The Society of Mind Analysis
Describing the relationship between the brain or another physical entity and the mind has become an important part of many diverse disciplines. Understanding how the brain gives rise to thought and intelligence is one of the most profound questions in contemporary science and one to which Marvin Minsky has devoted his career at the Artificial Intelligence Laboratory at the Massachusetts Institute of Technology (MIT). The Society of Mind is a compilation of his theories and ideas about intelligence and the possibility of its creation in machines.
One of the major points Minsky makes about intelligence is that commonsense reasoning is more difficult to describe and explain than is logical rule-bound reasoning. Thus, it has been easier to program computers to play chess than to build a children’s tower. When Minsky makes this assertion, however, he neglects to specify that computer chess-playing programs do not recognize the pieces through a visual mechanism nor do they physically grasp and move the pieces, which are part of the requirements for his computer program “Builder.”
The Society of Mind is primarily concerned with explaining how humans learn to think and solve problems. The theory of intelligence formulated here is indebted to the concepts of learning investigated by Jean Piaget. Small children, as they gain knowledge, motor skills, and the ability to understand sensory data and relate them to a meaningful reality, are the model for learning and intelligence. Minsky has collaborated with Seymour Papert, a student of Piaget, to recognize and formulate some understanding of the principles which underlie human learning strategies. One example discussed in detail throughout several essays is the way children learn to recognize amounts. For example, a five-year-old child will watch as water is poured from one container into a taller, thinner container and will say that the taller container now has more water in it. By the time a child is seven, he or she will typically realize that it is the same amount of water. Minsky places this specific and universal process into a large and complex “society-of-more.” The concept of “more” applies to many diverse elements, all of which need some organizing principles in order to relate to one another and to the larger society. This requires hierarchical systems which relate the lower level agencies to the higher level concepts. Minsky’s diagrams demonstrate Papert’s principle: “Some of the most crucial steps in mental growth are based not simply on acquiring new skills, but on acquiring new administrative ways to use what one already knows.”
Minsky presents the way children learn meaning as primarily through an experience of trial and error. His examples are drawn from a physical experience that the child has which relates to an understanding of a concept or a definition. Various aspects of learning can have an impact on understanding—including recognizing repetition and realizing that humans tend to see things as wholes, adding missing elements. Nevertheless, people learn not only from physical experience but also from language and other abstractions. Minsky tends to define language as an agency of its own which is connected with other agencies. He does acknowledge that language creates one’s image of reality, although he qualifies that statement: “Yet words themselves can’t be the substance of our thoughts. They have no meanings by themselves; they’re only special sorts of marks or sounds.” Language is one of the most powerful controlling agencies; it seems to regulate what consciousness does. This makes language one of the higher level, or managing, agencies in the society of the mind.
According to Minsky’s definition, a word “makes various agents change what various other agents do.” Words are tied by knowledge-lines to other agents, which are activated or changed by the use of the word. In fact, the meanings of words only become totally clear in relationship to other words; that is how the inherent ambiguity of language is handled. Language is also intimately associated with the way connections between agencies are made. Out of language arise expression and creativity. Minsky’s argument is that creativity and intelligence must originate from the language agency, which is in turn related to other smaller agencies. He expands the concept of meaning to describe how the ambiguity of all isolated knowledge is resolved through larger contextualizations. Stories and other verbal narratives have cohesion because they are related to a context that makes sense. Just as humans tend to see partial designs as complete forms, they tend to complete stories and narratives as well.
In order for a story to be truly understood, there must be recognizable connections between the various phrases, and these connections can also activate specific knowledge about contexts in the mind. In addition to these recognition agents which lurk in the mind, Minsky draws on Freudian psychology to describe suppressor agents and censor agents which prevent one from acting on certain thoughts or even from recognizing certain thoughts and ideas. He uses these agents to develop a definition of humor and to discuss how humor relates to learning. His discussion of jokes also is influenced by Freudian psychology and defines jokes primarily as stories which fool the censors. Minsky goes on to hypothesize that humor plays an essential role in learning because it allows one to escape the censors, which may be limiting new ideas and stifling creativity. Humor allows connections and analogies, which may be misleading or which may lead to new insight, to be explored. Clearly, in Minsky’s scheme humor and even laughter are built out of smaller agencies and serve a necessary function in the learning process.
The Society of Mind demonstrates how intelligence, problem solving, learning, and models of reality can all be built from lower level agencies. Minsky asserts that “minds are simply what brains do,” and he claims that brains are primarily machines. Yet he makes it clear that he means an extremely complex machine and even acknowledges that it is correct to deny that the brain is similar to a simplistic machine. Nevertheless, this mechanistic emphasis is also responsible for some of the omissions of the book as a complete presentation of intelligence and the learning process. He ignores all studies of animal intelligence and even seems to want to define intelligence in such a way that animals cannot possess it. For example, when discussing the problem solving inherent in an activity such as the dam building of beavers Minsky states that this knowledge is genetically based rather than truly learned. According to his theory, the ability of a beaver to build a dam is not a sign of intelligence or learning ability.
One other major limitation of the theory of mind as presented by Minsky is that it does not acknowledge some of the significant aspects of human memory. For example, some recollections are much more vivid than others; some are much easier to recall; some are sublimated entirely. In addition, humans are not always aware of the distinction between memories of actual events and fictional memories. Minsky simply does not discuss these aspects.