Our understanding of the world is communicated through language which helps shape the way we execute actions by helping us understand needs and address them. Often robots are built to address specific challenges in controlled environments. This leads to robots that cannot execute new tasks or fail to execute a known task in an unknown environment. Studying language and how it relates to actions like locomotion can help us extend the use of specialized robots to doing more tasks. It can also allow us to study the variance across languages with respect to action. These are just a few ideas that this research can contribute to. The main goal of this project is to use ML techniques to explore how cognitive boundaries shape the connection between language and representations of actions.