Grounding the Lexical Semantics of Verbs in Visual Perception using Force Dynamics and Event Logic