Multa renascentur quae iam cecidere cadentque quae nunc sunt in honore vocabula si volet usus quem penes arbitrium est et vis et norma loquendi. (Horace, Ars Poetica) Many a word yfalne shall eft arise And such as now bene held in hiest prise Will fall as fast, when vse and custome will Onely vmpiers of speach, for force and skill. (Puttenham's translation, 1589) 1. Introduction All through the history of dictionary-making and reflections on standard languages, usage has been one of the lexicographers' major stumbling-blocks. (1) Although it can be argued that the question of a linguistic norm is even more relevant in the fields of phonology and syntax, it is eminently important for vocabulary, too. To indicate the currency, degrees of acceptability and the geographical, social and stylistic restrictions of individual words is a problem common to both more prescriptive and more descriptive dictionaries. Johnson's aim to establish a standard for the English lexicon left but a few words unaffected in his great dictionary of 1755 and these had usage labels attached to such entries to warn readers against using them; modem dictionary editors sometimes use panels of experts to decide on the acceptability of words and their individual meanings. Whatever method for establishing usage is preferred, whether that of the editor's Sprachgefuhl or the combined judgment of experts (a method which appears to permit neatly graded degrees of acceptability), any indication of usage carries with it idiosyncratic elements as well as the restriction to a particular period, region and class-related opinion. It would be silly to complain about this, since it is the very essence of usage, that the labels used to indicate it have a certain degree of vagueness about them. The narrower the description of correct usage is, the more certain it is that the evidence will be homogeneous, and the more likely that the guidance will be taken as prescriptive. In a dictionary which has 'usage' as the first word of its title, readers will rightly expect this type of information to be central. The UDASEL project was in fact started with the explicit intention of concentrating on the fully accepted loanwords from English in 16 European languages, (2) rather than focussing on more narrowly etymological questions of word origins. As will be obvious (or become evident from my discussion), the difficulty of describing the usage of specific items in an individual language multiply when 16 sets of data are compared (which makes possible up to 240 binary comparisons). The basic decision as far as the currency of loanwords from English is concerned was not to rely on text corpora, for various reasons: a) such collections from recent texts were not available for most of the languages concerned; b) where such corpora existed, their representativeness and comparability was open to doubt; c) the frequencies elicited from them would not really be indicators of acceptability: for this the context and context would have to be investigated for each individual item, and judgments would therefore in any case be based, at least partially, on the interpreter's Sprachgefuhl. All this made it necessary, if the evidence was to be covered in a 'snapshot' manner for the early 1990s, to rely on the competence of educated users of the individual language, and leave judgments on the acceptability of English words to the specialist. This hard-won decision is supported by the fact that pilot studies with German informants (Gorlach 1994) have shown that speakers' views on degrees of acceptability conform to a greater extent than might have been expected, or feared. (3) 2. Degrees of acceptability After the pilot study was completed, the grading decided on was based on a combination of degrees of integration and of acceptability/currency: these two aspects are not always compatible, but coincide in most cases, and they were combined in order to prevent the system from becoming too complex for a general reader to take in and remember. …