IS THERE AN ACROSTIC IN BALBILLA, EPIGR. 31 BERNAND?*

The article presents a critical discussion of the interpretation of one of Balbilla’s epigrams carved on the foot of Memnon’s colossus at Egyptian Thebes (Bernand no. 31) proposed by Marialuigia Di Marzio in her article published in QUCC 2019. According to this interpretation, it contained the acrostic ΕΦΗΩΚΑΕ, which Di Marzio resolves as ΕΦΗ Ω ΚΑΕ(ΣΑΡ), suggesting a direct apostrophe to the emperor Hadrian on the part of the statue. However, there are compelling reasons to reject this intriguing proposal.

In 2019, Marialuigia Di Marzio published an article 1 suggesting a previously unnoticed acrostic in one of the epigrams by Julia Balbilla inscribed on the left leg of one of two colossal statues, originally carved in 1400 BCE to honour Amenhotep III, who at the end of the 1st century BCE was re-identified as Memnon, the son of the goddess Eos, the mythical king of Ethiopia killed at Troy by Achilles, but later bestowed with posthumous immortality by Zeus. 2 The statue was severely damaged in the first century CE and since then its stone base, probably expanding under the heat of the rising sun, began to make a mysterious sound at dawn, which was interpreted as Memnon's complaint to his mother.Hearing this sound was considered a sign of divine favour.During the first two centuries CE, crowds of pilgrims flocked to Thebes to witness the phenomenon (after two centuries, the statue grew silent again, probably due to efforts to repair the damage); many tourists / pilgrims commemorated their presence with inscriptions in Latin or Greek on the colossus.Hadrian and his consort Vibia Sabina visited the statue in November 130, and Balbilla accompanied them on that tour, both as a lady-in-waiting of Sabina and as a poetess to celebrate the emperor's encounter with the Memnon statue.
The standard edition of Memnon's inscriptions (107 in total) was prepared by the brothers André and Étienne Bernand. 3 The Balbilla epigram we are dealing with here is numbered 31 therein and reads as follows: A significant change from the Bernand brothers' edition is that the beginning of the verse in question ( 6) is read as <ἄμ>α<τ᾽> ἔχεσκε δ᾽5 instead of (φῶτ)α δ᾽ ἔχεσκε(ν)6 .Let me emphasise right away that I consider this a legitimate correction. 7Once this conjecture is adopted, the initial letters of the verses form the sequence ΕΦΗΩΚΑΕ, which Di Marzio resolves as ΕΦΗ Ω ΚΑΕ(ΣΑΡ).Her interpretation of it is noteworthy: 8 "The poetess in the last epigram [= 31 Bernand], the one that is only apparently the most straightforward and administrative, plays with the reader and sends them a message, which is an exceptional example of syncretism in several respects.The statue of Phamenoth / Memnon spoke and, while speaking Greek (ὦ Kαε), used the Latin form pronounced with the hard guttural (K), Cae(sar), instead of the normal Greek form Καῖ(σαρ), perhaps in homage to Hadrian's Latinity.The encomiastic intention is thus not only in the invocation with the appellation typical of Roman emperors, but also in the Latin incipit of the honorary title.The Greek-Egyptian statue seems to bow to the Emperor of Rome and does so in his own language". 9 Di Marzio is right in stating the linguistic uniqueness of the vocative KAE.It is not found in either Greek or Latin inscriptions.In the former, there appear the abbreviations Κ, ΚΑΙΣ 10 (and additionally on the coin inscriptions ΚΑΙ) 11 , in the latter CAES, CAE, C. 12 An argument for the intentional inclusion of the abbreviation of Καῖσαρ in the initial letters of verses 5-7 could be that KAE would match Κοιράνω<ι> at the beginning of verse 5, forming together with it a kind of 'homonymous gamma (or rectangular) acrostic' to comprise two common imperial titles: The point of overlap between the two words discussed here, allowing only one character (and thus forcing a compromise), is the 'corner' letter to represent the 'k' sound, written Κ in one language and C in the other, which would to some extent justify the unusual spelling of the Latin CAE as KAE.The bilingual nature of such an acrostic would not be surprising, as a parallel case can be 8 Di Marzio 2019: 116 (transl.J.D.). 9The view expressed synthetically in Di Marzio 2020: 68 n. 2: "Un messaggio breve, ma carico di sottintesi, nel quale il Colosso greco-egizio parla in greco, ma usa la forma latina Cae(sar) piuttosto che quella greca Καῖσαρ, in un probabile omaggio alla latinità di Adriano".Di Marzio (2019:  117) enhances her observations with sociolinguistic aspects that take into account the perspective of the average reader: "[u]na mistione di ele menti capace dunque di intercettare l'attenzione di qualsiasi viaggiatore ammirasse il colosso, quello greco ma anche quelli egizi o romani con una più o meno limitata alfabetizzazione greca".On the possible function of acrostic in alien geographical and chronological contexts, see Garulli 2013: 274: "In particular, the provenance of the acrostic poems examined from areas where Greek was not the only (or first) language might also suggest that the acrostics could have functioned as a socially-oriented abstract of the whole inscribed text, especially for readers who were not native or good speakers of the Greek language". 10McLean 2002: 53. 11Hartmann, MacDonald 1969: 31. 12https://www.forumancientcoins.com/dougsmith/abb.html(Abbreviations on Roman Imperial Coins).

ΡΩΜΑΙΩΝ ΜΑΝ ΩΣ
The author of this inscription, transliterating into Greek characters the Latin plural accusative Romanos, even took pains to mark the length of the 'o' in his writing, clearly demonstrating that he consciously introduced two languages into the acrostic.
The presence of the acrostic in Balbilla's poem also seems to be evidenced by its length: it spans the entire inscription (from the first to the last verse).Furthermore, it contains more than one word, more specifically a phrase with a verb, cf.verses 1-22 of the aforementioned Kalabsha inscription, where, in addition to the author's name and status, the verb ἔγραψα (18 -22) is inserted.Incidentally, in Balbilla the acrostichic verbum dicendi (ἔφη) has its counterparts in the concurrent verses: αὐδήσαντος (1), φώνα(ς) (2). 14t this point, however, I would like to signal my doubts.Di Marzio understands the final words of the acrostic ΕΦΗ Ω ΚΑΕ(ΣΑΡ) as an invocation preceded by a verbum dicendi15 (Diceva: 'O Cesare') 16 addressed to Hadrian by the colossus.17Although the acrostic is by its nature paratextual, it refers only to the work in which it is anchored, and the communication layout encoded in epigram 31 does not include the emperor.The idea of referring in a preceding epigram (here 28) to an acrostic found in one of the subsequent pieces (here 31) within the same 'chain' of epigrams, as developed by Di Marzio,18 although ingenious, is unlikely.Uttering a sound by Memnon is a theme running through all four epigrams of Balbilla, who in Epigr.31.1-2finally makes herself the recipient of his voice (Ἔκλυον αὐδήσαντος ἔγω ᾽πυ λίθω Βάλβιλλα / φώνα(ς) τᾶς θείας Μέμνονος ἢ Φαμένωθ).The comment contained in the acrostic, if any, would naturally refer to her.Theoretically, Balbilla could thus inform the emperor that she too had been honoured by the colossus with a sound made in her presence. 19Finally, it should be noted that the sound was inarticulate, so that the invocation "O Caesar!" simply could not come 'out of the lips' of Memnon.In Epigr.28.4 Memnon greeted Hadrian "as he was able" (χαίρην εἶπέ (ϝ)οι ὠς δύνατον), and immediately afterwards in this poem his voice is compared to the sound of beaten bronze (v.7-8 ὠς χάλκοιο τύπεντ[ο]ς ἴη Μέμνων πάλιν αὔδαν / ὀξύτονον). 20As regards apostrophes, the true ones are directed by visitors to the statue, which thus, as it were, is revived and personified. 21he plausibility of the presence of the acrostic is also sought by Di Marzio against the background of certain 'eccentric' features of Epigr.31, such as, inter alia, the lack of correct hexameter-pentameter alternation, with an isolated pentameter (v.7) closing a series of full couplets (vv.1-6). 22The scholar allows for the possibility that the piece was deliberately extended by one verse, since seven verses are needed to accommodate the invocation in the acrostic.The double dating in the final two pentameters 23 would then have been subordinated to this very purpose.In my view, this is an assumption made under a preconceived thesis.The second date bears all the hallmarks of an addition and has no parallel in the other Memnon inscriptions.The epigram proper was probably only six verses long.
There is, however, an even more serious reason to doubt the intentional stretching of the acrostic.As is well known, an acrostic, in order to be perceived by observers, must be visually distinguished by the positioning of its constituent letters at the beginning of successive verses.Epigr.31 by Balbilla, as inscribed on the colossus, does not fulfil this condition, since from the second letter of the last word of the fifth verse onwards its graphic layout is broken up into segments that do not coincide with hexameter and pentameter (see above the text quoted according to the Bernand brothers' edition, in which the vertical dashes indicate the boundaries of the individual lines).As a result, the beginnings of the inscription lines are arranged as follows: 24 This means that the postulated acrostic is partially hidden from the observer's view.If Balbilla had indeed intended to introduce an acrostic, she would have had the lines carved "normally", that is, in whole verses, even at the expense of reducing the size of the letters (the distinguishing feature of this particular inscription is the enlargement of the letters in comparison to neighbouring inscriptions).
In the situation outlined above, Di Marzio's proposal, though generally interesting and intriguing, must ultimately be considered unproven. 25