Sometimes interpolations and continuations can be very readily distinguished in the course of the operations for restoring a text of which there are several copies, when it so happens that some of these copies reproduce the primitive text as it was before any addition was made to it. But if all the copies are founded on previous copies which already contained the interpolations or continuations, recourse must be had to internal analysis. Is the style uniform throughout the document? Does the book breathe one and the same spirit from cover to cover? Are there no contradictions, no gaps in the sequence of ideas? In practice, when the continuators or interpolators have been men of well-marked personality and decided views, analysis will separate the original from the additions as cleanly as a pair of scissors. When the whole is written in a level, colourless style, the lines of division are not so easy to see; it is then better to confess the fact than to multiply hypotheses. III. The critical investigation of authorship is not finished as soon as a document has been accurately or approximately localised in space and time, and as much information as possible obtained about the author or authors.[85] Here is a book: we wish to ascertain the origin of the information contained in it, that is, to be in a position to appreciate its value; is it enough to know that it was written in 1890, at Paris, by So-and-so? Perhaps So-and-so copied slavishly, without mentioning the fact, an earlier work, written in 1850. The responsible guarantor of the borrowed parts is not So-and-so, but the author of 1850. Plagiarism, it is true, is now rare, forbidden by the law, and considered dishonourable; formerly it was common, tolerated, and unpunished. Many historical documents, with every appearance of originality, are nothing but unavowed repetitions of earlier documents, and historians occasionally experience, in this connection, remarkable disillusions. Certain passages in Eginhard, a ninth-century chronicler, are borrowed from Suetonius: they have nothing to do with the history of the ninth century; how if the fact had not been discovered? An event is attested three times, by three chroniclers; but these three attestations, which agree so admirably, are really only one if it is ascertained that two of the three chroniclers copied the third, or that the three parallel accounts have been drawn from one and the same source. Pontifical letters and Imperial charters of the middle ages contain eloquent passages which must not be taken seriously; they are part of the official style, and were copied word for word from chancery formularies. It belongs to the investigation of authorship to discover, as far as possible, the _sources_ utilised by the authors of documents.