As explained in Section 1.1, a system can be meaningfully considered as an "information system", if its internal activity can be abstractly analyzed in terms of its goals, actions, and beliefs, and its external activity can be abstractly analyzed as experience and behavior. When some information systems share physical and social environment and closely interact with one another, usually the whole community shows certain regularities that can be naturally described in this way. Examples of these systems include human organizations and communities (institute, company, city, country), animal herds (ant colonies, bee swarms), computer networks (the Internet, multi-agent systems), etc. On the other hand, a collection of random information systems often does not qualify as an information system as a whole. Though the boundary between an individual system and a social system is fuzzy (some communities behave more "like a single person" than the others), we can still say that in the former the components are more closely coupled, with more consistent beliefs and goals. On the other hand, the latter have non-negligible internal conflicts in beliefs and goals.
To maintain stability and efficiency, a community often has an internal social structure, in which different members play different roles, and make different contributions to the community as a whole. The communications among members are also strongly influenced by the social structure. Consequently, a feature (goal, action, or belief) of a community is usually not the sum or average of the corresponding features of its members, because it also depends on the structure of the community.
The intelligence of a community obvious relates to the intelligence of its members, though cannot be simply reduced to that. The intelligence of an ant colony (or bee swarm) is widely considered as much higher than the intelligence of its member, and it is not too hard to find reverse situation in human organizations. Once again, the intelligence of a community is influenced by its social structure, as well as by the intelligence of its individual member.
During the adaptation process of a collective intelligence, its goals, actions, and beliefs go through self-organization processes similar to that of an individual system (as described in Chapter 4). When a community becomes stable enough, the common goals, beliefs, and actions become its "culture", which works as an "external heredity" — though it is not built into the innate structure of the individual systems, it does act on each system, to shape it in a specific direction through socialization (as explained in Section 5.4). Changes in the goals, beliefs, and actions in the community are usually initiated by an individual, though most of the proposed changes fail to realize, or end up in very different forms. The successful changes are the ones accountable for the adaptation of the community.
Some of the common goals are outer-oriented, indicating the desired events or situations in the environment of the community. Some other common goals are inner-oriented, indicating the desired events or situations in the community itself.
The community goals regulate individual goals of members through socialization. In this process, some goals of a member are encouraged, while some other goals are discouraged, depending on the expected role of the member in the community, as well as the social norms (ideology, law, ethics, religion, custom, etc.) of the community.
As the community develops, the common goals produce derived goals, and the goal alienation phenomenon (discussed in Section 4.2) become inevitable. Consequently, for a community with long history and complex structure, some of its goals may not be attributed to any of its members, but should be considered as belonging to the community as a whole. When the conflict among community goals and member goals becomes too serious, the stability of the community is at risk.
In a collective intelligence, most of its common actions are learned from the experience of the system as a whole. In principle, it is similar to the learning process of individual systems (discussed in Section 4.3, in that it is the coordination of basic operations, just that here there are multiple members involved. The learned actions are distributively stored in the memory of the relevant members, and can be passed to other members though socialization and education.
Shared knowledge starts at knowledge about the communication language(s) used in the community, which makes communication possible. The vocabulary of the language provides a common framework, in which the environment and the community are described, so that experience can be shared, generalized, and accumulated. As a supplement, some knowledge can be demonstrated in the actions of members, so that the other members can acquire it by observation and exercise. Pure personal experience, that is hard to express in language and action, cannot be effectively communicated and shared.
The basic function of common belief is to connect goals to actions, either for a member or for the whole community. The primary form of common belief is "common-sense", the knowledge every (normal) member of the community should have, acquired either from personal experience or from socialization. More advanced form of common belief are "theories", which are organized knowledge on certain aspects of the environment and the community.