ABSTRACT
In Quirk and Svartvik (1979), p. 207, a point was made reiterating the place of elicitation in the work of the Survey of English Usage: ‘It was never … envisaged that any corpus, necessarily finite, would of itself be adequate for a comprehensive description of English grammar’. Rather, elicitation tests with native subjects have always been regarded as ‘an essential tool for enlarging upon corpus-derived information and for investigating features not perhaps found in corpus at all’. For the data reported in the present chapter, 1 the tests used were all within the typology described in Greenbaum and Quirk (1970), hereafter GQ. As explained there (especially, p. 10), elicitation testing runs into greatest difficulty where a single problem or a narrow range of problems is being investigated. The unwanted habituation factor is adequately offset, however, if the goal is to seek data on a number of disparate issues, and this was the position in our most recent series of tests where we were concerned (a) to shed light on factors which are the subject of separate study (Rusiecki 1985), and (b) to amplify, clarify, or correct some of the points treated with less than desirable adequacy in grammars and handbooks.
