Beyond translation: divine names and religious strategies in “asymmetrical” bilingual inscriptions from Regio I (Latium et Campania)

Introduction

Engraved on various media such as altars, stelae, plaques, statues or public monuments, ancient inscriptions serve as a reservoir and privileged domain for the study of the diverse forms of ritual interaction between mortals and divine beings. In different contexts of expression, gods are referred to by a plethora of names, meticulously chosen or crafted by human actors, depending on the occasion and purpose of religious communication within a specific socio-cultural milieu. The ancient names of the gods, which have been the focus of scholarly attention since the work of H. Usener, have enjoyed renewed interest since the 1990s,¹ leading to the development of various research projects, often resulting in databases dedicated to the study of ancient divine onomastics.² These efforts are motivated by the belief that the names of the gods offer valuable insights into the complex processes involved in defining, organising and functioning within the divine realms of ancient polytheisms. This takes place against the backdrop of Mediterranean antiquity, characterised by a wealth of cultural and linguistic exchanges, coexistences and interactions.³

Within this multicultural basin, the present article focuses on a specific selection of bilingual (Latin and Greek, and vice versa) sacred inscriptions from Regio I (Latium et Campania),⁴ which contain divine names shaped by different socio-cultural strategies. This area offers the richest evidence of such inscriptions across the entire peninsula, with a notable concentration in Rome during the Imperial Age – a city where people, languages, and cultures have intertwined and coexisted over the centuries. Against this background, Latin was the predominant language, while Greek served as an alternative language valued for its prestige. Moreover, it was widely adopted by several linguistic communities from the Eastern Mediterranean.

The bilingual (Latin-Greek) inscriptions discovered in the Italian peninsula, documented in various epigraphic collections and online databases, have so far received limited scholarly attention.⁵ For sacred dedications, this lack of scrutiny concerns the various forms of bilingualism they exhibit and the specific socio-cultural strategies, customs and needs that underlie the processes of divine naming in both languages. Recent contributions by G. Tozzi, focusing on bilingual (Latin-Greek) epitaphs from Rome, and by L. Pérez Yarza and C. Bonnet, investigating bilingual (Latin-Greek) inscriptions (funerary, sacred, defixiones) of Rome bearing divine names, deserve recognition for their thorough exploration of these two aspects in the context of the Urbs from the Late Republican era to Late Antiquity.⁶ Tozzi argues for a less rigid methodological framework and hyper-technical terminology, especially that codified by J. N. Adams.⁷ Building on this, Tozzi identifies four distinct types of bilingualism in the context of Rome:

1. bilingual inscriptions in the strict sense of the term, which feature an exact translation or present minor additions or slight variations in one of the two languages.
2. inscriptions with two different texts, one Greek and the other Latin, juxtaposed in different ways on the same support.
3. A Latin text that includes one or more isolated words or a whole Greek expression (and vice versa).
4. Latin texts written entirely in the Greek alphabet (alongside instances showing the reverse scenario).⁸

Tozzi also points out the importance of considering the socio-cultural background of these texts, stressing its pivotal role in the classification of bilingual inscriptions.

The significance of political and cultural needs, as well as social and personal situations, also guides the inquiry conducted by Pérez Yarza and Bonnet. Through a comprehensive analysis based on several case studies, they argue that the incorporation of Greek (or Latin) into an inscription in the other language transcended specialised fields and was not limited to “social exhibition” or “accessibility of technical contents”.⁹ They also identify another reason: “the need to define more adequately, even precisely, the names and attributes of the gods”. In this regard, they outline some strategies and criteria that influenced the translation or adaptation of divine names in both languages. These include attempts to provide a complementary message, as well as to better construct and verbalise a divine power from two different angles across the two languages.

The present contribution aims to build on these considerations in order to further the study and interpretation of bilingual Latin-Greek inscriptions within the Italian peninsula. Specifically, it intends to closely examine sacred dedications from Regio I in which the two texts show slight but noticeable differences or are entirely different in their invocation or address to the targeted deities. The examples selected for this study belong to the first and second group of bilingual inscriptions identified by Tozzi, which M. Corbier has called “dissimmetriche” (i.e. “asymmetrical”).¹⁰ The choice to confine the analysis to this typology of inscriptions is motivated by the fact that they do not testify to a systematic translation from one language, the primary, to another, the secondary, but to significant linguistic choices made consciously or unconsciously by the agents behind the epigraphic texts, which deserve to be highlighted.

In this sense, the present study engages directly with the theoretical reflections and taxonomic framework outlined in the introduction to this volume, by addressing the question of how multilingual expression operates within religious inscriptions – not merely as a matter of translation, but as a negotiation of religious discourse. As emphasised by L. Pérez Yarza, J. Herrera Rando and S. Bianchi Mancini in the introduction of this volume, bilingual texts often reveal how divine names and ritual vocabulary were shaped through intercultural interaction and adapted to specific socio-religious needs. The case studies discussed here aim to contribute to this broader discussion by showing how the linguistic nuances exhibited by bilingual inscriptions, with regard to the onomastic sequences of divine invocation, can illuminate the complex dynamics of language contact and religious communication. These texts elucidate diverse linguistic, and thus cultic, strategies used by human agents to verbalise or conceptualise a deity. The alternation or collaboration between Latin and Greek can indeed shed light on the multifaceted contextual approaches to the divine world, while also highlighting the underlying cultural and social background that these inscriptions enduringly reflect.

How to “Adapt” Divine Names Across Languages

Let us begin this study with a bilingual votive inscription engraved on a parallelepiped limestone altar measuring 135x80x70 cm. The piece was found re-employed in a modern villa in Sant’Angelo in Formis, a village located near Capua in Campania. It is currently kept in the Museo Provinciale Campano in Capua (Sala Mommsen) and the inscription reads:¹¹

Δεσποίνῃ Νεμέσει
καὶ Συννάοισι Θεοῖσιν
Ἀρριανός βωμὸν
τόνδε καθειδρύσατο.
Iustitiae, Nemesi, 5
Fatis, quam voverat aram,
numina sancta colens,
Cammarius posuit.

To Queen Nemesis and the Gods sharing the same temple, Arrianos raised this altar. To Justice, Nemesis, the Fata, honouring the holy gods, Cammarius raised the altar he had promised.

The inscription spans nearly the entire epigraphic field and is symmetrically arranged, with the Greek and Latin texts running for four lines each.¹² Despite its brevity, the text is an epigram composed of two regular elegiac couplets. Its dating to the 2nd c. AD, supported by both material evidence and palaeography, is further substantiated by the analysis of internal features such as language, prosody, and dedication formulation.¹³ The deliberate placement of the Greek text before the Latin is not accidental: it suggests that the dedicator was likely a Hellenophone with a proficient command of the other language. The name Ἀρριανός is notably attested in Greek inscriptions from Asia Minor during the Imperial period.¹⁴ Conversely, Cammarius, as C. Pepe¹⁵ remarks, the name that appears in the Latin text with no other documented occurrences, may be a nickname (signum) derived from cammarus (“shrimp”) or, more plausibly, a gentilicium instead of Camarius. The inclusion of this second onomastic element in the Latin text is clearly a deliberate choice. Just as the decision to convey the religious communication in two languages stems from the agent’s desire to assert his dual cultural identity, the use of a double name, which differs between the two texts, could be a sign of the acquired libertine status. However, this assumption is problematic due to the lack of mention of the tria nomina and the general changes in Latin onomastics in the 2nd c. AD resulting in the use of Greek cognomina among all social classes.¹⁶ Furthermore, his mastery of Latin is evidenced by the fact that the Latin text provides more detailed information about the nature of the dedication than its Greek counterpart. It specifies that the erection of the altar was the result of the fulfilment of a vow (quam voverat), a detail missing from the Greek text. Additionally, the Latin version expresses the agent’s religious devotion (numina sancta colens), a precision not found in the Greek text.

These observations already suggest that we are dealing with an “asymmetrical” bilingual inscription. Indeed, the Latin text does not simply serve as a direct “translation” of the Greek version. Rather, it provides a distinct set of information, indicating to bilingual readers (or to those who were proficient in one of the two languages) not only the agent’s attachment to his origins – as evidenced by his use of Greek as the primary language of the dedication – but also his integration and acculturation into the Latin social and religious milieu.

However, the most notable difference between the Greek and Latin texts of the dedication lies in the divine interlocutors to whom the altar is consecrated. Both versions mention the Greek goddess Nemesis.¹⁷ In the Greek text, she is preceded by an onomastic attribute, δεσποίνα (“Queen”), a term commonly used for female deities, and is invoked alongside “the Gods sharing the same temple” (Συννάοισι Θεοῖσιν), a formula pointing to a plurality of unspecified deities. In the Latin text, the same goddess, lacking onomastic attributes, is preceded and followed in asyndeton by Iustitia and the Fata, deities associated with the realms of justice and destiny, with whom Nemesis was also tied.¹⁸ It is plausible that the agent referred to them under the “synthetic” qualification of σύνναοι θεοί in the Greek text. The difference between the two divine addresses may indicate an intention to convey different types of information to distinct categories of readers, ensuring that the religious communication was comprehensible to both Greek and Latin audiences. The absence of an attempt to translate the σύνναοι θεοί formula from Greek into Latin likely stems from an awareness that technical idioms or cultic formulae cannot be translated directly between languages. In other words, the Greek formula may have seemed unclear from a Roman perspective, leading to a specification in the Latin text through deities (Iustitia, the Fata) who were “familiar” to a Latin reader by virtue of both their names and their Nemesis-related competencies. In this way, bilingualism in this inscription functions as a practical ritual tool, clarifying and refining the ritual invocation while demonstrating varied approaches to addressing divine powers through inclusive formulae and specific terms.

The Capua dedication can be compared with another bilingual votive inscription, this time addressed to Herakles/Hercules. The piece probably belonged to the base of a statue that has been lost together with its support, and its date remains unknown. Discovered in Pontecorvo (province of Frosinone in southern Lazio), corresponding to ancient Aquinum, the text reads as follows:¹⁹

Εὐχὴ
Ἡρακλῇ Θαλλοφόρῳ
Ἱερῷ Εὐακούστῳ
Λ(ούκιος) Κορνήλιος Λ(ουκίου) Κορνη=
λίου Πρειμιγενίου5
υἱὸς Παλ(ατίνα) Τερεντιανὸς
καὶ Λήμνιος ἀπελεύθερος
ἐποίησαν.
Herculi Pacifero
Invicto Sancto10
sacr(um)
voto suscepto
L(ucius) Cornelius L(uci) f(ilius)
Pal(atina) Terentianus
et Lemnius libertus15
merito libentes
fecerunt.
A vow to Herakles, the Olive-shoots Bearer, Holy and Willing to listen, Lucius Cornelius Terentianus, of the Palatine tribe, son of Lucius Cornelius Primigenius, and (his) freedman Lemnius, made (this). For Hercules Pacifer, Invincible and Venerable, by a vow made, Lucius Cornelius Terentianus, of the Palatine tribe, son of Lucius, and (his) freedman Lemnius, consecrated this properly and willingly.

The text commemorates the dedication of the monument to Herakles/Hercules by the Roman citizen Lucius Cornelius Terentianus and his freedman Lemnius, in fulfilment of a previous vow. The god also appears in another inscription engraved on a stele or limestone cippus found at Aquino, which testifies to the burial ground reserved for the devotees of Hercules Victor (ll. 1-5: loca sepulturae cultorum Herculis Victoris).²⁰ This inscription probably attests to a local cult of this deity and suggests the probable existence of a sanctuary in the area, possibly linked to the world of transhumance and trade. Unlike the Latin inscription from Aquino, the one from Pontecorvo consists of two texts, one in Greek and the other in Latin, of eight and nine lines respectively. As in the case of the Capua dedication, the decision to place the Greek text before the Latin is not accidental: one of the two devotees, Lucius Cornelius Terentianus, seems to be the same person to whom his mother, Terentia Thallusa, and her three sons (Secunda, Iusta and Marionianus) dedicated a marble funerary altar (101x65x43 cm), probably in the second half of the 1st c. AD.²¹ It was discovered in Rome near S. Maria in Portico and is now kept in Palazzo Altemps (inv. 420813).

The inscription mentions Terentianus’ father, called Lucius, and the Palatine tribe (as in the Pontecorvo text), additionally recording that Terentianus died at the age of 35 and had worked as a subordinate scriba aedilium et decemvirum. The mother’s onomastic (cf. the cognomen Thallusa) also suggests that she was probably a freedwoman or a migrant – even a migrant Roman citizen of Greek or Eastern origin – and that the cognomen of her son, Terentianus, did not come from adoption but from the maternal line. Thus, if the two individuals in the two inscriptions can be considered the same person, Terentianus’ knowledge and primary use of Greek could be tentatively explained by his family history and dual cultural identity. However, it is also plausible that Terentianus aimed to convey his religious communication in a culturally prestigious manner, which may have been facilitated by the use of Greek.

A comparison of the texts of the dedications reveals, first of all, minor discrepancies in the votive formulary, which once again underline the agent’s command of the Latin cultic vocabulary: in Greek, we find the more concise εὐχὴ (l. 1), while the Latin text refers more explicitly to a vow previously made and now fulfilled (voto suscepto, l. 12). There is also a subtle divergence in the dedication formula: the Greek text suffices with ἐποίησαν (l. 8), while the Latin has the more extended and precise expression sacrum (…) fecerunt (ll. 11 and 17), meaning “they consecrated/they offered”. However, it is primarily the attributes chosen to characterise the deity in both languages that offer us some interesting considerations. Notably, there is no direct correspondence or exact translation between the two divine onomastic sequences. The term θαλλοφόρος, which literally means “who bears young olive-shoots” (position 1 in the Greek onomastic sequence) and is attested only in this inscription in reference to a deity,²² is matched by pacifer (position 1 in the Latin onomastic sequence). The adjective is frequently associated with the image of Herakles/Hercules on imperial coins from the Postumus series onwards but does not appear elsewhere in the epigraphic context.²³ As pointed out by D. Potage, both are compound terms with the root *bher– and may be linked by the symbolic echo between the olive tree and the idea of peace.²⁴ The term ἱερῷ (position 2) is answered by the word sancto (position 3), which is a possible translation.²⁵ The intention was to emphasise the sacredness and the sanctity of the god through a praise epithet. In this regard, it is worth noting once again the rare occurrence of ἱερός in onomastic reference to a deity, as opposed to the more common ἅγιος, to which sanctus was supposed to correspond more directly. On the contrary, εὐακούστῳ (position 3) and invicto (position 2) do not correspond either in form or in meaning: the former relies on the god’s predisposition to listen, the latter on his appearance as a victorious warrior god. On closer examination, however, as Potages notes, the benevolent disposition of the deity who is inclined to grant requests could be expressed in Latin by pacifer: the god “who listens well to prayers” is the one who probably welcomes them, bringing his pax, his venia. At the same time, the idea of the deity’s invincibility could also be conveyed by θαλλοφόρος, in an echo, also symbolic, between the crown of branches worn by the winner of the games (olive for those of Olympia, laurel for those of Delphi) and the god’s “impregnable” nature.²⁶

It follows that in this dedication Herakles/Hercules is doubly characterised as the god who triumphs and the god who grants requests. These dual attributions, one probably derived from the local cult of the deity, as indicated by the title Victor attributed to the god in the Aquino inscription, and the other more directly related to the private dimension of the dedicator’s religious experience and expectation, are harmoniously integrated and reinforced in both Greek and Latin by a skilful selection of attributes. This selection seems to exploit their potential polysemy. The Latin text, which is presumably secondary both in its conception and in its physical arrangement on the monument, shows not only an effort of “translation”, which indicates a familiarity with the technical religious lexicon of the language, but also an attempt to integrate and interpret in Latin the terms used in the Greek divine onomastic sequence. This is exemplified by the term θαλλοφόρος, coined in an unprecedented way in the Greek text, to which some of the terms chosen to address the god in the Latin version can only be alluded to by cross-reference. Leveraging on the familiarity with the terms with which the god was frequently invoked in the Latin world (such as invictus, sanctus and pacifer, the latter at least found on coins),²⁷ the Latin text subtly adapts the attributes and qualities of the invoked god to the cultural context of the secondary language, presenting a functional representation of him to the Latin reader.²⁸

From Identity Integration to Claims of Cultural Prestige: Bilingual Inscriptions Bridging Institutional and Personal Realms

We now turn our attention to a pair of fragmentary bilingual inscriptions from Rome, specifically from the Regio XIV Transtiberim and the site of the ancient Horti Caesaris, later the Vigna Bonelli-Mangani area.²⁹ Here, there was a Palmyrene community, as evidenced by the extant epigraphic record in various mixtures of Greek, Latin and Palmyrene, together with a sanctuary and the local practice of the cult of Palmyrene gods.³⁰ Of particular interest among these inscriptions, which span both institutional and personal spheres, are two Palmyrenes, Heliodorus and C. Licinius, who recorded the construction of a temple (or temples?) in honour of the Palmyrene gods Bel, Yarhibol/Hierobolo and Malakbel/Malagbelo on two marble plaques dated to 116 AD:³¹

1.
Pro salute Imp(eratoris) [Caes(aris) Nervae Traiani Augusti]
C(aius) Licinius N[— et Heliodorus]
Palmyrenus [aedem Belo Iaribolo Malachbelo]
constitu[erunt sua pecunia?].

ʽΗλιόδωρος ὁ [Παλμυρηνὸς καὶ Γ(άϊος) Λικίνιος N—]5
τὸν ναὸν Βή[λῳ ᾿Ιαριβώλῳ Μαλαχβήλῳ Θεοῖς]
Παλμυρην[οῖς ἐκ τῶν ἰδίων ἀνέθηκαν].

For the safety of the Emperor [Caesar Nerva Traianus Augustus], Caius Licinius N[— and Heliodorus] the Palmyrene, built [the temple for Bel, Yarhibol, Malakbel with their own money?)]. Heliodorus the [Palmyrene and Caius Licinius N— built] the temple for Bel, [Yarhibol, Malakbel, gods] of Palmyra [with their own money?].

2.
[Pro salute Imp(eratoris) Caes(aris) Nervae Traiani Augusti]
[C(aius) Licinius N— et Heliodorus Palmyrenus]
[ae]dem Belo Iaṛ[ibolo Malachbelo].

[ʽΗ]λιόδω[ρος]
[ὁ Παλμυρην]ὸς καὶ Γά(ϊος) Λικίνος [N— Βήλῳ]5
[᾿Ιαριβώλῳ] Μαλαχβήλῳ Θε[οῖς Παλμυρηνοῖς].
L(ucio) Lamia C̣[armin]ịọ [Vetere co(n)s(ulibus)].

[For the safety of the Emperor Caesar Nerva Traianus Augustus, Caius Licinius N— and Heliodorus the Palmyrene], the temple for Bel, Yarhibol, [Malakbel]. Heliodorus the Palmyrene and Caius Licinius [N— for Bel, Yarhibol], Malakbel, Gods [of Palmyra]. [Under the consulship] of Lucius (Fundanius) Lamia (Elianus) and (Sextus) Carminus [Vetus].

The inscriptions are highly fragmentary, and it is only through comparison that the texts can be reconstructed in both cases. They are similar, but not identical, in the arrangement of the elements and formulae used in the two languages, with the Latin version preceding the Greek in both instances. Latin serves here as the formal and institutional language, due to the public location where these inscriptions were intended to be placed. This is evident from the typical Latin formula pro salute of the emperor at the beginning of the first inscription (which probably also appeared at the beginning of the second), as well as the code-switching for the consular year following the Greek version of the second inscription. Moreover, if we consider that Heliodorus (and perhaps Licinius as well) were speakers of Palmyrene Aramaic – as evidenced by the adjective Palmyrenus³² – their use of Latin would demonstrate their acquired mastery of this language (in addition to Greek) and thus their integration into the Roman cultural milieu. Greek, on the other hand, appears here as the “international” language common in Palmyra, or more generally as the lingua franca used in the Eastern Mediterranean, though not the language that strongly reflects a Palmyrene identity.

In this respect, it is noteworthy that, despite the fragmentary state of the two texts, while both the Greek and the Latin versions mention the Palmyrene origin of at least one of the two dedicators, this claim is not echoed in the divine interlocutors to which their cultic attention was directed. In Greek, they are collectively referred to as “Palmyrene gods” (Θεοῖς Παλμυρηνοῖς), a “topographic” marker seemingly missing from the Latin version. This slight discrepancy in the divine onomastic sequence between the Latin and Greek texts is not trivial. It can be seen as the result of a deliberate decision on the part of the agents, as J. N. Adams has pointed out, “not to complicate the Latin version with a reference to ‘Palmyrene gods’, which might have been lost on the local readers.”³³ In other words, this denomination is not “translated” into Latin because only Greek allows for a more precise delineation of the cultic background and/or the divine designation. Moreover, it was through the use of Greek and the omission in the Latin text of a strongly “Palmyrenean” charge of the invoked gods that, as Adams has observed, “the dedicators were able to pay lip service to their origins without presenting themselves as out-and-out aliens, given that Romans too were devotees of Greek culture and that the city abounded in Greek inscriptions of one sort or another”.³⁴

Another interesting pair of bilingual inscriptions can be found at the interface between the “official” and private dimensions. The pieces were engraved on two modestly sized marble altars (76x36x30 cm; 61x35x25 cm), which have identical features (in terms of shape, carving and calligraphy) and record dedications by the gens Papiria around the 2nd c. AD to Herakles/Hercules and Silvanus. The altars were discovered on the Esquiline, likely in the Orti Santarelli in 1663, and their texts read as follows:³⁵

1.
Herculi / Defensori / Papirii // Ἡρακλεῖ / Áλεξι/κάκωι / Παπείρι/οι.
To Hercules Protector, the Papirii (dedicated this) // To Herakles Averter of evil, the Papirii (dedicated this).

2.
Silvanus / Custodi / Papyrii // Σιλβανῶι / Φύλακι / Παπείριοι.
To Silvanus Guard, the Papirii (dedicated this). // To Silvanos Guard, the Papirii (dedicated this).

The two altars were likely installed in the private domus of the Papirii, a gens of senatorial rank. Basing his analysis on the site where they were found, F. Coarelli places the domus Papirii not far from the base of the Mons Cispius.³⁶ In both cases, the Latin text appears on the front face of the monument, while the Greek text appears on the back. This arrangement shows a preference for Latin, the more immediately visible language, even though the monuments were designed to be seen from both sides. The divine onomastic sequences follow a similar structure in both texts: the divine name has been paired with an onomastic attribute in Latin, and both are followed by their Greek “translation”. What emerges from the four texts is a mutual demand for the custody and protection of the home and the entire household, a prerogative shared by the two selected deities, Herakles/Hercules and Silvanus. In the latter case, the selected onomastic attributes (custos and φύλαξ) convey almost identical notions of protection given to a person or a place. For Herakles/Hercules, however, the Latin term defensor denotes a defensive-protective domain similar to that of Silvanus, while the Greek ἀλεξίκακος implies and demands protection from evil.³⁷

Once again, we have texts in a primary language, here Latin, where the Greek version, the secondary language which is supposed to reflect the first, is not an exact translation in all respects. Instead, the bilingual approach to the gods in these inscriptions allows another language to convey a particular nuance of the god’s protective powers. Thus, through the double text, the god is not only invoked for protection, but also to ward off evil. On the other hand, the widespread use of the epithet ἀλεξίκακος for Herakles in the Greek world indicates a technical understanding and familiarity with Greek religious and cultic vocabulary on the part of the dedicators.³⁸ As for the choice of such language, in the absence of concrete evidence, we can imagine that it was chosen by the gens as a symbol of high social status, given its cultural prestige.

“Complementing” and Enhancing a Divine Power through Bilingualism

The altars of the Papirii illustrate how bilingualism is often employed to convey religious communication in a more prestigious manner, while simultaneously constructing and achieving a more specialised meaning from this communication. A striking example of this dynamic can also be seen in a bilingual inscription, probably engraved in the 3rd c. AD, on a small marble tabula (20.5×26.5 cm) found in Rome (the exact location is unknown) and now preserved in the Vatican Museums (Galleria Lapidaria, 8, 54, inv. 5684). It records a dedication of an altar and a crater to Nemesis, inspired by a vision of her in a dream by Hermes, libertus Augusti and vilicus:³⁹

Μεγάλη Νέμεσις ἡ
βασιλεύουσα τοῦ κόσμ(ου)
Magna Ultrix Regina Urbis
ex visu
Hermes Aug(usti) lib(ertus) vilicus
eiusdem loci aram et
crateram cum basi bicapite
d(ono) d(edit).

The text begins with a two-line onomastic sequence praising the goddess in Greek, followed by a one-line Latin counterpart. The praise is separated from the agent’s name and the dedication formula, expressed only in Latin, by the technical expression ex visu. While this expression is often interpreted as a conventional epigraphic formula reflecting shared religious and social norms rather than an individual’s personal experience, its placement on a separate line and its position within the structure of the text may also invite a reading that gives greater weight to the experiential dimension. These interpretations are not mutually exclusive, and the layout of the inscription may be seen as open to both.

M. G. Schmidt suggests that the agent addressed the goddess Nemesis in both Greek and Latin in order to increase his chances of being heard by her.⁴⁰ Specifically, the use of code-switching in this context meets the practical need to supplement the ritual communication and to address the goddess exactly as instructed in the dream. In fact, the two laudatory formulae are not identical, but rather accentuate different facets of divine power, complementing each other intertextually: the Greek highlights the goddess’ universal aspect, while the Latin, in addition to labelling her as an “avenger”, contextualises her power within the dimension of the Urbs, where the dedicator operated. The agent’s bilingualism simultaneously indicates his likely Greek origins, as well as his wealth and elevated social status as an emancipated vilicus and former imperial servus who had risen to the status of libertus. By employing both languages, he demonstrates to society and to potential readers of the dedication his proficiency in both Greek and Latin, a sign of a well-educated individual.

The dedication to Hermes can be compared to another bilingual inscription discovered in Rome (again, the exact location is unknown) and currently on display in the Villa Albani (Portico Centrale, inv. 53). The inscription is engraved on a modestly sized marble herm (41x28x25 cm),⁴¹ dating to the 2nd c. AD, which was dedicated to Hermes by a certain Attis. Attis calls upon Hermes to protect himself, his family, and his friends:⁴²

〈:in fronte〉
‛Ερμῆς
Lucri repertor atque sermonis dator
infas palaestram protulit Cyllenius.
῎Αρτις τὸν ‛Ερμῆν εἵσαθ’‧ ‛Ερμῆς δ’ ’Αττίωι
[—]οι καὶ γένος φίλους θ’ ἅμα‧.5
[—]ε̣ἰσφέρω ‛Ερμῇ δ’ ἐπ’ ἄλειφα χέσομαι
φυ̣λ̣α̣τ̣τ̣ομένωι δῶμα τόδ’ ’Αττιάδων.
[—]ς ‛Ερμῆς δὲ Ι̣Ι̣ουν[—]νιος.
〈:in latere dextro〉
Interpres divum, caeli terraeq(ue)
meator
sermonem docui mortales atq(ue)
palaestram
[Caelorum incola toti]usque terrae 5
sermonis dator atq(ue) somniorum
Iovis nuntius et precum minister.
Hermes. Inventor of profit, giver of language, the child of Cyllene invented technical skills. Attis dedicated the herm; Hermes to Attis [—] and together his lineage and friends; [—] I bring to Hermes, and I will pour on him, who protects this house of the Attiades, the ointment. [—] Hermes (?). (I), messenger of the gods, mediator between heaven and earth, I taught mortals the language and technical skills. Inhabitant of the heavens and the whole earth, giver of language and dreams, messenger of Jupiter, and bearer of prayers.

Firstly, the inscription is clearly elaborate from a formal point of view. It features a bilingual polymetric text inscribed on two sides: on the front, there is a bilingual text in iambic (two Latin verses) and dactylic (Greek verses), while on the right side there is a Latin text composed of two hexameters and three Phalaecian hendecasyllables. This complex formal structure is further enriched by a sophisticated invocation. The first part, engraved on the front face, is written in the third person and offers a concise Latin description of the attributes of the invoked deity, whose name prominently stands out in Greek at the centre. The Latin text is followed by the Greek text, where we find the name of the dedicator, accompanied by a pledge, although damage to the monument makes it difficult to read. In return for the protection granted to him, his family, and his friends by the god – whose Greek onomastic attribute cannot be reconstructed – Attis ensured the performance of a ritual sacrifice, anointing the herm with ointment.⁴³ The dedication seems to end with Attis’ promise, but the additional Latin inscription on the right side of the monument introduces the god speaking in the first person, explaining his prerogatives in poetic form across several lines. While certain expressions are repeated between the first and second texts (i.e. sermonis dator…protulit palaestram / sermonem docui mortales atque palaestram), in the latter the god – initially described as the deity of gain and language – is addressed as a messenger who serves as an intermediary between mortals and gods, and as the inspirer of dreams. His ceaseless movement ensures his omnipresence both in heaven and throughout the entire earth.

The multifaceted portrayal of the divine figure is skilfully achieved through the creative interplay of two poetic Latin texts. Enriched with literary echoes and rare terms,⁴⁴ these inscriptions not only reflect the agent’s cultural background, but also convey his intention to endow the dedication with prestige and ritual precision, thereby effectively seeking the favour of the invoked power. The elaborate construction, alternating between first and third person, and the arrangement of the text demonstrate both Attis’ mastery of religious knowledge and his intention to display it in a practical way intended to please the god, with whom he claims a close relationship. This is constructed and set in stone through the elaboration of a direct address by the god, who seems to respond to Attis’ dedication, taking up his words while adding other elements. The two Latin texts are thus inseparable, artistically complementing one another to provide a fuller portrait of the divine power being addressed. In this context, Latin serves as a poetic language of prestige, providing a valuable medium for describing and highlighting the god and the prayer. Conversely, Greek – used both for the god’s name, prominently displayed at the centre of the front side of the monument, and for the actual request addressed to him – likely represented the language that Attis (whose cognomen alone suggests Greek origin) considered primary and most direct for personal religious communication.

Conclusion

The examination of the bilingual “asymmetrical” religious inscriptions from Regio I has allowed us to elucidate a variety of reasons that influenced the linguistic choices of speakers from different social backgrounds in multilingual environments when addressing divine powers in ritual contexts.

During the Imperial period, to which these inscriptions can be dated, Greek functioned both as a primary language for many speakers of the Eastern Mediterranean – chosen as the most direct language for personal addresses to the gods – and as a lingua franca or lingua ponte through which, as seen in the case of the Palmyrene Heliodorus, an identity not originally tied to Latin culture could be more readily conveyed to Roman audiences. Greek also served as a language of prestige, allowing high social classes such as the Papirii to express their culture and social distinction. However, Latin, as a means of expressing and displaying social and cultural integration, could fulfil this role as well. This is illustrated by the dedication of Attis, who deliberately chose Latin as the elevated poetic language suited to conveying a personal religious message.

With regard to the use of bilingualism in relation to divine names, the analysis of such “asymmetrical” inscriptions shows that the two versions (Latin and Greek, and vice versa) of the onomastic sequences adopted (as well as the dedication text itself) were not always identical and thus did not represent exact translations of one another. On the contrary, the dual linguistic code employed, whether the two texts were separate or integrated, functioned primarily as a valuable resource for creatively translating or adapting both elaborately constructed onomastic elements into the language with which one felt most comfortable and elements common to the linguistic-devotional practice of one culture (but not the other) into the secondary language. This facilitated the readers’ comprehension of the religious message of one or both languages. Moreover, the bilingual approach to the gods allowed for nuances or different information about the invoked deity’s powers and abilities to be conveyed in one language rather than the other, as well as through the integrated reading of both versions. A distinct definition of the god in Latin and Greek thus served as a consciously adopted strategy by the agents to construct and reinforce religious communication with greater completeness, precision, emphasis, and prominence. Conversely, the different divine onomastic sequences in the twolanguages, while conveying information about the divine instances, could at the same time reflect details about the identity, cultural background, and cultic practices of the agents behind the dedications, as in the case of the “Gods of Palmyra” or the topographical anchoring of Nemesis within the Urbs where Hermes operated.

In conclusion, the inscriptions examined illustrate how the bilingual approach to divine names often represented a dynamic effort to articulate and express a divine power from two different linguistic perspectives. Given the inherently plural nature of each deity, the verbalisation of the divine through diverse onomastic sequences necessarily involved multifaceted interpretations, shaped by the socio-cultural or intimately personal (multilingual) contexts in which the agents were embedded.

Abbreviations

Università degli Studi di Siena, ginevra.benedetti@unisi.it, ORCID: https://orcid.org/0000-0002-3682-7427. This contribution has been written as part of the project “Omnipotens. Manufacturing and Empowering Gods in Greco-Roman Antiquity” (OMEGA), 2022-2025, funded by the Spanish Ministry of Science and Innovation (ref. no. PID2021-127020NB-I00), within the National Plan for Scientific, Technical and Innovation Research (PEICTI 2021-2023), and led by Valentino Gasparini (Universidad Carlos III de Madrid).