Semantic network
Knowledge stored as a web of concepts linked by meaningful associations.
What it means
A semantic network is a model of how conceptual knowledge is organized in long-term memory: concepts are represented as nodes connected by labeled links that encode relationships such as category membership, properties, and associations, with related ideas closer together in the web. Early hierarchical versions stored each fact at the most economical level and predicted that verifying 'a canary can fly' should be faster than 'a canary can breathe' because the property is stored nearer the node; later spreading-activation versions relaxed the strict hierarchy, spacing concepts by associative strength and typicality to better fit human reaction times. Retrieval is modeled as activation spreading outward from an accessed node along its links, automatically priming nearby concepts, which accounts for semantic priming, the tip-of-the-tongue state, and patterns of intrusion and confusion. The framework underlies much of cognitive psychology's theory of meaning and influenced knowledge representation in artificial intelligence. It matters because it explains how meaning is structured, why related ideas come to mind together, and how that structure shapes search, inference, and creativity.
Examples
Asked to verify 'a robin is a bird,' people respond faster than for 'an ostrich is a bird,' because the typical robin sits closer to the bird node in the network.
Stuck on a crossword, you can feel that the answer starts with an S and involves sailing — activation is spreading near the right node without quite reaching the word.
Read the word 'doctor' and you recognise 'nurse' a fraction faster than 'butter' — activation has already spread outward to whatever sits closest in the web.
First described in Collins & Quillian (1969); Collins & Loftus (1975).