The second part of the functionality requirement deals with what are commonly called false positives. Work on specific methods relating to this is still in progress. It is clear at this stage that it will involve the re-use (double interpretation) of test material developed for requirement part I, i.e. sentences containing error forms but also the corrected versions. Additionally, we anticipate the use of larger amounts of naturally occurring text to identify unexpected errors. Finally, it is anticipated that some use will be made of the technique of error-spotting (designing `traps' for foreseen weaknesses of systems).