Linguistic experiential priors account for notation-dependent numerical representations

Area of Science:

• Cognitive Science
• Psycholinguistics
• Computational Linguistics

Background:

• The abstractness of number representation is a key debate in numerical cognition.
• Understanding how notation (e.g., digits vs. words) influences number representation is crucial.

Purpose of the Study:

• To investigate if linguistic usage patterns reflect number representation.
• To test if these linguistic patterns act as notation-dependent priors influencing behavior.
• To determine if number representation is abstract and notation-independent or experientially driven.

Main Methods:

• Utilized distributional semantic models on large text corpora to analyze number word and Arabic digit usage.
• Conducted computational experiments to identify distance effect-like relationships in linguistic data.
• Performed behavioral experiments using linguistic priors to predict performance in number comparison tasks.

Main Results:

• Linguistic distributional histories mirrored distance effects for both digits and number words.
• Notation-dependent linguistic priors better predicted behavioral performance than numerical distance alone.
• Evidence found for notation-dependent behaviors explained by linguistic priors.

Conclusions:

• The way numbers are used in language (distributional patterns) reflects how humans represent them.
• Findings support an experiential-driven model of number representation, challenging purely abstract or notation-independent views.
• Linguistic experience plays a significant role in shaping numerical cognition.