This article examines the proposal made by Torres-Martínez and argues that to understand the patterns of modalilty in English requires lexical and construction information. It first attempts to clarify what embodied cognition is and how it can be applied to argument structure constructions and suggests that it may not be appropriate to assume this in the argument structure constructions of modal auxiliary verbs. Then, this article further argues that the notion of agency or force may not readily apply to all types of argument structure constructions with modals, particular those with epistemic modality. Apart from some syntactic issues, we propose that learning modal meanings or uses require lexical information such as that offered in Frame Semantics and constructional information.