Rethinking the Idiomaticity Decomposability Hypothesis: Evidence from Distributional Learning
arXiv:2606.03817v1 Announce Type: new Abstract: Idioms can be analysed in terms of their decomposability, the extent to which constituent meanings contribute to the figurative whole. Decomposability is thought to predict syntactic flexibility. Usage-based accounts instead attribute idiom behaviour to distributional experience, such as speaker familiarity and predictability. We examine these views using contextualised language models as controlled distributional learners. We propose a model-internal measure of decomposability and relate it to human ratings, syntactic flexibility, and predictabi
The proliferation of sophisticated contextualized language models allows for empirical testing of long-standing linguistic hypotheses about idiom structure and learning.
Understanding how AI models learn and process complex linguistic phenomena like idioms provides insights into their cognitive capabilities and potential for more nuanced language generation and understanding.
This research provides a refined theoretical framework for understanding idiom behavior, moving beyond simple decomposability to include distributional learning.
- · AI researchers
- · NLP developers
- · Computational linguists
Improved understanding of language model internal representations of semantic and pragmatic meaning.
Development of more robust and less error-prone AI systems for text generation and translation, particularly with idiomatic expressions.
Enhanced ability to model human language acquisition and cognitive processes related to idiom understanding.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL