MEMORY | 1

Reading Notes 2135 – Midterm 2

Part 3| Memory

Chapter 6: The Acquisition of Memories and the WM system

Clive Wearing
 Accomplished musician
 When he was 47, herpes virus caused profound amnesia such that he has no episodic memories
 Lives only ‘in the moment’ bc he cant form new memories
 Still loves his wife, still a gr8 musician

Acquisition, storage and retrieval

 Acquisition
o Process of gaining information and placing it into memory
 Storage
o Hold info in memory until it is needed
 Retrieval
o Locate memory
 This method is the way most computers work  info (input) is provided to computer (acquisition phase),
which then goes into dormant form on the hard drive (storage phase) and finally info can be brought back
via search processes that hunts through the disk (retrieval phase)
o Side note: evidence suggests that what is ‘good learning depends on how memory is used later on

The route into memory

 For many years, theorizing in cog psych focused on memory acquisition
 Modal model by Atkinson & Shiffrin = model for this

Updating the modal model

 Information processing view of memory
 1) info arrives, is stored in sensory memory (for a second) – has ‘raw’ iconic and echoic memory
 2) process of selection + interpretation moves the info into short-term memory: where you hold info while
you’re working on it
 3) some info is transferred into long-term memory: a larger and more permanent storage place
o **caveats: modern approach limits action of sensory memory, and changes the term short-term
memory into working memory; WM is not a single storage space, it is a STATUS such that when
ideas are ‘in working memory’ it means they are currently activated and being worked on by a
specific set of operations
Working memory Long term memory
- Limited in size - Huge in size (bc it contains all your semantic
- Getting info into WM is easy (simply must and episodic knowledge)
think about some idea) - Getting info into LTM involves work
- Getting info out of WM is also easy - Getting info out of LTM is also hard bc it is
not linked to current focus of your thoughts

Working memory and long-term memory: one memory or two?

 Free recall: pts are given a list of words, one by one, then after asked to recall as many as they can
o Ppl usually recall 12-15 words; more likely to remember first few and last few

Explaining the recency effect

 The last few words you hear are still in consciousness, therefore in WM and are easier to recall
 Since it has a limited size, it is only capable of holding 5-6 words; any earlier words are pushed out by
these later words
 MEMORY | 2

Explaining the primacy effect

 It takes some work to get info into LTM, so it is likely this requires time and attention
 When pts hear the list of words, they do their best to memorize it (often through memory rehearsal)
o Basically words earlier in the list get more attention, as there is less to try to remember
o Once 6-7 words have been presented, pts need to divide their attention for all these words such
that each word receives only a small fraction of their focus

Testing claims about primacy and recency

 RECALL: recency comes from WORKING MEMORY; primacy comes from LTM
o Therefore modifying each process will cause either memory for recency or memory for primacy
words to be affected
 Use tweaked paradigm for WM: after list is presented to participants, delay recall by asking them to
perform some other task first (eg. immediately count backward by 3’s starting from 200, for 30 seconds)
o Eliminates recency effect!!! Does nothing for primacy effect bc that’s LTM
o Nothing happens if you only delay the recall and have no other activity occupying WM
 Use tweaked paradigm for LTM: slow down the presentation of list (not 1 per second, but longer)
o Pts have to spend longer on all the list items, which increases the likelihood of transfer into more
permanent storage and therefore improving recall for LTM items; no influence on WM
o Other things that influence LTM: familiar or more common words ease entry into LTM, improves
pre-recency retention but has not effect on recency
 fMRI scans show that memory for early items on a list depends on brain areas associated w LTM & HC;
memory for later items do not show this pattern

A closer look at WM
 diffs btw LTM and WM memory include: 1) size of stores 2) ease of entry 3) ease of retrieval 4)
dependency on current activity

The function of WM
 Coordination of several pieces of info is required for all mental activities
 You use WM whenever you have multiple ideas in your mind, multiple elements that you are seeking ot
combine or compare; NOTE: ppl differ in ‘holding capacity’ of their WM

Digit span
 One way to measure WM holding capacity
 People are asked to read a series of digits and must immediately repeat them back; keep being given a list
of different but longer numbers until they can no longer repeat it back correctly (around 7-8 items)
 WM span is usually 7 +/- 2 (aka 5-9)
 ‘items’ depend on how we chunk information
o E.g. given a list of letters, if we chunk the letters into syllables, we can still only remember 7ish
syllables, but many more letters
o E.g. one person who is a fan of track events will chunk 3,4,9,2 into 3 minutes, 49.2 seconds ….
They can therefore keep 7 finishing times in memory, which is 20-30 numbers!
 CAPACITY does not change, SSTRATEGY for chunking does

Operation span
 Modern conception of WM is more dynamic than WM = storage space
 Operation span allows us to test capacity in active operation instead
o Different ways to measure operation span exist, w types differing in what the operation is
o E.g. reading span: pts asked to read aloud a series of sentences and immediately after, the pts is
asked to recall each sentence’s final word
 First asked to read 2 sentences, then 3, then 4… until cannot do this correctly
 Someone w higher span has larger WM
 Consistent w this claim is the idea that ppl w greater WMC show an advantage
for reading comprehension, standardized academic tests, and more
 MEMORY | 3

The working memory system

 WM is not a single entity but it is a SYSTEM built of several components
 Central executive
o In charge of running all your mental operations
 Articulatory rehearsal loop
o Executive’s ‘helper’ used for storing verbal material
 Visuospatial buffer
o Executive’s ‘helper’ used for storing visual material

The central executive

 Same as executive control processes  needed to control the selection of ideas that are activate at any
moment in time
 WM, bc of this, is quite fragile as well: each shift in attention brings new info into WM which displaces
earlier times

Entering LT storage: the need for engagement

Two types of rehearsal

 Maintenance rehearsal
o Focus on the to-be-remembered items with little thought about what the items mean or how they
are related to each other
o A rote/mechanical process; provides no long term benefit whatsoever
 E.g. we have seen coins a gazillion times in our lives; however we never actually pay
attention to the coins such that when ppl are asked which way Lincoln’s head is facing on
the penny, ppl’s answers are close to random
 Relational/elaborative rehearsal
o Thinking about how these items relate to each other or to what you already know/what they mean
o Vastly superior to maintenance rehearsal for establishing info in memory

The need for active encoding

 Takes time to get info into LTM
 fMRI studies: show that when ppl study list of words and then later recall it
o Greater levels in HC and PFC were reliable associated w greater probabilities of retention later on
o Aka learning is NOT A PASSIVE PROCESS

Incidental learning, intentional learning and depth of processing

 STUDY: pts in one condition heard a list of 24 words and their task was to remember as many of those
words as they could  intentional learning (learning that is deliberate, expectation that memory is tested
later on)
o Pts in other condition heard the same 24 words but had no idea their memory would be later tested
 incidental learning (learning in absence of any intention to learn)
 Diff groups were asked to assess diff aspects of the words (e.g. contains letter e? how
many letters did each word contain? Consider each word and rate how pleasant it seems?
o Later all pts were tested ans asked to recall as many of the words as possible
o Results
 Poor performance for ‘find the e’ and ‘count the letters’ condition, but better in ‘how
pleasant?’ group (even just as good as intentional learning group!!)
 This suggests that intention is not what is important but instead on the level of processing!!
o Shallow processing: engaged in material in superficial fashion
 Worse recall
o Deep processing: processing that requires some thought about what the words MEAN
 Better recall
 TLDR words remembered are the same in conditions w shallow processing w INTENT to recall vs NO
INTENT to recall; and words remembered are the same w deep processing w INTENT to recall vs NO
INTENT to recall… what matters is how ppl approach the material they are seeing/hearing
 MEMORY | 4

The role of meaning and memory connections

 To remember what you’re learning you should use deep processing (attention to meaning)

Connections promote retrieval

 Task of learning is not a matter of placing info into LTM storage… learning also needs to establish some
appropriate indexing
 What facilitates memory retrieval? A great deal depends on memory CONNECTIONS
o Appropriate connections allow linking of contextual information
o This means that attention to meaning involves thinking about relationships, which helps you find
connections among your various ideas
o Therefore connections are the things that really matter for memory

Elaborate encoding promotes retrieval

 Craik and Tulving (1975)
o Pts were shown a word then shown a sentence w one word left out
o Task: decide if word fits into sentence (eg. word: chicken; sentence: “she cooked the _______”)
o After series of trials there was a surprise memory test, w pts asked to remember all the words
 Some sentences shown were simply; others were complex
 More elaborate sentences showed a larger memory benefit
o Suggest that DEEP and ELABORATE processing leads to better recall than deep processing itself
o Potentially the richness of this sentence offers the potential for many connections to be activated,
which then process potential retrieval paths

Organizing and memorizing

 Sometimes memory connections link the to-be remembered materials to other info already in memory
 in other cases, the connections link one aspect of the t-be-remembered material to another aspect of the
same material, which ensures that if any part of the material is recalled, then all will be recalled
 George Katona: argued that the key lies in ORGANIZATION
o Said the processes of organization and memorization are inseparable: you memorize well when
you discover the order within the material
o Aka links to connections again

Mnemonics
 These strategies simply provide some means of organizing the to-be-remembered material
o E.g. ROY G BIV = colours of the rainbow
 Others involve the use of ‘mental pictures’  must have a RELATIONSHIP
o Eg. trying to remember word pair of eagle-train, imaging eagle soaring w train in mouth
 MEMORY | 5

 Peg word systems: involves linking WORDS W NUMBERS which provides ‘peg words’ (numbers) in
memorizing something that you can ‘hang’ the materials to be remembered on these ‘pegs’
o Basically allows you to remember a list w no organization by imposing an organization on it by
using a skeleton or scaffold that is itself tightly organized
 Mnemonics are v helpful in general, but downside: you only pay attention to one aspect of the material you
are trying to memorize (eg. first letter of the word) such that you will spend less time thinking about other
possible connections to help you understand the material

Understanding and memorizing

 When we want to remember more complex and meaningful items, like memory for events, it is STILL
DEPENDENT on your being able to organize the material to be remembered
o Also dependent on understanding****
 A students tend to understand material v well
 C students tend to understand the material poorly
 Brandford and Johnson
o Given a passage about doing laundry
o Some given title; others not
o Those not given title were HELLA confused and could not remember after a delay (bc information
was not connected to past knowledge)

The study of memory acquisition

The contribution of the memorizer

 Memory (again) depends on connections among ideas
 These connections are fostered by the steps you take in your effort toward organizing and understanding
the materials you encounter
o Memories not established via sheer contact w items
 To know if someone will remember something, need to know what they were doing during info acquisition
o Also need to know if that person has ability to connect to-be-remembered material to past
knowledge (other memories) so you can ‘hook’ this new material
o Eg. why sports fans have easier time learning new facts about sports

The links among acquisition retrieval and storage

 Acquisition cannot be separated from claims about storage and retrieval

Cognitive psychology and education: how should I study?

 Intention to memorize (on its own) has no effect on memory
 You should focus your efforts on making sure you UNDERSTAND the material instead
o This is why it is often useful to quiz yourself w questions
 Do not rely on passive study strategies (eg. reading and re-reading notes or lectures)
 Useful to study w a friend – they can explain topics to you and vice versa  gives you a more active role in
learning information + allows you to enhance understanding through their help
 Memory is best if you spread your studying out across MULTIPLE occasions (spaced learning) rather than
massed learning (‘cramming”)
o Basically you may get new perspective every time you return to material, which enables new
connections
 Mnemonic helpful but at cost

Chapter summary 246

 MEMORY | 6

Chapter 7: Interconnections between Acquisition and Retrieval (pg. 248)

 Talks about HM again; MTL removal caused him to lose ability to form new, specific episodic memories
o His symptoms are the same as ppl who suffer from Korsakoff’s syndrome: these ppl are
profoundly amnesic, unable to recall events from their own lives
o Can test Korsakoff patients indirectly if we play them a song, and sk them which song they just
heard  patients will answer randomly
o BUT if we ask them ‘which melody do you prefer’ they are most likely to pick song they heard
before… shows that they somehow remember/ are influenced by the earlier experience
o These patients also get better w procedural memory tasks (but do not remember them explicitly)
 E.g. mirror drawing task
 This suggests there exist diff types of memory

Learning as preparation for retrieval

 There are diff ways to retrieve info from memory
o can RECALL info
o can RECOGNIZE info

Crucial role of retrieval paths

 TBT: when we learn, we make connections btw newly acquired info and info already in memory

Context-dependent learning
 Context-dependent learning: scuba diving experiment
o Ppl remembered material better when the context they learned it in matched the context they
recalled it in  aka under water vs on land
o BUT still had good recall ON LAND
 This has been replicated to match situations which students face (aka diff rooms)
 What matters is not he PHYSICAL context but the PSYCHOLOGICAL context
o As a result, we can use context reinstatement: strategy to re-create the thoughts and feelings of
the learning episode even if at time of recall you are in a very different place

Encoding specificity
 NOTE: the scuba divers remembered the words they’d learned AND something about the context in which
this learning occurred (aka if context left no trace in memory, there would be no way for a return tot eh
context to influence the divers later)
o E.g. study gave 2 groups diff sentences: 1) the man lifted the piano or 2) the man tuned the piano
 If pts were given sentence #1, then during recall if they were cued w ‘something heavy’
they could recall the sentence easier
 ^^^ this effect is known as encoding specificity: what we encoded is specific – not just the physical
stimulus as it was encountered, but the stimulus together with its context
o TLDR text argues that what goes into our memory is a record of the material we’ve encountered
AND ALSO a record of the connections we established during learning

The memory network

 Learning involves strengthening of memory connections
 These connections serve as retrieval paths, guiding you toward th einfo you seek
 Memory is thought of as a VAST NETWORK of ideas
o These ideas exist as nodes within the network
o These nodes are then tied to each other via connections called associations/associative links

Spreading activation
 Nodes become ACTIVATED when they have received a strong enough input signal
o Once the node is activated, it can in turn activate other notes that are connected
o TLDR nodes receive activation from their neighbors, and as more activation arrives at a particular
node, the activation level for that node increases until it reaches the node’s response threshold,
after which that node will fire
 MEMORY | 7

 Node fires  summons the attention to that node AND sends energy to its neighbors to
activate them too 😊
o Subthreshold activation: since threshold is thought to accumulate, 2 subthreshold inputs may add
together to bring the node to threshold!!
 Recency effect: if node was activated recently = warmed up = even weak input can be
sufficient to bring the node to threshold
 This whole process is known as spreading activation; this process is RANDOM, and starts at one point
but flows in all directions simultaneously

Retrieval cues
 If you get two retrieval cues,
allows spreading activation
from two sources
simultaneously which is enough
to life the nodes’ activation to
threshold levels

Context reinstatement
 Context reinstatement allows
contextual thoughts during
learning to become associated w
the materials being learned
 AKA info you seek in memory
is probably tied to the retrieval
cue you are given, BUT it is
possible that the info you seek
receives insufficient activation
from this source; HOEVER, the
info you seek may also be tied
in memory to thoughts that had
been triggered by the learning
context (e.g. being underwater,
a certain classroom, etc.) such
that if you are back in that
context during recall, the target
nodes can receive 2x input (activation from 2 diff sources) and this will help activate the target nodes)

Semantic priming
 Lexical decision task: pts are shown a series of letter seq on a computer screen
o Some are words, some are not (e.g. blar, plome)
o The pts are asked to hit yes button if it’s a word and no button if not a word!
o We can use pts speed of response on this task as an index of how quick they can locate the words
in their memories
 Meyer & Schvanveldt (1971)
o Showed pts PAIRS of letter strings, pts said yes if both strings were words and no otherwise
o Also sometimes if the strings were both words they were semantically related (nurse, doctor) or
not (lake, shoe)
 Hypothesized that if the words are related, spreading activation occurs quicker bc the
node already partially activated its neighboring word via subthreshold activation, and
response time would be faster  semantic priming
 Results confirmed their predictions; primed words were faster by almost 100ms is words
were related 😊
 NOTE: we have some level of control over the starting points for our memory searches via processes of
reasoning and executive control AND that once spreading activation has begun, ppl have the option of
‘shutting down’ some of this spread if they are convinced the wrong nodes were activated
 MEMORY | 8

Different forms of memory testing

 Recall: you are presented with a retrieval cue that broadly identifies the info you seek, and then you need
to come up w the info on your own
o Requires memory search bc you have to come up w the info on your own
o Depends HEAVILY therefore on memory connections
 Recognition: information is presented to you, and you have to decide whether it is the right info or not
o Often depends on a sense of familiarity
o Do NOT have source memory: any recollection of the source of your current knowledge but if
you do have a strong sense of familiarity, then you can make an inference about where that
familiarity comes from
 These modes of retrieval are fundamental diff from each other

Familiarity and source memory

 Source memory  memory connections that link the target material to thoughts about the setting in which
you encountered that material; these connections help you recall when and where you saw that person,
heard that song, smelled that perfume, etc.
 Familiarity and source memory are INDEPENDENT of each other
 It is possible for an even to be familiar without any source memory, and also have source memory without
any familiarity
o E.g. see an actor, cant remember what film they are in (familiarity without source mem)
o E.g. Capgras syndrome  accurate and detailed memories of the past sans familiarity
 Remember/know distinction:
o Remember: have source memory  can recall exactly the info and where they know it from
 fMRI shows activity in hippocampus
 BUT if during learning the hippocampal region is active, it is more likely that the person
will remember the stimulus when tested later
o Know: have familiarity  no source memory, just broad feeling that item was seen before
 fMRI shows activity in anterior parahippocampus
 ALSO if during learning the rhinal cortex is especially active, then the stimulus is likely
to seem familiar later on and thus trigger a ‘know’ response

Implicit memory

Memory without awareness

 Pts asked to read words w no indication that their mem would be tested later on
 Then some time later pts are given a lexical decision task (some of the words are duplicates from the
previous list)
o Wanted to see if first exposure somehow primed the pts for a second encounter
 Results
o Lexical decisions are quick if the person has already seen the test word: repetition priming
o Observed even when pts have no recollection of having encountered the stimulus words before
 This is called ‘memory without awareness’
 Word-stem completion: pts given 3 or 4 letters and are asked to complete the word (e.g. “cla- could be
completed w “clam” or “clatter”)
o Turns out that pts produce words that they’ve encountered recently; seen even if they do not
remember the words consciously
 Explicit memories: revealed via direct memory testing (which specifically urges you to rmbr the past)
o E.g. test of recall
 Implicit memories: revealed by indirect memory testing (often shown using priming effects)
o E.g. lexical decision, word-stem complete, and many other tasks

False fame
 Jacoby, Kelley, brown and jasechko
o Presented pts w a list of names to read out loud
 MEMORY | 9

o Told nothing about memory test (told test had to do w pronunciation)

o Then later shown new list of names, asked to rate each person on this list according to how famous
each was (list included real and fake ppl)
 Of the fake ppl, some of the names were from previous list and some were just new
o For some pts the famous list was presented right after the pronunciation list and for others there
was a 24 hr delay
o Results
 Ppl presented list right away noticed that names were familiar, but attributed the source to
the previous list of names
 Ppl presented the list 24 hrs later noticed the names were familiar, but since such a large
time delay they attributed it to them being famous

Implicit memory and the ‘illusion of truth’

 How broad is this potential for misinterpreting an implicit mem?
o Pts heard a series of statements and had to judge how interesting each statement was (some were
true; some were false)
o After that, pts were presented w more sentences, but had to judge the credibility of these
statements (certainly true  certainly false)
 Some of the sentences heard before were repeated in the ‘truth test’
o Results
 Sentences heard before were more likely to be accepted as true (familiarity ↑ credibility)
 Effect seen even when pts were warned in advance not to believe the sentences in first list
 Another experiment
o Same thing, this time pts were told half were made by women, and were true; other half were
made by men and were false (other pts told the reverse)
 E.g. jess said crocodiles sleep with eyes open; frank said mice can run 4 mi/hr on avrge
o Results: statements plainly identified as false when they wre first heard still created an illusion of
truth!!! In second pt of experiment!! Bc ppl couldn’t attribute them to male of female

Attributing implicit memory to the wrong source

 Implicit memory can influence and bias us (e.g. politics, marketplace)
 Often times we can have source confusion: when an item invokes a sense of familiarity, but we attribute
the source of where we know that item incorrectly

Theoretical treatments of implicit memory

 We are better at remembering that something is familiar over remembering WHY it is familiar
 Pts even prefer a previously presented stimulus to a novel stimulus, even if they have no idea that their
preference is being guided by memory  basically called ‘memory without awareness’

Processing fluency
 Process:
o Stimulus arrives in front of our eyes
o Triggers certain detectors
o These trigger still other detectors
o Etc. etc. until you recognize the object
 “flow of activation” that moves from detector to detector, called the processing pathway: the seq of
detectors and their CONNECTIONS that the activation flows through in recognizing a specific stimulus
 Remembering often involves the activation of a node, and this node triggers other, nearby nodes so that
they become activated, and those trigger other nodes… etc. etc.
 THE USE OF THE processing pathway STRENGTHENS that pathway bc the baseline activation level of
nodes or detectors increases if the nodes or detectors have bene used freq in the past/recently
 TLDR the use of a pathway increases the processing fluency of that pathway (the ease and speed w which
the pathway will carry activation) … This explains implicit memory effects

The nature of familiarity

 MEMORY | 10

 Familiarity is more like a conclusion you draw rather than a feeling trigger by a stimulus
 Evidence suggests that a stimulus will seem familiar when the following occurs:
o 1) you have encountered the stimulus before
o 2) bc of that prior encounter, you are now faster and more efficient in your processing of that
stimulus (aka processing fluency)
o 3) you detect that fluency, which leads you to register the stimulus as distinctive or special
o 4) you try to figure out WHY the stimulus seems special, and you reach a particular conclusion:
the stimulus has a distinctive quality BC it’s a stimulus you have met before somewhere
o 5) you draw a further conclusion about when and where you encountered the stimulus
 NOTE: none of these steps happen consciously, also can go astray such that an illusion of
familiarity is produced

The hierarchy of memory types

 We are often influenced by
the past without being aware
of that influence
 Two types of memories:

Explicit memory: conscious and

deliberate, can be subdivided into
episodic memories and semantic
memory

Implicit memory: divided into 4

subcategories: procedural memory,
priming, perceptual learning and
classical conditioning

Amnesia
 Retrograde amnesia: caused often by blows to the head, person unable to recall events prior to blow
 Anterograde amnesia: person cannot recall anything after the blow
o OFTEN TIMES BOTH CO-OCCUR TOGETHER

Disrupted episodic memory but spared semantic memory

 Clive Wearing: episodic memory disrupted, but memory for generic info (e.g. loving his wife) was fine
 Other ppl show opposite effect
 MEMORY | 11

o These cases show double dissociation that demands a distinction btw episodic and semantic mem

Anterograde amnesia
 H.M. suffered severe memory loss bc MTL were removed in brain surgery
 He had anterograde amnesia, not retrograde; he could no longer hold anything in his LTM
o Could hold a convo unless it was interrupted
 Korsakoff’s syndrome: longtime alcoholism causes thiamine deficiency (vitamin B1) which causes severe
anterograde amnesia, like H.M.

Anterograde amnesia; what kind of memory is disrupted?

 1911  Swiss psychologist named Edouard Claparede reported that he was introduced to a young woman
suffering from Korsakoff’s amnesia
o He reached over to shaek her hand, but had a pin in it such that when they shook hands he gave
her a painful pinprick
o The next day, they “re-met” an dhe reached out to introduce himself to shake hands, but just
before she pulled away and refused to shake his hand. When asked why, after some confusion the
patient simply said vaguely “sometimes pins are hidden in ppl’s hands” #deep
 Basically only explicit memory is fucked up in kosrakoff’s such that they confabulate/ make up responses
to things their implicit memory remembers

Can there be explicit memory without implicit?

 Explicit mem is independent of implicit mem
 Experiment: brain damage to hippocampus but not amygdala and another damage to amyg but not HC
o Exposed to series of trials in which particular stimulus (blue light) reliable followed by a loud boat
horn WHILE ANOTHER STIMULUS (green, yellow or red) were not
 Later pts were exposed to blue light on its own and their galvanic skin response were
tested and then were asked directly: which colour was followed by the horn?
o Results
 Pt w HC damage showed fear rxn via skin conductance response (like control) but could
not answer q
 Pt w amyg damage showed no fear rxn but able to report that blue was associated w horn

Optimal learning
 How we test memory (e.g. direct vs direct test of memory) reflects diff types of memory
 Ideal form of learning is one that is ‘in tune w’ the approach to the material you’ll need later
o E.g. if you are tested explicitly, you want to learn that material in a fashion that prepares you for
that form of retrieval
o Prob: during learning, you don’t often know how you will be approaching the material later… best
strategy for learning would be to use MULTIPLE PERSPECTIVES

Cognitive psychology and the law: the “cognitive interview”

 Are there steps to help witnesses remember events?
 Some tried efforts include hypnosis, memory-enhancing medication, etc.
 More promising approach: cognitive interview (relies on context reinstatement)

Chapter summary

Chapter 8: Remembering Complex events

 Hyperthymesia: super autobiographical recall

Memory errors, memory gaps

 Sometimes we remember things perfectly, other times we don’t

Memory errors: some initial examples

 In 1992, an El Al cargo plan crashed into a building, killed 43 ppl
 MEMORY | 12

 10 months later, researchers questioned 193 Dutch ppl about the crash, asked them “did you see the TV
film of the moment the plane hit the apartment building?”
o More than ½ said yes, even though there was no such film
 Follow up study: investigators surveyed another 93 ppl about plane crash and asked about the video again
o This time, 2/3 of ppl said they saw the video and many of them provided specific details
 Another study (that used a shorter delay) found that pts memory was influenced by TYPICALITY EFFECT
o Pts asked to wait in academic office
o Taken out of office, asked to describe what its like
o Found that ppl expected office to contain shelves w books (for eg.) but this office had no books
 Descriptions were in line w expectations and not w reality!!

Memory errors: a hypothesis

 Your mem for an episode becomes more and more interwoven w other thoughts you’ve had about the
episode, and it becomes difficult to keep track of which elements are linked to the episode bc they were, in
truth, part of the episode itself, and which are linked merely bc they were associated with the episode in
your thoughts… this produces transplant errors, in which elements that wre part of your thinking get
misremembered as if they were actually part of the original experience

Understanding both helps and hurts memory

 Memory connections both help and hurt recollection
o Help  bc connections (retrieval paths) enable location of info in memory
o Hurt  bc they sometimes make it hard to see where the remembered episode stops and other,
related knowledge begins
 This results in intrusion errors: errors in which other knowledge intrudes into a
remembered event

The DRM procedure

 Word lists can show the effect of memory connections helping and hurting memory
o E.g. list: bed, rest, awake, tired, dream, wake, snooze, blanket, doze etc…
 All related to sleep, but word sleep is not there BUT pts v often remember it being there
 Pts are just as confident that word sleep is there as other words they acc heard
 DRM procedure: eg. above^^; yields a lot of memory errors, even if pts are warned before

Schematic knowledge
 Schema: summarize the broad pattern of what is normal in a situation
o help guide your actions, also help when the time comes to recall how an event unfolded
o this is bc we often have gaps in our recollection, such that we rely on our schema to fill them in

Evidence for schematic knowledge

 schema help us, but can also hurt us by promoting errors in perception and memory
 Errors produced by schema are predictable bc they tell us what is ordinary/typical in a given situation:
o E.g. Frederick Bartlett presented pts w a story about Native Americans
 Tested his pts after, they did well in recalling the gist of a story but made many errors in
recalling the particulars
 Details omitted tended to be ones that made little sense to Bartlett;s British pts, such that
unfamiliar aspects were changed into more familiar aspects to increase logic
o E.g. schema of office before, makes sense why ppl misremembered shelf of books!
The cost of memory errors
 We rely heavily on memories…turns out they are not super accurate
 Troubling  e.g. eyewitness testimony putting wrong guy in prison ☹

Eyewitness errors
 How often do eyewitnesses make mistakes?
 Now there are over 320 cases of these DNA exonerations of ppl convicted but were not guilty!!!
o Exonerees had spent over 12 yrs in jail, on average, for crimes they did not commit
 MEMORY | 13

o Most were bc of eyewitness errors ☹

Planting false memories

 Loftus & Palmer (1974)
o Showed pts a series of pics depicting car crash
 Sked “how fast were cars going when they hit/smash each other?”
 If said smash, pts rated much faster
 1 week later, pts asked in a neutral way if they had seen broken glass in the pics;
pts who were asked w the word ‘hit’ remembered (correctly) no glass – pts w
word “smash” remembered glass
 Many other experiments duplicate this finding; all cause pt to experience an event and then expose them to
misleading suggestion about how the event unfolded… then some time passes… pts memory is tested and
outcome is always the same: large # of pts end up incorporating the false suggestion into their memory
o NOTE: easier to plant plausible memories than implausible ones
o NOTE: false memories are more easily planted if the research pts don’t just head about the false
event, but are urged to imagine how the suggested event unfolded (called imagination inflation)
 Overall, v easy to alter ppls memories for past events

Are there limits on the misinformation effect?

 Misinformation effect: memory errors that result bc of some form of post-event misinformation
 We can use even one word (e.g. hit vs smash) to alter ppls memories
 Larger scale:
o Students were told that the investigators were trying to learn how diff ppl remember the same
experience; pts were then given a list of events (they were told) that had been reported by their
parents
 Some were real and some were fake
 Results: pts could easily remember real events; 25% of them were able to ‘remember’
fake events and supply details of this entirely fictitious episode

Errors encouraged through ‘evidence’

 Other researchers provide pts w ‘evidence’ in support of their bogus memory
o E.g. got real childhood photo and altered it to put the pt in a hot air balloon
o E.g. pts showed the ppl pics of their second grade class, which was sometimes enough to prove to
the pts they really did have info on them as kids

False memories outside the laboratory

 Are there limits to false memory?
 Tldr yes; this part is pointless wtf

Avoiding memory errors

 Overall, our memories are generally accurate

Memory confidence
 Confidence of memories matters bc ppl put more trust in confident memories than hesitant ones
 There is little relationship btw how certain someone is vs how accurate their recollection is
 Why is this? 1) our confidence in memory is often influenced by factors that have no impact on memory
accuracy (e.g. giving positive feedback after a task increases confidence)
o Other researchers look at emotion when ppl are recalling a memory, response speed, and so forth
o There is v weak (if any) links to memory and these markers

Forgetting

The causes of forgetting

 1) One prominent reason we forget is bc of failure in acquisition
 2) Another is retention interval (amount of time that elapses btw initial learning and subsequent retrieval)
 MEMORY | 14

o As this interval grows, you are likely to forget more of the earlier event  “forgetting curve”
 3a) decay: w the passage of time, memories fade or erode
 3b) interference theory: passage of time is correlated w forgetting but does not cause forgetting; instead,
time just allows the opportunity for new learning, and new learning disrupts old learning
 4) retrieval failure: the greater the retention interval, the greater the likelihood that your perspective has
changed, the greater the likelihood of retrieval failure
 All these hypotheses are correct!!!!
o Memories decay w passage of time
o A lot of forgetting also has to do w retrieval failure
 Sometimes retrieval is partial
 E.g. TOT phenomenon: can recall the starting letter of the sought after word
and approx. what it sounds like
 Memory interference
o In many cases, new info gets interwoven w older info, producing a risk of confusion (central for
misinformation effect)

Undoing forgetting
 Is it possible to undo forgetting?
 Hypnosis, certain drugs
o NOTE: neither of these
techniques actually
improves memory
o E.g. under hypnosis ppl
may be just willing to
SAY more in order to
comply with the
hypnotists instructions
o E.g. drugs given to
improve memory largely
work as sedatives, and put
the person in a less
guarded state of mind
 Procedures to diminish forgetting:
cognitive interview
o Designed to maximize
quantity and accuracy of
info obtained form
eyewitnesses to crimes
o Builds on the fact that
retrieval of memories
from LTM is more likely
if suitable cue is provided
o Therefore interviewer offers many diff retrieval cues on the idea that one will be successful
 TO LIMIT FORGETTING: just revisit the info… each ‘visit’ seems to refresh memory such that forgetting
is much less likely
Memory: an overall assessment
 Literally restate everyting this tb is so dum

Autobiographical memory
 Do the 3 factors – involvement in a remembered event, emotion, and long delay – matter in any way for
episodic memories??
 Autobiographical memory: memory that each of us has for events in our lives; this memory plays an
important role in shaping how we think about ourselves and thus how we behave

Memory and the self

 MEMORY | 15

 Having involvement in an event, rather than passively witnessing it has a large effect on memory
o This is (in part) to do with the fact tha info relevant ot he self is better remembered than info that
is not self relevant
 Still can have mem errors wrt bc it is a mix of genuine recall and schema-based reconstruction
o E.g. ppl often have a self-schema of being consistent and stable; therefore will remember past
relationships differently and unwittingly distort their personal history to match this
 Also to do w the fact ppl want to keep a positive view o fhte self
 E.g. study showed that students remembered correctly 89% of ‘A’ grades and only
remembered 29% of the ‘D’ grades

Memory and emotion

 Emotion is important in autobiographical experiences
 At biological lvl, emotional arousal seems to promote the process of memory consolidation
o This process happens unconsciously
o If it is interrupted for whatever reason no memory is established for later recall
 Factors that promote consolidation (and therefore later retrieval)
o SLEEP
o Emotion  triggers amygdala, which in turn increases activity in HC
 Also emotional even is more likely to be important to you, guaranteeing that you’ll pay
close attention
 Emotion changes what you pay attention to within an event, which will alter what is
remembered or not (often ‘narrowing’ of attention occurs)

Flashbulb memories
 These are memories that are extraordinarily clear, of typically highly emotional events
o E.g. first hear the news of Kennedy’s assassination
 These memories often contain substantial errors
o This occurs bc memories are MALLEABLE; every time we discuss a memory, it serves as a
memory rehearsal which (tbt) may promote memory accuracy or introduce new information

Traumatic memories
 Most traumatic events are well remembered for many years  victims often are plagued by a cruel
enhancement of memory
 This is bc of CONSOLIDATION: promoted by conditions of bodily arousal
o However not all traumatic events are always well remembered!
 i.e. these events may be accompanied by sleep deprivation, head injuries, substance abuse
etc. which can disrupt memory, as well as stress associated w the event itself

Repression and ‘recovered’ memories

 controversial; thought now that psychologists actually promote the appearance of these ‘repressed’ mem bc
they are often re-remembered during therapy session
 some memories do seem to be repressed (due to retrieval failure)
 ALSO some recovered memories may be false memories!!

Long, LT remembering
 TB: longer retention intervals tend to be associated w greater amount of forgetting!!
 Some memories from long ago are v accurate
o E.g. pts brought back and shown pics of their graduation; showed v little forgetting
 Answers were 9% correct for 3 months – 7 years after grad!!
 What about stuff we learn in classes?
o Study tested students memory for concepts they learned in cog psych classes
 Found that forgetting of names and specific concepts were observed during the first 3 yrs
of the course, after which performance stabilized such that students tested 10 yrs after
still remembered a fair amount!!!
 MEMORY | 16

 Means that how much mem fades depends on how well these memories were established in the first place

How general are the principles of memory?

 Ppl tend to remember v little from before age 3
 V clear and detailed memories of late adolescence and early adulthood (reminiscence bump)
 Autobiographical memories can last us a long time; much more likely to last if person revists the target
memories (reduces forgetting)
o The same thing can be said of non-autobiographical memories tho
o Both types of memories are open to error
 Idea: certain principles apply to all mem in general, regardless of what is being remembered:
o All depend on connections, which facilitate interference (cause memories to blur)
o Connections can fade w passage of time, producing memory gaps
o Gaps are likely to be filled in via reconstruction based on generic knowledge
 ^^all true for memories of recent events emotional memories or memories of calm events,
memories for complex episodes or mem for simple word lists
 Rules for implicit memory is diff

Cognitive psychology and education: remembering for the LT

 Meh not important

Chapter summary

Part 4| Knowledge

Chapter 9: Concepts and generic knowledge (pg. 332)

 “Funes the Memorious  Argentinian book, Funes never forgets anything BUT he is immensely distressed
by his memorial ability
o Says he has trouble thinking bc he remembers so much about how people differ that he cant focus
on what they have in common ☹
 Actual case: Solomon Shereshevsky  had enhanced autobiographical memory; he too had a (-) exp w this

Understanding concepts
 We need concepts in order to have knowledge, and we need knowledge in order to function
 Hypothesis: understand a concept is analogous to knowing a dictionary definition (this is a v basic idea, we
need to make it more complex to fully encompass concepts as a whole

Definitions: what is a dog?

 Everyone knows what a dog is! Ty textbook!
 Defines dog
 We use this definition in straightforward ways
 Ludwig Wittgenstein: argued that simple terms we all use every day actually don’t have definitions
o E.g. game  no one definition to encompass all attributes of every game
o Basically says that commonplace, every day items have definitions to every attempted def’n

Family resemblance
 Turns out we cant say that a “dog is a creature that has fur and 4 legs and barks”
o Bc there are exceptions, such that Wittgenstein proposed that members of a category have a
family resemblance to one another
 No defining features for everyone in a family
 There are commonalities btw ppl in a family
 Basically we take an ‘ideal’ member of a family; all other members share at
least one feature w this ideal such that there are no features shared by all items
in a category, there are characteristic features for each category
 Family resemblance is a matter of degree, not all or none
 MEMORY | 17

 Elenor Rosch: created the prototype theory out of Wittgensteins research

Prototypes and typicality effects

 Prototype theory: best way to identify a category = specify the ‘perfect’ type (center) than the boundaries
o Have some ideal depiction of each category, and compare wrt this ideal

Prototypes and graded membership

 Membership in a category depends on resemblance to the prototype, and resemblance is a matter of degree
 Graded membership: objects closer to the prototype are ‘better’ members of the category than objects
farther from the prototype

Testing the prototype motion

 Use sentence verification task  pts are presented w a succession of sentences, and asked to indicate
whether each sentence is true or false
o Response is slower for sentences like ‘a penguin is a bird’ vs ‘a robin is a bird’ bc pts use their
response by comparing the thing mentioned to their prototype for that category
o When there is close similarity btw test case and the prototype, the pts make their decisions quickly
 Use production task  ask ppl to name as many birds or dogs as they can
o Via prototype theory, they will do this by first locating the prototype in memory, then asking what
resembles that prototype such that (for e.g.) birds closer to the prototype should be mentioned
first; birds farther from the prototype later on
 Use rating task  pts given list of members in a category (e.g. birds) and asked to list which ones are
“birdier” (use prototype as base for these judgements!)
Basic level categories
 Turns out that not only are category members privileged, certain types of category are also privileged
o Rosch aruves that there is a ‘natural’ level of categorization that is not too specific nor general
 This is called basic level categorization: represented using a single word
o E.g. basic level categorization: chair; higher level categorization: lawn chair, kitchen chair
 We also have easier time explaining what basic level categories (e.g. chairs) have in common vs more
superordinate categories (e.g. furniture)
 Children learning to talk often acquire basic level terms earlier than either the most specific subcategories
or the more general, more encompassing categories
o Aka basic level categories seem to reflect a natural way to categorize the objects in our world

Exemplars
 Prototype theory isn’t the only way to think about graded membership

Analogies from remembered exemplars

 Exemplar-based reasoning: similar to prototype theory, except instead of comparing each new item to a
specific ‘ideal item’, we compare each new item to all exemplars that we know of the category
o Aka we categorize objects by comparing them to a mentally represented ‘standard’ that is
provided by whatever example(s) of the category that comes to mind

Explaining typicality data w an exemplar model

 How we organize fruit – more typical = banana, apple, etc. – vs less typical – grapefruit, dragon fruit

A combination of exemplars and prototypes

 Prototypes provide an economical representation of whats typical for a category
 Exemplars provide information that’s lost from the prototype – including info about the variability within
the category
o E.g. category ‘gifts’ can be fine tuned for ‘gifts for four year old’ and even further to ‘gifts for four
year old that recently broke her wrist’
o This is easy to explain w exemplar theory
 Both exemplar and prototype theory are useful in many diff ways
o E.g. someone may know a lot about horses, therefore have many exemplars in their mind
 MEMORY | 18

 We also can COMBINE the models bc they are similar in crucial ways

The difficulties with categorizing via resemblance

 For most concepts, definitions are not available (this does not seem to be a big deal bc we can rely on a mix
of prototypes and exemplars)
 TYPICALITY also plays a large role in ppl’s thinking
 ….some results do not fit into this thinking tho:

The differences btw typicality and categorization

 Sometimes typicality and category membership do not go hand in hand:
o Armstrong, Gletiman and Gleitman (1983)
 Gave their pts this instruction: we all know that some numbers are even-er than others. I
want you to rate each of the numbers on this list for how good an example it is for the
category ‘even number’
 Results: pts were just as consistent as when given categories like bird or fruit
 E.g. judged 4 as more of an even number than 4354
o Lol wtf is the point of this
 How are category judgements made when they DON’T rely on typicality?
o Preschool children were asked to make something into a ‘coffeepot’, a ‘raccoon’, etc.
o Then kids were asked if it was possible to turn a toaster into a coffeepot; kids said yes
o Then kids were asked if one can turn a skunk into a raccoon; kids said no
 Aka ppl reason differently w naturally occurring items than manufactured items

The complexity of similarity

 Judgements about categories are often influenced by typically
 Sometimes, category judgements are independent of typicality (e.g. Moby Dick is a special type of whale,
but still a whale OR counterfeit money is not really money, even though it looks and acts like it)
 If you do not focus on category membership based on typicality, you instead focus on attributes that you
believe are essential for each category, which then depend on your BELIEFS about that category!
o This is bc prototype and exemplar theory are based on RESEMBLANCE
 However using the notion of ‘resemblance from shared properties’ doesn’t work by itself
o W a little creativity, can have thousands of properties shared by plums and lawn mowers, for eg.
 Aka resemblance does depend on shared properties, but only for important, essential properties

Concepts as theories

Explanatory theories
 Our understanding of a concept seems to involve a network of interwoven beliefs linking the target concept
to other concepts (e.g. to understand what a raccoon is, you need to know what parents, life cycles,
hereditary, etc. are)
 We need a more holistic approach where we place more emphasis on the interrelationships among concepts
o Many authors therefore suggest that each of us have something that we can think of as a ‘theory’
about raccoons – what they are, how they act, why they are as they are
o The theories serve the same function: to provide crucial knowledge base that we rely on in our
thinking about an object, event or category

The function of explanatory theories

 Our implicit theories influence how we categorize things
o E.g. being drunk  if you see someone jump into a pool at a party in full clothes, you may put
them in this category. HOWEVER jumping into a pool does not fit into a prototype of being
drunk… it does fit into our theory of drunken behaviour, allowing us to think thorugh what being
drunk will cause someone to do or not do
 Our theories allow us to think of new things (e.g. can an airplane fly if it was made of wood?)
 Our theories also affect how quickly we learn new concepts
 MEMORY | 19

Inferences based on theories

 Categorization enables us to apply our general knowledge to new cases we encounter
o E.g. meet new dog, instantly know a lot bc we have theories about dogs
 Early research indicated that inferences about categories are guided by typicality
o E.g. study told pts new fact about robins; they were willing to extend that to ducks. BUT if told
the fact about ducks, would not extend that to robins bc people make inferences from typical
case to whole category, but not from atypical case

Different profiles for different concepts

 Implication: ppl may think about diff concepts in diff ways
o Aka most people believe that natural kinds (groups of objects that exist naturally in the world)
are there bc of forces of nature, consistent across the years
o Means their properties are seen as stable
 Things are different for artifacts(objects made by humans)
o Meaning we reason differently about natural kinds and artifacts bc we have diff beliefs abt them
o E.g. kids said toasters can be turned into coffeepots, but skunks cannot be turned into raccoons
 Role of beliefs is super important (e.g. some concepts involve goal-driven categories: diet foods)
 fMRI shows that diff parts of the brain are activated when ppl think of living vs nonliving thigs
 brain damage cases:
o lose specific ability to name certain objects or answer simple questions about those objects
o suggest that separate brain systems are responsible for diff types of conceptual knowledge
 In healthy brains, when ppl think of specific categories their sensory and motor cortex lights up, suggesting
that abstract conceptual knowledge is intertwined w knowledge of what objects look and sound like
The knowledge network
Traveling through the network to retrieve knowledge
 Associative links don’t just tie together the various bits of knowledge, they also help REPRESENT the
knowledge
o E.g. G. Washington was an American president
 Associative link exists btw WASHINGTON node and PRESIDENT node
 Retrieving information from network occurs via spreading activation
 Collins & Quillian (1969)
o Sentence verification task: pts shown sentences like ‘a robin is a bird’ or ‘cats have claws’ or
‘cats have hearts’ (true sentences) vs “a cat is a bird’ (false sentences)
 pts had to hit true or false buttons as quick as possible
o results  concepts that had to travel less (e.g. nodes btw CATS and CLAWS was faster than
CATS and HEARTS) due to efficacy of model (cognitive economy)
 however, non-redundancy does not always hold using this model
o e.g. property of ‘having feathers’ should be associated w BIRDS node rather than ROBIN node,
PIGEON node, etc…. but pts respond v quick to ‘peacocks have feathers’
 still, we can often predict the speed of knowledge access by counting the number of nodes pts must traverse
in answer a question

Propositional networks
 Need more than simple associations btw concepts (represent diff btw ‘sam has a dog’ & ‘sam is a dog’)
 Early theory  introduced diff types of associative links
o ISA: represent equivalence
o HASA: represent possessive relations
 Later theories  John Anderson and his theory of propositions:
o The smallest units of knowledge that can be either true or false
o Eg. ‘children love candy’ = proposition; ‘love candy’ is not
 MEMORY | 20

 PIC:
o Ellipses identify the propositions
themselves
o Associations connect an ellipse to
the ideas that are the propositions
constituents
o The associations are labelled, but
only in general terms (which
enables us to distinguish the
proposition ‘dogs chase cats’ vs
‘cats chase dogs’)
 THIS IS AN EXAMPLE OF LOCALIST
REPRESENTATION:
o Individual ideas are represented w
local representations
o Each node = one idea

Distributed processing
 Distributed representations is when each
idea is represented by a means of specific
pattern of activation across the network 😊
(not by specific nodes)
o Means certain nodes can fire for diff ideas, but it is the node wrt the PATTERN of nodes activated
that allows for that idea to come forth
 Also called parallel distributed processing (PDP)
o Ppl argue this makes biological sense bc our brain works via parallel processing
 HERE KNOWLEDGE IS SEEN AS A POTENTIAL, NOT A STATE (eg. have WASHINGTON node and
PRESIDENT node, connections are in place for this)… learning is an adjustment of connection weights