Document Type


Publication Date



Computer Sciences


Noreen Herzfeld


Do you remember your first lesson in metaphor? Most likely, you learned that it is a comparison of one thing to another, to give a more descriptive account of something. "He was an iceberg" or "time is money" are obvious metaphors. They are regarded as simply vivid expressions to explain our ideas. However, George Lakoff, professor of linguistics at the University of California, Berkeley, and Mark Johnson, professor of philosophy at the University of Oregon, assert that metaphors are much more than literary devices. In their 1980 book, Metaphors We Live By, they claim that the way we think, act, and understand our world is rooted in metaphor. Their theory represents shift in the focus on metaphor, taking it from merely poetic substitution to a bold statement regarding human understanding. The theory, then, provides an important application to one of artificial intelligence's main facets: natural language processing. It provides a theory that encompasses broad understanding and context, and it can be computationally modeled, both important aspects to natural language processing. At the same time, a growing field fosters numerous opinions, and there are many fundamental issues about metaphor that remain unresolved. This thesis investigates these two very different sides to the partnership of metaphor and artificial intelligence.