Memory Errors and False Memories PDF

Week 6 Memory (cont.) Memory Errors ​ Memory is NOT like a video camera ○​ We often remember only the ‘gist’ of what we experience rather than the veridical details ○​ Our memory is influenced by other things: beliefs, expectations, assump...

Week 6 Memory (cont.) Memory Errors ​ Memory is NOT like a video camera ○​ We often remember only the ‘gist’ of what we experience rather than the veridical details ○​ Our memory is influenced by other things: beliefs, expectations, assumptions, etc. Memory is a Reconstructive Process ​ Reassemble pieces from fragments ​ We take whatever comes back to mind and fill in the gaps ​ Retrieving a past event engages the same brain mechanisms as imagining a future event ​ Patients with hippocampal damage show an impaired ability to envision the future ○​ Asked hippocampal amnesic patients to imagine future events ○​ Amnesic patients had significantly lower details in their imagination ○​ Evidence that the hippocampus is important for envisioning the future because we are drawing from our present experiences to come up with some scenario in the future ​ Memory is useful to predict the future, make good choices, learn from mistakes, and facilitate positive outcomes Misinformation Effects ​ Misleading information can affect memory for the actual event ​ Important for courtroom/eye-witnesses testimony (from memory) ​ Dr. Elizabeth Loftus Memory Can Be Distorted By Post-Event Information ​ Classic study by Loftus and Palmer (1974): ○​ Subjects were shown a video depicting a car accident ○​ Then given the following question: “How fast were the vehicles going when they ______” ○​ Different subjects were asked questions that differed in the “magnitude” of the final word ○​ The possible words were: ​ Contacted, Hit, Bumped, Collided, and Smashed ​ Speed estimate changed depending on what word the subjects got ​ Effect of changing the intensity of the verbs  ​ Estimates of car speed could be distorted by leading questions about the accident ○​ More dramatic the verb = more exaggerated memory of speed (estimated faster speed) Misinformation Effects ​ Did the question affect memory for other aspects of the accident? ​ They asked subjects more questions about the accident Loftus et al. (1978) ​ Showed participants a slideshow/video where the car hits pedestrian ​ Then subjects got a long list of questions ○​ Were there trees in the scene? ○​ Did the red car turn right at the intersection? ○​ Was accident victim wearing a plaid shirt? ○​ Was there a light pole in the images? ○​ Did another car pass the red car while it was at the stop sign? ○​ Was the police car white? ​ Immediately after viewing these slides, subjects filled out a questionnaire of 20 questions ​ For half of the subjects, Question #17 was: "Did another car pass the red Datsun while it was stopped at the stop sign?"  ​ For the other half, the same question was asked with the words "stop sign" replaced with "yield sign" ​ Some time later, shown 15 pairs of slides and asked to judge which image in each pair was the one they originally had seen ​ The critical trial was the one where these two images appeared: ​ Subjects given misleading information after encoding had more false memories for the details of the visual scene ​ Control group was right 75% of the time ​ Misinformation group was right 41% of the time ○​ 59% of the time they got it wrong because hearing the question about the “yield sign” influenced their memory ​ Shows that question you get can override your memory of what you actually saw Planted Memory Study: Lost in a Shopping Mall ​ Loftus & Pickrell (1995) ○​ Interviewed subjects about different aspects of their lives ○​ The experimenters talked to relatives in the subject’s lives ○​ Participants read one-paragraph stories about three events that actually happened to them and one that did not (lost in a shopping mall) ○​ Then interviewed about each of these memories ○​ 6 of 24 subjects developed false memories for the mall scenario by the third interview ○​ 25% of the subjects falsely recalled getting lost in the mall and reconstruct false details about it Planted Memory Study: Accident at a Family Wedding ​ Convinced subjects they had an accident of spilling punch at a wedding ​ 3% of participants provided false recall in 1st interview ​ 27% of participants provided false recall in 2nd interview ​ False recalls sometimes takes time to emerge Planted Memory Study: Vicious Animal Attack  ​ Convinced subjects they were attacked by an animal when younger ​ 26% of participants “recovered” a complete memory for the false experience, and an additional ​ 30% of participants recalled some aspects of the experience Planted Memory Study: Doctored Photos of Hot Air Balloon Ride ​ Photoshopped subjects’ photos into a hot air balloon ​ After 3 interviews, 50% of participants created completely false or partial false memories of going on a hot air balloon ride as a kid Planted Memory Study: Met Bugs Bunny at Disneyland ​ Sometimes related memories or events are causing the false recall ​ Bugs Bunny is not a Disney character ​ 16% of subjects were led to believe that they met Bugs Bunny at Disneyland Recipe for Creating a False Memory ​ Source monitoring framework 1.​ Subjects must accept that the suggested event is plausible 2.​ Subjects must create contextual information for the event, such as an image and a narrative 3.​ Subjects must commit a source monitoring error, wrongly attributing their memory construction to personal experience a.​ Seen it somewhere else, dreamt about it, heard about, etc. False Memories Can Even Lead to Confessions to a Crime One Didn’t Commit ​ Contributing Cause of Wrongful Convictions (first 325 DNA exonerations) ○​ Total is more than 100% because wrongful convictions can have more than one cause ​ One potential solution to eyewitness memory errors: ○​ Better instructions to the jury about the fallibility of human memory ​ “Eyewitness identification evidence must be scrutinized carefully. Human beings have the ability to recognize other people from past experiences and to identify them at a later time, but research has shown that there are risks of making  mistaken identifications. That research has focused on the nature of memory and the factors that affect the reliability of eyewitness identifications. ​ Human memory is not foolproof. Research has revealed that human memory is not like a video recording that a witness need only replay to remember what happened. Memory is far more complex.” Concepts and Knowledge Carving Up the World Into Categories ​ Categorization- process through which ideas and objects are recognized, differentiated, classified, and understood ​ Concepts- our mental representations of categories How are categories organized? ​ Category- set of items that are grouped together on the basis of something ​ Gain mental efficiency with categorization The Basic Level of Categorization ​ Eleanor Rosch argued that there is a basic level of categorization ○​ Neither too general nor too specific ○​ Tends to be used in speaking and reasoning about categories (intuitive category) ​ Here, “chair” is the basic-level category, as opposed to “furniture” (more general, or superordinate) or “wooden desk chair” (more specific, or subordinate)  ​ Easier to explain what features are common to members of basic-level categories than for other levels How are categories stored in memory? ​ Classical theory- explicit rules for category membership ​ Prototype theory- prototypes (mental average of all category members) ​ Exemplar theory- individual instances ○​ Pull memories of other things to mind Classical Theory ​ A category is defined in terms of necessary and sufficient features ○​ Necessary: has to be there ○​ Sufficient: all that you need ​ These features define the category ​ This representation is abstract ○​ It does not store any information about specific exemplars ​ Ex: square is defined by the following features: ○​ Closed figure ○​ Four sides ○​ Sides equal in length ○​ Equal angles Criticisms of the Classical Theory ​ Defining features often can’t be found ○​ You can often remove any particular feature and some object will still be a category member ​ Non-necessary features affect categorization ○​ e.g., Which of these shapes is a parallelogram? ​ Technically they’re all parallelograms ​ But most people think the 2nd shape is a better parallelogram than others ​ Ex: what is bird? ○​ Penguin is an acception ​ Has flippers, don’t eat worms, don’t build nests ​ Violation of any one defining feature does not seem to affect our categorization What is a game? ​ Ludwig Wittgenstein’s (1953) famous critique: ○​ What is the necessary feature of the concept of a game? ​ Competition between people or groups? ​ Solitaire: you're just trying to beat yourself ​ Has a winner? ​ Jumping rope: no winner ​ Provides amusement or diversion  ​ Are professional athletes amused or diverted? ○​ For many categories there are no clear defining features ​ Various members share various features, but there is no single feature that is necessary ​ Classical theory doesn’t work Family Resemblance ​ Ludwig Wittgenstein proposed that members of a category have a family resemblance to each other ​ Dark hair, glasses, a mustache, and a big nose are typical for this family but do not define the family ​ General set of features that tend to be present or overlap but don't have to be present Typicality Effects ​ ​Some members of a category are more “typical” than others ○​ They are verified more quickly (sentence verification task) ○​ e.g., “A robin is a bird” is faster than “A penguin is a bird” ○​ But, these differences are related to non-necessary features (e.g., ability to fly), which are not included in the classical theory ​ Mean scores on 7-point ratings ○​ Also applies to reaction times in sentence verification tasks ○​ Bat is not even a bird: suggests that people think it has traits that are bird-like ​ Typicality ratings may be based on the total number of typical features ​ This also explains other frequent errors: e.g., whales & dolphins as “fish” ​ Production task: ○​ How many examples of fruit can you name in the next 30 seconds? ○​ Which ones are high on the list? ​ Strong exemplars: apple, orange ○​ Lower on the list? ​ Weaker exemplars: kiwi, tomato, olive, avocado ○​ What is the “best one”? ​ Graded membership: some fruits are “fruitier” than others”  ​ Matter of degree ​ Ex: what is an even number? ○​ Very clear rule for what is even/odd ​ Classical theory should handle it well ○​ Yet people rate certain numbers to be more odd or even ​ 3 is more odd/highly typical than 7 ​ 4 is more even/highly typical than 18 ​ People may think of which number they see more ​ Typicality is a very core attribute of categorization Typicality and Generalization ​ Rips, 1975 ​ Penguins can catch disease X All birds can catch disease X ​ Robins can catch disease X All birds can catch disease X ​ When a typical category member has a particular trait, you’re much more likely to think that that attribute would be present for all member ○​ Affects generalization Probabilistic Theories of Categorization ​ Failure of classical theory led to proposal that category representation may be probabilistic rather than deterministic (rule-based) ○​ 2 approaches: ​ Prototype theory- category judgments are made by comparing a new exemplar to the prototype ​ Don’t have prototype in the brain but people can still conjure up a prototype by averaging up generic things ​ Exemplar theory- category judgments are made by comparing a new exemplar to all the old exemplars of a category or to the exemplar that most readily comes to mind ​ Actually have prototype representations that are conceptual knowledge, stored in mind, and extracted ​ Both are based on the idea of similarity Prototype Theory ​ Imagine two birds perched on a tree branch ​ When we think about a category, we are really thinking about the prototype for the category ​ Categories are represented by the average of all members of the category ○​ e.g., the concept “bird” is represented by a prototype that is very similar to a robin or sparrow and quite different from an ostrich or penguin ○​ The prototype does not need to exist in the real world  ​ The category representation is abstract ○​ Does not store information about specific exemplars Problems with Prototype Theory ​ Assumes that information about individual instances is not used to guide categorization ​ However, people do seem to store information about individual exemplars, and can sometimes be influenced by these specific exemplars ​ Also, prototype theory doesn’t have a way of taking into account the variance of a given category ​ Some categories are highly variable ○​ Ex: pizza, tomatoes, chairs ○​ Many different types Exemplar Theory ​ Concepts are represented by all of the exemplars that have been experienced ○​ The category “bird” is represented by memories of all previous experiences of birds ○​ When we categorize something, we compare it one or more exemplars retrieved from memory, and decide the category based on the most similar exemplars ​ The category representation is concrete ○​ Labels ○​ There is not necessarily a summary of the category Problems with Exemplar Theory ​ Assumes that many individual exemplars are stored in memory without “blending” ​ Has trouble accounting for people’s ability to extract general properties of categories so as to allow classification of new instances Problems with Both Prototype and Exemplar Views ​ Both rely heavily on the idea of similarity ​ What’s wrong with similarity? ​ Take a blackbird, remove all of its feathers and clothe it in a suit of bat skin ○​ Is it more similar to a bat or a blackbird? ​ Similar to each but in different ways ○​ Would you categorize it as a bat or a blackbird? ​ Similarity is always relative ○​ There are an infinite number of ways in which two things can be similar ○​ We require some way of knowing what features are being compared Theory-Based Categorization ​ We know much more about categories than a list of their features or their values in dimensional space  ​ Categories provide explanations for how things work in the world ○​ The same way that theories provide explanations for scientific phenomena ○​ They center on causal relations between entities in the world ○​ Theories guide perception by leading us to believe that particular features are interesting and others are not ​ What is a drunk? ○​ It’s late at night, you are walking home when you see someone jump into a swimming pool fully clothed ○​ You think “that guy must be drunk!” ○​ Why do you think that? ○​ Classical categorization ​ Stinky breath ​ Wobbling walk ​ Impaired speech ​ … ​ Maybe one entry is “jumps in pools at night”? ​ Theory-based categorization- concept of drunk involves a theory of impaired judgment, which explains the man’s behavior, so you induce he must be drunk ○​ Knowledge-based causal theory Summary: Approaches to Categorization ​ Classical Theory: concepts have definitions (necessary & sufficient conditions) ​ Prototype theory: there is a summary representation for each category ​ Exemplar theory: no summary representation- a concept is a collection of individual instances ​ Theory-based categorization: categories include knowledge-based causal explanations Ch. 9: Concepts and Generic Knowledge Understanding Concepts ​ Family resemblance- members of a category resemble one another ○​ No defining features that all members share ​ Can identify characteristic features (that most have) for each category that enable you to recognize it Prototype and Typicality Effects ​ Prototype- ideal depiction, best example, average ​ Resemblance is a master of degree ​ Typicality- how much they resemble the prototype ​ Graded membership- objects closer to the prototype (high typicality) are better members of the category  ​ Sentence verification task- participants were presented with a series of sentences and they had to indicate whether the sentence was true or false ○​ Participants chose their responses by comparing the thing (ex: penguin) mentioned to their prototype (ex: bird) for that category ○​ Close similarity between the test case and prototype → quicker response ○​ Distance between the test case and prototype → slower response ​ Production task- ask people to name as many examples as possible ○​ According to prototype view they’ll do this task first by locating their example bird or dog prototype in memory and what resembles the prototype ○​ Start at the center of category (prototype) → quicker response ○​ Then work their way outward (farther from prototype) → slower response ​ Rating task- participants must evaluate item or category ○​ How typical something is within a category and give it a rating ○​ Rates items as less when they are farther from prototype ○​ Suggest that they compare item to prototype ​ Basic-level categorization- not too specific and not too general ○​ Ex: chair, apple ○​ If asked to identify object, more likely to use basic-level categories ○​ Natural way to categorize Exemplars ​ Exemplar-based reasoning- draws on knowledge about specific category members (exemplars) rather than drawing on more general information about overall category ​ Exemplar and prototype theories: an object before your eyes triggers some representation in memory ○​ Asses resemblance between the conceptual knowledge supplied by memory and object Difficulties with Categorizing via Resemblance ​ In using a prototype or exemplar, you need to make a judgment of resemblance ○​ Decide how much some candidate object resemble the prototype or exemplar ​ These depend on your being able to focus on features that are essential for each category ​ Decisions about what’s essential/important vary from category to category and vary on your beliefs about that category Broader Role of Conceptual Knowledge ​ Categories enable you to extend your knowledge in important ways ​ People were willing to make inferences from the typical case to the whole category, but not from an atypical case to the category  ​ Prototype vs. stereotype ○​ Prototype: summary of your experience ○​ Stereotype: acquires through social channels Diversity of Concepts ​ Different types of concepts are represented in different brain area ​ Anomia- disorder in which people lose the ability to name certain objects ​ Hub and spoke model- how elements that make up conceptual knowledge are linked and coordinated ○​ “Hub” is supported by tissues in the anterior temporal lobes, connects and integrates information from many brain areas ○​ “Spoke” is represented in more specialized brain regions so visual information relevant to the concept is stored in visual areas, relevant action information is stored in motor areas, etc. Knowledge Network ​ Propositions- smallest units of knowledge that can be either true or false ​ Connectionist networks- distributed representations in which each idea is represented not by a certain set of nodes but instead by a pattern of activation across the network ​ Parallel distributed processing (PDP)- system of handling information in which many steps happen at once and various aspects of the problem are represented only in a distributed way ○​ Knowledge is contained in memory visa distributed representations that requires disturbed processes Week 7 Language Properties of Language 1.​ Symbolic: makes up of arbitrary relation between sounds and meaning 2.​ Discrete infinity: a finite set of elements can generate a (potentially) infinite set of ‘meanings’ 3.​ Structure dependence: meaning is conferred through a specific arrangement of symbols 4.​ Displacement: language allows referring to ideas/elements that are not “there” 5.​ Organized at multiple levels: sounds, words, sentences, paragraphs, and text Symbolic ​ Language is arbitrary ​ Referent- actual object, action, or event in the world that a word refers to ○​ Ex: sign language, braille  ​ We even use different words to refer to the same thing within the English language (dialects) ○​ Ex: soda, pop, coke, soft drink Discrete Infinity (Generativity) ​ From a discrete set of units (words), we can generate seemingly infinite meaning ​ Ex: ○​ John hates cheese ○​ My roommate heard a rumor that John hates cheese ○​ I told her that my roommate heard a rumor that John hates cheese ○​ … ​ The only limitation is the working memory capacity and attention span of reader/listener ○​ Gets too long to keep in mind Structure Dependence ​ Language is governed by rules that impart meaning and define which combinations of elements are acceptable and which are not ​ Ex: ○​ John kissed Mary ○​ Mary kissed John ○​ Kissed John Mary*​ (* means not grammatically correct) ○​ John Mary kissed* ​ Knowing how the word is used in the language and what property the word has grammatically Displacement ​ Language allows us to think of, and communicate about, ‘things’ beyond what is immediately ‘sensed’ ​ Ex: ○​ “On Sunday night, your next quiz will be due.” ○​ “Dark matter is a type of matter which neither emits nor scatters light or other electromagnetic radiation and is estimated to constitute 83% of matter in the universe” ○​ “Imagine no possessions. I wonder if you can. No need for greed or hunger. A brotherhood of man. Imagine all the people. Sharing all the world...” Levels of Language Representation Phonemes ​ Phonemes- smallest unit of speech that can be used to distinguish one utterance from another (in a given language) ​ In English, phonemes are made of a consonant or vowel sound (44 in total) ​ Different languages employ different sets of phonemes (and different types, such as ‘clicks’) ​ Children appear to be sensitive to any set of phonemes at birth ○​ Learn through experience ○​ Children can learn any language they’re exposed to ○​ Harder to learn phonemes as you get older ​ Phonemes are produced by modulating the flow of air from the lungs to the mouth and nose ​ Phonemes can be classified according to specific features ​ Voicing ○​ Whether vocal folds vibrate ([z], [d], [b], [v]) ○​ Or not ([s], [t], [p], [f]) ​ Manner of production ○​ Whether air is fully stopped ([b], [p], [d], [t]) ○​ Or merely restricted ([z], [s], [v], [f]) ​ Place of articulation ○​ Where in the mouth the air is restricted: ○​ Closing of lips ([b], [p]) ○​ Top teeth against bottom lip ([v], [f]) ○​ Tongue behind upper teeth ([d], [t], [z], [s]) ​ Normal speaking rate is 150 words per minute ○​ ~15 phonemes per second ​ A major challenge for speech recognition is phoneme segmentation ○​ Coarticulation: pronunciation of a phoneme is changed by the following phoneme (blending a phoneme by the next one being produced) ​ Between phonemes in a word: truth vs. tooth ​ Between phonemes in different words: “can’t handle” Phoneme: Segmentation Errors ​ Ex:  ○​ “This guy is falling! This guy is falling!” ○​ Sounds like… “The sky is falling! The sky is falling!” ​ To disambiguate it, you need context ​ The reason you hear one and not the other is how you segment phonemes Categorical Perception of Phonemes ​ Voice onset time (VOT)- time between the beginning of the pronunciation of the word and the onset of the vibration of the vocal cords ○​ “ba” your vocal cords vibrate right from the start ○​ “pa” your vocal chords do not vibrate until after a short delay ​ Continuous = actual sounds ​ Discrete = actual perception ​ Experiment: Liberman et al. (1957) ○​ Used computer to acoustically manipulate the VOT, in systematic increments ○​ Sounds jumped in a discrete way (not a gradient) ​ Our categorization of phonemes shows abrupt boundaries, even when there is no corresponding abrupt change in the stimuli themselves ​ This phenomenon is referred to as categorical perception Morphemes ​ Morpheme- smallest unit of meaning within a language ​ Morphemes can be divided into: root words and affixes (i.e., prefix, suffix) ​ Ex: ○​ “dog” single morpheme (dog) ○​ “dogs” 2 morphemes (dog & -s [number]) ○​ “studied” 2 morphemes (study & -ed [tense]) ○​ “restudied” 3 morphemes ○​ “related” only 2 morphemes Syntax ​ Syntax- systematic way in which [categories of] words can be combined and sequenced to generate meaningful phrases and sentences ○​ Rules apply to grammatical categories (nouns, verbs, prepositions, adjectives, adverbs) ​ One kind of syntactic rule is a phrase-structure rule- constraint that governs the pattern of branching in a phrase-structure tree ​ One such rule specifies that a sentence must contain a noun phrase (NP) and a verb phrase (VP) Types of Phrase Structure Rules ​ Descriptive rules- characterize the language as it is ordinarily used by fluent speakers ○​ Linguistics aims to provide a descriptive grammar of language ○​ Syntax aims to provide descriptive rules of language ​ Prescriptive rules - standards for how language “ought” to be used ​ For ex: ​ Don’t start a sentence with And or Because ​ Don’t end a sentence with a preposition ​ Don’t split an infinitive ○​ We should keep a healthy dose of skepticism about prescriptive rules ​ This type of English I just can’t put up with ​ Up with this type of English I just cannot put ○​ Guidance of how to use language well Phrase Structure Organization Aids the Reader ​ The way the lines are organized are consistent with your internal model of the phase structure ​ For someone to follow, breaking it up in a way that’s more intuitive is more effective than breaking it up like this Critical Distinction Between Syntax and Semantics ​ Syntax and semantic are independent ​ “Colorless green ideas sleep furiously” (Noam Chomsky) ○​ Incoherent in semantic level ○​ BUT we still process grammar ○​ Compare with “Furiously sleep ideas green colorless.”* ​ Not grammatically correct ○​ Mind is building phrase structure ​ “Twas brillig and the shlithy toves did gyre and gimble in the wabe” (L. Carroll ‘Jabberwocky’) ○​ Illustrates that sentences can be syntactically correct, even when meaningless ​ EEG data (event-related potential)  ○​ Ex: “He drinks his coffee with cream and dog” ​ What’s happening in your brain when you experience a semantic violation (doesn’t make sense with your knowledge of the sentence) ​ “Dog” does not belong in sentence ​ Detect earlier than syntax ○​ Ex: “He prefers to solve problems herself” ​ What’s happening in your brain when you experience a syntactic violation ​ “He” does not agree with “herself”​ Syntactic Ambiguity ​ Sentences can be interpreted in different ways ○​ Map onto phrase structure ​ The girl looked at the boy with the telescope ○​ Sentence does not give you enough to ambiguate it ​ Ex: poorly worded headlines ○​ Complaints about NBA referees growing ugly ​ Real meaning: complaints are growing ugly ​ Misinterpretation: NBA referees are becoming uglier ○​ Dr. Ruth to talk about sex with newspaper editor ​ Real meaning: talking to newspaper editor about sex ​ Misinterpretation: sex with newspaper editor ○​ Juvenile court tries shooting defendant ​ Real meaning: court is having trial about a shooting ​ Misinterpretation: juvenile court trying to shoot defendant  ○​ Squad helps dog bite victim ​ Real meaning: squad trying to help victim ​ Misinterpretation: squad trying to help the dog bite the victim ○​ NJ judge to rule on nude beach ​ Real meaning: judge sitting on nude beach issuing rulings ​ Misinterpretation: judge is ruling whether nude beach would be permanent or shutdown Garden Path Sentences ​ When we perceive a sentence, we must parse the sentence’s syntactic structure ​ Garden-path sentence initially suggests one interpretation, which turns out to be wrong ​ Ex: ○​ The secretary applauded for his efforts was soon promoted ​ “The secretary applauded”: think secretary is actually clapping ​ But if you read the rest of the sentence, “the secretary was applauded for his efforts” ​ Adding words can help interpret sentences differently (ambiguate) ○​ Fat people eat accumulates ​ The fat that people eat accumulates in the body ○​ The man whistling tunes pianos ​ The man who is whistling tunes pianos ○​ Because he ran the second mile went quickly ​ He ran the second mile, that’s why it went quickly Background Knowledge Aids Sentence Parsing ​ Use gestures to help people parse sentences ○​ Help ambiguate ​ Parsing- dividing sentence into proper phrase structure to understand it ​ Ex: ○​ The detectives examined by the reporter revealed the truth about the robbery ​ Garden-path sentence ​ Reassign “examine” verb, detectives are not the ones examining but being the ones examined by the reporters ​ “Detectives” can be subject or object of the verb ○​ The evidence examined by the reporter revealed the truth about the robbery ​ NOT garden-path sentence ○​ We know that evidence is NOT capable of examining anything Extralinguistic Context Aids Sentence Parsing ​ Extralinguistic context- factors outside of language itself  ​ The sentence “Put the apple on the towel into the box” is a garden-path sentence, unless the sentence is uttered with the appropriate visual context ○​ But without the picture, it doesn’t make sense Prosody Aids Sentence Parsing ​ Prosody- patterns of pauses and pitch changes that characterize speech production ​ It is used to: ○​ Emphasize elements of a sentence ○​ Highlight the sentence’s intended structure ○​ Signal the difference between a question and an assertion ​ Ex: ○​ “No dogs are here” implies that there are no dogs ○​ But by putting bigger space and changing the way “dogs” are spoken, changes the meaning of the sentence ​ “No. Dogs are here” ○​ “I’m sorry?” with higher frequency to indicate a question ○​ “I’m sorry” with a lower pitch indicates an apology Critical Period for Language Development ​ There seems to be a critical period during which language develops readily and after which language acquisition is more difficult and less successful ​ The extraordinary cases of Victor, the Wild Child of Aveyron (France, 1800), and Genie (United States, 1970) seem to support the critical period hypothesis ​ Other evidence for critical period comes from: ○​ Studying the effects of damage to language areas in the brain (children recover more readily than adults) ○​ Studying the ages at which a second language is acquired ○​ The longer you wait, you won’t be good at it compared to someone who’s spoken the language since childhood Language in the Brain ​ Arcuate fasciculus- white matter tract (bundle of axons) connecting Broca’s and Wernicke’s Areas Broca’s Aphasia ​ Deficit in speech production but not comprehension ​ Lesion in inferior frontal lobe (stroke) ​ Characteristics: ○​ Halting speech ○​ Tendency to repeat phrases or words ○​ Disordered syntax/grammar ○​ Disordered structure of individual words ○​ Comprehension intact Wernicke’s Aphasia ​ Deficit in comprehension but not speech production ​ Lesion in superior temporal lobe ​ Characteristics: ○​ Fluent speech ○​ Little spontaneous repetition ○​ Adequate syntax/grammar ○​ Contrived or inappropriate words ○​ Comprehension not intact Modern Comparison ​ Look at maximum overlap ​ Broca’s patients ○​ Lesions overlap of 36  ​ Wernicke’s patients ○​ Lesion overlap of 11 Other Types of Aphasia ​ Conduction aphasia: ○​ Preserved comprehension ○​ Spontaneous speech has proper syntax and semantics ○​ Impaired repetition & paraphasic errors (phonemes and syllables will be dropped or misplaced) ​ Global aphasia: ○​ Huge left hemisphere stroke ○​ Nearly complete loss of comprehension and production of speech ○​ Some single stereotyped words might still be retained Classic Model Needs an Update ​ Double dissociation ​ Broca’s area and Wernicke’s area are just smalls parts in the language network ​ Language has components in frontal lobe, parietal lobe, temporal lobe, and little in occipital lobe ○​ And different for different people ○​ Amodal  Ch. 10: Language Organization of Language ​ Sentences- coherent sequence of words that express the speaker’s intended meaning ​ Morphemes- smallest language of units that carry meaning ​ Phonemes- smallest units of sound that serve to distinguish words in a language ​ Language can also be organized in another way ○​ Within each of these levels, people can combine and recombine the units to produce novel utterances ​ Assembling phonemes into new morphemes or assembling words into new phrases Phonology ​ Voicing- buzzing sort of vibration produced by rapid opening and closing of vocal folds ​ Vocal folds- two flaps of muscular tissue in larynx ​ Manner of production- speaker momentarily obstructs the flow of air out of the lungs to produce a speech sound ​ Speech segmentation- “slice” stream of speech into segments to identify the phoneme ​ Coarticulation- producing speech involves the overlap of phonemes, you don’t utter one phoneme at a time ​ Phonemic restoration effect- people hear phonemes that are actually not presented but highly likely in the context ​ Categorical perception- people are much better at hearing the differences between categories of sounds than they are at hearing the variations within a category of sound Morphemes and Words ​ Generativity- capacity to create an endless series of new combinations all built from a small set of fundamental units Syntax ​ Syntax- rules that govern the structure of a phrase or sentence (grammar) ​ Phrase-structure rules- define overall organization of the sentence and determine how various elements are linked to one another ​ Prescriptive rules- describe how something is “supposed to be” ​ Descriptive rules- characterize language as it’s ordinarily used by fluent speakers and listeners Sentence Parsing ​ To understand a sentence, a listener needs to parse the sentence determining each word’s syntactic role  ​ Garden-path sentences- initially leads reader to one interpretation but requires a change in this understanding to comprehend the full sentence ○​ People parse a sentence as they see/hear each word leading to parsing errors that must be later repaired Prosody ​ Extralinguistic context- physical and social setting in which you hear or see sentences ​ Prosody- pitch and rhythm cues Pragmatics ​ Pragmatic rules- govern how people actually use language Biological Roots of Language ​ Broca’s aphasia- damage to left frontal lobe ○​ Produces nonfluent aphasia: can understand but can’t speak ​ Wernicke’s aphasia- damage temporal lobe ○​ Produces fluent aphasia: can speak but can’t understand ​ Children are sensitive to patterns in language that they hear ​ Specific-language impairment (SLI)- disorder in which individuals have normal intelligence but experience problems in learning rules of language ​ Overregularization errors- produces a form that is consistent with a broad pattern even though that pattern does not apply to the current utterance Language and Thought ​ Linguistic relativity- people who speak different languages think differently ​ Language shapes thoughts and a way a thought is formulated into words can affect how you think about the thought’s content ​ Language can call your attention to a category which makes it likely that you will have experience in thinking about the category ○​ Promotes fluency Discussion #4 Language What are the 5 Main Properties of Language? 1.​ Symbolic 2.​ Discrete infinity (generativity) 3.​ Structure dependence 4.​ Displacement 5.​ Multiple levels of organization Symbolic ​ Association between the sound “dog” and the object arbitrary ​ Could be any other sound (e.g., perro)  ​ We often use different words to refer to the same thing Discrete Infinity (Generativity) ​ Generativity- capacity to create an endless series of new combinations, all build from the same fundamental units (words in dictionary) ​ Ex: run-on sentence Displacement ​ Language allows us to thin of and communicate about things that are beyond what is currently being senses (or things that may not exist) Structure Dependence ​ Ex: ○​ John killed Bill ○​ Bill killed John ○​ Killed John Bill ○​ Bill John killed Levels of Language ​ Phonology (phonemes)- smallest unit of speech that can be used to distinguish an utterance from another (in a given language) ​ Morphology (morphemes)- smallest unit of meaning within a language ​ Syntax- systematic way in which (categories of) of words can be combined and sequenced to generate meaningful phrases and sentences Phonemes ​ Smallest distinct unit of speech ​ In English, phonemes are made of consonant or vowel ​ Different languages employ different sets of phonemes (and different types, such as “clicks”) ​ Coarticulation- pronunciation of a phoneme is changed by the following phoneme ○​ Ex: I dunno Morphemes ​ Small unit of meaning within a language ​ Morphemes be split into 2 kinds: root words and affixes (prefix and suffix) ​ Ex: ○​ “Dog”: single morpheme ○​ “Dogs”: 2 morphemes (dog and -s [number]) ○​ “Studied”: 2 morphemes (study and -ed [tense]) ○​ “Restudied”: 3 morphemes ○​ “Relate”: only 1 morpheme Syntax ​ Transformation grammar- sentence can be rearranged to express new meanings (relates to structure dependence)  ○​ Ex: “Mark drank a cup of tea” vs. “a cup of tea drank Mark?” ​ Recursion ○​ Sentence = noun + verb + sentence ○​ Ex: “he went to the park” ○​ Recursive sentence: “i thought that he went to the park” Syntax Ambiguity ​ Syntactic ambiguity- sentence may be interpreted in multiple ways due to ambiguous sentence structure ​ Ex: “squad helps dog bite victim” ​ Extralinguistic context can help with understanding of sentence ○​ Explain and give example ​ Factors outside of language itself can affect sentence understanding/comprehension Context ​ Extralinguistic context- factors outside of language itself can affect sentence parsing ​ Prosody- patterns of pauses or pitch changes that characterize speech production ○​ Used for emphasis, highlighting sentence’s intended structure, assert question/statement ○​ Ex: “I never said she stole my money” ​ Pragmatics- how we use words, context in which we say them, and what’s left unsaid ○​ Ex: “class ends in two weeks” Language Questions 1.​ Which of the following statements is TRUE about phonemes? a.​ A phoneme is the smallest unit of speech that can be used to distinguish one utterance from another. b.​ A phoneme is the smallest unit of meaning within a language. c.​ Which phonemes we can hear is determined at birth. d.​ All languages share the same basic set of phonemes, but combine them in different ways, which is what makes languages unique. e.​ Like in vision, the perception of phonemes can be ambiguous so that, like categories, there are fuzzy boundaries between whether a given sound is one or another phoneme 2.​ Which of the following statements is TRUE about sentences? a.​ Syntax is unambiguous and determines the meaning. b.​ Garden Path Sentences demonstrate that we follow one interpretational path to arrive at the correct meaning of a sentence. c.​ Sentence meanings are determined solely by the morphemes and syntactic structure. d.​ The way that we pronounce a sentence can alter its meaning.  e.​ If a sentence is syntactically correct, it must be meaningful. 3.​ If someone asks you what your favorite thing is about this class and you respond with “it ends in two weeks”, what aspect of language are you using to communicate your true meaning? a.​ Phonetics b.​ Semantics c.​ Syntax d.​ Prosody e.​ Pragmatics Why are categories important? ​ Categorization- process through which ideas and objects are recognized, differentiated, classified, and understood Why do we categorize? ​ Cognitive economy ○​ Enable thought and communication ○​ Minimize memory storage (more efficient) ​ Generalization ○​ Inductive inference Categories vs. Concepts ​ Concept- mental representation of a category ​ Category- a set of things that are grouped together by virtue of some shared attributes ​ Ex: ○​ Tallness is a concept ○​ Tall people that we know is a category How do categories arise? ​ Item set that is grouped by some dimension/feature ​ Many possible types of categories ○​ Natural: groupings that occur naturally (cats, flowers) ○​ Artifact: designed/invented (phones, houses) ○​ Stable: generally consistent over time ○​ Ad-Hoc: unstable categories defined for a special purpose or within a specific context “Basic Level” of Categorization ​ Neither too general nor too specific ​ Tends to be used in speaking and reasoning about categories ​ It is easier to explain what features are common to members of basic-level categories than for other levels ○​ Tools vs. hammers Classical Theory  ​ Categories can be defined by necessary (has to be there) and sufficient (all that you need) features, so members are either “all or none” ​ If an enclosed shape has 4 equal-length sides and four right angles it is a square ​ If not, it is not a square Issues with Classical Theory: Typicality Effect ​ Production task: how many examples of birds can you name in the next 30 seconds? ​ Typicality effect- some members (or even nonmembers) of a category are more “typical” than others ○​ Robins are more typical birds than penguins or ostriches ​ Ludwig Wittgenstein (1953) ○​ Many categories have no clearly defined features that work for every member ○​ Proposed instead “family resemblance” Probabilistic Theories of Categorization ​ Failure of classical theory led to proposal that category representation may be probabilistic rather than deterministic (rule-based) ○​ Two approaches ​ Prototype theory ​ Exemplar theory ​ Both are based on the idea of similarity Prototype Theory ​ Imagine two birds perched on a tree branch ​ When we think about a category, we are really thinking about the prototype for the category ​ Categories are represented by the average of all members of the category ○​ E.g., the concept “bird” is represented by a prototype that is very similar to a robin or sparrow and quite different from an ostrich or penguin ○​ The prototype need not exist in the real world ​ The category representation is abstract ○​ Does not store information about specific exemplars Prototypes and Graded Membership ​ Graded membership- objects closer to a prototype are “better” members of the category than objects further from the prototype ​ Some birds are “birdier” than others Problems with Prototype Theory ​ Some categories are highly variable; then, some members might look very different from the averaged prototype ​ Prototype theory assumes that information about individual instances is not stored, or at least is not used to guide categorization judgments  ○​ But people do seem to store information about individual exemplars, and can sometimes be influenced by these specific exemplars The Exemplar Theory ​ Concepts are represented by all of the exemplars that have been experienced ○​ The category “bird” is represented by memories of all previous experiences, or ‘examples’, of birds ○​ When we categorize something, we compare it one or more exemplars retrieved from memory, and decide the category based on the most similar exemplars ​ The category representation is concrete ○​ There is not necessarily a summary of the category Problems with Exemplar Theory ​ Problem: Assumes that many individual exemplars are stored in memory without “blending” ​ Problem: Has trouble accounting for people’s ability to extract general properties of categories to allow classification of new instances Prototype vs. Exemplar Theory ​ Prototype theory: category judgments are made by comparing a new exemplar to the prototype ​ Exemplar theory: category judgments are made by comparing a new exemplar to all the old exemplars of a category or to the exemplar that most readily comes to mind ​ When there is high variability within a category, a new potential category member may have both ○​ Low similarity with all instances of that category (exemplar model) ○​ Low similarity with the averaged prototype of that category (prototype model) Problems with Both Prototype and Exemplar Theory ​ Both rely heavily on the idea of similarity ​ Take a blackbird, remove all of its feathers and clothe it in a suit of bat skin ○​ Is it more similar to a bat or a blackbird? ○​ Would you categorize it as a bat or a blackbird? ​ Similarity is always relative ○​ There are an infinite number of ways in which two things can be similar ○​ We require some way of knowing what features are being compared Theory-Based Categorization ​ We know much more about categories than a list of their features or their values in dimensional space ​ Categories provide explanations for how things work in the world ○​ The same way that theories provide explanations for scientific phenomena ○​ They center on causal relations between entities in the world  ○​ Categories include causal explanations! ○​ Theories guide perception by leading us to believe that particular features are interesting and others are not Knowledge Network ​ Is a prerequisite for category judgments ​ Representation of knowledge is stored in a semantic network of associated information ​ Reaction times go up for longer associative paths The Collins & Quillian Hierarchical Propositional Model ​ Representation: ○​ Nodes are meaningful concepts ○​ Links/arrows are propositions ​ Proposition: ○​ A simple idea unit that denotes the relationship between two concepts (“can” , “has” , “is”, etc.) ○​ Can be used to represent properties (i.e., “robins lay blue eggs”) ​ Hierarchy: ○​ Information about superordinate categories is only stored at that level Hierarchical vs. Revised Model ​ Hierarchical model- rigid, nested structures ○​ Takes time to navigate through structure to retrieve info ​ Revised model- flexible, direct connections ○​ Accounts for strength of associations ○​ Spreading activation triggers retrieval processes for associated information ​ How would you explain the longer reaction time taken to categorize a penguin as a bird compared to robin in terms of the knowledge network? ○​ “Robin” might have closer (direct), stronger connections to “bird” than “penguin” ○​ “Penguin” is less typical and shares fewer features with the most common birds ​ Fewer shared features mean less overlap in activation ○​ The activation spread might take longer to reach the “penguin” node Advantages of this Approach ​ Economy of storage ○​ Propositions that are true of most/all members of a category can be stored just once ○​ Ex: most birds can fly, so you do not need to store the proposition that robins, sparrows, crows, etc. can fly ​ Generalization  ○​ Propositions that are true of a superordinate category, such as birds, can typically generalize to newly learned category members ○​ Ex: I learn about a new bird – the kingfisher – I can make the generalization that it can probably fly Week 8 Mental Imagery: Debating the Nature of Our ‘Mental Theater” Francis Galton (1880) ​ Had subjects use introspection to study mental imagery ​ Pioneered the use of the questionnaire as a research tool ​ Subjects’ self-reports suggested they could inspect mental images as pictures ​ Example instructions given to his subjects: ○​ "Before addressing yourself to any of the Questions on the opposite page, think of some definite object -- suppose it is your breakfast-table as you sat down to it this morning -- and consider carefully the picture that rises before your mind's eye.” 1.​ Illumination -- Is the image dim or fairly clear? Is its brightness comparable to that of the actual scene? 2.​ Definition -- Are all the objects pretty well defined at the same time, or is the place of sharpest definition at any one moment more contracted than it is in a real scene? 3.​ Colouring -- Are the colors of the china, of the toast, bread-crust, mustard, meat, parsley, or whatever may have been on the table, quite distinct and natural? Aphantasia vs. Hyperphantasia ​ Aphantasia- not able to picture/imagine things ○​ Not vivid ​ Hyperphantasia- able to picture/imagine things vividly ○​ More than average person What are mental images? ​ Mental imagery may involve any of the sensory modalities ○​ Imagine a taste, a sound, a touch, a smell ​ Consequential for mental health ○​ PTSD: strong negative intrusive imagery ○​ Depression: insufficient or weak positive imagery Great Imagery Debate ​ Two differing viewpoints as to the nature of mental images ○​ Analog viewpoint: visual mental images are analogous to pictures in the head (“functional equivalence”)  ​ Mental images are an approximation ​ Championed by Stephen Kosslyn ​ Visual mental images are functionally equivalent with pictures: 1.​ Preserves metric spatial information 2.​ Changes with viewpoint 3.​ Empty space explicitly represented 4.​ Experienced using spatial attention ○​ Propositional viewpoint: although we may believe we experience images as “pictures,” the underlying mental representations are actually non-pictorial abstract concepts ​ Mental images are actually like an illusion ​ Championed by Zenon Pylyshyn ​ Mental imagery is guided by symbolic or linguistic representations ​ e.g., structural description ○​ List of features and relationships ○​ e.g., dog: pointy ears on-top-of head, head on-top-of body, four legs beneath body, fur covering ears, legs, body In a Nutshell… ​ Analog Theory ○​ Kosslyn ○​ Image is a picture ​ Propositional Theory ○​ Pylyshyn ○​ Image is a description ​ 3 sided shape ​ 3 angles of 60˚ ​ Small lines bisecting Propositional Representations ​ Thinking of it in words and their meanings Analog Representations ​ Anecdotal evidence: ○​ The experience of imagining feels very much like seeing a picture in one’s mind ○​ But can the existence of analog mental images be demonstrated experimentally? Mental Rotation ​ Experiment: Shepard & Metzler (1971) ○​ Task: ○​ Are objects on left and right same? ​ NO ​ Mentally rotate to decide if they’re the same ​ More example trials: Image and Rotations ​ Reaction times got slower with increasing rotation angle Shepard & Metzler’s Results ​ Reaction time increases linearly with distance ​ As if have image in head that can gradually update Analogous to real rotation ​ These results are observed even without instruction to use imagery, suggesting that  participants spontaneously form mental images and scan them ​ The more you need to rotate something in your mind, the longer it will take fMRI BOLD Data from Mental Rotation Task ​ More dorsal stream brain activity needed for a 40o rotation than for no rotation at all, and even more activity needed for an 80o or a 120o rotation ​ Further apart angle is, the more active the brain region is Propositional Theory’s Response ​ Elaborate structural descriptions can explain rotation results ○​ Angles of features are specified ​ e.g., partial description of Shepard & Metzler stimuli: Let’s Test Your Visual Imagery Ability! ​ Experiment: Reed & Johnsen (1975) ○​ Form a mental image of this picture ○​ Which of the pictures on the next slide are part of this picture? ​ Reed & Johnsen (1975) found that subjects only succeeded 55% of the time ○​ Suggested that visual mental images often lack sufficient details for direct comparisons of part-whole relationships ○​ Perhaps subjects don’t really store images as “pictures” Imagery & Ambiguous Figures ​ Experiment: Chambers & Reisberg (1985)  ○​ Showed ambiguous figures for 5 seconds and asked for first interpretation ○​ Removed picture, asked people to form a mental image ○​ Duck-rabbit illusion ​ Results ○​ People were unable to discover a second interpretation from the image ○​ Then drew the figure from memory and could then find the other interpretation ​ Could find that it could be a rabbit from their own drawing not from original mental image of the picture ​ Conclusion ○​ A propositional code may override the imaginal code in some circumstances ○​ Something limited about mental image Slezak Figures ​ Pick one of these animals and memorize what it looks like ​ Now rotate it in your mind by 90 degrees clockwise and visualize what it would look like ​ NONE of Slezak’s subjects could identify the new animal produced by the mental rotation ​ Argued that mental images are intrinsically bound to an structural interpretation ○​ Consistent with propositional theory (processing with list of features) Scaling of Mental Images ​ Experiment: Kosslyn (1975) ○​ Examined how participants scan and use mental images ○​ Some participants imagine an elephant next to a rabbit ○​ Others imagine a rabbit next to a bee ○​ Then answer questions about the rabbit: ​ Does the rabbit have whiskers? ​ Does the rabbit have ears? ​ Does the rabbit have a beak? ○​ Reaction time to answer is measured ​ If RT’s are the same in both conditions, then verbal code (propositional)  ​ If RT varies depending on the condition, then pictorial code (analog) ​ Just as in visual images, level of detail in mental images can vary ​ Results: it took longer to respond to rabbits paired with elephants than to rabbits paired with bees Kosslyn’s Image Scanning Experiment (1978) ​ Task: 1.​ Memorize map 2.​ Map taken away 3.​ Focus attention on named location 4.​ Hear other location 5.​ Press button when attention at 2nd location ​ Results ○​ “Scanning” farther takes longer ○​ Support for analog representation hypothesis ​ Mental images are internal representations that operate in a way that is analogous to the functioning of the perception of physical objects ​ Functional equivalence Resolution of the imagery debate? ​ Lots of evidence for analog visual images ​ But propositional theory never completely ruled out ○​ Assumes that the mind can rapidly work with extremely elaborate structural descriptions ​ Can neuroscientific data weigh in on the debate? Early Neuroimaging Evidence (fMRI) for Involvement of Visual Processing Areas During Visual Imagery ​ People were lying in the scanner with during periods where the stimulus was on, off, and imagined ​ Observed at BOLD signal in primary visual cortex (1st stop at cortex for processing)  ​ Simulus on: V1 is activated ​ Simulus off: no activation/baseline ​ Simulus is imagined: V1 is activated, almost as high as when stimulus is on ○​ Top-down, no bottom-up ○​ Evidence that visual imagery is picture-like, uses visual cortex Decoding Individual Mental Images From fMRI Activity ​ Hear name of famous person or building and participant has to imagine them ​ Measured BOLD signal in parahippocampal place area (PPA) and fusiform face area (FFA) over time ​ Each arrow was one trial where they were told to imagine something ○​ Ex: in the first trial, FFA activity is high and PPA activity is low so they are imaging a face ​ 85% correct classification at telling whether the participant was imaging a face or building based on the activity of the PPA or FFA ○​ Subject was not seeing anything ​ Neural evidence that when we imagine, we’re activating the same areas for seeing More fMRI Evidence ​ People vary widely in their imagery abilities ​ V1 activity during visual imagery correlates with individual differences in imagery vividness Decoding Much Finer Grained Mental Images from fMRI Activity ​ Decoding- scientists train a computer program by showing it brain-scan data associated with seeing certain images ○​ Once it has built a database of activity patterns, it can be tested with images the participant hasn’t necessarily seen before ○​ During testing, the program must guess the object viewed on the basis of what it has learned about similar patterns of activity ​ Look at activity pattern (voxels) between different items ○​ Each item does NOT have an area of activation ​ Experiment: Stokes et al., 2009 ○​ Imagery task: participants had to imagine looking at the letter “X” or “O” ​ They would stare at the screen with a dot in the center, no letters appeared ​ They would hear a tone corresponding to the letter they had to imagine ○​ Perceptual task: participants actually looked at the letters “X” and “O” ​ Used the data to train the model ​ Is there a neural pattern in their brain (V1) that differentiates “X” or “O” ​ Apply model to the imagery task and see if it can make a prediction Decoding Imagery of X’s vs. O’s ​ Find area that differentiates “X” or “O” ​ Imagery decoding results: activity patterns in left lateral occipital region could predict what the subject was imagining 62% of the time (significantly better than chance) ​ Lateral occipital region showed the most robust effect when imagining Decoding More Complex Mental Imagines: Imagining Specific Paintings​ ​ Naselaris et al. (2015) ​ Subjects memorized 5 different paintings​ ​ In the scanner, they were asked to imagine one of the paintings ​ Experimenters classified which painting they were imagining and trained it on perception when subjects are actually seeing the painting ○​ Then they know which painting they were seeing or imagining based on V1 activity Decoding More Complex Mental Imagines: Using Deep Learning Algorithms (Convolutional Neural Networks)​ ​ Participants were seeing or imagining an image​ ​ Experimenters were able to reconstruct what the participants were seeing ​ Experiment: Shen et al., 2019 ○​ Based on imagery, not perfect reconstruction ○​ But takes brain activity pattern and converts it back to what the person might have been thinking about Decoding More Complex Mental Imagines: fMRI Data From Rissman Lab​ ​ Participants learned their way in 3 different virtual environments ​ In the scanner, they were either watching videos of someone navigating the virtual environments or they were mentally imagining  ​ Able to decode which of the 3 environments they were viewing ○​ V1, occipital lobe, parietal lobe ​ Able to decode which of the 3 environments they were imagining ○​ Parietal lobe (spatial attention) ​ Able to decode areas that can do both ○​ Medial parietal lobe Reconstructing Visual Experiences from fMRI Brain Activity Patterns Measured During Movie Viewing ​ Showed participants video clips in the scanner ​ Reconstructed the clips of what they might be seeing based on brain activity patterns Neural Decoding of Visual Imagery During Sleep ​ Evidence came out that brain scans done during sleep could decode the contents of dreams ○​ No bottom-up perception ​ Able to decode about 25 different attributes the person could have been dreaming about ​ Limitations: ethics of mind-reading ○​ Invade privacy of thoughts ​ *debate: for the large part, imagery is built upon of perceptions Making Judgements Making Judgments in Everyday Life ​ Judgment- process through which people draw conclusions from the evidence they encounter ​ How frequently does a given event occur? ​ How likely is a given thing to be in a certain category or have certain properties? When Making Judgments, We Use Heuristics! ​ Heuristics- “rule of thumb” or mental shortcut ​ Often based on past experience  ​ Does a good job most of the time ​ Save us time and energy ​ Not guaranteed to be correct…and errors tell us important things… Heuristics for Human Judgment ​ Amos Tversky (1937-1996) ​ Daniel Kahneman ○​ 2002 Nobel Prize in Economics Tale of 2 Systems ​ System #1 ○​ Intuitive ○​ Automatic/rigid ○​ Immediate/fast ○​ Parallel ○​ Most heuristics are produced by this system ​ System #2 ○​ Analytical ○​ Controlled ○​ Consciously monitored ○​ Rule-governed ○​ Serial (one step at a time) ○​ More cognitively demanding ○​ Flexible ○​ We generally make little use of this system ​ Ex: big decisions ​ System #1 rapidly generates intuitive answers, which can be monitored/evaluated by System #2, though this latter stage may be rare ​ Can hold back System #1 and allow System #2 to catch up to us and accomplish judgements more accurately ​ Ex: ○​ Emily’s father has 3 daughters. The first two are named April and May. ​ What is his third daughter named? Emily ​ Most people would say June ​ Framing effect Availability Heuristic ​ Availability heuristic- judged probability of an event is related to how easily that event can be brought to mind ​ Influenced by a given event being: ○​ Recent ○​ Frequent ○​ Extreme  ○​ Vivid ○​ Negative Availability: Role of Media Coverage ​ Judgments are clearly influenced by how often each factor appears in the media ​ While homicide and fire are often reported, kidney disease doesn’t get as much press ​ Ex: ○​ After reading an article about lottery winners, you start to overestimate your own likelihood of winning the jackpot, you start spending more money than you should each week on lottery tickets ○​ After seeing news reports about people losing their jobs, you might start to believe that you are in danger of being laid-off. You start lying awake in bed each night worrying that you are about to be fired. ○​ After seeing news stories about high-profile child abductions, you begin to believe that such tragedies are quite common, you refuse to let your child play outside by herself and never let her leave your sight. ○​ After seeing several television programs on shark attacks, you start to think that such incidences are relatively common. When you go on vacation, you refuse to swim in the ocean because you believe the probability of a shark attack is high. ​ Ex: ○​ The availability heuristic may lead us to believe that we always do the housework ourselves Wide Range of Availability Effects ​ Schwarz et al. (1991) ​ Asked subjects to recall examples from their lives in which they had acted in an assertive fashion ○​ Half of the subjects asked to recall 6 examples ○​ Half of the subjects asked to recall 12 examples ​ Then participants were asked some more general questions, including how assertive they thought they were ○​ The group that had been asked to generate 6 examples rated themselves as more assertive ○​ 12 is too many for most people to think of, feel hard to do ○​ If you’re not able to reach 12 examples, you’ll start feeling like you’re not that assertive Craig Fox (2006) Experiment ​ How soliciting more criticism can actually boost your course ratings ○​ PS…but only if your class is good! ​ List 2 ways in which the course could be improved (east task) OR list 10 ways (relatively more difficult task)  ○​ Listing 2 ways can lead you to think that the class is actually good ​ Which results in higher ratings? ○​ People who had to list 10 ways gave overall higher ratings Representativeness Heuristic ​ “Steve is very shy and withdrawn, invariably helpful, but with little interest in people, or in the world of reality. A meek and tidy soul, he has a need for order and structure and a passion for detail” ○​ Steve more likely to be a librarian ○​ Most people pick librarian, since these qualities are representative of our stereotype of a librarian ○​ In fact, Steve is 83 times more likely to be a salesman than a librarian ○​ There are more than 15 million salespeople in the US and only 10,000 librarians ○​ Base-rate fallacy: fail to take into account the base-rate ​ If you know nothing, you’re more likely to be influenced by what you read ​ Another experiment: Kahneman & Tversky (1973) ○​ Showed subjects personality descriptions allegedly sampled at random from a group of 100 professionals (lawyers & engineers) ○​ Subjects told: 70% lawyers, 30% engineers (BASE RATE) ○​ Description of an individual: ​ Dick is a 30-year old married man with no children. A man of high ability and high motivation, he promises to be quite successful in his field. He is well-liked by his colleagues ​ Designed not to favor one field over another ​ Subjects predict 50/50 ○​ Base-rate neglect ​ Representativeness heuristic- people judge probabilities based on the degree that the situation is similar to, or representative of, their stereotypes or knowledge ​ They do this even when there is other information that a rational person would use to make the best possible decision ​ Base-rate neglect: tendency to ignore the “prior probability” of an event ○​ People make incorrect judgments even when they are explicitly informed about the base rates ​ The representativeness heuristic may lead us to believe that smoking must be okay for your health based on one example (anecdotal evidence or “man who” stories) ​ We often assume that what is true of one instance of the category must be true of the category as a whole ​ Ex:  ○​ Flier from the American Smokers Party (image of 100 year old smoking), using this one woman to mock anti-smoking efforts and illustrate that smoking won’t necessarily kill you​ ○​ But in reality, smoking causes 1 out of every 5 deaths! And smokers die on average 15 years sooner than nonsmokers! ​ (plus 1.3 million more deaths per year from secondhand smoke) ​ Ex: ○​ Despite ample evidence, some people don’t believe in global warming… ○​ But you’re succumbing to the representativeness heuristic if you mock the idea of global warming just because of a few days of unusually cold winter weather… ○​ Misunderstanding between weather and climate ​ Using representativeness heuristic: taking assessment of current weather trend and extrapolating from that ​ Think cold day doesn’t mean global warming ○​ Actual data ​ Ex: another classic empirical study of the representativeness heuristic ○​ Kahneman and Tversky (1983) presented participants with the following description: ​ Linda is 31 years old; she’s single, outspoken, and very bright. She majored in philosophy. As a student, she was deeply consumed with issues of discrimination and social justice, and participated in anti-nuclear demonstrations. ○​ How likely are the following statements? (A = highly likely; E = unlikely) ​ Linda works in a bookstore and takes yoga classes. E ​ Linda is a bank teller. E ​ Linda is a member of the League of Women Voters. A ​ Linda is an insurance salesperson. E ​ Linda is a bank teller who is active in the feminist movement. A ○​ Conjunction fallacy  ​ Results: participants tend to estimate “Linda is a bank teller who is active in the feminist movement” as more likely than “Linda is a bank teller.” ​ Description represents one’s stereotypes of a feminist ​ People ignore basic probability principles ​ 2 events in combination cannot be more likely than just one of the constituent events ​ People treat the questions as independent; logical error Why We Make These Errors of Judgment: Attribute Substitution ​ Available heuristic ○​ Frequency of occurrence in the world ​ If something easily comes to mind, it must be frequent ​ If it’s hard to think of cases, it must be infrequent ​ Judge things of how easily it comes to memory ​ Very useful but doesn’t always get us the right answer ​ Other factors that influence memory (news coverage, selective nature we tend to encode, etc.) ​ Representative heuristic ○​ Probability of en event being in a category or having certain properties ​ People tend to use what’s available to them, resemblance of event that’s described ​ Use what you know as a basis to make the judgment ​ Many categories are homogenous so it usually works well ​ But there's exceptions and stereotypes can leave us astray More Representativeness… ​ If you flip a coin 6 times, which of the following sequences would be a more likely outcome? ○​ H-T-H-T-T-H ○​ H-H-H-H-H-H  ○​ Equally likely ​ Okay, let’s try another one… ○​ T-H-H-T-H-T ○​ T-T-T-H-H-H ○​ Equally likely Misconceptions of Chance ​ People expect that a sequence of events generated by a random process will be representative of a longer random sequence ○​ In flipping a coin, people think H-T-H-T-T-H is more likely than H-H-H-T-T-T (which seems nonrandom) or H-H-H-H-H-H (which seems like an unfair coin) ​ Gambler’s Fallacy ○​ After a long run of red on the roulette wheel, people think that black is now “due to happen” ​ Chance is viewed as a self-correcting process ​ It doesn’t matter what has happened in the past ​ Ex: casinos and roulette ○​ Casinos often post the last 20 numbers that occurred ○​ People try to use this information to make their bets ​ They might expect that a number is “due” to come up ​ But there are no patterns in roulette Anchoring and Adjustment ​ How do people make numerical estimations? ○​ They typically start by anchoring on some salient number and then adjusting that number ​ Ex: ○​ What percentage of countries in the UN are in Africa? [A roulette wheel is spun to get an initial starting point.] ​ Wheel spins 10 → median estimate is 25 (too low) ​ Wheel spins 65 → median estimate is 45 (too high) ​ True answer: 53 ​ Ex: ○​ Did Gandhi live past the age of 120? ​ NO! ​ How old was Gandhi when he died? ​ Average guess: 67 ○​ Did Gandhi live past the age of 9? ​ OF COURSE! ​ How old was Gandhi when he died? ​ Average guess: 50  ○​ True answer: 78 ○​ It depends on how you ask the question! Practical Applications ​ Jury damage awards: tend to anchor on the plaintiff’s initial claim (encouraging unreasonably high claims) ​ Used car sales: state a high initial price, then quickly offer a “discount” ​ Ex: infomercials where they keep cutting the price and adding more and more gifts for free with the purchase Ch. 11: Visual Knowledge Visual Imagery ​ People describe their thoughts in a variety of ways ○​ Formulates in words ○​ Sequence of ideas that lack concrete form ○​ Sequence of pictures or sounds ​ Galton used self-report data ○​ Asked people to describe their images and rate them for vividness ○​ People differed in the capacity for visual imagery Chronometric Studies of Imagery ​ Chronometric studies- “time-measuring”, studies that measure the amount of time a task takes ○​ Pattern of what information is more available and what is less available in an image closely matches the pattern of what is available in an actual picture ​ Imaging-scanning procedure- determine how long “travel” takes across a mental image ○​ Participants can across their images at a constant rate ​ Same for a task that requires you to “zoom in” or “zoom out” of a mental image ○​ Response times are proportional to the amount of zoom required ​ “Travel” in the imaged world resembles travel in the real world ​ Mental rotation task- determine if the shapes are the same or differ by mentally rotating and aligning them Visual Imagery and the Brain ​ Visual imagery involves mechanisms that overlap with those used for visual perception ​ Imaging one thing can make it difficult to perceive something else ​ Imaging the appropriate target can prime a subsequent perception Individual Differences in Imagery ​ 2 types of imagery  ○​ Visual ​ Ex: colors ○​ Spatial ​ Mental rotation tasks relies on the right hemisphere and shows that men are faster than women ​ Mental folding shows no pattern ​ Ex: navigation ​ Can also use either one for a mental image ○​ Depends on individual's ability ​ Eidetic imagery- person can retain long-lasting and detailed visual images as if they were still physically present ○​ Photographic ○​ Very rare Images are not Pictures ​ Mental images are picture-like ​ Percepts- mental representations of the stimuli you’re perceiving ○​ Organized depictions and unambiguous in a way pictures are not ○​ Perceptual reference frame that guides the interpretation of the images Long-term Visual Memory ​ Images containing more parts take longer to create ​ Images containing more details take longer to create ​ Some information about visual appearance or spatial arrangement is stored in LTM ○​ Ex: location of cities may be stores in terms of propositions rather than a mental map (“Montreal is in Canada; Canada is north of the US”) ​ Imagery improves memory ○​ Materials that evoke imagery are easier to remember ​ Ex: word lists are more readily recalled if the words are easily imaged ​ Instructions to form images help people memorize ○​ Imagery mnemonics help recall performance ​ Only if the mnemonics show objects interacting in some way ​ Dual coding- imageable materials will be doubly represented in memory ○​ The word itself and the corresponding pictures will be remembered ○​ Storing information in both a verbal format and format that encodes appearances (doubles the chances of recalling the material) ○​ Suggests that you have 2 types of information in LTM: symbolic materials and imagery-based materials ​ Boundary extension- tendency for people to remember pictures as being less “zoomed in” than they actually were Diversity of Knowledge ​ Image-based memories are influenced by the same principles as other memories ​ Ex: memory for pictures is influenced by schematic knowledge Discussion #5 Language (cont.) Syntax ​ Syntax-first approach- theory that the parsing of a sentence is first derived based on principles of grammar alone, without regard to the meaning of the words ​ But experiments have shown that we do not rely on grammar alone to parse sentences ​ Minimal attachment: prefer interpretations that involve fewer syntactic nodes or complexity ​ Late closure: as long as it makes sense, we keep attaching the incoming words to the phrase we are processing ○​ Ex: garden-path sentences Phonological Level— McGurk Effect ​ McGurk effect- perceptual phenomenon that illustrates how our brain integrates visual and auditory information during speech perception ○​ It occurs when conflicting auditory and visual speech cues are presented, leading to a combined perception that is different from either the sound or the visual input alone ​ Ex: ○​ Auditory input: The sound “ba” ○​ Visual input: A video of someone mouthing “ga” ○​ Perceived sound: “da” (a fusion of the auditory and visual inputs) ​ This effect highlights the brain’s reliance on multisensory integration for interpreting speech, demonstrating that what we see can influence what we hear ​ It is significant in cognitive psychology as it shows how sensory modalities work together to shape perception, rather than functioning independently ​ The person is making the same sound, but we might hear it differently ​ Why? The mouth makes characteristic movements that can be used to determine which phoneme a person is pronouncing, which can change our interpretation of the sound ​ Takeaway - we use visual information to help us identify sounds ○​ There is much noise in the environment, so we rely on many types of information to identify sounds among noise (including visual cues like mouth movement) Broca’s and Wernicke’s Aphasia  ​ Broca’s area ○​ Region of the brain concerned with speech production ○​ Left inferior frontal lobe ​ Wernicke’s area ○​ Region of the brain concerned with speech comprehension ○​ Left posterior superior temporal lobe Language Impairments: Broca’s Aphasia ​ Patients struggles to produce grammatical speech ​ Sentences are short, usually consist of just nouns and verbs ​ Yet reading and verbal comprehension largely unaffected ​ Lack of complex syntactic structures ○​ Rarely use prepositional phrases (above, less than, etc.) Language Impairments: Wernicke’s Aphasia ​ Patients struggle to comprehend sentences (written or spoken) ​ Sentences mostly grammatical, but often nonsensical ○​ Either because the combination of words is nonsensical… ○​ Or they introduce new words (neologisms) Language and Thought ​ Sapir-Whorf hypothesis ○​ Strong interpretation (linguistic determinism) ​ Thoughts and behavior are determined by language ​ The language you speak determines the concepts and categories that you use, and as a result, shapes what you can think about ​ No solid evidence that certain languages forbid a speaker from thinking about certain concepts! ○​ Milder interpretation (linguistic relativity) ​ Thoughts and behavior are influenced by language ​ Language influences what we pay attention to, and this shapes experience, which influences how we think ​ Suggests that the language we speak influences how we think and perceive the world. It has two main versions: 1.​ Linguistic determinism: language determines thought, meaning we can only think within the constraints of our language (strong claim, largely discredited) a.​ Not completely true 2.​ Linguistic relativity: language influences thought by shaping attention and perception, guiding how we interpret experiences ​ Ex: speakers of languages with distinct color terms (e.g., Russian’s separate terms for light and dark blue) perceive and categorize colors differently than speakers of languages without these distinctions  ​ Ex: Alaskan native words for snow ○​ Often used to illustrate linguistic relativity ○​ How language adapts to environmental and cultural needs, shaping cognitive processes ○​ These languages have many terms for snow due to their polysynthetic structure, which combines morphemes to describe specific snow types (e.g., “falling snow,” “snow on the ground”) ○​ This reflects the importance of snow in their daily lives and survival ○​ Language influences how we categorize and focus on distinctions in the environment ​ For speakers of Eskimo-Aleut languages, these snow-related terms may direct attention to specific features of snow that are less emphasized in other languages ​ Ex: color terms/Berinmo Color Naming ○​ Languages vary in the number of basic color terms they have (e.g., English has 11; some languages have fewer) ○​ These differences affect how speakers perceive and group colors ○​ The Berinmo language, spoken by a small group in Papua New Guinea, categorizes colors differently than languages like English ​ For instance, the Berinmo language lacks distinct terms for some color boundaries (e.g., “blue” vs. “green”) but instead has unique terms for boundaries not recognized in English (e.g., “nol” vs. “wor”). ○​ Speakers of Berinmo were better at distinguishing colors that cross the “nol-wor” boundary than those that cross the English “blue-green” boundary ○​ Conversely, English speakers showed better discrimination for colors crossing the “blue-green” boundary than for colors crossing the Berinmo “nol-wor” boundary ○​ In memory experiments, participants were more likely to misremember colors within the same linguistic category (e.g., “blue-blue”) compared to colors from different linguistic categories (e.g., “blue-green”), suggesting that linguistic categories influence memory and perception Does language influence perception of color? ​ Do Berinmo speakers perceive color differently? ​ If categorical effects are restricted to linguistic boundaries, these groups should show different responses across the two category boundaries (green-blue and nol-wor) ​ If categorical effects are determined by the universal properties of the visual system, then both populations should show the same response patterns Recognition Memory Task  ​ Recognition memory task- tests a person’s ability to recognize previously seen stimuli. In cognitive psychology, it is often used to study how language influences memory ​ Subjects were given a specific Maunsell color chip to remember ​ After a 30 second delay, they were given two target chips (the old one and a new one) and had to recognize the original chip ​ Results: ○​ Both English and Berinmo speakers showed better performance when the two test colors were associated with different color words in their respective languages ​ Berinmo speakers do best on wor-nol trials ​ English speakers do best on blue-green trials Communicative View vs. Cognitive View ​ Communicative view- language does not affect the way we think ○​ We just use language to communicate our thoughts ​ Cognitive view- language determines or affects the way we think ○​ Specific effects: how does our specific language affect our thinking (e.g., Spanish vs Mandarin) ○​ General effects: how does having a language (as opposed to not having any language) affect our thinking? Cognitive View (general): Spelke ​ Humans are similar to other species in some ways, and different in other ways (cognitive achievements) ​ What makes humans different? ​ Spelke’s theory is that language is the thing that makes us different ​ Language allows us to achieve things other species can’t ​ Spelke suggests that language plays a crucial role in integrating and expanding the core knowledge systems humans are born with ​ While these systems are innate and domain-specific, Spelke argues that language enables cross-domain integration, allowing humans to combine information from different cognitive systems to form more complex and abstract ideas ​ Spelke’s emphasizes that core knowledge systems provide the building blocks for cognition, but language enables humans to combine and extend these systems, leading to more sophisticated and uniquely human abilities, such as abstract reasoning and cultural learning Grammatical Gender Effects ​ Grammatical gender effects- refer to how the gender assigned to nouns in some languages (e.g., Spanish, German) influences perception and thought about those objects ​ English ○​ Does not assign a gender to animate vs. inanimate objects  ​ Spanish (as well as many other languages) ○​ Marks gender with morphological info carried by pronouns, determiners, nouns, and adjectives ​ “una niña alta” vs. “un niño alto” ​ “luna” = feminine, “sol” = masculine ​ Lera Boroditsky’s 2003 Experiment: ○​ Do people include gender in their conceptual representations of objects? ○​ Are people’s ideas about the genders of objects influenced by grammatical genders assigned in their native languages? ​ Boroditsky et al. (2003) experiment: ○​ Taught Spanish and German speakers object-name pairs (e.g., apple-Patrick) ​ Name was either consistent or inconsistent with the grammatical gender of the object in their native language ○​ Measured memory for the pairs; all testing was in English ○​ Both Spanish and German speakers remembered object-name pairs better when the gender of the proper name given to an object was consistent with the grammatical gender of the object name in their native language than when the two genders were inconsistent ​ Another Boroditsky experiment: ○​ Spanish and German speakers asked to write down the first 3 adjectives that came to mind to describe various objects ○​ More masculine properties produced for masculine objects (defined by native language), and more feminine properties produced for feminine objects ○​ Ex: key ​ Masculine in German, feminine in Spanish ​ German: hard, heavy, jagged, metal, serrated ​ Spanish: golden, intricate, little, lovely, shiny ○​ Grammatical gender focuses speakers of different languages on different aspects of objects Bilingualism and Cognition ​ Kovelman found that that investigated how bilingualism affects language processing and brain function, while bilingual children may face initial delays in single-language development, they benefit from enhanced brain function and cognitive control in the long term ​ Bilingual children may initially have smaller vocabularies but catch up over time (Kovelman et al., 2008) ​ Some research suggests bilinguals are better at task-switching and avoiding distractions ​ Bilinguals show increased activation in language-related brain areas, such as the  left inferior fronta