A Linguistic-Sensorimotor Model of the Basic-Level Advantage in Category Verification

Area of Science:

• Cognitive Science
• Computational Linguistics
• Neuroscience

Background:

• Humans excel at basic-level object categorization (e.g., dog) over superordinate (e.g., animal).
• Existing theories propose linguistic-distributional and sensorimotor relationships underlie this basic-level advantage.
• These proposed mechanisms lacked formal computational validation.

Purpose of the Study:

• To develop and test a computational model of category verification.
• To investigate the interaction of linguistic distributional information and sensorimotor experience in conceptual systems.
• To determine if the model replicates the human basic-level advantage.

Main Methods:

• Developed a computational model simulating a full-size adult conceptual system.
• Integrated linguistic distributional data and sensorimotor experience within the model.
• Conducted simulations across multiple datasets to test category verification performance.

Main Results:

• The computational model achieved human-comparable accuracy in category verification tasks.
• The model's operation inherently produced the basic-level advantage phenomenon.
• Higher overlap in linguistic and sensorimotor information between concepts correlated with easier categorization.

Conclusions:

• Findings support the linguistic-sensorimotor preparation account of the basic-level advantage.
• The study validates the role of integrated linguistic and sensorimotor information in conceptual systems.
• The model provides a computational framework for understanding human categorization behavior.