The research project of Artificial Intelligence, AI, was formally established in America at a Dartmouth College workshop in New Hampshire in the summer of 1956. The project’s goal: configure a certain type of machine, “AI’s machine”, to exhibit human-like intelligence. This hoped intelligence would be exhibited in the form of written language emitted by the machine. The August 1955 flier for the workshop put it this way:
“The study is to proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it. An attempt will be made to find how to make machines use language, form abstractions and concepts, solve kinds of problems now reserved for humans, and improve themselves. We think that a significant advance can be made in one or more of these problems if a carefully selected group of scientists work on it together for a summer.”
In order to understand and pursue this goal, the project used a set of ideas that comprised a conceptual framework, a set largely unchanged in AI today. And the above extract from the 1955 flier strongly indicates a core part of this framework.
In examining a conceptual framework, one might start with elementary ideas then work up to compounds, since then part of the needed work will already have been done. The idea of the symbol is considered elementary to various significant AI concepts. For instance, that of simulation, knowledge, and problem solving.
In due course, we hope to discover why AI uses the idea of the symbol at all, then whether this concept helps or hinders (or is neutral to) an accurate understanding of human-like intelligence. At present, however, it’s enough to note that the project does indeed use the concept and also relies upon it.
In examining this use, I’d like to start at the beginning of the project, identify when the concept was first used, who first used it, and what the term was used to refer to.
Most discussions of early AI rightly focus on renowned British theoretical mathematician and WWII code breaker Alan Turing (1912-1954). In 1936 he invented an abstract machine. What “abstract” means here will be discussed a little later. During WWII this abstract design was realized (so it is said) as a practical device. Then after the war in the late 1940s and early 1950s Turing founded the AI project with its goal of configuring the practical machine to exhibit human-like intelligence. Hence the practical machine came first. In fact its development inspired the founding of the AI project and its dominant theory of mind. This theory says the mind operates on the same functional principles as the machine. Thus AI both in practice and in theory was spawned by a new machine.
The 1956 summer workshop was preceded by several years of informal research. American professors Herbert A. Simon and Allen Newell, for instance in 1954 wrote and tested a machine configuration they called Logic Theorist, then presented this at the workshop.
AI, in order to configure its machine, needed to understand it. Hence some ideas of the project are also ideas that explain the machine. The notion of the symbol is one of these. This notion was key to understanding the machine, and AI naturally adopted the same idea to understand the mind.
Apart from the single case of 17th Century British philosopher Thomas Hobbes (1588-1679), 20th Century British mathematician Alan Turing was the first in the surviving history of Western thought to propose that the mind operates internally in the same functional manner as the WWII machine. This proposal is typically traced back to Turing’s 1947 lecture to the London Mathematical Society about a version of the machine he was working on at the time, called the ACE. Turing explains,
“Machines such as the ACE may be regarded as practical versions of [my abstract 1936] machine.“
The practical machine could alter its own configuration and,
“When this happens I feel that one is obliged to regard the machine as showing intelligence.”
So to understanding AI’s conceptual of the symbol we start with AI’s machine (such as the ACE). And in understanding AI’s machine, we start with Turing’s abstract 1936 design. This seems to neatly divide examination of AI’s idea of symbol into:
- Turing’s abstract design, now called the Turing machine
- The WWII “practical version”, or realization, of this design
- The concepts AI uses to understand the WWII machine
- The detailed electrical nature of the WWII machine
- Conclusions about AI’s concept of symbol