The SEE project is also using natural language processing to develop a more refined linguistic understanding of how humans think about causation. The existing model only uses the word