Perceptual Symbol Systems - PowerPoint PPT Presentation

1 / 19
About This Presentation
Title:

Perceptual Symbol Systems

Description:

PSS Core Properties. P-syms - Stand in as referents ... O - Olfactory. A - Auditory. L - Language. S - Spatial. E - Emotional. Other systems certainly exist ... – PowerPoint PPT presentation

Number of Views:188
Avg rating:3.0/5.0
Slides: 20
Provided by: jamest45
Learn more at: https://ed.stanford.edu
Category:

less

Transcript and Presenter's Notes

Title: Perceptual Symbol Systems


1
Perceptual Symbol Systems
  • L. W. Barsalou (1999), Brain and Behavior
    Science, 22, 577-660.

An overview of Barsalous Perceptual Symbol
System theory
James Sulzen June 6, 2001
Psychology 264 Gordon Bower Spring 2001
2
Terminology
  • P-state - Perceptual state
  • A combination of multimodal percepts constituting
    an experience or some aspect of perception of the
    real world.
  • P-sym (or p-symbol) - Perceptual symbol
  • Recalled or constructed subset of p-states which
    symbolically stands for a referent of some sort
  • PSS - Perceptual Symbol System
  • System of symbols and processes which operate on
    them to produce cognitive processes.
  • Frame
  • Simulation

3
Amodal Symbol Systems
  • Amodal systems transduce p-states into amodal
    equivalents.

4
Amodal vs. Modal Issues
  • Evidence Little/no direct evidence for amodal
    systems
  • Neuroscience Psychology Much evidence for
    modal processing
  • Awkwardness Certain computations are very
    amodally awkward (i.e. spatio-temporal)
  • Transduction and symbol grounding Amodal
    systems are essentially arbitrary.
    (Grounding/associating amodal representations
    back to perceptual ones just ultimately rates to
    make the amodal systems redundant.)
  • Too much power Amodal systems ultimately are too
    powerful They can explain anything. Modal
    systems have a priori limitations (modal ones do
    not). Modal systems are open to falsifiability
    and provocative hypothesis

5
PSS Core Properties
  • P-syms - Stand in as referents
  • Frames - These combine p-syms frames with
    relationships
  • Simulators - dynamically bind other elements
  • Language - Can stand for or drive other elements

6
PSS Core Properties
  • P-syms
  • Have a neural representation / substrate
  • Are schematic representations of p-states (see
    Fig. 1)
  • Arise from repeated exposure neuronal
    recruitment / exclusion
  • Are inherently multimodal (including
    proprioception and introspection)
  • Selective attention operates to extract p-sym
    attributes
  • Frames
  • Organize p-syms (i.e., define primitive
    relationships - up/down, in/out)
  • Combine together
  • Predicates
  • Attribute-value bindings
  • Constraints
  • Recursion

7
Multimodal Organization of Knowledge
Sensory registers
V- vision G - Gustatory H - Haptic K -
Kinesthetic proprioceptive O -
Olfactory A - Auditory L - Language S -
Spatial E - Emotional Other systems certainly
exist
- Each modality can be thought of as a
representational system - Each provides certain
affordances
8
Frames
Establishing an initial frame for car after
processing a first instance.
Evolution after processing a second instance.
Creating a simulation of second instance from
frame in B
9
Simulators
  • Simulators
  • This is an area requiring more development in
    the theory
  • Frames a set of p-syms, frames, simulators to
    create higher-level structures
  • Temporally dynamic integration of other elements
  • Can be dynamically constructed, modified,
    componentized, and executed
  • Comparable to mental models, schema, concepts,
    etc.
  • Simulation
  • This is the execution of a simulator
  • Are always sketchy and incomplete - are never
    veridical
  • Idealization occurs (i.e., Gestaltist principles
    apply)
  • Example Categorization - if a category
    simulator can produce a satisfactory simulation
    of a perceived entity, then the entity belongs in
    the category (p. 587)
  • Derived properties of simulators
  • Productivity Can be combinatorially and
    recursively combined
  • Propositions Can be bound to individual entities
    framing?
  • Variability Implement variable embodiment
  • Abstraction Can combine physical and
    introspective events to represent abstractions

10
Simulator Productivity
Object categories
Spatial relationships
Combinatorialproductivity
Recursiveproductivity
11
Prop- ositions
Proposition representation (the balloon is above
the cloud)
Complex hierarchicalproposition (balloon above
cloud)
Alternative proposition (balloon below cloud)
simulators
simulations
Perceived situations
12
Language
  • Can be linked any other element (p-sym, frame,
    simulator)
  • Can be used to construct and control simulators
  • Language is probably the amodal basis of
    cognition and memory?

13
PSS Symbol Manipulation
  • Barsalou essentially shows that p-syms, frames,
    simulators, etc can
  • Be composed, associated, combined, subtracted,
    and so on
  • They constitute a symbol manipulation system
  • He does not demonstrate a formal correspondence
    between his PSS and an amodal symbol system(such
    as a semantic knowledge network)
  • Such a demonstration would constitute a form of
    completeness proof
  • Language as a representational system and can
    serve the role of an amodal symbol system

14
Abstractions
  • Metaphor (anger gt liquid exploding out of a
    container)
  • Not adequate in of itself to represent all
    abstractions
  • Abstraction are constructed from three main
    elements
  • Framing Are framed against the background of a
    simulation
  • Selectivity Selective attention highlights the
    core content of the concept in the context of the
    simulation
  • Introspective symbols (i-syms) These are central
    to representing abstract concepts (i-syms are the
    internal sense of an experience or internal
    perception)
  • Methodology for identifying abstractions
  • Find a frame that characterizes the abstraction
  • Identify seeming p-syms and i-syms
  • Identify the focal element that characterizes the
    abstraction

15
Truth Falsity
Mapping succeedstruth
  • Simulated event seq. frames the concept
  • The abstraction is only a focal part of the
    simulation (i.e., the outcome that construes the
    concept)
  • Introspective symbols are central to the
    construing of the meaning of the concept
  • After many mappings, a simulator develops for
    truth

Mapping failsfalsity

After many mappings abstractionsimulatorbecomes
established
16
Anger
  • Anger involves
  • Appraisal of an initiating event, and that the
    agents goal is blocked by the event
  • Intense affective states
  • Behavioral responses (i.e., disapproval, revenge,
    and redirecting goals)
  • Core component Blocked goal
  • A goal is a simulated state that the agent
    desires to achieve
  • A blocked goal is a failed mapping when it is
    expected to successfully map
  • Similar to falsity
  • A simulated situation fails to map to a perceived
    situation
  • Except that affective states and behavioral
    responses are also associated with anger
  • Lie
  • A statement induces a simulation purported to be
    true that is actually not(i.e., simulation is
    negative in liars simulation, but false in the
    deceiveds)

(P. 602)
17
Disjunction
Attempted reconstruction
Partial recall
Original event
18
Transformations
19
Ad hocCategories
Write a Comment
User Comments (0)
About PowerShow.com