The attempt to locate and remove interpolations – non-authorial matter that has made its way into a manuscript tradition – is essentially a subdivision of conjecture. To bracket as interpolated words that are transmitted in all manuscripts is on a par with printing a reading that is the product of editorial conjecture in place of the reading(s) of all manuscripts. For several reasons, though, the pursuit of interpolation has long been an especially contentious area of critical practice. For one thing, it involves a drastic intervention on the critic's part (especially if the alleged interpolation consists of whole lines of text), thereby raising in an acute form the tension between respecting the transmitted text and subjecting it to careful scrutiny. Even critics who accept the need for conjecture sometimes baulk at efforts to diagnose interpolation, all the more so when the text in question enjoys a privileged place in the canon. If, as I suggested in an earlier chapter, critical choices are justified to the degree that they can be made convincing, the decision to excise material transmitted in all manuscripts creates a particularly heavy burden of persuasion for a critic.

Another reason why the study of interpolation has been a site of dispute is that it has often proven difficult to practice in moderation, and its excesses have fuelled resistance to the enterprise as a whole. Quickness to suspect interpolation was a distinctive feature of the hyperscepticism that animated much textual criticism in the second half of the nineteenth century – the age of August Nauck's Euripides (see p. 34) – and an equally extreme aversion to interpolation was typical of the conservative reaction that followed.