This paper addresses the issue of systematic and large-scale sense-tagging, i.e. the problem of attaching each occurrence of a word in a text to one and only one sense in a particular dictionary. The paper reports empirical findings in the field and specifically examines two well-known contributions critically. Kilgarriff (1993) shows that the task of sense-tagging cannot be carried out systematically by humans; Yarowski (1995) claims good results at systematic sense-tagging.