Summary Task 9 - Collective Intelligence

COLLECTIVE INTELLIGENCE
THE MIND – CHAPTER 12: BODIES, THE WORLD & DYNAMIC SYSTEMS (THAGARD)

 3 challenges – body, world, dynamic systems – accuse CRUM of
 Focusing too much on mental representation
 Neglecting the fact that thought is not a solitary, disembodied occurrence, but
occurs in individuals who interact with a physical world


BODIES IN THE WORLD

 Body & world challenge – thinking is not just in the head
 CRUM seems to restrict thinking to computational processing occurring in the
mind

EMBODIMENT & DIRECT PERCEPTION

 How do people interact with the world?
 Info needs to be conveyed from the world to the mind through senses


CRUM  Inferential view of perception
 Treats perception as involving the inferential construction of representations
that capture features of the world

 Inferential view – rejected  learn about the world more directly, our perceptual
apparatus so attuned to world that info is directly conveys to brain without requiring
computations on representations
 Physical sensory apparatus – one of the contributors to our ability to interact with
the world
 Body-based relations in language – up & down, left & right, in & out
 Many key aspects of human thinking depend on body we have & how it is attuned
to world
 Basic concepts we use to categorise world derived from way our sensory systems
detect the overall part-whole structure of the world
 Support: conceptual representations are grounded in specific sensory modalities
 Example: representation of car not abstract, verbal symbol BUT involves neurons in
brain’s visual areas that capture & re-enact sensory experiences of cars

BEING-IN-THE-WORLD (HEIDEGGER)

 How do people hammer in a nail?


CRUM  Representational view
 Start to consider what kinds of representation we have of a hammer & nail
 Hammering takes place bc we are able to do computational operations on

, these representations that somehow get translated into physical action of
hammering the nail

 Representational view – rejected  we function in the world simply bc we are a part of it
 “being-in-the-world” – conveys that people can perform tasks just by virtue of their
physical skills, without any kinds of representation
 AI trying to formalise & represent knowledge = hopeless  our intelligence is
non-representational
 Antirepresentational view of cognition
 Found favour in some AI researchers
 “embedded computing” (Smith) – avoids representational load by emphasising
interaction with world rather than internal processing

ROBOTICS (BROOKS)

 Embodiment – major theme in research in robotics
 Brooks – builds simple machines that have capacity to learn about their environments
 Insect-like robots – multiple processors that enable them to learn to walk by
interacting with the environment WITHOUT representation techniques used in
CRUM
 Designed to be able to recognise & respond to faces BUT contain no rules /
concepts for reasoning about people

SITUATED ACTION (SUCHMAN & LAVE)

 Cognitive psychologists – examined human thinking on artificial tasks
 Problem solving in realistic context not so heavily dependent on mental
representations BUT depends on direct interaction with world & other people
 People are thinking through interaction with the world
 CRUM incapable of appreciating the subtle, contextual ways in which people deal with
the world

INTENTIONALITY (SEARLE)

 Relation between mind & world – crucial objection to CRUM
 Mental states – intended to represent the world, they possess intentionality (they are
about sth)
 Computer could never have intentionality
 Computer manipulates symbols  lacking in understanding
 Symbols that people operate with (e.g., letters/words) have semantic meanings 
have intentionality
 Representations in computers are independent of the world  lack intentionality
 Symbols they use have no meaning for them