Salve Meeting 2003

The ‘homunculus’ fallacy: when philosophy comes real


Gábor Szirtes
http://people.inf.elte.hu/gszirtes
ELTE Faculty of Informatics
Dept of Information Systems
NIPG group

 


The problem of representation

The brain computes
Computing means encoding information through changes of state variables
Encoding is a transformation of the perception of the (partially) observed environment into a ‘representation’


The Homunculus Fallacy

How do we know that this is a phone?

 





Working Hypothesis

Fallacy can be resolved:
Let us turn the infinite regression into a loop:
The input (!) ‘makes sense’ if it can be reconstructed by means of the representation

Auto-association (remember?)



Biology or Engineering?

B: ‘Cause-oriented’
E: ‘Goal-oriented’
Link: Reverse engineering?

Strong AI: states that a computer with the right program would be mental. The brain implements the software of human mind. It can be duplicated (Mind body problem resolved…)


Turing Test

Operational definition?
All purpose Vacuum Cleaner (Gunderson,1964)
Anthropomorphism (Millar,1973)
Behaviorism (Block, 1981)
Chinese Room (Searle, 1980)


Against Searle

Slow hardware (Dennett, 87)
Other minds
Other forms of intentionality (Carleton, 84)
System reply (Rorty, 84) But: observer-related intentionality
Wrong causal relation to the world (Fodor, 80)


(No) conclusion

Underlying properties of a (seemingly) cognitive
System

Goal-oriented
Interacting
Dynamic
Emerging complexity

What is missing?