Cognitive Psychology and its Implications -fifth edition -by John Anderson

Back to Website

Pages: -5th version

Chapter 1

Introspection: trained observers report contents of consciousness under carefully controlled situations àstudied by wurnt

àbased on free-association

Empiricism

Empiricism: based on experience/environment

-Wurnt

Nativism

Nativism: inborn stuff

Behaviorism: unconcerned w/ working of mind: only external b/h

Information-processing analysis.

Information-processing analysis: how info is processed in mind

Sternberg paradigm: subjects are shown small # of digits (1-6) àmeasure how long it took them to recognize the digit.

àperceive stimulus (encode it) compare it to the new stimulus àmake decision àgenerate a response

-upwards slope à38 milliseconds longer for each added digit to the # series

-addition of time per unit of search!!!e

  1. speaks of info processing -no speak of neural representation/location in brain
  2. information = symbolic character – comparing symbol 9 to symbol 3. no talk of neural representation of the symbol
  3. computer is sternberg’s metaphor (turin metaphor)
  4. time = critical variable, since info processing takes place in stages

criticism:

  1. relevance: how relevant it is in the real world àoutside the lab
  2. sufficiency: the sternberg paradigm deals w/ simple tasks, and not complex tasks. àother theories, called cognitive architectures: describe how complete cognitive system function
  3. necessity: the sternberg paradism can predict data, but it doesn’t describe how the brain actually processes.


Cognitive neuroscience: how the brain works, to cognitive processes.

Neuron: a cell which accumulates/transmits electric pulses

-100 Billion in brain àmost processing = through interaction

Nerve cell

-Soma: name for the cell body of the nerve

-Dendrite: arms leading the impulse to the nerve

-Axons: leading impulse away from the nerves

-Axon hillock: the little hill where the axon starts

-Myelin sheath: covers some axons

-Synapse: where an axon and dendrite meet

-Axon terminal: end of axon, which releases the neurotransmitter

-Neurotransmitter: chemicals released by the axon which act on the dendrite membrane to change its action potential



-average soma/dendrite has 1000 synapses, and each axon synapses to about 1000 neurons

Action potential: when positive sodium ions enter the nerve momentarily making it more positive than the outside

-info in brain – 1 out of 2 quantities

  1. membrane potential: could be more or less negative
  2. rate of firing: # of impulses –not pattern’ -greater rate of firing àmore infl on synapting cells




Medulla: controls breathing/swallowing/digestion/heart breat

Cerebellum: motor coordination/voluntary movements

Thalamus: relay station for motoric/sensory info from lower areas to cortex

Hypothalamus: expressions for basic drives

Limbic system: border areas b/w cortec and lower structures

àcontains hippocampus

hippocampus: vital to memory

cerebral cortex/neo-cortex: external layer of brain -amount of wrinkles àthe diff b/w humans and lower mammal.

Neocortex

Crossed over –right hemisphere controls the left side of body/vise versa

4 parts of neocortex



-52 Brodman areas of brain

-hemispheres are somewhat specialized


-corpus callasum: connects b/w hemispheres

àsplit-brain patients: to avoid seizures, corpus callosum is cut

broca’s area (grammatical area) and wernickle’s area (meaning of words) 2 areas in left area critical for speech

àdamage to them: aphasia: severe impairment of speech

Topographical organization

Topographic organization: info processing is organized spatially

ài.e. adjacent area of retina processed adjacently in brain

-some areas are size-proportionally over-represented in brain –more sensitive areas.

Electroencephalography (EEG)

-records electrical potential of scalp

àmeasures large groups of active neurons

àshows waves

Event-related potential (ERP):

-averaged responses to a specific stimulus

àgood temporal resolution, but hard to tell what area of brain is working

-2 techniques which are good in location but not in temporal (timing) of the neural activity

  1. PET (position emission topography): injecting harmless radioactive stuff to blood and see its effects in brain
  2. fMRI: (functional magnetic resonance imaging): better spatial resolution than PET/less intrusive

Connectionism

Connectionism: the method of connecting various processes of brain to build models of how brain deals w/ higher level thinking, like judging people

àexample: parallel distributed processing (PDP)

parallel distributed processing:

example:

-When list of people w/ various groups of classifications, there is a central area of memory which remembers a series of connections to specific choices in the various variables.

example

I want to retrieve info about Jim: I have to connect the variable about:

ài.e. what a certain group of people are like

Chapter 2 - perception

vision

Visual agnosia: inability to recognise objects, and not because of intellectual or sensory loss.

2 groups of visual agnosia:


-->the kind of visual agnosia depends on how high the level of damage to the brain is.

Cornea: bends the light when light first reaches outside of eye

Aqueous humor: liquid inside pupil -->b/w cornea and pupil

pupil: the hole to the eyeball: stretched by the Iris

Iris: muscle increasing or decreasing he pupil - has the color of Eye

Lens further bends the light

Retina: the sheat of photoreceptors

Fovea: center of retinal -->not exactly where optic nerve leaves eye

Bitreous humor: liquid inside the eyeball

Cones: for day

Rods: for night

-Both optic nerves meet at optic chiasm

-->there, nasal part of retina (peripheral/sides of plain of view) crosses over and outside part of retina (center/nasal part of sight) doesn't cross over

2 stations along the way:

Lateral geniculate nucleus -details/recognizing objects -'what'

Superior colliculus -localization of objects in space -'where'

-->they connect to cortical visual areas (i.e. V1)

Ganglion cells/lateral geniculate nucleus: either on-off or off-on cells (on-center =off-sides or off-center=on-sides)

In primary cortical cells: 2 kinds of responses:

Edge detectors: respond positively to light at 1 side of a line but negatively to another side of a line

-->probably made from 2 rows of basal ganglios together -one side reacting negatively while 1 positively (see p.43)

Bar detectors: respond positively to light at center of the straight area (streight/long/narrow strips) and negatively to the outsides

-->probably made from row of single ganglion cells: all of them having center on while ouside off meands that this cell is stimulated: there is a line!! (see p. 43)

-->see p. 42 for bar/edge

-both ege/bar detectors each respond to:

* specific part of visual plain

* specific orientation

* specific width

-the bar/edge detectors are probably lines up in a hyporcolumn in the primary visual cortex

Feature maps: spatial representations of field of ight in each area of the brain which processes the diff. aspects of sight (i.e. movement/color/form)

Depth/surfaces

-we see a 2-D image in the retina and we have to make it 3-D, so we infer the depth

* texture gradient: more tightly packed: further

* stereopsis: the diff. b/w the imge seen by each eye

* motion parallax: further things move slower than fast things: can even be noticed while slightly moving head or eyes!

question:

how does brain process

2 1/2-D sketch: a drawing made to identify various visual features to the viewer

3-D model: representation of objects in a scene.

Object perception:

-brain uses rules to make objects out of visual stimulus. those rules appear to be innate.

gestalt principles of organization: principles of the gestalt school which the brain organizes stimulus into objects -->see p. 47-8

* proximity: when there is a pattern of lets say, lines, then the closer lines are assumed to be together

* good continuum: lines crossing each other assued to be contining straight and not having the sharp turn.

* closure: when object is assumed to be an object since it is closed

* overlap: when 1 object occludes another, i.e. circle over another, almost full circle, bottom object is assumed to b also a circle, even though it could be another shape.

-harder to perceive things contrary to gestalt rules: i.e.: ThIsSenTeNcEiShArDtOrEaD.

Aqueous humor: liquid inside pupil -->b/w cornea and pupil

Visual-Matching Models

Template Matching: brain compares visually perceived pattern to various stored patterns

-problems could occur when there is a diff b/w sight atd template's:


-->though humans have great flexibility and can recognize many variation of the template (i.e. can compare diff. shapes/orientations/sizes/places along pattern)

Feature analysis: stimuli is a combination pf elemental features.

-i.e. regardless of size/shape, the visual stimulus corresponding to the letter A is thought of as a combination of / \ and -

-->kind of like edge/bar detectors

-when we see an unclear stimulus, we extract or combine features to make complete objects

Object recognition: recognizing object categories to recognize more complex objects:

* Recognition-by-components theory:

  1. brain segments object to basic sub-objects (in early visual processing)
  2. Classify the objects (using Geons: [Geometric ion] - see p. 54)

-->biederman - 36 basic geons -->each could vary in size/proportion

3) having identified pieces making up the object, the brain recognizes the pattarnsd on

geons as the object

-nly edges need to recognize ions -->not color/texture/small details

-->geons - just like feature analysis of the letters

Biederman/Beiring/Ju/Blickle study 1985

componant deletion: when a componant is deleted from the drawing of an object.

midsegment deletion: the lines of the drawinf are dotted, and not continuous:

-study shown: with faster flashes of drawings, drawings w/ midsegment deletion is easier to recognize, while when the flashes slow down a bit, then they b/c just as easy -and maybe component deletion a bit easier to detect (the real object)

Speech recognition

-hard to segment word àit sounds like the words are segmented, but it is really 1 long stream of words

-intraword segmentation is also hard

àidentification of phonemes: sound unit i.e. what syllables are made of:bat = [b], [a] , [t]

identification problems:

1) segmentation of phonemes

àunlike written words, spoken words have phonemes run into each other, making it hard to identify

àindividual diff. even w/I the same dialect/accent

2) coarticulation:

-the word is made by 1 movement of the vocal tract (making it harder to identify)

-speech perception is in the left temporal lobe

feature analysis of speech

-feature-analysis and feature-combination seems to underlie speech perception, just like visual recognition

features of phonemes include:


Example: p/b/t/d à4 consonants


-those features are used to recognize the phonemes

àmore confusion w/ more diff. in features:

-the diff. b/w P and T is only in place of articulation/ P and D is voicing and place of articulation. Therefore more mistakes b/w P and T than P and D.

Acoustic diff:

-voiced phoneme (i.e. B) is immediate sounded w/ release of lips, whereas voiceless phonemes (i.e. P) is heard a few milliseconds later)

àtiming of sound seems to be the cue as to the diff! (i.e. b/w B and P)

categorical perception: perceiving groups of stimulus w/o gradient b/w them. àhard to see diff. w/I the group

context/pattern recognition

bottom-up: perception using the stimulus àpattern

top-down: processing using high-level general knowledge to interpret the low level-perceptual units àespecially used in ambiguous cases, like then the H looks like an A in the word the àthe prior knowledge is that there is a word THE but no word TAE

àcontext

top-down: can also be seen if lets say, each 3rd letter is replaced by an x

word superiority effect: if letter or word is flashed fast, and then 2 alternatives were given as to what was flashed, the words guesses would have less mistakes. I.e. D vs. K or WORD vs. WORM

Context and speech

Phoneme-restoration effect: filling in missing phonemes with phonemes according to context, in an automatic fashion (kinds of top-down processing)

-* was the brief beep instead of the sound:

-same thing happens with other over-learnt stimulus – i.e. in faces w/ missing features.

àvery few features are needed to recognize a drawn face

Massaro’s FLMP model of combination of context and feature information

-4 quadrants , each having a series of words… gradient from e to c. In each quadrant, there is a varying amounts of contextual info. Quadrants ranging from:

  1. only e made the word
  2. only c made the word
  3. both made a word
  4. non made a word

-as one went down the quadrant, looked increasingly like the possible word

logical method of perception

àfuzzy logical model of perception

-since there are 2 sources of info (stimulus/context), they are probabilistically different.

probabilistically different: evidence for features of the stimulus is independent of the context

PDP model of letter Recognition

McClelland/Rumelhart

-depends heavily on nerve system: the features (nodes) of letters are activated is seen and inhibited if not present in the letters. Then the letters excite the words an inhibits the word which doesn’t contain the given letter. Words can also activate letters (downwards) to support its recognition. Recognition of the word inhibits other words. According to this model, the word superiority effect is b/c the word supports is component letters.

Conclusion

Light energyàfeature extractionàprimal sketch(initial sense of info) àdepth infoà2 ½D sketchàgestalt principles of organizationà3-D sketchàfeature combination/contextual infoàRECOGNITION

Chapter 3

Parallelism: using diff systems for diff. tasks

àlimited amount possible àoften, after a while = serial bottleneck

ài.e. in motoric acts, only 1 can be done

serial bottleneck: the point where it is no longer possible to process diff. things in parallel

Early selection/Late selection

-models explaining when in the stimulus processing the bottleneck is

auditory attention

dichotic selection task: task revolving listening to 2 ears

àthus processing 2 things w/ same system

-very little is processed from unattended ear

-i.e.:


Filter theory –broadbent

-filter selection: early selection theory based on physical characteristics

ài.e. listening only to L. ear

àinfo is processed until bottleneck is reached

hearing: temporal cortex, in Sylvain Fissure

ànear 41/42 of Brodmann’s areas

àshows enhanced response to listened-to ear

àin cases where we choose by of physical features (i.e. which ear)

àthough we can also filter semantic contents

ài.e. hearing our name

example:

-we can jump /w ears to make out a sentence

Attenuation theory/Late selection theory

Attenuation theory: b/c of physical properties, some messages weakened, but not eliminated.

Late selection theory

-info processed unattenuated

àprocessing system can process both. The problem is reacting to both.

àwe decide which to react to, either based on semantics or on ear of

-semantic selection/detection = easier in this theory than in attenuation theory


-unattenuated stimuli can be remembered, but just for brief moments

àechoic memory

Visual attention

-visual acuity is highest in the small part of the visual field (fovea)

-->by moving eye, we fixate on what we want to give maximum attention to.

spotlight metaphor

-the visual attention could vary in angles it spans. narrowing its span gives maximum processing, yet when there is something in other parts of the visual field, beyond spotlight (i.e. an unexpected thing) it takes time to respond/process it.

Experiment -Leberge

-participant told to focus on center of figure (in a way that figure would be in the fovea) -shown fast figures of ++7++ and asked whether it was a 7 or distracters, (T or Z). -->reacted faster if 7 is in the middle than n sides (i.e. +7+++)

-->conclusion: attention spanned differentially across the visual field

-in complex visual situations, we need to shift our visual attention

-experiment -Neisser/Becklen

-2 TV shows superimposed on each other: one of a hand-slapping game, one of a ball game

-->participants could monitor one and note odd events (like players shaking hands) yet could not monitor both without difficulty/missing great deal of critical events.

note: both physical and content cues were used:

1) physical: moving eyes so main thing will fall on fovea

2) content: using the content to know where to move eye to

Neural basis of visual attention

-just like auditory attention: in visual attention visual attention directed to a spatial location enhances cortical sign from that location

-->i.e. in higher-order feature (i.e. attending to chairs and not tables), the response didn't last for more than 200ms.

conclusion:

-we can assume that people select in a 'early selection model' -->on basis on physical properties, especially on the basis of location

Visual sensory memory

-if given array of items to remember, we would need to look at 1 location at a time

-->i.e. if given a series of letters arranged in rows: need to l letter at a time to remember them

Experiment:

-card with letters flashed at participant: they could remember a few, but noted that there were more that they didn't manage to process

Experiment #2 - sperling

-tone was given right after row was shown (high for 1st row, medium for middle, low for bottom row) -subjects asked to remember 1 row, while tone came right after the display.

àremembered more, since the tone helped them focus.

Partial-Report procedure: subjects were asked to report partial series of the whole amount, which was asked to remember

Whole-report procedure: subjects were asked to recall everything that was asked of them to remember.

Visual sensory store: a memory that can effectively hold all the info in the visual display

-Called iconic memory, like there is echoic memory in auditory system

Iconic memory: brief memory of visual field after it has been seen. if not processed fast, it would be forgotten

Sperling paradigm: how fast attention can process/encode visual iconic memory.

Pattern recognition and Attention

Feature-integration theory: people focus on a stimuli b/f they can process in into a pattern.


example: participants asked to find a T in a pool o 30 Y’s and I’s. they subject need only find the cross-bar of the T which distinguishes it from all the I’s and the Y’s

àtook them longer to find a T in a pool of Z’s and I’s

àless features to distinguish among.

-in stimuli outside focus of attention, we mix up features (i.e. pink T, yellow S and green F à many mistakes =, include pink S)

-people can attend to specific features. Some claim that people can focus on more than feature at a time, but the more specific features = more noisier/less accurate

Neglect of the visual field:

-visual attention to a specific location results in enhanced accticity in the right portion primary visual cortex

-shift in attention really takes place in the posterior parietal lobe/mid brain area called pulvinar.

Neglect: problem shifting attention to a certain receptive field

Unilateral visual neglect: more serious kind of neglect where the side is ignored

àmight not shave 1 side of his face

Object-Based Attention

-sometimes, easier to focus on object than on spatial location

ài.e. in comparing 2 icons of 2 object figures – easier to compare the diff. on same obj, than on the other obj, even though the 2 differing parts were closer in space

àconclusion: easier to focus on obj. than on spatial features

another object based attention technique

Inhibition of return: harder to return to look at particular region of space once we looked there: we are inborn to have searching techniques for objects!!!!!

-some people have neglect to 1 side of visual field

-some have neglect to 1 side of objects

A central bottleneck

We know that each modality has a bottleneck, but, so do 2 tasks of diff. 2 modalities

-SOA (stimulus onset asynchrony):how much time b/w the 2 stimulus

àshorter SOA = more interference of 1st task to 2nd task

-Therefore, one can say that some tasks -->can not be done totally in parallel, and must finish 1 task b/f starting the next

-one can do 2 tasks at once, if only 1 is being thought about (and the other one is automated.

-->i.e. driving is automated, speaking is thought about

Automaticity

automaticity: when you practice - you reduce central cognitive component until it is automatic.

3 processes in automated b/h

  1. perceptual systems
  2. central cognition
  3. motor systems

-central cognition = often the bottleneck, since it is required for coordination.

àAutomaticity eliminates need for central cognition, thus reducing bottleneck

-using paralel motor/perceptual systems = no conflict

stroop effect

stroop effect: the effect of seeing words and recognizing them immediately

stroop experiment:

-with the colored words: much faster answered color words, where their ink color matched than when didn’t match. also many more errors in naming the ink color àhard to inhibit the word, even when asked to name the ink and not the word name of the color.

àshows the automated element of reading.

warner/Polansky Experiment

-Rows w/ diff. # of digits of #.

àeasier to say the digit name rather than the # of digits

ànumber recognition is much more automated than counting.

Chapter 4 –perception-based knowledge representations

Knowledge-representation: proposals of how diff. types of info are encoded and processed

Dual-code theory

Dual-code theory: we remember things better if we assign a visual image to the verbal thing you want to remember (link visual/audio for enhanced memory)

Comparisons of verbal vs. visual processing

  1. Santa’s experiment showed a cue-card w/ 3 shapes. Then asked people to say if the next cue card they’re shown is the same shapes (even if diff. place on card) -->vs. other shapes

-->reacted to shapes faster than words of those letters

-->b/c in words, they are encoded left-right/top-bottom (linear order)

  1. while rehearsing a word jingle or how to get from your house to friends house

-->same mechanisms for internal and external processing of stimuli

Nature of knowledge representations

-knowledge of brain localization shows how processing works

-prob: instead of understanding the connection, we see one unscientific homunculus being who also sees and hears internal things

Visual and Spatial Imagery

Mental Images: representations of things (not visual imagery)

Mental Rotation: same image, but diff rotation (2-D Representation images if 3D Items)

àmore angles of rotation àmore reaction time

2 Kinds of Rotation

  1. object on page – turn the page (picture plane)
  2. object into the page – Depth

à3D Rep. is in th ebraib

àneural mental process – analegeous to physical rotation

image scanning: we look through an object (i.e. house) mentally to find its critical info (i.e. # of windows)

2 experiments

Brooks

  1. had to go through the lines of f and say ‘yes/no’ to whether it was top or bottem extreme or not.
  1. you have a sentence in memory while answering if each word is a noun or not

àB/c already using that part of brain

àreally b/c of spatial inference

Conclusion: interference in stimuli if conflicting special process

Comparison of visual quantities

-when asked to mentally compare visual info (i.e. obj)

àless diff. b/w obj. = more time diff.

àsame w/ real objects

àmental/perception processing/discrimination/comparison

2 types of imaging

  1. spatial: independent of visual aspects
  1. visual/experiancial (i.e. color): specific visual details = in the temporal lobe








àalso true to switch to diff. images of anambigous image àthough mentally hard

hierarchical structure of images:

-people break an image into specific subpart most of the time

àat each level, the unit is called a chunk

àsee p.122-123

-also hierarchical organization of special images

ài.e. people broke up a room into idiosyncratic sub-areas to remember what is in it

cognitive maps


Map distortions

-When map is congruent – i.e. border curves

àmore mistakes than no border, or straight border

Serial-order encoding

-for example, the alphabet, using the alphabet song

a b/ c d/ e f/ g /h i/ j k /l m /n o/ p/etc…

\/ \/ \/ \/ \/

0 0 0 0 0

\/ \/ \/

0 0 0

\/

0

Note: bigger pause when asked next letter when it is in the next chunk!!!

Chapter 5 – meaning-based knowledge representations

Chapter4 – knowledge representation retaining the original event, b/c of perception.

Meaning-based knowledge representations: representations that retain the meaning, but not the exact detailed perception

Verbal

-In verbal sentences, we have a tendency to extract the meaning out of what we hear, and forget the style/specific words

àwe do not remember specific wording styles unless especially cued to do so

àyet styles memory is weaker than meaning memory

-testing warned vs. unwarned about recallingàstylistic diff vs. no meaning diff

Visual

-Sometimes, visual memory is stronger than verbal memory

study:


-studies show that they really remember the meaning of the picture, not the full pic:

i.e. study:

-8 pics. In a series. Then shown other pictures

conclusion: meaning is of the essence!!!

Retention of detail vs. meaning

-evidence that first, perceptual detail is perceived, then when image is interpreted for meaning, the perceptual detailed are quickly forgotten.

Study: picture shown, then reversed and shown

Conclusion: memory for details [i.e. phrasing of sentence] is available initially but then quickly forgotten whereas memory for meaning is retained

Implication:

Mnemonic technique: making a meaningful imagery of some sort to enhance memory of something hard to remember

Prepositional representations

Prepositional representations: the meaningful structure remains after the perception details are gone

-->preposition: smallest unit of independently standing assertion.

i.e.: Lincon, who was president of the USA during a bitter war, freed the slaves

  1. Lincon was president of the USA during a war
  2. The war was bitter
  3. Lincon freed the slaves

-each of those simpler sentences must be true so that the whole sentence is true.

propositions about how we notate them in memory:

  1. relation: organizes (i.e. verb) the arguments: (details
  1. (bitter, war)
  2. (free, Lincoln, slaves)

study shows: we’re can remember which ones we encounter, we’re insensitive to actual combination of propositions

-people might remember the propositions that they encounter but not the combination of propositions, which means that they might add a proposition that they really didn’t hear.

propositional networks:

propositional network: how the propositions are tied into each other.

USA



object


war time agent

Lincoln



relation agent

president of object

slaves


Subject relation



Relation freed

bitter

‘Lincoln who was president during a bitter war, freed the slaves’

nodes: the propositions, relationships, arguments of the sentence

Links: the arrows, b/c they connect the nodes

-could be bigger networks(see p. 150 imp!)

conceptual knowledge

-significant abstraction away from the original experience

àmade from general categories from specific experiences

-important for things like prediction: if I say dog, the other person knows exactly what I am talking about… saves words.

How are categories formed?

2 theories:

  1. semantic networks (similar to propositional networks):
  2. Schemas:

Semantic networks:

isa links: linking nodes for 2 categories:


Animal Has skin

Can move around

Eats

Breath

Bird has wings Fish

can fly

Has feathers




Canary can sing ostrich long/thin legs

is yellow Tall

Rules

1) more frequently encountered fact àit will be stored, even if it belongs to a more superordinate concept

  1. more frequently encountered fact about a concept is, the more strongly the fact will be associated w/ the concept. more associated fact/concept àfaster verification

3) facts that are not directly stored w/I a concept = takes more time

-> more isa links = more reaction time

i.e. cannery - can sing = less time than canary has skin

schemas

schemas: representational structure of a thing

House

isa: building

parts: rooms

Materials: woods/brick/stone

Function: human dwelling

Shape: rectilinear/triangle

size: 100-10,00 sq. feet

Slots: attribute variables: i.e. size/function/etc…

propositions: what specific things have whereas schemas: what things generally have in common.

àisa slot: points to the superset: to the thing above it

Schemas can also be generalized: its parts could also be said to part of the superset

-we can understand constraints (i.e. house underground has no windows) or exceptions (house w/o roof)

psychological reality of schema

-object is understood to have same characteristics of its schema unless specifically contradicted.

àmemory is enhances and even sometimes distorted to incl. things in schema

degree of category membership

-schemas allow for variation w/I the group

-even w/o default values àcan have a schema àthough much disagrement

ài.e. not clear whether stoke/happiness is a disease.

àhard to define disease!

àmany disagreements.


Events schemas:

Scripts: stereotypical things that are supposed to occur/be done

-->people recall parts of a story that really didn’t occur, but was part of their personal script for that event

-->if order of script oddly reversed, then when recalling story, high tendency to recall them in right order -->shows strong tendency of the scripts

-->b/c of schemas, some events are left out in stories, yet assumed to have occoured (like writing a check.

Abstraction theory vs. instance theory

Abstraction theories: make an abstract concept out of the specific cases/experiences

-->i.e. schemas -->prototype = the generalized concept

Instance theories: remember a specific case(s) w/ generalizations made from it -->no specific concept -->comparing the specific case to the paradigm case

Learning Schemas in a Neural Network

-4 symptoms gives, for people to judge them for 2 hypothetical diseases – 1 is 3 times more common than the other

  1. how many input (pieces of knowledge) X 2) output neurons (possible categories)= 3) synaptic associations being learnt (actual category used)


4 X 2 = 8 synapses being learns

-->more common = higher probability, since more synaptic connections

Categories in the brain

-temporal lobe damage = recognition of specific things = related to their subcategories

chapter 6

ebbinghaus: taught himself a series of nonsense consonant-vowel-consonant combinations

àhow long it would take him to learn 13 syllables like that twice w/o error












Percent of time saved ↑ / days of practice à

Rise and fall of short-term memory

Short term memory: the kind of memory where the iconic/auditory memory enters, in affended to.

àif rehearsed, it could go into the relatively permanent LT memory

Memory span: capacity of the specific kind of memory (ST/LT)

àusually 7/8 digits

-evidence of ST memory as processing station for LT memory (beyond rehearsal) = more time is ST memory = more LT memory

-others argue = not time in ST memory but rather depth of processing:

à depth of processing: only If rehearsed in a meaningful way

problem:

evidence:that ST is not necessarily station to LT àinfo could go directly from sensory input to LT àas long as we process the info in a way conducive to creating a Long term trace

(experiment: passive rehearsal for 2/6/18 secs àshow no relationship to time 11/7/13%)

--

-some speak of a memory span or working memory

others (badderly) speaks of a articulatory loop: we keep however much info we can rehearse in a fixed amount of time (usually 1.5-2 secs.)

  1. wit/sum/harm/bay/top
  2. university/opportunity/aluminum/constitutional/auditorium

-both have 5 words, but # 2 longer to say: more % of mistakes

badderly

we have

  1. phonological loop i.e. articulatory loop (see above) ànot ST memory but system for keeping info available.
  2. Visuospatial sketchpad
  3. Central executive: where they interact àit can take/put info into any one of those 2 slave systems or translate from one to the other!

i.e. if given Question 37 X 28 àyou might visually do it or verbally say it. Central command keeps things like task (i.e. multiplication)

frontal cortex: plays a role in working memory.

i.e. if monkey shown 2 options: one of them, food is hidden. Then eyes covered for 10 secs. then monkey found the food. If frontal cortex lesions: couldn’t find it.

àmonkey’s brains are not identical in areas but close

Activation/LT memory

Theories (similar to each other)

ACT = memory traces b/c active w/ associated concepts are presented

SAM = using contextual cues used to make images (i.e. memory traces) more or less familiar (active)

Activation: determines probability of access and rates of access

2 factors:

  1. Level of activation
  2. Amount of practice

àlittle difference w/ short time-delay effect b/w practice and non-practice, but w/ long-term delay, the more practice = faster retrieval

spreading activation

Spreading activation: the idea that activation spreads along paths of a network

ài.e. dog connected to the concept of bone

Therefore, priming: the setting into awareness of a certain connection/network b/w things

i.e.

- if said


-whereas control group: (no priming)


associative priming: priming of knowledge that occurs unconsciously.

i.e. faster to recognize that what is seen is a meaningful word if: the 2 given words are related, rather than non-related or 1 nonsense

Practice and strength

-activation determines how accessible that memory is: activation level could change quickly (w/ milliseconds)

àTraces could also lose activation

Strength: each time we use a memory trace, it increases a little in its strength.

àmore gradual than activation

àrate of improvement decrease (in % improvement) w/ time

Power function: amount of practice is raised to a power

Power law of learning: memory performance improves as a function of practice àin a logarithmic rate

àlinear decrease in response time

Long term potentiation

long-term potentiation: more responsiveness of a neuron b/c of bast stimulation

àoccurs in hippocampus/cortical areas

àmore practice àdiminishing increase (% change upwards)

àLTP = neural learning.

Factors infl. learning:

-not only what kind of practice but how we process it.

ài.e. depth of processing/looking for meaning.

Elaborative processing

Elaborative processing: adding additional info in order to remember better.

i.e. if given either ‘doctor hater the lawyer’ or ‘doctor hated the lawyer b/c of a malpractice suit’

àlater if asked doctor hated ___’ more likely to remember ‘lawyer’ b/c of additional info in memory being a cue.

note: closer to info: better elaboration:

i.e. fat man saw the sign: 1) the was 2 feet tall’ 2) that said ‘be careful, ice’ - #2 is stronger àmore directly connected)

techniques for studying textual material:

-if you have a pre-set cue –i.e. questions that force you to think through the text you’re reading

àbetter learning

PQ4R Technique

  1. Preview: find the general topics of the chapter
  2. Questions: make up questions about each subsection (using the headings as bases for question)
  3. Read: the chapter carefully in order to answer the given questions
  4. Reflect: as you’re reading the text – think of examples/relate info to prior knowledge
  5. Recite: the given info. if you can’t –reread the trouble areas
  6. Review: go through chapter once you’ve finished it. try to answer the questions that you made up. if you can’t answer them – reread trouble areas

àthis technique helps b/c it sets prior organiziors (of info) in mind

study: question making = most learning process àthen, 2nd best is reading to answer questions àthen just reading

àreading w/ q. in mind àbetter than answering q. afterwards.

conclusion of idea: 2 things help learning

  1. question generating
  2. question answering

meaningful vs. non-meaningful elaborations

-more elaborate processing will result in better learning even if learning less meaningful material

i.e. study:

-more memory of the upside down sentence: more processing, regardless of meaningfulness

study 2:

  1. if one generated a word to pair another: leanrt it better that if just read a given pair
  2. synonym is remembered better than rhyme àstill some room for meaning in the elaboration, though the elaborate processing is most important.

Incidental vs. intentional learning

-level of processing and not the intent (or lack of intent) decides the learning.

i.e. if people processed the meaning of words: more learning than just checked for a specific letter

àthough in intentional learning: person might actively take more steps in order to achieve learning

Flashbulb Memories/Self-reference Effect

Flashbulb memory: something special/traumatic/etc… that happened to you, or to something that you feel related to àthat you remember for long time in great/accurate detail

reference effect: people rememborer more information that refers to themselves. (also for people that ne feels close to

-memory for verbal material is assocated w/ greater activation on the L. frontal cortex/L. hippocampal region

Note: until we spoke about issues of encoding àin memory, we’ve also got to deal w/ retention/retreival


chapter 7

electrical stimulation of temporal lobe: brought forgotten, that normally couldn’t be retrieved, i.e. memories of childhood: portions of the neural network that spreading activation couldn’t reach.

-in forgotten memories àsomething is still retained.

-.i.e. in experiment of remembering pairs of #/word

conclusion: even when people think that they forgot something, sensitive tests can fid evidence of some of those memories

Retention function

-with delay, we can recognize less and less a thing we saw

ànegatively accelerated: the improvement % slows down

ài.e. first day, improve 50%, then next 40% etc…

Power law of forgetting: retention function is generally power functions àmore delay, more forgetting, but at a decreasing rate

àlearning also has a power law, but of practice function w/ diminished improvement w/ practice. Retention function shows diminished loss w/ delay

-In ST memory, first few sec. most memory, then fast decrease until stable àdorsn’t really prove existance of ST memory b/c all retention functions are similar in that regard. (whether seconds or days)

-the decay in strength of memory (level of LTP) =in strength of synapses

Bahrick experiment:

-memory decays w/ time in a consistent pattern (regardless of h.m. encoding/depth of processing). Then steadies off. small decline at end (b/c old age)

conclusion: strength of memory trace decays as a power of a power function of the retention interval

Decay theory: as seen above, the theory which claims that memory decays through decaying of neural networks

-competitor to that theory: interference theory

Interference theory

-another thing which strongly impacts memory is interference: interfering of material

fan effect

-when we give people a stimulus (i.e. the word cat), activation spreads to all of its associated concepts.

àmore associated items àless activation to each part

experiment: give people 4 items to remember:

  1. unique person in a unique place (1-1)
  2. unique person in a place common to another (1-2)
  3. person appearing in 2 locations in a unique place (2-1)
  4. person appearing in 2 location in a place common to another (2-2)

-more facts about a person = more delay in recognition

àmore time to spread activation to the whole network

-(see bottom of p 209)

i.e. lawyer in the park

  1. presentation of the probe will activate the presentation of the sources a) lawyer, b) in c) park
  2. paths go from those sources to the nodes (ore nodes àless activation spreads to each one of them)
  3. those activations will converge in the propositional nodes, giving an overall propositional activation
  4. the proposition will be recognized in an amount of time inversely related to level of activation

conclusion: more facts associated w/ a concept = slower retrieval of any of those facts

Interference w/ pre-experimental memories

-people give:

  1. things true in the experiment
  2. true facts about a real life thing
  3. things false both in real life and in the experiment

-Fastest reaction to the true facts (#2). Second fastest reaction is to the experimental true facts (#1) and then slowest response time to false memories (#3)

conclusion: material learnt outside lab can be interfered with by material learnt in the lab

Interference and decay

2 things we saw that produce forgetting:

  1. decay of trace strength
  2. interference from other memories

-some think that decay is really interference

àb/c people remember more is the delay is sleep = less interference, but still same amount of time of delay to allow the decay

àbut really the key is time of day in which the learning took place:

-learn at night/evening

àat highest arousal time

conclusion: also decay and also interference takes place

Interference and redundancy

experiment:

3 groups: each given something else to remember:

  1. 1 fact
  2. many facts about a person, yet each fact being unconnected to each other (i.e. back trouble/hated kids)
  3. Many related/relevant facts about the person (i.e. Mozart went to Paris b/c of romance/improved musically in Paris) àrelated through the place àwhat/why in Paris

Recall speed (frm fastest to slowest:

  1. Irrelevant facts (#2)
  2. Single facts (#1)
  3. Relevant facts (#3)

Conclusion: learning redundant info about something is not interference and may even help memory

Retrieval and inference

-i.e. in the irrelevant facts case: if they could not retrieve 1 of the facts, they had a chance of retrieving other unrelated facts about the person/thing/concept

àmake inferences

experiment:

  1. group pair of sentences that have meaning diff.
  2. pother group given sentence and another only implying what was said in the 1st sentence… they thought that the implying sentence is the real sentence

inference as a bias for remembering text

-if given a text, and then the same text w/ another label 9i.e. another name) people well infer diff. things based on knowledge of the specific person

example: story of a stubborn kid who at age 8, the parents couldn’t find an appropriate mental institution

-evidence for increase increasing inference w/ decaying memory

àpeople do the inferences during the test ài.e. if told that the person was really Helen Keller at time of test

àstill made the errors

conclusion: people use what they remember to infer what might have been part of the text

Plausible retrieval

Plausible retrieval-we retrieve info that is likely meant yet is not explicit

example

-manager of a fries store. he loved them. he thought that they were delicious –he got them for free.

Assumption: he got them for free from the store!

subjects given: incorrect, 2) plausible 3) explicitly written(exact recall) sentence

-with time, exact recall sentences had longer reaction time since it got weaker, whereas plausible sentence had less reaction time:

other study

-more facts about a person = less reaction to a plausible statement, yet longer reaction time to explicit fact about him àdoesn’t depend on retrieving a certain fact.

conclusion: people try to judge plausibly rather than retrieve exact facts

Interaction of elaboration and inferential reconstruction


-elaborate processing doesn’t only increase memory but also memory of inferred things.

-i.e. story about Dr. who, after the checkup, tells college student that what he thought was confirmed

-we expect the in the intro group to have more inferences about the inferred theme

-those inferences that are really not in the text are not errors!. i.e. in real life (i.e. exam) we expect them to remember those kinds of inferences as if it was read material!

Advertisers also capitalize on those inferences.

Memory Errors

-sometimes, it is logical to infer, but sometimes it is important not to, i.e. in eyewitnes testimony

False memory syndrome: many false memories induced by psychotherapist (i.e. of past sexual abuse)

àsometimes, hard to distinguish b/w memory and imagination

conclusion: serious errors of memory could occur b/c people fail to separate actual experience w/ inferred/imagines stuff.

Associative structure and retrieval

-sometimes, a blocked memory comes along when we give it prompts.

study: given animal names. if paired w/ the group it is from (i.e. mammal), more recall

organization and recall

-giving a mechanism for cueing individual pieces of info (i.e. organization)

ài.e. remembering things like lists/orders into a hierarchy helps

Minerals


Metals stones











Rare common alloys Precious Masonry

platinum aluminum bronze Sapphire Limestone

silver copper steel Emerald granite

gold lead brass Diamond marble

Method of Loci: using a fixed sequence of locations to cue retrieval of memories


-other ways of hierarchy organization also help

Effects of Encoding context

-context acts as a cue since it is recorded into memory- trace that records the memory

i.e. physical context/mood context

-i.e. what teacher/where

àstill, not always is contextual info saves… depends on h.m. degree of contextual cues is integrated into memory

mood congruence: easier to remember happy memories in happy state/sad memories in sad states

state-dependant learning: easier to recall info is return to physical/emotional state that they learnt the info in.

i.e. alcoholics can’t remember where they hid the alcohol, when they’re drunk

--.yet it has a debilitating effect

conclusion

  1. people show better memory if external/internal contexts are matched
  2. intoxication usually has a debilitating memory effect


effects of other material on the context

to-be-learned material infl. by other ‘to-be-learned’ info

ài.e. if told to remember the 2nd word, and then told ‘sky- blue’, remembered more often when the word ‘sky’ was in the context of the other learned word in the pair. (even when only tested for the 2nd word

steps of experiment showing contextual cues as helpers to learn:

  1. Original study: told 2 words, while only tested on 2nd word
  2. Generate/Recognize: subjects given a word and asked them to generate related words to b/c the remembering pair (i.e. white-snow). then had to indicate which of the 4 words was the once chosen: got is 54% of time
  3. cued recall: given original context words: many words not recognized when given as free associations (beforehand)

-recognition is stronger than recall. can’t recall if you can’t recognize it.

Encoding-specificity principle: better recall on test if similarity of encoding to original encoding

conclusion: better word memory if words are tested in the context of same word that they were studied in

Hippocampal formation and amnesia

Amnesia: memory loss, due to brain damage àin hippocampal formation

-hippocampal formation: critical for memory formations

àdamage to it = prob. learning associations: especially of diff. elements

àhippocampal formation: in temporal lobe

reasons for hippocampal problems

  1. surgery (i.e. b/c of severe epilepsy
  2. blow to head
  3. alcoholism (i.e. korsakoff syndrome)
  4. brain infection

2 kinds of amnesia

  1. retrograde amnesia: total forgetting of old things
  2. Anterograde amnesia: inability to learn new things

-sometimes blow to head = not a permanent damage:

àbut still has a set pattern of recovering

Set pattern:

1)distorted memory until 2 yrs. bf. accident àthen total retrograde amnesia (RA) until coma) after accident total anterograde amnesia (AA)

  1. patchy memory 4 yrs b/f accident (partial RA). total RT 1 yr b/f the accident. 3 months of total AA. then, a few memories recalled
  2. Final: normal memmory until 2 weeks b/f accident – total RA and total AA 3.5months after. then memory is precise again

-RA and AA recover together, even though 1 might be more severe than the other

interesting to note:

  1. memory deficit is never complete àalsways some form of learning
  2. hippocampal critical in the new memories àold memories are in the cerebral cortex. then while still relatively fresh àmemory needs hippocampal support b/f b/c independent on the cerebral cortex.
  3. AA can retain memories for a few secs: can remember the guy’s name and use it a few times b/f he forgets it


Implicit vs. explicit memory

Dissociated memory: memory that are conscious

implicit memory: knowledge that can’t be described: keyboard

dissociations: the contrast/diff. b/w explicit and implicit memory ài.e. keyboard typing is one example

àmemory tests showed diff. results = therefore must be the diff. kinds of memory

-Amnesiacs show total dissociation: might be able to learn a list of words àcan’t recall them, but could put them in onto a word-completion task, but are unable to gain conscious access to it why they did it.

conclusion

-implicit ways to show that amnesiacs have some sort of memory for events

implicit vs. explicit in normal subjects

-certain variables infl. diff: implicit/explicit memory

i.e. study: word learning

-no context (of other word) àcontext àgeneration of context

imp! Implicit memory decreased w/ elaboration while, as expected, explicit memory increased w/ memory

àpriming: mostly in implicit, non-context situations: we rely most on perceptual encoding

-no diff. in recognizing meaningful vs. non-meaningful words

àwhether explicit or subliminal (implicit)

conclusion

-elaborative processing facilitates explicit memories but not implicit memories

procedural memory

procedural memory: knowledge of how to perform certain acts

study: amnesiacs had to keep switching the # of workers in order to maintain sugar output b/w 8000-10,000

àcan learn how to implicitly control the fctory w/o explicit knowledge

requant distinction:

Declarative knowledge: explicit knowledge that we can report/consciously aware of

procedural knowledge: knowledge of how to do things. usually explicit

conclusion: peple can dev. good ways for doing tasks w/o explaining what they are doing

Chapter 9 –expertise

expertise: could only be achieved w/ practice

  1. more practice = better (no pain, no gain)
  2. as probs. b/c more complicated, the diff. between novice and relative expert increase


general characteristics of skill acquisition

3 stages of skill acquisition

  1. cognitive stage: declarative encoding: I consciously know the info but have not internalized in procedurally: I have to think about it as I do it: i.e. learning to shift gears: I declaratively know where the gears are and I have to think about when/how to use themàslows down use
  2. Associative stage: errors are gradually detected and eliminated. (i.e. coordination of release of clutch w/ gas, in order not to kill the engine)àprocedural knowledge, w/ errors (still some declarative knowledge might be around ài.e. I can speak a language fluently, yet think consciously of some of its rules)
  3. Autonomous stage: utomated/rapid: few attentional resources ài.e. while driving, can have a conversation, yet not remember any of the traffic they experienced

2 elements:

  1. speed
  2. accuracy

power law of learning

-we saw in chapter 6 àretreival of simple associations improve according to a power law

àso does complex associations, like driving.

logarithmic scale

  1. the more decrease in time à more practice needed
  2. more time needed to get more experience/practiced àto produce more actions

-retention of learnt skill is retained even w/ years of abstinence

àmight need a short warm-up

conclusion

-performance of a cognitive skill improves as a power function of practice. only shows modest declines over long retention intervals.

Nature of Expertise

-many expertise studies work on comparing people at diff. levels of then same domain of expertise (i.e. medical students vs. stage vs. experienced dr.)

proceduralization

proceduralization: the switch from declarative o procedural knowledge

-degree of relying on declarative vs. procedural

1) processing:


  1. speeding up
  2. steps being processed/matched up in a single step and not in parts of the whole

Tactical Learning

Tactical Learning: learning a series/string of operations

àspeeding up = change from computation to retreival

-at first àprefrontal areas (anterior cingulate gyrus area àa medial frontal area àmedial =in interior of cortex)àthen shifts w/ better learning t/w back areas of frontal areas. at first cerebellum [motor learning], later hippocampus [learning]

strategic learning

strategic learning: learning to organize one’s problemsolving techniques

-at first backwards thinking: first thinking about the goals/subgoals, and then the stratagy to get each one of the steps.

-later àforwards thinking: thinking about what is missing: already able to not think of each missing/each subgoal

àgood for things like geometry/physics but not for things like programming, where one needs to go through backwards reasoning all the time

àbreath for experts/depth for novices

Conclusion: each domain ha sits own way of acquiring problem-solving techniques

Problem perception

Experts: -richer perceptual features for encoding problems

ài.e. use deeper categorizations of problems/more definitive rules (oi.e. X ray = not tumor, but collapsed lunge

Chess

-experts are just as good in seeing ahead as novices àthey just already remember chunks of moves and therefore can foresee many more probs.

conclusion

experts can recognize chunks in problems which are patterns of elements that repest over problems

long term memory and expertise

-also LT memory improved w/ expertise ànot only Working memory


note: this is b/c expertise helps chunk info into more convenient chunks

Deliberate practice

-necessary for acquiring expertise in anything

Ericsson: deliberate practice: motivation to learn – not just to perform: monitor performance to correct performance and try to reduce deviance

àtallet is peripheral to practice!!!

Transfer of skills


-very little transfer of skills to similar domains and virtually non to diff. domains [i.e. spellingàmath]

ài.e. street vending kids had hard time doing paper-work math, yet could figure out costs in head

Theory of identical Elements

Theory of identical elements: brain is not train ed by formal doctrines of general facilities, but rather specific habits/associations

àprovides narrow responses for specific stimuli

àThorndike: there is transference of skills of similar stimuli-response elements

prob: it is not the stimuli but the abstract concepts which allow for some transference (i.e. calculus for geometry/economic prob: as long as there are common facts!!!)

ànot on surface level, but things w/ same logical structure do have transfer of skills!

Negative transfer: 1 skill inhibits another àvery rare

àexists in memory but not in skills

1 skills example: using a calculator makes you forget a algebra shortcut

Educational implications

Componential analysis: teach the subparts of the whole in order for the student to see the whole

Mastery knowledge: makes sure the student knows how to masker all the subcomponents

Conclusion: get student to identify the underlying knowledge elements and get student to master them all

Intelligent tutoring systems

Intelligent tutoring system: using computers to interact w/ student in order to tutor them (i.e. LISP)

Chapter 10

Logic – formal study of what it takes for an argument to be correct.

-intuitively, we use logic if we have enough info

ài.e. if we know that pressing 15 on the microwave gives 15 sec. of heating, then pressing 30 gives 30 secs. of heating.

Kinds of logical reasoning

Deductive reasoning: certain conclusions based on their premises

Inductive reasoning: probabilistically stemming from their premise

deductive

Fred = Lisa’s bro/Mary = Lisa’s daughter ètherefore, freed is Mary’s uncle

Conditional statement: an if a statement

-If A is true, than B

Antecedent -the if part

Consequent: the then statement èwhat is conditional on the antecedent

Modus ponens: inference logic: if A = B and A is rue, than B follows

    1. if A understood the book, she’s get a good grade
    2. she understood the book:

modus tollens: A implies B àtherefore if B is false, then A is also false

    1. if A understood the book, she’s get a good grade
    2. she got a good grade

-those conclusions are not compelling, but aren’t treated as certain either

conditional syllogisms: if A happens than B. A does [not] happen therefore X happens

[doesn’t happen]

sideways U = sign of implication A U B = A implies b. if ~A

~ = negation sign

Mistakes

Denial of antecedent: if A U B and if A is not true àwe wrongly think that can know B

àB is not necessarily but sufficiently dependant on A

Affirmation of the consequent: we assume that we can know the consequent (A) based on the antecedent (b)


sources of these mistakes:

-Assumption of Biconditional

àBiconditional: if and only if

ànecessary and sufficient

-other reason we choose the way we do, in logical statements = probability

Wason selection task

-cards showing symbols

-condition: if vowel = even # on other side

-had to turn over card to show if rule = correct

àselection task

Cards: E/K/ 4/7


conclusion: Wason card selection could be explained if we assume that people chose cards thy thought were informative w/ probabilistic model

Permission schema of the conditional

Permission schema: connective if statement doesn’t have to be logical/probable statement but also an ought statement

ài.e. if he is drinking beer, he must be over 19

note: must be familiar w/ the rule: i.e. therefore know what/why it is there. If person is unfamiliar w/ the rule, or its rationale, he will not know how to apply it.

àExperiments might also give rationale behind the rule.

Permission schema

-Sometimes we use ‘if’ to say what has ought to be.

àif he is drinking, then he must be over 19’

àlogical use of permission schema is based on familiarity w/ the concept of permission [vs. a senseless rule] and not w/ specific rule

conclusion: in detecting social contract was violated, àlarge proportion of logically correct choices

conclusion:

-if has many diff. interpretations.

for example:

  1. probabilistic
  2. logicians

-training in logic doesn’t necessarily make people us it!

reasoning about quantifiers

logical quantifiers: i.e. ‘all’ or ‘some’

-sometimes when we say ‘all’ we mean most/probably

à’all philosophers read books’

categorical syllogisms (early logic research)

All A = All B

All B = All C

Therefore, All A = C

Some A = Some B

Some B = Some C

Therefore Some A= some C

(truth: Therefore some A doesn’t necessarily mean some C)

àOften, people accept false syllogisms, like #2

Atmosphere hypothesis

-we’re not totally indiscriminant!

Atmosphere hypothesis: the logical terms (some/all/no/not) create the answering tendency ->


Conclusion: people are biased to accept conclusions w/ same quantifiers as their premises

Process explanation

Mental model theory-people make mental images of the situation and sometimes fail to consider alternate images that might actually be more reflective of the situation

Inductive reasoning

Inductive reasoning: to try to get most probably ànot certain

Bayes’s Theorem: combining the different probabilities to come up with an assumption (using mathematical formula)

Prior probability: less likely the hypothesis was b/f the evidence, so too, it will be less likely after the evidence (i.e. b/fàlow statistical chances of being burgled b/f the event, so too, chances are low now)àalternate is higher!

Conditional Probability: type of evidence is true if the hypothesis is true; i.e. if I believe that if the door is open (i.e. evidence) then high chances of being burgled (hypothesis)

Posterior probability: the probability that the hypothesis is true, after considering the evidence.

Prescriptive model: (normative model) – evaluating the probability of a hypothesis

Descriptive model: what people actually do

Base rate neglect

-people don’t take prior probability in account enough

à paying too much attention to the evidence

Conservatism: paying too little attention to the evidence

Correspondence to Bayes’s Theorem w/ experience

probability matching: people choose among alternatives according to success.

People’s processing of probabilities correspond to Bayes’s theorem, their experience-based b/h doesn’t no

Judgments of probabilities


-despite normative value, ($ [value] X odds) people choose the bigger odds, yet smaller values

subjective utility:

value









àtakes double the amount to double the utility ($ x odds)

conclusion: people make decisions under uncertainty in terms of subjective utilities and subjective probabilities

framing effects

framing effects: people’s b/h changes according to what they think the utility curve

i.e. a discount from $125 to $120 is less worth a discount from $15 to $10

chapter 11 -linguistics

linguistics: attempt to characterize nature of language

Productivity: the fact that in any language –infinite # of utterances

Regularity: systematic usage of the language utterances

àvery few combo of words really acceptable

Grammar: set rules of linguistic

3 kinds of rules:


linguistic intuition: judgment of language

àimplicit knowledge

àalso deals w/ ambiguity:

ài.e.: ‘they are cooking apples:’

phrase structure: to recognize a sentence




Noun phrase Verb phrase





Article adj. Noun verb phrase









The brave dog saved the drowning child

Ambiguous sentence

Noun phrase Verb phrase



Pronoun Verb Noun Phrase



Aux verb noun



They are cooking apples



Noun phrase verb phrase




Pronoun verb adj noun




They are cooking apples


Rewrite rules: node’s labels àtheir categories

àcould be used to rewrite sentences

Pauses structures in speech: intuitive use meaningful phrasing units

àtendency to use smallest bundler of words

speech errors

-tendency to correct whole phrase when error is detected

àpoint: sentence is a psych reality

àonly a phrase of the sentence would be repeated as correct

spooner: mixed up sounds to make them funny

àusually happening w/I a phrase

examples:





Transformation: moving elements w/I a sentence (from normal position) in more complex sentences

àsome restrictions apply

relationship b/w language and thought

behaviorist proposal:

Watson: everything is just a conditioned response

whorfian hypothesis of language determinism

language determinism: language determines how person sees world

whorf: impressed at how diff. languages emphasize their word structure

Color experiment

Conclusion: language helps us communicate ideas àbut does not infl. our thoughts

Is language infl. by thought?

Conclusion: in many ways, the structure of the language does reflect the structure of how our minds process the world

Examples:

  1. thought came b/f language àit is or b/c a tool to communicate
  2. sentence structure: split up to reflects unit ofthought: i.e. “the tall boy/…/” =1 idea
  3. Order of Sov (Subject/Object/Verb)

àconclusion: thought infl. language

modularity of language

modularity:

approach: language and thought are independent

àlanguage functions separately from other cognitions

àused to communicate, yet is not infl. by general cognition

questions

Language acquisition

-grammar = implicit

à10 yr-old knows tons of rules


Issues of rules/case of past tense:

-do they learn a rule (i.e. add ed to make is passed) or just associate the two (kick+ed)?

  1. first: irregulars are thought by the kid to be correct
  2. first –learn to associate word +ed àover-generalize the irregularities (i.e. singed’)

connectionist model:

Inputs: verb àgoes though network layers until right verb output is produced (see p.378)

i.e.: go though

  1. feature of root word
  2. feature of past tense

àphonological rep of past tense

aphasia:


-PET studies of unimpaired regular people also show:


àthough debate whether regulars are learn through rule based or through association base

quality of input

-kids don’t get instruction on their first language/parents regularly don’t correct speech!

Mothers: tend to speak in clear/short sentences and w/ clear/exaggerated intonation

àeven though not all mothers do this, all kids learn!!!

Conclusion: instruction quality is not critical to the learning of speech

-Deaf kids = make up sign language!!!

àconclusion: deaf kids master language w/ little direct instruction, at young age!

critical period for language acquisition

2-11 yrs: easiest to learn

àolder kids (i.e. above 11) learn faster (initially)

àyet lack fine mastering of fine point

àsame w/ 1st languageàdeaf adults learn sign language slower than youths

language universals: chomsky: we have innate knowledge of limits of possibilities of characteristics of language/grammar

àlanguage breaking those rules is unlearnable: i.e. adjectives and nouns separated by too much/hold word for too much!!! (language who’s rules could be learnt = natural language)

Critics of chomsky: maybe it is cognitive prob. and not innate?

Transferring part of sentence: arbitrary constraints or the movement of parts of a phrase

Parameter setting:

  1. some languages: 1st person chooses how to phrase sentence (i.e. Finnish)
  2. some languages allow to drop the pronoun

àlearning the language is setting its unique parameters

àchildhood mistakes = mixing up rules (i.e. dropping pronoun in a non-pronoun language)

Cognitive Textbook

Cognitive Psychology and its Implications -fifth edition -by John Anderson

Back to Website

Pages: -5th version

Chapter 1

Introspection: trained observers report contents of consciousness under carefully controlled situations àstudied by wurnt

àbased on free-association

Empiricism

Empiricism: based on experience/environment

-Wurnt

Nativism

Nativism: inborn stuff

Behaviorism: unconcerned w/ working of mind: only external b/h

Information-processing analysis.

Information-processing analysis: how info is processed in mind

Sternberg paradigm: subjects are shown small # of digits (1-6) àmeasure how long it took them to recognize the digit.

àperceive stimulus (encode it) compare it to the new stimulus àmake decision àgenerate a response

-upwards slope à38 milliseconds longer for each added digit to the # series

-addition of time per unit of search!!!e

  1. speaks of info processing -no speak of neural representation/location in brain
  2. information = symbolic character – comparing symbol 9 to symbol 3. no talk of neural representation of the symbol
  3. computer is sternberg’s metaphor (turin metaphor)
  4. time = critical variable, since info processing takes place in stages

criticism:

  1. relevance: how relevant it is in the real world àoutside the lab
  2. sufficiency: the sternberg paradigm deals w/ simple tasks, and not complex tasks. àother theories, called cognitive architectures: describe how complete cognitive system function
  3. necessity: the sternberg paradism can predict data, but it doesn’t describe how the brain actually processes.


Cognitive neuroscience: how the brain works, to cognitive processes.

Neuron: a cell which accumulates/transmits electric pulses

-100 Billion in brain àmost processing = through interaction

Nerve cell

-Soma: name for the cell body of the nerve

-Dendrite: arms leading the impulse to the nerve

-Axons: leading impulse away from the nerves

-Axon hillock: the little hill where the axon starts

-Myelin sheath: covers some axons

-Synapse: where an axon and dendrite meet

-Axon terminal: end of axon, which releases the neurotransmitter

-Neurotransmitter: chemicals released by the axon which act on the dendrite membrane to change its action potential



-average soma/dendrite has 1000 synapses, and each axon synapses to about 1000 neurons

Action potential: when positive sodium ions enter the nerve momentarily making it more positive than the outside

-info in brain – 1 out of 2 quantities

  1. membrane potential: could be more or less negative
  2. rate of firing: # of impulses –not pattern’ -greater rate of firing àmore infl on synapting cells




Medulla: controls breathing/swallowing/digestion/heart breat

Cerebellum: motor coordination/voluntary movements

Thalamus: relay station for motoric/sensory info from lower areas to cortex

Hypothalamus: expressions for basic drives

Limbic system: border areas b/w cortec and lower structures

àcontains hippocampus

hippocampus: vital to memory

cerebral cortex/neo-cortex: external layer of brain -amount of wrinkles àthe diff b/w humans and lower mammal.

Neocortex

Crossed over –right hemisphere controls the left side of body/vise versa

4 parts of neocortex



-52 Brodman areas of brain

-hemispheres are somewhat specialized


-corpus callasum: connects b/w hemispheres

àsplit-brain patients: to avoid seizures, corpus callosum is cut

broca’s area (grammatical area) and wernickle’s area (meaning of words) 2 areas in left area critical for speech

àdamage to them: aphasia: severe impairment of speech

Topographical organization

Topographic organization: info processing is organized spatially

ài.e. adjacent area of retina processed adjacently in brain

-some areas are size-proportionally over-represented in brain –more sensitive areas.

Electroencephalography (EEG)

-records electrical potential of scalp

àmeasures large groups of active neurons

àshows waves

Event-related potential (ERP):

-averaged responses to a specific stimulus

àgood temporal resolution, but hard to tell what area of brain is working

-2 techniques which are good in location but not in temporal (timing) of the neural activity

  1. PET (position emission topography): injecting harmless radioactive stuff to blood and see its effects in brain
  2. fMRI: (functional magnetic resonance imaging): better spatial resolution than PET/less intrusive

Connectionism

Connectionism: the method of connecting various processes of brain to build models of how brain deals w/ higher level thinking, like judging people

àexample: parallel distributed processing (PDP)

parallel distributed processing:

example:

-When list of people w/ various groups of classifications, there is a central area of memory which remembers a series of connections to specific choices in the various variables.

example

I want to retrieve info about Jim: I have to connect the variable about:

ài.e. what a certain group of people are like

Chapter 2 - perception

vision

Visual agnosia: inability to recognise objects, and not because of intellectual or sensory loss.

2 groups of visual agnosia:


-->the kind of visual agnosia depends on how high the level of damage to the brain is.

Cornea: bends the light when light first reaches outside of eye

Aqueous humor: liquid inside pupil -->b/w cornea and pupil

pupil: the hole to the eyeball: stretched by the Iris

Iris: muscle increasing or decreasing he pupil - has the color of Eye

Lens further bends the light

Retina: the sheat of photoreceptors

Fovea: center of retinal -->not exactly where optic nerve leaves eye

Bitreous humor: liquid inside the eyeball

Cones: for day

Rods: for night

-Both optic nerves meet at optic chiasm

-->there, nasal part of retina (peripheral/sides of plain of view) crosses over and outside part of retina (center/nasal part of sight) doesn't cross over

2 stations along the way:

Lateral geniculate nucleus -details/recognizing objects -'what'

Superior colliculus -localization of objects in space -'where'

-->they connect to cortical visual areas (i.e. V1)

Ganglion cells/lateral geniculate nucleus: either on-off or off-on cells (on-center =off-sides or off-center=on-sides)

In primary cortical cells: 2 kinds of responses:

Edge detectors: respond positively to light at 1 side of a line but negatively to another side of a line

-->probably made from 2 rows of basal ganglios together -one side reacting negatively while 1 positively (see p.43)

Bar detectors: respond positively to light at center of the straight area (streight/long/narrow strips) and negatively to the outsides

-->probably made from row of single ganglion cells: all of them having center on while ouside off meands that this cell is stimulated: there is a line!! (see p. 43)

-->see p. 42 for bar/edge

-both ege/bar detectors each respond to:

* specific part of visual plain

* specific orientation

* specific width

-the bar/edge detectors are probably lines up in a hyporcolumn in the primary visual cortex

Feature maps: spatial representations of field of ight in each area of the brain which processes the diff. aspects of sight (i.e. movement/color/form)

Depth/surfaces

-we see a 2-D image in the retina and we have to make it 3-D, so we infer the depth

* texture gradient: more tightly packed: further

* stereopsis: the diff. b/w the imge seen by each eye

* motion parallax: further things move slower than fast things: can even be noticed while slightly moving head or eyes!

question:

how does brain process

2 1/2-D sketch: a drawing made to identify various visual features to the viewer

3-D model: representation of objects in a scene.

Object perception:

-brain uses rules to make objects out of visual stimulus. those rules appear to be innate.

gestalt principles of organization: principles of the gestalt school which the brain organizes stimulus into objects -->see p. 47-8

* proximity: when there is a pattern of lets say, lines, then the closer lines are assumed to be together

* good continuum: lines crossing each other assued to be contining straight and not having the sharp turn.

* closure: when object is assumed to be an object since it is closed

* overlap: when 1 object occludes another, i.e. circle over another, almost full circle, bottom object is assumed to b also a circle, even though it could be another shape.

-harder to perceive things contrary to gestalt rules: i.e.: ThIsSenTeNcEiShArDtOrEaD.

Aqueous humor: liquid inside pupil -->b/w cornea and pupil

Visual-Matching Models

Template Matching: brain compares visually perceived pattern to various stored patterns

-problems could occur when there is a diff b/w sight atd template's:


-->though humans have great flexibility and can recognize many variation of the template (i.e. can compare diff. shapes/orientations/sizes/places along pattern)

Feature analysis: stimuli is a combination pf elemental features.

-i.e. regardless of size/shape, the visual stimulus corresponding to the letter A is thought of as a combination of / \ and -

-->kind of like edge/bar detectors

-when we see an unclear stimulus, we extract or combine features to make complete objects

Object recognition: recognizing object categories to recognize more complex objects:

* Recognition-by-components theory:

  1. brain segments object to basic sub-objects (in early visual processing)
  2. Classify the objects (using Geons: [Geometric ion] - see p. 54)

-->biederman - 36 basic geons -->each could vary in size/proportion

3) having identified pieces making up the object, the brain recognizes the pattarnsd on

geons as the object

-nly edges need to recognize ions -->not color/texture/small details

-->geons - just like feature analysis of the letters

Biederman/Beiring/Ju/Blickle study 1985

componant deletion: when a componant is deleted from the drawing of an object.

midsegment deletion: the lines of the drawinf are dotted, and not continuous:

-study shown: with faster flashes of drawings, drawings w/ midsegment deletion is easier to recognize, while when the flashes slow down a bit, then they b/c just as easy -and maybe component deletion a bit easier to detect (the real object)

Speech recognition

-hard to segment word àit sounds like the words are segmented, but it is really 1 long stream of words

-intraword segmentation is also hard

àidentification of phonemes: sound unit i.e. what syllables are made of:bat = [b], [a] , [t]

identification problems:

1) segmentation of phonemes

àunlike written words, spoken words have phonemes run into each other, making it hard to identify

àindividual diff. even w/I the same dialect/accent

2) coarticulation:

-the word is made by 1 movement of the vocal tract (making it harder to identify)

-speech perception is in the left temporal lobe

feature analysis of speech

-feature-analysis and feature-combination seems to underlie speech perception, just like visual recognition

features of phonemes include:


Example: p/b/t/d à4 consonants


-those features are used to recognize the phonemes

àmore confusion w/ more diff. in features:

-the diff. b/w P and T is only in place of articulation/ P and D is voicing and place of articulation. Therefore more mistakes b/w P and T than P and D.

Acoustic diff:

-voiced phoneme (i.e. B) is immediate sounded w/ release of lips, whereas voiceless phonemes (i.e. P) is heard a few milliseconds later)

àtiming of sound seems to be the cue as to the diff! (i.e. b/w B and P)

categorical perception: perceiving groups of stimulus w/o gradient b/w them. àhard to see diff. w/I the group

context/pattern recognition

bottom-up: perception using the stimulus àpattern

top-down: processing using high-level general knowledge to interpret the low level-perceptual units àespecially used in ambiguous cases, like then the H looks like an A in the word the àthe prior knowledge is that there is a word THE but no word TAE

àcontext

top-down: can also be seen if lets say, each 3rd letter is replaced by an x

word superiority effect: if letter or word is flashed fast, and then 2 alternatives were given as to what was flashed, the words guesses would have less mistakes. I.e. D vs. K or WORD vs. WORM

Context and speech

Phoneme-restoration effect: filling in missing phonemes with phonemes according to context, in an automatic fashion (kinds of top-down processing)

-* was the brief beep instead of the sound:

-same thing happens with other over-learnt stimulus – i.e. in faces w/ missing features.

àvery few features are needed to recognize a drawn face

Massaro’s FLMP model of combination of context and feature information

-4 quadrants , each having a series of words… gradient from e to c. In each quadrant, there is a varying amounts of contextual info. Quadrants ranging from:

  1. only e made the word
  2. only c made the word
  3. both made a word
  4. non made a word

-as one went down the quadrant, looked increasingly like the possible word

logical method of perception

àfuzzy logical model of perception

-since there are 2 sources of info (stimulus/context), they are probabilistically different.

probabilistically different: evidence for features of the stimulus is independent of the context

PDP model of letter Recognition

McClelland/Rumelhart

-depends heavily on nerve system: the features (nodes) of letters are activated is seen and inhibited if not present in the letters. Then the letters excite the words an inhibits the word which doesn’t contain the given letter. Words can also activate letters (downwards) to support its recognition. Recognition of the word inhibits other words. According to this model, the word superiority effect is b/c the word supports is component letters.

Conclusion

Light energyàfeature extractionàprimal sketch(initial sense of info) àdepth infoà2 ½D sketchàgestalt principles of organizationà3-D sketchàfeature combination/contextual infoàRECOGNITION

Chapter 3

Parallelism: using diff systems for diff. tasks

àlimited amount possible àoften, after a while = serial bottleneck

ài.e. in motoric acts, only 1 can be done

serial bottleneck: the point where it is no longer possible to process diff. things in parallel

Early selection/Late selection

-models explaining when in the stimulus processing the bottleneck is

auditory attention

dichotic selection task: task revolving listening to 2 ears

àthus processing 2 things w/ same system

-very little is processed from unattended ear

-i.e.:


Filter theory –broadbent

-filter selection: early selection theory based on physical characteristics

ài.e. listening only to L. ear

àinfo is processed until bottleneck is reached

hearing: temporal cortex, in Sylvain Fissure

ànear 41/42 of Brodmann’s areas

àshows enhanced response to listened-to ear

àin cases where we choose by of physical features (i.e. which ear)

àthough we can also filter semantic contents

ài.e. hearing our name

example:

-we can jump /w ears to make out a sentence

Attenuation theory/Late selection theory

Attenuation theory: b/c of physical properties, some messages weakened, but not eliminated.

Late selection theory

-info processed unattenuated

àprocessing system can process both. The problem is reacting to both.

àwe decide which to react to, either based on semantics or on ear of

-semantic selection/detection = easier in this theory than in attenuation theory


-unattenuated stimuli can be remembered, but just for brief moments

àechoic memory

Visual attention

-visual acuity is highest in the small part of the visual field (fovea)

-->by moving eye, we fixate on what we want to give maximum attention to.

spotlight metaphor

-the visual attention could vary in angles it spans. narrowing its span gives maximum processing, yet when there is something in other parts of the visual field, beyond spotlight (i.e. an unexpected thing) it takes time to respond/process it.

Experiment -Leberge

-participant told to focus on center of figure (in a way that figure would be in the fovea) -shown fast figures of ++7++ and asked whether it was a 7 or distracters, (T or Z). -->reacted faster if 7 is in the middle than n sides (i.e. +7+++)

-->conclusion: attention spanned differentially across the visual field

-in complex visual situations, we need to shift our visual attention

-experiment -Neisser/Becklen

-2 TV shows superimposed on each other: one of a hand-slapping game, one of a ball game

-->participants could monitor one and note odd events (like players shaking hands) yet could not monitor both without difficulty/missing great deal of critical events.

note: both physical and content cues were used:

1) physical: moving eyes so main thing will fall on fovea

2) content: using the content to know where to move eye to

Neural basis of visual attention

-just like auditory attention: in visual attention visual attention directed to a spatial location enhances cortical sign from that location

-->i.e. in higher-order feature (i.e. attending to chairs and not tables), the response didn't last for more than 200ms.

conclusion:

-we can assume that people select in a 'early selection model' -->on basis on physical properties, especially on the basis of location

Visual sensory memory

-if given array of items to remember, we would need to look at 1 location at a time

-->i.e. if given a series of letters arranged in rows: need to l letter at a time to remember them

Experiment:

-card with letters flashed at participant: they could remember a few, but noted that there were more that they didn't manage to process

Experiment #2 - sperling

-tone was given right after row was shown (high for 1st row, medium for middle, low for bottom row) -subjects asked to remember 1 row, while tone came right after the display.

àremembered more, since the tone helped them focus.

Partial-Report procedure: subjects were asked to report partial series of the whole amount, which was asked to remember

Whole-report procedure: subjects were asked to recall everything that was asked of them to remember.

Visual sensory store: a memory that can effectively hold all the info in the visual display

-Called iconic memory, like there is echoic memory in auditory system

Iconic memory: brief memory of visual field after it has been seen. if not processed fast, it would be forgotten

Sperling paradigm: how fast attention can process/encode visual iconic memory.

Pattern recognition and Attention

Feature-integration theory: people focus on a stimuli b/f they can process in into a pattern.


example: participants asked to find a T in a pool o 30 Y’s and I’s. they subject need only find the cross-bar of the T which distinguishes it from all the I’s and the Y’s

àtook them longer to find a T in a pool of Z’s and I’s

àless features to distinguish among.

-in stimuli outside focus of attention, we mix up features (i.e. pink T, yellow S and green F à many mistakes =, include pink S)

-people can attend to specific features. Some claim that people can focus on more than feature at a time, but the more specific features = more noisier/less accurate

Neglect of the visual field:

-visual attention to a specific location results in enhanced accticity in the right portion primary visual cortex

-shift in attention really takes place in the posterior parietal lobe/mid brain area called pulvinar.

Neglect: problem shifting attention to a certain receptive field

Unilateral visual neglect: more serious kind of neglect where the side is ignored

àmight not shave 1 side of his face

Object-Based Attention

-sometimes, easier to focus on object than on spatial location

ài.e. in comparing 2 icons of 2 object figures – easier to compare the diff. on same obj, than on the other obj, even though the 2 differing parts were closer in space

àconclusion: easier to focus on obj. than on spatial features

another object based attention technique

Inhibition of return: harder to return to look at particular region of space once we looked there: we are inborn to have searching techniques for objects!!!!!

-some people have neglect to 1 side of visual field

-some have neglect to 1 side of objects

A central bottleneck

We know that each modality has a bottleneck, but, so do 2 tasks of diff. 2 modalities

-SOA (stimulus onset asynchrony):how much time b/w the 2 stimulus

àshorter SOA = more interference of 1st task to 2nd task

-Therefore, one can say that some tasks -->can not be done totally in parallel, and must finish 1 task b/f starting the next

-one can do 2 tasks at once, if only 1 is being thought about (and the other one is automated.

-->i.e. driving is automated, speaking is thought about

Automaticity

automaticity: when you practice - you reduce central cognitive component until it is automatic.

3 processes in automated b/h

  1. perceptual systems
  2. central cognition
  3. motor systems

-central cognition = often the bottleneck, since it is required for coordination.

àAutomaticity eliminates need for central cognition, thus reducing bottleneck

-using paralel motor/perceptual systems = no conflict

stroop effect

stroop effect: the effect of seeing words and recognizing them immediately

stroop experiment:

-with the colored words: much faster answered color words, where their ink color matched than when didn’t match. also many more errors in naming the ink color àhard to inhibit the word, even when asked to name the ink and not the word name of the color.

àshows the automated element of reading.

warner/Polansky Experiment

-Rows w/ diff. # of digits of #.

àeasier to say the digit name rather than the # of digits

ànumber recognition is much more automated than counting.

Chapter 4 –perception-based knowledge representations

Knowledge-representation: proposals of how diff. types of info are encoded and processed

Dual-code theory

Dual-code theory: we remember things better if we assign a visual image to the verbal thing you want to remember (link visual/audio for enhanced memory)

Comparisons of verbal vs. visual processing

  1. Santa’s experiment showed a cue-card w/ 3 shapes. Then asked people to say if the next cue card they’re shown is the same shapes (even if diff. place on card) -->vs. other shapes

-->reacted to shapes faster than words of those letters

-->b/c in words, they are encoded left-right/top-bottom (linear order)

  1. while rehearsing a word jingle or how to get from your house to friends house

-->same mechanisms for internal and external processing of stimuli

Nature of knowledge representations

-knowledge of brain localization shows how processing works

-prob: instead of understanding the connection, we see one unscientific homunculus being who also sees and hears internal things

Visual and Spatial Imagery

Mental Images: representations of things (not visual imagery)

Mental Rotation: same image, but diff rotation (2-D Representation images if 3D Items)

àmore angles of rotation àmore reaction time

2 Kinds of Rotation

  1. object on page – turn the page (picture plane)
  2. object into the page – Depth

à3D Rep. is in th ebraib

àneural mental process – analegeous to physical rotation

image scanning: we look through an object (i.e. house) mentally to find its critical info (i.e. # of windows)

2 experiments

Brooks

  1. had to go through the lines of f and say ‘yes/no’ to whether it was top or bottem extreme or not.
  1. you have a sentence in memory while answering if each word is a noun or not

àB/c already using that part of brain

àreally b/c of spatial inference

Conclusion: interference in stimuli if conflicting special process

Comparison of visual quantities

-when asked to mentally compare visual info (i.e. obj)

àless diff. b/w obj. = more time diff.

àsame w/ real objects

àmental/perception processing/discrimination/comparison

2 types of imaging

  1. spatial: independent of visual aspects
  1. visual/experiancial (i.e. color): specific visual details = in the temporal lobe








àalso true to switch to diff. images of anambigous image àthough mentally hard

hierarchical structure of images:

-people break an image into specific subpart most of the time

àat each level, the unit is called a chunk

àsee p.122-123

-also hierarchical organization of special images

ài.e. people broke up a room into idiosyncratic sub-areas to remember what is in it

cognitive maps


Map distortions

-When map is congruent – i.e. border curves

àmore mistakes than no border, or straight border

Serial-order encoding

-for example, the alphabet, using the alphabet song

a b/ c d/ e f/ g /h i/ j k /l m /n o/ p/etc…

\/ \/ \/ \/ \/

0 0 0 0 0

\/ \/ \/

0 0 0

\/

0

Note: bigger pause when asked next letter when it is in the next chunk!!!

Chapter 5 – meaning-based knowledge representations

Chapter4 – knowledge representation retaining the original event, b/c of perception.

Meaning-based knowledge representations: representations that retain the meaning, but not the exact detailed perception

Verbal

-In verbal sentences, we have a tendency to extract the meaning out of what we hear, and forget the style/specific words

àwe do not remember specific wording styles unless especially cued to do so

àyet styles memory is weaker than meaning memory

-testing warned vs. unwarned about recallingàstylistic diff vs. no meaning diff

Visual

-Sometimes, visual memory is stronger than verbal memory

study:


-studies show that they really remember the meaning of the picture, not the full pic:

i.e. study:

-8 pics. In a series. Then shown other pictures

conclusion: meaning is of the essence!!!

Retention of detail vs. meaning

-evidence that first, perceptual detail is perceived, then when image is interpreted for meaning, the perceptual detailed are quickly forgotten.

Study: picture shown, then reversed and shown

Conclusion: memory for details [i.e. phrasing of sentence] is available initially but then quickly forgotten whereas memory for meaning is retained

Implication:

Mnemonic technique: making a meaningful imagery of some sort to enhance memory of something hard to remember

Prepositional representations

Prepositional representations: the meaningful structure remains after the perception details are gone

-->preposition: smallest unit of independently standing assertion.

i.e.: Lincon, who was president of the USA during a bitter war, freed the slaves

  1. Lincon was president of the USA during a war
  2. The war was bitter
  3. Lincon freed the slaves

-each of those simpler sentences must be true so that the whole sentence is true.

propositions about how we notate them in memory:

  1. relation: organizes (i.e. verb) the arguments: (details
  1. (bitter, war)
  2. (free, Lincoln, slaves)

study shows: we’re can remember which ones we encounter, we’re insensitive to actual combination of propositions

-people might remember the propositions that they encounter but not the combination of propositions, which means that they might add a proposition that they really didn’t hear.

propositional networks:

propositional network: how the propositions are tied into each other.

USA



object


war time agent

Lincoln



relation agent

president of object

slaves


Subject relation



Relation freed

bitter

‘Lincoln who was president during a bitter war, freed the slaves’

nodes: the propositions, relationships, arguments of the sentence

Links: the arrows, b/c they connect the nodes

-could be bigger networks(see p. 150 imp!)

conceptual knowledge

-significant abstraction away from the original experience

àmade from general categories from specific experiences

-important for things like prediction: if I say dog, the other person knows exactly what I am talking about… saves words.

How are categories formed?

2 theories:

  1. semantic networks (similar to propositional networks):
  2. Schemas:

Semantic networks:

isa links: linking nodes for 2 categories:


Animal Has skin

Can move around

Eats

Breath

Bird has wings Fish

can fly

Has feathers




Canary can sing ostrich long/thin legs

is yellow Tall

Rules

1) more frequently encountered fact àit will be stored, even if it belongs to a more superordinate concept

  1. more frequently encountered fact about a concept is, the more strongly the fact will be associated w/ the concept. more associated fact/concept àfaster verification

3) facts that are not directly stored w/I a concept = takes more time

-> more isa links = more reaction time

i.e. cannery - can sing = less time than canary has skin

schemas

schemas: representational structure of a thing

House

isa: building

parts: rooms

Materials: woods/brick/stone

Function: human dwelling

Shape: rectilinear/triangle

size: 100-10,00 sq. feet

Slots: attribute variables: i.e. size/function/etc…

propositions: what specific things have whereas schemas: what things generally have in common.

àisa slot: points to the superset: to the thing above it

Schemas can also be generalized: its parts could also be said to part of the superset

-we can understand constraints (i.e. house underground has no windows) or exceptions (house w/o roof)

psychological reality of schema

-object is understood to have same characteristics of its schema unless specifically contradicted.

àmemory is enhances and even sometimes distorted to incl. things in schema

degree of category membership

-schemas allow for variation w/I the group

-even w/o default values àcan have a schema àthough much disagrement

ài.e. not clear whether stoke/happiness is a disease.

àhard to define disease!

àmany disagreements.


Events schemas:

Scripts: stereotypical things that are supposed to occur/be done

-->people recall parts of a story that really didn’t occur, but was part of their personal script for that event

-->if order of script oddly reversed, then when recalling story, high tendency to recall them in right order -->shows strong tendency of the scripts

-->b/c of schemas, some events are left out in stories, yet assumed to have occoured (like writing a check.

Abstraction theory vs. instance theory

Abstraction theories: make an abstract concept out of the specific cases/experiences

-->i.e. schemas -->prototype = the generalized concept

Instance theories: remember a specific case(s) w/ generalizations made from it -->no specific concept -->comparing the specific case to the paradigm case

Learning Schemas in a Neural Network

-4 symptoms gives, for people to judge them for 2 hypothetical diseases – 1 is 3 times more common than the other

  1. how many input (pieces of knowledge) X 2) output neurons (possible categories)= 3) synaptic associations being learnt (actual category used)


4 X 2 = 8 synapses being learns

-->more common = higher probability, since more synaptic connections

Categories in the brain

-temporal lobe damage = recognition of specific things = related to their subcategories

chapter 6

ebbinghaus: taught himself a series of nonsense consonant-vowel-consonant combinations

àhow long it would take him to learn 13 syllables like that twice w/o error












Percent of time saved ↑ / days of practice à

Rise and fall of short-term memory

Short term memory: the kind of memory where the iconic/auditory memory enters, in affended to.

àif rehearsed, it could go into the relatively permanent LT memory

Memory span: capacity of the specific kind of memory (ST/LT)

àusually 7/8 digits

-evidence of ST memory as processing station for LT memory (beyond rehearsal) = more time is ST memory = more LT memory

-others argue = not time in ST memory but rather depth of processing:

à depth of processing: only If rehearsed in a meaningful way

problem:

evidence:that ST is not necessarily station to LT àinfo could go directly from sensory input to LT àas long as we process the info in a way conducive to creating a Long term trace

(experiment: passive rehearsal for 2/6/18 secs àshow no relationship to time 11/7/13%)

--

-some speak of a memory span or working memory

others (badderly) speaks of a articulatory loop: we keep however much info we can rehearse in a fixed amount of time (usually 1.5-2 secs.)

  1. wit/sum/harm/bay/top
  2. university/opportunity/aluminum/constitutional/auditorium

-both have 5 words, but # 2 longer to say: more % of mistakes

badderly

we have

  1. phonological loop i.e. articulatory loop (see above) ànot ST memory but system for keeping info available.
  2. Visuospatial sketchpad
  3. Central executive: where they interact àit can take/put info into any one of those 2 slave systems or translate from one to the other!

i.e. if given Question 37 X 28 àyou might visually do it or verbally say it. Central command keeps things like task (i.e. multiplication)

frontal cortex: plays a role in working memory.

i.e. if monkey shown 2 options: one of them, food is hidden. Then eyes covered for 10 secs. then monkey found the food. If frontal cortex lesions: couldn’t find it.

àmonkey’s brains are not identical in areas but close

Activation/LT memory

Theories (similar to each other)

ACT = memory traces b/c active w/ associated concepts are presented

SAM = using contextual cues used to make images (i.e. memory traces) more or less familiar (active)

Activation: determines probability of access and rates of access

2 factors:

  1. Level of activation
  2. Amount of practice

àlittle difference w/ short time-delay effect b/w practice and non-practice, but w/ long-term delay, the more practice = faster retrieval

spreading activation

Spreading activation: the idea that activation spreads along paths of a network

ài.e. dog connected to the concept of bone

Therefore, priming: the setting into awareness of a certain connection/network b/w things

i.e.

- if said


-whereas control group: (no priming)


associative priming: priming of knowledge that occurs unconsciously.

i.e. faster to recognize that what is seen is a meaningful word if: the 2 given words are related, rather than non-related or 1 nonsense

Practice and strength

-activation determines how accessible that memory is: activation level could change quickly (w/ milliseconds)

àTraces could also lose activation

Strength: each time we use a memory trace, it increases a little in its strength.

àmore gradual than activation

àrate of improvement decrease (in % improvement) w/ time

Power function: amount of practice is raised to a power

Power law of learning: memory performance improves as a function of practice àin a logarithmic rate

àlinear decrease in response time

Long term potentiation

long-term potentiation: more responsiveness of a neuron b/c of bast stimulation

àoccurs in hippocampus/cortical areas

àmore practice àdiminishing increase (% change upwards)

àLTP = neural learning.

Factors infl. learning:

-not only what kind of practice but how we process it.

ài.e. depth of processing/looking for meaning.

Elaborative processing

Elaborative processing: adding additional info in order to remember better.

i.e. if given either ‘doctor hater the lawyer’ or ‘doctor hated the lawyer b/c of a malpractice suit’

àlater if asked doctor hated ___’ more likely to remember ‘lawyer’ b/c of additional info in memory being a cue.

note: closer to info: better elaboration:

i.e. fat man saw the sign: 1) the was 2 feet tall’ 2) that said ‘be careful, ice’ - #2 is stronger àmore directly connected)

techniques for studying textual material:

-if you have a pre-set cue –i.e. questions that force you to think through the text you’re reading

àbetter learning

PQ4R Technique

  1. Preview: find the general topics of the chapter
  2. Questions: make up questions about each subsection (using the headings as bases for question)
  3. Read: the chapter carefully in order to answer the given questions
  4. Reflect: as you’re reading the text – think of examples/relate info to prior knowledge
  5. Recite: the given info. if you can’t –reread the trouble areas
  6. Review: go through chapter once you’ve finished it. try to answer the questions that you made up. if you can’t answer them – reread trouble areas

àthis technique helps b/c it sets prior organiziors (of info) in mind

study: question making = most learning process àthen, 2nd best is reading to answer questions àthen just reading

àreading w/ q. in mind àbetter than answering q. afterwards.

conclusion of idea: 2 things help learning

  1. question generating
  2. question answering

meaningful vs. non-meaningful elaborations

-more elaborate processing will result in better learning even if learning less meaningful material

i.e. study:

-more memory of the upside down sentence: more processing, regardless of meaningfulness

study 2:

  1. if one generated a word to pair another: leanrt it better that if just read a given pair
  2. synonym is remembered better than rhyme àstill some room for meaning in the elaboration, though the elaborate processing is most important.

Incidental vs. intentional learning

-level of processing and not the intent (or lack of intent) decides the learning.

i.e. if people processed the meaning of words: more learning than just checked for a specific letter

àthough in intentional learning: person might actively take more steps in order to achieve learning

Flashbulb Memories/Self-reference Effect

Flashbulb memory: something special/traumatic/etc… that happened to you, or to something that you feel related to àthat you remember for long time in great/accurate detail

reference effect: people rememborer more information that refers to themselves. (also for people that ne feels close to

-memory for verbal material is assocated w/ greater activation on the L. frontal cortex/L. hippocampal region

Note: until we spoke about issues of encoding àin memory, we’ve also got to deal w/ retention/retreival


chapter 7

electrical stimulation of temporal lobe: brought forgotten, that normally couldn’t be retrieved, i.e. memories of childhood: portions of the neural network that spreading activation couldn’t reach.

-in forgotten memories àsomething is still retained.

-.i.e. in experiment of remembering pairs of #/word

conclusion: even when people think that they forgot something, sensitive tests can fid evidence of some of those memories

Retention function

-with delay, we can recognize less and less a thing we saw

ànegatively accelerated: the improvement % slows down

ài.e. first day, improve 50%, then next 40% etc…

Power law of forgetting: retention function is generally power functions àmore delay, more forgetting, but at a decreasing rate

àlearning also has a power law, but of practice function w/ diminished improvement w/ practice. Retention function shows diminished loss w/ delay

-In ST memory, first few sec. most memory, then fast decrease until stable àdorsn’t really prove existance of ST memory b/c all retention functions are similar in that regard. (whether seconds or days)

-the decay in strength of memory (level of LTP) =in strength of synapses

Bahrick experiment:

-memory decays w/ time in a consistent pattern (regardless of h.m. encoding/depth of processing). Then steadies off. small decline at end (b/c old age)

conclusion: strength of memory trace decays as a power of a power function of the retention interval

Decay theory: as seen above, the theory which claims that memory decays through decaying of neural networks

-competitor to that theory: interference theory

Interference theory

-another thing which strongly impacts memory is interference: interfering of material

fan effect

-when we give people a stimulus (i.e. the word cat), activation spreads to all of its associated concepts.

àmore associated items àless activation to each part

experiment: give people 4 items to remember:

  1. unique person in a unique place (1-1)
  2. unique person in a place common to another (1-2)
  3. person appearing in 2 locations in a unique place (2-1)
  4. person appearing in 2 location in a place common to another (2-2)

-more facts about a person = more delay in recognition

àmore time to spread activation to the whole network

-(see bottom of p 209)

i.e. lawyer in the park

  1. presentation of the probe will activate the presentation of the sources a) lawyer, b) in c) park
  2. paths go from those sources to the nodes (ore nodes àless activation spreads to each one of them)
  3. those activations will converge in the propositional nodes, giving an overall propositional activation
  4. the proposition will be recognized in an amount of time inversely related to level of activation

conclusion: more facts associated w/ a concept = slower retrieval of any of those facts

Interference w/ pre-experimental memories

-people give:

  1. things true in the experiment
  2. true facts about a real life thing
  3. things false both in real life and in the experiment

-Fastest reaction to the true facts (#2). Second fastest reaction is to the experimental true facts (#1) and then slowest response time to false memories (#3)

conclusion: material learnt outside lab can be interfered with by material learnt in the lab

Interference and decay

2 things we saw that produce forgetting:

  1. decay of trace strength
  2. interference from other memories

-some think that decay is really interference

àb/c people remember more is the delay is sleep = less interference, but still same amount of time of delay to allow the decay

àbut really the key is time of day in which the learning took place:

-learn at night/evening

àat highest arousal time

conclusion: also decay and also interference takes place

Interference and redundancy

experiment:

3 groups: each given something else to remember:

  1. 1 fact
  2. many facts about a person, yet each fact being unconnected to each other (i.e. back trouble/hated kids)
  3. Many related/relevant facts about the person (i.e. Mozart went to Paris b/c of romance/improved musically in Paris) àrelated through the place àwhat/why in Paris

Recall speed (frm fastest to slowest:

  1. Irrelevant facts (#2)
  2. Single facts (#1)
  3. Relevant facts (#3)

Conclusion: learning redundant info about something is not interference and may even help memory

Retrieval and inference

-i.e. in the irrelevant facts case: if they could not retrieve 1 of the facts, they had a chance of retrieving other unrelated facts about the person/thing/concept

àmake inferences

experiment:

  1. group pair of sentences that have meaning diff.
  2. pother group given sentence and another only implying what was said in the 1st sentence… they thought that the implying sentence is the real sentence

inference as a bias for remembering text

-if given a text, and then the same text w/ another label 9i.e. another name) people well infer diff. things based on knowledge of the specific person

example: story of a stubborn kid who at age 8, the parents couldn’t find an appropriate mental institution

-evidence for increase increasing inference w/ decaying memory

àpeople do the inferences during the test ài.e. if told that the person was really Helen Keller at time of test

àstill made the errors

conclusion: people use what they remember to infer what might have been part of the text

Plausible retrieval

Plausible retrieval-we retrieve info that is likely meant yet is not explicit

example

-manager of a fries store. he loved them. he thought that they were delicious –he got them for free.

Assumption: he got them for free from the store!

subjects given: incorrect, 2) plausible 3) explicitly written(exact recall) sentence

-with time, exact recall sentences had longer reaction time since it got weaker, whereas plausible sentence had less reaction time:

other study

-more facts about a person = less reaction to a plausible statement, yet longer reaction time to explicit fact about him àdoesn’t depend on retrieving a certain fact.

conclusion: people try to judge plausibly rather than retrieve exact facts

Interaction of elaboration and inferential reconstruction


-elaborate processing doesn’t only increase memory but also memory of inferred things.

-i.e. story about Dr. who, after the checkup, tells college student that what he thought was confirmed

-we expect the in the intro group to have more inferences about the inferred theme

-those inferences that are really not in the text are not errors!. i.e. in real life (i.e. exam) we expect them to remember those kinds of inferences as if it was read material!

Advertisers also capitalize on those inferences.

Memory Errors

-sometimes, it is logical to infer, but sometimes it is important not to, i.e. in eyewitnes testimony

False memory syndrome: many false memories induced by psychotherapist (i.e. of past sexual abuse)

àsometimes, hard to distinguish b/w memory and imagination

conclusion: serious errors of memory could occur b/c people fail to separate actual experience w/ inferred/imagines stuff.

Associative structure and retrieval

-sometimes, a blocked memory comes along when we give it prompts.

study: given animal names. if paired w/ the group it is from (i.e. mammal), more recall

organization and recall

-giving a mechanism for cueing individual pieces of info (i.e. organization)

ài.e. remembering things like lists/orders into a hierarchy helps

Minerals


Metals stones











Rare common alloys Precious Masonry

platinum aluminum bronze Sapphire Limestone

silver copper steel Emerald granite

gold lead brass Diamond marble

Method of Loci: using a fixed sequence of locations to cue retrieval of memories


-other ways of hierarchy organization also help

Effects of Encoding context

-context acts as a cue since it is recorded into memory- trace that records the memory

i.e. physical context/mood context

-i.e. what teacher/where

àstill, not always is contextual info saves… depends on h.m. degree of contextual cues is integrated into memory

mood congruence: easier to remember happy memories in happy state/sad memories in sad states

state-dependant learning: easier to recall info is return to physical/emotional state that they learnt the info in.

i.e. alcoholics can’t remember where they hid the alcohol, when they’re drunk

--.yet it has a debilitating effect

conclusion

  1. people show better memory if external/internal contexts are matched
  2. intoxication usually has a debilitating memory effect


effects of other material on the context

to-be-learned material infl. by other ‘to-be-learned’ info

ài.e. if told to remember the 2nd word, and then told ‘sky- blue’, remembered more often when the word ‘sky’ was in the context of the other learned word in the pair. (even when only tested for the 2nd word

steps of experiment showing contextual cues as helpers to learn:

  1. Original study: told 2 words, while only tested on 2nd word
  2. Generate/Recognize: subjects given a word and asked them to generate related words to b/c the remembering pair (i.e. white-snow). then had to indicate which of the 4 words was the once chosen: got is 54% of time
  3. cued recall: given original context words: many words not recognized when given as free associations (beforehand)

-recognition is stronger than recall. can’t recall if you can’t recognize it.

Encoding-specificity principle: better recall on test if similarity of encoding to original encoding

conclusion: better word memory if words are tested in the context of same word that they were studied in

Hippocampal formation and amnesia

Amnesia: memory loss, due to brain damage àin hippocampal formation

-hippocampal formation: critical for memory formations

àdamage to it = prob. learning associations: especially of diff. elements

àhippocampal formation: in temporal lobe

reasons for hippocampal problems

  1. surgery (i.e. b/c of severe epilepsy
  2. blow to head
  3. alcoholism (i.e. korsakoff syndrome)
  4. brain infection

2 kinds of amnesia

  1. retrograde amnesia: total forgetting of old things
  2. Anterograde amnesia: inability to learn new things

-sometimes blow to head = not a permanent damage:

àbut still has a set pattern of recovering

Set pattern:

1)distorted memory until 2 yrs. bf. accident àthen total retrograde amnesia (RA) until coma) after accident total anterograde amnesia (AA)

  1. patchy memory 4 yrs b/f accident (partial RA). total RT 1 yr b/f the accident. 3 months of total AA. then, a few memories recalled
  2. Final: normal memmory until 2 weeks b/f accident – total RA and total AA 3.5months after. then memory is precise again

-RA and AA recover together, even though 1 might be more severe than the other

interesting to note:

  1. memory deficit is never complete àalsways some form of learning
  2. hippocampal critical in the new memories àold memories are in the cerebral cortex. then while still relatively fresh àmemory needs hippocampal support b/f b/c independent on the cerebral cortex.
  3. AA can retain memories for a few secs: can remember the guy’s name and use it a few times b/f he forgets it


Implicit vs. explicit memory

Dissociated memory: memory that are conscious

implicit memory: knowledge that can’t be described: keyboard

dissociations: the contrast/diff. b/w explicit and implicit memory ài.e. keyboard typing is one example

àmemory tests showed diff. results = therefore must be the diff. kinds of memory

-Amnesiacs show total dissociation: might be able to learn a list of words àcan’t recall them, but could put them in onto a word-completion task, but are unable to gain conscious access to it why they did it.

conclusion

-implicit ways to show that amnesiacs have some sort of memory for events

implicit vs. explicit in normal subjects

-certain variables infl. diff: implicit/explicit memory

i.e. study: word learning

-no context (of other word) àcontext àgeneration of context

imp! Implicit memory decreased w/ elaboration while, as expected, explicit memory increased w/ memory

àpriming: mostly in implicit, non-context situations: we rely most on perceptual encoding

-no diff. in recognizing meaningful vs. non-meaningful words

àwhether explicit or subliminal (implicit)

conclusion

-elaborative processing facilitates explicit memories but not implicit memories

procedural memory

procedural memory: knowledge of how to perform certain acts

study: amnesiacs had to keep switching the # of workers in order to maintain sugar output b/w 8000-10,000

àcan learn how to implicitly control the fctory w/o explicit knowledge

requant distinction:

Declarative knowledge: explicit knowledge that we can report/consciously aware of

procedural knowledge: knowledge of how to do things. usually explicit

conclusion: peple can dev. good ways for doing tasks w/o explaining what they are doing

Chapter 9 –expertise

expertise: could only be achieved w/ practice

  1. more practice = better (no pain, no gain)
  2. as probs. b/c more complicated, the diff. between novice and relative expert increase


general characteristics of skill acquisition

3 stages of skill acquisition

  1. cognitive stage: declarative encoding: I consciously know the info but have not internalized in procedurally: I have to think about it as I do it: i.e. learning to shift gears: I declaratively know where the gears are and I have to think about when/how to use themàslows down use
  2. Associative stage: errors are gradually detected and eliminated. (i.e. coordination of release of clutch w/ gas, in order not to kill the engine)àprocedural knowledge, w/ errors (still some declarative knowledge might be around ài.e. I can speak a language fluently, yet think consciously of some of its rules)
  3. Autonomous stage: utomated/rapid: few attentional resources ài.e. while driving, can have a conversation, yet not remember any of the traffic they experienced

2 elements:

  1. speed
  2. accuracy

power law of learning

-we saw in chapter 6 àretreival of simple associations improve according to a power law

àso does complex associations, like driving.

logarithmic scale

  1. the more decrease in time à more practice needed
  2. more time needed to get more experience/practiced àto produce more actions

-retention of learnt skill is retained even w/ years of abstinence

àmight need a short warm-up

conclusion

-performance of a cognitive skill improves as a power function of practice. only shows modest declines over long retention intervals.

Nature of Expertise

-many expertise studies work on comparing people at diff. levels of then same domain of expertise (i.e. medical students vs. stage vs. experienced dr.)

proceduralization

proceduralization: the switch from declarative o procedural knowledge

-degree of relying on declarative vs. procedural

1) processing:


  1. speeding up
  2. steps being processed/matched up in a single step and not in parts of the whole

Tactical Learning

Tactical Learning: learning a series/string of operations

àspeeding up = change from computation to retreival

-at first àprefrontal areas (anterior cingulate gyrus area àa medial frontal area àmedial =in interior of cortex)àthen shifts w/ better learning t/w back areas of frontal areas. at first cerebellum [motor learning], later hippocampus [learning]

strategic learning

strategic learning: learning to organize one’s problemsolving techniques

-at first backwards thinking: first thinking about the goals/subgoals, and then the stratagy to get each one of the steps.

-later àforwards thinking: thinking about what is missing: already able to not think of each missing/each subgoal

àgood for things like geometry/physics but not for things like programming, where one needs to go through backwards reasoning all the time

àbreath for experts/depth for novices

Conclusion: each domain ha sits own way of acquiring problem-solving techniques

Problem perception

Experts: -richer perceptual features for encoding problems

ài.e. use deeper categorizations of problems/more definitive rules (oi.e. X ray = not tumor, but collapsed lunge

Chess

-experts are just as good in seeing ahead as novices àthey just already remember chunks of moves and therefore can foresee many more probs.

conclusion

experts can recognize chunks in problems which are patterns of elements that repest over problems

long term memory and expertise

-also LT memory improved w/ expertise ànot only Working memory


note: this is b/c expertise helps chunk info into more convenient chunks

Deliberate practice

-necessary for acquiring expertise in anything

Ericsson: deliberate practice: motivation to learn – not just to perform: monitor performance to correct performance and try to reduce deviance

àtallet is peripheral to practice!!!

Transfer of skills


-very little transfer of skills to similar domains and virtually non to diff. domains [i.e. spellingàmath]

ài.e. street vending kids had hard time doing paper-work math, yet could figure out costs in head

Theory of identical Elements

Theory of identical elements: brain is not train ed by formal doctrines of general facilities, but rather specific habits/associations

àprovides narrow responses for specific stimuli

àThorndike: there is transference of skills of similar stimuli-response elements

prob: it is not the stimuli but the abstract concepts which allow for some transference (i.e. calculus for geometry/economic prob: as long as there are common facts!!!)

ànot on surface level, but things w/ same logical structure do have transfer of skills!

Negative transfer: 1 skill inhibits another àvery rare

àexists in memory but not in skills

1 skills example: using a calculator makes you forget a algebra shortcut

Educational implications

Componential analysis: teach the subparts of the whole in order for the student to see the whole

Mastery knowledge: makes sure the student knows how to masker all the subcomponents

Conclusion: get student to identify the underlying knowledge elements and get student to master them all

Intelligent tutoring systems

Intelligent tutoring system: using computers to interact w/ student in order to tutor them (i.e. LISP)

Chapter 10

Logic – formal study of what it takes for an argument to be correct.

-intuitively, we use logic if we have enough info

ài.e. if we know that pressing 15 on the microwave gives 15 sec. of heating, then pressing 30 gives 30 secs. of heating.

Kinds of logical reasoning

Deductive reasoning: certain conclusions based on their premises

Inductive reasoning: probabilistically stemming from their premise

deductive

Fred = Lisa’s bro/Mary = Lisa’s daughter ètherefore, freed is Mary’s uncle

Conditional statement: an if a statement

-If A is true, than B

Antecedent -the if part

Consequent: the then statement èwhat is conditional on the antecedent

Modus ponens: inference logic: if A = B and A is rue, than B follows

    1. if A understood the book, she’s get a good grade
    2. she understood the book:

modus tollens: A implies B àtherefore if B is false, then A is also false

    1. if A understood the book, she’s get a good grade
    2. she got a good grade

-those conclusions are not compelling, but aren’t treated as certain either

conditional syllogisms: if A happens than B. A does [not] happen therefore X happens

[doesn’t happen]

sideways U = sign of implication A U B = A implies b. if ~A

~ = negation sign

Mistakes

Denial of antecedent: if A U B and if A is not true àwe wrongly think that can know B

àB is not necessarily but sufficiently dependant on A

Affirmation of the consequent: we assume that we can know the consequent (A) based on the antecedent (b)


sources of these mistakes:

-Assumption of Biconditional

àBiconditional: if and only if

ànecessary and sufficient

-other reason we choose the way we do, in logical statements = probability

Wason selection task

-cards showing symbols

-condition: if vowel = even # on other side

-had to turn over card to show if rule = correct

àselection task

Cards: E/K/ 4/7


conclusion: Wason card selection could be explained if we assume that people chose cards thy thought were informative w/ probabilistic model

Permission schema of the conditional

Permission schema: connective if statement doesn’t have to be logical/probable statement but also an ought statement

ài.e. if he is drinking beer, he must be over 19

note: must be familiar w/ the rule: i.e. therefore know what/why it is there. If person is unfamiliar w/ the rule, or its rationale, he will not know how to apply it.

àExperiments might also give rationale behind the rule.

Permission schema

-Sometimes we use ‘if’ to say what has ought to be.

àif he is drinking, then he must be over 19’

àlogical use of permission schema is based on familiarity w/ the concept of permission [vs. a senseless rule] and not w/ specific rule

conclusion: in detecting social contract was violated, àlarge proportion of logically correct choices

conclusion:

-if has many diff. interpretations.

for example:

  1. probabilistic
  2. logicians

-training in logic doesn’t necessarily make people us it!

reasoning about quantifiers

logical quantifiers: i.e. ‘all’ or ‘some’

-sometimes when we say ‘all’ we mean most/probably

à’all philosophers read books’

categorical syllogisms (early logic research)

All A = All B

All B = All C

Therefore, All A = C

Some A = Some B

Some B = Some C

Therefore Some A= some C

(truth: Therefore some A doesn’t necessarily mean some C)

àOften, people accept false syllogisms, like #2

Atmosphere hypothesis

-we’re not totally indiscriminant!

Atmosphere hypothesis: the logical terms (some/all/no/not) create the answering tendency ->


Conclusion: people are biased to accept conclusions w/ same quantifiers as their premises

Process explanation

Mental model theory-people make mental images of the situation and sometimes fail to consider alternate images that might actually be more reflective of the situation

Inductive reasoning

Inductive reasoning: to try to get most probably ànot certain

Bayes’s Theorem: combining the different probabilities to come up with an assumption (using mathematical formula)

Prior probability: less likely the hypothesis was b/f the evidence, so too, it will be less likely after the evidence (i.e. b/fàlow statistical chances of being burgled b/f the event, so too, chances are low now)àalternate is higher!

Conditional Probability: type of evidence is true if the hypothesis is true; i.e. if I believe that if the door is open (i.e. evidence) then high chances of being burgled (hypothesis)

Posterior probability: the probability that the hypothesis is true, after considering the evidence.

Prescriptive model: (normative model) – evaluating the probability of a hypothesis

Descriptive model: what people actually do

Base rate neglect

-people don’t take prior probability in account enough

à paying too much attention to the evidence

Conservatism: paying too little attention to the evidence

Correspondence to Bayes’s Theorem w/ experience

probability matching: people choose among alternatives according to success.

People’s processing of probabilities correspond to Bayes’s theorem, their experience-based b/h doesn’t no

Judgments of probabilities


-despite normative value, ($ [value] X odds) people choose the bigger odds, yet smaller values

subjective utility:

value









àtakes double the amount to double the utility ($ x odds)

conclusion: people make decisions under uncertainty in terms of subjective utilities and subjective probabilities

framing effects

framing effects: people’s b/h changes according to what they think the utility curve

i.e. a discount from $125 to $120 is less worth a discount from $15 to $10

chapter 11 -linguistics

linguistics: attempt to characterize nature of language

Productivity: the fact that in any language –infinite # of utterances

Regularity: systematic usage of the language utterances

àvery few combo of words really acceptable

Grammar: set rules of linguistic

3 kinds of rules:


linguistic intuition: judgment of language

àimplicit knowledge

àalso deals w/ ambiguity:

ài.e.: ‘they are cooking apples:’

phrase structure: to recognize a sentence




Noun phrase Verb phrase





Article adj. Noun verb phrase









The brave dog saved the drowning child

Ambiguous sentence

Noun phrase Verb phrase



Pronoun Verb Noun Phrase



Aux verb noun



They are cooking apples



Noun phrase verb phrase




Pronoun verb adj noun




They are cooking apples


Rewrite rules: node’s labels àtheir categories

àcould be used to rewrite sentences

Pauses structures in speech: intuitive use meaningful phrasing units

àtendency to use smallest bundler of words

speech errors

-tendency to correct whole phrase when error is detected

àpoint: sentence is a psych reality

àonly a phrase of the sentence would be repeated as correct

spooner: mixed up sounds to make them funny

àusually happening w/I a phrase

examples:





Transformation: moving elements w/I a sentence (from normal position) in more complex sentences

àsome restrictions apply

relationship b/w language and thought

behaviorist proposal:

Watson: everything is just a conditioned response

whorfian hypothesis of language determinism

language determinism: language determines how person sees world

whorf: impressed at how diff. languages emphasize their word structure

Color experiment

Conclusion: language helps us communicate ideas àbut does not infl. our thoughts

Is language infl. by thought?

Conclusion: in many ways, the structure of the language does reflect the structure of how our minds process the world

Examples:

  1. thought came b/f language àit is or b/c a tool to communicate
  2. sentence structure: split up to reflects unit ofthought: i.e. “the tall boy/…/” =1 idea
  3. Order of Sov (Subject/Object/Verb)

àconclusion: thought infl. language

modularity of language

modularity:

approach: language and thought are independent

àlanguage functions separately from other cognitions

àused to communicate, yet is not infl. by general cognition

questions

Language acquisition

-grammar = implicit

à10 yr-old knows tons of rules


Issues of rules/case of past tense:

-do they learn a rule (i.e. add ed to make is passed) or just associate the two (kick+ed)?

  1. first: irregulars are thought by the kid to be correct
  2. first –learn to associate word +ed àover-generalize the irregularities (i.e. singed’)

connectionist model:

Inputs: verb àgoes though network layers until right verb output is produced (see p.378)

i.e.: go though

  1. feature of root word
  2. feature of past tense

àphonological rep of past tense

aphasia:


-PET studies of unimpaired regular people also show:


àthough debate whether regulars are learn through rule based or through association base

quality of input

-kids don’t get instruction on their first language/parents regularly don’t correct speech!

Mothers: tend to speak in clear/short sentences and w/ clear/exaggerated intonation

àeven though not all mothers do this, all kids learn!!!

Conclusion: instruction quality is not critical to the learning of speech

-Deaf kids = make up sign language!!!

àconclusion: deaf kids master language w/ little direct instruction, at young age!

critical period for language acquisition

2-11 yrs: easiest to learn

àolder kids (i.e. above 11) learn faster (initially)

àyet lack fine mastering of fine point

àsame w/ 1st languageàdeaf adults learn sign language slower than youths

language universals: chomsky: we have innate knowledge of limits of possibilities of characteristics of language/grammar

àlanguage breaking those rules is unlearnable: i.e. adjectives and nouns separated by too much/hold word for too much!!! (language who’s rules could be learnt = natural language)

Critics of chomsky: maybe it is cognitive prob. and not innate?

Transferring part of sentence: arbitrary constraints or the movement of parts of a phrase

Parameter setting:

  1. some languages: 1st person chooses how to phrase sentence (i.e. Finnish)
  2. some languages allow to drop the pronoun

àlearning the language is setting its unique parameters

àchildhood mistakes = mixing up rules (i.e. dropping pronoun in a non-pronoun language)


Locations of visitors to this page