Exam 2 Flashcards
semantic memory
- knowledge of the world
- not tied to a particular event (decontextualized)
- organized by meaning or content
-ie: what is taller? grass or trees?
episodic
- something you experience
- tied to particular event (contextualized)
- organized temporally
- amnesia affects episodic memory much more than semantic memory
- ie: what did you have for breakfast today?
decontextualization
as certain events/objects keep recurring, you knowledge of them is less tied to the contexts in which you encountered them—> become semantic memories
K.C. from text
lost episodic but not semantic memory
Butters & Cormak (1986)
- Amnesic through Korsakoff’s syndrome
- written autobiography–> tested him on semantic memory as recorded in autobio
- recognized scientists mentioned
Lewis & Anderson (1976): Fantasy Facts
-taught subjects fantasy facts about famous people
-had subjects verify sentences that described real/fantasy facts
prof: goldstein: semantic + episodic memory interact so that episodic memories become more like semantic memories over time
ie: old memories are “known,”; recent memories are “remembered”
prof murphy: the principles that govern episodic memory are the same as those that govern semantic memory
-as context is lost–> old episodic memories act like semantic memories –> theyre in the same system
Conclusion:
- episodic & semantic memories do not seem to be in different systems
- amnesia affects both
- memory for one affects the other
- some older episodic memories seem very much like semantic memories
Why do SM and EM seem to differ?
- contents often differ
- SM is usually much more familiar + more engrained, encountered across many contexts
- more resistant to loss + less tied to a specific event BUT that
- rules of encoding + retrieval seem the same for both
explicit memory
-direct tests; declarative memory; conscious memory
you are asked to remember something or to reproduce something that happened earlier
-depends on perceptual details
-affected by amnesia
-relational`
implicit memory
- indirect tests; procedural memory; unconscious memory; skill learning
- you are asked to do some other task and an earlier event might influence your performance on that task
- does not depend on perceptual details
- generally not affected by amnesia
- not relational
Cohen & Squire (1980): Classic implicit memory study of mirror-reading
1) What kind of memory is this?
2) Are people asked to recall or recognize something
3) Does earlier experience affect their performance on the task
4) How do you prove that 2 kinds of memory are in different systems
-when asked to read the words twice, subject improves
1) implicit– repetition priming
2) no
3) yes
4) typically try to find dissociations (variables that affect one kind of memory but not the other (has been found in amnesia))
retrograde amnesia
loss of memory for info before injury
anterograde amnesia
loss of memory for new info post injury
Graf, Squire, Mandler (1984)
subjects (normal + amnesic) read a list of words–> took a test for word stem
-were told to either finish the stem with a word on the list (cued recall)
-or finish the stem with the first word that comes mind (Stem completion task)
Results:
-Amnesic subjects do terribly on the recall test; normal subjects did ok
-both groups of subjects perform well on the “first word that comes to mind”
-even amnesic subjects tend to give words from the first list
-but they didn’t forget the words
-this is another dissociation
propaganda effect
subjects are more likely to rate statements they have read/heard before as being true
classical conditioning (implicit)
occurs when the following 2 stimuli are paired
1) neural stimulus
2) conditioning stimulus
Jacob & Dallas (1981) : typical levels of processing experiment
- subjects answered questions about each word, sound, letter, sound, meaning–> should lead to better memory
- at test they either got a recognition memory test
- or identified briefly-presented words (“perceptual identification”)
Weldon & Roediger (1987)
- in general, people remember pics better than words
- but what if you had an implicit task that was better suited to verbal materials
- word fragment completion test
- subjects studied pictures and words–> then took a recall test (explicit–>pictures better) or a word fragment completion task (implicit–>words better)
- another dissociation
- as a general rule, implicit tasks are sensitive to the exact way a stimulus was presented but explicit memory isnt (as much)
procedural memory
- knowing how
- memory is not in a special store but is part of the proedures (skills) that you learn
declarative memory
knowing that
- requires the hippocampus to be encoded
- inherently relational
- relations btwn the parts of a single event (e.g.- the words on a sentence/on list)
- relations between items + context (e.g.- this word was said 10 min ago by..)
Chun & Jiang (1998): Implicit Contextual Learning
-robust memory for visual context exists to guide
spatial attention
-Targets appearing in learned configurations were detected more
quickly
-compares conscious access of memories explanation with relational explanation
-driven by incidentally learned associations between spatial configurations (context) and target locations
-amnesics show impaired relational learning
-hippocampal memory system may be important for explicit and explicit measures of relational memory in humans
implicit tasks are not relational, explicit tasks are relational
models of memory
- not an explanation of a single phenomenon but a theory of all memory that attempts to explain most phenomena at once
- formal
- mathematical/computationa
- # of models rely on associations as the building blocks of memory (2 elements being paired together, or in chains(if you forget one link of the chain, can’t remember the rest of the chain))
How are memory lists structured?
Context
| | | |
word 1—word 2—word 3—word 4
SAM: Search of Associative Memory (Shiffrin & Raaijmakers 1992)
- each item=image
- each image consists of a set of features, including the context, the name +property of the item, associations to other images
- an item gets from STM–> LTM the more time it takes in STM (ie- the more it is rehearsed)
Remembering in SAM/
- at retrieval, a cue activates the items that are similar to it or associated w it
- the diff items “fight it out” (compete) + the one with the most activation wins, if its activation reaches a criterion
what is the cause of all forgetting?
1) retrieval failure: think about the “bad” condition in encoding specificity experiments; easier as cues get better
2) decay: breaking down, may be due to time/gradual; hard to say its just decay that is causing memory loss
3) interference: proactive (old get in the way of new + retrograde(new gets in the way of old)
1) spacing effect
2) primacy effect
1) -given a set # of study trials, memory is better the more spread out they are
- spaced >massed
- SAM: more contextual cues
2) -more things in working memory the better you’re going to put things into LTM
formal models of memory
- help to show that our theories are internally consistent
- help us work out specific details b/c your cpu doesnt understand words like “similar”/”long time”
What are some drawbacks of FMM ?
- can be difficult to understand
- you have to specify details even if you don’t know them
Understanding + remembering discourse & events
- without knowledge on the topic, could be impossible to correctly understand a passage/event
- people draw inferences in order to understand + remember those inferences
Result: much that we remember was not actually seen/heard– contextual
-constructive processes of memory
constructive processes of memory
much that we remember was not actually seen/heard
memory= constructive
refers to what you do @ encoding to fill in gaps in a story or event or pic
memory=reconstructive
refers to what you do @ retrieval to fill in gaps in your memory
schema
a way of representing knowledge
- consists of “slots” (basic dimensions that describe the thing)
- consists of “fillers” (features that specify the dimensions of slots)
- fillers are ordered by typicality or frequency
- theme: people remember the interpreted content, not exactly what they head or read
- memory for exact wording of spoken or read material is bad. memory for meaning is good or at least better
script
kind of schema describing an event with a regular order