The editions by Therese Fuhrer of Contra Academicos (hereafter Acad.) and De ordine (hereafter ord.), and by Simone Adam of De Beata uita (hereafter beat. u.) are the most recent works of textual criticism of Augustine’s dialogues of Cassiciacum.1 Their study brings some new conclusions about the classification of manuscripts by group and affirms their high degree of contamination. However, the editors neither give a clear picture of the transmission nor take stock of the contamination issue: they say that they have examined all the manuscripts, but they do not give any list (p. X); they say that the transmission is contaminated, but they do not share any evidence with the reader (p. XIV, XVIII-XIX, XLVIII); they provide critical texts based mainly on a few manuscripts from one group of the tradition, but they just say that these (contaminated) manuscripts have an homogenous or a better text (p. XIII, XIX). We would have expected stronger bases and further explanations.
However, it is always useful to have a new edition, especially in the case of the dialogues of Cassiciacum. Indeed, among the many editions of these dialogues published during the twentieth century, some did not actually improve the text, as they were largely based on the textus receptus of the Maurist edition (1679) as reproduced in the Patrologia Latina by J.-P. Migne.2 Only the editions of P. Knöll in the CSEL in 1922 (reprinted in 1962), of W.M. Green in the Stromata Patristica and Mediaevalia in 1956 (reprinted in the CCSL in 1970), and of J. Doignon, for beat. u. and ord., in the Bibliothèque Augustinienne in 1986 and 1997 are based on the classification of manuscripts. Nevertheless, the methodology employed by these editors to construct their texts varied in quality and led to opposing conclusions. In order to better compare T. Fuhrer's and S. Adam’s editions with previous editions, it would have been preferable to see a complete list of variants in the Preface, instead of some variants printed in the critical apparatus (p. XLVIII),
Editing Acad., beat. u. and ord. together is relevant since these three works partly share the same manuscript tradition. Actually, Acad. and ord. are usually copied together, whereas beat. u. often has a specific tradition (perhaps due to the special interest in this work in the monastery of Hippo, p. X). The distribution of the editorial work between T. Fuhrer and S. Adam is therefore coherent. Editing these three works was a real challenge, because the editors had to cope with massive contamination. Unfortunately, in this situation, there is little an editor can do, apart from thoroughly and carefully explaining his choices. I praise the honesty of the editors on the nature of their edition (“this edition’s text can be understood as a proposed reconstruction—or rather construction—of a possible version of the author’s text”, p. XLIX) and on their principles (“we should avoid further contaminating the text by combining variants from different textual traditions with another at whim”, p. XLIX, which was the major bias of J. Doignon’s editions). The critical apparatus, of semi-negative type (becoming positive when clarification is needed), is realistic and convincing, and it matches the principles laid out on p. XLIX-L. This apparatus is not exhaustive, but it is in no way a bad choice. Moreover, I should like to fully support the editors’ point of view p. L: “The critical apparatus should not, however, document all features of each manuscript; thus, the vast majority of unique readings or additions that are characteristic of a given manuscript are not included in the apparatus”.
However, I am not comfortable with the way the philological presentation is outlined. Indeed, the presentation of the Preface (II. Transmission, p. IX-XLV; III. Editions, p. XLV-XLVIII; IV. This text, p. XLVIII-LII), and especially the section II, could have been clearer. The presentation of the tradition arranged by century (p. XI-XII) is somehow puzzling at the beginning of the section devoted to transmission, because the reader has not yet become aware of the difficulty of the matter. If placed after the pages XII-XLV (on the data of the transmission), it would have been a helpful complement to catch the philological features of each period. More confusing is the description of the manuscripts after the pages where their connections and groups are presented: one is forced to switch constantly between the sigla used p. XIII-XXI and their elucidation p. XXII-XLV. Putting a list, even limited to some basic information (signature, date, origin and / or provenance) before discussing the actual transmission would have been better. I am not comfortable either with some assertions in the Preface, for instance, p. XLIX: “This edition attempts to get as close as possible to the original composition of the author. This attempt is, however, made in knowledge that total success is not possible due to paucity of textual witnesses—that is, the high number of lost manuscripts”. No matter how many witnesses, lost or not, there are to reconstruct the author's text: the problem with manuscripts is not a matter of quantity, but quality. As far as manuscripts are concerned, the recensio codicum is based on all the witnesses, that is to say around one hundred manuscripts copied between the ninth and fifteenth centuries3; the number of witnesses meant that the editors chose what they considered to be the best manuscripts on the basis of selective collations (p. XI). However, instead of giving the complete list of collated witnesses, for Acad. and ord. T. Fuhrer preferred to refer to the (partial) lists of the Knöll and Green editions. This means that the reader has to trust the editor when she says that all manuscripts have been collated. For beat. u. S. Adam refers to her dissertation. With few exceptions (the details of which are not given), collations were carried out from reproductions (p. XI). Besides, whether a second direct collation has been made on the manuscripts used for the construction of the text is not said.
Regarding Acad. and ord., T. Fuhrer faced two main issues: one is the early contamination of what she calls the Φ-group (p. XIV), the other is the problem regarding the tradition. Fundamental to T. Fuhrer’s edition is the following postulate: the additional words in the γ/τ/σ tradition (listed n. 12, 127, 128 and 129) are interpolations from which the Φ-tradition’s text is free. On this matter, T. Fuhrer agrees with the position of Knöll and Green, and not with Doignon who considered these interpolations as parts of the original text. Unfortunately, there is a misprint in the Preface: some words are missing in the sentence (p. XVII, line 22) explaining this situation, and it prevents the reader from understanding the problem. On the classification and contamination of manuscripts, T. Fuhrer marks an actual philological break with the previous editions: on the one hand, the editions of Knöll and Green concluded, for Acad. and ord., that the manuscripts were divided into two families with no contamination; on the other hand, Doignon had already noticed that the tradition of ord. was contaminated but he stated that the manuscripts the so-called Φ-group is made of was an interpolated text. According to T. Fuhrer, the Φ-group provides a rather good text, and therefore the construction of her critical text is based on it. Nevertheless, T. Fuhrer has occasionally printed the variants of the γ/τ/σ tradition when no satisfying solution was offered by the manuscripts she had first chosen,4 which is the least bad alternative.5 Now, that being said, I have some reservations. First, T. Fuhrer concludes that the tradition of Acad. and ord. is an open recension, so no stemma codicum can be drawn (p. XIX). This assertion is, I feel, a bit hasty: open recension does not necessarily mean no possible stemma. No global stemma indeed, but still some groups and connections can be established. In fact, contrary to what T. Fuhrer said, she created some partial stemmata. By the way, if no direct dependence of β from C can be established (p. XVI), it is unwise to make it appear on the partial stemmata (p. XVIII). In addition, if no global stemma can be proposed, it is better not to place the siglum ω standing for the archetype on the partial stemma I of the τ-tradition and its connection to the Φ-tradition (p. XVIII). It just so happens that my second point deals with the archetype: although there are some traces of archetypal errors,6 T. Fuhrer does not say a word about the existence of an archetype. Thirdly, and this point is more important, this is what T. Fuhrer writes about the contamination of the Φ-group: “The divergences within the Φ-tradition make it clear that the scribes of these manuscripts were copying from exemplars which were themselves composed from a mixture of sources and/or themselves employed copying practices which mixed exemplars to produce contaminated texts”. Giving some examples that support this assertion would have been highly expected. The contamination issue should have been one of the major achievement of this edition, but evidence is lacking.
As far as beat. u. is concerned, S. Adam had to face to an extensive cross-contamination, which made her renounce to build a stemma (p. XIX). The text is based on thirteen manuscripts. However, these thirteen manuscripts do not form a group because they share some readings (p. XIX): they are gathered together because they do not share the readings of two other groups represented by the sigla Γ and Λ. According to the editor, all these thirteen manuscripts put together provide an “overwhelmingly good text” (p. XIX), which contrasts with the over-contaminated text of the Γ-group and the smoothed one of the Λ-group. When S. Adam says that “the classic common error method” cannot be used “to determine specific genealogical relationships between manuscripts” (p. XIX) in such a contaminated tradition, she is right. However, she cannot consider these thirteen manuscripts as a group. Moreover, there is no further information about the relationships between the thirteen manuscripts and the way S. Adam managed to choose the readings when these witnesses disagree. I hope that these justifications are provided in her dissertation and will be given in a forthcoming paper.
The editions of T. Fuhrer and S. Adam give new insights about the transmission of the three works, Contra Academicos, De beata uita and De ordine, and they provide a new Latin text in keeping with the principles set out in the Preface. However, some further justification is required, and some assertions that involved the methodology used give an uneasy feeling. Contamination is an important issue in philology, and these editions are an encouragement to have a closer examination of this matter.
1. Furthermore, there is an edition pending of Contra Academicos by Anne-Isabelle Bouton-Touboulic in the Bibliothèque Augustinienne.
2. V. Capánaga 1947; R. Jolivet 1948; D. Gentili 1970; M. Betteni/G. Catapano/G. Reale 2006.
3. Among the excerpts that have also been used for constituting the text of ord., T. Fuhrer indicates Cheltenham, Phillips (sic) Collection (Robinson Trust), 16278 (p. XLIV). This reference to the bibliotheca Phillippica is no longer correct: the manuscript was held in the J. Paul Getty Museum Collection until 1988. The last record we have is when it was sold at Sotheby’s, London, June 18, 2002, lot 38. The editor should have specified that she knew the variants of this excerpt through the previous editions. Furthermore, there is a mistake n. 197: Phillipps is the correct spelling.
4. One example is significant enough: in Acad. 3, 10, the γσ tradition has preserved (or has made a good correction ?) the correct reading falsa, against falli from Φ+(Rac). However, I do not understand why in Acad. 3, 1 the classical perfect indicative uacauit given by the γσ-tradition has been preferred to the late antique form, uacuit, given by Φ+(PacRac).
5. Some of the corrections T. Fuhrer made on the text of Acad. were already printed in the German translation and commentary she published in 1997.
6. E.g. Acad. 2, 21 uicturum] coni. Schaüblin uictorem codd.; Acad. 2, 41 dixisti] Fuh1 (cf. p. 40, 26-27) dixi codd.