This project examines the patterns of use of fun and funny in American and British English using a data-driven phraseological approach to corpus linguistics to show similarities and differences between their use and shades of meaning. It consists of a theoretical part, presenting and exploring the field of phraseology advocated by authors like Sinclair (1991), Hunston and Francis (2000), and Stubbs (2001), and discussing central terms and definitions. In the empirical part, this project thoroughly maps out the patterns of fun and funny using two large-scale, balanced corpora, The Corpus of Contemporary American English (COCA) and The British National Corpus (BNC). Both qualitative and quantitative research methods are used in exploring the corpora. The findings are also compared to a dictionary and a grammar. One of the main findings is that fun is a central adjective in COCA and a peripheral one in the BNC. This is reflected in differences in its collocations, colligations and other patterns of fun in the corpora. Funny is found to have a continuum of meanings, and it is also found to be used without evaluative force in introductions or ‘frames’ of the discourse. The difference between the use of fun and funny often resides in the patterns they prefer, and they are found to evaluate different types of words.