited, as standard models typically focus on how cooperation evolves among payoff-driven individuals [ 25 , 26 ] . Since cheap talk does not alter the game’s payoff structure, it has no effect on evolutionary outcomes under conventional assumptions. As a result, most theoretical work has focused on binding communication, where signals are tied to material consequences such as costs for deception or rewards for honesty [ 27 , 28 , 29 ] . These models aim to identify conditions under which cooperation can be evolutionarily stable across diverse contexts [ 30 , 31 , 32 ] . Nevertheless, individuals may differ in how they process and respond to signals, even when the signals carry no direct payoff consequences [ 33 ] . Such variation can stem from underlying cognitive processes, as described by dual-process theories, which distinguish between fast, intuitive strategies and slower, deliberative ones [ 34 , 35 ] . While deliberation can enable more strategic responses, it also imposes cognitive costs. These costs create fitness trade-offs, opening the possibility of modelling responses to non-binding communication as heritable traits subject to evolutionary selection. To examine this possibility, we develop an evolutionary model that explores the role of non-binding communication in promoting cooperation among self-interested individuals, inspired by dual-process theories of cognition [ 34 ] . The game unfolds in two stages: in the first, players choose whether to signal an intention to cooperate; in the second, they choose to cooperate or defect. We consider four strategies. The two intuitive types are unconditional cooperators ( UC ), who always signal and cooperate, and unconditional defectors ( UD ), who never signal and always defect. The two deliberative types, which incur a cognitive cost, are conditional cooperators ( CC ), who cooperate only if their opponent also signals, and strategic defectors ( CD ), who signal but always defect. Through evolutionary simulations, we confirm that in well-mixed populations, non-binding communication alone cannot sustain cooperation in one-shot, anonymous interactions. In contrast, in structured populations, cheap talk promotes cooperation through intricate dynamic pathways shaped by the cost of social reasoning. When this cost falls within an intermediate range, cooperation peaks, driven by context-dependent patterns of cyclic dominance. Outside this regime, cooperation may persist through baseline network reciprocity, but otherwise collapses. These results provide an evolutionary explanation for how costless communication can support cooperation in structured populations and offer a general framework for studying cheap talk in diverse social contexts. 2 Models and Methods 2.1 Two-stage game with cheap talk We formulate cheap talk within a two-stage game framework, integrating pre-game intention-signalling and in-game decision-making. In the pre-game stage, players simultaneously decide to signal a cooperative intention ( S ) or remain silent ( N ). This communication is “cheap talk”, as the signal itself carries no direct cost and is non-enforceable. After observing each other’s signals, players independently choose to either cooperate ( C ) or defect ( D ). These decisions take place in the context of a social dilemma, which we introduce via the classic Donation game. In the Donation game, players choosing C incur a cost c to provide a benefit b ( b>c ) to the co-player, while players choosing D incur no cost and provide no benefit. Thus, mutual cooperation generates a reward R=b-c , mutual defection incurs a punishment P=0 for each player, and unilateral actions lead to an asymmetry: a cooperator receives a sucker’s payoff S=-c , while a defector receives a temptation payoff T=b . After rescaling and substituting, the payoffs become R=1 , S=-r , T=1+r , and P=0 , a special case of the Prisoner’s Dilemma game [ 36 ] . Here r>0 quantifies the dilemma strength [ 37 ] . A player’s strategy in the game is determined by three components: their action in the first stage ( S , N ), their response if the co-player signals ( C , D ), and their response if the co-player remains silent ( C , D ). This results in eight possible strategies summarized in Table A1 . To reduce complexity, our model focuses on four archetypal strategies, reflecting distinct cognitive processes: intuitive and deliberative . • UC (unconditional cooperation), where players always signal cooperative intention in the pre-game stage and always cooperate in the game,