Editions of any sort are integral to that trans-historical contact that underpins modern study. The human sciences depend upon documentary access, and on a general belief that our access to such documents is accurate. The model for producing such a formalised presentation of a text derives from the reproduction of historical documents: one first finds, and then accurately transcribes what remains, most typically a unique record (for example, a charter conveying land to a monastery). This renders this portion of the archive available for those interpretative acts which render historical study possible.

What of texts in more than one copy, which includes the archive accessed by those in literary studies across a range of languages, as well as philosophy, theology, law, and other disciplines? The difficulty might be illustrated by the medieval history of that most central cultural text, the Latin Bible. To put matters crassly, its source was believed unimpeachable, the text a revelation from God herself. But copies of the Bible might vary widely, because, by definition, they are the product of the errant human agents who had been responsible for promulgating the text. The standard text, called the ‘Vulgate Bible’, had been produced by St Jerome in Palestine at the turn of the fourth and fifth centuries. However, the majority of circulating copies, and the usual text, was that promulgated from Paris during the thirteenth century, and in the intervening 800 years, widespread (and well-recognised) variation had entered the text. How was one to find what the author might have intended? Or to correct whatever version one might have received?

Here in the pre-print era there was a particular difficulty. In general, print produces identical copies of the same (although there are always matters of detail). Since the coming of the press, published editions have typically followed a line of least resistance. When a text is reprinted, it customarily follows that in an earlier edition, and print texts most usually offer a vertical succession of textual forms – the later ones typically introducing variations not the property of earlier ones. As a result, finding what the author had intended, editing these, is a relatively simple procedure; one discovers the latest edition in which the author can be said to have been involved, and follows it, unless certainly wrong (e.g., an obvious typographical error).