This paper addresses the question of whether it is possible tosense-tag systematically, and on a large scale, and how we shouldassess progress so far. That is to say, how to attach each occurrenceof a word in a text to one and only one sense in a dictionary – aparticular dictionary of course, and that is part of the problem. Thepaper does not propose a solution to the question, though we havereported empirical findings elsewhere (Cowie et al., 1992;Wilks et al., 1996; Wilks and Stevenson, 1997), and intend to continue andrefine that work. The point of this paper is to examine two well-knowncontributions critically: The first (Kilgarriff, 1993), which is widelytaken to show that the task, as defined, cannot be carried outsystematically by humans and, secondly (Yarowsky, 1995), which claimsstrikingly good results at doing exactly that.