2/27/12

Organizaon of knowledge

• Marie’s phone rings. She answers. A woman’s voice shrieks, “They’re making a movie about Schemas and Scripts them!” • Marie is confused. • Helps if you know that – Caller was vet receponist – Marie has two cats named Shaggy and Velma • You have to give people enough context.

Organizaon of knowledge Organizaon of knowledge

• Test example • What we know • Schema (pl. “schemata” or “schemas”) – John possessed a copy of – John picked up a test – Knowledge about complex situaons from the TA. the test. – – He worked for an hour • What we “know” Helps you understand the current situaon and twenty minutes. – John took an exam. • You’re not trying to remember a list of events to report – He le feeling extremely – He worked on the exam. later to a scienst (as in a serial recall experiment)— worried. – He was worried about • You want to get what you need out of the situaon his exam performance. – Top-down knowledge (fill-in)

The last me you dined out… Schemas

• Can you remember… • Test example • Assumpons are filled in from Less – What was waitperson’s name? – John picked up a schemas schema- test from the TA. relevant – Meaning = input + acvated knowledge – How they took your order, word for word? – He worked for an • (Hmm, what if you acvate the wrong hour and twenty knowledge?) – Was the service good? minutes. – You’re acvely construcng meaning Schema- – He le feeling relevant – What you ate? – To understand is to come up with an integrated extremely representaon. – Whether the food contained metal shreds? worried.

1 2/27/12

Schemas Schemas

• Acvang different knowledge results in a • What they are different understanding – Your knowledge about how the world works – Anderson et al. (1977): prisoner story (based on your experiences) • Could also possibly be construed as wrestling – General--about type of situaon (not token/ • Test on people who are/aren’t acvang wrestling episode) knowledge a lot – Structured relaonships, not just set of facts – Are (PE majors): 64% “wrestling” responses – Not (music majors): 28% “wrestling” responses – Used to understand the world

Example: CLOTHING schema Schemas

Slots Values • Can have embedded schemas • Torso covering: • T-shirt – Going-to-dinner schema Purchase-stuff schema • Leg covering: • Jeans Nothing – Ice-cream-truck schema • Head: • Nothing Jeans • Benefits of schemas – Infer things that aren’t directly observed • Feet: • Sandals , sneakers, pumps – Predict upcoming stuff • Jan was at a party talking to a very aracve individual. • Slots are specific and • She then noced a ring on that person’s le hand. contain defaults • What is Jan going to do?

Schemas Markman & Gentner

• Influences on memory Show 2 pictures Show 2 pictures – Place schemas Man, who Man, who Woman with has dropped has dropped Girl looks at • Dorm room; grad student office camera cigaree, cigaree, xmas tree, photographs • Are there books in your TA’s office? paints paints holding dripping boy scout – Brewer & Treyens (1981): 30% say “yes” when no books were woman woman candle actually present Rate similarity Rate similarity Picture-taking schema acvated Arson schema acvated – Markman & Gentner (1997) • Acvated schemas by juxtaposing similar pictures… Now recall picture A given a cue

Cue: woman cigaree Cue: woman cigaree Mem: GOOD BAD Mem: BAD GOOD

2 2/27/12

Schemas Schemas

• Influences on memory • Influences on memory – Place schemas (Brewer & Treyens, 1981) – Place schemas (Brewer & Treyens, 1981) – Markman & Gentner (1997) – Markman & Gentner (1997) • Acvated schemas led to different encoding of • Acvated schemas led to different encoding of depicted events depicted events • Poor encoding of schema-irrelevant events – Boutla et al. (2004)

Schemas Scripts

• Stereotypes • Specific type of schema – Scienst • Used for stereotyped event sequences • Test tubes and symbols – Going to dinner, geng ice cream, taking exam • Caucasian (?) • Male • Contains: – Boutla et al. – Set of ordered acons • Of those using pronoun in discussing this on a problem – Causal links between events set the first me this class was taught, all used he/him • E.g. p depends on good service • Boutla is a woman!

Scripts Scripts

• Evidence for scripts (Bower et al. 1979) • Evidence for scripts (Bower et al. 1979) – Study 1 – Study 2 – Presented 6-acon passages – Present 10 lists of acons – Later, gave tles & asked to recall exactly – Some lists in order, others out of order • Correctly recalled: 3 out of 6 – Asked people to recall acons • Filled-in: 1 extra fact that didn’t take place • In order lists: 50% correct order at recall – Recall was based on familiar series of events • Out of order lists: only 18% correct order – Recall was structured around familiar order

3 2/27/12

Scripts Schemas

• Problems • General problems – What about things that don’t have a parcularly – Slots: stereotyped order of occurrence? • funcon dineOut(diner, food_type, transportaon) • Going to the bathroom at a restaurant – Dependence between slots • Going-to-dinner schema – If diner is person A, Thai food – If diner is person B, nothing with meat • Beer captured by PDP-type models

Is memory accurate?

When memory goes bad

Reconstrucve memory False memory: your data N=219 • In recalling an event, somemes other stuff is False memories! 0.9 recalled with it that’s not part of it 0.8 – Esp. for complex events, may put mulple pieces 0.7 together--reconstruct 0.6 *** 0.5 – Errors when you probe with cues from part of a 0.4 recollecon to retrieve the rest 0.3

Average FM per trial 0.2 0.1 0 Unrelated distractor Special distractor Error types 194 out of 219 of you made more “special” than unrelated errors.

4 2/27/12

Reconstrucve memory Reconstrucve memory

• Example: • Error-prone memories (episodes) – Actual event: – Poorly encoded ones • Not processing lecture much • dinner (Cuban) & movie (Wordplay) w/Julia – – Probe: movies seen with German friends Ones similar to other memories (≈ encoding cues) • » Wordplay Winter cog sci lectures • German postdoc friends in Philly » An Inconvenient Truth – Pull up wrong movie – Not recent » Remember: Cuban & Inconvenient Truth w/Julia • General picture: if you can retrieve only bits and pieces, you fill in to get a whole memory

Reconstrucve memory Reconstrucve memory

• Effects of retrieval cues – Anderson & Pichert (1978) • Stereotypes (e.g. • Parcipants read burglar/home buyer story sciensts, band nerds) • Asked to recall details from one perspecve – 64% perspecve-relevant facts recalled – Guide retrieval of events – 46% other-perspecve facts recalled • How badly injured was this – (I.e., perspecve maers) person? • Then asked to recall from the other perspecve – Probably encoding effects – Another 10% of facts suddenly came to mind! too – Retrieval alone can “jog” memory • Expectaon that band – Tversky & Marsh (2000) nerds are a lile clumsy • Recalling from a perspecve can alter memory itself

www.xkcd.com

Reconstrucve memory Memory issues in real life

Big point: remembering isn’t just about pulling • Eyewitness tesmony an experience out of a lile pigeonhole in – Misinformaon effect your mind. You filter it through the rest of your world knowledge. • Flashbulb memories • False memory (Which usually works, but can somemes get you in trouble.)

5 2/27/12

Eyewitness tesmony Eyewitness tesmony

• Assumpons • The misinformaon effect (Lous, Burns, & Miller, 1978) – It’s accurate – Slide show of car accident – Certainty and accuracy are correlated – Half saw YIELD sign, half saw STOP sign • Data – Quesonnaire • Misleading queson (“stopped at stop/yield sign?”) – 75000 suspects ID’ed per year • No misinformaon (“stopped at intersecon?”) – Somemes right, but not always – Pick YIELD slide or STOP slide • No misinformaon: 85% correct • One problem: Misleading quesons • Misleading-Q group: 38% correct :-( – Memory has been overwrien/revised

Eyewitness tesmony Misinformaon effect

• The misinformaon effect (Lous, Burns, & Miller, • Source confusion explanaon 1978) – Like trace interference – How does this happen? • Overwring (“destrucve updang”) – Original memory is there, but not clear where it • Source confusion came from • Misinformaon acceptance – Lindsay & Johnson (1989): • If given a misleading suggeson, it is also recalled and it may be incorrectly remembered as the thing you saw

Misinformaon effect Misinformaon effect

• Misinformaon acceptance • Lous: overwring – You’re totally aware that you didn’t know • McCloskey & Zaragoza (1985): misinformaon – But you assume that the misleading informaon acceptance was correct • Why would a lawyer say something inaccurate?

6 2/27/12

Misinformaon effect Misinformaon effect

McCloskey & Zaragoza (1985) • But Lindsay (1990): source confusions • Set-up – Subjects in Lous, M&Z weren’t aware • To test non-awareness of where info came from: – See the from under hammer – Event happens – Narrave w/ or w/o misleading screwdriver – Misleading narrave happens • Test – 48 hours go by… – Lous version: hammer vs. screwdriver – Just before test: “narrave was made up” • If they know the source, should dismiss misleading info – Modified version: hammer vs. wrench • But they didn’t--oen recalled narrave informaon!! • Result: screwdriver misleads, wrench not • Source confusions do occur • M&Z: misinformaon acceptance – Retrieval cues? (Screwdriver > wrench) – Does not support overwring

Screwdriver must not overwrite because hammer “sll there”

Misinformaon effect Memory issues in real life

• Effect is widely accepted • Eyewitness tesmony • Underlying explanaon, less so – Misinformaon effect – Some sll hold to overwring • Flashbulb memories – More likely: source confusion • False memory • Real and fake info compete for recognion • Plausibility important too – “The car stopped at the stop sign, then a flying saucer hit the pedestrian. The car then fled the scene.”

Flashbulb memories Flashbulb memories

• Memories that “stand out” from others • Indelibility: Brown & Kulik (1977) – College admission – Asked N=80 about JFK • What were you doing when you found out? – First date – 79/80 remembered – Historical events (easier to invesgate) • 13-year delay (1963-1977) • JFK – Argued for disnct biological mechanism for storing • Challenger surprising memories • 9/11 • Would have been crucial to survival • Lile forgeng • Really as “indelible” as they seem? • Highly detailed, including circumstances surrounding

7 2/27/12

Flashbulb memories

• Neisser & Harsch (1992): not so fast. – Morning aer 1986 Challenger explosion When I first heard about the explosion I was sing on my freshman dorm room with my roommate and we were • Quesonnaire to ugrads at Emory watching TV. It came on a news flash and we were both totally – What happened? shocked. I was really upset and I went upstairs to talk to a – What were you doing? friend of mine and then I called my parents. Neisser & Harsch, 1992, p. 9 / Textbook p. 234 – Who told you? (and so on) – Aer 2.5 years, recontacted to be in study • Only 25% recalled taking quesonnaire

Flashbulb memories

• Neisser & Harsch (1992): not so fast.

When I first heard about the explosion I was sing on my – Morning aer 1986 Challenger explosion freshman dorm room with my roommate and we were • Quesonnaire to ugrads at Emory watching TV. It came on a news flash and we were both totally – What happened? shocked. I was really upset and I went upstairs to talk to a – What were you doing? friend of mine and then I called my parents. – Who told you? (and so on) Neisser & Harsch, 1992, p. 9 / Textbook p. 234 – Aer 2.5 years, recontacted to be in study • Only 25% recalled taking quesonnaire • Only 3/44 had perfect recall (assuming q’aire true) • Lile relaon between confidence & accuracy

Actual: found out in class, felt sad, watched TV for details

Flashbulb memories Flashbulb memories

• Objecons to N&H • Maybe no special mechanism – Conway et al (1994) – Important, so likely to get rehearsed a lot • Challenger explosion not consequenal for people • Tested memory for Thatcher’s 1990 resignaon – Very unusual--less interference – UK, US, Denmark – Strong emoonal tone may affect memorability – Tested at 2 weeks & 11 months » UK: 86% highly accurate » US, Denmark: 29% • Good challenge • But maybe less well encoded to begin with

8 2/27/12

Memory issues in real life Recovered memories

• Eyewitness tesmony • Terrible event forgoen for many years – Misinformaon effect • Somemes necessitates criminal prosecuon • Flashbulb memories – Statute of limitaons excepon--criminal doesn’t • False memory benefit from causing witness trauma • Somemes the accused protests

Recovered memories Recovered memories

• Agreed by all pares • Why aren’t these extra memorable? – Child abuse (or other criminal acts) are frequent, – Repression. and should be punished • Is repression real? – But also, innocent people shouldn’t be punished – Maybe. – Somemes no way to verify/falsify • Proponents: once retrieved, very accurate and vivid – Again, hard to verify – Remember that confidence ≠ reality

Recovered memories Recovered memories

• Opponents • False memories induced – No evidence that memories real – Geng lost in a mall (but not enema) – Therapists (trying to be helpful) may ask – Details of childhood crib misleading quesons or encourage erroneous reconstrucon of events • Alien abducon vicms: Clancy et al. 02 – We know recall can be inaccurate – More suggesble • Response: Sure, but lab experiences can’t achieve – More prone to false memory effect (see Ch. 6) ecological validity • – Emoonal trauma very strong Familiarity = fame – Extended abuse ≠ stop sign!

9 2/27/12

Memory issues in real life

• Eyewitness tesmony – Misinformaon effect Memory & Metamemory • Flashbulb memories • Recovered memory • False confession

A bit more on recovered memories Final issue: false confession

• Suggesbility • Causes: – Emoonal stress – Porter & colleagues: – Social pressure • Lost in mall (15% implanted) – Suggeson • Enema (0% implanted) • Distrust memory enough • Serious childhood animal aack: about 30% – If you think it’s possible to repress horrific memory, you – Not a minor incident might believe you’ve done it – A large proporon of people are suggesble • “Interrogave suggesbility” (Gudjonsson) – Tested people who made confession, later retracted – They were more prone to suggesve quesons

Final issue: false confession Memory issues in real life

• Experimental false confession (Kassin) • Eyewitness tesmony – “Type spoken leers--don’t hit ALT!” – • Fast or slow Misinformaon effect – Experimenter: false accusaon • Flashbulb memories – Subject: no, I didn’t 1 • Recovered memory – Confederate: “I saw you do it!” ( /2 subj’s) – Overall, 70% signed a confession Low certainty of own • memory False confession – If typing fast and confederate, all • And 35% had a detailed recollecon about it!! Misleading queson • But again, problem of scale – ALT key isn’t exactly a dagger

10 2/27/12

Memory issues in real life Hypnosis and memory

• Common threads: • Used in therapy, on eyewitnesses – Great confidence • Scienfic findings – Possible inaccuracy – Does not improve list recall • Driest lab test possible • Hypnozed subjects recall more vs. controls – Stuff on the list – Stuff NOT on the list • Overall, no improvement (+signal but +noise too) • Somemes not as well as control subjects who are encouraged strongly to try their best

Hypnosis and memory Hypnosis and memory

• Used in therapy, on eyewitnesses • Used in therapy, on eyewitnesses • Scienfic findings • Scienfic findings – Does not improve list recall – Does not improve list recall – Does increase confidence in memory – Does increase confidence in memory • • Some states exclude tesmony

Alternaves to hypnosis “Special” memories?

• Are there beer ways? • Eyewitness tesmony – Cognive interview (Fisher & Geiselman) – Misinformaon effect • No misinformaon provided • “Report everything” (not just specific quesons) • Flashbulb memories • Ask witness to reinstate context • Recovered memory • Ask witness to take different perspecve • False confession – Results: beer recall, with a lile bit of erroneously-recalled info Upshot: there’s no such thing as a free lunch. – Used to train English & Welsh police These memories are prey to everything that “normal” memories are suscepble to, despite the subjecve feeling of certainty that oen surrounds them.

11 2/27/12

Really, really good memory

• What about those people who can memorize Extraordinary memory phone books? – Yes, they exist. and metamemory – No, they’re not fundamentally different from anyone else.

Really, really good memory Really, really good memory

• Individual differences • Strategies we’ve already discussed – Aren’t we all working with the same equipment? – Pay aenon at encoding! • If you don’t think about someone’s name, you won’t remember it – Factors later • Movaon to learn – Rehearse in mulple ways • More interested in material • Relate to knowledge (looks like a Roxanne I went to grade school • Just plain beer memory with) – Evidence: • Elaborate (imagine her at La Jolla Cove where there were rocks and sand) • Strategies – Set up a good retrieval plan • Prior knowledge • Imagine (or pracce) giving your report in auditorium • Retrieval cues will beer match encoding cues

Really, really good memory Memory experts

• Exisng schemas/domain knowledge • 7 ± 2 is normal STM limit – Spilich et al. (1979): baseball • Chunking helps (even if not meaningful) – Memory for melodies • Chase & Ericsson (1981)

• In an unfamiliar domain – Things go “in one ear and out the other” – Expert performance seems ‘magical’ to you

12 2/27/12

Memory experts Memory experts

• Chase & Ericsson (1981) • – Trained people to chunk Chase & Ericsson (1981): HOW? – Subject 1: runner, chunked into running mes for Incoming: 2141034084750 different races; got up to 80 digits 2141034084750 Huff, puff – Crazily-talented person? – Subject 2: trained w/same strategy, got up to 40 digits with ≈ same rate of improvement marathon 10 miles LTM LTM 100-yd mile dash

No strategy Strategy

Memory experts Memory experts

• Chase & Ericsson (1981) • Mnemonists – Great, but so what? – Use mnemonics – – Wouldn’t generalize to (e.g.) leer sequences or Examples: • Bizarre images grocery lists – Problem set 3 grows legs, leaps into your bag before Thursday – Your chunks have to match what you’re morning class • Method of loci (for ordered things) memorizing – Mentally navigate a familiar path – Put the to-be-remembered things at points along path • PPMDAS, EGBDF (order of operaons, lines on staff) – Point: you aren’t going to forget sentence order – Banks off of knowledge of grammar

Memory experts But wait… 1 2 3 4 2 3 4 5 • Luria (1968): “S” 4 5 6 7 … • Are there real memory experts? – 70 words in memory span task • And backwards • And the next/previous word, given any word • Part 1 – Synaesthete (A is blue, etc.)--rich encoding • Part 2 – Had to also use method of loci and others – Quit journalism to become professional mnemonist – Somemes didn’t pick up on simple consistencies

Thanks to a student from last year for bringing this to the class’ aenon.

13 2/27/12

Thought quesons Metamemory

• Are these people qualitavely different from • Knowing what you know (and how well you people with “regular” memory? know) • Think about a sports-obsessed or music-obsessed friend. Would astounding content knowledge in – Do I know James Bond’s phone number? their domain of interest surprise you? – Is it reasonable to forget my keys? • Does somebody’s life provide an unusually – Do I need to read over my notes again? coherent reference frame? • What if people mentally rehearse the events in • Are we good at this? their lives constantly—could that constant – E.g. false confidence in memory rehearsal lead to beer encoding? – Usually we are good – How would you test this?

Metamemory skill Metamemory skill

• Nelson et al. (1994): good metamemory • Infer well-encodedness by property of – Parcipants learned new (Swahili) words memory – Experimental group esmated how well each – Level of detail word was learned – Speed of recollecon – Experimenters used these esmates to increase/ • Big area of interest decrease study of poorly/well-learned words – Experimental group > control group

14