Transcript Introduction to Cogsci

Introduction to Cogsci
April 07, 2005
Central Theme
• Cognitive Science was occuppied with the
algorithmic level for much of its history:
successive manipulation of semantically evaluable
symbolic representations.
• Cognitive Science today: two major new trends:
The three levels
•
•
•
•
Computational
Algorithmic: input, computation, output
Implementation
One way of putting lesson from before:
once you start to deal with real systems in
real worlds, then things look different.
Clark’s problems with Marr
• Central problems, p. 85:
• Distinctions among levels not always clear
• Process of discovering should be deeply
informed by neuroscience
Process of Discovery
• Biological evolution is liberated by being
able to be mess
– Hand case
– Perceptual adaptation can be “motor specific”
(87)
Interactive vision
• 88: 4 points
Mirror Neurons
• Mirror neurons and knowledge of other’s emotions
Mirror Neurons
• Fire in response to action in others
• Constitute in part a mirroring of a motor
routine
• Connected to the limbic system and so to
emotional reactions
• Explain potentially why facial imitation
produces emotional synchronization
Summary
• “… the brain is revealed not as (primarily)
an engine of reason or quiet deliberation,
but as an organ of environmentally situated
control.” (95)
Unclear Distinctions Among
Levels
• Can get an upward cascade of influences in
which isolating ‘the right task’ is
problematic
• Many different levels of implementation
• Timing washes out in algorithmic
description
• Biology need not give us neat
decomposition
Reverse Engineering (Ch. 6)
• 2 examples in which behavior is revealed to
have a basis very different from what one
might have thought:
– Cricket: does not have general representational
and computational abilities; has special purpose
strategies effective in environment.
– Flocks of birds: no one has a plan, just three
simple rules (108)