This book explores a virtually untapped, yet fascinating research area: television dialogue. It reports on a study comparing the language of the American situation comedy Friends to natural conversation. Transcripts of the television show and the American English conversation portion of the Longman Grammar Corpus provide the data for this corpus-based investigation, which combines Douglas Biber's multidimensional methodology with a frequency-based analysis of close to 100 linguistic features. As a natural offshoot of the research design, this study offers a comprehensive description of the most common linguistic features characterizing natural conversation. Illustrated with numerous dialogue extracts from Friends and conversation, topics such as vague, emotional, and informal language are discussed. This book will be an important resource not only for researchers and students specializing in discourse analysis, register variation, and corpus linguistics, but also anyone interested in conversational language and television dialogue.