to him are contained in manuscripts executed in the eleventh century, we have in this circumstance an excellent proof that the attribution is ill-founded: no document of which there exists a copy in eleventh-century handwriting can be posterior in date to the eleventh century. Then we examine the language. It is known that certain forms have only been used in certain places and at certain dates. Most forgers have betrayed themselves by ignorance of facts of this kind; they let slip modern words or phrases. It has been possible to establish the fact that certain Phœnician inscriptions, found in South America, were earlier than a certain German dissertation on a point of Phœnician syntax. In the case of official instruments we examine the formulæ. If a document which purports to be a Merovingian charter does not exhibit the ordinary formulæ of genuine Merovingian charters it must be spurious. Lastly, we note all the positive data which occur in the document—the facts which are mentioned or alluded to. When these facts are otherwise known, from sources which a forger could not have had at his disposal, the bonâ fides of the document is established, and the date fixed approximately between the most recent event of which the author shows knowledge, and the next following event which he does not mention but would have done if he had known of it. Arguments may also be founded on the circumstance that particular facts are mentioned with approval, or particular opinions expressed, and help us to make a conjectural estimate of the status, the environment, and the character of the author.
When the internal analysis of a document is carefully performed, it generally gives us a tolerably accurate notion of its