• 1 Zappata 1996, 193-96, no. 46, pl. 32 (image). AE 1953, no. 26=CCID 433. Fowlkes-Childs 2012, 215-17 (...)
• 2 On Doliche, Winter 2011, 2017. On Jupiter Dolichenus, Collar 2011, 2012, 2013, 79-145; Blömer and W (...)

1Sometime in the 2^nd century or so, a soldier named Caius Iulius Dionysius, “Syrian by origin” (natione Surus), made a votive offering to his “ancestral Commagenian god” (pater[no deo] Comogeno) in the city of Rome. Joining him were his sons Iulius Romanus and Iulius Aquila and a priest named Sabeo.1 The god that Dionysius, a north Syrian, celebrated as his ancestral Commagenian divinity was probably Jupiter Dolichenus, who originated from a north Syrian cult site at Doliche and had become popular among soldiers, whether Syrian or not.2 But why did Dionysius opt to identify his divinity as the Commagenian god, a phrasing rarely attested in the Roman empire? What governed the premise that Jupiter Dolichenus was the ancestral god of people from a north Syrian region, when his standard name associated him with a local cult site within it?

• 3 Tert., Apol. 24.4 and ad nat. 2.8; Alpass 2015, 379.
• 4 IAquileia, nos. 3, 5-6, 105-7, 108-13, 115-26, 133-33, 138-51, 154-57.
• 5 Cadotte 2007, 25-112; Shaw 2013, 214.
• 6 Cadotte 2007, 609-10. A god named Pluto Variccala was worshipped at Thabraca; CIL 8.17330 (Cadotte (...)
• 7 But for “Dacian” gods, see IDR 3.2.209, 216-17; 3.3.103; 3.5.43, 46. The naming of the “Pontic Moth (...)
• 8 IG II³ 4, 1121; IG  II³ 4, 1127; IGLS 6.2731, Zeus Heliopolitanus and the Egyptian god. I thank Ala (...)

4Even so, the statements of the north African Christian Tertullian, active c. 200 CE, imply consistency. He indicates that certain provinces or their cities (civitates) promoted gods as their unique patrons. In two similar passages, he links Atargatis to Syria, Dusares (and Obodas) to Arabia, Belenus to the inhabitants of Noricum, Caelestis to Africa, and “minor kings” (reguli) or Varsutina to Mauretania.3 With the exception of Belenus, a god attested at Aquileia and in Noricum,4 Tertullian’s statements focus on provinces or regions of the Near East or north Africa that purportedly had a “provincial god.” Tertullian probably had decent information on the province of Africa Proconsularis, where he resided. Caelestis is a common label for the Phoenician goddess Tanit in Latin inscriptions of north Africa.5 Even so, very rarely does she appear as “the African goddess.” Varsutina is hardly attested in inscriptions, if at all.6 If Tertullian is correct, it could mean that the epigraphic record unevenly represents a uniform phenomenon of Roman provincial religious practice. Yet, such uneven representation may point to the inconsistency of the phenomenon itself. Gods perhaps rarely earned such recognition, and only among limited numbers of worshippers.7 If the Syrian goddess attracted wide attention, others did not. Even Isis and Sarapis, whom ancient worshippers certainly associated with Egypt, are almost never specified in surviving inscriptions as “the Egyptian” god or goddess.8

• 9 Henrichs 2019 [1976] 241-46. The Syrian goddess famously acquired many traits and epithets of Kybel (...)

5In light of such issues, this article explores the social mechanisms that endowed gods of Near Eastern origin with ethnic or regional labels like the Syrian goddess, the Commagenian god, or the Arabian god. It also explores, comparatively, the dedications made to “the Maurian gods” or goddesses occasionally identified as Maurian, Punic, or African in north Africa. The amplified inter-regional contact of the Hellenistic and Roman Mediterranean surely laid the groundwork for such labeling. As observed in scholarship, people who encountered foreign gods naturally applied their traditional worshippers’ ethnic or regional labels to them, along with standard epithets or attributes for elevating eastern gods.9 Yet, as this article emphasizes, the labeling also reflected forms of social differentiation intrinsic to the ethnic and provincial categories that Roman imperial management fostered in the Roman Near East or among expatriate Syrians. Perhaps more important, it sometimes emphasized the ethnic differences that existed among a god’s diverse worshippers. Some worshippers could claim greater legitimacy than other worshippers by framing a god as “ancestral” to them alone.

• 10 For religious markets, see Bendlin 2000.

6The forms that these differentiations could take were many. In some cases, ethnic labeling pointed to how people were identifying their own “ancestral” gods as the patrons of their ethnic or provincial populations, even as they promoted these same gods’ worship and “marketability” among foreigners or outsiders.10 In others, it represents how worshippers were conceptualizing the regional or ethnic foreignness of gods introduced to them by other peoples. Very often, such phrasing celebrated common Syrian, Arabian, or north African regional origins among worshippers and their gods. But at times, it could distinguish worshippers descended from Greek and Italian settlers, or at least people operating within their municipal paradigms, from longstanding Aramaean, Arabian, or Libyan populations. Finally, ethnic labeling sometimes circulated among people of relatively equitable stature. The invocation of a Commagenian god among soldiers and connected civilians is arguably one example. In some instances, however, it reflects how a social group empowered or privileged by imperial structuring appropriated a divinity in symbolically violent ways. After all, some people who celebrated the “Syrian goddess” were non-Syrians, like Athenians on Hellenistic Delos or various residents of Roman Italy, who had displaced Syrians as the priests and participants of her local cult. Others who did so were Roman soldiers, veterans, and colonists (or their descendants) who had adopted a goddess worshipped by local Syrians and reconstituted her according to their own perspectives.

• 11 Egea 2015.
• 12 For example, Bruneau 1970, 466-73; Baslez 1977, 67-97; Hörig 1984, 1568-70; Will 1985; Lightfoot 20 (...)
• 13 Lightfoot 2003 is the foundational work.
• 14 For the attestations, Hörig 1984; Lightfoot 2003, 537-38; Gavrilović Vitas 2021, 151-64.

7The most famous and longstanding example of a divinity described by an ethnic or provincial name is “the Syrian goddess,” known among Syrians primarily as Atargatis. Her most notable cult site and pilgrimage center has been immortalized by Lucian of Samosata. The goddess’ worship at Hierapolis-Manbog from the Iron Age to Roman times,11 her celebration as “Holy Aphrodite” on Hellenistic Delos,12 and Lucian’s portrayal of her cult site and its ethnic dimensions have attracted a wealth of treatments.13 In light of the immense scholarship on the Syrian goddess, we limit the discussion to how worshippers identified Atargatis as “the Syrian goddess” or a “Syrian Aphrodite.”14

• 15 Wilcken 1912, no. 101.
• 16 OGIS 733.

8Atargatis’ reputation as a “Syrian goddess” preceded Roman imperial intervention in the eastern Mediterranean and originated in Hellenistic social contexts. Even so, it largely took shape within social and religious frameworks generated by contact between Syrian worshippers and non-Syrian audiences, especially as Syrians migrated (or were violently displaced) to various parts of the eastern Mediterranean. Some early attestations for Atargatis as “the Syrian goddess” come from Pelusium in Egypt. In a papyrus from 222 BCE, a woman named Asia indicates that her husband Machata had been maintaining a small shrine to the Syrian goddess and Aphrodite Berenike on his property.15 His children or descendants apparently still worshipped her as such a generation later.16 In their inscription, Machata’s descendants identified themselves as Macedonians. Even so, on the basis of Asia’s name, commentators have long observed that she probably originated from the vicinity of north Syria and had introduced the worship of Atargatis to her husband. Machata may have been involved in the military campaigns in north Syria for which Ptolemy III is well known, and he conceivably came into contact with his wife Asia in that area (her social status at the time remains unknown, but she had possibly been enslaved and named “Asia” accordingly). At Pelusium, Malchata and Asia reconfigured Atargatis as the Syrian goddess, whether due to their family members’ non-Syrian origins or to describe the goddess to local Greeks and Egyptians at Pelusium. Whatever the scenario, certain people at Pelusium were reckoning the divinity as “the Syrian goddess” amid contact between Syrian and non-Syrian populations and, apparently, the movement of Syrians overseas. These conditions enabled the transfer of Atargatis’ cult from Syrians to non-Syrians.

• 17 Will 1985. See also n. 12.
• 18 Siebert 1968, 360, with commentary.
• 19 IDelos 2226.
• 20 IDelos 2247, 2257-58, and 2280. On the priesthoods, Lightfoot 2003, 44-45.
• 21 IDelos 2257-58, with 2281-82.

9We witness a similar pattern at 2^nd-century BCE Delos, where many north Syrians had expatriated and established a sanctuary for Atargatis and her consort Hadad.17 Their activity represents what is arguably the best-documented context for the worship of Atargatis in the Hellenistic period. It offers ample evidence for how worshippers increasingly invoked her as a “Syrian” divinity as her cult incrementally integrated Athenian worshippers and eventually priests. In the earliest dedicatory inscription from the Syrian sanctuary, which seems to predate the Roman transfer of Delos to Athenian control in 166 BCE, a priest named Nikon and his wife, the priestess (hiereia) Onesako, commemorated their construction of an oikos on behalf of the koinon of Syrian thiasitoi.18 After Delos’ transfer to Athenian control, Syrians apparently officiated the cult to Atargatis for several decades. In 128/27, a Hierapolitan named Achaios, dedicated (or restored) the naos, an adjacent oikos, and an altar for Atargatis and Hadad.19 For the next decade, a spate of dedications indicates that Atargatis’ priests were Hierapolitan.20 Even so, worshippers at the sanctuary were diverse, as two dedications by an Athenian demonstrate.21 The general picture of the sanctuary during this period (before 118 BCE) is that of Syrians officiating the priesthoods for their ancestral goddess on behalf of diverse worshippers, including Athenians.

• 22 Parker 2017, 162-63.
• 23 IDelos 2247.
• 24 IDelos 2248.
• 25 IDelos 2237 and 2248.
• 26 IDelos 2245, 2250-52.

10A noteworthy feature of the dedications made at the sanctuary on Delos during the tenure of Syrian priests, while being overwhelmingly inscribed in Greek, is that they refer to their patron divinities by their more traditional Aramaic names Hadad and Atargatis (or variants thereof). Otherwise, they sometimes invoke Atargatis anonymously as a “holy” goddess.22 In one example, a person venerated Hadad and Atargatis, his “ancestral divinities” (theoi patrioi), while a Hierapolitan was serving as priest.23 Yet, inscriptions of this period do not refer to Atargatis as “holy Aphrodite” or “the Syrian goddess.” Even Athenians and Italians who worshipped at the temple apparently invoked the gods by their Aramaic names. Only starting in 112 BCE, when Athenians increasingly served as priests of the goddess, did references to “holy Aphrodite” appear.24 Intriguingly, in a couple such attestations, her consort is called by his traditional name of Hadad.25 After 107, dedications to “Syrian goddess” become prominent, and the goddess’ traditional name disappears.26 Altogether, the recognition of Atargatis as the “Syrian goddess” at Delos happened as her cult incrementally involved non-Syrian priests and worshippers, particularly Athenians whose privileged position at Delos was being buttressed by Roman imperial intervention. Moreover, such non-Syrians, especially the Athenians, noticeably did not refer to holy Aphrodite or the Syrian goddess as their ancestral divinity even as they worshipped and celebrated her. They understood Atargatis to be a foreign goddess, the ancestral divinity of a foreign people whose cult they now overwhelmingly regulated.

• 27 For this material, Hörig 1984, 1565-69. The remarkable inscription recently discovered near Larisa, (...)
• 28 IG 9².95-110, with Hörig 1984, 1565-66; Valmin 1928-1929, nos. 1-2. Paus, 4.31.1 visited the temple (...)
• 29 I.Beroia no. 19.
• 30 I.Beroia, nos. 52-53.
• 31 Baslez 1999, 234.
• 32 Plut., Arat. 18.2.
• 33 Hörig 1984, 1565-66.
• 34 Lightfoot 2003, 537-38.

11While not as thoroughly documented, the residents of municipalities in mainland Greece and Macedon also worshipped “the Syrian goddess” during the Hellenistic period.27 Sometimes they even established a temple for her. For example, the goddess was worshipped at Physteion in Aetolia (as Syrian Aphrodite) by the 3^rd or 2^nd century BCE, and at Thuria in the Peloponnese, a temple was dedicated to the “Syrian goddess” by the 2^nd century BCE.28 Around the same time, a man in Macedonian Beroia made a dedication to Atargatis Soteira,29 though in subsequent centuries she attracted notice as “the Syrian goddess.”30 How the goddess arrived in these places is unclear, though various forms of connectivity were potentially in play. In some instances Macedonian veterans returning from Syria may have brought her cult.31 Also, as Syrians increasingly settled in the Aegean sea or joined the armies of Greco-Macedonian empires, commercial or labor aspirations may have brought them to coastal areas of the Peloponnese and Macedon. Plutarch reports that Syrians (or Greeks from Syria) were affiliated with the Antigonid garrison at 3^rd-century Acrocorinth.32 Furthermore, the fact that worshippers conceived of the goddess as overseeing slave manumissions in Greece and Macedon may suggest that her cult traveled with the movements of enslaved Syrians.33 However such movement and settlement occurred, it seems that the name “Syrian goddess” or “Syrian Aphrodite” became more popular as her cults increasingly attracted non-Syrian worshippers from local communities. In general, these non-Syrian worshippers conceived of “the Syrian goddess” as the ancestral divinity of a foreign Syrian people, though the vulgarization Diasura that emerged over time may reflect how some such worshippers had disassociated her from her Syrian origins.34

• 35 Beek 2016 on the eastern connections; Padilla-Peralta 2017 on slave religiosity.
• 36 IG 14.9. Syracuse.

12The presence of the Syrian goddess in Sicily bears some significance insofar as textual sources link her to a major slave rebellion on the island in the 130s BCE. According to Diodorus (34.2), a slave from Apamaea named Eunous organized a rebellion after he had visions of the divinity. Diodorus describes her as the Syrian goddess, though Eunous probably knew her simply as Atargatis. The rebellion reminds us that the worship of Atargatis traveled in part due to the violence of the slave trade.35 Of course, by every indication she accrued worshippers in Sicily who were not of Syrian origin and described her as the Syrian goddess, as a Latin inscription from Syracuse attests.36 If the goddess’ initial presence in Sicily during the 2^nd century BCE reflects the violent relocation of the enslaved from Syria, her cult eventually made the transfer to free worshippers from the local population. Yet, such transfer happened through the literal and symbolic violence wrought by human trafficking and enslavement. In this context of domination, free people of increasingly non-Syrian origin were ascribing the name “Syrian goddess” to a divinity brought to Sicily by enslaved Syrian bodies.

• 37 CIL 6.115-16, 399, 32462; 9.6099, 10.1554; Suet., Nero 56. Hörig 1984, 1572-73 for the Italian mate (...)
• 38 Though note IDR 2.346 (an Aramaic inscription with Atargatis’ name?).
• 39 RIB 1792.
• 40 RIB 1791. For an inscription associating Julia Domna with Caelestis, CIL 13.6671.

13The epigraphic evidence from the Roman imperial period overwhelmingly points to the worship of a “Syrian goddess” among people who were not Syrian by origin but who had maintained direct contact or formed social bonds with Syrians. In Rome and elsewhere in Italy, dedications to “the Syrian goddess” or its derivatives were standard.37 The same pertains to other regions.38 During the 160s at Carvoran in Britain, the prefect of the cohors I Hamiorum made an altar dedication to Dea Suria.39 About 50 years later, a military tribune named Donatianus, apparently from north Africa, celebrated her in an inscribed iambic poem to the Syrian goddess. It conflated her with Ceres (a popular goddess in north Africa) and perhaps even Julia Domna.40 Donatianus’ unit is unclear, but the presence of the cohors I Hamiorum at Carvoran was probably significant in exposing him to the Syrian goddess’ cult.

• 41 For the (relatively rare) pairing of Jupiter Dolichenus and the Syrian goddess, with priests of Syr (...)
• 42 IDR 3.5.221. See Petolescu 2006, 461-71 and Haensch 2006, 209 on sacred personnel. On Syrian cults (...)
• 43 Gavrilović Vitas 1987, 2019, and now 2021, 151-64.
• 44 CCID 430.

14As the Syrian goddess earned a reputation as a primary patron divinity of Syrians, her cult sometimes circulated in a relatively equitable way among soldiers in the Roman army, some being Syrian and others not. In such contexts, she was sometimes paired with other popular gods of Syrian origin whose worship also circulated throughout the Roman empire and eventually attracted non-Syrian worshippers or even priests; Jupiter Dolichenus is a prime example.41 Yet, worshippers identified these gods in terms that specified the local cult site from which they originated while the Syrian goddess appeared with her more regionalized appellation (even if Jupiter Dolichenus could sometimes be deemed “the Commagenian god,” as discussed in the next section). For such reasons, a 2^nd-century dedication raised at Apulum in Dacia for Jupiter Dolichenus merits attention. As its inscription states, the dedicator was a man named Flavius, son of Barhadadus, who was priest of Jupiter Dolichenus for the legio XIII Gemina. The dedication celebrated both “Jupiter Optimus Maximus Dolichenus” and the Syrian goddess (Dea Syria), great Caelestis.42 In this instance, the name of Flavius’s father Barhadadus points to north Syrian descent, and Flavius was presumably a member of a north Syrian expatriate community resident in certain cities of Dacia during the 2^nd century. As such, he was serving as priest to gods in whose sacred rites he was expert, but for a local legion whose soldiers had become exposed to Jupiter Dolichenus and the Syrian goddess through their military service. In other words, the priest was a man of Syrian ancestry who accommodated mostly non-Syrian worshippers when he invoked his goddess as “the Syrian goddess.” A possible comparative example is that of Iulius Apollonaris, celebrated as the priest of Jupiter Dolichenus and the Syrian goddess in an epitaph from Glamija, in Moesia Superior.43 Even so, we occasionally encounter evidence elsewhere for the pairing of Jupiter Dolichenus and the Syrian goddess from the vantage point of ostensible non-Syrians. In Rome, a freedman who celebrated Jupiter Dolichenus and “Iuno Assyriae Reg(ina) Dolichena” exhibits no Syrian background.44

• 45 Andrade 2013, 2014.
• 46 RPC 3.3446-57.
• 47 On this phenomenon, Lightfoot 2003, 37-38 (international perspective on 38).
• 48 Sartre 2001, 480-82; Pollard 2000, 39-42, 257-68.
• 49 IGLS 6.2929, with 2936; Hajjar 1977, no. 211=Yon and Aliquot 2016, no. 36. On cults at Niha and its (...)
• 50 IGLS 11.45.
• 51 Lightfoot 2003, 52; Cumont 1926, nos. 85-121; TEAD 3, nos. 144-66; Frye et al. 1955, 128-37; Duchât (...)

15 The worship of “the Syrian goddess” was apparently rooted in the premise among certain worshippers that Atargatis was a (or the) patron divinity of the Syrian ethnos, whether defined as people of Aramaic ancestry or, eventually, provincial residents of Roman Syria generally.45 But it was also contingent on forms of social contact that transferred her cult from Syrians to non-Syrian worshippers, who to some degree emphasized her foreignness or maybe even exoticized her. In addition to the widespread evidence for non-Syrians invoking her as “the Syrian goddess,” the Syrians who also did so were anticipating non-Syrians who venerated her. Intriguingly, the phenomenon did not merely involve the expatriation of Syrians throughout the Mediterranean, but it also pertained to the settlement of Italians or soldiers in certain parts of Syria under Roman rule. This may in part explain attestations for “the Syrian goddess” on the coins of Hierapolis itself starting with the reign of Trajan, when Hierapolis first started to mint municipal coins.46 To be sure, this phenomenon was related to the Hierapolitans’ promotion of their home city as a site of pilgrimage for a variety of people and the adoption of “an international perspective” on Atargatis herself.47 Yet many Roman legions and accompanying auxiliaries were stationed in north Syria during the 1^st century CE to defend the Euphrates frontier,48 and the fact that that the earliest issues of municipal coins celebrating “the Syrian goddess” date to Trajan’s reign suggests that she was already attracting worship as such in earlier periods. “The Syrian goddess’” also noticeably appears in dedications in parts of Lebanon where the settlement of Italians or veterans was intensive in comparison to other parts of Roman Syria. The bilingual inscriptions celebrating Dea Suria Nihathe(na) in the mountains of Lebanon and to Dea Syria Geranensis at Beirut arguably reflect how a localized cult for a popular goddess was attracting the attention of Latin-speaking worshippers descended from the Italian veterans settled at Beirut, Heliopolis, and adjacent areas of the Bekaa valley. Intriguingly, both dedications referred to Atargatis in Greek and to the Syrian goddess in Latin.49 In Mt. Hermon, located just to the south of the Bekaa valley, Atargatis could be identified in a Greek inscription simultaneously by her name and as “the Syrian goddess of the Hieropolitans.” Such alternative forms of labeling apparently reflect the retention of Atargatis’ name by the longstanding local population and the popularity of “the Syrian goddess” among soldiers, veterans, or colonists who had settled in the vicinity more recently and had come into contact with a cult for her there.50 A striking contrast can be found at Parthian-era Dura-Europos, where Atargatis is never called “the Syrian goddess” despite being worshipped by diverse residents.51

• 52 Syr. D. 1, 22, in Lightfoot 2003, whose treatment provides important commentary throughout. Also El (...)

16Such premises bring us to the most famous source for the worship of the Syrian goddess from antiquity. In On the Syrian Goddess, Lucian famously portrays himself as a Syrian who is describing the patron divinity of the “Assyrians” for a foreign audience. He thus identifies Atargatis as “the Assyrian Hera” without mentioning her actual name, and he also states that she could not be likened to a single Greek divinity since she instead encapsulated aspects of most Greek goddesses.52 Despite the text’s fundamentally intricate nature, its treatment is very consistent with how Syrian provincials portrayed Atargatis to Greek and Roman “outsiders” as a Syrian version of a Greek goddess, or simply as “the Syrian goddess.” As the inscriptions show, people of Syrian backgrounds often promoted her as a Syrian Aphrodite or a Syrian goddess to foreign worshippers, who responded in kind. This happened most often outside of Syria itself, in places to which Syrians had brought their goddess. But as Roman imperialism intensified the presence of Latin-speaking soldiers and veterans in north Syria and parts of the Bekaa valley, “the Syrian goddess” could be celebrated there too.

• 53 Winter 2011; Facella 2006; Merz and Tieleman 2012; Reidel 2018; Blömer et al. 2021. See IGLS 1.42-4 (...)
• 54 Versluys, 2012, 56-59. But see IG 14.885; RIB 758.

17The evidence for the existence of a divinity defined as the “Commagenian god” is not ample. Perhaps this should not surprise. The client kingdom of Commagene was located in a part of north Syria traditionally called Kummuh, and it apparently became an identifiable district after the region joined the province of Syria.53 Even so, “Commagenian” was not used to denote ethnic or regional identity by north Syrians or anyone else.54 Yet, units of the Roman army recruited in north Syria were often identified as Commagenian. This concept thus determined how certain north Syrians who joined the Roman army classified themselves, and the formulation of a “Commagenian god” accordingly happened in military contexts. Worshippers were overwhelmingly soldiers, their dependents and relatives, or people resident in regions where north Syrian units or soldiers were deployed.

• 55 Kennedy 1980, 97-101; Tentea 2012, 41-45. For this unit and the Commagenian god, Collar 2011, 240-4 (...)
• 56 IDR 3.3.67.
• 57 IDR 3.3.138. Haynes 2013, 229-30; Stoll 2007, 470; Gavrilović Vitas 2021, 114-17. An altar for Turm (...)
• 58 IDR 3.3.298 (Ampelum); IDR 3.3.299. For “Suri negotiatores” and Jupiter Dolichenus, IDR 3.5.218 (Ap (...)
• 59 CIL 3.10243=CCID 208. A man from Doliche was in the province in 228. CIL 3.3490=CCID 185. On names (...)

18The movement of the cohors II Flavia Commagenorum equitata sagittariorum and its religious worship figures prominently here. Raised in north Syria and stationed in Moesia during the late 1^st century and into the opening decades of the second, it was consistently deployed at Micia in Dacia thereafter.55 Its soldiers in the vicinity made dedications to Jupiter Dolichenus, Jupiter Turmazgades, and other gods with north Syrian origins. For example, at Micia, one of them made a dedication for Jupiter Dolichenus,56 and another dedicated an altar to Jupiter Turmazgades.57 Over time, it seems that the descendants of Syrian veterans in nearby locations in Dacia, whether military or civilian, began to label a god brought to the region by a “Commagenian” unit as the ancestral god of “Commagenians,” or otherwise north Syrians. For example, civilians raised dedications for Jupiter Dolichenus as an explicitly Commagenian god at Ampelum; their names often suggest north Syrian ancestry.58 The decurion of a city in Pannonia Inferior (but located at the threshold of Dacia), a man named Apollonius, referred to a Commagenian god in similar terms, and he also identified the god as paternus, thus suggesting that he claimed some north Syrian background or ancestry.59 Altogether, the discernable pattern is that soldiers or civilians worshippers with linkages to military contexts in Dacia were sometimes treating Jupiter Dolichenus as a “Commagenian god” insofar as he was a key divinity of soldiers in units initially mustered in north Syria and labeled “Commagenians” by the Roman army. Yet, they only called him “ancestral” when they claimed some sort of north Syrian background or ancestry. The labeling communicated their north Syrian origins and a unique authenticity as the god’s worshippers not necessarily possessed by others who venerated him.

• 60 Zappata 1996, 193-96, no. 46, pl. 32 (image) offers the preferred reading. Fowlkes-Childs 2012, 215 (...)
• 61 For description and image, see Zappata 1996, 193-96, no. 46, pl. 32 (image). In the same volume, Bi (...)
• 62 Zappata 1996, 125-27, no. 15, pl. 7, with 115-25, nos. 13-14, Pl. 5-6.

19Outside Dacia, the rare celebrations of a “Commagenian god” also betray close linkages to a military context, the recruitment of north Syrians, and some sort of affiliation with Jupiter Dolichenus. In some cases, the worshipper was clearly Syrian by origin and considered the divinity to be an ancestral divinity or the god of an ancestral homeland. We have already witnessed how Caius Iulius Dionysius, a soldier in the classis praetoria Misenensis at Rome, made a dedication in which he identified himself as Syrian and celebrated his “ancestral” (paternus) Commagenian (Comogenus) god.60 The inscription does not refer to Jupiter Dolichenus in explicit terms, but the base still bears traces of what can be plausibly reconstructed as the bull on which Jupiter Dolichenus normally stands in the iconography.61 At the same time, the cult to Jupiter Dolichenus clearly was adopted by non-Syrians, and some of these also identified him as a “Commagenian god.” For example, a figure named Aurelius Magnesius was involved in dedications to both Iupiter Dolichenus and the “great god of the Commagenians” at Rome in the dolicheum on the Aventine. His father apparently officiated as priest there. But despite his clear veneration of the “Commagenian” god, the dedicator seems to have conceived of him as the ancestral divinity of a foreign people.62

• 63 On the movement of the cult for Jupiter Dolichenus, Collar 2013, 79-145 and Blömer and Winter 2011.

20The treatment of Jupiter Dolichenus as “the Commagenian god” did not circulate widely. Apparently rooted in a tendency for the Roman army to classify north Syrians as “Commagenian” and to organize them into military units with the name, its attestations are localized to Rome, parts of Dacia, and occasional outliers. Yet, the formulation was undoubtedly significant to the north Syrians that invoked Jupiter Dolichenus as their ancestral “Commagenian god” during the 2^nd century, when Jupiter Dolichenus was becoming increasingly popular among soldiers and, by extension, civilians at Rome or in Dacia.63 Some north Syrians were identifying Jupiter Dolichenus emphatically as “their ancestral god” in ways that distinguished them from non-Syrians who also worshipped him. Yet, in some cases, north Syrians familiar with the Commagenian regional category promoted by the Roman army took the additional step of identifying him as both ancestral and Commagenian. Some non-Syrians even adopted this formulation too, while framing “the Commagenian god” as a foreign divinity.

• 64 Smith and Day 1989, 131-32; Lichtenberger 2003, 184, 311.

21Unlike the gods so far discussed, the Arabian god (theos Arabikos) apparently only surfaces in the province of Arabia, and in select cities there. The majority of existing attestations are localized to Gerasa, where the Arabian god is occasionally described by the epithets hagios and epekoos. The god is also apparently attested once at Pella, in a slightly different form: theos ouranios Arabios.64 On the basis of Tertullian, one would expect this divinity to be Dusares.

• 65 IDelos 2311, found in context with 2307-10, 2312 (2308 for origins from Iamnia). Lichtenberger 2003 (...)
• 66 Lichtenberger 2003, 223; Healey 2001, 142-43.

22Yet, at Gerasa, where the theos Arabikos is most widely attested, a divinity called Pakeidas appears in several inscriptions, often with the same epithet of hagios. Moreover, attestations to Pakeidas and the Arabian god happened in the same complex south of the temple of Artemis. This would suggest that Pakeidas was the theos Arabikos. But who was Pakeidas? To date, the god has only been attested at Gerasa, a site that has produced no surviving dedications to Dusares. A possible exception is from Delos, where a person made a dedication to “Pakeidokosos” in a sacred space containing dedications to gods, or from people, with origins in Palestine.65 Some surmise that Pakeidokosos was the Idumaean god Qos, whose name had been attached to the Semitic root “pqd” in ways that denoted that the god was a commander or overseer. It is perhaps possible that this Idumaean god had been brought to Gerasa and invoked in a similar way there, but it seems more likely that we are witnessing a parallel phenomenon. People at Gerasa were invoking a different god with the Semitic root “pqd.”66

• 67 Alpass 2015, with Acqua 2015 on the coins of Adraa. Tertullian claims that Arabians worshipped Obad (...)
• 68 For their local framing, Lichtenberger 2003, 224-25. On Dusares’ origins, Healey 2001, 80-82.

23Dusares is not attested at Gerasa. His worship throughout the province of Arabia, which is attested as far north as Bostra and Adraa, was clearly inspired by how the Nabataean dynasty promoted him throughout its territories.67 Yet, dedications to Pakeidas and “the Arabian god,” and not Dusares, surface at Gerasa, which was never integrated in the Nabataean state. The existing evidence allows for no secure solution. Even so, Dusares, a god promoted and arguably made famous through Nabataean royal patronage, had probably earned widespread recognition as a primary regional or ancestral divinity of Arabians (as Tertullian claimed), whereas the Gerasenes were promoting their local Arabian god Pakeidas in somewhat similar terms (but with much less regional success).68 If the Nabataean dynasty had facilitated the worship of Dusares far beyond the palace city of Petra, the Gerasenes were not able to promote their divinity to the same stature, even if they called him “the Arabian god.”

• 69 For the formation of the province and Gerasa’s situation in the 1^st century, Sartre 2001, 730-32. G (...)
• 70 This article accepts that “Arabian” was foremost a provincial label or a formulation of Greek and R (...)
• 71 Gerasa harbored a Greek settlement, or had been classified as a Greek polis, by the mid-Hellenistic (...)
• 72 IGerasa 17-18.
• 73 De Vaux 1951. Another inscription from Gerasa’s hinterland is Gatier 1982, no. 3.
• 74 Emphasized as early as De Vaux, 1951, 23-24.
• 75 Sartre 1976 is the inscription.
• 76 IGerasa 19-22: 150 (21) and 155 (22); Lichtenberger 2003, 221-25. The father of Demetrios from IGer (...)
• 77 Gatier 1982, no. 3.

24What factors governed the tendency for Gerasenes to call Pakeidas “the Arabian god?” The dated dedications to Pakeidas at Gerasa were noticeably made before the province of Arabia was established. Those to the Arabian god were made in the decades after the province was formed and Gerasa was transferred from Syria to Arabia.69 This suggests that the formation of the province of Arabia, and the identification of its inhabitants as Arabian provincials, had an impact.70 It also seems, however, that people who claimed Greek ethnic or municipal backgrounds but who were Arabian by provincial classification deemed Pakeidas an ancestral god of Arabia’s indigenous inhabitants. They thus called him the Arabian god (and not Pakeidas) through a hierarchical appropriation that may have occluded the god’s traditional name in the long-term.71 The dedications that refer specifically to Pakeidas were made by Ameros, son of Ragelos, the bearer of Arabic names, and a slave (of Arabian background?) named Tyche by her master.72 By contrast, the dedicators to “the Arabian god” overwhelmingly have Greek names. The earliest dated dedication from Gerasa’s hinterland, made in 126-127,73 records the names of Leontas and Prepon for the dedicator and his father. When Leontas honored “the Arabian god,” he apparently did so from the vantage point of an ethnic Greek (who was also a provincial Arabian).74 Leontas’ dedication, intriguingly, was made roughly when the Gerasenes credited the governor of Arabia with overseeing homonoia in an honorific inscription, which suggests that some sort of dissension had recently occurred.75 As for Gerasa’s urban landscape, the earliest dedications to the Arabian god found there date to 150-155.76 Intriguingly, the dedicators overwhelmingly have Greek names and invoke the health of the imperial family. The conceptualization implies that worshippers were treating “the Arabian god” (presumably Pakeidas) as the patron divinity of the Arabian province formed by Roman management and as the ancestral divinity of the province’s indigenous ethnic population. Over time, this persisting appellation gained some popularity among locals of Arabian descent.77 Even so, the phenomenon did not attract widespread appeal, and external observers like Tertullian instead deemed Dusares, who had attracted more regional attention under the Nabataeans, as the god of the Arabians.

• 78 Camps 1990, especially 150-53.
• 79 CIL 8.2637-41, with others listed by Camps 1990, 149.
• 80 Camps 1990.
• 81 Hobson 2016, 111-14; Le Bohec 2005, 53-59; Briand-Ponsart 2011, 166-86. For cities, Lepelley 1979-1 (...)
• 82 Le Bohec 1989; 2005, 97-108, 120-26.
• 83 Hamdoune 2011, esp. 52; Coltelloni-Trannou 2011, 89-105: Laporte 2011, 115-29.

25Starting in the mid-2^nd century, local divinities whom civilians knew by their traditional Libyan names were being classified as “Maurian gods” (dii Mauri) by worshippers. According to the definitive treatment, these were overwhelmingly imperial magistrates and soldiers, oftentimes with origins outside North Africa.78 By all indications, this is accurate. Lambaesis, a military hub, has produced more attestations than any site.79 Regionally Numidia and, to a lesser extent, Mauretania Caesarensis have produced the most such attestations. This is probably linked to how surviving dedications to gods with traditional African names and backgrounds overwhelmingly come from these provinces or regions.80 Surprisingly both types of dedications are absent from Mauretania Tingitania even though it had longstanding Libyan populations. It is unclear why this is, but several possible factors are worth highlighting. One is that Italian and veteran settlers, extensions of Roman citizenship, and the organization of existing cities according to Italian municipal models were not distributed evenly throughout Roman north Africa. By the mid-2^nd century, the province Africa Proconsularis, including the region of Numidia which constituted its western end, had been intensively settled by Italian settlers, thus placing the area’s more indigenous populations in contact with imperial newcomers and their descendants. This heavily urbanized area encountered frequent grants of Roman citizenship and the organization of settlements as Italian coloniae or municipia.81 It also had the amplest military presence. By the 2^nd century, Lambaesis in particular harbored a legion of troops and accompanying auxiliaries. This made Numidia a region where different ethnic populations and language groups maintained intensive contact.82 Despite some auxiliary units and municipal centers, Mauretania Tingitania never witnessed the same scale of Italian settlement, municipal structuring, military deployment, or connected urban life as Africa Proconsularis and Numidia, or even Mauretania Caesarensis.83

• 84 Briand-Ponsart 2011; Coltelloni-Trannou 2011; Brett and Fentress 1996, 25-42.
• 85 On Punic and Libyan north Africa, Quinn and Vella 2014.

26Another factor was how people in different north African zones conceived of their region’s heritage as either Libyan or Phoenician/Punic. Numidia had been part of the province of Africa Proconsularis during the 1^st and 2^nd centuries, but its inhabitants, like those of the Mauretanian provinces, leaned more toward a Libyan framework for their region’s ancestral past than the Phoenician/Punic focus farther east in Africa Proconsularis and Tripolitania.84 In theory, references to “Maurian” gods distinguished them from both Italian and Punic gods and identified them as the ancestral divinities of African Libyans.85

• 86 Briand-Ponsart 2011, 187-88.
• 87 CIL 8.2637.
• 88 CIL 8.8435
• 89 CIL 8.9327.
• 90 CIL 8.9195=AE 1993, no. 1781.
• 91 AE 1968, 590.
• 92 CIL 8.20251-52.
• 93 M. Caes. 2.3.5; Keulen 2014, 131-37; Apuleius, Apol. 24: Mattiacci 2014. Shaw 2014, 530-32.

27Significantly, the earliest attested worship of the Dii Mauri happened well after the province of Mauritania Caesarensis had taken shape. It also largely occurred after Numidia was made into its own province by the Severan dynasty.86 This explains why Tertullian, a contemporary of the Severans, noted that the Maurians worshipped their “rulers” (reguli). Yet, worshippers apparently framed the Maurian gods as the ancestral divinities of ethnic Libyan peoples or communities, even if they could also conceive of them as patron gods of the provincials in Numidia and Mauretania more generally. Most surviving dedications belong to governors and officers who occupied brief posts in north Africa. These apparently worshipped the Maurian gods as the ancestral divinities of “foreign” ethnic or provincial populations through a form of hierarchical appropriation that often occluded the gods’ indigenous names. Thus, we list examples of dedications by a legate named Fuscinus at Lambaesis in the late 150s,87 a praeses named Saturninus at Sitifis in the 250s,88 and a prefect of the classis Germanica at Caesarea in the 3^rd century.89 Similarly, an equestrian official named Donatus who had governed a local Libyan gens made a dedication at Rapidum in the 3^rd century.90 Even so, occasionally municipal elites or other civilians adopted the formulation. In 203-4 at Musti in Numidia, Perellius Saturninus celebrated the Dii Mauri Castores and identified himself as the duumvir of his municipality.91 At Satafis, a local named Saturninus ornamented a temple; another inscription celebrates the numen Maurorum.92 The writings of Fronto and Apuleius show that provincial Africans who emphasized Italian ethnic or municipal origins could distinguish themselves from those who were Libyan or African through ethnic lineage, or vice versa.93 Residents of North Africa who worshipped the Maurian gods apparently associated them with “foreign” ancestral populations, whether because of their descent from Italian settlers or their Italian-modelled municipal life.

• 94 AE 1935, no. 39, Cadotte 2007, no. 338; IAlg. 1.2033, Cadotte 2007, no. 344.
• 95 AE 1953, 48; Cadotte 2007, no. 87.
• 96 Tert., ad uxorem 6.4; de exhortatione castitatis 13.2.

28Despite the patterns governing the worship of the Dii Mauri, we also witness that civilians in north Africa worshipped female divinities that they classified as African, Maurian, or even Punic and associated with the Roman empress. The occasions in which Ceres receives ethnic epithets among worshippers in north Africa warrant comment. At Thagura and Madouros, located at the western (Numidian) end of Africa Proconsularis, Ceres is called Ceres Maurusia (Maurian Ceres) by dedications. At Thagura, the dedicators are a married couple who also label her Augusta; the wife was apparently her priestess. At Madauros, a basket carrier for the goddess (canistraria) also described her as Augusta.94 At Mactaris, we hear of the Cereres Punicae for whom a certain Nonnia Primitiva was priestess.95 Tertullian refers to a Ceres Africana.96 Significantly, people who applied these epithets to Ceres were ostensibly of north African regional background, as opposed to visiting Roman government officials. Their ancestral origins are obscure, but they clearly believed that the Cereres worshipped in their province, and by its ancestral population, were distinctively Maurian-African (in Numidia) or Punic-African (farther east in Africa Proconsularis).

• 97 Cadotte 2007, 25-112.
• 98 AE 1980, 0595bis=2003, 0948. On this inscription, Canto 2003, esp. 308-10.

29We can remark similarly about Caelestis in Africa Proconsularis, which was the Punic goddess Tanit.97 Her ubiquitous presence in the province and other parts of North Africa surely explains why Tertullian, from Carthage, considered her the goddess of Africa. Since dedications to her heavily populate the epigraphy of north Africa, worshippers presumably deemed her a primary divinity of African populations. Yet, only rarely do inscriptions call her “African” explicitly. At Lugo in Spain, a freedman of Augustus made a dedication to Africa Caelestis. The name of the freedman, Saturninus, suggests north African origins and membership in the household of Septimius Severus.98 While in Spain, he perhaps made explicit what many people in north Africa like Tertullian presumed: their Caelestis was an African divinity, and even the quintessential Punic or African one.

• 99 On Phoenician categories, see now Quinn 2018; on African or Libyan, Shaw 2014 and Briand-Ponsart an (...)

30Certainly, explicit attestations of gods that label them Maurian, Punic, or African are not widespread. It is difficult to say whether they point to a conceptual formulation common among worshippers of north African gods and goddesses, as Tertullian’s testimony suggests. Yet, the formulation does reflect the widespread sharpening of “Maurian,” “Punic,” or African ethnic and regional consciousnesses engendered by Roman imperial intervention and how Roman authorities classified the various civic and tribal populations of north Africa. After the 1^st century, provincials in eastern Africa Proconsularis and Tripolitania increasingly embraced premises of a Punic past as their “African” past. By contrast, people in both the Mauretanian provinces and Numidia conceived of a Maurian heritage that they defined as “African.”99 As Italians and veterans settled in these regions, they situated themselves in Maurian-African or Punic-African regional categories even if they did not trace those particular ancestries. Due to such ethnic and provincial conceptualizations, worshippers considered some longstanding gods of north Africa to be distinctively Maurian, Phoenician/Punic, or African.