CH. 9 CP
Conceptual Knowledge
Understanding Categorization
Introduction to Categories
Objects are placed into categories based on shared characteristics and attributes.
Examples: people, buildings, street scenes.
Category understanding enables recognition, inference about properties, and ease of communication.
Categories provide a framework for organizing vast information.
Basic Properties of Concepts and Categories
Functional Role of Categories
Categorization aids in interpreting the environment and guides action (e.g., using jam, bread, knife).
Categories offer knowledge pointers to predict behaviors of objects (e.g., fan behavior based on appearance).
How Are Objects Placed into Categories?
Definitional Approach
Initially considered a viable means to categorize objects (e.g., geometric definitions).
Issue: Natural objects don't fit neatly into definitions due to variability.
Example: A "chair" can vary greatly in form and function.
Family Resemblance
Introduced by Wittgenstein to explain categorization variations.
Objects within a category resemble each other but might not meet strict definitions (e.g., chairs).
Prototype Approach
Prototype Formation
Prototypes are average representations of a category.
E.g., typical bird prototype based on those commonly seen (sparrows, robins).
Typicality Effects
High typicality indicates close resemblance to the prototype; low typicality indicates variance.
Prototypical objects are identified more rapidly in tasks such as naming and verification.
Family Resemblance Revisited
Prototypical objects will share more defining characteristics with other category members (e.g., chair vs. telephone).
The Exemplar Approach
Exemplar Definition
Unlike prototypes, which are the average, exemplars are actual instances encountered.
Better accounts for atypical instances (e.g., a penguin as a bird).
Combined Approach
Suggests people may utilize both prototypes and exemplars depending on context.
Hierarchical Organization of Categories
Basic Level Categories
Example levels: Superordinate (Furniture), Basic (Chair), Subordinate (Kitchen Chair).
Basic level categories are preferred due to balancing detail and generalization (e.g., more specific than superordinate, less detailed than subordinate).
Rosch's Research
People predominantly use basic categories in naming and understanding.
Knowledge is context-dependent, as seen in expertise contexts (e.g., ornithologists vs non-experts).
Network Models of Categorization
Semantic Networks
Categories represented in interconnected hierarchies of nodes and links.
Nodes denote concepts; links denote relationships.
Collin and Quillian's model: Reaction time increases with distance in the network.
Spreading Activation
Activation of one concept triggers related concepts in the network.
Support from priming studies indicates faster responses for closely associated items.
Connectionist Models
Connectionism
Concepts represented through distributed activation across a network.
Adjustments in connection weights based on learning and feedback (backpropagation).
Training Example
Networks adjust based on various presentations (e.g., birds like canaries or robins).
Brain Representation of Concepts
Four Approaches
Sensory-Functional Hypothesis: Knowledge of living things obtained through sensory attributes; artifacts primarily through function.
Multiple-Factor Approach: Distributed representation with multiple attributes beyond sensory or functional.
Semantic Category Approach: Specific neural circuits for innate categories crucial for survival.
Embodied Approach: Knowledge based on reactivation of sensory and motor processes upon interaction.
Hub and Spoke Model
Anterior Temporal Lobe (ATL) as a hub, integrating information from various specialized areas (spokes). Different functions demonstrate unique activation patterns in response to stimuli.
Conclusion
Categorization is a cognitive process with diverse approaches and models that help explain how we understand, learn about, and interact with the world around us. Each of these models contributes to a broader understanding of human conceptual knowledge.