Notes on Symbolic AI

Haugland offers a description of an actual practical system which later on a computer system ought adhere to (solve, reproduce, etc.). The type of tasks presented are formal systems which are token-manipulative: tokens are manipulated according to definite rules, contain formal tokens and starting positions and a set of formal rules. The systems are entirely self-contained, digital in the sense that the results are coherent with and do not stray from whatever is possible within the systems and medium independent in that the task does not rely on any specific medium.

To solve such systems, Haugland presents Automatic formal systems. Important aspects of such systems are that they are based on general purpose software which is able to create specific automatic formal systems itself. The general computer follows its own rules with regard to all of its tokens; but the program tokens are so arranged that the net effect is to manipulate the configurations implementing the tokens of the special computer in exactly the way required by its rules. Universal machines are thus computers which can perform any formal task, given enough time and/or hardware and may use simulation to get to the desired output through the backdoor. Algorithms and heuristics define the automatic rearrangement of tokens.

Of course, the challenge and the solution presented here on a meta level beg the question if mental tasks performed by humans can be defined in formal systems or simulated as such. One such approach is GOFAI, which manipulates symbols (tokens which meaning) in a way as to produce something which is recognizable as human reasoning, but thereby of course onloading all the problems that come along once one tries to formalize systems based on human language. Recalling Haugland’s challenge, GOFAI uses a set of tokens/symbols, internal states and rules to get to it’s result. The largest issue for GOFAI has been large scale knowledge representation of common-sense knowledge and all those little bits of knowledge which humans seemingly automatically know in a situation, though it’s not explicitly inherent.  


Haugeland, John (1997) “What is Mind Design?,” in: Mind Design II Philosophy, Psychology, and Artificial Intelligence. MIT Press: 1–29.

Mitchell, Melanie. (2019) Artificial Intelligence. Farrar Strauss & Giroux. Chapter 1: “The Roots of Artificial Intelligence”


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s