Metaphors abound in human language. We are constantly using one kind of thing to understand another – we talk easily about “political landslides” or being “boiling mad” or having “come to a crossroads” in our lives. For computers to understand this, they must understand metaphors, yet this entails knowing not just language, but a great deal about the world and the culture that we live in. Argamon and his team are working on developing a system to identify and explain metaphors by analyzing enormous amounts of textual data to find statistical patterns to help us connect metaphorical and non-metaphorical usages, and so enable us to understand the meaning of those metaphors. The methods they are developing are not specific to any one language or culture, so by applying them to various languages and cultures they can uncover the similarities and differences between how different cultures understand the world we live in.