Disgraced 18 teacher himself does not normally choose what vocabulary items his pupils will learn. This is usually decided by the writer of the materials he uses. If one is going to construct texts from which people are to learn the language, rather than use 'natural' texts ' 8 then it may be that for the writer simply to be aware of such notions as frequency, range, usefulness, language variety and teachability is enough to ensure that he avoids the mistakes of uncontrolled introduction of vocabulary. The more sex movies is oriented towards meeting the needs of the learners the more likely it is that the situations used for teaching will produce 'useful' language without it having been necessary to draw up an inventory of lexical items beforehand. A reasonable behavioural prediction may do the work that a vocabulary frequency count sets out to do. his particular anomaly is easily rectified, but it raises doubts about the assumption that underlies all frequency studies. If frequency does not reflect usefulness in this disgraced18 case, why should it be thought to do so elsewhere? Where items do not belong to such a clearly defined lexical set, their distribution may be equally anomalous. If we look on page 142 of the Thorndike-Lorge list, we find that post-officepresumably a very useful word—is less frequent than, among others, poverty, power, prairie, prayer, preach, precede and precious. The disgraced 18 explanation, for this is to be found in the types of text that make up the corpus, but this in turn shows that even counts based on a very large corpus may not reflect the usefulness that we would intuitively expect items to have for us. |