Reading...
Play button
Play button
Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
112 Cards in this Set
- Front
- Back
|
Mental Imagery: what is it?
|
is representation of
something not currently being sensed by sensory systems |
|
|
Historical perspectives on mental imagery
|
Introspectionists
Behaviorists Cognitive revolution |
|
|
Introspectionists
|
– Posited that imagery represented thought
|
|
|
3 Theories - Mental Imagery
|
Dual-Coding Model
Propositional Theory Functional-Equivalence Theory |
|
|
Dual-Coding Model
|
• We use imagery for concrete words but not
abstract words – Memory performance for word pair task • Performance best when both word types matched • Abstract – abstract (Democracy-intellect) • Concrete – concrete (Tree-pencil) • Suggest dual coding for visual and verbal informa;on |
|
|
Dual coding model (Paivio)
|
Abstract nouns:
can be coded only verbally. Concrete nouns: can be coded verbally or in terms of images. |
|
|
Propositional Theory
|
• Mental representa;ons
– not stored in form of images – resemble abstract proposi;ons and they take the form of a “rela;on (argument)” • Red (car) • Typed (Camille, keyboard) • Mental imagery comprised of abstract proposi;ons – Proposi;on = meaning underlying rela;onships among concepts – Represent conceptual objects and rela;ons in a form not specific to any language or sensory modality – “a universal, amodal, mentalese” |
|
|
Kosslyn’s counter to propositional theory (1980)
|
Proposi'on Analog
Rela;on No dis;nct rela;on Syntax No syntax Truth value Truth value only when described Abstract Concrete Not Spa;al Spa;al (if an image is anything, it’s spa;al) |
|
|
Functional-Equivalence Theory
|
We represent and use visual imagery in a way that is functionally equivalent to physical percepts
|
|
|
Evidence: Mental Imagery
|
Mental Rotation: Shepard & Metzler(1971)
studies and what they suggest Image Scaling Easier to distinguish featural details of large objects compared to small objects Respond more quickly to questions about features of large objects we observe than small objects Kosslyn's approach Image Scanning Studies |
|
|
Mental Rotation: Shepard & Metzler(1971)
|
• Rotate figures 0°-180°
• RTs increase as func;on of degree of rota;on analogous to rota;ng the objects in “real space” • Remarkably consistent rela;on between angle of rota;on and RT Mental Rota;on • Same effects for digits, leQers, body parts • Suggests – visual images have aQributes of actual objects in the world – take up “mental space” in the same way that physical objects take up physical space in the world – we manipulate them the same way that we manipulate objects in real world |
|
|
Image Scaling
|
• In visual percep;on, there are limita;ons on
resolu;on – Easier to dis;nguish featural details of large objects compared to small objects – Respond more quickly to ques;ons about features of large objects we observe than small objects |
|
|
Image Scaling: Kosslyn's approach
|
Kosslyn (1975) - Ss imagine different pairs of
animals to manipulate rela;ve size – Blue whale + dolphin – Dolphin + shrimp |
|
|
Image Scanning
|
In percep;on takes longer ;me to scan longer
distance vs. shorter distances • Study & learn map to criterion |
|
|
Limitations of Functional Equivalence Hypothesis
|
Reed (1974) – is there
a parallelogram in the Star of David? • Chambers & Reisberg (1985) –ambiguous figures |
|
|
What is imagery for?
|
Memory, Make implicit knowledge conscious, Prepare for future actions,
Problem solving |
|
|
What is language for
|
Communication
|
|
|
In what kind of communication environment did human languages evolve?
|
Spoken, not wri[en
– Physical proximity of speaker and listener – Probably one-to-one or one-to-few |
|
|
Properties of Language
|
Communicative
Arbitrarily symbolic Regularly structured Structured at multiple levels Generative Dynamic |
|
|
Conversational as “joint action,”
|
like shaking hands
Rapid corrections All kinds of non-word feedback (e.g., “uh-huh”) Conventions are established |
|
|
Grice's Conversational Maxims
|
Quan=ty
– Be informa=ve – But not too informa=ve • Quality – Tell the truth • Rela=on – Be relevant • Manner – Be clear • A “normal” u[erance sa=sfies these maxims – Viola=ng these maxims produces either social or communica=ve difficul=es – What are some examples of viola=ons of these? • Regional/cultural differences in prac=ces and maxims |
|
|
Example evidence in favor of rule-based view of language:
|
U-shaped development
– Children start out producing the correct irregular forms, e.g., “went” – Then over-generalize a rule, e.g. “goed” – Then come back to the correct irregular form – Over-applica=on taken as evidence for rule-following |
|
|
Counter-argument to rule-based view (Language)
|
Systems with no explicit representa=on of the rule
anywhere in the system (e.g., connec=onist networks) can show u-shaped development as well |
|
|
More modern alternative view (Language)
|
What is learned are not rules, but statistical tendencies
|
|
|
Structure of Language
|
Phonology
Morphology Morpheme and Lexicon Syntax Semantics Discourse |
|
|
Phonology
|
Smallest unit of speech
– Different languages use 20-80 (English ~ 40) |
|
|
Morphology
|
– Morpheme - the smallest unit that denotes meaning
• Root words, Suffixes, Prefixes • Cake Chair Boy Pre- Non- Un- -ly -ist -ness – Lexicon - En=re set of morphemes for a language |
|
|
Syntax
|
– Rules used to to combined words to form a sentence
-describes the gramma=cal structure of language – The rules for forming sentences from words • In English, these are mostly rules about word order – There are some rules involving inflec=on • Such as adding “-ed” to form the past tense – In many languages, most of the rules involve inflec=on |
|
|
Seman=cs
|
– The study of meaning
|
|
|
Discourse
|
– The study of interac=ons between the context and language
|
|
|
Modularity of language
|
• This is the no=on that language is separate
from the rest of human cogni=on, that there is a “language module” • Evidence in favor – Cri=cal period for learning language – However this evidence is not strong |
|
|
Language as Thought
|
Language is one of the key areas of inquiry
that led to the downfall of behaviorism – Very hard to imagine language without internal representa=on or state • One of the founders of the behaviorist movement (Watson) claimed that all so-called “thinking” was simply “talking to yourself” – That is, explicit behavior |
|
|
Sapir-Whorf Hypothesis
|
Whorf was a student of Sapir’s who took the original
ideas and fleshed them out further • A famous case: counterfactuals in Chinese – A psychologist who knew some Chinese found it difficult to express counterfactuals when speaking Chinese • Rela=ng to or expressing what has not happened or is not the case, e.g. If kangaroos had no tails, they would topple over – Did studies on the ability of Chinese speakers to reason about counterfactuals • Showed advantage for English speakers – Chinese student at Harvard found this odd and reviewed s=muli; terrible Chinese transla=on |
|
|
Why speech perception is hard
|
Co-ar=cula=on
– Phonemes overlap, producing one phoneme affects produc=on of following phonemes • Non-invariance – Sound pa[ern for phoneme constantly changes • Segmenta=on – Speech is con=nuous stream with few periods of silence |
|
|
How to we resolve ambiguities (Speech)
|
Phoneme restora=on effect
• Categorical percep=on • McGurk effect |
|
|
Phoneme Restoration: Warren & Warren (1970)
|
Warren & Warren (1970)
– It was found that the *eel was on the axle – It was found that the *eel was on the shoe – It was found that the *eel was on the orange – It was found that the *eel was on the table – * was a cough but it was heard as the missing phoneme implied by the context • Contextual factors influence our percep=on |
|
|
Categorical Percep=on
|
• Voice onset =me (VOT)
– The =me between the beginning of the pronuncia=on of the word and the onset of the vibra=on of the vocal chords • Example “ba” and “pa” – "ba" your vocal chords vibrate right from the start – "pa" your vocal chords do not vibrate un=l aker a short delay – The sounds "ba" and "pa" differ on the con$nuous dimension of VOT – How can listeners differen=ate between /p/ & /b/? • Where does one begin and the other end? |
|
|
Voice onset time (VOT)
|
– The =me between the beginning of the pronuncia=on of the
word and the onset of the vibra=on of the vocal chords |
|
|
McGurk & MacDonald (1976)
|
• To resolve ambigui=es
– We use both visual and auditory cues to perceive speech. • Developed clever tes=ng paradigm to prove this |
|
|
McGurk effect
|
• Lip movements = “ga”
• Soundtrack = “ba” • What do you hear? • McGurk & MacDonald (1976) found that people make a comprised sound “da” – Synchrony of visual and auditory percep=ons – Perceiving neither what was said nor what was seen |
|
|
Theoretical Models ofSpeech Perception
|
• Motor Theory
• Cohort Theory |
|
|
Motor Theory
Lieberman et al. (1967) |
– Use processes involved in producing speech to
perceive speech – Implicit ar=culatory knowledge aids percep=on – Addresses “invariance problem” |
|
|
Motor Theory
Liberman & Whalen (2000) |
auditory signal not only matched to an acous=c
template of phonemes – phonemes also recognized by inferring ar=culatory movements necessary to produce sounds – What motor movements would be necessary to create these sounds? – Some neuroimaging evidence that motor system is ac=ve during speech percep=on |
|
|
Cohort Theory:
Marslen-Wilson and Welsh (1978) |
• As we hear speech, set up a cohort of possible
words • Words eliminated from set un=l target remains • Allows for interac=on of sensory informa=on and contextual informa=on |
|
|
Perception of Discourse
|
• People can oken understand sentences purely on the basis
of seman=cs, even if syntax is incorrect – Children’s u[erances are oken intelligible even if syntax is incorrect • Correct syntax does not guarantee meaningfulness – Famous example: “Colorless green ideas sleep furiously.” • However, humans cannot rely on seman=cs alone, consider: – Mike wondered if he failed the exam. – He wondered if Mike failed the exam. • Most of the evidence is that people use both syntax and seman=cs to comprehend sentences |
|
|
Who do we resolve ambiguities in discourse?
|
Minimal Attachment
|
|
|
Minimal Attachment
|
• Principle of minimal a[achment: Bias to
a[ach to old structures – (a) The spy saw the cop with binoculars but the cop didn’t see him. – (b) The spy saw the cop with a revolver but the cop didn’t see him. • According to minimal a[achment, (a) should be easier to parse than (b) • Experimental results support this no=on |
|
|
Why do we investigate speech errors?
|
“The inner workings of a highly complex
system are oken revealed by the way in which the system breaks down.” Dell (1986) • ”…and they have no disregard for human life." -George W. Bush, Washington, D.C., July 15, 2008 |
|
|
Types of Speech Errors
|
Shift
Exchange Anticipation Perseveration Addition Deletion Substitution Blend |
|
|
Shift
|
- one speech segment disappears from its appropriate location and appears
somewhere else That’s so she’ll be ready in case she decide to hits it (decides to hit it). |
|
|
Exchange
|
- double shifts, two linguistic units exchange places
Fancy ge~ng your model renosed (nose remodeled). |
|
|
Anticipation
|
-later segment takes place of an earlier segment
– Bake my bike (take my bike). |
|
|
Perseveration
|
- earlier segment replaces later item
– He pulled a pantrum (tantrum) |
|
|
Addition
|
- addition of linguistic material
– I didn’t explain this charefully enough. |
|
|
Deletion
|
- deletion of linguistic material
– I’ll just get up and mu[er intelligibly. (un-) |
|
|
Substitution
|
- one segment replaced by intrusion of another
– At low speeds it’s too light (heavy). |
|
|
Blend
|
-more than one word being considered, two items “blend” into single
item – That child is looking to be spaddled. (spanked/ paddled?) |
|
|
Tip of the Tongue (TOT) (Brown & McNeill 1966)
|
• Know what one wishes to say, but unable to
retrieve corresponding spoken form of output • Inducing TOT states (Brown & McNeill 1966) – “a naviga=onal instrument used in measuring angular distances, especially the al=tude of sun, moon, and stars at sea” • Oken aware of gender, first le[er, number of syllables, length of word |
|
|
Universal Stages of language
|
Cooing
Babbling One-word utterances Two-word utterances Basic adult structure |
|
|
Methods of studying Categorical Perception Among Infants
|
High amplitude sucking procedure
Preferential Looking Procedure |
|
|
Critical Periods of language Acquisition (Pinker, 1994)
|
Genie; Chelsea; Isabelle
|
|
|
Cri=cal Periods
|
Acquisi=on of normal language is guaranteed
for healthy children in normal environments up to the age of 6 and then steadily declines (Pinker, 1994) |
|
|
Genie
(Fromkin et al. 1974; Cur=ss, 1977) |
Normal birth, but severe linguis=c depriva=on
• From age 20 months un=l 13 years, isolated in small room, strapped to chair • When discovered, virtually no language • Learned some language, but syntax was grossly impaired • Used few func=on words • Failed to acquire inflec=onal morphology (i.e., rules forming past tense) – I like hear music ice cream truck. – Think about Mama love Genie. |
|
|
Chelsea
(Cur=ss, 1989) |
Born deaf in northern California
• Diagnosed as mentally retarded w/o recogni=on of deafness • Grew up without language • Age 31 discovered by neurologist, hearing restored • Learned 2000 words, reads, writes – The small a the hat. – Richard eat peppers hot. – Orange Tim car in. – Banana the eat. – I Wanda be drive come. – Breakfast ea=ng girl. |
|
|
Isabelle
(Davis, 1947; Mason, 1942) |
Reared from infancy with minimum a[en=on
from deaf-mute mother un=l age 6 ½ • Passed through normal phases of language development at accelerated rate • Aker 18 months, was linguis=cally competent |
|
|
Process of Language Acquisition
|
Hypothesis Tes=ng
• Imita=on • Learning Theory • Motherese • Language Acquisi=on Device |
|
|
Adversarial problem solving
|
In tradi?onal view, this is problem solving when every other
move is made by an opponent (or adversary) • Changes the nature of the search – Must consider what opponent will do, not just what states are good • Ability to make accurate predic?ons of opponent’s moves becomes paramount • Currently, huge differences in how the best machines do this and how the best people do this – Best humans tend to evaluate few moves, but do a good job of evalua?ng only ones which are good to start with – Computers do this by evalua?ng many (e.g. billions) of moves |
|
|
Perspectives on relation between decision making and problem solving (Allen R.
Solem, 1992) |
Decision-maker is a problem solver with available alternatives from which to
choose. Decision-making processes and those in problem solving differ. |
|
|
Problem Solving as Search: Newell and Simon (1972) Human Problem Solving
|
The problem space consists of a set of states
One of those states is the initial state One of those states is the current state One (or more) of those states is (are) the goal state(s) A state is transformed to a new state by the application of an operator |
|
|
Assumptions (Problem Solving)
|
Problem solver has to be able to represent the states
Problem solver has to have knowledge about operators These can be non-trivial |
|
|
Algorithms
|
Depth-first search
Breadth-first search Brute force |
|
|
Heuristics
|
-Problem solving “rules of thumb”
– May drama?cally cut down the amount of the problem space explored – Not always guaranteed to find a solu?on |
|
|
Weak vs. Strong Methods
|
When the solver will have a lot of knowledge about
how to proceed – Methods which rely on a lot of space-specific knowledge are called “strong” methods – (More on this in the context of exper?se) • When the solver has very limited specific knowledge and relies on more general or “weak” strategies – Means-end analysis – Hill-climbing – Generate-and-test – Depth-first search |
|
|
Means-end analysis: The General Problem-Solver (GPS: Newell and Simon,
1956) |
• Find the largest difference between the goal state and the ini?al state
• The General Problem-Solver (GPS: Newell and Simon, 1956) – One of the first-ever AI programs – Based en?rely on means-end analysis |
|
|
Hill-climbing: Water Jug problem
|
• A simple but powerful and pervasive heuris?c
• Always take the move that results in the maximum reduc?on of difference between the current state and the goal state • How is that different than MEA? • This is surprisingly effec?ve in many domains • But it can and does fail |
|
|
Generate-and-test
|
Some method for coming up with candidate solu?ons
• Some method for evalua?ng the validity of the solu?on generated – Does it meet all requirements of the goal state? • Easy to observe people doing this for logic puzzles that require some state – John is sikng to the ler of Mary and to the right of another woman. Frank, being ler-handed, must sit at the end. Jane is sikng… and so on and so forth • Ranges from dumb trial-and-error to very sophis?cated generators and testers |
|
|
Depth-first search
|
NEED TO COMPLETE
|
|
|
Isomorphs: tower of hanoi (TOH)
|
Two problems that share the same problem space but have
different surface features are called “isomorphs” • A fair amount of work has been done on isomorphs – Especially tower of hanoi (TOH) isomorphs |
|
|
Solving by Analogy
|
A common method of problem-solving
– Remember (or look up) a similar but already-solved problem – Modify that solu?on slightly to fit the current problem • Steps to using analogy – Retrieval of source – Map new situa?on to source – Extension—use source & mapping to make predic?ons |
|
|
Solving by Analogy
(Studies) |
-Tumor Ray
-Dictator Invade • For this study – When given problem in isola?on, only 10% solve – 30% solve if they’ve read the other story – But, if given the “hint” to make use of a previous solu?on, 90% solve – So, what does this mean? • Problem is generally one of access, not applica?on – A good analogy needs structural similarity for mapping – We usually focus on surface similari?es • Is solving by analogy a “weak” method? |
|
|
Solving by Analogy
(Examples) |
Surface vs. deep structure
– Surface structure is the set of non-relevant features • For example, algebra problem is about golf balls – Deep structure refers to the solu?on methods • What equa?ons actually apply in what ways • People, especially novices in a domain, tend to retrieve poten?al analogs based on surface structure – This is oren unhelpful • Want retrieval based on deep structure |
|
|
Surface vs. deep structure
|
– Surface structure is the set of non-relevant features
• For example, algebra problem is about golf balls – Deep structure refers to the solu?on methods • What equa?ons actually apply in what ways |
|
|
What do novices tend to do?
|
NEED TO COMPLETE
|
|
|
Insight problems
|
When the solu?on to the problem occurs suddenly in conscious
– Single operator – You see it or you don’t – Some problems don’t seem to fit this conceptualiza?on very well • Further supported by popular myths – Coleridge wrote Xanadu in a single (drug-induced?) sikng • Not true, there are known drars – Kekule’s benzene ring came in one brief flash of insight where he visualized a snake ea?ng its tail • This story was first told some 35 years later • S?ll, problems like the Duncker’s ray problem “feel” different – Inability to formulate the problem space or operators |
|
|
Incuba?on
|
• For “insight” problems
– Seem to have no clear state or operators – Does the search conceptualiza?on fail here? • An alterna?ve view – When no progress is made in a problem space, form a new space to work out the issue (e.g., no operators) • Rigidity issues seem par?cularly strong in these sub-spaces • “Incuba?on” is based on the no?on that going away from the problem for a while helps • Feeling of insight may be like confidence in the accuracy of recall |
|
|
Einstellung
|
– Predisposi?on to solve a given problem in a specific manner even
though there are "beYer" or more appropriate methods of solving the problem. • Thus crea?ng nega?ve effect of previous experience when solving new problems |
|
|
Func?onal fixedness
|
– Cogni?ve bias that limits a person to using an object only in the way it
is tradi?onally used. |
|
|
Self-Explana?on
|
• Many textbooks contain worked-out examples
• Protocol studies show that high performers do something different when they encounter such examples • What they do – Work through the steps outlined and try to generate explana?ons for why each step was taken – Especially steps they didn’t understand on first reading • Further studies show that training in this technique improves problem solving performance |
|
|
Novice vs. Expert
|
• Novices
– Weak methods – Lots of search • Experts – Strong methods over weak methods – Knowledge about the structure of the problem space – Use of analogy is usually correct – Very liYle search – The search that does occur is very directed – Oren have difficulty ar?cula?ng what they know |
|
|
Decision Making:
If people have to choose between options, what rules do they use? |
• What people actually do (descrip?ve)
|
|
|
Prescriptive
|
Rationality – internally consistent decision rules
Normative - Because of possibility of gain Expected value and expected utility |
|
|
What are some of the methods for eliciting decisions?
|
Method 1: Which would you pick?
• Method 2: What’s the lowest price at which you would sell the chance to play this game? |
|
|
Things that impact Decision Making
|
• Framing
– Way that op?ons are presented influences the selec?on of an op?on (Tversky & Kahneman, 1981) – Risk averse for gain – Risk seeking for losses • Psychic budget • Sunk cost |
|
|
Methods for Decision Making
|
Algorithms
Heuristics Representativeness Availability Heuristic Anchoring and adjustment Satisficing Elimination by Aspect |
|
|
Psychic Budgets
|
concern how we mentally
categorize money we have spent or are contempla?ng spending. • The categories don’t always correspond to reality, and so can lead to odd decisions. |
|
|
Sunk costs
|
is ?me, money, or other
investment that is irretrievably spent, and therefore should not affect current decision making, and yet does. |
|
|
• Algorithms
|
– Always get correct solu?on
– Time and resource intensive |
|
|
• Heuris?cs
|
– Reduces cogni?ve load
– Increases error – Types • Representa?veness • Availability • Anchoring and adjustment • Sa?sficing • Elimina?on by Aspect |
|
|
• Availability
|
is used to judge the probability of
events • You use availability by trying to call to mind examples: the more examples you think of, the more probable you judge the event to be. |
|
|
Availability Heuris?c
|
• Human judgments of probability are ouen governed by
the heuris?c of “availability” – Examples of rare events are ouen vivid and compelling, and therefore easier to recall – This slants judgment of probability • Implica?ons of portrayal of groups on TV/movies • Racial and ethnic - African Americans over-represented as criminals on TV • Gender – Women represented as “non-career” – Middle age men represented as clueless -suggests you’ll believe what is represented |
|
|
Anchoring
|
Describes a common human strategy called
“anchoring” or “anchor-and-adjust” |
|
|
Sa?sficing
(Simon) |
• Rather than op?mizing, do this:
– Set some threshold of “good enough” – Examine an op?on • If it’s above the threshold, stop • If not, move to the next op?on and repeat • Example: – Have to select a project partner in a class of 50 – Hard to quan?fy all the acributes that are important or even to rank all the alterna?ves – Snag the first acceptable person in the class • Par?cularly good when: – Differences betwe |
|
|
Elimina?on by Aspects
|
• Each op?on has a set of acributes, each with some value
– Order the acributes in terms of importance • Start with the most important acribute – Eliminate all op?ons that are not acceptable on this acribute • Then go to the next most important acribute – Repeat un?l only one op?on is leu • Variant of this called “lexicographic semiorder” – Take the most important acribute – Select the op?on that has the best value on that acribute • Common for “big” decisions |
|
|
Informa?on we ignore
|
• Sample Size
• Base Rates |
|
|
Bayes’s theorem
|
describes the normative procedure for how to combine
probabilistic evidence |
|
|
Normative Reasoning:
|
Human Deduction
Effects of content Cheater detection Causal knowledge |
|
|
Human Deduction
|
• Reasoning in an argument is valid if the
argument's conclusion must be true when the premises (the reasons given to support that conclusion) are true. – Premise 1: All humans are mortal. – Premise 2: Socrates is a human. – Conclusion: Socrates is mortal. • Validity doesn’t necessarily macer – Premise 1: If green is a color, then grass poisons cows. – Premise 2: Green is a color. – Conclusion: Grass poisons cows. |
|
|
Human Induc?on
|
-is not determinis?c
– It is probabilis?c • People generally have difficulty reasoning about probabili?es – For example, on the forthcoming homework, which you haven’t seen, rate the probability that it will take you • Less than 1 hour • From 1 to 3 hours • From 3 to 5 hours • 5 or more hours • Now add those up. Do they add up to exactly 1 (or 100%)? |
|
|
Wason-Laird Task
|
Every card has a lecer on one side and a number on the
other • The premise: If there is a vowel on one side of a card, then there is an even number on the other side • Ques?on: How many cards need to be flipped to verify that this is true? – And which ones? A B 2 3 |
|
|
Case-based reasoning
|
• Perhaps we are merely good at scenarios with
which we are familiar. Such theories are called cased-based reasoning theories. We simply remember what worked before. |
|
|
Story Model (Pennington & Has?e)
|
• Developed to explain individual juror decisions
– jurors typically receive large amounts of evidence • ouen contradictory • in essen?ally random order • Three stage process – Story Construc?on (org. info. into coherent mental representa?on); stories with causal links – Verdict representa?on--four axes • Iden2ty (i.e. was the defendant the one?) • mental state of the defendant at the ?me • circumstances during the event • ac2ons taken by the defendant – Story Classifica?on • story constructed in step 1 is matched to the verdicts represented in step 2 • central element is goodness of fit between the story and the various verdicts • verdict with the best fit to the story is the one chosen by the juror |
|
|
Reasoning About Evidence
|
• People generally exhibit a “confirma?on bias”
• People tend to seek evidence that is consistent with their hypothesis – This is backwards, disconfirming evidence is more informa?ve • People tend to view ambiguous evidence as consistent – Tilts balance when weighing consistent and inconsistent evidence |
