Encyclopedia · Subject area
Concepts, frames and cascades
How do minds represent concepts, and how do word meanings combine? This subject area presents the main answers discussed in cognitive science, linguistics and philosophy — frames, cascades, conceptual spaces, prototypes, decomposition and atomism, formal and cognitive semantics — and the experiments that test them. It rests on the open-access volume Concepts, Frames and Cascades in Semantics, Cognition and Ontology (Springer, 2021) and relates its ideas to ontologies, knowledge graphs and structured documentation.
By knowledge.aitechdoc.world · Last reviewed
Articles
- FramesFrames: concepts as recursive attribute–value structures (Barsalou), their formal definition, uses in semantics and their relation to ontologies.Read the article
- Cascades and level-generationCascades: Goldman’s level-generation read as multilevel categorization, c-constitution, cascades as relations between frames, and their uses.Read the article
- Conceptual spacesConceptual spaces: concepts as regions in spaces of quality dimensions (Gärdenfors), convexity as naturalness, similarity as distance.Read the article
- Prototypes and typicalityPrototypes and typicality: graded membership, the modification effect, typical members as extremes, and frame-based category learning.Read the article
- Decomposition and atomismDecomposition vs. atomism: whether word meanings have structure, the analytic–synthetic problem, and frame-based decomposition.Read the article
- Formal and cognitive semanticsFormal vs. cognitive semantics: truth conditions and compositionality, the extensional turn, and bridges such as TTR record types.Read the article
- Cognitive structures in experimentExperimental tests of cognitive structures: reversal learning, social reinforcement in rats, action verbs and motor activity, and representational explanation.Read the article
Why concept theory matters for information work
Ontologies, knowledge graphs, taxonomies and structured documentation all make decisions about concepts: which classes exist, which properties a thing has, which properties are definitional and which are merely typical, and how one description of an event relates to another. Theories of concept representation study the same questions for the human mind. They do not prescribe how to build an ontology, but they explain why some models feel natural to readers, where definitions break down and why one occurrence can be described correctly on several levels.
The source work
The volume Concepts, Frames and Cascades in Semantics, Cognition and Ontology, edited by Sebastian Löbner, Thomas Gamerschlag, Tobias Kalenscher, Markus Schrenk and Henk Zeevat (Springer, 2021), collects twenty contributions from linguistics, philosophy and psychology. Many were presented at the conference “Cognitive Structures” (CoSt16) in Düsseldorf in 2016, organized by the Collaborative Research Centre 991 of the Deutsche Forschungsgemeinschaft. The book is open access under CC BY 4.0. The articles here summarize and compare its ideas in our own words and credit each chapter they draw on.
Comparison
| Type | What is modeled | Core structure | Key question | Relation to ontologies |
|---|---|---|---|---|
| Frames | The content of a concept | Central node, functional attributes, values, constraints | Which attributes does the concept have? | Close to classes with typed properties |
| Cascades and level-generation | One occurrence under several descriptions | Frames on levels linked by c-constitution | What does this act amount to, under these circumstances? | A relation between descriptions that standard vocabularies lack |
| Conceptual spaces | Similarity and value ranges | Quality dimensions, distance, convex regions | How similar are two instances? | Value ranges and measures behind properties |
| Prototypes and typicality | Typical members and defaults | Prototype, typicality gradient | How typical is this member? | Defaults and overriding in inheritance |
| Decomposition and atomism | Whether meanings have parts | Components or atoms plus inferences | What belongs to the meaning itself? | Definitional vs. typical properties |
| Formal and cognitive semantics | Truth conditions or concepts | Models, types, record types | When is it true — or what is represented? | Model-theoretic basis of description logics |
| Cognitive structures in experiment | Evidence for structures | Behavioral and EEG studies | Do the structures show in behavior? | None directly; limits of transfer |
How the domains interoperate
The introduction to the volume suggests that the notions are connected by integration or combination rather than divided. Cascades integrate several frames into one structure. The attributes of a frame take their values in conceptual spaces, and the regions a concept occupies there behave much like prototypes. Decomposition decides how much of a concept a frame should spell out; atomism moves part of it into inference. Formal semantics supplies compositionality and, through type theory, a formal home for frames. The experiments test whether such structures guide behavior.
Frames and cascades
A frame describes one level; a cascade relates frames on several levels. In practice the frame answers “what is this, with which properties?” and the cascade “what does it also amount to, in this context?”.
Frames, spaces and prototypes
Frames give the attributes, conceptual spaces their geometry, prototypes the defaults. Together they explain structured, graded and context-dependent categorization.
Relation to ontologies and knowledge graphs
Knowledge graphs and information models use classes, properties and constraints much as frames use nodes and attributes, and inherit the model-theoretic semantics of formal semantics through RDF and description logics. Three ideas from concept theory are less common in such models: the distinction between definitional and typical properties (decomposition and prototypes), value ranges with a similarity measure (conceptual spaces) and the relation between descriptions of one occurrence on different levels (cascades). They are relevant wherever a model has to evolve — when terms shift in meaning, when new instances do not fit the defaults, or when the same event matters differently to different roles. None of these ideas replaces an agreed vocabulary or its governance.
Source work and licence
Sebastian Löbner, Thomas Gamerschlag, Tobias Kalenscher, Markus Schrenk, Henk Zeevat (eds.), Concepts, Frames and Cascades in Semantics, Cognition and Ontology. Language, Cognition, and Mind, vol. 7. Springer (Springer Nature Switzerland AG), 2021. ISBN 978-3-030-50200-3. doi.org/10.1007/978-3-030-50200-3
Open access under the CC BY 4.0 licence. The articles of this subject area summarize, compare and recontextualize the volume’s ideas in our own words and structure; no text, figures or tables are reproduced. Positions are attributed to the chapter authors; the comparisons and the relation to ontologies and documentation are ours.