The Nature of Mental Representations How is knowledge represented in the Mind?

1
The Nature of Mental RepresentationsHow is
knowledge represented in the Mind?
2
(No Transcript)
3
Think carefully about the people depicted. What
do they look like? Now compare each of their
voices How do these representations exist in your
mind?
4
The Nature of Mental Representation
What does your representation of a penny look
like?
5
One of the greatest problems remaining is
understanding the nature of mental representation
Knowledge Structures Declarative Memory facts
that can be stated (Knowing that) Procedural
knowledge procedures that can be
implemented(Knowing how) How to study mental
representation Experiments manipulating
represented knowledge Neuropsychology
6
Cognitive MapsTolman 1930s Experiments by
Tolman demonstrated that representations exist

• Group 1 Rats had to learn maze, reward for
  getting from start box to end box was food
• Made more errors than group 1
• Group 2 no reinforcement for getting to end box
• These rats learned the maze without making errors
• Group 3 no reward for 10 days. On 11th day
  food placed in end box for first time.
• With one reinforcement ran maze as well as group
  1

7
What do Tolmans Maps look like?
Edward C. Tolman (1886-1959)
Latent learning goes against standard
behavioristic principles, which claim that
learning comes only from outcomes
8
Cognitive MapsTolman 1930s

• Conclusions Learning a cognitive map
• An internal representation of the maze

9
Cognitive MapsBees vonFrisch 1967

• Studied behavior of bees returning to hive after
  locating nectar
• Can form imaginal maps
• Can use a symbolic form of communication
• Different patterns of dances represent different
  meanings

10
Cognitive MapsBees vonFrisch 1967

• Round dance source less than 100 yards from hive
• Figure 8 dance greater distances

11
The Age-old Debate
Propositional versus Analogical Representation
12
Representational Systems Require An orderly
mapping between the Represented and Representing
worlds A mental representation is a system of
symbols, conscious or unconscious, that are
isomorphic to some aspect of the environment,
used to make behavior-generating decisions that
anticipate events and relations in that
environment
C.R. GallistelRutgers University
13
Words are the images of things
The Greek poet Simonides, c. 500 B.C.
14
How about Book?
15
Theories of Mental Representation
16
Propositional
17
Analogical
?

Hammer
18
Propositional Representations for Circle
A series of points all of which are equidistant
from a central point (Arbitrary code
English, Geometry) 101101101101101010
(Arbitrary code Binary Digit) x2 y2 r2
(Arbitrary code Analytical Geometry)
19
Analogical Representation for Circle
20
Analogical Representations
Thermometers Photographs
21
Cognitive mapsHumans

• Landmark knowledge features imaginal and
  propositional
• Route road knowledge specific pathways
  procedural and declarative knowledge
• Survey knowledge estimated distances between
  landmarks imaginally or propositionally(number)

22
Cognitive mapsHumans
23
Imagery

• Mental representation of things not currently
  being sensed by the sense organs
• Often a recollection of something
• Can imagine things never experienced
• Imagine what it would be like to jump from a
  plane
• Can imagine things that dont exist
• Imagine that Bush wasnt the current president
• Can use any of the sensory modalities
• Most research targets visual imagery

24
KosslynWe use imagery to solve problems

• Which is a darker red An apple or a cherry?
• How many windows in your bedroom?
• How do you get to the subway from here?

25
(No Transcript)
26
(No Transcript)
27
Hypothesis When people rotate a mental image,
the farther they have to rotate it, the longer
they will take.Method Subjects were shown
pairs of two-dimensional drawings of
three-dimensional objects and were asked whether
the drawings in each pair represented the same
object in a different orientation or whether they
represented different objects.Results 97
Accuracy. The time it took the subjects to
provide the answer was proportional to the
angular difference between the two drawings.For
example, if the second image was rotated 60
degrees, the subject would take twice as long
than if it was just rotated 30 degrees.
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
Experiment

• Close eyes and imagine a cube floating in space
  in front of you
• Now Grasp the Left Front bottom corner of the
  cube with your left hand and grasp the right back
  top corner of the cube with your right hand.
• Now Rotate the cube so that the corner in your
  left hand is directly below the corner in your
  right hand
• Describe

32
(No Transcript)
33
Using positron emission tomography (PET), an
axial brain slice taken 56 mm above the anterior
commissureposterior commissure (ACPC) line
shows activation in the primary motor cortex (M1)
when data from the external-action condition were
subtracted from those of the internal-action
condition. Depending on the strategy used, motor
regions of the brain are recruited during mental
rotation. The result also shows that the strategy
used to accomplish a given task can vary
according to previous training, and can be
adopted voluntarily.
34
Other evidence suggesting that representations
are analogical1) Mental images of large
objects (e.g., an elephant) take longer to scan
from left to right than smaller objects2)
Details are more difficult to work out in mental
images of small things (fruit fly) than in large
objects (airplane). For example, Elephant head
to tail last longer than Pig head to Pig
tail.Similar with map experiments.
35
Distance in image is proportional to time spent
scanning
36
Kosslyn 1975

• Imagine
• elephant and a rabbit
• Rabbit and a fly
• Rabbit and elephant sized fly
• Rabbit and a fly sized elephant
• TASK describe details of objects
• RESULTS took longer to describe smaller object
  of pair
• Perceptual resolution better for larger objects

37
Evidence for Analogical Scanning
Stephen Kosslyn (right) and one of his
collaborators
38

• Perception longer to scan longer distance than
  short distances
• Experiment to test imaginal scanning
• Shown a map of an imaginary island with landmarks
• Studied until could reproduce from memory
• Critical phase
• Hear name of an object (tree)
• Picture map, mentally scan directly to object and
  press a key when have arrived to the object
• Hear name of second object, scan to that objects
  location
• Measured RT

39
Image scaling
40
Findings Distance RT are Positively
Related
41
Image scaling

• Spatial representations can be scanned in the
  same way that physical percepts can be scanned
• Our strategies and responses for imaginal
  scanning are expected to be functionally
  equivalent to those we use for perceptual scanning

42
Auditory imaginal representation

• Intons-Peterson et al 1992
• Subjects take longer to mentally shift a sound
  upward in pitch from the low purring of a cat to
  a high pitched ringing of a phone than to switch
  from purr to clock tick.

43
Zenon Pylyshyn
Problem Is it all due to task
instructions?Demand Characteristics
44
Is the mental image of Kosslyns island really
map-like?
What is farther west, Reno or San Diego?
45
(No Transcript)
46
Neuropsychological evidence

• Georgopoulos et al 1989
• Single cell recordings in motor cortex of monkeys
• Monkey trained to physically move a handle
  perpendicular to and counterclockwise to a target
  light
• During rotation cortical activity recorded
• Later take away handle, cortical activity
  recorded
• RESULTS activity in motor cortex showed same
  cortical cells respond when monkeys are
  anticipating the movements of the rotations

47
L.H.Farahs patient

• Visual imagery(color and shape) distinct from
  spatial imagery(depth and distance)
• LH head injury at age 18
• Lesions in left temporo-occipital regions, right
  temporal, right inferior temporal
• Anomia- could see but could not apply verbal
  label
• Bad at describing color and shape imaginal info
• Good with spatial manipulations

48
What processes do we use to represent and
manipulate knowledge?

• How we represent info influences how easily we
  can perform cognitive tasks
• CMLIXLVIII

49
What processes do we use to represent and
manipulate knowledge?

• How we represent info influences how easily we
  can perform cognitive tasks
• CMLIXLVIII
• Easier as 95958

50
Declarative vs. Procedural

• Declarative knowing that
• Procedural knowing how

51

• Write normal signature as quickly as possible
  From first letter of first name to last letter of
  last name.

52

• Now turn over paper
• Write your signature backward starting with last
  letter of last name to first letter of first name

53
Declarative vs. Procedural

• For both signatures had declarative knowledge
  about what comes next to each letter
• Only had procedural for the normal writing of name

54
Concepts

• Idea about something fundamental unit of
  symbolic knowledge
• Relates to other concepts
• Red, round, fruit
• How do we organize concepts
• Categories organization based on common
  features or similarity to a prototype

55
Schemas

• Can include other schemas
• Encompass typical general facts that can vary
  from one instance to another
• Vary in their degree of abstraction
• Often have if, then relationships embedded
• If a glass falls on to a hard surface, then
• It should break

56
Schematic relationships

• Young child calls out mama! after he falls
  from the jungle gym at the playground
• Who is he calling to- a 75 yr old woman, a 45
  year old man, and a 30 year old nun or 25 year
  old woman?
• Utilize schemas to answer this questions

57
Procedural knowledge

• Problem solving
• Improved by practice until becomes automatic
• Others have made a distinction from declarative
  knowledge as non declarative
• Includes perceptual motor and cognitive skills
  procedural
• Simple associative knowledge CC
• Priming

58
Experiment1/2 class close eyes
59

• ZAZIP
• GASPETHIT
• POCH YUSE
• OWCH MINE
• ILCHI
• ACOT

60
Other half of class close eyes
61

• TECKAJ
• STEV
• ASTEREW
• OLACK
• ZELBAR
• ACOT

62

• PIZZA
• SPAGHETTI
• CHOP SUEY
• CHOW MEIN
• CHILI
• TACO OR COAT

• JACKET
• VEST
• SWEATER
• CLOAK
• BLAZER
• TACO OR COAT

63
Declarative/ Procedural differences

• Amnesiacs can learn new skill but cant
  express, cant consciously remember that they
  have been learning it.
• Show improvements in
• skill tasks mirror drawing

64

• One representation is part of non declarative
  knowledge, trying to make it explicit can be very
  difficult

65
Karl S. Lashley and the Problem of Serial Order
66
Musicians play at a rate so fast that feedback
cannot serve as a stimulus for the next note.
Pigeons were taught to learn an arbitrary
sequence of colorsred?green?yellow?blueSucce
ssive Chaining paradigmS(A)R(A)?
S(B)R(B)?S(C)R(C)?S(D)R(D)In successive
chaining, each cue is responded to individually
Serial Order Its Representation
67
No Possibility of Chaining
Simultaneous Chaining ParadigmS(A)S(B)S(C)S(D
)? R(A)? R(B)?R(C)?R(D)All items are presented
simultaneously, and there is no differential
feedback during the execution.
68
(No Transcript)
69
(No Transcript)
70
(No Transcript)
71
(No Transcript)
72
Herb Terrace and one of his students at Columbia
University
73
(No Transcript)
74
Results It takes longer for Rhesus monkeys to
learn a new list if items are repeated in a
different ordinal position.This provides
evidence that Rhesus monkeys internally represent
ordinal position
75
Still, one of the greatest problems confronting
psychology is the nature of mental representation
76
http//www.psych.uni.edu/psychexperiments/Exps/lab
experiments.htm http//www.cog.brown.edu/tarr/sti
muli.html www.sims.berkeley.edu/sack/
CM/detailed-intro.htm