DOC PREVIEW
U of U PSY 3120 - Memory - Categorization and Organization

This preview shows page 1-2 out of 7 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 7 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 7 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 7 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

Page 1Memory: Categorization and OrganizationMarch 22, 2001More than just typicality--levelsBasic levelcategory: most inclusive but members still share most of their features. E.g., all “birds” are “winged,” “lay eggs,” “sings” etc. Superordinate level category: one level more abstract. E.g., “animals” do not all share features: some are “winged” some are not; some are “tailed” some are not; some are “warm-blooded” some are not, etc. Subordinate level category: less abstract than basic level. E.g., “wrens” are all very similarMore examplesMusical instrument Guitar Folk guitar Classical guitar Piano Grand piano Upright piano Fruit Peach Freestone peach Cling peach Grapes Concord grapes Green seedless grapes Tree Maple Silver maple Sugar maple Birch River Birch White Birch Oak White oak Red oak Musical instrument Guitar Folk guitar Classical guitar Piano Grand piano Upright piano Fruit Peach Freestone peach Cling peach Grapes Concord grapes Green seedless grapes Tree Maple Silver maple Sugar maple Birch River Birch White Birch Oak White oak Red oak Superordinate Basic Subordinate SubordinateLevelsBasic level categories are argued to be psychologically privileged:• We usually use in language• Faster to verify an object name at the basic level than at the superordinate or subordinate level.Note that basic level depends on expertise. If you’re an expert, what is a subordinate level for everyone else may be a basic level for you.What type of representation can capture these typicality effects?The big answer has been probabilistic, or similarity models. •Probabilistic: category membership is a matter of probability, not all-or-none. •Similarity: because the likelihood of being a member of a category is calculated by computing the similarity of the exemplar to the categoryPage 2Similarity model 1: prototypesStart with dot patterns = prototypesCreate exemplars by randomly moving some dots.Study patterns until can categorize correctlyOne week later, recognition test for old, new, and prototype.Subjects were as good on prototype as on old items.How a prototype model worksWhen you see a new exemplar you compare it to the prototype; the more similar it is to the prototype, the more confident you are that it belongs to that category.Thus, when you see the prototype dot pattern, you’re very confident that you know its category.Exemplar modelIt almost feels like the prototype model hasto be right. How else can you account for the success in recognizing the prototype so well?Here’s how. . . . Exemplar modelNote: Exemplar model also accounts for the fact that you can recognize specific dogs, not just the prototypical dog.Exemplar modelHow do you get typicality effects out of the exemplar model?Stored in memory are lots of exemplars of birds that sit in branches and go “tweet tweet.”Which is right?This is unclear. Research strategy has been mathematical psychology, and all of the models are pretty successful.Page 3More general: problems with similarity models• How to select features?• Similarity depends on context• At times, we do seem to categorize via rules (as the classical view suggested)Feature selection: Compare an elephant and a grapeSimilarBoth cannot jump 6 feetBoth are not found on the moonBoth are not yellowDifferentDifferent sizesDifferent weightsAnimal vs. fruitIs there a principled way to select a set of features?Similarity depends on contextA B C DE F GH44% 42%14%12% 80%8%Subjects asked to judge similarity of A to (B, C or D) or of E to (F, G, or H). % similar to A or E.Note that B and F are the same, but % differ greatly, based on other faces present (i.e., context).Categorization based on rules“The object is 3 inches in diameter. Is the object a pizza or a quarter?” Pizza: 63% Quarter: 37%This shows that people are sensitive to the fact that a quarter is not free to vary in size. Note that a quarter is more similar in size to a 3 inch object.Either rule or similarity? Allen & BrooksSome told to memorize exemplars. Others told to categorize using a rule . At test they see creatures that follow the rule to define them as one creature, but they are more similar to the other creature.Rule = at least two of (long legs, angular body, spots)->builderResultsWhen confronted with builders who were similar to diggers, subjects’ categorization depended on the instructions they had received:Told to memorize: categorized based on similarityTold the rule: categorized based on the rule.Next question: when do people use rules, and when similarity?Page 4PossibilityWe don’t really know the answer to that, but one possibility is that they use rules when there are no features that are characteristic of any category.When characteristic information is available, then people will use similarityBack to Pizza and QuarterThe object is 3 inches in diameter. Is the object a pizza or a quarter?” The object is 3 inches in diameter and silver colored. Is the object a pizza or a quarter?” Memory: OrganizationThe problemHow do you get the information needed from memory?How do you find the right book in a big storehouse?An obvious option is to number all the books. This is called an addressing system.Computer hard drives also use addressing systems.Page 5Your mind clearly does not work this way.1. More books means slower search2. Slight errors mean possibly big mistakesIt doesn’t take us longer to remember things as we learn more. When we make a mistake it’s typically a near miss.Even more impressive. . .If you dropped a raw egg from the Eiffel tower, would it break?How do you know that?You probably have not observed an egg being dropped from the Eiffel tower, so what happens is not in memory. How, then, do you answer the question?Your memory system provides relevant facts:• Eiffel tower is tall• Eggs dropped from a great height will break.Theories of memory designed to account for. . .Ready access to lots of informationAccess to relevant information not directly storedNetwork TheoriesNodes: concepts (red, candy, bird)Activation: some level of “energy”Links: relationships between conceptsCollins & Quillian Hierarchical ModelLiving thingAnimalPlantBirdFishChickenCanaryis ais ais ais ais ais asingyellowcaniswingshasfeathershasskinmovehascanbreatheeatmustmust1. Accounts for property inheritance (“does a canary breathe?”)2. Accounts for identity questions (“is a canary


View Full Document

U of U PSY 3120 - Memory - Categorization and Organization

Download Memory - Categorization and Organization
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Memory - Categorization and Organization and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Memory - Categorization and Organization 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?