The role played by language in our cognitive lives is a topic at the centre of contemporary debates in cognitive (neuro)science. In this paper we illustrate and compare two theories that offer embodied explanations of this role: the WAT (words as social tools) and the LENS (language is an embodied neuroenhancement and scaffold) theories. WAT and LENS differ from other current proposals, because they connect the impact of the neurologically realized language system on our cognition to the ways in which language shapes our interaction with the physical and social environment. Examining these theories together, their tenets and supporting evidence, sharpens our understanding of each, but also contributes to a better understanding of the contribution that language might make to the acquisition, representation and use of abstract concepts. Here we focus on how language provides a source of inner grounding, especially metacognition and inner speech, and supports the flexibility of our thought. Overall, the paper outlines a promising research program focused on the importance of language to abstract concepts within the context of a flexible, multimodal, and multilevel conception of embodied cognition.