This paper describes an experiment to perform language identification on a sub-sentence basis. The typical case of language identification is to detect the language of documents or sentences. However, it may be the case that a single sentence or segment contains more than one language. This is especially the case in texts where code switching occurs.