during more than fourteen centuries down to the invention of printing and its application to Greek literature.
7. Every transcription of any kind of writing involves the chance of the introduction of some errors: and even if the transcript is revised by comparison with its exemplar or immediate original, there is no absolute security that all the errors will be corrected. When the transcript becomes itself the parent of other copies, one or more, its errors are for the most part reproduced. Those only are likely to be removed which at once strike the eye of a transcriber as mere blunders destructive of sense, and even in these cases he will often go astray in making what seems to him the obvious correction. In addition to inherited deviations from the original, each fresh transcript is liable to contain fresh errors, to be transmitted in like manner to its own descendants.
8. The nature and amount of the corruption of text thus generated and propagated depends to a great extent on the peculiarities of the book itself, the estimation in which it is held, and the uses to which it is applied. The rate cannot always be uniform: the professional training of scribes can rarely obliterate individual differences of accuracy and conscientiousness, and moreover the current standard of exactness will vary at different times and places and in different grades of cultivation. The number of transcriptions, and consequent opportunities of corruption, cannot be accurately measured by difference of date, for at any date a transcript might be made either from a contemporary manuscript or from one written any number of centuries before. But these inequalities do not render it less true that repeated transcription involves multiplication of error; and the consequent presumption that a relatively late text is likely to be a relatively corrupt text
