Very often, it makes sense to treat a community of information systems as an information system, and such a community intelligence share many, though not all, properties with individual intelligence systems.
Section 6.1. A community as a system
A community is a group of information systems with shared physical and social environment. When multiple system closely interact with one another, often the whole community deserves to be described as an information system. Such examples include an art colony, a government, or a computer network.
When a community as a whole is treated as an information system, its goals, actions, and beliefs are related to those of its members, though in a complicated manner. A community can have various internal social structures, in which members play different roles, and the information processing activities follow different procedures.
The boundary between an individual system and a social system is fuzzy, though 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.
Common beliefs take the forms of common sense, custom, religion, or science. Common actions are developed through cooperation of the members, as well as through the development of technology, which is tool building and applying at the level of a society. Common goals take the form of ethics, ideology, or religion, which normally is consistent with the individual goals overall.
When a social system becomes stable enough, the common goals, beliefs, and actions become its "culture", which is 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. Changes in the goals, beliefs, and actions in the community often initiates by an individual, though most of the proposed changes fail to realize, or end up in very different forms.
Section 6.2. Common knowledge
In a community, common knowledge is the knowledge that can be passed from system to system, and from generation to generation. By nature, common knowledge must be expressible in a certain language, or can be demonstrated in action, so that the other systems can obtain it. For this reason, pure personal experience is hard to fully get into it.
Theories are generalized and structured common knowledge, which can be applied to concrete, often novel, situations. Also, a theory is a compact and economic way to represent a huge amount of knowledge, by allowing conclusions to be derived from, not merely included in, the given knowledge, according to a "logic" (i.e., valid inference rules) associated with the theory.
Two major types of theory are "closed" and "open", respectively. The fundamental difference between these two types of theory is their attitude toward new experience. Though both can be seen as compressed form of some past experience, a closed theory forbids revisions caused by new evidence, while an open theory allows such revisions.
The structure and logic of a closed theory work in such a way that it is prevented from possible revisions from new evidence. One common method to do so is to declare its core judgments are "absolute truth", by which the truth value of all other statements are judged. Another method is to avoid concrete predictions about future events, so that no matter whatever evidence there is, it can always be explained as consistent with the core.
On the contrary, an open theory is always open to possible revisions. It also has "core statements" and "peripheral statements", and the former is supported by more past experience, therefore harder to be revised. Even though, it must acknowledge such a possibility, and shows how it can happen when its predication is inconsistent with future experience. In short, content of such a theory is "predicative knowledge", in the sense that it makes concrete predictions about future experience, so that it can be tested, and revised if the predication turns out to be wrong.
Compared to our previous discussion of individual systems, "closed theories" corresponds to "instinctive systems", while "open theories" corresponds to "intelligent systems". They are good for different environments. In a relatively stable environment, a closed theory may provide an efficient way to unify a community and direct its behaviors. Systems in the society can solve their doubts and disagreements by comparing them with the central dogma of the theory, and the behaviors of other systems are predictable to a large extent. On the other hand, an open theory is needed in situations that are novel in nature, and change from time to time, so that past experience does not always dependable.
Section 6.3. The nature of science
Science is open theories based on organized common beliefs that can be used to link actions to goals in the future for a community. The development of science follows the same logic as the self-organization of beliefs in an individual intelligent system.
Science is common knowledge, so it must be able to be expressed in some language, to be communicated among systems. Also, it must be accepted by many individuals in the community as consistent with their experience. A scientific theory must be "objective", not "subjective", in the sense that it should be from the view point of average member of a society, rather than only from the view point of specific members in the society.
Science is not a collection of unrelated statements. It consists of theories, and each of which is some knowledge organized together. A well-organized theory should be as close to an axiomatic system as possible, which means that
- It contains clearly defined concepts, whose extension and intension are sharp;
- It contains a small number of principles or laws, which describe the relations among the concepts, and are well-supported by past experience;
- It contains effective procedures to deal with practical problems in the corresponding domain.
This type of systemization is designed, because all information systems that use a scientific theory are restricted by resources.
A scientific theory is the result of deliberate mental works, and it does not directly come from experience. For given experience, the possible theories one can build is usually not unique, though not all of them are equally good in terms of systemization.
A scientific theory must make concrete prediction about future events, and be ready for revision if the predictions are different from future observations. Therefore, it cannot consists of general statements that explain everything but predict nothing.
Section 6.4. Empirical science and abstract science
An empirical science keeps direct relation to the experience of the systems. As a theory, it captures the stable items and patterns in experience as concepts, and stable relations among the concepts as statements. An empirical science usually follows non-axiomatic logic. Many problems discussed in the philosophy of science can be addressed using the logic in NARS.
With such a theory, a system can predict future events, and behaves accordingly to achieve its goals. If a prediction is disconfirmed by a following observation, the trust to the theory is more or less decreased, and some revision is carried out in the theory. Usually the general statements in the theory is revised less than the concrete ones.
A scientific theory focusing on a small domain is easier to become well-organized than a theory that deals with a complicated domain, because of the variations of the experience and the limit on the processing capacity of a system. Consequently, natural sciences qualified better than social sciences as "science". Philosophy is the study of the most general domain in our experience. As a result, few philosophical theory is organized so well, that it deserves to be called a "science" in the usual sense of the word. Nevertheless, the difference is usually a matter of degree.
Abstract science is not directly based on experience. Instead, it consists of a convention system, where meaning and truth are defined internally in a closed world. The examples of this type of science include various kinds of mathematics and formal theories. An abstract science usually uses pure-axiomatic logic to present its result, though non-axiomatic logic is needed when the theory is developed and used. The resulting theory is quite different from an empirical science, in that it has no tolerance to ill-defined concept, uncertain conclusion, and internal inconsistency.
Usually, an abstract science is also developed with practical problems in mind. However, when the theory is presented, its empirical relations with experience are omitted or ignored. The foundation of such a system is a given set of axioms, which is assumed to be true, and they give meaning to the primary concepts appearing in them. According to such a foundation, the meaning of other concepts are defined, and the truth of other statements is decided, according to whether they can be derived from the axioms according to the given inference rules.
When an abstract science is used in empirical science or daily life to solve practical problems, an "interpretation" step is needed to build a one-to-one mapping between the formal concepts in the theory and the concrete concepts in the domain, in such a way that satisfies the axioms and assumptions of the theory. Then, according to the implication relations within the theory, concrete conclusions about the domain can be derived.
Science 6.5. Relations among scientific theories
Two theories can co-existing with a generalization-specialization relation, in the sense that the domain of one is a component in the domain of the other. These two theories are consistent as long as their relevant conclusions are consistent. For a specific object, relation, or event, it always can be describe on different levels of generality, but a specific description on a certain level is not always translatable to other levels without changing the information in it. In general, neither the more general theory nor the more specific theory have higher privilege in such a relationship, in the sense that the other theory must be reduced to it or be consistent with it.
Another form of co-existing theories is that one theory is the object of study of another theory. In this case, the latter is usually called a "meta-theory" of the former. If somehow a theory and its meta-theory have overlap, we get the a "strange loop", that is, some kind of self-reference. Sometimes this should be avoid, because it leads to paradox. However, in some situations it is desired or inevitable. For example, a theory on philosophy of science must be applicable to itself, as far as it is proposed as a scientific theory.
For a given domain, due to the difference in motivation, experience, background, and so on, different people (or systems) may build different theories. They compete with each other in their descriptions about the domain and their predictions about future situation. According to the traditional opinion, science is to reveal the "truth" lying in the objective world, therefore, competing theories cannot all be correct (though may be all wrong). However, if theories are seen as summaries of experience, this conclusion does not hold anymore. Usually, in a debate all parties have some "truth" in their arguments, since they are viewing the situation from different angels, and no one is absolutely correct. However, it does not mean that all the competing theories are equally good.
The conflicts among theories usually cannot be solved by merging them together into a "summarized theory", because their conceptual systems are very often incompatible. For the same reason, the people in different schools often cannot solve their difference by debate. In complex domains, it is unlikely to find a "decisive test" that can settle the competition once for all, and succeeds and failures are often temporary.
In scientific research, the attempt to keep all competing theories alive and the attempt to unify them into a single theory are both valid, and valuable for the progress of science.
Science 6.6. The science of intelligence
This last section provides a self-referential summary of the book by applying the theory about science into the domain of intelligence.
A scientific theory of intelligence should consists of a set of well-defined concepts, organized in a hierarchy, as well as statements about the relations among these concepts. The theory should make concrete predictions about what will happen and what won't happen, and give concrete instructions on how to solve various types of problem in the domain. The theory should be relatively simple, so as to be easy to understand and to use.
Such a theory should explain the observations in human intelligence, animal intelligence, and community intelligence. Therefore, it should be consistent with our knowledge about human psuchology, though it is not an attempt to replace psychology. Instead, it should separate the human-specific content of psychology from the human-independent aspect of intelligence. Similarly, this theory should be consistent with the general principles generalized from the study of linguistics, logic, philosophy, anthropology, sociology, education, and other related domains. This theory should allow alien intelligence as a possibility.
Such a theory should provides guideline for the development of artificial intelligence. It should explain the successes and failures in the history of AI, and propose solutions for the pending theoretical problems. Even so, this theory cannot fully replace a theory especially targeting the AI field by taking computer-specific properties into consideration.
The theory presented in this E-book is such an attempt.