The Textual Transmission of Aristotle's De anima

This TIDA subproject studies the textual transmission of Aristotle’s De anima in Greek, Arabic, Latin and Hebrew. The aim is to produce a reliable critical edition of the text that takes into consideration the set of all relevant witnesses.

A critical edition of an ancient text aims at reconstructing the text written by the author on the basis of the extant evidence. The first step is the identification of every document that provides some testimony to the text: this may include Greek medieval manuscripts, ancient papyri, translations in ancient or medieval languages such as Arabic or Latin, commentaries (especially ancient ones) and other exegetical productions, early Renaissance printed editions, and more generally any source that reflects a determinate state of the text with a reasonable degree of accuracy. The whole evidence is then collected and processed so as to facilitate comparisons, for instance by recording the readings of a given manuscript every time its text de-parts from a reference text (‘collating’). The second step is the critical examination of these witnesses in order to map out their genealogical relationships. Part of its goal is to identify dependence-relationships between textual witnesses, so as to exclude witnesses which are found to be derivative with regard to other extant witnesses and to provide no evidence of their own. Its results are usually summarized graphically by means of a stemma. The third and final step is the edition itself. Every variant reading, once its precise authority has been evaluated, is examined in the light of the transmission and of the author’s language, style and thought.

In the case of Aristotle’s De anima, the evidence includes, among others, around a hundred Greek manuscripts, ranging roughly from the 10th to the 15th century, three Arabic translations, two medieval Latin translations (12th-13th centuries), Alexander of Aphrodisias’ own work on the soul as well as his lost commentary to Aristotle’s treatise (ca. 200 CE), a paraphrasis by Themistius (4th century CE) ancient commentaries by Ps.-Simplicius and Philoponus (6th century CE). Half of these Greek manuscripts have not been properly examined, most of the Arabic and Latin translations have not received scientific editions yet and the issue of their relationship to the Greek transmission has not been addressed.

The situation is particularly dramatic regarding the Arabic translations of the treatise. Aristotle’s De anima is one of the few texts in the medieval Graeco-Arabic translation corpus for which we have for virtually every passage of the text the evidence — extant or reconstructible — of two Arabic translations: one complete and two partially complete but which complement each other. (The text of the latter two transla-tions must however be reconstructed for the most part on the basis of the indirect tradition, esp. the Arabo-Hebrew and Arabo-Latin translations.) A rigorous reconstruction of the readings underlying the Arabic translations allows us to evaluate two further lost Greek textual witnesses for virtually every pas-sage of the text. The Greek exemplars of the Arabic translations are older than any of the extant witnesses of the direct tradition and their stemmatic position within the Greek tradition has hitherto not been properly determined.

Modern editions of Aristotle’s De anima are all based on Bekker’s seminal work that culminated with the publication in 1831 of the first edition of Aristotle’s complete works to be exclusively based on a critical examination of the manuscript evidence. Bekker relies on eight Greek manuscripts to reconstruct Aristotle’s text for De anima: E (Paris. gr. 1853), S (Laurent. Plut. 81.1), T (Vat. gr. 256), U (Vat. gr. 260), V (Vat. gr. 266), W (Vat. gr. 1026), X (Ambros. H 50 sup.) and for the third book only L (Vat. gr. 253). He gives no indication as to his assessment of their respective value or his method for reconstructing the text. Bekker’s edition soon came under criticism on the part of Trendelenburg, who insisted on the potential value as textual witnesses of ancient commentaries and early Renaissance editions, both of which had not been taken into account by Bekker. Trendelenburg’s own edition, published only two years later in 1833, makes use for the first time of the paraphrasis by Themistius and the commentaries by Ps.-Simplicius and Philoponus, as well as one more manuscript (Paris. gr. 2034) and two editions (the 1497 Aldina and the 1587 Sylburgiana), and emphasizes the importance of the text of manuscript Paris. gr. 1853 (E).

Three decades later, Torstrik, himself a former student of Trendelenburg, published a new edition of the treatise in 1862 in which he provided two ground-breaking hypotheses. 1. Torstrik realized that the text of Paris. gr. 1853 (E) had been replaced by a new version for book II, with the consequence that the value of its testimony for book II should not be automatically put on a par with its value for books I and III. 2. Torstrik also claimed that the available manuscript evidence is to be divided into two groups: A, which comprises the original part of manuscript E and manuscript L, and B, which includes every other manu-script known at the time. No witness of the A-text for book II has been identified so far. This split within the transmission made such an impression on Torstrik that he ventured the claim that A and B were two authentic versions of the same treatise, as if Aristotle had himself revised his own text. This claim was attacked by Bonitz and subsequently retracted by Torstrik.

Torstrik’s hypotheses were further refined in Förster’s 1912 edition of the treatise. Förster examined 49 manuscripts of the treatise, claiming that almost all of them were of no significant value – with one single exception: manuscript T (Vat. gr. 256) turned out to be a copy of C (Paris. Coislin. 386). Thanks to the developments of the Commentaria in Aristotelem Graeca under the leadership of Diels, Förster was able to take advantage of much-improved scientific editions of the indirect ancient evidence. He refined Tostrik’s view about the structure of the transmission into a stemma codicum presenting the relationships between the nine manuscripts which form the textual basis of his edition.

In the meantime, a “new” manuscript of the treatise, Vat. gr. 1339 (P) had come to light when Rabe no-ticed the peculiarities of its text of book II, which he transcribed and had published in 1894. Förster took note of its odd readings in an appendix, but did not address the question of its situation within the transmission. No editor has been able to give a clear assessment of the nature of its testimony yet.

Our knowledge of the history of the transmission of Aristotle’s De anima has barely made any progress since 1912. De Corte, after traveling through various European libraries, pointed out in the 1930s a few interesting manuscripts which had been rejected by Förster. W. D. Ross published two editions of the treatise in Oxford (1955 & 1961) which, as far as the reconstruction of the transmission goes, are entirely based on Förster’s work. Siwek studied some 65 manuscripts of the treatise and went on to publish an edition as well (his second one, as he had actually been working on the treatise since the 1930s). According to Siwek, the tradition is composed of 9 families that are more or less independent from each other. Unfortunately, his results are marred by considerable uncertainty due to his peculiar methods. Jannone prepared a new Greek text for the 1966 French Budé volume, but chose to focus almost exclusively on Italian manuscripts.

The TIDA project aims at providing a complete reconstruction of the history of the transmission of Aristotle’s De anima. The following three key issues have been identified:

1. Are there extant witnesses belonging to the A-group for book II?
2. What is the relationship between (a) the Greek manuscript transmission of the treatise and (b) indirect textual testimonies, especially (b1) ancient Greek exegetical works, (b2) medieval Graeco-Arabic, and (b3) medieval Graeco-Latin translations?
3. What is the status of manuscript Vat. gr. 1339 (P)?