DOC PREVIEW
BU PSYC 358 - Concepts and Knowledge Representation
Type Lecture Note
Pages 4

This preview shows page 1 out of 4 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 4 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 4 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

Econ 101 1st Edition Lecture 18 Outline of Last Lecture II. Mental Representationa. Definitionb. Difficulty of studying the ideac. Distinctionsd. Featural representationse. Propositional RepresentationsOutline of Current Lecture III. Concepts and Knowledge Representationa. Multiple Levelsi. Superordinateii. Basiciii. Subordinateb. Taxonomic EconomyCurrent LectureDay 19 – Concepts and Knowledge Representation4/2/15Question of the Day: how is knowledge organized and structured?-For exam: pg. 260-271 won’t be the basis for an essay questionDo we have concepts that are restricted or do we have knowledge that represents ideas.-How much goes in the category and how much knowledge lies between the categories The basic level and how knowledge affects categories (Rosch)-Some categories are privileged within a hierarchical organization (fig. 9.8Animal SuperordinateDog Bear BasicCollie Poodle Grizzlie Polar SubordinateWhy have multiple levels? (fig. 9.9)Property inheritanceThese notes represent a detailed interpretation of the professor’s lecture. GradeBuddy is best used as a supplement to your own notes, not as a substitute.-don’t have to store the info at every level-If you know something is a dog and that there is something true of all dogs, you don’t have to store that info about a poodle or collie. This is called Cognitive Economy.Different levels are useful for different purposes-Inferential power (# of common features)-Within-category similarity vs. between-category difference-At the highest level of the super ordinate. They don’t have a lot ofsimilarity, they are uniquely separated from other animals. The purpose is to separate categories and within the subordinate category there is very little difference between things. Its is basically saying all poodles are going to be alike. Privilege level is acompromise between the two levels. It’s the basic level. The cleanest level in terms of giving a satisfying breakdown between what it means to be in the category and not be in the categoryEvidence for the basic level (Rosch)-Lexicalization in language-We don’t always have words at the subordinate level, but we always have words for the superordinate level-Naming performance (fig. 9.10)-When we show people an object, they tend to use the name used at the basic level. When you show someone a chair, they just call it a chair, not a rocking chairor a 3-legged chair.-Commonality of shape and parts-The basic level was the highest level where the physical similarity is still held. Its where things still seem to be similar-Developmental order of acquisition-Kids first learn basic levelsBUT, tends to vary systematically depending on:-Expertise (fig. 9.11)-As you start to get more and more knowledge about something, the basic level can shift. You are able to make different subordinates for things. If you are an expert on birds, bird becomes a superordinate level for you. Then there would be finches as a basic level and goldfinch as a subordinate level.-Culture-The Itza maya tend to categorize more like U.S. experts also, the cultures project inferences differently. Less reliance on similarity based taxonomyBeyond the use of similarity for taxonomic categorization-Goal-derived categories (Barsalou, 1983)-e.g. pets, rewards-Have an organizing basis that is different than Rosch-Graded membership but violates family resemblance-Goodness based on central tendency around ideals-Ad-hoc’ categories-Are not necessarily stable or regular forms of knowledge but can be generated very quickly-“Things to use instead of a hammer”-“Things to take out of house in case of fire”-Goal derived categories can be generated on-line-Not necessarily based on stored examples form experience-Relational categories-Based on a certain relationship between things-E.g. boundary, revengeProblems with Categorizing by Resemblance-The ‘pizza vs. quarter’ study (Rips, 1989)-Imagine a three-inch round object-Is it more similar to be a pizza or a quarter?-People say Quarter-Is it more likely to be a pizza or a quarter?-People say Pizza-Failure of similarity. When people are thinking about similarity, they are also thinking about variabilityFurther Complexities with Similarity/Resemblance-Similarity is context sensitive-Similarity of things depend on the context-Rate similarity given a label context (Barsalou, 1982)-Similarity of dog and cat?-Similarity of dog and guinea pig-Similarity of dog and raccoon?-Similarity of dog and wolf?-In context of ‘pet’ other options seem more similar-From perspective of…-Secretaries or truck drivers or schoolchildren-How typical a beverage is: coffee/milk/tea-Perspective of a child, milk is most typical-Perspective of an adult, coffee or tea might be most typical-Rating similarity in a comparison context (Tversky, 1977)Italy, SwitzerlandItaly, Switzerland, Brazil-Italy and Switzerland are seen as more similar when compared to BrazilItaly, Switzerland, Sweden-Switzerland is seen as less similar to Italy when compared Sweden-Similarity is vacuous without specifying respects (Goodman, 1972)-Similarity of: lawn mower and plum-Could use context of backyard-Knowledge needed to guide evaluation of resemblanceConcepts as ‘theories’ (Murphy & Medin, 1985)-Applying a concept is more like reasoning than matching-Explaining available information in terms of causal principles-E.g. a fully clothed man jumps into a swimming pool-How is he categorized? He is categorized as drunk-Concepts are based on common features-But they are picked out from the top-down by a causal theory-Concepts depend on relations between and within-So once again…-What is a horse? What is a table?The case for theories underlying concepts-Visibility of potential features-e.g. vehicle construction materials-Projectability of features-Consider an example of a novel animal that-Is brown-Is tall-Has a tail-Does tricks-Has two hearts-Has two stomachs-A novel animal has an armored exoskeleton, is it fast or slow?-You would assume it was slow because the armor would weigh it down-Correlational knowledge, but must be supported by explanation-Features are theory-embedded (Medin & Shoben)-Is it still a banana if it is not curved?-Most people would say yes-Is it still a boomerang if it is not curved?-Most people would say no because the curvedness is more fundamental to the structure of the designGrey vs. black vs. white (Hair versus clouds)-In what context is more similar-What are the


View Full Document

BU PSYC 358 - Concepts and Knowledge Representation

Type: Lecture Note
Pages: 4
Download Concepts and Knowledge Representation
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 Concepts and Knowledge Representation 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 Concepts and Knowledge Representation 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?