Schemas and Scripts Organizacon of Knowledge
Total Page:16
File Type:pdf, Size:1020Kb
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 recep6onist – 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 leO feeling extremely – He worked on the exam. later to a scien6st (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 6me you dined out… Schemas • Can you remember… • Test example • Assump6ons are filled in from Less – What was waitperson’s name? – John picked up a schemas schema- test from the TA. relevant – Meaning = input + ac:vated knowledge – How they took your order, word for word? – He worked for an • (Hmm, what if you ac6vate the wrong hour and twenty knowledge?) – Was the service good? minutes. – You’re acvely construc:ng meaning Schema- – He leO feeling relevant – What you ate? – To understand is to come up with an integrated extremely representa:on. – Whether the food contained metal shreds? worried. 1 2/27/12 Schemas Schemas • Ac6vang 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 ac6vang 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 arac6ve individual. • Slots are specific and • She then no6ced a ring on that person’s leO 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 cigare9e, cigare9e, 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) • Ac6vated schemas by juxtaposing similar pictures… Now recall picture A given a cue Cue: woman cigarele Cue: woman cigarele 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) • Ac6vated schemas led to different encoding of • Ac6vated 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 – Scien6st • 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 ac6ons • Of those using pronoun in discussing this on a problem – Causal links between events set the first 6me 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-ac6on passages – Present 10 lists of ac6ons – Later, gave 6tles & asked to recall exactly – Some lists in order, others out of order • Correctly recalled: 3 out of 6 – Asked people to recall ac6ons • 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 par6cularly – 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 • Beler captured by PDP-type models Is memory accurate? When memory goes bad Reconstruc6ve memory False memory: your data N=219 • In recalling an event, some6mes other stuff is False memories! 0.9 recalled with it that’s not part of it 0.8 – Esp. for complex events, may put mul6ple pieces 0.7 together--reconstruct 0.6 *** 0.5 – Errors when you probe with cues from part of a 0.4 recollec6on 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 Reconstruc6ve memory Reconstruc6ve 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 Reconstruc6ve memory Reconstruc6ve memory • Effects of retrieval cues – Anderson & Pichert (1978) • Stereotypes (e.g. • Par6cipants read burglar/home buyer story scien6sts, band nerds) • Asked to recall details from one perspec6ve – 64% perspec6ve-relevant facts recalled – guide retrieval of events – 46% other-perspec6ve facts recalled • How badly injured was this – (I.e., perspec6ve maers) person? • Then asked to recall from the other perspec6ve – 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 lille clumsy • Recalling from a perspec6ve can alter memory itself www.xkcd.com Reconstruc6ve memory Memory issues in real life Big point: remembering isn’t just about pulling • Eyewitness tes6mony an experience out of a lille 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 some6mes get you in trouble.) 5 2/27/12 Eyewitness tes6mony Eyewitness tes6mony • Assump6ons • The misinformaon effect (LoOus, 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 – Ques6onnaire • Misleading ques6on (“stopped at stop/yield sign?”) – 75000 suspects ID’ed per year • No misinformaon (“stopped at intersec6on?”) – Some6mes right, but not always – Pick YIELD slide or STOP slide • No misinformaon: 85% correct • One problem: Misleading quesAons • Misleading-Q group: 38% correct :-( – Memory has been overwriIen/revised Eyewitness tes6mony Misinformaon effect • The misinformaon effect (LoOus, Burns, & Miller, • Source confusion explanaon 1978) – Like trace interference – How does this happen? • Overwri6ng (“destruc6ve updang”) – Original memory is there, but not clear where it • Source confusion came from • Misinformaon acceptance – Lindsay & Johnson (1989): • If given a misleading sugges6on, 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 LoOus, M&Z weren’t aware • To test non-awareness of where info came from: – See theO 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--oOen recalled narrave informaon!! • Result: screwdriver misleads, wrench not • Source confusions do occur • M&Z: misinformaon acceptance – Retrieval cues? (Screwdriver > wrench) – Does not support overwri6ng Screwdriver must not overwrite because hammer “sll there” Misinformaon effect Memory issues in real life • Effect is widely accepted • Eyewitness tes6mony • Underlying explanaon, less