Spring 1991




Master of Arts (MA)

Primary Subject Area

Philosophy; Religion, General; Religion, Philosophy of


Advocates of strong artificial intelligence believe

that properly programmed computers can go beyond the

simulation of intelligent acts so as to instantiate and

exhibit true intelligence, that is, intelligence equivalent

to that of man. In this thesis, I consider three problems

for Searle.

First, John Searle's well-known thought experiment of

the "Chinese Room" is used to reestablish the syntax-semantics

distinction and to show how this distinction

applies to computer programs. I review the Chinese Room,

consider a variety of objections to it, and then expand on

the key points in Searle's work.

Second, 1 examine the Frame Problem in artificial

intelligence, a question made popular by Daniel Dennett.

Rational agents have the ability to adjust their conceptual

schemes and update their noetic web of beliefs so as to

maintain a representation of the world. This ability is

easily observed but not well understood. I argue that

computers lack this ability altogether. The Frame Problem

examines this deficiency and programming techniques designed

to overcome it.

Third, the Overseer problem examines the need for

artificial systems to have a rational agent in place who

designates a given task and determines when that task is

successfully completed by the system. I argue that as long

as this need exists, artificial systems cannot be considered

intelligent in an unequivocal sense.