Abstract
This paper shows how tensor-based distributional semantics can be incorporated into Minimalist Grammars (MGs). We embed the MG feature calculus with a tensor algebra and give a joint tensor-based representation where compositional semantics is guided by the minimalist syntax. By bridging syntactic and semantic operations in tensor spaces, we aim to contribute to the broader enterprise of neurosymbolic approaches to linguistic cognition.
Type
Publication
Society for Computation in Lingistics
