Two principal kinds of emendation are normally applied to the books of the ancients, to free them from the many and varied flaws and stains that they have contracted on their long journey into barbarism, and restore them more nearly to their ancient and original form. The one entails more effort and, I might almost say, misery; the other, more leisurely delight. Each, if rightly applied, is useful; but one is more useful. Take someone, even someone poorly equipped with the best aids, who gives us a writer restored to a more correct form, either by conjecture or by the use of a few manuscripts; even if he removes just thirty warts, and leaves a hundred, no one will deny that he has rendered service to literature. And this used to be the way of things, especially in the days when manuscripts had not long begun to be printed, and it was widely expected that new aids would soon appear. Many have imitated this custom from then on, down to our own time, even for those writers who abounded with critical evidence and aids of all sorts. In fact, few authors of new recensions are so diligent and willing to work that they collect all the variant readings from what are often obscure and scattered sources, and especially from old exemplars, and then compare these with the standard text so that they can set about a consistent emendation of it. On the contrary, they generally stop short only when a difficult thought or an error obvious at first sight presents itself, and then they consult variant readings or an old exemplar. But these oracles are usually unresponsive except to those who consult them regularly. A similar method is applied by those scholars — sometimes very learned and expert — who hold that every emendation should depend on the credibility of a selection of manuscripts, or who edit texts, as they are called, in accordance with one exemplar, as if those manuscripts had been destined by the fates to save their author.

A true, continuous, and systematic recension differs greatly from this frivolous and desultory method. In the latter we want only to cure indiscriminately the wounds that are conspicuous or are revealed by some manuscript or other. We pass over more [readings] which are good and passable as regards sense, but no better than the worst as regards authority. But a true recension, attended by the full complement of useful instruments, seeks out the author's true handiwork at every point. It examines in order the witnesses for every reading, not only for those that are suspect. It changes, only for the most serious reasons, readings that all of these approve. It accepts, only when they are supported by witnesses, others that are worthy in themselves of the author and accurate and elegant in their form. Not uncommonly, then, when the witnesses require it, a true recension replaces attractive readings with less attractive ones. It takes off bandages and lays bare the sores. Finally, it cures not only manifest ills, as bad doctors do, but hidden ones too. This method certainly has a place for natural talent and the art of conjecture, but as the credibility of every ancient text rests entirely on the purity of its sources, we must strive above all — and can hardly do so without talent — to examine the properties and individual nature of the sources from which each writer's text is derived; to judge each of the various witnesses, once they are set out by classes and families, by its character; and to learn to follow their voices, and gestures, so to speak, with cunning, but without bias. Indeed, in many cases both the critic, and anyone who would undertake a historical investigation, must emulate the prudent custom of a good judge, who slowly examines the testimony of the witnesses, and gathers all the evidence for their truthfulness, before he ventures to put forward his own conjecture about the case. And indeed, it is impossible that one who relies on a few codices of the common sort and practices conjecture, however cleverly, can often arrive at the genuine text. In resolving questions of law, no amount of talent can make up for a want of wills and documents. Similarly, the acutest talent labors in vain on historical and critical questions, unless it is tempered and controlled by diligent use of manuscripts. In fact these two activities differ more in name than in kind, and they are bound by the same rules of judgment. It is proper that you should attach more weight to talent than to costly parchments. But even genius badly needs to have access to as many codices as possible, so that its judgment about the true reading may rely on their testimony and its divination may find aid of many kinds. The more often, then, that the manuscripts of any author undergo collation, the more a true and consistent recension takes place. For those authors who still suffer from the lack of useful materials, I am inclined to think that the multiplication of forms of their texts harms rather than helps these studies, which are more than tedious enough in themselves. Such authors can presumably be given a sort of review (recognitio), not what is truly called a recension (recensio).

But in general no one disputes any of these points. In Homer, however, the oldest poet, doubts clearly exist as to whether so much weight should be given to the authority of such recent manuscripts. For none of them is even so old as the latest Latin writers. Those that date before the twelfth or eleventh century are few and far between. This doubt may carry the implication that these sources cannot enable us to restore Homer's work to the genuine, pure form which first poured from his divine lips. If so, I shall say later how willingly I follow this school of thought and line of reasoning. But when I consider the fates of ancient books, sometimes unexpectedly favorable, and see that we have Herodotus, Plato, Xenophon, others of the same period, whom we have received from the hands of the same scribes, restored so nearly to their original luster — then I can find no reason, unless perhaps the texts of the ancient bards that have come down to us are worse than the rest, why we should trust the latter so much less than the former. Moreover, thanks to the Alexandrian critics, who flourished after [Herodotus, Plato, and Xenophon], we doubtless read a Homeric text more correct in many passages than the one that they themselves could read. Finally, it is surprising how many of the variant readings found in authors who quote Homer — variants of any significance, and which are not simply errors resulting from a faulty recollection [of the passage] — are found in almost identical form in those manuscripts. For newness in manuscripts is no more a vice than youth in men. In this case, too, age does not always bring wisdom. Insofar as each follows an old and good authority well, it is a good witness. True, I attach somewhat more weight to the apparatus...