The Julian Poems of C.P. Cavafy G.W. Bowersock
The publication of five poems, hitherto unknown, by the poet Cavafy all on the theme of Julian the Apostate and the projected edition of his reading notes on Gibbons Decline and Fall illuminate not only the poets obsession with Julian but also the principles of his craft.1 It is perhaps not too presumptuous for an historian to attempt an assessment of Cavafys work on Julian in the light of the rich new material, for he is said himself to have declared that he was an historical poet: ... E .2 This remark finds confirmation in a critique from May 1927, written earlier by the poet or by a sympathetic associate, with a threefold categorization of Cavafys oeuvre into H, I, and .3 Of the sensual and historical categories it is said that there is sometimes so great a degree of overlapping that classification becomes difficult but not impossible.4 Several of the Julian poems prove this point.
     The new poems may now be added to the group of Julianic pieces already known, seven in number. The grand total of Cavafys poems on the apostate emperor is therefore twelve. It becomes evident that no other historical topic preoccupied this writer to such an extent as Julian, and any interpreter is bound to ask why that emperor held so great a fascination. The question is particularly important since it has long been clear from the previously published poems that Cavafy did not much care for Julian. He shared none of the late romantic admiration for the last of the pagan rulers. Cavafy appears to have been obsessed with removing the glamour and exposing the fraud of this hero of latter-day pagans. There seems to be a paradox in so hostile a treatment from a Greek who was among the first in modern times to write brilliant poetry on sensuality and sexual encounters. At issue here is precisely the blending of sensual and historical matter.
     Most of the Julian poems can be dated, although only five of them were actually published in the poets lifetime. The dates are revealing:5
1. O I M
   written November 1896, published posthumously
2. M
   probably written March 1917 as a revision of a poem from September 1892,
   published August 1926
3. A
   written April 1920, unfinished and unpublished hitherto
4. O
   written May 1920, unfinished and unpublished hitherto
5. O I,
   published September 1923, no exact date of composition
6. H I
   written December 1923, unfinished and unpublished hitherto
7. O I N
   published January 1924, no exact date of composition
8. Hunc deorum templa reparaturum
   written March 1926, unfinished and unpublished hitherto
9. O I A
   published November 1926, no exact date of composition
10. O
   published January 1928, no exact date of composition
11. Untitled poem beginning E
   no exact date of composition, unfinished and unpublished hitherto
12. E A
   written c. November 1932 and April 1933, published posthumously.
     It is clear from the foregoing list that only two of the twelve poems on Julian can be dated before 1920; of these two one was probably reworked as late as 1917, and even this was kept from publication until 1926. As far as can be told, all the other poems were composed between 1920 and the poets death in April 1933. It would scarcely be an exaggeration to say that the last decade and a quarter of his life were the period of his greatest involvement with the life of Julian.
     And yet the interest had surfaced early. The satirical account of Julians fright at the mysteries and the potent sign of the cross which Julian made by reflex is anchored to November 1896, the very time in which Cavafy was engaged in his critical reading of Gibbons Decline and Fall. As Diana Haas has admirably shown in her examination of Cavafys annotations to Gibbon,6 the decade of the 1890s was important for Cavafys mastery of Greek history in the Roman and Byzantine ages. She has shown how carefully he compared points in Gibbon with parallel discussions in Paparrigopoulos history of the Greek people, and she has drawn attention to his particular concern with the early Christian church. His repudiation of Gibbons snide remark on Gregory of Nazianzus evokes his positive judgment of Gregory in an article on Byzantine poets, published in 1892.7 Moreover, it seems likely from the extended citations of Gregory and Theodoret in the annotations to Gibbon, now that these can be seen in full, that Cavafy had access to original texts of at least both ecclesiatical writers at that time.
     Cavafy must have found that his researches into the early Church spoke to some degree to his own personal needs. He was drawn to Christianity but found his greatest solace in the 1890s not so much in the Church itself as an institution as in the solitary struggles of its pioneers. Few could read his annotation in Gibbon on Simeon Stylites without appreciating the deep feeling with which he wrote it: This great, this wonderful saint is surely an object to be singled out in ecclesiastical history for admiration and study. He has been, perhaps, the only man who has dared to be really alone.8 The impress of this figure can be seen in the poem on Simeon from July 1917. From a thematic list of his early poems Cavafys interest in the early Church can be readily inferred. Under the rubric A A X he lists a group of poems, largely lost, but including the extant works on Julian at the mysteries and Simeon.9 The lost poems all date from the 1890s.10
     In these years and on into the first decade of the twentieth century Cavafys experience of Christianity was complicated by feelings of guilt and distress over his sexual nature, which he did his best to confront alone. He wrote a series of private confessions about his solitary erotic passion. There is material, still unpublished, which confirms that Cavafy was genuinely tormented by what was apparently frequent masturbation.11 By 1911, however, he seems to have reconciled himself to his homosexuality, to have sought partners in fulfilling it, and to have resolved not only to write about it in his poetry but actually to publish the poems.12 It is with T , published in November 1911, that Cavafy publicly declared himself a sensualist. Although the poem is not yet explicitly homosexual as later poems were to be, it is nonetheless a striking departure for Cavafy. It is especially notable for the conjunction of his historical interests with his advocacy of sensuality. The speaker is a young Syrian in the reign of Constans and Constantius, therefore precisely between A.D. 340 and 350, the years of Julians adolescence. These were the years in which Julian was raised a Christian and became a pagan, the years to which Cavafy devoted three of his later Julian poems (nos. 4, 6, and 7 in the foregoing list). The young Syrian of this epoch is described as partly pagan and partly christianized: , . He proclaims that he will not fear his passions, he will satisfy his most daring erotic proclivities. He repeats that he will not be afraid because he is confident that if he is called upon to be ascetic he will have the power to be so. The appearance of this poem and the end of the confessional notes mark a new stage in Cavafys life and oeuvre. He is moving, with the help of historical analogues, toward a reconciliation of his sexuality and his Christianity. The Syrian in T was partly Christian but still sensual, just as the Christian Myris in a poem of 1929 set in Alexandria of the year A.D. 340 had rejoiced in the love of a pagan. Inevitably Cavafy would have asked himself what impact Julian would have had on the Greek world of that Syrian youth or of Myris lover. This was a world in which pagans and Christians could associate easily with one another in unhindered pursuit of the sensual life. It was the avowed aim of Julian, the ascetic pagan, to put an end to all that.13
     It may be argued, therefore, that Cavafys return to his historical interests of the 1890s was an important part of his adjustment to his homosexual nature as well as to his Christian sympathies. The adjustment, which began in 1911, led to the elaboration of the erotic category of his work and its conspicuous overlapping with the historical category. The newly published poems contribute substantially to enlarging our understanding of the way in which Cavafy worked on his historical poetry, and it will be helpful to proceed to a close look at these unfinished pieces for what they reveal about both Cavafys preoccupations and his methods. What is learned can serve in turn to provide a more profound interpretation of the Julian poems already known.
     Cavafys remarkable note on the Athanasius poem is a major addition to our knowledge. Here, on the drafts of a work of April 1920, he appended a note over nine full years later to the effect that he could not find the ancient source for the incident on which his poem was based.14 He had found in Butchers history of the Egyptian church the story of the two monks who learned of Julians death by extra-sensory perception while they were in a boat with Athanasius on the Nile.15 It was a good story, well suited to Cavafys manner; but when after nine years he had not located it in the Patrologia of Migne, either in volume 67 or in volume 82 (to which he obviously had access by 1929), he declared that unless the source could be traced somewhere else the poem could not stand.16 So firm a commitment to an historical source must be rare indeed in the annals of poetry.
     There are other examples of similarly scrupulous scholarship in Cavafy. Among the new poems is one on the saving of the infant Julian during the massacre of his family after the death of Constantine the Great. As Renata Lavagnini has perceptively pointed out, the kernel of this poem is the confrontation of Julians salvation at the hands of Christians with his own later ingratitude as expressed in the remark (from Julians oration to Helios the King).17 Cavafy first drafted a poem involving the salvation of Julian together with his half-brother Gallus. He may, as Lavagnini suggests, have worked under the influence of the wording in Gregory of Nazianzus, but he may equally have read hastily in Allards Julien IApostat, Mais cette protection [against Constantius soldiers] naurait peut-être pas suffi à les sauver.18 In any case, while studying Allards pages more closely, he observed that Allard mentions Julian alone as having been rescued by the Christians: Des hommes dévoués enlevèrent secrètement Julien. Cavafys note, O Allard I, explains why he undertook to revise his poem so as to record the salvation of only one prince instead of two. He was desperately keen to be historically accurate.
     Even in the titles of his poems Cavafy strove for exactitude. The folder containing his unique poem on Julians career as a commander in Gaul bears the Latin title, Hunc deorum templis. Presumably he was writing from memory what he intended to be a quotation from Ammianus account of the blind old woman in Vienne, but when he took the trouble to check the text of Ammianus he discovered that he had erred. He then revised the title to give the correct Latin quotation, Hunc deorum templa reparaturum.19
     A comparable scrupulosity over a title can be detected in the poem O I M, which had previously borne the title O I E.20 It has seemed clear to most readers that the initial inspiration for this work of 1896 was Cavafys study of Gibbon at that very time. The episode of Julians making the sign of the cross when he encountered demons in an underground cavern occurs in Gregory of Nazianzus, whose original text was in all probability familiar to Cavafy.21 But it was Gibbon who inferred from Gregory that Julian was at Eleusis: He [Julian] obtained the privilege of a solemn initiation into the mysteries of Eleusis....22 Hence the title O I E. With the indisputable evidence we now have of Cavafys study of Allard in regard to the massacre of Julians family, it becomes almost certain that his study of the same author led to his alteration of the title of his earliest Julian poem. Allard argued at length against the supposition that Julian was initiated into the Eleusinian Mysteries: Les historiens modernes disent presque tous que Julien fut alors initié aux mystères dÉleusis. Cela ne résulte pas clairement du texte dEunape.... Il me semble difficile que Julien ait Allard été initié.... Nulle part il [Julien] ne laisse entendre quil ait reçu linitiation dEleusis....23 Hence a new title for an old poem.
     In what is perhaps the most memorable of the five new poems, the evocation of the boy Julian on the plains of Troy in the company of the Christian Bishop Pegasius, the materials published by Renata Lavagnini again prove the poets scrupulous concern for scholarship. The primary inspiration in this case must have been Allard, who saw well the implications of Julians own spare testimony about the episode from his youth: On peut, sans un trop grand effort dimagination, se figurer létat des deux personnages que le hasard avait ainsi mis en présence. Julien, aux allures de levêque, a surpris ses pensées secrètes: il attache sur lui un regard pénétrant et lui pose des questions captieuses. Pégase connaissait sans doute par la renommée les vrais sentiments de Julien....24 Although Cavafy undoubtedly knew the actual letter of Julian in which the meeting with Pegasius (who later converted to paganism) was described, he found in Allard a congenial speculation about the nature of the encounter. The boy and the man, both nominally Christians, questioning each other amid the pagan shrines of Troy, would have sensed indirectly one anothers true disposition. To the religious ambivalence of the scene and the hypocrisy of the players Cavafy added in the work of 1920 a palpable atmosphere of paedophilia by insisting on hidden revelations which the boy and the man surely divined in each other. This was a theme ideally suited to Cavafys taste and talent; it blended almost perfectly the categories of sensuality and history. And yet Cavafy troubled, ten years after working on the poem, to re-examine the incident at Troy by copying out Bidezs treatment of it in La vie de lEmpereur Julien, which appeared in 1930.25 So once again we can see Cavafys passion for scholarly acceptability.
     In a well-known remark, cited by Malanos, Cavafy is said to have observed that two of his poems remained incomplete for lack of a copy of Gregory of Nazianzus.26 From what can now be seen of his methods such a remark is by no means as implausible as it might once have seemed. Furthermore, among the seven pieces on Julian which he chose not to publish (nos. 1, 3, 4, 6, 8, 11, and 12 in the present listing) exactly two depend upon Gregory as the principal ancient source. These are no. 1 on Julian at the Mysteries and no. 6 on the salvation of the infant Julian from the murderous soldiers of Constantius. While it has become clear that Cavafy paid close attention to modern scholarly writing, it has become no less clear that he insisted on verifying his facts by reference to the ancient texts. Just as this means that it would be wrong to speak of a single source for one of his poems (such as Gibbon, Allard, Gregory, or Julian himself), it is wrong to discredit the importance he assigned to the original texts for poems initially inspired by secondary sources. The poem on the mysteries was certainly the result of reading secondary literature (Gibbon), and the poem on the saving of Julian may have been (Gibbon again, or an overly hasty first reading of Allard). In any case, it was Allards work which drew Cavafys attention to historical difficulties in his treatment of the two incidents, and he might naturally have wished to check Gregory, the ancient source for both, before proceeding. It seems likely from Cavafys allusion to the Patrologia of Migne in the note attached to his draft on Athanasius that at least by November of 1929 he had access to the volumes of Migne, which would have included Gregorys invectives against Julian. If the remark quoted by Malanos is genuine, it must accordingly belong to a time before November 1929 but probably after the work on H I in December 1923.
     The five poems on Julian published in Cavafys lifetime (nos. 2, 5, 7, 9, and 10) were presumably released to his public because they met his scholarly criteria. No less than three of the five include verbatim quotations from the surviving writings of Julian himself. These are no. 5, which opens with a substantial citation from the emperors letter to Theodorus on the neglect of pagan gods;27 no. 9, to which Cavafy affixed a quotation from the Misopogon on the Antiochenes predilection for Christ and Constantius;28 and no. 10, which is in fact little more than a repetition of a comment by Julian preserved in the ecclesiastical historian Sozomen together with the Christians witty reply to him (also as given by Sozomen).29
     The main point of no. 7, mocking Julians reading of the Scriptures at the very moment of his growing infatuation with Neoplatonist theurgy, depends upon the testimony of Gregory of Nazianzus.30 While it is probable, as Diana Haas implies, that Gibbon was the primary inspiration for this poem and that some of the introductory material about Julians teachers derives ultimately from Eunapius and the ecclesiastical historians, the appearance of the young prince as a lector in Nicomedia is told only by Gregory.31 To judge from Cavafys procedure elsewhere, it is highly unlikely that he would have released this poem without satisfying himself that it conformed to Gregorys account. Curiously, the two poems which appear to owe their inspiration to Gibbon, namely this one (no. 7) and the piece on the mysteries (no. 1), both depend upon Gregory as the principal ancient source. Since Cavafy seems to have had access to Gregorys work in the 1890s when he was reading Gibbon (and when he is known to have written the poem on the mysteries), it is tempting to suppose that the poem on Julian in Nicomedia, although published in 1924, may have been drafted or outlined at that same time. The long delay until publication would closely parallel the fate of the poem on a great procession of priests and laymen (no. 2), published in 1926.
     That poem poses more problems than the four other Julian pieces which Cavafy chose to make public. In view of the prominence given to the cross it is reasonable to identify the work with a version of O listed as one of the early poems on A A X. This early poem on the cross was probably written in September 1892 and then revised in March 1917 only a few years before Cavafys most conspicuous period of preoccupation with Julian began.32 The subject reflects the poets immersion in ecclesiastical history during the 1890s. Its description of the Christian celebration at Antioch after Julians death depends principally upon Theodoret but may also reflect passages in Sozomen and Gregory.33 The actual ceremony is imaginary, but the spirit of the Antiochene Christians at the time comes through memorably and accurately. Of course Jovian himself was to have trouble with the Antiochenes just as Julian had, but it was hardly to Cavafys purpose to dwell upon that particular irony. The salutation of Jovian with which the poem concludes is unsettling for anyone acquainted with the history, and Cavafy certainly was acquainted with it. But the historicity of the poem remains intact and provides the occasion Cavafy wanted for exalting what really obsessed him here, the cross, which is equally important as a symbol in another poem from the 1890s, O I M. It was not, however, until nine more years had passed after the revision of O that Cavafy could bring himself to publish M . But it is worth emphasizing that it appeared in the very same year as another poem on Julian and the Antiochenes (no. 9) and preceded it by just a few months.
     The recurrent motifs of the cross and of Antioch lead one to observe that all of Cavafys Julian poems are restricted to a relatively small number of topics from the life of the emperor. No less than three pieces (nos. 2, 9, and 12) are explicitly concerned with the reaction of the people of Antioch to Julian, and a fourth (no. 5) is based on letters written from Antioch. The fact of Julians death is the basis of three poems (nos. 2, 3, and 11), of which two (nos. 2 and 3) depict an immediate response to the news, while the third (no. 11), in dating the scene to the first year of Theodosius solely by an allusion to the number of years after Julians death, attempts to suggest how rapidly the emperors posturing became old fashioned and absurd. Four of the poems are devoted to Julians childhood (nos. 1, 4, 5, and 7), and a fifth (no. 8) is based upon a report from the very first year of the young princes public career (A.D. 355). Indeed, poem no. 8, with its setting in Gallic Vienne, is a reminder that Cavafy nowhere else exploited an episode from the six years that Julian spent in Gaul.
     The omission of the crucial years in Gaul, apart from the legend about his arrival there, cannot have been due to the poets ignorance. He was far too avid a reader of Gibbon and Allard, not to mention Julian himself. It is understandable that Cavafy might not have found inspiration in Julians battles against obscure tribes and chieftains, but it is not immediately obvious why he took no interest in Julians dependent relation upon Salutius or in the strange episode of the proclamation of Julian as emperor at Paris.34 On balance it seems best to assume that the poet had little taste for the western ambience of Gaul and preferred to concentrate on Julian in a Greek context. But even this hypothesis does not explain another notable omission in the work of one who knew the history so well. There is nothing in Cavafy about Julian in Constantinople during the period after Constantius death and before the journey to Antioch. This was undeniably a Greek environment, and there were rich materials for Cavafys genius for example, Julians ostentatious appearance as a mourner at Constantius funeral.35 The travesty of justice at the Chalcedon trials might also have engaged Cavafys muse: after all, even Ammianus Marcellinus, who admired Julian, had to admit that Justice herself shed tears on that occasion.36
     With most poets it would not be very profitable to ponder historical subjects they choose to omit. But for Cavafy, an avowedly historical poet who was obsessed with Julian, it is worth taking into account that, in spite of turning to Julian twelve times, he concerned himself with a rather small number of topics from the range of those that were possible. They are, as has already emerged, Julians childhood, Julian at Antioch, and Julians death. The common denominator for every single one of the poems what links the principal motifs together is Christianity. All of the poems, in one way or another, address the issue of Julians encounter with the Christians. It seems evident that this was all that Cavafy really cared about in Julians career. Weaknesses of character which the emperor displayed toward other pagans, as in Paris or in Constantinople, held no fascination for him. Similarly the dramatic effort which Julian made to rebuild the temple at Jerusalem and the equally dramatic termination of that effort amid miraculous balls of flame appear nowhere in Cavafys oeuvre.37 Christianity was the real obsession.
     The characteristics of Julian which Cavafy chose to underscore are hypocrisy (nos. 1, 4, 6, 7) and puritanical intolerance (nos. 5, 9, 10, 11, 12). The sources provide ample justification for characterizing the emperor in this way, even if many writers have preferred different assessments. Julian was an ascetic who demanded strict adherence to the principles of his new pagan church. He had learned about church organization from his upbringing as a Christian, and it is clear from his writings that he intended to surpass the Christians at their own game.38 Meanwhile the Christians of Julians time were, after all, for the most part yesterdays pagans. They had not changed their way of life all that much.39 At Antioch they went to the theater and the chariot races, and they celebrated their festivals as they had before. When Julian entered the city he heard the ill omened wailing of the festival of Adonis.40 Cavafy understood all this rather better than most historians, in part, of course, because Christian Antioch with its traditional style of life was the kind of city in which he longed to live. He had no problem with the old paganism, robust and free (at least in his view). His poem I, first composed in the same year as O I M, shows a deep feeling for Greek paganism. It was Julians anti-Christian paganism that Cavafy could not abide. What Cavafy praised in I he could find among the Christians of Antioch. The city was a profoundly important symbol for him: its people were immoral (), but their life was delectable (). And they were Christian.
     They were also Greek. Cavafys , written in the year 1927 and therefore right in the midst of his Julian poems, provides the appropriate commentary. The city of Antioch is proud of its buildings and streets, its kings, its wise men, its merchants. But it has far more to be proud of: it is Greek from ancient times, . It has often been observed that the Greekness of Cavafy is that of the diaspora Alexandria, Constantinople, and (in imagination) Antioch. Greeks of the mainland have tended to admire Julian as their courageous if unsuccessful champion, and on the whole they still do. A young Athenian scholar of today, who has done considerable research on Julian and will shortly publish a book on him, has recently written, Kavafy ... chose to use his poetic gift in order to turn his most private predilections into an art theory, and in Julians personality he discovered the great negative symbol that would convey to initiates his cult of the Beautiful. Identifying with the Antiochenes of the fourth century, the twentieth-century Alexandrian despises, almost hates, the ascetic emperor for his superior indifference to all that glitters, even gold... Theirs [the Antiochenes] was certainly not a perfectly genuine form of Hellenism, and Julian was not the first intellectual to denounce their .41 This writer goes on to commend the emperor as honest, compassionate, friendly, and fond of family life. Cavafy remains the better historian, but the issues are as alive now as they were in his day or in Julians.
     The Christians of his time could not forgive Julian for arrogating Hellenism Greekness to the pagan cause. By forbidding Christians to teach the great Greek classics he tried to cut them off from their crucial heritage. Gregory of Nazianzus was quick to protest and as eloquent as always: no one trained in Hellenic traditions could be denied being a Greek.42 Cavafy from Alexandria must have sensed a kindred spirit in Gregory, the Christian from Cappadocia. But the tide was turning. The very word would soon be the current word for pagan and would no longer serve for a Greek. Cavafy, like Gregory, belonged to the far-flung community of Greek culture beyond Greece and embracing Christianity. Cavafy was E; just as King Antiochus of Commagene in a poem of 1923: Y , E.43
     Permissive Christianity, then, appears to be the fundamental interest of Cavafy in handling the various Julian episodes. To be a Christian did not preclude being a pagan in the old sense, like the young Syrian in T , nor did it preclude a romance with a pagan like Myris lover. In Antioch Cavafy found the resolution of the problem he began to solve in 1911 when he started to make his erotic verse public. It was no accident that the historical and sensual categories of his oeuvre tended to merge at times, as the writer of May 1927 observed; for Cavafy was able to interpret his own eroticism in terms of historical examples that preserved for him what he probably found more important than anything else: his Christianity and his consciousness of being Greek. In the Julian poems he struggled for historical accuracy because it was clearly imperative for him to know that there really had been a world that could accommodate a sensualist, both Christian and Greek.
1. The five new poems appear in the present issue of BMGS in an editio princeps prepared by Renata Lavagnini. Also of use to me was Diana Haass unpublished article, Cavafys Reading Notes on Gibbons Decline and Fall. I am grateful to both these scholars for their generosity in showing me their work before publication; and I am likewise grateful to my friend and former colleague, George Savidis, for first drawing my attention to the new poems in the Cavafy archive. Savidis and I gave them their world première at a talk to the Shop Club of Harvard University on 20 Dec. 1979. Savidis son, Manolis Savidis, undertook, with his fathers permission and mine, to offer a study of the Julian poems as his term paper in my last Harvard course, Historical Studies B-11 (The Christianization of the Roman World), in the spring of 1980. His paper was full of valuable observations and citations, proving that a mastery of Cavafys work continues in the Savidis family. I am glad to acknowledge my debt to this paper. Finally I owe thanks to Edmund Keeley for his helpful comments on the present study as well as for more wide-ranging discussion of Cavafy.
2. G. Lechonitis, K , 2nd ed. (Athens, 1977), pp. 19-20.
3. G. P. Savidis, O K (Athens, 1966), pp. 209-10. Cf. the translation of this critique in Edmund Keeley, Cavafys Alexandria (Cambridge, Mass., 1976), pp. 186-7.
4. Ibid.
5. For the dates of the published poems, see the annotations by G. P. Savidis ad loc. in C. P. Cavafy: Collected Poems (Princeton, 1975), translated by E. Keeley and P. Sherrard, edited by G. P. Savidis. For the dates of the new poems, see R. Lavagninis edition in the present issue of BMGS.
6. Diana Haas hopes to publish her article in the near future.
7. Gibbon had referred to the loose invectives of Gregory, on which Cavafy responded, with accompanying quotations, No artist the word is not misplaced here had spoken so boldly before, For the article of 1892: C. P. Cavafy, (Athens, 1963), p. 49, as noted by Haas in her article.
8. Cited by Haas.
9. G. P. Savidis, ed., K.. K, A 1882-1923 (Athens, 1968), p, 172. On Cavafys Christianity, see R. Liddell, Cavafy (London, 1974), pp. 205-6 (wearing a cross around his neck, waiting for the patriarchal procession on Good Friday, the last sacraments) as well as Keeley, op. cit. (n. 3 above), p. 184. Savidis paper on Cavafys Christianity is fundamental, (Athens, 1973), pp. 115-20.
10. E.g., H K (1892), O or variants of this title (1892), (1892), O (1898).
11. O K (n. 3 above), p. 195. On the confessional notes, see also Liddell (n. 9 above), pp. 72-3. The unpublished material, in the possession of George Savidis, is cited by his son Manolis in stressing the importance of late 1911 as a turning point in Cavafys life. In April 1913, Cavafy turned 50. This seems to have been a mid-life crisis with liberating results.
12. On the nature of Cavafys work and sexual life from 1911, see Liddell (n. 9), pp. 155-71. The importance of the change in 1911-12 seems to be generally recognized. Cf. G. Seferis, A Poets Journal: Days of 1945-1951 (Cambridge, Mass., 1974), p. 139: Up to a fairly advanced age (maturity), Cavafy seems to remain at a very low level; he seems to be unable to rise above a certain very mediocre ceiling (as it is called in aviation, ceiling, plafond). What happens at and beyond a certain point? How does he cross that threshold? Heres a question that interests me not only about Cavafy but in general. Cf. Seferis , 3rd ed. (Athens, 1974), I, p. 328 on Cavafys work after 1910 as one and the same poem in progress.
13. Cf. Seferis, A Poets Journal (n. 12 above), p. 137: Julian represents a problem for Cavafy, he is a splinter on the horizon.... He is worse than a problem; he is a sort of illegal competitor.
14. Migne 67 ( ) 82 () Butcher. E , B A. A, .
15. E. L. Butcher, The Story of the Church in Egypt (London, 1897), I, p. 185.
16. In fact, the story is in MPG, XXVI, cols. 980C-81C.
17. Julian, 131A (p. 101 Budé).
18. Gregory in MPG XXXV, col. 549; P. Allard, Julien lApostat I (Paris, 1900), p. 263. It is worth noting that Gibbon, Decline and Fall, chap. 19 ad init. gives Julians age at the time of the massacre as six; Allard also gives it as six, but Migne, in a note, says seven. Socrates (MPG, LXVII, 369) and Sozomen (LXVII, 1213) both give the age as eight. Cavafys poem has Julian at six.
19. Amm. Marc. 15.8.22: Tunc anus quaedam orba luminibus... exclamavit hunc deorum templa reparaturum. In the long form of the title Cavafys handwriting is unclear: templa or templis could be read, but the presence of reparaturum seems decisive to me.
20. Cf. G. P. Savidis in Cavafys Collected Poems (n. 5 above), p. 430.
21. MPG, XXXV, cols. 577 ff.
22. E. Gibbon, Decline and Fall, ch. 23, Everyman ed., vol. 2, p. 367.
23. Allard (n. 18 above), vol. I, pp. 330-2. In her note 64 Diana Haas rightly raised the possibility of Allards discussion as the explanation of Cavafys change of title.
24. Allard (n. 18), vol. I, p. 347. Cf. Julian, Epist. 19, 79B (pp. 85-7 Budé).
25. Lavagnini, p. 76 (above).
26. T. Malanos, O K.. K (Athens, 1957), p. 123: , N A. Cf. Liddell, op. cit. (n. 9 above), p. 197.
27. Julian, Epist. 89B (p. 154 Budé): ... does not mean seeing contempt but rather seeing neglect (indifference, slighting) of the cults of pagan deities. Poem no. 5 is also based on Julians remarks to Arsacius in Epist. 84B (pp. 144-7 Budé).
28. Julian, Misop. 357A.
29. Sozomen 5.18.
30. Greg. Naz. in MPG, XXXV, cols. 551 and 632.
31. See notes 65 and 66 in Diana Haass article. Compare Gregorys words (MPG, XXXV, 551) and with Cavafys .
32. Cf. Savidis in Collected Poems (n. 5 above), p. 424; also A (n. 9 above), p. 172.
33. Theodoret 3.22. Cf. Sozomen 6.4 and Greg. in MPG, XXXV. cols. 708-12, all cited by Diana Haas in her note 69.
34. For Salutius, cf. G. W. Bowersock, Julian the Apostate (Cambridge, Mass. 1978), pp. 44-5; for the proclamation at Paris, ibid., pp. 46-54.
35. See Libanius, Orat. 18.120; Mamertinus, Panegyric 27.5; and Greg. Naz. in MPG, XXXV, col. 685.
36. Amm. Marc. 22.3.7: ipsa mihi videtur flesse Iustitia.
37. On this affair, Bowersock, op. cit. (n. 34 above), pp. 88-90.
38. See W. Koch, Comment lempereur Julien tâcha de fonder une église païenne, Revue belge de philologie et dhistoire, VI (1927) and VII (1928), published in four installments.
39. See G. W. Bowersock in Gibbon et Rome à la lumiere de lhistoriographie moderne, Publications de la Faculté des Lettres de lUniversité de Lausanne, XXII (1977), pp. 210-12.
40. Amm. Marc. 22.9.15.
41. Polymnia Athanassiadi-Fowden, JTS, XXX (1979), 331-2.
42. Greg. Naz. in MPG, XXXV, col. 536A-B. The partisan effort to identify Hellenism with paganism may well go back to the Neoplatonist teachers Porphyry (on whom Cavafy wrote an early poem, now lost [note 10 above]) and lamblichus.
43. Malanos reported that Cavafy once said to Stratis Tsirkas, E E. , , E, E: T. Malanos, K (Alexandria, 1935), p. 56. On this remark see E. Keeley, op. cit. (n. 3 above). p. 111.

G.W. Bowersock, The Julian Poems of C.P. Cavafy, Byzantine and Modern Greek Studies 7 (1981)