Observations on the organisation and original extent of 
Cyril, 'Contra Julianum' and Julian 'Contra Galilaeos'

by

Wolfram KINZIG and Michael CHRONZ

(Translated by Roger Pearse)

The "final, great Apologies of Augustine, Theodoret of Cyrrhus, and Cyril of Alexandria" bring to an end the era in which Christianity is transformed into the state religion.  Unfortunately the long vindication written against the emperor Julian by the Alexandrian patriarch has only come down to us in an incomplete form.  Uncertainty also prevails on the original extent of the Contra Iulianum (hereafter: CI), as well as on the dimensions of the work combatted, the Contra Galilaeos of the emperor Julian the Apostate (hereafter: CG).  The intention of what follows is to clarify these two interlocking problems.

Since the question of the extent of CI is usually answered in the literature by reference to the extent of CG, a brief overview of the latter is required.

1.  The original extent of Contra Galilaeos

In 362-3 AD, during the preparation for the Persian expedition in which he would perish, the emperor Julian wrote at Antioch his work against the Christians, whose title was probably κατα τῶν Γαλιλαιῶν.   This has not come down to us, and can only be reconstructed in part from quotations in Christian authors.  By far the majority of these (85%) are in CI.  The date at which the work was lost cannot be determined.

We have two contradictory statements about the extent of the work.  Jerome twice refers to seven books, in two different works: in Letter 70 (to Magnus) in 397-8, and in his Commentary on Hosea in 406.  Neumann saw that a mistake in transmission of VII for III is out of the question, in view of the two witnesses.  Instead he preferred to see it as a mistake by Jerome. Perhaps he did not have Julian's work before him, but one of the now lost Christian refutations (perhaps those of Theodore of Mopsuestia or Philip of Side), which were in seven books.  When referring to the work, he wrongly equated the number of books in CG with the number in the refutation.

The other statement is in CI.  In order to understand this, we must bear in mind the transmission of the text of CI.  The Prosphonema and the first ten books of CI are more or less completely transmitted in 14 manuscripts (see table 1):

Christoph Riedweg has investigated the manuscript tradition.  For the new edition of CI, not only must the main manuscripts GV and FME be considered, but also the glosses in IBHQ must be collated, and the Latin translation of Oecolampadius also taken into account.  For the prosphonema, we are restricted to G and FME, since V does not contain it, plus the glosses from BQ only, as this interesting preface is not in I or H, and, finally, the as yet unexamined early editions, particularly that of Oecolampadius.

In G, as in all the Greek editions, Julian wrote three books against the Christians: Εκηῶν τοινυν ευπηυα τὲν γλῶτταν ηο κρατιστοσ Ιουλιανοσ κατετηὲχεν αυτὲν του παντῶν ηὲμῶν Σῶτὲροσ Κηριστου? καὶ δὲ καὶ τρια συγγεγραπηὲ βιβλια κατα τῶν ηαγιῶν ευαγγελιῶν καὶ κατα τὲσ ευαγουσ τῶν Κηριστιανῶν τηρὲσκειασ (PG 76, 508BC/SC 322, 106, 14-17). In B, τρια is added in the margin by a fourth, probably Western, hand.  Q is a 17th century manuscript, which has evidently benefitted from text-critical comparison with other manuscripts of CI, and possibly was created to serve the edition of Aubert.  It is surrounded by square brackets in the text, while in the margin is a note that it is not found in all the manuscripts.  In the lost Codex Capnioneus the text included the numeral, as the translation of Oecolampadius witnesses. Riedweg assumed, probably correctly, that the entries in BQ and ? came from G. In addition, Michael Glykas bears witness to the numeral in the middle of the 12th century in his World Chronicle, which contains this passage from CI transcribed almost word for word.  But in codices FME which must derive from the same archetype as the others, the numeral is missing.

We will follow Neumann in rejecting the testimony of Jerome.  But what about the number of three books?  Julian himself gives us a pointer.  Cyril quotes a passage from CG in CI at the beginning of book 8, in which Julian seeks to prove that the Christians in fact contradict Moses and the Prophets.  Specifically, Julian wants to show that the Christian interpretation of Deut. 18:18 and Gen. 49:10 is wrong: Moses did not prophesy the birth of Christ in these passages.  The second passage promises a prince  from Judah, but in Christian theory Jesus was begotten by the Holy Spirit, and likewise the fiction of the descent of Joseph from Judah fails, because the evangelists Matthew and Luke contradict each other in their genealogy of Jesus.  After the retort of Cyril, follows the continuation of the quotation from Julian:

αλλα περι μεν τουτου μελλοντεσ εν τῶ δευτερῶ συγγραμματι το αλὲτηες ακριβῶσ εχεταζειν ηυπερτιτηεμετηα.

Neumann has shown that in Julian's day συγγραμμα or βιβλος (like βιβλιον) can mean "writing, work", and "part of a work, book" with passages from Origen, Herodian, Eusebius and Theodoret.  According to the TLG, the word appears 6 times in Julian's works.  In one place (Themistio phil. 7,10 [260D]), Julian introduces a quotation from Aristotle's Politics (III, 15, 1286B) as follows: "Πηὲσι δε ηο ανὲρ εν τοισ Πολιτικοισ συγγραμμασιν..." 'Here by συγγραμμα the "books" of the Politics are meant.'

Julian thus shifts, in the quotation above from CG, therefore, the argument about the gospels from the second book of his work: the quotation itself must come from the first book of the CG, in which Julian attacked the Christian use of the Old Testament.  Strictly the numeral ηετερον means that Julian planned at least three books against the Galilaeans, otherwise he would have had to speak of συγγραμμα.  This information from Cyril himself fits the number indicated in the prosphonema.  Julian, therefore, seems to have actually published at least three books.

In CI, Cyril in book 1 begins with a general defence of Christianity, and in books 2-10 engages with the attack of Julian on the Christian understanding of the Old Testament.  At the beginning of the second book he refers explicitly to the λογοσ πρῶτος of his adversary.  At no point does he indicate that he is refuting a second or third book of the work.  Also, in the quotations from CG, no transition to a new book is recognizable.

From the fact that Julian promised a critique of the  gospels in his second book, and that the indirectly transmitted fragments of books 11-19 of CI concern themselves exclusively with this subject, Neumann inferred that the third book of CG concerned the Acts and the Letters of the New Testament, and thus that Cyril must have written books 21-30 of CI on this.

Neumann prints two short texts as fragments of book 3 of CG.  The first comes from CI and contains no direct quotation from CG, but only the remark that Julian scoffed at the hope of eternal life of the Christians.  Neumann seeks to relate this to the words of Paul (1 Thess. 4:13 ff), although they do not seem very close.

The second passage he takes from the Suda, which quotes a sentence of Julian's under the keyword απονοια, but without attributing it to CG.  Jean Bidez and Franz Cumont doubted this assignment, because the Suda quotes Julian often, but never from CG.  Also the sentence seems to have nothing to do with the Christians, or the New Testament.

There are thus no fragments of CG known which can be certainly assigned to the third book.  The existence of this book need not be doubted, however.

2. Contra Iulianum

Cyril's work in the manuscripts bears the title: Υπερ τὲσ τῶν Κηριστιανῶν ευαγουσ τηρὲσκειασ προσ τα του εν ατηειοσ Ιουλιανου. The date of composition, time of publication, and the reason for writing are not usually clarified in the literature.

Cyril was born in 380 AD, and became patriarch of Alexandria in 412 until his death in 444.  Several attempts have been made to pin down the date of composition and publication of the CI.  The only external reference for the date of publication is in Theodoret of Cyrrhus († ca. 466) in letter 83 (in 448) to Cyril's successor, Dioscuros († 451): 

I think that your holiness already knows that Cyril of blessed and holy memory often wrote letters to us.  And when sending what he had written against Julian at Antioch, and also his work on the scapegoat, he asked the blessed John, bishop of Antioch, to show this to the teachers in the Orient.  And the blessed John responded to this letter by sending us the books, and after we had read them, we admired them and sent Cyril of blessed memory a letter.  And he wrote back and assured us of the accuracy and his affection, and these letters we have. 

Since John ascended the episcopal throne of Antioch in 428 or 429, and died in 441 or 442, Cyril must have sent him the two works mentioned within this period.  But on the other hand, since the Nestorian controversies took up most of Cyril's time from 425 on, he might well have composed CI before then.  Previous attempts to specify the date of composition and publication further have been contradictory and unconvincing.

The most recent suggestion for the date is that of Markus Vinzent.  He proceeds from the certainly dated Festal letters of Cyril, which in each case discuss current conditions in Egypt.  In letters 1, 3-6 and 9-10 (written between 414 and 417) Cyril addresses the remaining pagans and semi-Christians with the relationship of Christianity and Hellenism.  In the festal letters 12-14 ( 424-426), the main them of "Hellenisation" within and without Christianity reaches its high point.  Thereafter, in letters 15-18, from 427-430, the subject noticeably fades away and in letters 19-30 (431-442) disappears entirely.

The Festal letters 12-16 (424-428) contain repeated resemblances to CI, sometimes even word-for-word parallels.  Already in 424 in festal letter 12 there are parallels to many books and the prosphonema of CI.  On the basis of these convincing arguments, Vinzent proposes that the work was completed between 423 and 428, most likely at the latest by Easter 424.

It is noticeable that CI first circulates some sixty years after CG.  Cyril himself gives the confusion caused to numerous believers by the work of Julian as the reason for his retort in his letter of dedication to Theodosius II.  Both weak and strong believers were impressed by the apparent literary learning of Julian, and the pagans made use of his arguments against the Christians which, so they proclaimed, no Christian teacher had been able to refute.  These statements of Cyril agree with the findings of Markus Vinzent based on the Easter festal letters which indicate that in the first two decades of the 5th century there were numerous baptisms carried out into a syncretism of Christianity and paganism.

Perhaps an investigation into the religious policy of Theodosius II and the relationship between Cyril and the emperor would throw some light on these issues.  Such an investigation remains to be done, and cannot be undertaken in the context of the current study.

3. The extent of CI

In order to determine the original length of CI, we must first clarify the direct and indirect transmission of the text and the organisation of the book.

3.1 The printed editions

Only the first ten books are well-known and directly transmitted.  In the printed editions they are numbered sequentially.  Oecolampadius in his Latin translation labelled them "liber primus", etc.  But he did not take this numbering from his Greek manuscript, the lost Codex Capnioneus in the possession of his teacher Reuchlin; instead he introduced it, as he indicates at the start of book 6 (f. 46D): 

In exemplari praescribebatur liber Secundus, nimirum pars secunda operis, quae et ipsa quinque libros continet.

Therefore the Codex Capnioneus was divided into pentads.

Borbonius printed only book 1 of CI, without the prosphonema.  He titled it βιβλιον, which gives no indications about the manuscript from which he prepared it.

Aubert numbers the ten books sequentially  (Λογοσ προτος/Liber primus, etc).  At the beginning of his Variae lectiones he writes:

Dividuntur hi libri in aliis aliter codicibus. Sunt enim qui in 2. libros, singulos in V. tomos totum hoc adversus Iulianum opus dividunt.

Also he thus knew of manuscripts, in which the ten books were divided into two pentads of five books each.  Unfortunately he neither indicates which manuscript he used, nor the source of his preferred numbering system.  This sequence of the books appears again in vol. 76 of the Patrologia Graeca.

3.2 The direct transmission of books I-X in the manuscripts.

The prosphonema ( = Pr.) and the first ten books of CI are more or less completely transmitted in fourteen manuscripts, which however only contain parts of the transmission (c.f. p. 31f).

The surviving manuscripts, without weighing their importance, have the following contents:

5 - the prophonema and books I-X (FMEBQ),
2 - contain only books I-V (VI),
1 - because of damage only books IV-X (H),
1 - because of damage only the prosphonema and book I (G)

Codex V seems to have been copied from a manuscript which contained only the first five books, as the note τελος at the end of book five suggests.  The text follows a copy of Cyril, Contra Anthropomorphitas.  In codex B book V of CI is likewise followed by Contra Anthropomorphitas and then by books VI-X of CI.  This codex is contaminated by readings from more than one manuscript family.  For the first five books, it follows V, while for books VI-X, missing in V, it follows M.  The organisation of the books in these manuscripts is given in table 2.

3.3 The indirect transmission

Excerpts from both the first ten books and also books XI-XIX of CI are transmitted indirectly in Greek and Syriac florilegia and catena.  These will now be presented, as they throw light on the internal organisation of the text.

Florilegia

1. Florilegium Cyrillianum (before 483)

The Florilegium Cyrillianum is a pro-Chalcedonian anthology from the works of Cyril, compiled in Alexandria before 483.  It is only transmitted to us indirectly.  Between 508 and 511 it was quoted at length verbatim by Severus of Antioch in the first part of his Philalethes, and then corrected and refuted by comparison in the second part with the original works.   At this time it contained five quotations from CI under no. 176-180, three of them from the lost books XII-XIV.  Later Severus in his Apology for the Philalethes repeated part of the florilegium, and his refutation of it.  Both works of Severus are preserved only in Syriac translations of the second or third decade of the 6th century, the Philalethes in cod. Vat. syr. 139 (8th century), the Apology in cod. Vat. syr. 140 (6th century).  These two Syriac codices are the oldest witnesses for the indirect transmission of the CI.  Before the loss of the Greek Philalethes (probably in the 6th or 7th century) the majority of the text of the florilegium contained in it was copied.  In this form it exists in three Greek manuscripts of the 13-15th centuries.

2.  The Dyotheletic florilegium of the Roman legates to the 6th ecumenical council of Constantinople, 680-81.

At the 6th ecumenical council of Constantinople the Papal legates submitted a dyotheletic florilegium, which among other things contained two quotations from book XII of the CI.  The florilegium was probably compiled shortly before the council by Greek monks in Rome.  At the tenth session of the council the florilegium was read out in full, when the individual quotations were compared with the text of the original works in the patriarchal library in Constantinople.  Since this proceeding was applied to the quotations from CI, another copy of CI which was still complete at least as far as book XII must have existed at that time.

3.  The Christological florilegium of codex Athos Vatopedianus 507 (after 685)

This florilegium is concerned with 13 testimonia which comprise a comprehensive addition to the Doctrina Patrum de Incarnatione Verbi,  which it follows in the manuscript.  The terminus post quem for the Vatopedi florilegium is therefore the date of the Doctrina Patrum, finished in 685, which it presupposes.  The terminus ante quem is the writing of the manuscript at the beginning of the 12th century.  However Maximus the Confessor seems to be the most recent writer quoted, so the date of the compilation may have to be set quite soon after the Doctrina patrum.  The compiler is unknown.  Since he takes his material rather obviously from other compilations, we need not suppose that a copy of CI was available to him.

The text of the excerpts from CI corresponds approximately to that contained in the two sections of the legates florilegium of 681/82, but it contains the bridging text between the two pieces which the latter omits.  The text cannot therefore be derived from that florilegium.  But since it exhibits gaps when compared to the Roman florilegium, a dependence in the other direction is also impossible.

4.  John Damascene, Hiera (Sacra Parallela - before 754)

The original form of this large florilegium carried the title ἰερα and comprised three books.  The first contained testimonia on God, the second on man, and the third on virtues and vices.  In the first two books the excerpts are arranged alphabetically by keyword, in the third each virtue and vice is paired.  The work has not come down to us in the original form, however, but only in later recensions.  It is attributed to John Damascene in numerous manuscripts.

Only one recension has been completed published, the Florilegium Vaticanum, by Michel Lequien from codex Vaticanus graecus 1236 (saec. XV), under the title Sacra Parallela.  From another recension, the Florilegium Rupefucaldinum, he prints only the index and some excerpts from the sole manuscript (cod. Berolin. gr. 46 [Phillips  1450]; saec. XII).  The title of the work is only found in the lemmata of the manuscript, and added by Lequien with some other excerpts take from Vat. gr. 1236.  Both recensions contain parts of all three books of the Hiera.  The edition of Lequien offers 17 excerpts from the lost books of CI out of both manuscripts, including two pairs of doublets (of which one pair is wrongly assigned to book V, when it probably belongs to book XV), 1 wrongly assigned to book I (probably from book XI), 3 without indication of which book they come from, and 1 from Chrysostom wrongly attributed.

From the Recensio Vaticana of the second book of Hiera, the forword, index and some excerpts were published by Angelo Mai from cod. Vat. gr. 1553 (saec. X) under the title Λεοντιου πρεσβυτερου και Ιῶαννου τῶν Ηιερῶν βιβλιον δευτερον.  This includes 24 testimonia coming from the lost books of CI, among which inadvertently appear 1 from book V (instead of XV?) 2 from book 7 (instead of XVII?), while on the other hand 2 excerpts from book XIV are taken for passages from book IV or book VI.

If the nearly unanimous ascription in the manuscripts of the Hiera to John Damascene is correct, there must have been accessible in the middle of the 8th century in Palestine a copy of CI which extended as far as book XVIII.  Since the editions of Lequien and Mai are inadequate, new and more reliable material could be obtained from a future critical edition of the Hiera.

5. Florilegium Patmiacum (11th century) 

The Florilegium Patmiacum belongs to the Florilegia Sacro-Profana and was edited by Étienne Sargologos from five manuscripts, among which the Patmiacus gr, 6 (saec. XI) is the oldest and most important.  In 56 chapters it offers excerpts on the theme of Christian living, mainly taken from other Florilegia sacro-profana and different recensions of the Hiera, and from catena, from Christian-ascetic, Jewish and pagan authors.  The anonymous compiler was dated by the editor to the 11th century.

In the selections the author of the excerpt, but not the title of the work, is usually indicated.  The editor added the missing titles from 21 manuscripts of other compilations, in which he found the same excerpts.  Altogether the florilegium contains 8 excerpts from CI, 6 of them from the lost books.  Of these one each is also printed by Lequien and Mai from the recensions of the Hiera.  Sargologos has identified five of the six remaining passages in unpublished manuscripts of different recensions of the Hiera.  Since the material is not taken directly from the CI, but via the Hiera, the Florilegium Patmiacum offers no further data to indicate the date at which books XI and after of the CI were lost. 

Catena and bible commentaries

1. The so-called Commentary on Mark of Victor of Antioch (CPG C125)

The edition of Pierre Poussin mixes the two recensions of the Commentary together and with other manuscripts.  The single scholion in the edition that comes from the CI is found only in cod. Vat. gr. 1692 (saec. XIII) on Mk 15,23. This manuscript contains an anonymous catena on Mark, whose author is not identical with Victor of Antioch.  Joseph Reuss places the emergence of this scholion at the beginning of the 6th century.

2. Catena on the letters of Paul, Typus Vaticanus (CPG C160)

The part of a catena on 1 Cor. printed by John A. Cramer from Paris. gr. 227 (saec. XVI) comes from the 10th century, based on an anonymous catena of the 7-8th century.  In this a quotation from book IV of the CI is found against 1 Cor. 15:24.

3. Catena of Ps.-Oecumenius on the letters of Paul (CPG C165)

These catena do not go back to Oecumenius, but in their original recension come from an anonymous writer at the end of the 8th century and were extended by another anonymous writer in the 9th or 10th centuries.  One of the sources is the Typus Vaticanus above.  Altogether at least five recensions of the catena of Ps.Oecumenius are known.

The two quotations from CI printed by Migne come from the very corrupt cod. Paris. gr. 219 (saec. XI), the only representative of an independent, extended Typus of these catena.  One is on 1 Cor. 15:24 and is taken from book IV of CI.  It is identical to the quotation on the same passage in Typus Vaticanus (above).  The other is against 2 Cor. 4:4, and comes from book I of CI.

John A. Cramer has published the extended Typus of catena from cod. Paris. gr. 223 (saec.XI).  He prints a lemma (also on 2 Cor. 4:4) which refers to book I of CI.  The text given is not there, however, and obviously does not relate to the passage in question.  The associated text is thus not in the manuscript or the edition.  Perhaps however it concerns the same quotation which cod. Paris. gr. 219 gives on this same passage.

Of the quotation from CI on 1 Cor. 15:24, which he edited from the Typus Vaticanus (above), Cramer gives only the lemma from cod. Oxon. Bodl. Roe 16 (saec. X), which belongs to the same normal type of the Catena of Ps.Oecumenius. 

4. Catena of Andreas on the Acts of the Apostles (CPG C150)

The edition of Cramer follows the cod. Oxon. Coll. Novi 58 (saec. XII) and offers some additions from Paris. Coisl. 25 (saec. X) . The Andreas mentioned in the latter manuscript is probably not the catenist, but merely the scribe.  This is the only indication of date in this manuscript.  Altogether there are three quotations from CI.

5.  The Commentary on Revelation of Arethas of Caesarea

Arethas wrote his Commentary on Revelation after becoming a bishop, probably not before 913.  For long stretches Arethas reproduces the commentary of his predecessor Andreas of Caesarea (bishop between 563-614).  The edition of the Commentary of Arethas by John A. Cramer is based on the cod. Oxon. Baroccianus 3 (saec. XI). This contains the only reference in the work to CI.  In the commentary on Rev. 7:8, the commentator quotes Lk. 23:34 and contradicts Cyril, who in book XIII of the CI denies the authenticity of this verse.  Since the passage is missing from the commentary of Andreas, it clearly belongs to the commentary of Arethas alone.

6.  Catena of Nicholas Muzalon on Isaiah 1-16 (CPG C62)

After his resignation as Archbishop of Cyprus, Nicholas IV. Muzalon was an abbot in Constantinople around 1110, and from 1147-1151 patriarch of that city.  While probably still archbishop of Cyprus he compiled a catena on Isaiah (as far as chapter 16), of which only the foreword has been published.  In the unpublished portion of the catena he quotes at least once from CI.

3.4 Copies and locations of copies of CI

From the above remarks it follows that at least three (complete?) copies of CI once existed.

Since Cyril of Alexandria dedicated his work against Julian to the emperor Theodosius II, a copy of the work must have come to Constantinople, and into the palace.  A further copy obviously remained in the patriarchate in Alexandria.  To the knowledge of Theodoret, a third copy was sent by Cyril to John of Antioch, to "show it to the teachers in the Orient" (see above, p.36f).

The Florilegium Cyrillianum must have been compiled by Chalcedonian circles from the copy of Cyril's work in the patriarchate at Alexandria around 482, perhaps even earlier as an answer to the florilegium of Timothy Aelurus in the third quarter of the 5th century (see. p.40 f).

Between 508 and 511 Severus of Antioch had at Constantinople access to a copy of CI, from which he corrected and supplemented the testimonia of the Florilegium Cyrillianum (cf. p. 40f).  Due to his good relations with the court under the emperor Anastasius it may be accepted that he made use of the complimentary copy in the palace library.  In what form he had access to CI during the compilation of the Apology for the Philalethes in his exile in Egypt, whether in direct or indirect form, it is impossible to say.

Regarding the copy which Cyril had sent to John of Antioch, it remains unclear whether in the region of Antioch other copies were made for the use of Eastern "Church teachers" and whether Theodoret returned the copy when it was sent to him.

A copy of CI appears in 682 at the 6th ecumenical council, once again in Constantinople, this time in the patriarchate (see. p. 41f).  There is no reason to suppose that it had arrived here from the palace in the meantime.  More likely Cyril sent a copy not only to the emperor and to John of Antioch, but also to the patriarch of the capital.  Previous to the council, the Greek compilers of the dyotheletic florilegium probably working in Rome must also have had access to CI.  Since they only included two passages, it cannot be said with certainty whether they had access to the text through the direct transmission (cf. p. 41f).

John Damascene often quotes from the CI in the Hiera (cf. p. 43 f).  Because of their frequency and self-sufficiency, these excerpts may originate from the direct transmission.  Since they extend as far as book XIX ( = Frg. 49 Neumann), the conclusion is that Damascene in the first half of the 8th century had access to a copy of CI containing four pentads.  Whether this copy belonged to the library of the patriarchate of Jerusalem, or the Convent of Mar Saba, or whether it was in the patriarchate of Antioch is impossible to say, any more than whether it was the same copy sent by Cyril to John of Antioch. 

To the three certain copies, in the patriarchate of Alexandria, in the imperial palace at Constantinople, and in the patriarchate of Antioch, we may probably add two more in the patriarchate of Constantinople and in Palestine, and perhaps a further one in Rome.  Considering the length of the CI, it is not very likely that very many copies ever were made.

3.5 The organisation of CI

The occurrence of the different lemma-types in the direct and indirect transmissions is apparent from tables 2-4.  For the direct transmission of the text only codd. GV and FME together with Q are considered relevant.  For reasons of clarity the locations of the lemmata in the works for the indirect transmission are not indicated in detail.  

Four different types of book organisation are found:

1. Type A only indicated pentads (τευκηὲ).
2. Type a only counted all the books in numerical sequence (τομοι, λογοι, βιβλια, κεφαλαια), without indicating the pentads.
3. Type Aa indicated the pentads (βιβλια or βιβλοι) and sequential counting of all books (λογοι).
4. Type Ab indicated the pentads (βιβλια), but counted books within each pentad as 1-5 (τομοι or λογοι).

These findings permit the following conclusions:

Direct transmission (see table 2)

1.  Codex G can tell us nothing about the organisation of the book.  The manuscript is mutilated at the beginning of the prosphonema and at the end of book I.  It offers neither an explicit for the prosphonema nor an incipit for book I.  However space is left for the latter, as for the first initial.  Clearly the incipit and the initials were to be rubricated later, which, however, did not happen.

2.  V is uncertain, since it contains only the first five books.

3.  The remaining relevant codices of the direct transmission (FMEQ) follow type Ab.  The terminology is constant: books as βιβλια; pentads always as τομοι.

4.  However two of these manuscripts (MQ) exhibit in the second pentad as well as type Ab a different numeration:

In cod. M books VI to IX are also numbered as for type a (VI becomes βιβλιον, VII-IX designated as τομοι).  This addition numeration is in another hand, and at the edge of the folios.  It is secondary.

In cod. Q books VII-IX are numbered in Greek as for type Ab, but also in Latin following type a ("Liber VII", etc).  This numbering is likewise secondary.

5.  The subscription of book X (τελος) in F and its incipit also suggest that certainly for codd. FM and probably for codd. EQ the (indirect) ancestor copy ended here.  The (indirect) ancestor of cod. V contained -- according to the closing subscription τελος -- only five books.  From this we can infer, using all the manuscript witnesses, that the text was originally delivered in codices, which each contained a pentad.  In the one case there were two codices, in the other only one.

Indirect transmission (see tables 3-4)

1.  Only the Christological florilegium of Vatopedi and the Catena of Nicholas Muzalon follow the type Ab of the direct transmission.

2.  The oldest witnesses, the Florilegium Cyrillianum (before 483) and the dyotheletic florilegium of the Roman legates to the Third Council of Constantinople (680/681) follow type Aa.  The designate the pentads with βιβλιον or βιβλος, the "book" as λογος.  It is unclear whether the indication of the pentad was only introduced secondarily by the florilegist for practical reasons, or whether it was present in the lost manuscripts of the direct transmission that they used.  Certain indications suggest the latter.  On the other hand Cyril himself repeatedly refers to the individual books of his work as λογοι.  However the arrangement into groups of five, running through the numbered Logoi (type Aa) must come from the manuscripts in the patriarchates, and so in the long run from Cyril himself.

3.  The following florilegia and catena observe type a, in which the "books" are indicated as λογος or βιβλιον or τομος only with a cardinal number:
- the Florilegium Rupefucaldinum and the Recensio Vaticana of the Hiera, which both go back to the Hiera (Sacra Parallela) of John Damascene;
- the Florilegium Patmiacum, which probably also draws its quotations from CI from the Hiera;
- the so-called Commentary on Mark of Victor of Antioch;
- the catena on the letters of Paul of the Typus Vaticanus;
- the Commentary on Revelation of Arethas. 

The catena of Andreas on the Acts of the Apostles only follows type a (τομος Β̕) in one of three quotations from the CI, while the other two follow type A.

4.  Type A is thus represented only twice, in the catena of Andreas on the Acts of the Apostles.  Book IX and X are labelled τευχος; the word τευχος is met here only.

3.6 Conclusions

1. Type a is probably a modification of type Aa.  This leaves us with types Aa and Ab, leaving aside the very rare type A.  Type Aa can be traced back to the 5-6th century and has an Alexandrian origin (Florilegium Cyrillianum). This type is also met with in Constantinople in 681-2.  Type Ab is uniformly present in the direct transmission of the first two pentads.  This suggests strongly that the direct transmission goes back to a copy which cannot have been taken directly from the copies in Alexandria and Constantinople.

2.  The fragments we have extend as far as book XIX.  It can be concluded that the books were bound divided into four pentads and each pentad had its own codex.  From the original four codices two were lost before the start of the direct transmission.  Therefore there is evidence for four pentads, which probably included twenty books.

3.  The indirect tradition prior to 483 (Florilegium Cyrillianum) is only using four pentads.  If a fifth and sixth pentad had existed, as Neumann assumed, they would probably have been quoted since they would probably likewise have contained material which would be usable in controversy against the monophysites.  Since the florilegium was probably made in the secretariat of the Alexandrian patriarchate where the Opera Omnia of Cyril might still have been available, a few decades after his death, the existence of books XXI-XXX is improbable.  Also Severus, who used a copy of CI around 510 in Constantinople, does not mention these books anywhere.  The same applies to the 6th ecumenical council  of Constantinople in 681-2, at which a copy of CI was consulted that contained the third pentad at least, and for all later works which contain quotations from CI.

3.7 Date of the loss of books XIff.

Terminus post quem: the last florilegium, which quotes from books XI-XIX, is the probably 11th century Florilegium Patmiacum (from book XVI).  However the compiler draws these quotations from older compilations.  It seems safe to say that John Damascene († vor 754) was still able to make use of the third and fourth pentad of CI during the composition of the Hiera, since he quotes from books up to book XIX extensively. 

Terminus ante quem: the beginning of the direct transmission (13th century) probably forms the terminus ante quem for the loss of the third and fourth pentads.  The scribe of cod. M (fol. 109r; 14th century) remarks:

Σὲμειῶσαι ὀτι και δευτερον βιβλιον ὀ αλιτὲριοσ (sc. Julian) εγραψε, και ζὲτὲτεον και τας προς αυτον του πατρος (sc. Cyril) αντιρρὲσεις.

Bessarion, to whom this codex belonged, from which he had cod. N copied, remarks on fol. 153r:

Ταυτα δε (sic) εισιν τῶν κατα τὲς ιταλαιας διατηὲκὲς Ιουλιανου λογῶν αντιρρὲσεισ μονον. Τα γαρ προς τα κατα του αγιου ευαγγελιου ειρὲμενα, ουχ ευρεταισ πῶ.

The third and fourth pentads were thus lost between the 8th and 13th centuries.  Through the publication of fragments of books XI-XIX from florilegia and catena, we are in a somewhat more fortunate position than the cardinal of the 15th century.

The full article with footnotes can be found in: Adolf Martin Ritter/Wolfgang Wischmeyer/Wolfram Kinzig (Hrsg.), ... zur Zeit oder Unzeit. Studien zur spätantiken Theologie-, Geistes- und Kunstgeschichte und ihrer Nachwirkung. Hans Georg Thümmel zu Ehren, TASHT 9, Mandelbachtal/Cambridge 2004, 29-62.

Posted originally at Academici without Greek (supplied later)

Constructive feedback is welcomed to Roger Pearse. Corrections and additions are very welcome.