Chapter 9: Knowledge and Category Representation

Vocabulary practice flashcards covering category learning, classical views, semantic networks, prototypes, exemplars, and the neurobiology of categorization.

Categorization

The cognitive process of identifying "a thing" as "something" and putting objects together into organized groups (Goldstone & Kersten, 2003).

Category Learning

The mental process of creating a mental representation for a specific category (Goldstone & Kersten, 2003).

Classical View (Bruner, Goodnow, & Austin, 1956)

A rule-based approach asserting that category representations are summary descriptions of an entire class defined by individually necessary and jointly sufficient features, with no disjunctive concepts.

Individually Necessary Features

A requirement of the classical view stating that every single member of a category must possess each specified feature to belong to that category.

Jointly Sufficient Features

A requirement of the classical view stating that possessing the entire specified set of features guarantees category membership without needing any additional traits.

Disjunctive Concepts

Categories defined by an "OR" rule rather than an "AND" rule (such as a baseball "strike" being "called" or "swinging"), which contradicts the classical view (Rosch et al., 1976).

Semantic Network View (Collins & Quillian, 1969)

A hierarchical model of semantic memory where nodes represent categories, links indicate relations (such as "is_a"), and properties are attached to specific nodes.

Cognitive Economy

The storage principle in hierarchical semantic networks where shared features are stored only once at the highest possible category node and inherited by subsets, minimizing redundant storage.

Overriding Properties

The semantic network mechanism that allows a lower-level subset node to store an exceptional property that supersedes an inherited general property (e.g., an ostrich "Can't fly" overriding a bird's "Can fly").

Sentence Verification Technique

An experimental paradigm where participants rapidly respond "Yes" or "No" to superset statements ("X is a Y") or property statements ("X has property P") to measure semantic retrieval times.

Superset Sentences ($S$)

Sentence verification trials testing category membership at varying hierarchical distances, such as $S_0$ ("A shark is a shark"), $S_1$ ("A shark is a fish"), and $S_2$ ("A shark is an animal").

Property Sentences ($P$)

Sentence verification trials testing feature possession at varying hierarchical levels, such as $P_0$ ("A shark can bite"), $P_1$ ("A shark can swim"), and $P_2$ ("A shark eats").

Spreading Activation

The process by which the activation of one semantic node automatically spreads along relational links toward associated nodes, explaining conceptual priming.

Lexical Decision Task (Meyer & Schvaneveldt, 1971)

An experiment requiring participants to judge whether a pair of letter strings contains two real words; reaction times were significantly faster for semantically associated pairs (e.g., "Bread"-"Wheat") than unassociated pairs (e.g., "Chair"-"Money").

Typicality Effect

The psychological finding that items judged as more representative of a category are verified and processed faster than less representative members (e.g., apple vs. pomegranate).

Superordinate Level

The highest, most general level in a hierarchical taxonomy (e.g., "Furniture" or "Vehicle"), characterized by the lowest number of listed attributes.

Basic Level

The psychologically privileged intermediate category level (e.g., "Chair" or "Car") providing the optimal amount of information, learned first by children and used preferentially in free-naming.

Subordinate Level

The lowest, most specific taxonomic category level (e.g., "Kitchen chair" or "Pickup"), which provides minimal additional attributes compared to the basic level and yields the slowest categorization times.

Second-Generation Semantic Networks (Collins & Loftus, 1975)

A modified semantic network approach abandoning rigid hierarchical structures in favor of psychological distance, where shorter links reflect closer associations based on personal experience.

Family Resemblances (Rosch & Mervis, 1975)

A categorization principle based on the number of overlapping characteristic features shared among category members rather than any single necessary attribute.

Prototype Model

A similarity-based model where categories are represented by an abstract summary representation derived by averaging previously encountered members (most frequent discrete traits or mathematical average of continuous traits).

Color Priming Paradigm (Rosch, 1975b)

An experiment demonstrating that hearing a category prime ("green") speeds verification of highly prototypical "good" color examples ($610\,\text{ms}$) substantially more than poor examples ($780\,\text{ms}$).

Exemplar Model

A similarity-based approach proposing that categories are represented by sets or subsets of all previously encountered concrete instances rather than by an abstract summary representation.

COVIS (Ashby et al., 1998)

A multiple-systems neuropsychological theory proposing two competing category-learning systems: an explicit hypothesis-testing system and an implicit procedural learning system.

SPEED Model

A neurobiological computational theory explaining how rule-guided categorizations transition from initial explicit processing to automaticity with extensive practice (Ashby et al., 2007; Kovacs et al., 2021).

Button-Switch Interference

A diagnostic test for automaticity where switching motor response keys causes high response time and accuracy costs after extensive verbal category practice, showing that automatic categorization relies on procedural motor mappings.

Rostrocaudal Cortical Organization (Hélie et al., 2010)

The neural recoding process where verbal rules initially learned via the caudate nucleus and ventrolateral prefrontal cortex are progressively recoded caudally to premotor cortex and dorsally across extensive practice.