pith. machine review for the scientific record. sign in

arxiv: 2604.25175 · v1 · submitted 2026-04-28 · ⚛️ physics.soc-ph

Recognition: unknown

Indirect reciprocity beyond pairwise interactions

Feng Fu, Hongwei Zheng, Junyu Lu, Longzhao Liu, Ming Wei, Shaoting Tang, Xin Wang, Yishen Jiang

Authors on Pith no claims yet

Pith reviewed 2026-05-07 14:38 UTC · model grok-4.3

classification ⚛️ physics.soc-ph
keywords indirect reciprocitygroup cooperationreputation dynamicsbistabilityhysteresismultiplayerleading eight norms
0
0 comments X

The pith

Stable group cooperation requires the rule 'all good, help; one bad, halt' based on collective reputations.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The authors create a general framework for indirect reciprocity when interactions involve groups rather than pairs. They identify that cooperation remains stable only when people cooperate if every group member has a good reputation and stop if any one has a bad reputation. This organizing principle proves both necessary and sufficient and matches known norms for two-person cases. Group judgment leads to bistable dynamics with a tipping point, allowing both cooperation and defection to persist depending on starting conditions. The framework also serves as a test for whether AI systems follow the same logic when given group-level information.

Core claim

Stable group cooperation obeys a simple organizing principle: `all good, help; one bad, halt'. This rule is both necessary and sufficient for cooperation to emerge, and it recovers the classical leading eight norms in the pairwise limit. Group structure fundamentally changes reputation dynamics: unlike pairwise models, which are monostable, multiplayer systems exhibit bistability and hysteresis, with a critical tipping point separating cooperative and defective regimes.

What carries the argument

The organizing principle 'all good, help; one bad, halt', which determines action based on the reputational profile of the entire group rather than individuals separately.

If this is right

  • Cooperation emerges and persists stably in groups when this rule governs decisions.
  • Reputation dynamics in groups are bistable with hysteresis, unlike the monostable pairwise case.
  • A critical tipping point exists that can shift the system between cooperative and defective states.
  • The rule generalizes to recover the leading eight norms when reduced to pairwise interactions.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The bistability suggests that once a group tips into cooperation, it may resist small defects, but crossing the point leads to collapse.
  • This principle offers a benchmark for designing institutions or AI agents that promote collective action in real scenarios like public goods provision.
  • Human experiments varying group reputation information could test for the predicted hysteresis effects.

Load-bearing premise

Moral judgments and cooperation decisions are made based on the reputational profile of the full group, following update rules that render the principle necessary and sufficient.

What would settle it

A demonstration of stable cooperation in a multiplayer setting despite the presence of one individual with bad reputation would falsify the necessity of halting under the 'one bad' condition.

read the original abstract

Cooperation in groups underpins collective responses to challenges from climate governance to public goods provision, yet how moral evaluation sustains it remains poorly understood. Indirect reciprocity -- cooperating to build a good reputation -- is well characterized for pairwise interactions, but real collective action requires individuals to be judged against the reputational profile of an entire group. Here we develop a general framework for multiplayer indirect reciprocity and show that stable group cooperation obeys a simple organizing principle: `all good, help; one bad, halt'. This rule is both necessary and sufficient for cooperation to emerge, and it recovers the classical leading eight norms in the pairwise limit. We further show that group structure fundamentally changes reputation dynamics: unlike pairwise models, which are monostable, multiplayer systems exhibit bistability and hysteresis, with a critical tipping point separating cooperative and defective regimes. Assessment of the latent norms of large language models reveals that they shift toward punitive defection when provided with richer social information, yet fail to follow the full logic of `all good, help; one bad, halt'. Our results establish a unifying principle for reputation-based cooperation in groups and provide a benchmark for evaluating cooperative alignment in artificial intelligence.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 4 minor

Summary. The manuscript develops a general framework for indirect reciprocity in multiplayer interactions, where individuals are evaluated based on the reputational profile of their entire group rather than pairwise encounters. It identifies a simple organizing principle ('all good, help; one bad, halt') that is claimed to be both necessary and sufficient for the emergence and stability of group cooperation. The rule recovers the classical leading eight norms as a special case in the pairwise limit. The analysis further shows that group structure induces bistability and hysteresis in reputation dynamics, with a critical tipping point separating cooperative and defective regimes, in contrast to the monostable behavior of pairwise models. The paper also evaluates latent norms in large language models, finding they shift toward punitive defection with richer social information but do not fully implement the identified rule.

Significance. If the derivations hold, the work offers a unifying principle for reputation-based cooperation beyond the pairwise setting that has dominated indirect reciprocity research. The necessity/sufficiency claim, the exact recovery of the leading eight norms, and the demonstration of bistability/hysteresis constitute substantive advances with implications for modeling collective action problems such as public goods provision and climate governance. The LLM assessment provides a concrete benchmark for evaluating cooperative alignment in AI systems.

minor comments (4)
  1. The abstract and introduction refer to the rule as 'necessary and sufficient,' but the precise stability conditions (e.g., invasion criteria or Lyapunov functions) used to establish necessity should be stated explicitly in the main text rather than deferred to supplementary material.
  2. Figure captions for the bistability diagrams should include the exact parameter values at which the critical tipping point occurs and clarify whether the hysteresis loop is robust to small perturbations in group size or observation accuracy.
  3. The pairwise-limit reduction to the leading eight norms is asserted; a short appendix table mapping the multiplayer rule to each of the eight norms under the appropriate limit would improve traceability.
  4. The LLM evaluation section would benefit from reporting the exact prompt templates and the number of trials per condition to allow replication of the shift toward punitive defection.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive assessment of the manuscript and for recommending minor revision. The provided summary accurately reflects the core contributions, including the general framework for multiplayer indirect reciprocity, the necessity and sufficiency of the 'all good, help; one bad, halt' rule, recovery of the leading eight norms, the emergence of bistability and hysteresis, and the evaluation of latent norms in large language models.

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

The paper develops a general framework for multiplayer indirect reciprocity and derives the 'all good, help; one bad, halt' rule as the unique strategy profile that is necessary and sufficient for stable cooperation under the model's payoff and reputation-update structure. This derivation is presented as emerging from the stability analysis rather than presupposed, with the pairwise limit serving as a direct reduction to recover the leading eight norms. No load-bearing steps reduce by construction to fitted parameters, self-citations, or ansatzes imported from prior author work; the central claims rest on the internal consistency of the group-level reputation aggregation model. The framework is self-contained against its stated assumptions, with no evidence of self-definitional loops or renaming of known results as new derivations.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Only the abstract is available, so specific free parameters, axioms, and invented entities cannot be identified. The framework likely relies on standard evolutionary game theory assumptions about reputation updates and introduces group-level reputational assessment as a core modeling choice.

pith-pipeline@v0.9.0 · 5514 in / 1336 out tokens · 131431 ms · 2026-05-07T14:38:53.353983+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

51 extracted references · 4 canonical work pages

  1. [1]

    Alexander, R.The Biology of Moral Systems (Routledge, 2017)

    2017

  2. [2]

    & Richerson, P

    Boyd, R. & Richerson, P. J. The evolution of indirect reciprocity.Social Networks11, 213–236 (1989)

    1989

  3. [3]

    Nowak, M. A. & Sigmund, K. The dynamics of indirect reciprocity.Journal of Theoretical Biology194, 561–574 (1998)

    1998

  4. [4]

    Nowak, M. A. & Sigmund, K. Evolution of indirect reciprocity.Nature437, 1291–1298 (2005)

    2005

  5. [5]

    A review of theoretical studies on indirect reciprocity.Games11, 27 (2020)

    Okada, I. A review of theoretical studies on indirect reciprocity.Games11, 27 (2020)

    2020

  6. [6]

    & Iwasa, Y

    Ohtsuki, H. & Iwasa, Y. How should we define goodness?—reputation dynamics in indirect reci- procity.Journal of Theoretical Biology231, 107–120 (2004)

    2004

  7. [7]

    & Nowak, M

    Hilbe, C., Schmid, L., Tkadlec, J., Chatterjee, K. & Nowak, M. A. Indirect reciprocity with private, noisy, and incomplete information.Proceedings of the National Academy of Sciences115, 12241– 12246 (2018)

    2018

  8. [8]

    Nature634, 883–889 (2024)

    Michel-Mata, S.et al.The evolution of private reputations in information-abundant landscapes. Nature634, 883–889 (2024)

    2024

  9. [9]

    & Chatterjee, K

    Schmid, L., Ekbatani, F., Hilbe, C. & Chatterjee, K. Quantitative assessment can stabilize indirect reciprocity under imperfect information.Nature Communications14, 2086 (2023). 8

    2086

  10. [10]

    & Nowak, M

    Ohtsuki, H., Iwasa, Y. & Nowak, M. A. Indirect reciprocity provides only a narrow margin of effi- ciency for costly punishment.Nature457, 79–82 (2009)

    2009

  11. [11]

    Ostrom, E.Governing the Commons: The Evolution of Institutions for Collective Action (Cambridge University Press, 1990)

    1990

  12. [12]

    Ledyard, J. O.et al. Public Goods: A Sur- vey of Experimental Research(Division of the Humanities and Social Sciences, California Inst. of Technology, 1994)

    1994

  13. [13]

    & G ¨achter, S

    Fehr, E. & G ¨achter, S. Cooperation and pun- ishment in public goods experiments.American Economic Review90, 980–994 (2000)

    2000

  14. [14]

    G., Dreber, A., Ellingsen, T., Fudenberg, D

    Rand, D. G., Dreber, A., Ellingsen, T., Fudenberg, D. & Nowak, M. A. Positive interactions promote public cooperation.Science325, 1272–1275 (2009)

    2009

  15. [15]

    & Iwasa, Y

    Ohtsuki, H. & Iwasa, Y. The leading eight: social norms that can maintain cooperation by indirect reciprocity.Journal of Theoretical Biology239, 435–444 (2006)

    2006

  16. [16]

    & Fischbacher, U

    Fehr, E. & Fischbacher, U. The nature of human altruism.Nature425, 785–791 (2003)

    2003

  17. [17]

    & Vaish, A

    Tomasello, M. & Vaish, A. Origins of human cooperation and morality.Annual Review of Psychology64, 231–255 (2013)

    2013

  18. [18]

    The tragedy of the commons.Science 162, 1243–1248 (1968)

    Hardin, G. The tragedy of the commons.Science 162, 1243–1248 (1968)

    1968

  19. [19]

    Trivers, R. L. The evolution of reciprocal altru- ism.The Quarterly Review of Biology46, 35–57 (1971)

    1971

  20. [20]

    Nowak, M. A. Five rules for the evolution of cooperation.Science314, 1560–1563 (2006)

    2006

  21. [21]

    & Small, D

    Silver, I. & Small, D. A. Put your mouth where your money is: A field experiment encouraging donors to share about charity.Marketing Science 43, 392–406 (2024)

    2024

  22. [22]

    Yoeli, E., Hoffman, M., Rand, D. G. & Nowak, M. A. Powering up with indirect reciprocity in a large-scale field experiment.Proceedings of the National Academy of Sciences110, 10424–10429 (2013)

    2013

  23. [23]

    Dawson, D.et al.The role of collegiality in academic review, promotion, and tenure.PloS one17, e0265506 (2022)

    2022

  24. [24]

    & Sigmund, K

    Brandt, H. & Sigmund, K. Indirect reciprocity, image scoring, and moral hazard.Proceedings of the National Academy of Sciences102, 2666– 2670 (2005)

    2005

  25. [25]

    & Boyd, R

    Panchanathan, K. & Boyd, R. A tale of two defectors: the importance of standing for evolu- tion of indirect reciprocity.Journal of Theoretical Biology224, 115–126 (2003)

    2003

  26. [26]

    M.et al.The psychological foun- dations of reputation-based cooperation.Philo- sophical Transactions of the Royal Society B376, 20200287 (2021)

    Manrique, H. M.et al.The psychological foun- dations of reputation-based cooperation.Philo- sophical Transactions of the Royal Society B376, 20200287 (2021)

    2021

  27. [27]

    Nowak, M. A. & Sigmund, K. Evolution of indi- rect reciprocity by image scoring.Nature393, 573–577 (1998)

    1998

  28. [28]

    & Milinski, M

    Wedekind, C. & Milinski, M. Cooperation through image scoring in humans.Science288, 850–852 (2000)

    2000

  29. [29]

    Milinski, M., Semmann, D., Bakker, T. C. & Krambeck, H.-J. Cooperation through indirect reciprocity: image scoring or standing strategy? Proceedings of the Royal Society of London. Series B: Biological Sciences268, 2495–2501 (2001)

    2001

  30. [30]

    P., Santos, F

    Santos, F. P., Santos, F. C. & Pacheco, J. M. Social norm complexity and past reputations in the evo- lution of cooperation.Nature555, 242–245 (2018)

    2018

  31. [31]

    Raihani, N. J. & Bshary, R. The reputation of punishers.Trends in Ecology & Evolution30, 98–103 (2015)

    2015

  32. [32]

    M., Santos, F

    Pacheco, J. M., Santos, F. C. & Chalub, F. A. C. Stern-judging: A simple, successful norm which promotes cooperation under indirect reciprocity. PLoS Computational Biology2, e178 (2006)

    2006

  33. [33]

    & Gintis, H.A Cooperative Species: Human Reciprocity and Its Evolution(Princeton University Press, Princeton, NJ, 2011)

    Bowles, S. & Gintis, H.A Cooperative Species: Human Reciprocity and Its Evolution(Princeton University Press, Princeton, NJ, 2011)

    2011

  34. [34]

    Boehm, C.Moral Origins: The Evolution of Virtue, Altruism, and Shame(Soft Skull Press, 2012). 9

    2012

  35. [35]

    & Akiyama, E

    Suzuki, S. & Akiyama, E. Three-person game facilitates indirect reciprocity under image scor- ing.Journal of Theoretical Biology249, 93–100 (2007)

    2007

  36. [36]

    Wei, M.et al.Indirect reciprocity in the public goods game with collective reputations.Journal of the Royal Society Interface22(2025)

    2025

  37. [37]

    Dafoe, A.et al.Cooperative ai: machines must learn to find common ground.Nature593, 33–36 (2021)

    2021

  38. [38]

    Jin, Z.et al.When to make exceptions: Explor- ing language models as accounts of human moral judgment.Advances in Neural Information Processing Systems35, 28458–28473 (2022)

    2022

  39. [39]

    & Lieder, F

    Cheung, V., Maier, M. & Lieder, F. Large lan- guage models show amplified cognitive biases in moral decision-making.Proceedings of the National Academy of Sciences122, e2412015122 (2025)

    2025

  40. [40]

    Wang, Y.et al.Aligning large language mod- els with human: A survey.arXiv preprint arXiv:2307.12966(2023)

    work page arXiv 2023

  41. [41]

    Pal, S.et al.Strategies of cooperation and defec- tion in five large language models.arXiv preprint arXiv:2601.09849(2026)

    work page arXiv 2026

  42. [42]

    Akata, E.et al.Playing repeated games with large language models.Nature Human Behaviour 1–11 (2025)

    2025

  43. [43]

    & Christakis, N

    Fu, F., Chen, X. & Christakis, N. A. On the opti- mal integration of intelligent agents into network systems to steer cooperation.Proceedings of the National Academy of Sciences123, e2537939123 (2026)

    2026

  44. [44]

    & Hughes, E

    Vallinder, A. & Hughes, E. Cultural evolution of cooperation among llm agents.arXiv preprint arXiv:2412.10270(2024)

    work page arXiv 2024

  45. [45]

    S., Samson, L., Ghebreab, S

    Pires, A. S., Samson, L., Ghebreab, S. & San- tos, F. P. How large language models judge and influence human cooperation.arXiv preprint arXiv:2507.00088(2025)

    work page arXiv 2025

  46. [46]

    Fishman, M. A. Indirect reciprocity among imperfect individuals.Journal of Theoretical Biology225, 285–292 (2003)

    2003

  47. [47]

    & Mashima, R

    Takahashi, N. & Mashima, R. The impor- tance of subjectivity in perceptual errors on the emergence of indirect reciprocity.Journal of Theoretical Biology243, 418–436 (2006)

    2006

  48. [48]

    L., Kessinger, T

    Radzvilavicius, A. L., Kessinger, T. A. & Plotkin, J. B. Adherence to public institutions that foster cooperation.Nature Communications12, 3567 (2021)

    2021

  49. [49]

    Philosophical Transactions of the Royal Society B: Biological Sciences376(2021)

    Wu, J.et al.Honesty and dishonesty in gos- sip strategies: a fitness interdependence analysis. Philosophical Transactions of the Royal Society B: Biological Sciences376(2021)

    2021

  50. [50]

    & Nakai, Y

    Okada, I., Sasaki, T. & Nakai, Y. A solu- tion for private assessment in indirect reciprocity using solitary observation.Journal of Theoretical Biology455, 7–15 (2018). 10 Methods In this section, we summarize our modeling framework and the methods used to obtain all the results, includ- ing both theoretical analyses and experiments with LLMs. Further det...

    2018

  51. [51]

    Moreover, according to Eq

    = 1. Moreover, according to Eq. (13), a(d, p)pair can resist invasion by ALLD mutants if R >1 +ζ 3,(14) where ζ3 = 2θ(p|2pALLD) 3θ(p|3p)−2θ(p|2pALLD) .(15) When errors are present, we would like their impact on cooperation to be kept as small as possible. Hence- forth, a successful ESS pair should both maintain a high level of cooperation in the presence ...

    2025