Hostname: page-component-cd9895bd7-p9bg8 Total loading time: 0 Render date: 2024-12-28T00:11:03.547Z Has data issue: false hasContentIssue false
  • English
  • Français

The occasion of Paul the Silentiary's Ekphrasis of S. Sophia

Published online by Cambridge University Press:  11 February 2009

Mary Whitby
Mary Whitby
Affiliation:
Oxford
Get access Rights & Permissions [Opens in a new window]

Extract

The ‘turgid archaisms’ of Paul the Silentiary's style have ensured that his two hexameter Ekphrases, describing the Emperor Justinian's sixth-century church of S. Sophia in Constantinople and its ambo, have lately attracted little interest, except among art historians who seek to extract nuggets of architectural information. On the other hand, the eighty or so pagan epigrams by Paul which are preserved in the Palatine and Planudean Anthologies have received attention in recent years both because of their literary interest and for the social and historical information which they contain. In my opinion, the much more substantial Ekphrases likewise deserve to be examined as literary and historical documents. The title Ekphrasis disguises the historical interest of the works: unlike the majority of the epigrams, these poems are no mere literary exercises, but official, public works, undoubtedly commissioned by Justinian, and delivered on specific and identifiable occasions. Only the central portion of the major poem describing S. Sophia comprises technical architectural ekphrasis; this is preceded and followed by panegyrical material appropriate to the occasion of recitation which together takes up almost half the total length of the poem (462 lines out of 1029). Here the Emperor Justinian, patron of the church, and Eutychius, Patriarch of Constantinople, are praised, and the events leading up to the occasion of the poem sketched. This topical part of the work provides evidence for the ceremonial which accompanied the poem's recitation and demonstrates the type of imperial propaganda pertinent to the end of the reign of Justinian. Moreover, an explanation for a limited number of stylistic flaws in a work which, by contemporary standards at least, is of high literary quality, may lie in the recognition that Paul was obliged to complete his poem in time for a specific occasion. It is with these three occasional aspects of the work that I shall here be concerned.

Type
Research Article
Information
The Classical Quarterly , Volume 35 , Issue 1 , May 1985 , pp. 215 - 228
Copyright
Copyright © The Classical Association 1985

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Majeska, George, ‘Notes on the Archeology of St Sophia at Constantinople: the green marble bands on the floor’, DOP 32 (1978), 299Google Scholar, referring to Paul's Ambo. Majeska echoes the unfavourable judgement of several earlier critics.

2 The standard edition is Friedländer, Paul, Johannes von Gaza und Paulus Silentiarius, Kunstbeschreibungen justinianischer Zeit (Leipzig and Berlin, 1912; repr. Olms, Hildesheim and New York, 1969)Google Scholar. New edition, using Friedländer's text, ed. Veh, O., in appendix to Procopius de Aedificiis (Munich, 1977)Google Scholar, adding German translation, and archaeological commentary by W. Pülhorn. English translation of the architectural ekphrases only by Mango, C. A., The Art of the Byzantine Empire, 312–1453 (Englewood Cliffs, New Jersey, 1972), pp. 80–96Google Scholar.

3 Edition of the epigrams, with Introduction, literary commentary and Italian translation, by Viansino, G., Paolo Silenziario Epigrammi (Turin, 1963)Google Scholar, with review by Cameron, Averil, JHS 86 (1966), 210 f.CrossRefGoogle Scholar; Averil, and Cameron, Alan, ‘The Cycle of Agathias’, JHS 86 (1966), 6–25, esp. 17–19Google Scholar, and ‘Further thoughts on the Cycle of Agathias’, JHS 87 (1967), 131CrossRefGoogle Scholar; McCail, R. C., ‘The Cycle of Agathias: new identifications scrutinised’, JHS 89 (1969), 87–96, esp. 94CrossRefGoogle Scholar; Yardley, J. C., ‘Paulus Silentiarius, Ovid and Propertius’, CQ n.s. 30 (1980), 239–43CrossRefGoogle Scholar; etc.

4 The importance of Paul's poem in these two respects has been recognised by Cameron, Averil, ‘Images of authority: elites and icons in late sixth-century Byzantium’, Past and Present 84 (1979), p. 9 and n. 25CrossRefGoogle Scholar.

5 Cf. Friedländer, pp. 109 f.

6 Procopius, , Bella 1. 24. 9Google Scholar, id.Aed. 1. 1. 21; Malalas, , Chronographia (ed. Dindorf, L., Bonn, 1831), p. 474. 17–19Google Scholar; Chronicon Paschale (ed. Dindorf, L., Bonn, 1832), pp. 621. 20–622. 2Google Scholar; Theophanes, , Chronographia (ed. de Boor, C., Leipzig, 1883), p. 181. 27–30Google Scholar.

7 Dedication in a.d. 537: Mal. 479. 21–2, Theoph. 217. 17–22. Earthquake of 557: Agathias, Historiae 5. 3–9 (ed. Keydell, R., Berlin, 1967)Google Scholar, Mal. 488. 20–489. 10, Theoph. 231. 14–232. 6. Collapse of dome: Mal. 489. 19–490. 5, Theoph. 232. 27–233. 3. Paul, (Descr. 186–203)Google Scholar and Agathias, (Hist. 5. 9. 1–3)Google Scholar both conflate the two last events and imply that the dome collapsed at the time of the earthquake; on Agathias, see Cameron, Averil, Agathias (Oxford, 1970)Google Scholar, Appendix C.

8 Re-dedication in 562: Mal. 495. 9–16, Theoph. 238. 18–24, Paul Sil. Descr. 315–49. Extension of festivities until Epiphany: Anon. Diegesis S. Sophiae ch. 27 (ed. Preger, T., Scriptores Originum Constantinopolitanarum, [Leipzig, 1901–1907], i. 105. 9–11Google Scholar), an unreliable source which is, however, in this instance supported by Paul's reference to the double extension of the festivities (Descr. 74–80). Friedländer (p. 110) suggested, on the basis of the opening lines of the poem (1 f.), that Paul may have recited on the Feast of Epiphany, but the reference is not conclusive. The opening lines (1–4) of Paul's poem on the ambo indicate that it was recited separately, in celebration of the subsequent completion of the ambo of S. Sophia.

9 Examples of hexameter panegyrics with an iambic prologue are found from the fourth to the sixth centuries. See Friedländer, pp. 119 ff.; Viljamaa, T., Studies in Greek Encomiastic Poetry of the Early Byzantine Period (Soc. Scient. Fenn., Comm. human. litt., Vol. 42. 4, Helsinki, 1968), pp. 68 ff.Google Scholar; Cameron, Alan, ‘Pap. Ant. III. 115 and the iambic prologue in late Greek poetry’, CQ n.s. 20 (1970), 119 fCrossRefGoogle Scholar. A possible precedent for a double prologue is Laud. Beryt. II (Heitsch, E., Die griechischen Dichterfragmente der römischen Kaiserzeit, i (2nd ed., Göttingen, 1963), No. xxxi, pp. 98 f.Google Scholar). Here an iambic prologue is followed by a piece in elegiacs, while a hexameter passage apparently followed; cf. line 40. In imitation of Paul, Corippus prefaced his Latin poem on Justin II with two hexameter panegyrics (one for the emperor and one for Corippus' patron), and Agathias composed a triple preface for his Cycle of poems (A.P. 4. 3: an iambic appeal for the audience's goodwill, a hexameter encomium of the Emperor and an elegiac piece on the endurance of literary monuments); cf. Corippus, Flavius CresconiusIn laudem Iustini Augusti minoris, ed. Cameron, Averil (London, 1976), pp. 118 fGoogle Scholar. S. Antès, however, argues that Corippus' panegyric of his patron Anastasius was composed independently, before the preface in praise of the Emperor and Books 1–3; see Corippe, Éloge de l'empereur Justin II (ed. Budé, , Paris, 1981), pp. xvii–xx and ci–ciiGoogle Scholar.

10 See the examples collected by Friedländer, Viljamaa and Alan Cameron (as cited in note sup.), although Pap. Ant. III. 115 is a panegyric.

11 Lines 66 f. πρ⋯ς δ⋯ τ⋯ν νεὼν|ἤδη βαδίζειν βουλομένοις θαρρεῖν δίδου As a classicising writer, Paul naturally avoided explicitly Christian terminology and so called S. Sophia a ‘temple’ ; cf. Cameron, Averil, ‘Procopius and the church of S. Sophia’, Harvard Theological Review 58 (1965), 161–3CrossRefGoogle Scholar, and Agathias, pp. 75–88.

12 Cf. line 134 οὐκο⋯ν ⋯π⋯νειμι πρ⋯ς βασιλέα τ⋯ν μέγαν, pointing the transition to Paul's praise of Justinian in the opening hexameters (135 ff.). Similar phrases using a verb of motion commonly mark the sequence of development in prose rhetoric.

13 Lines 81 ff. ⋯κ τ⋯ς ⋯στίας|τ⋯ς το⋯ βασιλέως ⋯ς βασιλέως ⋯στίαν|το⋯ παμμεγίστου; 85 τ⋯ν γερ⋯ν ⋯ προστ⋯της, 88 τ⋯ν ἱερ⋯ν τ⋯ν προστ⋯της. Justinian is called τ⋯ν γερ⋯ν ⋯ προστ⋯της for the sake of the pun (achieved by the softened pronunciation of the gamma of γερ⋯ν) with the appellation of Eutychius. Linguistic play is a common feature of the iambic prologue, which was written in a lively style so as to enlist the audience's goodwill for the serious poem to follow; see Viljamaa, op. cit. in n. 9, pp. 82 f.

14 The lemma reads: τούτων λεχθέντων ⋯ν τῷ παλατίῳ διῃρέθη ⋯ ⋯κρόασις, κα⋯ ⋯λέχθη τ⋯ λοιπ⋯ ⋯ν τῷ ⋯πισκοπείῳ, ⋯π⋯ Εὐτυχίου το⋯ ⋯γιωτ⋯του πατρι⋯ρχου, προλεχθέντων τ⋯ν ὑποκειμένων ἰ⋯μβων.

15 The poem survives only in one manuscript, the tenth-century Heidelberg cod. gr. 23 (the source of the earlier books of the Palatine Anthology). This lemma does not appear as a marginal note, but is inserted into the main body of the ms. in the first hand. This suggests that it was already in the exemplar from which the tenth-century scribe was copying. A lemma at the beginning of Paul's Ambo refers to the performance of the poem in the patriarchate (⋯ν τῷ πατριαρχείῳ). It may be significant that our lemma uses the term ⋯πισκοπεῖον rather than the more correct πατριαρχεῖον: if the motive was stylistic, that is to avoid the repetition of ⋯ν τῷ πατριαρχείῳ, ⋯π⋯ Εὐτυχίου το⋯…πατρι⋯ρχου, it may be that the lemma was written by the fastidious Paul himself.

16 Chronicon Paschale (ed. Dindorf, L., Bonn, 1832), p. 727. 7 ffGoogle Scholar.

17 For a more popular poem composed for the re-dedication of S. Sophia, see Trypanis, C. A., Fourteen Early Byzantine Cantica (Wiener byz. Studien, Band V, Vienna, 1968), No. xii, pp. 139–47Google Scholar.

18 Contra Lethaby, W. R. and Swainson, H., The Church of Sancta Sophia Constantinople (London and New York, 1894), p. 34Google Scholar. The precise location of the patriarchal palace in a.d. 562/3 is uncertain. According to the contemporary Syriac historian John of Ephesus (Ecclesiastical History 2. 34, trans. Smith, R. Payne [Oxford, 1860], p. 145Google Scholar), the palace was magnificently rebuilt by the Patriarch John Scholasticus (a.d. 565–77) after it had been destroyed by fire. John's palace has now been identified with a suite of rooms, decorated with mosaic, which survives above the south-west vestibule of S. Sophia. See Cormack, Robin and Hawkins, Ernest J. W., ‘The mosaics of St Sophia at Istanbul: the rooms above the southwest vestibule and ramp’, DOP 31 (1977), 175–251Google Scholar. (I am grateful to the CQ reader for this reference.) Cormack and Hawkins suggest (pp. 199–202) that the fire mentioned by John of Ephesus is probably that of December 563 (not that of the 532 Nika Riot as previously postulated), and they discuss the possible location of the palace in the period 532–63, noting that a Latin version of the Acts of the Second Council of Constantinople in May 553 specifies that the 168 bishops met in secretario venerabilibus Episcopis huius regiae civitatis, which indicates that the Patriarch did not lack official apartments at this time. Paul's words (see nn. 11 and 13 sup., and cf. Descr. 414) suggest that they were located close to S. Sophia.

19 As suggested by Friedländer, p. 110. I agree with Fr.'s reconstruction of the occasion.

20 As stated in the lemma; see n. 14 sup. It is Eutychius' goodwill which Paul seeks in the second prologue; cf. line 89 ἔστω δ⋯ καὐτ⋯ς εὐμενής.

21 See Averil Cameron, art. cit. in n. 4 sup.

22 But see Friedländer, pp. 99–101, for prose panegyrics of bishops responsible for founding new churches. In the seventh century, George of Pisidia praised the Patriarch Sergius (a.d. 610–38), his patron, in iambic verse, e.g. de Vanitate Vitae 231 ff. (ed. Migne, J-P., Patrologia Graeca 92. 1598Google Scholar); cf. Hexaemeron 1 ff. (PG. 92. 1425 ff.) and 1869 ff. (col. 1576 ff.; here in conjunction with reference to the Emperor Heraclius, lines 1845 ff., col. 1575 f.).

23 E.g. Proc. Aed. 1. 1. 22 f., Agath. Hist. 5. 9. 2, Paul Sil. Descr. 110 f. The anonymous Diegesis 28 (p. 105. 15 ff. Preger; see n. 8. sup.) puts the collapse and rebuilding of S. Sophia in the reign of Justin II, but this is erroneous.

24 The Christian emphasis is already marked in Corippus' panegyric on Justin II, composed soon after the latter's accession in a.d. 565; see Averil Cameron, art. cit. in n. 4 sup., pp. 10 ff. The change is illustrated by the following contrast: in Paul, the goddess Roma appears to Justinian and urges him to rebuild S. Sophia (lines 219 ff.); in Corippus, the Christian Virgin appears to Justin in a dream and presses him to assume the imperial throne (Laud. Just. 1. 32 ff.).

25 Sources written after Justinian's death (in a.d. 565) comment on the neglect of his later years, particularly in military and financial matters; cf. Cod. Just. Nov. 148 (a.d. 566) praef.; Cor. Laud. Just. 2. 260 ff.; Agath. Hist. 5. 13. 7–14. 4. The great military victories in Italy and North Africa came before the last decade of the reign, while in a.d. 561 a Fifty-Year Peace was agreed with Persia, over whom Justinian had failed to achieve a decisive victory; see Stein, E., Histoire du Bas-Empire, ii (Paris, 1949), 516 ff.Google Scholar, Bury, J. B., History of the Later Roman Empire, A.D. 395–A.D. 565, ii (London, 1923), 120 ffGoogle Scholar. Surveys of Justinian's last years: Stein, op. cit. ii. 777 f., Jones, A. H. M., The Later Roman Empire, i (Oxford, 1964), 298 ff., esp. 301 fGoogle Scholar.

26 Lines 135 f. οὐδ' ⋯π⋯ νίκην|ἔσπερον ἠ⋯ Λίβυσσαν ⋯πείγομαι; 138 Μηδοɸόνων ⋯βόητα μένοι κλέα σήμερον ἔργων; 142–4 παντ⋯ς ὑπερκύδαντος ὑπέρτερον οἶκον ⋯γ⋯νος|…,ᾦ ὕπο μούνῳ|π⋯ν κλέος ὑψορόɸοιο κατώκλασε θέσκελον ἔργου.

27 Like other poets of the period; cf. A.P. 4. 3B. 1 ff., 9. 641. I f. (both Agath.), ib. 16. 72. 1 ff. (Anon.); Cor. Laud. Just. praef. 27 ff., with Cameron's n. on 30 f.

28 The scene is reminiscent of Claudian, de bello Gildonico 17 ff. and its imitation in Sidonius Apollinaris, Panegyricus Aviti 45 ff., where a distraught Roma appeals to Jupiter.

29 Cf. Cameron on Cor. Laud. Just. 3. 309.

1 Contemporary examples are Proc. Aed. 1. 1. 10, 16, on Justinian's clemency to the conspirators Arsaces and Artabanes in a.d. 548; Cor. Laud. Just. 1. 60 f., 4. 348 ff. and A.P. 4. 3B. 53 f. (Agath.), probably all referring to the conspiracy of Aetherius and Addaeus in a.d. 566; see Cameron on Cor. loc. citt., and Agathias, pp. 14 f.

31 The panegyric concludes with 7½ lines (58–65) of compliment to the dead Empress Theodora. The opening passage (3–21) is a survey of Justinian's achievements in partnership with Christ, the last few lines of which (18 ff.) prepare for the extended treatment of the conspiracy. Paul reverts to the theme of conspirators and Justinian's clemency in more general terms in his concluding panegyric of the Emperor (937–58).

32 Cf. Cameron on Cor. Laud. Just. 1. 60 f.

33 Iambic prologues are written in the resolved trimeters of Attic comedy; see Maas, P., Greek Metre (trans. Lloyd-Jones, H., Oxford, 1962), sec. 114Google Scholar.

34 The dative ᾧ in this phrase is vague and ambiguous. I favour the interpretation proposed to me by Dr J. D. C. Frendo : observing that the verbs of 29–30 correspond to the three virtues attributed to Justinian in 22–3 (γνοὺς and μαθὼν recalling γνώμη, ⋯καρτέρησας recalling εὐψυχία and πεπίστευκας πίστις), he suggests that ᾧ is a neuter referring back generally to the combination of these three virtues, which together secure victory for Justinian in all things.

35 He is not named, the usual practice in panegyric; cf. Cameron, , Agathias, p. 15Google Scholar. The suicide is mentioned again, Descr. 937–9; cf. note 31 sup.

36 Cf. Call. Ap. 25 ff.; also Il. 17. 98 f., A.P. 10. 91 (Palladas), Nicephorus, , Hist. Syn. (ed. de Boor, , Leipzig, 1880), p. 6. 11–13Google Scholar.

37 Fullest account, Malalas, , Excerpta de insidiis, fr. 49 (ed. de Boor, , Berlin, 1905, p. 173. 30 ff.)Google Scholar; also Malalas, pp. 493. 1–495. 5 Bonn, Theophanes, pp. 237. 15–238. 18 de Boor.

38 Others later implicated in the plot are called ⋯ ⋯ργυροπρ⋯της. On this basis, Averil Cameron (on Cor. Laud. Just. 2. 361 f.) suggests that this was a bankers' plot, reflecting the financial crisis at the end of Justinian's reign and consequent strained relations between government and bankers.

39 Great Hall or state room, probably the ‘Triclinium of the Nineteen Couches’. The grand new Chrysotriclinium is attributed by the sources to Justin II; see most recently Cameron, Averil, ‘The artistic patronage of Justin II’, Byzantion 50 (1980), 76Google Scholar. Bury, however, suggested that Justinian may have designed the Chrysotriclinium buildings as part of the rebuilding of the palace area after the destruction of the Nika Riot (op. cit. in note 25 sup., ii. 54, n. 4).

40 Mal. Exc. de ins. p. 174. 19; cf. Theoph. p. 238. 3 f.

41 It is, of course, indirectly attributable to God's protection of Justinian, since the hand of God may be seen in Ablabius' revelation of the plot. The chronographers prefix the reference to Ablabius' betrayal with the phrase το⋯ θεο⋯ [οὕτως] εὐδοκήσαντος (Mal. Exc. de ins. p. 174. 9, Mal. P. 493. 12, Theoph. p. 237. 25).

42 P. 239. 17 f.

43 Agath. Hist. 5. 15. 7 ff., Theoph. p. 233. 18 ff.

44 Agath. Hist. 5. 20. 5. On the bias of Agathias' whole account of this episode, maximising Belisarius' role, see Cameron, , Agathias, pp. 49 fGoogle Scholar.

45 Belisarius' disgrace at the end of his life (he died in March 565, Theoph. p. 240. 24–6) is reflected in medieval legend, according to which the great general ended his days as a blind beggar in Constantinople; cf. Wagner, W., Medieval Greek Texts (London, 1870, repr. 1972), pp. 116 ff.Google Scholar; Bury, op. cit. in note 25 sup., ii. 69; Stein, op. cit. in note 25 sup., ii. 779 f., n. 5.

46 Enumerated in lines 75 f.: ἄπας ⋯ δ⋯μος εὐθύς, ⋯ γερουσία, | οἱ τ⋯ν μέσον ζηλο⋯ντες ⋯σɸαλ⋯ βίον…. The second group is the senate. The first could be either technical ‘circus factions’ or general ‘plebs’. The third expression perhaps has more point if δ⋯μος is taken to refer to the circus factions; then ‘those who pursue the safe middle life’ would be the clergy, who, as a result of Justinian's ecclesiastical legislation, were ‘strictly forbidden to indulge in the pastimes of attending horse-races or visiting the theatres’ (Bury, op. cit. in note 25 sup., ii. 361), and hence could be termed ‘secure from’ the passions generated in the Hippodrome. I am indebted to Dr J. D. C. Frendo for this suggestion.

47 See note 33 sup. The accentual Byzantine dodecasyllable, a standard metre of later poetry, is based on the iambic trimeter.

48 For example, proparoxytone words are avoided at the line-end. For detailed analysis of Nonnus' metrics, see Keydell, R.'s edition of the Dionysiaca, i (Berlin, 1959), 35*–42*Google Scholar; summary in F. Vian's Budé edition, i (Paris, 1976), pp. 1–1v. Paul admits only six of the nine permutations of the hexameter found in the Dionysiaca and never writes two consecutive spondees. See Friedländer, pp. 117–19; Merian-Genast, J., De Paulo Silentiario byzantino (Diss., Leipzig, 1889), pp. 41–86Google Scholar; Caiazzo, Carla, ‘L'esametro in Paolo Silenziario’, JÖB 32(3)Google Scholar (1982; = Acts of 16th Intern. Byz. Congress, Vol. ii (3)), pp. 335–43.

49 Ed. Heitsch, op. cit. in note 9 sup., pp. 127–52.

50 For, example, John of Gaza; see Friedländer, p. 112.

51 Viansino has explored the literary affinities of the epigrams, and Yardley stresses Paul's knowledge of earlier Greek literature (loc. citt. in note 3 sup.), but the high literary quality of the Descr., although recognised by Friedländer (pp. 123 f.), has hitherto been insufficiently explored and appreciated. It is punctilious, for example, in the correct use of particles. For an instance of recherché Homeric scholarship, see McCail, R. C., ‘ΛΑΩ: two testimonia in later Greek poetry’, CQ n.s. 20 (1970), 306–8CrossRefGoogle Scholar and, for an Aristophanic allusion in the iambic prologue, ‘Κυαμοτρὼξ Ἀττικός in Paulus Silentiarius, Descriptio 125: no allusion to Simplicius’, PCPS n.s. 16 (1970), 79–82Google Scholar.

52 Cf. Call. Ap. 58 τετραέτης τ⋯ πρ⋯τα θεμείλια Φοῖβος ἔπηξε. This imitation of Callimachus is the only instance in the Descr. of the lengthening of a monosyllable (τ⋯) by mute and liquid; see Merian-Genast, op. cit. in note 48 sup., pp. 68 f.

53 I tentatively retain Du Cange's conjecture, accepted by all editors, for the manuscript's metrically defective reading ⋯δόνητον κρηπίδι θεμείλῳ. But Ludwich, Arthur, Textkritische Noten zu Paulus Silentiarius (Königsberg, 1913), pp. 10 f.Google Scholar, correctly objected that Paul does not elsewhere use εὔκρηπις and that he uses θεμεῖλον elsewhere in the plural only. These objections can be countered, but Ludwich's conjecture ⋯π⋯ κρηπῖδι θεμείλων (or ⋯ν⋯ κρηπῖδι θεμείλων) is none the less attractive.

54 Anthemius of Tralles, judged by Procopius (Aed. 1. 1. 24) to be superior in his profession not only to his contemporaries but also to his predecessors. In the building of S. Sophia, he collaborated with Isidorus of Miletus, as Paul mentions, line 552 f.; cf. Proc. Aed. 1. 1. 24, 50, 70.

55 Other suggested subjects are unsatisfactory. Du Cange extracted οἱ τοῖχοι from τοίχοις (272) as the subject of ⋯κλασεν, but the plural subject does not marry with the singular verb; he found a subject for ⋯πεστήριξεν by combining the expression from ⋯λλ' (275) with the one which follows in lines 276–8, but this entails other difficulties. Meineke (in notes to C. W. Kortüm's German translation of Paul's poem, appended to Salzenberg, W., Altchristliche Baudenkmale von Constantinopel, Berlin [1854], p. xi n. 16Google Scholar) proposed σθένος (272) as the subject of both verbs, but its use with ἴχνος ⋯πεστήριξεν would involve an unlikely personification.

56 I have emended Friedländer's punctuation of this line.

57 The brief notices of the chronographers for the two encaeniae of S. Sophia suggest that both involved similar processions headed by the Patriarch; cf. Theoph. p. 217. 16–20 and ib. 238. 18–24. Paul does, however, specifically recall (in lines 350 f.) the psalm which the chronographers record was sung at the second encaenia.

58 See Ludwich, op. cit. in note 53 sup., pp. 14 f., on Paul's sparing and discriminating use of the particle τε.

59 Cf. line 2 θεός τε κα⋯ βασιλεὺς σεμνύνεται

60 E.g. A.P. 9. 423. 4 (Bianor) ⋯λβον Πακτωλο⋯ ῥεύματι δεξ⋯μεναι; Nonn. D. 34. 213 Λυδ⋯ν ἄσπετον ὅλβον, ὅσον Πακτωλ⋯ς ⋯έξει, al.

61 Od. 18. 19; cf. Hes. Th. 420, Manetho 2(1). 221, Greg. Naz. Carm. 1. 1. 4. 83 (PG. 37. 422A), etc.

62 There may be a further example of inadvertent repetition earlier in the poem. At Descr. 286 and 289, the noun ⋯μμα is reiterated in eadem sede, and likewise the verb ἰδὼν/ἰδεῖν (284, 288, 290). These repetitions do not appear to have any stylistic purpose, although they do occur within an elaborate synkrisis which is framed by the parallel expressions π⋯ς βροτ⋯ς (286) and π⋯ς κόρος (299).

63 In Paul's poem, there is no shadow of the events of a.d. 565, when Eutychius was deposed by Justinian for refusing to sign the Emperor's aphthartodocetist edict; see Stein, op. cit. in note 25 sup., ii. 687 with 688 n. 1.

64 Building achievements are a standard element of imperial panegyric in the category of ‘peacetime achievements’, e.g. Vell. Pat. 2. 130, Pliny, Pan. 51, Prisc. Pan. 184 f., Proc. Gaz. Pan. 17–21, Proc. Aed. 1. 1. 12, 17 ff., and passim.

65 Cf. Friedländer, p. 124, who quotes (n. 1) the judgement of Joseph Scaliger on Paul (in a letter to Salmasius),… ‘Silentiarius, vitio saeculi sui, quae tum virtus erat, usus est. Strepitus verborum, ambitus sententiarum, compositio Dithyrambis atidacior…. Quod uno verbo exponere poterat, maluit binis, trinis versiculis producere’.

66 Cf. Viansino, op. cit. in note 3 sup., p. xvii.

67 Cf., for example, Themistius, Or. 19 (i. 337. 7 ff. Downey), al.; Agapetus, Cap. 35 (PG. 86. 1176A), al.

68 See section II above.

69 158 f. ῥινοτόρῳ δούλωσεν ⋯πείρονα β⋯ρβαρον αἰχμῇ,|⋯ɸρα τεῖς ⋯δμ⋯τα λόɸον κλίνειε λεπ⋯δνοις.

70 E.g. lines 208–13. This synkrisis is preceded by a gnomic statement (207), parallel to that of 956.

71 The epigrammatist illustrates his point by three analogies, the sting of the bee, the whip which directs the proud horse and the staff which controls the herd of swine. The last concludes with the same half-line as Paul's 958. The resemblance is unlikely to be accidental. There is no means of dating the epigram, beyond the fact that its style and language are in the manner of Nonnus. I am inclined to believe that Paul is adapting the epigram (rather than vice versa) because Paul's argument, that the bull is more amenable to persuasion than to force, is less obvious than that of the epigrammatist, and might have been intended as a refutation of the latter. The fierce and belligerent bull is a better analogy than the swine for the rebel, to whom Paul's synkrisis refers, its seduction by the pipe being as surprising and apparently uncharacteristic as the rebel's conversion to allegiance.

72 See Jannaris, A. N., An Historical Greek Grammar (London, 1897), sec. 2092, p. 488 and Appendix iv, sec. 7, pp. 554 f.Google Scholar; cf. Mitsakis, K., The Language of Romanos the Melodist (Byz. Archiv Heft 11, Munich, 1967), sec. 284, p. 143Google Scholar.

73 The use of ὑπό + dative instead of a simple dat. instr. originates in Homer; see Friedländer, p. 115, and cf. Chantraine, P., Grammaire homérique, ii (Paris, 1953), 140 f., sec. 208Google Scholar; it is not uncommon in post-classical poetry, e.g. A.R. 2. 26, A.P. 5. 74. 2 (Rufin.), Nonn. D. 34. 137, al. For the use of ⋯πό instead of ὑπό in late prose, see LSJ s.v. ⋯πό III. 4, Blass, F., Debrunner, A. and Rehkopf, F., Grammatik des neutestarnentlichen Griechisch (14th ed., Göttingen, 1976), sec. 210. 1 & 2, p. 171Google Scholar, Jannaris, op. cit. in note sup., sec. 1507 f., pp. 369 f. Paul's expression does, however, include an idea of cause or means, a context in which ⋯πό is regularly used; see LSJ s.v. III. 6. For Nonnus, see Peek, W., Lexikon zu den Dionysiaka des Nonnos (Berlin, 1968–1975)Google Scholar, s.v. ⋯πό II. 1, col. 159 f., Keydell, op. cit. in note 48 sup., i. 63*. Paul uses a similar ⋯πό in line 988.

74 See Mitsakis, op. cit. in note 72 sup., sec. 67, p. 37. I am indebted to Dr J. D. C. Frendo for both this point and that of note 72. Paul uses similar compound superlatives in his iambics, e.g. παγκρ⋯τιστος (22), παντ⋯ριστος (60), παμμέγιστος (83, 101,111).

75 Agath. Hist. 5. 6. 7 … γένους τε κοσμούμενος δόξῃ κα⋯ πλο⋯τον ἄɸθονον ⋯κ προγόνων διαδεξ⋯μενος, ὅμως παιδεία γε αὐτῷ κα⋯ λόγων ἄσκησις διεσπούδαστο, κα⋯ ⋯π⋯ τοῖσδε μ⋯λλον ηὔχει κα⋯ ⋯σεμνύνετο.

* I am much indebted to the inspiration and guidance of Dr R. C. McCail and to the constructive criticism of Dr Michael Whitby.