Priscian's Ars was written in Constantinople in 526–527. It is the most complete grammatical work of the Greco-Latin tradition, and the first to divide grammatical exposition into phonetics/phonology, morphology and syntax. In addition to the two grammars authored by Donatus, it has served as a model for the development of the first vernacular grammars in Europe. Yet despite its great importance, it was last edited by Martin Hertz as long ago as 1855–1859. In 2015, however, Georg Olms/Weidmann launched a new edition incorporated as volume 13 in the admirable collection Collectanea Grammatica Latina, directed by Giuseppe Morelli(†) and Mario De Nono. Part II.1 of vol. 13, now published here, presents the text of the second part of Book 18 of the Ars as established by Michela Rosellini. It is divided into three sections: introduction (v–cxlix), critical text (7–115), indexes (119–59). Vol. 13 will eventually also include the commentary on the second part of Book 18 (II.2) and also the edition and commentary of its first part (I).

The decision to start the editorial project with the latest book of the Ars respects the particularities of its content and transmission. As for the content, Book 18 is an alphabetical lexicon of parallel Greek and Latin constructions, constituting not only a heterogeneous part in the Ars, but a unicum in ancient grammatical literature as a whole (v–vi). As for the transmission, Book 18 has been transmitted by a smaller number of manuscripts than the rest of the Ars due to the difficulty its parts in Greek posed to Western copyists (xxxi–xxxiii, xxxv). It thus allows one to understand more easily the genealogical structure of the manuscript tradition, and establish more directly the stemma codicum (xxv–xxvi).

R.’s edition is based on the ninth- and tenth-century manuscripts of the Ars. The reason for this is that the manuscripts preserving the Greek parts are restricted to that period, since later on Greek was no longer copied by Western copyists (xxvi–xxvii). In addition, the ninth- and tenth-century manuscripts are divided into three independent families, which allow for the reconstruction of their ancestor. As all later manuscripts can be related to one of those families, they offer no new contributions to the reconstruction of the text (xxvii).

In her long and meticulous introductory study, R. explains the scriptio of the Greek text (xxxvi–xxxviii), and describes one-by-one the manuscripts she has used according to the following aspects: (1) location and date; (2) characters and handwriting; (3) content (texts by Priscian and others); (4) eventual occurrence of glosses; (5) particularities of the transcription; (6) additions (xxxix–xciii). Finally she offers a reconstruction of the archetype (xciii–xcvii, cxvii–cxxxiii) and establishes the stemma codicum (c–cxvii).

In reconstructing the archetype, R. demonstrates that some mistakes belong not to the manuscript tradition, but to the archetype of Book 18 (xciii–xcvii). Some mistakes in the Greek text (relating to the quotations and spelling) were already in the Greek source used by Priscian; some mistakes in the Latin text (inaccurate citations with metric mistakes, repetition or displacement of lemmata) are attributable to Priscian himself (xciii–xciv). For Book 18 is actually an unfinished book, and what we see in it is the rough material Priscian collected for further elaboration (xcvii–xcviii), but he lacked the time to correct or complete it later (xciv, xcix). In establishing the stemma codicum, R. analyses the relationship between the manuscripts, dividing them into three families, and producing a diagram of their stemma (c–cxvii).

R. reviews the previous editions, from the editio princeps by Vindelino da Spira (Venice, 1470) down to Hertz's edition (Leipzig, 1855–1859) (cxxxiii–cxl), paying particular attention to the editorial treatment given to the Greek parts of Book 18. Hertz was the first to establish the text exclusively from manuscript evidence. What is more, he had at his disposal manuscripts belonging to all branches of the manuscript tradition, so that he was able to establish the text with certainty (cxxxix–cxl).

R.’s edition, in turn, rests on the assumption that Book 18 was left incomplete and unrevised by Priscian. Thus, whereas others have suggested ordering the material of Book 18 that Priscian himself left unfinished according to lexical principles, R. evidences the errors of the original, that is, gaps in quotations, the disorder of lemmata, repetitions in transcription (cxl–cxli). The critical apparatus is smaller than in Hertz's edition, because R. has established a genealogical plan of the manuscript tradition that allows her to eliminate singular readings, and because the modern philological documentation enables her to review and resolve many readings which Hertz felt obliged to indicate as uncertain (cxlvi–cxlvii). The second apparatus, in turn, is larger. It supplies the gaps in the quotations, and provides the corrections made by learned scribes in the manuscripts (cxlviii).

The Latin text is followed by four indexes: (1) index of words in the order of appearance in the text; (2) index of words according to the Greek lemmata; (3) index of Latin words; (4) index of quotations from Greek and Latin writers.

Over the past thirty years scholars have surveyed the entire manuscript material of Priscian's works in European and North American libraries. R.’s edition can be seen as the linchpin of this enterprise. Her critical principles not only allow her to establish the text in a solid and innovative way, but also help scholars to gain deeper understanding of the genesis of the Ars.