CRANV2 Introduction chapter Cyril Cleverdon Michael Keen Cranfield An investigation supported by a grant to Aslib by the National Science Foundation. Use, reproduction, or publication, in whole or in part, is permitted for any purpose of the United States Government. -! CHAPTER 1 Int rodu ct ion H two scientists disagree on any issue, and the issue is within the ambit of science, then it must be possibIe for them to agree on a procedure which they can both accept as a critical test of their points of difference. For reasons of personality they may not be able to get together to work out such a test procedure, but it must exist as a possibility. If such a critical test cannot be imagined as possible, then the issue between them is not a scientific issue. The scientific method does not vary with the subject-matter, but is the same irrespective of its results and basically the same in all the sciences. L.T. Wilkins: Social Deviance, page 4. In reviewing the final report on Cranfield I, N.D. Stevens (Ref. 1) described it as being 'extremely complex'; even after 'several careful readings' he found parts of it 'still bewildering', and said that 'there are so many side issues that the author neglects the clear and detailed presentation of the main headings; the reader finds himself sidetracked by these, or other interesting diversions'. Since this reviewer was by no means the only person who made such comments, it has been our particular endeavour in this report to make quite clear what has been done, how it has been done and what has been the outcome, even though at times this has lead to what some people may consider undue verbosity and repetition. In one respect, this project is easier to report, for, being in a more concentrated field, it does not raise many of the side-issues - such as indexing times, indexer qualifications, etc. which came up in Cranfield I and which were sufficiently interesting to sidetrack the reader, On the other hand, to those who have been involved in Cranfield II, the earlier project seems to have been child's play to what has now been attempted, and the complexity of the present work is inevitably reflected in what has to be reported. To those readers who, like ourselves, tend to view with dismay the many papers on information retrieval which consist substantially of some twenty pages of mathematics, we can only apologise that it has become necessary to introduce a number of equations into this volume. However, it is certain that there is no mathematics