Blurb:

Although raw text corpora have been felt to be "inadequate for many of the questions Systemic Functional Linguistics asks" (Honnibal 2004: 7), annotation also creates problems, not only for its costs in terms of time and money (Wu 2009: 142), but also due to technical and philosophical issues. This paper will look at issues surrounding automatic versus manual annotation, and whether SFL corpus annotation is worthwhile. The results of a test on two taggers (UAM Corpus Tool and Halliday Center Tagger) will then be discussed.