Bryn Mawr Classical Review

BMCR 2019.05.17 on the BMCR blog

Bryn Mawr Classical Review 2019.05.17

Marcus Deufert, Prolegomena zur editio teubneriana des Lukrez. Untersuchungen zur antiken Literatur und Geschichte, 124.   Berlin; Boston:  De Gruyter, 2017.  Pp. 265.  ISBN 9783110549980.  €109,95.  

Reviewed by Nicoletta Bruno, 'Thesaurus Linguae Latinae', München (

Table of Contents

Marcus Deufert, in over twenty years of outstanding study dedicated to Lucretius, is perhaps best known for his bold and controversial position on interpolation. He has interpreted many of Lucretius’s textual inconsistencies – repetitions, incomplete verses, gaps – as proof of extensive editorial activity soon after the poet’s death. Many Lucretian scholars have agreed that the De rerum natura (DRN) was interpolated at some point between its publication and the earliest medieval witnesses, the two famous codices Vossiani (O and Q), which date back to the 9th century. The suspicion of spurious verses arose during the Renaissance, but the first scholar who theorized that the DRN had been improved and polished by a later author was Heinrich Eichstädt in the early 19th century. In 1958 Gerhard Müller identified 270 interpolated verses.1 In 1975 Konrad Müller published his controversial critical text, which expunged 223 verses throughout the poem.2 Deufert in 19963 proposed the deletion of 368 verses, but in the preface of his Kritischer Kommentar (2018), he explains that in his Teubner edition, scheduled for Spring 2019, he eliminates 220 verses.

The Prolegomena zur editio teubneriana des Lukrez explain Deufert’s text of Lucretius’ poem and shed light on the main problems of the textual tradition of DRN. Deufert explains his editorial reasoning and defines, specifically and with admirable precision, the criteria he has followed in drawing up the critical apparatus.

Deufert’s position on interpolation is perhaps the most striking difference between him and David Butterfield, the editor of the future Lucretius OCT, who writes: “the hardened practitioner of Interpolationsforschung argues that Lucretius worked his poem up to a level of finish and perfection and as a result produced a wholly consistent and complete unit. (…) If it can be accepted that Lucretius did not finish the work, and therefore could let himself nod in composition, can it still be demonstrated that the poem was augmented by an over-active reader or an over-zealous editor?”4 The primary reason for these differences of opinion is essentially the following: although there is strong evidence regarding the dubious authenticity of a substantial number of verses, such as the cases of literal repetitions of long portions of text, we have no extra-textual evidence to show that there were one or more interpolators.

A secondary reason is that we do not have the strong evidence for the incompleteness of the Lucretian text that is found in the case of Virgil. There is a report that the DRN was annotated by the grammarian Marcus Valerius Probus in the second half of the first century AD, as we know from the Anecdoton Parisinum (Paris BN lat. 7530 ff. 23r-29r), a short treatise about the use of critical signs developed by Alexandrian scholars and used in Rome for different Latin texts by different scholars, amongst them Probus. Furthermore, Jerome states that he had commentaries on Lucretius (Adversus Rufinum 1, 16), and these commentaries might perhaps be related to Probus’ work. This does not constitute sufficient proof of interpolation. Nevertheless, Deufert has tried to prove that the use of critical signs in the text of Lucretius at an early stage of its transmission can be reconstructed from a strange particularity in the extant manuscripts, and he connects these with the interpolations: several verses in the archetype of all manuscripts, which can be reconstructed from O (Oblongus) and Γ (the lost model of Q and GVU), appear to have been rubricated as tituli. Deufert suspects that this arose through misunderstanding of critical signs.5 As Deufert shows at length in chapter III (Lukrezische Paratexte und ihre Darbietung in der Edition, pp. 177-203), at least nine verses of the Lucretian archetype were written as if they were tituli, and problems of authenticity are connected to eight of them. Deufert argues that the rubrication of the lines was not a mere accident; accordingly, it is highly probable that there is a correlation between interpolation and rubrication.

In contrast to Enrico Flores and in agreement with Michael Reeve and David Butterfield, Deufert finds that the tradition of Lucretius is unitary.6 In other words, the codices antiquiores of the ninth century, the lost codex Poggianus and the codices Itali (15th century) derive from a single archetype (= Ω). It is likely that the Poggianus is not a direct copy of the Oblongus O (Leidensis Voss. Lat. F 30, saec. IX) but descends from it by way of a lost intermediate copy. Therefore, only the readings of the Carolingian manuscripts have independent value, while the Itali are reduced to the status of transmitters of conjectures, which are numerous and persuasive. The differences between these two positions are obviously fundamental and have important consequences for constituting the text of the DRN.7

In the first chapter of the Prolegomena Deufert explains that the oldest tradition of the DRN is represented by three codices of the Carolingian age (here and elsewhere the abbreviations are those of Deufert, pp. XV-XVI): O (Leidensis Voss. Lat. F 30, saec. IX), Q (Leidensis Vossianus Lat. Q 94, saec. IX) and GVU: respectively, G (Schedae Gottorpienses Hafniae servatae, saec. IX), V (Schedae Vindobonenses priores, saec. IX), U (Schedae Vindobonenses posteriores, saec. IX). The three witnesses, O, Q and GVU, descend from a single archetype (Ω), from which are derived, on the one hand, O, and, on the other hand, Γ (the lost fons of Q and of the schedae GVS). The lineage is proved by Deufert’s accurate analysis of conjunctive errors (Bindefehler) and separative errors (Trennfehler) among the three codices antiquiores. Deufert uses the collective sigla Γ (= agreement of Q and GVU, that is, the second hyparchetype of the tradition next to O), and Ω (agreement of Γ and O, which constitutes the archetype) in carrying out the mechanical operation of the eliminatio lectionum singularium. For the Carolingian tradition, Deufert (pp. 17-65) proposes to eliminate: the errores singulares of the witnesses, the readings that are erroneous conjectures of medieval scribes, as well as their variants and the orthographic mistakes. At the end of the book, we also find ten color reproductions of the codices O and Q (pp. 259-265) and of eight other manuscripts (the last is a sheet of the ms. G), which give a solid idea of the interventions of the various correctors (in particular Dungal, the Irish corrector of the ninth century, who made many interventions on O, indicated with OD) as well as the subscriptiones, the tituli and the indices (pp. 323-324).

The second chapter of the Prolegomena (pp. 66-176) is devoted to the humanistic textual tradition. Deufert agrees (with Reeve and Butterfield, in disagreement with Flores) with the hypothesis that the so-called lost codex Poggianus, which can be reconstructed from the Itali, does not represent an independent branch of the textual tradition, but derived from the Oblongus (O). This argument impressively reduces the codices Itali to a repository of useful and remarkable conjectures.

In the third chapter (pp. 177-203), Deufert devotes particular attention to the paratexts in Carolingian manuscripts: the subscriptiones at the end of the poem, the tituli between the verses and the indices (only at the beginning of Books 4 to 6). The manuscripts O, G, V and U all have three types of paratext, written in capital letters in red ink; in the codex Q the subscriptiones and the tituli are missing, but the indices remain. Blank space is left for them where they should have been added by the rubricator. The subscriptiones and the tituli (but not the indices) must have already been in the archetype (Ω), capitalized (probably in red ink) to distinguish them from the poetic text.

Deufert is aware of the importance of the paratexts as evidence for the ancient reception of the DRN and he is rightly convinced of the need to report them to modern readers. Deufert admits that the tituli and indices do not date back to Lucretius, but were added later by a reader, a corrector. Moreover, the editing of the tituli seems to precede the editing of the indices: the latter seem to be a subsequent reworking. Deufert (p. 187) follows Lachmann and plans to publish all the subscriptiones and the tituli together in a special appendix also addressing their order in the manuscripts.

The volume ends with an exhaustive fourth chapter entitled Orthographica (pp. 204-248). Here Deufert wants to clarify the principles that an editor has to follow in confronting the abundant orthographic inconsistency in manuscripts. Secondly, he aims to offer a rich and detailed analysis of exemplary cases of the specific orthographic difficulties that are involved in the most complex interpretative problems.

Thanks to Prolegomena, the Teubner edition (in press), and the Kritischer Kommentar, Deufert’s name is destined to leave a lasting impression on Lucretian scholarship, for his incomparable methodological accuracy and his deep and mature knowledge of the DRN’s textual transmission and related problems.


1.   Eichstädt, Heinrich (ed.). T. Lucreti Cari De Rerum Natura Libri Sex. Vol. 1. Leipzig 1801. Müller, Gerhard. "Die Problematik des Lucreztextes seit Lachmann". Philologus 102 (1958) 247-83.
2.   T. Lucreti Cari De Rerum Natura Libri Sex. Conradus Müller recensuit et adnotavit. Zürich 1975.
3.   Deufert, M. Pseudo-Lukrezisches im Lukrez. Die unechten Verse in Lukrezens „De rerum natura“. Berlin-New York 1996.
4.   Butterfield, David J. "Lucretius Auctus? The Question of Interpolation in the De rerum natura" in Fakes and Forgers of Classical Literature. Ergo decipiatur!, edited by Javier Martínez. Leiden-Boston, 2014, 29-30.
5.   Deufert, Marcus. "Overlooked Manuscript Evidence for Interpolations in Lucretius?: The Rubricated Lines" in Latin Literature and its Transmission, edited by Richard Hunter and Stephen P. Oakley. Cambridge 2016, 70.
6.   Reeve, Michael. "The Italian Tradition of Lucretius Revisited". Aevum 79 (2005) 115-64. Butterfield, David J. The Early Textual History of Lucretius’ De rerum natura. Cambridge 2013, 210-11. The most recent critical edition of Lucretius’ De rerum natura is edited by Enrico Flores (Titus Lucretius Carus, De rerum natura, Critical Edition with Introduction and Version. Vol. I: Libri I-III, Napoli 2002; vol II: Book IV, Napoli 2004; Volume III: Books V-VI, Napoli 2009). This edition is based on a systematic collation of both the Carolingian period manuscripts and the whole of the Humanistic tradition (and this collation is its most substantial innovation). Flores considers the tradition of the Itali to be independent of that of the codices antiquiores and necessary for the constitutio textus of the DRN.
7.   Deufert investigated the entire manuscript tradition of the DRN, from the Carolingian period to the Renaissance, but only part of his collation will appear in the apparatus criticus of his forthcoming edition.

Read comments on this review or add a comment on the BMCR blog

Read Latest
About BMCR
Review for BMCR
Support BMCR

BMCR, Bryn Mawr College, 101 N. Merion Ave., Bryn Mawr, PA 19010