W3 Semantic Memory Flashcards
Semantic memory
General world knowledge including objects, people, concepts and words.
What predicts what happens next based on regularities in the world?
Categories, schemata and scripts.
Structure: 2 Options
Sematic Memory
Option 1 = Separately stored representations of information and their various relations, problem = not economical.
Option 2 = Storing representations and their relations in a more economical network.
Hierarchical network model
Collins & Quilian’s, 1969
Access of concept representations through spreading activation between nodes via their connecting paths.
Limitation: does not account for semantic relatedness
Semantic dementia
Syndrome of progressive deterioration in semantic memory, leading to the loss of knowledge about objects, people, concepts, and words.
Categorization
Sematic Memory
Semantic memory enables us to form representations of categories (e.g. “dog”) based on regularities in the world, thereby allowing us to make predictions about what will happen next.
Classical theory of categorization
Categories are defined by necessary and sufficient features. (e.g., separating off and even numbers because odd numbers cannot be divided evenly into groups of two.)
Critcism of Classical theory
Family resemblance, central tendency, graded membership
Family resemblance
Critisms of Classical Theory of Categorization
different members of a category can share different features. This doesn’t take into account of central tendency (average ideal).
Central tendency
Critisms of Classical Theory of Categorization
categories exhibit an average ideal
Graded membership
Critisms of Classical Theory of Categorization
Some members are more typical for a category than others.
Typicality rating
Measuring Categorization
Rank the following chairs from being the best example to being the worst example of a chair: DV = average rank or rating.
= we grade things that don’t need to be, graded membership exists even for odd numbers. They are all odd but still have different ratings.
Exemplar production
Measuring Categorization
Recall as many pieces of furniture as you can. (e.g., chair, desk, cupboard, bed, drawer, lamp.) DVs = frequency pf production and/or position in the production.
Category membership varification
Measuring Categorization
(e.g., is this an exemplar of category? Furniture = carpet, bird = robin, fish = shark) DVs: accuracy of responses and/or reaction times.
Prototype theory (categorization)
Modern Theories of Categorization
Categories are determined by a mental representation that is a weighted average of all category members. This prototype may or may not be an actual entity.
Common features = four legs, furry, tail. Distinctive features = barks, is omnivore.
This theory cannot explain how people can tell the sizes of categories (Many types of dogs, fewer types of elephants.). It also cannot explain How can people add new members to a category.
Address issues with Prototype Theory
Exemplar Theory
Modern Theories of Categorization
Categories consist of separate representations of the physical features of experienced examples of the category. People can tell category sizes, people can add new members.
But it cannot explain = How can people retrieve all category members to define a category if retrieval is based on category membership (theoretical circularity)?
Explanation based theory
Modern Theories of Categorization
Categories are based on common causal characteristics rather than physical features.
In previous accounts waterfowl = animals with webbed feet. Explanation-based account waterfowl = animals that swim.
Categories can be created ad hoc using world knowledge and explanations. (things with a distinctive smell, perfume, coffee)
Barsalou’s 1983 experiments
Do ad hoc categories (e.g. things with a distinctive smell) have the same features as common categories (e.g. fruit)?
Family resemblance, Central tendency, Graded membership
High average agreement among participants regarding category membership, typicality of members, and production of exemplars.
Ad hoc categories are similar to common categories in that they exhibit family resemblance, central tendency, and graded membership.
Schemata
Forming Concepts
Semantic memory also enables us to form schemata (e.g. “buying things”) that capture commonly encountered aspects of life. They can be thought of as explanation-based event categories.
Scripts
Forming Concepts
Semantic memory also enables us to form scripts (e.g. “eat in a restaurant”) that capture the order of events for common aspects of life. They can be thought of as temporally ordered schemata.
(Alba & Hashers, 1983)
Five primary schema processes
Schema Process
Encode = 1- selection, 2- Abstraction, 3- Intrepretation, 4- Integration
Retrieve = 5- Reconstruction
Encode (1)
Selection
Five primary schema processes
of information central to a schema. Matching preference against supply, quality, and prices.
Bransford & Johnson’s (1972) experiments = schema activation benefits encoding of schema-relevant information
Encode (2)
Abstraction
Five primary schema processes
Carmichael et al.’s (1932) experiments
The surface form of information (e.g. ambiguities in pictures, verbatim wording) is converted into a more abstract representation that captures the meaning but is schema-consistent.
Encode (3)
Interpretation
Five primary schema processes
Johnson et al.’s (1973) experiment = Interpretation is used to “fill in the gaps” in a story with schema-consistent information.
Encode (4)
Intergration
Five primary schema processes
Integration of information is used to form schema-consistent holistic representations.
Bransfored et al., 1972 experiments
Retrieval
Reconstruction
Five primary schema processes
Bartlett’s (1932)
Brewer & Treyens (1981)
Bartlett’s (1932) experiments: British students studied the Native American tale The War of the Ghosts and were asked to recall it after days, weeks, months, or years. Details were reconstructed to be simplified and fit cultural schema. (canoes = boats, paddles = rowing, protagonist from “Egulac” = British warrior.)
Brewer & Treyens (1981) experiments: Participants were waiting in a graduate student office and later asked to recall everything they could remember about the room.
Participants recalled: Correctly = Chair, Desk, Poster, Skull. Incorrectly = Books, Filing cabinet.
Prototype approach
idea that concepts are represented based on a typical (common) instance of that concept.
Exemplar approach
: idea that concepts are represented based on exemplars of the category that one has experienced previously.
Basic-level concept
level of concept hierarchy where common objects reside
Believe in Yourself
You’ve got this!