could have been obtained in carrying out a research that, even if it does not reveal anything new, at least confirms what is already known, and makes the researcher realize what he knew before only in a vague way, and on the evidence of others. We all know in a general way that Milton uses a higher percentage of Latinized words than does Defoe. But if we take the trouble to select two thousand words consecutively from any part of the Areopagitica, and two thousand words consecutively from any part of Robinson Crusoe, and classify the words as (1) Latinized, (2) Saxon, and (3) those neither Latin nor Saxon in origin, we get a quantitative result from which we can say with greater exactness how the two vocabularies stand to one another. Results of this kind very often surprise the investigators: nearly always they suggest facts that had not before been suspected.
A particularly useful exercise at the early stages of your practice in research is the verification or testing of results obtained by others. The advantage of this exercise is that you have a sort of standard by which to judge whether you are keeping fairly near the truth. If your results are widely different from those of your predecessor, you have the alluring hunt for the big error you have made, with, of course, just the delightful possibility that the error was made by the other fellow.
[ 263 ]
