Kimery R. Levering, Kenneth J. Kurtz
Concepts: Structure and Acquisition
- Concepts emerge from the discovery of fundamental similarities between category members. They are the building blocks of thought as they connect perception to memory and allow for reasoning about unknown properties.
- Some theorists assert that concepts are abstractions of experienced category members, either in the form of definitional rules for membership (classical view) or sets of commonalities or averages that hold in most cases (prototype approach).
- In contrast to abstraction, some theories assume that concepts are simply stored information about individual examples that have been associated over time with category labels (exemplar approach).
- The theory-driven view focuses on the role of concepts in explanation and considers them to be embedded in rich theoretical systems of knowledge that inform our determination of what things are above and beyond how similar features are to previous examples.
- Concept learning is most studied through a classification task in which examples are displayed and guesses followed by feedback result in learners developing knowledge of what differentiates between members of more than one category.
- Category learning tasks outside of the traditional classification task (observation, inference, use) often result in more robust knowledge of the internal structure of a concept.
- Learning concepts based largely on one’s own organization and in a self-directed way can result in better learning.
- While we can think of categories at many levels, there is evidence of a basic level (e.g., dog) that is favored over other levels (e.g., mammal or pit bull), perhaps because of its compromise between generality and specificity.
- In addition to taxonomic categories, categories can be created on the fly (ad hoc), created based on relevant tasks (goal-derived), or based on relationships between features (relational categories).