The properties of the formal system vs the way we

• The properties of the formal system vs. the way we interpret those properties
• interpretation: a specification of what the tokens of a formal system mean
• semantics: what tokens mean, stand for or represent
• syntax: formal and structural properties of the system
• the syntax is the study of the grammar of the tokens of a formal system (e.g., What are the tokens? What
moves are legal?)
• the semantics is the study of the meaning of the token
How do those two lives come together?
According to the computational approach, “if you take care of the syntax, the semantics will take care
of itself ” (206)
• Formal logic: truth preservation is guaranteed by simply attending to the formal properties symbols
◦ ex) algebra
▪ rules of algebra are truth-preserving

▪ if the theorems are assume to be true, and if the rules of algebra are truth-preserving, then any
application of a rule on a theorem will give rise to another theorem which is also true
• impressive result: what seemed to be a characteristic feature of human minds – can be assumed of
computers as well
◦ revisiting the computational approach: minds are computer of a certain sort
◦ but minds are not only good at following rules or solving problems; they are also capable of acting
in a way that consistently makes sense
• So, if minds are computers, then it must be shown that there are some automatic formal systems that can
be interpreted in a way that they consistently make sense
In addition to being truth-preserving, the formal system must be:
a. Rational: i.e., it must be capable of (i) easily generating obvious, logical, and commonsensical consequences,
and (ii) eliminating inconsistencies
b. capable of reliably interact with the world
c. cooperative in communication
d. “Knowledgeable” of certain assumptions of language and discourse
Haugeland:
“Interpreting an automatic formal system is finding a way of construing its outputs such that they
consistently […] many other considerations are important as well” (209).
Summary
• The computations approach to the mind holds that intelligent beings are computers
• That is, intelligent beings are automatic formal systems that can be interpreted in a way they
consistently make sense;
• If “making sense” was equivalent to “being truth-preserving” then the computational approach would
have been vindicated;
• But the two terms are not equivalent
Two objections to the computational approach
1. Objection 1: Hollow Shell Strategy: Minds (or intelligent beings) are not semantic engines.
• Haugeland's response: Consciousness is mysterious. So how do we know that genuine understand isn't
possible without consciousness? And how do we know that semantic engines are not conscious?