PGG

A recent letter to Nature looking at communal activities and whether/how/if they emerge from individual dynamics clued me in to some of the ideas in current research on this topic.

The model these authors use (and appears to be standard) is the Public Goods Game (PGG). In this, a group of actors each chooses to cooperate at some cost 'c' and the benefit of the cooperation is distributed across the whole group, both cooperators and defectors. Defecting is still the dominant strategy because you gain the benefit without paying the cost. The question is, how do you get a group to cooperate under these conditions?

One idea is that there are 'punishers' who identify and punish the defectors, making defecting no longer a viable strategy. The authors are tackling the question of how you get punishers into a community and come up with some conditions under which the most common stable equilibrium is a mix of cooperators and punishers.

Some points of basic difference in thinking:
* The PGG game is a nice extension to communal activities. And my idea of having enforced social norms seems to imply the existence of some sort of enforcer (punisher). How social norm enforcers get into the system seems more basic than my thinking.
* But it doesn't seem to consider the idea of competing communities. Those would seem to enforce a certain kind of punishment as well. If your community doesn't cooperate and gain the benefits, a neighboring community (tribe) will out-compete you.
* The PGG game also implicitly leaves the D-D cell as a very bad thing. I think this is the typical assessment, but it points out to me that maybe it's not that simple. In these games Defect is the opposite of Cooperate. But Compete is also the opposite the Cooperate. It's not synonymous with Defect, but it overlaps a lot. And groups who compete get global social benefits as well under certain conditions.
* Which leads me to think that Symmetry is maybe the key idea for me. You don't just need C-C outcomes, but the idea is that you want C-C and D-D outcomes and avoid the C-D cells. If you can make the case for this, then homogeniety of decision processes becomes high utility.
* The PGG model (and maybe most game theory) doesn't really dig into the idea of the individual decision arising from a series of cognitive operations. Nor is there any learning. I suppose you could say that learning to become a cooperator from being a defector is identical to replacement in their models. But if there are environmental context effects on learning, these might affect that process (i.e., change the replacement function). Both the idea of homogenizing cognition across individuals and the effect of reinforcement learning in symmetry-dominated groups (which will lead to C being rewarded over D and increase altruism) require doing some cognitive psychology on the actors.
* The issue of the virtuous D-D conditions may be complex. It seems to me to be the place where capitalism occurs -- competition for ideas, information, valuation seems to emerge from D-D interactions. I have a strong intuition that a high percentage of dyadic capitalist interactions are D-D. E.g., prices set on supply/demand rather than value to the purchaser, any "trading" that is based on estimates of future value (the transaction should only occur when the seller and buyer have different estimates). Is this an unusual idea? How common or useful are virtuous D-D interactions versus C-C interactions in capitalist interactions (buy-and-hold long from an IPO or bond is a C-C, you cooperate in providing capital, there's a communal gain from the investment growth and the rewards benefit all participators).
* The the communal action model of the PGG, D-D can be better than C-D if the group action has some non-linearity to it, e.g., it requires a certain number of C actors to gain the benefit. Then if some Cooperate but not enough to gain the benefit, it's a communal loss. This is a very minor version of the "virtuous D-D" where D-D > C-D.


The abstract from the Nature paper:

Social diversity promotes the emergence of cooperation in public goods games

Francisco C. Santos¹, Marta D. Santos² & Jorge M. Pacheco²

1. Institut de Recherches Interdisciplinaires et de Développements en Intelligence Artificielle (IRIDIA), Computer and Decision Engineering Department, Université Libre de Bruxelles, B-1050 Brussels, Belgium
2. ATP-group, Centro de Fisica Teórica e Computacional (CFTC) and Departamento de Física da Universidade de Lisboa, Complexo Interdisciplinar, Av. Prof. Gama Pinto 2, 1649-003 Lisboa, Portugal

Correspondence to: Jorge M. Pacheco² Correspondence and requests for materials should be addressed to J.M.P. (Email: pacheco@cii.fc.ul.pt).

Top of page

Abstract

Humans often cooperate in public goods games^{1, 2, 3} and situations ranging from family issues to global warming^{4, 5}. However, evolutionary game theory predicts^{4, 6} that the temptation to forgo the public good mostly wins over collective cooperative action, and this is often also seen in economic experiments⁷. Here we show how social diversity provides an escape from this apparent paradox. Up to now, individuals have been treated as equivalent in all respects^{4, 8}, in sharp contrast with real-life situations, where diversity is ubiquitous. We introduce social diversity by means of heterogeneous graphs and show that cooperation is promoted by the diversity associated with the number and size of the public goods game in which each individual participates and with the individual contribution to each such game. When social ties follow a scale-free distribution⁹, cooperation is enhanced whenever all individuals are expected to contribute a fixed amount irrespective of the plethora of public goods games in which they engage. Our results may help to explain the emergence of cooperation in the absence of mechanisms based on individual reputation and punishment^{10, 11, 12}. Combining social diversity with reputation and punishment will provide instrumental clues on the self-organization of social communities and their economical implications.