Reciprocity as a Supplement to Trust and Reputation. An interesting overview to the ‘evolution of cooperation’ is provided by Mui [207]. In the context of LiquidPub, these theories may be used for defining the various incentives for peers to cooperate or defect (such as lying when rating others or providing unreliable reviews). Four theories are presented by Mui [207]: the group selection theory, the kinship theory, the reciprocation theory, and the social learning theory. The group selection theory states that cooperation amongst individuals is not consistent with Xxxxxx’s theory. As an alternative, it suggests that cooperation is a result of natural selection being based on the group level: species, community, etc. However, strife is often observed within species. This weakens the group theory, suggesting the need for alternative theories. The kinship theory states that individuals are ready to cooperate and sacrifice themselves for their kins, e.g. parents and children. The closer the relationship with the kin, the more altruism and less aggression is shown, even if it was over one’s own personal benefits. However, the main problem with this theory is that it does not answer the relatedness and competition that sometimes exist among kins. The reciprocation theory states that ‘reciprocal altruism’ is the reason behind individuals sac- rificing their personal gain for the good of others. As such, all cooperative behaviours are viewed as postponing immediate personal gain to benefit from future reciprocations by others. Well de- fined equations are presented for computing reciprocation. Furthermore, this is the only theory (amongst the presented four) that has been verified by experimental work, proving that ‘image scoring does play a role in actual human cooperation’. Finally, the theory of social learning suggests that in the absence of kinship and reciproca- tion, cooperation is based on ‘cultural transmission’. This implies that individuals learn the most dominant behaviour in their community. However, no experimental work has either proved or disproved this theory. As for the computational trust model, Mui [207] introduces the concept of reciprocity into its proposed computational model. While trust is defined as ‘a subjective ex- pectation an agent has about another’s future behaviour’ and reputation as the ‘perception that an agent has of another’s intentions and norms’, reciprocity is defined as the ‘mutual exchange of deeds, such as favour or revenge’. The relation between these three concepts is ...
