Concepts & Categories

Concepts & Categories - Concepts& Categories...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Concepts & Categories Concepts Concepts: an abstract notion or idea, a mental Concepts: symbol, bearers of meaning. Concepts are abstract in that they omit differences in the bearers of their extension. extension Categorization: Process by which ideas & objects are Categorization: recognized & understood; how people group items, organize these items in memory, & understand their extension. extension. Psychologists interested in HOW people represent Psychologists categories categories Function of Categories? Function Ability to categorize serves as a kind of Ability cognitive shorthand: cognitive Helps organize our knowledge Permits linguistic label for classes of objects, in Permits order to communicate about them order Permits categorical inference (if apples are sweet, Permits and this is an apple, it must be sweet) and I. Category Basics I. A. Ontological Categories A. Ontological 1. Nominal Kinds: dependent on a definition 1. Nominal Examples: Examples: What is a prime number? How to define a bachelor? Are these Bachelors? Are Pope Benedict? A recently Divorced Man? Pope Jean-Paul Sartre (who cohabitated with Simone de Beauvior most of his life)? Simone Best example? Best Or….? Or….? 2. Natural Kinds: that appear naturally in 2. Natural that world world Animals Plants Rocks Minerals 3. Artifacts: Human-made, typically to serve 3. Artifacts Human-made, some function some Furniture Vehicles 4. Further Distinctions thought to have an innate basis 4. are Living vs. Non-living, or Animate vs. Inanimate are Animate Living Inanimate Humans Animals Plants/flowers Fruits Vegetables Rocks Musical Instrumts. Tools Furniture Vehicles Robots Robots Non-Living Inanim. Inanim. (Motion) (Motion) B. Hierarchical Categories Hierarchical ANIMAL DOG Chihuahua Collie Poodle Salmon Clownfish Shark FISH BIRD Canary Robin Macaw 1. Levels 1. Levels a. Superordinate: ANIMAL a. Superordinate ANIMAL b. Basic Level: DOG, FISH, BIRD b. Basic c. Subordinate: Salmon, Clownfish, Shark c. Subordinate II. Feature Based Theories Feature Class Exercise: 1. List all attributes of triangle 2. List as many BIRDS as come to mind in 2 2. minutes A. Classical View Classical 1. Necessary & Jointly Sufficient Features 1. Necessary Necessary: Must have a given feature (or set of Must features) features) Jointly Sufficient: If have ALL the criterial features, If MUST be in the category MUST 2. Implicit Assumptions: 2. Implicit Firm Category Boundaries: a token is either IN or Firm OUT OUT All Members Equal Which is the “Best” Triangle? Which Problems with the Classical View? Problems 1. Even strictly-defined nominal 1. categories appear to have a graded structure (some members are “better” than others) than E.g., Triangle 2. Necessary & JS Features? Game (Wittgenstein) Bachelor B. Family Resemblance Theory (Rosch) Family 1. Graded Structure 1. Graded Prototype: Best Best example of a category (most central; but is NOT a chimeric averaged prototype!!) averaged Peripheral Members: least characteristic members; may be considered members of similar categories similar a. Linguistic Hedges a. “Technically, a whale is a mammal.” “Loosely speaking, a tomato is a vegetable.” b. Rosch (1975) asked subjects to rate “goodness-of-example” members of 6 superordinate categories: Furniture Vehicle Fruits Weapons Vegetables Clothing Rank ordered 20 members per category: Furniture Vehicle 1 Chair Car 2 Sofa Truck 3 Table Bus 4 Dresser Motorcycle 5 Desk Train .. 10 Piano Tractor 11 Cushion Cart 12 Mirror Wheelchair .. 18 Closet Skates 19 Vase Wheelbarrow 20 Telephone Elevator Fruit Orange Apple Banana Peach Pear Grapefruit Pineapple Blueberry Coconut Tomato Olive c. Armstrong, Gleitman & Gleitman (1985) asked Ss to rate the “goodness of example” of both: both: 1-Prototype categories from Rosch (fruit) 2-Well-defined categories: odd numbers, plane 2-Well-defined figures, females figures, --Graded structure found for BOTH prototype --Graded & well-defined categories (even though subjects reported knowing that well-defined categories were Either/Or membership) categories Armstrong et al. (1985) then had Ss engage in a Armstrong category verification task: category A banana is a fruit. An avocado is a fruit. An A carrot is a fruit. 57 is an odd number. 13 is an odd number. 42 is an odd number. T T F T T F Prototypical items (in yellow) had lower RTs (and Prototypical were thus responded to more quickly)* quickly)* *some have claimed this shows the limitations of a resemblance view of *some concepts—category membership can be divorced from typicality ratings. concepts—category 2. Emphasis on Characteristic Features Emphasis a. there are no necessary & jointly sufficient a. features for categories features b. rather, there are overlapping characteristic b. features among members (features which are common to many, but not all, members of the not members set); e.g., set); “has peel” of FRUIT “runs on gasoline” of VEHICLES 3. Family Resemblance Family a. Categories structured according to overlapping a. [characteristic] features, rather than sets of N & JS features features b. Rosch & Mervis (1975, Exp. 1): 1-Ss were given the name of one exemplar from each 1-Ss of 6 superord. categories, and asked to list attributes of that object of e.g., BICYCLES have 2 wheels, pedals, e.g., handlebars, used for transportation, don’t have engines engines Attributes per category received a weight of 120 (depending on number of exemplars within 20 that category which exemplified that attribute) that Examples Of 20 vehicles, _13__ have wheels Of 20 vehicles, _11__ run on gasoline Of 20 vehicles, _1__ are living 3- FINDINGS: Very few attributes common 3- FINDINGS Very to ALL members of a category to Prototype items had highest number of shared Prototype characteristic features characteristic 4. Fuzzy Boundaries Fuzzy a. Because of emphasis on overlap of characteristic a. features, some peripheral members may share attributes with other, similar categories attributes --e.g. whales considered fish; tomatoes considered --e.g. vegetables vegetables b. Thus, peripheral members share some family b. resemblance to members of other categories resemblance 5. Similarity-based Comparison: Similarity-based Prototype as Standard Prototype When confronted with novel item, mentally When tabulate its resemblance to prototypes of competing categories competing e.g., is an ostrich a BIRD or MAMMAL? Compare to SPARROW vs. DOG C. Exemplar Theory (Hintzman) 1. Feature theory which claims we (implicitly) store 1. information about every exemplar of a category to which we’ve been exposed. New examples automatically activate similar exemplars. automatically 2. Exemplar as Standard: Novel items can be 2. Exemplar Novel compared to EITHER 1-an exemplar-- e.g., OSTRICH compared to FLAMINGO 1-an & thus categorized as a BIRD, or: or: 2- a prototype (based on mental calculation of most 2characteristic features, summed across exemplars) characteristic 3. Typical instances are recognized faster and 3. more easily because they activate MORE exemplars (and thus achieve threshold more quickly). quickly). 4. Variability within categories is preserved. III. Knowledge-Based Theories Knowledge-Based A. Psychological Essentialism Psychological (Keil, Medin, S. Gelman) 1. People act as if they know that many concepts 1. have underlying essences. These essences then predict more obvious features, and help explain correlations of features e.g., has a heart, has legs, warm-blooded, has hair, are all e.g., has are features of mammals. Warm-blooded & has hair are Warm-blooded has correlated because of need to maintain warmth. correlated e.g., genetic essences for natural kind categories (BIRD); e.g., genetic functional essences for artifacts; chemical essences functional 2. Essences may serve merely as 2. “placeholders:” We need not fully understand genetics to know that animals have genetic essences. We can know THAT something has an essence without knowing what that essence may be (but we’re confident that experts know). know). 3. Categories thus have internal coherence, 3. which permits category inference. which From Susan Gelman From Essentialism is the view that certain categories (e.g., Essentialism women, racial groups, dinosaurs, original Picasso artwork) have an underlying reality or true nature that one cannot observe directly. Furthermore, this underlying reality (or "essence") is thought to give objects their identity, and to be responsible for similarities that category members share…recent psychological studies converge to suggest that essentialism is a reasoning heuristic that is readily available to both children and adults. available 4. Developmental trend from 4. Family Resemblance-style categories to belief in essences categories E.g., Keil’s transformation E.g., experiments (raccoon to skunk) experiments Younger children swayed by a change Younger in the external, characteristic features in Older children required more Older fundamental, essential changes (e.g., in parenthood) in B. Other Knowledge-Based Views Other 1. Labov’s Cup/Bowl 1. Experiment Experiment Neutral Coffee Food (Rice) Flowers Cup vs. Bowl Cup 2. Barsalou’s Ad Hoc 2. Categories: we don’t Categories we have statically stored “great gifts for Mother’s Day” but can easily construct such a category on-line, based on our goals. on IV. Further Research IV. A. Characteristic vs. Essential Features Characteristic 1. Malt (1994): Water is the quintessential 1. Malt Water “essential” category. Psychological Essentialism predicts that the relative quantity of pure H20 in substances should determine whether they are categorized as “water” or not. whether a. Methods Methods Experiment 1: examples of water & related liquids examples collected from media use, a computer search, and from students. Students then judged average % of H20 in each of 43 “waters” and 55 “non-waters.” Students then asked to judge percentage of water Students composing each of the Waters & Non-waters composing Examples & Average Judged % of H2O Pure Water Rain Water Tap Water Babbling Brook Water Sewer Water (lowest) Salt Water Tea Urine Windex Blood Bleach Bleach Cool Whip (lowest) 98.1 90.0 87.7 85.6 67.0 68.8 90.1 79.1 70.1 68.7 51.3 51.3 41.8) b. Conclusions b. Conclusions “Water” does not refer only to liquids deemed Water” to be entirely composed of H2O; average of 67% for non-waters vs. 87% for waters. Only 5% of Waters judged to be 90% H2O or Only above. above. Even most of the non-waters deemed to be Even constituted of a majority of water (over 50%). constituted Essentialist view cannot fully explain the Essentialist tremendous overlap in perceived % of H2O between Waters & Non-waters, nor that most Waters deemed to be <90% H2O. Waters c. Typicality Ratings Typicality Experiment 2: On a 1-7 scale, subjects asked On to judge the typicality of 43 waters: to Drinking water Rain water Salt water Pond water Pond Muddy water Radiator water 6.5 6.1 5.7 4.9 4.9 3.7 2.7 (lowest) Chemical makeup is somewhat important in Chemical determining whether something is Water or Non-water, BUT Non-water, Other factors which seem to be important Other when judging “Waters:” when Familiarity Location/Source How used in human lives Role played in human lives B. Artificial Stimuli & Concept Learning Artificial PLUROSIS Patient Skin A.B. rash A.B. C.D. rash C.D. E.F. sores E.F. G.H. rash G.H. I.J. spots I.J. K.L. oily K.L. MULTINOMIA Patient Skin O.P. scaly O.P. Q.R. scaly Q.R. S.T. oily S.T. U.V. scaly U.V. W.X. sores W.X. Y.Z. spots Y.Z. Eyes dry red dry sunken dry cloudy Eyes puffy sunken puffy cloudy puffy red Breathing shallow rapid rapid uneven wheezing rapid Breathing shallow labored labored uneven wheezing labored 1. Tests for a Graded Structure Tests Prototype-High Plurosis: rash, dry, rapid Multinomia: scaly, puffy, labored Prototype-Medium Plurosis: rash, dry, uneven Multinomia: scaly, puffy, uneven scaly, 2. Tests for an exemplar effect 2. Tests Plurosis: spots, ?, wheezing Multinomia: sores, ?, wheezing 3. Attribute weighting: which feature is 3. Attribute which weighted most heavily in categorization weighted V. Category Specific Impairments Category A. Theories Theories 1. Category Specific Impairment (Caramazza 1. Category & Shelton, 1998): Functionally distinct category representation is based on an evolutionary need to distinguish living from non-living items (and within LIVING, non-living LIVING Animate from Inanimate things). Animate Representation of Living vs. Non-living objects is Representation subserved by neuroanatomically distinct areas subserved Deficits for living things (Basso et al. 1988; Deficits Bunn et al., 1988; DeRenzi & Lucchelli, 1994; Hart & Gordon, 1992; Hillis & Caramazza, 1991; Laiacona et al. 1997; Moss et al., 1997; Sartori & Job, 1988; Warrington & Shallice, 1984) 1984) Non-living: deficits for Artifacts/man-made deficits objects (Warrington & McCarthy, 1987; Hillis & Caramazza, 1991; Moss & Tyler, 2000; Sacchett & Humphreys, 1992; Warrington & McCarthy, 1983). Living things may be further dissociated into Living Animate (animals) vs. Inanimate (fruits, vegetables, flowers): vegetables, deficit for fruits & vegetables (Hart et al., 1985; deficit Sampson & Pillon, 2003; Sheridan & Humprheys, 1993); animals intact 1993); deficit for animals (Caramazza & Shelton, 1998), deficit but fruits/vegetables intact but Deficits for living things, also musical instruments 2. Sensory/Functional Distinction (Warrington 2. & McCarthy, 1984; 1987; Farah & McClelland model, 1991) model, a. Differential weighting of features per category a. Differential type: Objects/Artifacts are mainly distinguished by function Objects/Artifacts function (hammer vs. a wrench); living things are mainly distinguished by sensory living sensory characteristics (e.g., cabbage vs. cauliflower). b. Suggests that patients who have trouble b. with living things should also have trouble with other items distinguished visually (e.g, foods & musical instruments; e.g., Sartori & Job; Warrington & McCarthy, 1984). Job; (However, some patients impaired at living some things are NOT impaired for musical instruments (DeRenzi & Lucchelli, 1994) or for foods (Hart & Gordon, 1992).) for 3. “Visual Crowding” View: claims that living 3. “Visual claims things may be most often impaired because the degree of overlap in visual (and other) features among living things is greater than the featural overlap among non-living things. overlap Thus, less discriminability means that living things Thus, (including fruits, vegetables) are more likely to be impaired B. Case Studies Case 1. Hillis & Caramazza (1991): Double 1. Dissociation Data Dissociation P.S. was impaired at naming tasks and P.S. comprehension/definition: 47% Animals, Fruits, Vegetables; 93% Non-Living items (bilateral temporal & L frontal damage) (bilateral J.J. was impaired: 12% naming non-living J.J. objects; 81% animals objects; (L temporal, basal ganglia damage) 2. Kolinksy et al. (2002): 2. Kolinksy E.R.: deficits for Living, biological entities E.R.: across tasks (in naming, definition, matching, drawing from memory, questionnaires); in visual or verbal presentation; and when presentation and responses were verbal, pointing, or visually responses matching Recognition task data (controlled for familiarity) Biological Biological Insects 12.5% correct Vegetables 20% Birds 22 Fish/Mollusks 33 Reptiles 40 Fruits 46 Mammals 48 Artifact Artifact Musical Instr. 72 Toys 72 Furniture 79 Buildings/parts 86 Vehicles/parts 89 Clothes/jewelry 91 Tools/weapons 100 --biggest deficits are for less typical, infrequent, and unfamiliar items --biggest 3. Patient data doesn’t always fall neatly into 3. Domain-Specific of Sensory/Functional classifications. Some patients are impaired when tested one way (e.g., picture recognition) but not when tested another (e.g., definition) but 4. Neuropsychological Implications 4. Neuropsychological a. Deficit for living things most associated a. with Herpes Simplex Encephalitis, bilateral (or Left) antero-medial & inferior temporal lobe damage (Gainotti et al., 1995; Pietrini et al., 1988). 1988). b. Deficits for artifacts/non-living items is b. most associated with largely left-hemisphere frontoparietal damage (Gainotti et al., 1995) & left posterior medial temporal lobe (Moore & Price, 1999). Price, 3. Further case studies 3. Further Deficits for living things (Basso et al. 1988; Bunn Deficits et al., 1988; DeRenzi & Lucchelli, 1994; Hart & Gordon, 1992; Hillis & Caramazza, 1991; Laiacona et al. 1997; Moss et al., 1997; Sartori & Job, 1988; Warrington & Shallice, 1984) Job, Deficits for Artifacts/man-made objects Deficits (Warrington & McCarthy, 1987; Hillis & Caramazza, 1991; Moss & Tyler, 2000; Sacchett & Humphreys, 1992; Warrington & McCarthy, 1983). c. Further Distinctions within “Living” Category deficit for fruits & vegetables (Hart et al., 1985; Sampson deficit & Pillon, 2003; Sheridan & Humprheys, 1993); animals intact intact deficit for animals (Caramazza & Shelton, 1998), but deficit fruits/vegetables intact fruits/vegetables ...
View Full Document

This note was uploaded on 12/17/2010 for the course PSY 43785 taught by Professor Reed during the Spring '10 term at University of Texas.

Ask a homework question - tutors are online