Sicut dicit are the first two words in any number of medieval scholarly works. The third is invariably the name of some authoritative writer – Aristoteles, Tullius, Boethius etc. Then follows, optionally, in Topicis or the like and, finally, what the auctor said. The ubiquity of this incipit testifies to the fundamental role of authoritative writings in medieval culture, and many of the works that start Sicut dicit are commentaries on an authoritative book, though usually not the one referred to in the incipit.
Commentaries come in many guises. Some are expositiones that divide the text under consideration into parts, the content of each of which is then analysed and explained, the explanation often involving paraphrase and sometimes discussions of problematic issues (dubia). Others, so-called question commentaries, consist entirely of discussions of selected problems. And then there are the scholia, the unstructured or loosely structured heaps of notes that accompany authoritative texts in manuscripts.
Each of these genres of exegetic works presents its own problems for editors, and this volume's four essays about Greek texts and eight about Latin ones vividly illustrate just how many challenges such works pose to their editors.
A fundamental problem is that commentators rarely start from scratch. Usually they build on predecessors and include excerpts from them in their work, often verbatim and mostly without indicating the source. Once completed, a commentary may be revised by the author, and others may not only use it as a quarry but also revise it, the result being texts with a fluid identity. Sicut dicit primarily deals with texts suffering from an identity crisis and the question how an editor is to treat such patients.
A classical example is Servius’ commentary on Virgil, which, so I learned as a youth, comes in two main variants, a shorter original one and an expanded one (‘Servius Danielis’), most of whose extra material had been quarried in Donatus’ now lost commentary. J.H. Brusuelas has now taught me that, while what I learned was not quite wrong, the situation is much more complicated, inter alia because there have not been watertight bulkheads separating the two versions in the course of transmission, and so it is less than obvious how to present the text(s) in an edition.
Medieval Latin question commentaries might seem to be relatively unproblematic: they generally have exactly one author, even when the name is Anonymus, and most of them have come down to us in just one version. However, exceptions are not rare. I. Costa presents a case of a Quaestiones super Rhetoricam Aristotelis transmitted in two versions, μ and Μ, that reflect the author's own manuscript, Θ, at different stages of development, plus a manuscript, K, produced by someone who had access both to Θ and to a report of an extra question that the author included in an oral course of lectures. Roughly speaking, the text content is: μ + additions = M, and M + addition = K; so there are no prohibitive obstacles to presenting the three stages of the work in one edition.
M. McVaugh discusses Latin commentaries on Hippocrates’ Aphorisms. He presents a case in which we have both a student's notes from a master's oral presentation of quaestiones on the Aphorisms and what looks like a magisterially approved report of a later repetition of the course. The editorial problems that this situation raises are surmountable, but a serious question remains: does it make sense to publish a Latin commentary on the Aphorisms without accompanying it with an edition of Galenus Latinus’ on the same text, as the medieval work is barely understandable except in the light of Galen's with which it is in constant dialogue? If not, a normal-size editorial task becomes almost Herculean.
The Greek text of Galen's commentary on the Aphorisms is the subject of G. Ecca's contribution, and more precisely the Hippocratic lemmata. Similarly, L. Ferroni and G. van Riel discuss the Platonic lemmata in Proclus’ commentary on the Timaeus. Ideally, a lemma in a commentary should be unanimously transmitted and be identical with a unanimously transmitted piece of the source text and be identical with what the commentator quotes or implies. But there may be lack of unanimity here or there or in both places, and it may be unclear which reading is being commented on. The hazards on the way of transmission are many. A lemma may be ‘corrected’ with the help of a MS of the source text, a full-text lemma in MS a may be shortened to the ‘Α ἕως τοῦ Β’ format when a is used as an exemplar for MS b, only for a full text to be inserted again from a different source when b is copied to produce c; so one may end up with a lemma that differs seriously from what the commentator comments on. The two articles provide good examples of situations where it is not obvious what is going on. Unravelling the history of lemmata is important because, correctly reconstructed, they may yield precious information about the history of the source text, whereas an adulterated lemma thought to be authentic may lead editors of the source text astray (and this has happened).
C. Brockmann analyses two passages from Aristotle's Posterior Analytics and on philosophical grounds argues forcefully for the conjecture <δι>ότι at 1.31.88a1 and tentatively for a more invasive one at 88a16. He also offers an impressive analysis of the manuscript situation and tries to buttress his arguments with the testimony of two commentators: Philoponus (sixth century) apparently read διότι at 88a1, and his twelfth-century(?) successor Leon Magentenos seems to presuppose the same reading. But Leon carries little weight as a witness to a reading not attested in our Aristotelian MSS – ultimately, his interpretation of the passage almost certainly depends on Philoponus’.
One big sub-genre of the quaestiones is the Sentences commentary, the mastodont work by which a medieval bachelor of theology proved he was ripe to become a master. Such works often exist in more than one version, reflecting different stages of the development of the text and/or a process of abbreviation, due to the author or to someone else. Moreover, on closer inspection the text of any given section will often reveal itself to be a matryoshka, the outermost doll being the work of the purported author, but containing in it another doll, a text borrowed from a predecessor, that may in turn contain a third doll, and so on. Ideally, an editor should detect all the hidden dolls and somehow signal their existence to the reader. But how? M. Brînzei and C. Schabel ably present the problem, but have no easy solutions to offer.
There are no easy solutions either for editors of legal texts, as shown with several examples by S. Menzinger, who lays great (perhaps a little too much) stress on how little many medievals cared about authorship and how many books remained open to being added to and developing as long as the author lived or even longer.
The problems created by texts-within-texts are acute in the case of Liber Glosarum, a vast dictionary-cum-encyclopaedia with some 55,000 entries produced in or about the eighth century. As M. Giani explains, it consists of more or less redacted, sometimes mangled, extracts from earlier authors, and, like a modern encyclopaedia, it must have had many collaborators. Presumably, there was an editor-in-chief, but, if so, he did little to make his staff follow uniform procedures; hence, even if an archetype of the Liber can be reconstructed, there is no way to tell whether it is legitimate to correct its deviations from the source texts, even when they are obvious errors.
When it comes to instability over time, few genres can match bodies of scholia, the annotation that accompanies an authoritative text without forming a continuous commentary (though the border line to exposition commentaries is porous). It is supposed that the bodies of scholia on literary texts in Byzantine manuscripts often started as extracts from commentaries; but in their detached state they became anonymous, and, as F. Montana puts it in the article about Homeric scholia, ‘low authorship entails low textual identity’. In the case of scholia on the Iliad, editors have distinguished three classes, each representing a separate selection-cum-redaction of ancient material, but in the transmission they have become mixed up, and it is no trivial task to identify some stage in the life of such fluid texts and to say ‘This is what I want my edition to show’. Montana suggests that the situation is different with scientific texts, where ancient commentaries survived into Byzantine times, but my own experience with scholia on Aristotle is that they pose many of the same problems as the literary ones.
Papyri preserve fragments of the sort of ancient companions to the Iliad used for early compilations of scholia, but, as L. Pagani points out, it may be difficult to tell exactly which type of companion a fragment represents. Sometimes a fragment overlaps with a medieval scholium, which may, with some caution, be used to fill in holes in the papyrus, but one cannot jump to the conclusion that the work from which the fragment comes was the one from which the scholium was originally excerpted. In antiquity, too, material could flow from one work to another.
Homeric scholia are tough enough, but nothing beats Glossa ordinaria, the vast compilation of notes on the Latin Bible that grew and developed in Protean ways from the early twelfth century onwards, a huge number of manuscripts attesting to it in one form or another – an editor's nightmare. A. Andrée suggests one way to make a sensible edition of at least some parts of the Glossa is to reconstruct the text that a particular teacher of theology, say Peter Comestor, used, and in some cases this seems possible. I have one objection against a text sample. On p. 157 Andrée writes Nota in euangelio duo continet<ur> principaliter: fidei institutionem et morum informationem. Something is wrong, but it is not the active continet, as shown by the accusatives institutionem et … informationem that are explicative of duo. So, we have the object of continet but no subject.
Printed books are not ideal for presenting protean texts in a readable format, and several of the contributors to Sicut dicit suggest that digital editions might solve some of the problems. They surely will allow editors to present a lot of information without making it an impediment to the basic reading of the text in the way a swollen apparatus or a complicated system of parentheses and different typefaces does. They are hardly a panacea for all the problems laid bare in this volume, but whoever heard of a genuine panacea?
Sicut dicit raises more questions than it answers, but it is a rich book. It does not distract readers with misprints; and, though most of the contributors are not native speakers of English, awkward formulations are few and insignificant.
