Research
Words that make SENSE: Sensorimotor Norms in Learned Lexical Token Representations
arXiv:2602.00469v2 Announce Type: replace-cross Abstract: While word embeddings derive meaning from co-occurrence patterns, human language understanding is grounded in sensory and motor experience. We
arXiv:2602.00469v2 Announce Type: replace-cross Abstract: While word embeddings derive meaning from co-occurrence patterns, human language understanding is grounded in sensory and motor experience. We present ext{SENSE} (extbf{S}ext{ensorimotor } extbf{E}ext{mbedding } extbf{N}ext{orm } extbf{S}ext{coring } extbf{E}ext{ngine}), a learned projection model that predicts Lancaster sensorimotor norms from word lexical embeddings. We also conducted a behavioral study where 281 participants selected which among candidate nonce words evoked specific sensorimotor associations, finding statistically significant correlations between human selection rates and ext{SENSE} ratings across 6 of the 11 modalities. Sublexical analysis of these nonce words selection rates revealed systematic phonosthemic patterns for the interoceptive norm, suggesting a path towards computationally proposing candidate phonosthemes from text data.
Related
- LLMs Struggle with Abstract Meaning Comprehension More Than Expected
- Do BERT Embeddings Encode Narrative Dimensions? A Token-Level Probing Analysis of Time, Space, Causality, and Character in Fiction
- Investigating the structure of emotions by analyzing similarity and association of emotion words
- Between Century and Poet: Graph-Based Lexical Semantic Change in Persian Poetry
Source: arXiv cs.AI | 2026-04-24