Dancing with Deer: A Constructional Perspective on MWEs in the Era of LLMs

TL;DR

This paper advocates for a construction grammar approach to understanding multiword expressions, demonstrating how both humans and large language models can generalize novel expressions, but only humans can reason over combined expressions due to richer cognitive representations.

Contribution

It introduces a constructional perspective on MWEs, provides case studies in English and Arapaho, and compares human and LLM capabilities in learning and reasoning about MWEs.

Findings

Both models and humans can generalize from a single exposure to novel MWEs.

Humans can reason over combined MWEs, unlike LLMs.

Constructional templates effectively represent MWEs across languages.

Abstract

In this chapter, we argue for the benefits of understanding multiword expressions from the perspective of usage-based, construction grammar approaches. We begin with a historical overview of how construction grammar was developed in order to account for idiomatic expressions using the same grammatical machinery as the non-idiomatic structures of language. We cover a comprehensive description of constructions, which are pairings of meaning with form of any size (morpheme, word, phrase), as well as how constructional approaches treat the acquisition and generalization of constructions. We describe a successful case study leveraging constructional templates for representing multiword expressions in English PropBank. Because constructions can be at any level or unit of form, we then illustrate the benefit of a constructional representation of multi-meaningful morphosyntactic unit constructions in Arapaho, a highly polysynthetic and agglutinating language. We include a second case study leveraging constructional templates for representing these multi-morphemic expressions in Uniform Meaning Representation. Finally, we demonstrate the similarities and differences between a usage-based explanation of a speaker learning a novel multiword expression, such as "dancing with deer," and that of a large language model. We present experiments showing that both models and speakers can generalize the meaning of novel multiword expressions based on a single exposure of usage. However, only speakers can reason over the combination of two such expressions, as this requires comparison of the novel forms to a speaker's lifetime of stored constructional exemplars, which are rich with cross-modal details.