Page:The Library, volume 5, series 3.djvu/214

 202 PROBLEMS OF THE ENGLISH been taken of the intrinsic merit of the readings. If the scheme proposed is correft, or anything like correft, then it is futile to argue that a reading found say in D only, or even in D W K, if opposed by a concensus of P H B, is original, however plausible it may be. It may be ' corredt,' it may be what the author originally wrote, but, apart from con- flation, it can only have got into its place by emen- dation, it cannot have been the reading of the archetype. Where the relation of the manuscripts is uncertain we have, of course, to rely upon the general plausibility of their readings for determining their value, but it is always a precarious test. There are, it is true, in different cases various degrees of improbability in supposing a scribe or corrector to have been responsible for a given reading, and sometimes the improbability is very great, but, logically at least, we are never justified, however strongly we may feel that a certain reading is original, in assuming that it must have been that of the archetype. I may here state the result of a critical considera- tion of the differences between P and ?. In a number of cases P is certainly corrupt, but in most of these the reading of ? seems as likely to be an emendation as to be original. There remain, how- ever, two or three in which emendation appears very improbable, and which, therefore, confirm our previous assumption that P is not the parent of the cyclic group. There is one warning which should perhaps be borne in mind. Our results stridly apply to the Antichrist play alone. Since we decided that the