Cooperation occurs amongst individuals embedded in a social environment. Consequently, cooperative interactions involve a variety of persistent social influences such as the dynamics of partner choice and reward division. To test for the effects of such dynamics, we conducted cooperation experiments in a captive population of Japanese macaques (Macaca fuscata, N = 164) using a modified version of the loose-string paradigm in an open-experiment design. We show that in addition to becoming more proficient cooperators over the course of the experiments, some of the macaques showed sensitivity to the presence of potential partners and adjusted their behavior accordingly. Furthermore, following an unequal reward division, individuals receiving a lesser reward were more likely to display aggressive and stress-related behaviors. Our experiments demonstrate that Japanese macaques have some understanding of the contingencies involved in cooperation as well as a sensitivity to the subsequent reward division suggestive of an aversion to inequity.
This article describes the transition of an undirected active network, used for numerical and logical analysis, and made up of communicating operators and nodes, and being self-extensible, into a network of semantic structures, and the extensions necessary to handle language. The network, called Active Structure, was infused with knowledge in the form of tens of thousands of definitions taken from a human-readable dictionary. Where the dictionary fell short, from circularity, lack of a hierarchical form for its senses and sub senses, or inadequate detail around verbs, the entries were extensively curated. The result is a machine which has access to every meaning of every word in its vocabulary, and needs to manage the right meaning for every use of a word, a task that our unconscious minds now handle. To demonstrate that such a network can aspire to human-level tasks, we have begun loading an Anti-Money Laundering Act so that the machine can devise and test strategies against money launderers. There is a higher goal being aimed at, the breaking of the Four Pieces Limit. That is, the network is not intended to save money on human cognitive labor, but to raise an absurdly low limit on it.