A geometric semantic model and Parts-of-Sense Inference annotation framework

Area of Science:

• Computational Linguistics
• Cognitive Science
• Semantics

Background:

• Traditional semantic models often overlook subtle meaning nuances.
• Cross-linguistic analysis requires robust tools for fine-grained semantic representation.

Purpose of the Study:

• Introduce a geometric semantic model for multidimensional meaning representation.
• Develop a novel annotation framework for detailed, cross-linguistic semantic analysis.
• Evaluate the applicability of the Parts-of-Sense Inference (POSI) framework across diverse languages.

Main Methods:

• Developed a geometric semantic model for fine-grained representations.
• Created a Parts-of-Sense Inference (POSI) annotation framework with 135 interpretable codes.
• Annotated semantic expressions in four typologically diverse languages.

Main Results:

• The POSI framework successfully annotates subtle semantic attributes.
• Demonstrated cross-linguistic and cross-structural applicability of the model.
• Achieved interpretable and cognitively plausible semantic representations.

Conclusions:

• The proposed geometric semantic model offers a robust tool for meaning inference.
• The POSI framework facilitates deeper investigation into language processing.
• This approach enhances understanding of semantic representation across languages.