A Taxonomic Analysis of Abstraction

Area of Science:

• Cognitive Psychology
• Cognitive Science
• Psychology

Background:

• Knowledge representation is multifaceted, existing across various levels of abstraction.
• Existing theories on abstraction's cognitive impact lack a unifying framework.
• Understanding abstraction is crucial for explaining perception, comprehension, memory, and thought.

Purpose of the Study:

• To propose a unifying taxonomy for understanding abstraction in cognitive processes.
• To differentiate between abstract entities, attribute abstraction, and abstract ideas.
• To explore how concrete and abstract mental representations differ.

Main Methods:

• Developed a taxonomy based on three senses of abstraction: abstract entities, attribute focus, and category generalization.
• Applied the taxonomy to analyze mental representations of instances, attributes, and categories.
• Examined the interplay of categorization strategies with abstraction at different levels.

Main Results:

• Distinguished concrete representations (modal images, equivalent attributes, exemplars) from abstract ones (amodal propositions, distinctive attributes, rules/prototypes).
• Demonstrated how abstraction influences perception, comprehension, categorization, memory, and thought.
• Showcased the application of the proposed taxonomy to words, pictures, problems, and hierarchical levels.

Conclusions:

• The proposed taxonomy offers a common conceptual framework for diverse theories of abstraction.
• Abstraction operates distinctly at the entity, attribute, and category levels, impacting mental representations.
• Further research can extend this framework to explore hierarchical abstraction and its implications.