Divine Marriage

Amongst the diverse religious statuary of the Roman provinces exists a body of sculpture which can provide unique insights into the processes of religious transformation. The sculpture in question depicts a gendered divine couple; a male typically of classical origin, and a female of a native, non-classical origin. These pairings, known as theogamy, or divine marriage, often feature the classical deity Mercury, who was frequently, although not exclusively, coupled with the goddess Rosmerta. Indicative of the popularity of this pairing, representations of Mercury and Rosmerta have been recovered from across the Rhone Valley (Esperandieu 1931), including at Glanum (Salviat 1979), Autun, Pagny-la-Ville in Burgundy (Esperandieu 1910), Metz (Green 1989) and Bierstadt (Esperandieu 1931). In addition to this eminent and popular pairing, the goddess Sirona was worshipped alongside Apollo at a number of Treveran sites, featuring on a number of statues and inscriptions (Watson 2005: 87). Mars Cicollius and Rosmerta (de Vries 1961: 141), and Mars Loucetius and Nemetona (Green 1989: 45) were similarly worshipped as divine pairings.

Whilst most widely evident in Germany and in central and eastern Gaul, the recovery of sculptures from the Cotswolds (Henig et al. 1993), Bath (Cunliffe and Fulford 1982) and Hadrian’s Wall (Phillips 1977; Coulston and Phillips 1988), is indicative of its practice in Roman Britain (see Figure 1). The collection of sculptures pertaining to divine marriage in Roman Britain is anepigraphic, and whilst classical deities were readily identifiable according to an established visual vocabulary of representation, archaeologists turned to the continental counterparts in an attempt to identify the female consort. A number of Gallic divine marriages were accompanied by epigraphic dedications to Rosmerta, known as ‘the Great Provider’, and feature the deity with attributes associated with fecundity. The mirroring of these attributes in a number of the sculptures of divine marriage in Roman Britain prompted the identification of the female deity as Rosmerta. Although there is neither certainty nor agreement on origins of the female component of the pair (Green 1998: 27; Yeates 2008; Webster 2015: 139–141), what is widely accepted is that this deity is certainly not Roman.

Figure 1 

Mercury and Rosmerta from the Shakespeare Inn, Gloucester (Drawing: Jane Webster, unpublished. Reproduced with permission).

Traditionally regarded as the apogee of Romano-Celtic syncretism, typical readings of the divine couplings positioned the classical deity as the dominant partner. In some instances, this was due to the presence of epigraphic dedications which addressed the classical deity alone (Green 1989: 43; Watson 2005: 86). In others, it was due simply to the pervasive idea of male dominance and the paternalism of Rome. This cross-religious coupling was hailed as the ultimate expression of a simplistic, benign syncretism between two polytheistic religions where ‘the ruling element was Roman’ (Haverfield 1923: 67). Inspite of the more recent reorientations of our understandings of divine marriage (Webster 1997; 2001; Aldhouse-Green 2003, 2012), Joseph Murphy’s (2011) interpretations of Santerian religion has led me to question whether we have been overlooking important facets of divine marriage. The approach adopted by Murphy has provided an exciting new perspective on cross-cultural interactions in the religious sphere, and, when applied to the field of Roman archaeology may furnish equally exciting new insights into the phenomenon of divine marriage.

New Perspectives: Santerian Religion and the Way of the Saints

In 1511, the first of 702,000 African slaves arrived in Cuba, many whom were of Yoruban origin (Lefever 1996: 319). The enslaved, as elsewhere in the colonies, were prohibited from practising their indigenous religion, which included the worship of divine intermediaries named Orishas. In spite of this oppression, many of their religious traditions survived in the religion which came to be practiced by the slave population, known as Santeria or ‘the way of the saints’. The development of this religion and the interaction between indigenous and non-indigenous elements is central to Joseph Murphy’s (2011) inquiry, and in drawing out some of Murphy’s key findings, it is perhaps most useful to begin with Santerian altars. These altars symbolise the cultural creativity that can arise from religious interactions, with images of Catholic saints placed alongside the symbols and attributes of an Orisha. Ostensibly, the shrine is simply devoted to a Catholic saint, as images of saints often provide the only example of religious anthropomorphic representation. The lack of anthropomorphic representation of the Orishas does not however mean that the Orishas are not represented. Each Orisha has a panoply of associated accoutrements and paraphernalia, including colours, plants, animals, foods and beads (Brody 1993: 7). These attributes can be found adorning the shrine of their associated Orisha. An altar to Chango will often feature his sacred stone, the otan, housed in a decorative wooden bowl. An assortment of accessories associated with Chango, including an axe (oche), beaded ram horns (ogues) and drums typically complete a Chango altar (Murphy 2011: 146). On a Creole Chango altar these objects are frequently accompanied by an image of the Catholic Saint Barbara (Brody 1993: 7–8).

The juxtaposing of two belief systems on one altar was understood by Murphy (2011: 158) to reflect ‘a complex religious world where Orishas and saints could be seen as both the same and different’. Their association, he argued, was a way to simultaneously differentiate and connect two very different religious and cultural systems. The Saint-Orisha associations were far from random. Rather than operating simply as a veneer through which to continue the worship of indigenous deities, they were thoughtful and selective. Parallels were drawn between Chango and Saint Barbara for a multitude of reasons, including their association with fire and lightening. Whilst such parallels resulted in an intimate association between the two deities, crucially, these deities, according to Murphy, were not simply considered the equivalent of one another, but one component of each other. How and why a saint and Orisha could be seen as both the same and different is due, according to Murphy (2011: 158), to the very nature of religion and religious expression:

‘…religious symbols are irruptions of the sacred realm onto that of the profane, the world of absolute being into the temporal and conditioned world we live in. To express this extraordinary coming together, religious people use paradox in an attempt to capture and experience the ineffable transcendence of the sacred world. The results are meaningful but non-rational, even impossible juxtapositions.’

These ‘impossible juxtapositions’ are evident in Santerian religion according to Murphy, in that Saints can be Orishas for example, whilst maintaining their distinct identities, just as the Orisha Chango for example could be both male and female. The Saint-Orisha interactions were argued to enhance the spiritual experience of a follower, creating new beliefs systems rooted in both the new and the traditional, and were held to form part of a body of knowledge that became increasingly available to religious adherents following initiation. Followers of the religion were introduced to hidden meanings and different levels of knowledge depending upon their level of initiation, and Murphy thus links the different aspects of divinities to particular social and spiritual standings, suggesting that ‘Saint Barbara is the outer level, the outer face of Chango, appropriate to public, Catholic levels of discourse’. Chango, on the other hand ‘is inner knowledge appropriate to the more intimate world of Lucumi ceremony and aspiration’ (Murphy 2011: 160).

This acknowledgement that varying levels of religious knowledge existed according to initiate status is an important process in understanding the dynamic and multivocal perceptions of the divine within a given cultural group. For some worshippers, the saints may have genuinely operated as a mask for Orisha worship. For others, depending on their initiate status and knowledge, the saints were seen more as aspects of an Orisha whilst maintaining individual identities. Whilst the acknowledgement that coloniser-colonised belief systems co-existed is far from new, the suggestion that indigenous and non-indigenous deities may have come to be seen as one component of another is an intriguing concept that provides a new window through which to view religious interaction.

Revealing Iron Age Religions and Cosmologies

Whether Murphy’s insights can provide a useful conceptual framework for understanding the phenomenon of divine marriage rests upon the existence of a correspondence between belief systems, that is to say, whether Santerian beliefs regarding divinities as the same-but-different were compatible to some degree with the Iron Age belief systems which contributed to nascent religions of the provinces. This is not to suggest a direct comparison of religious beliefs is necessary, but that a loosely shared system of beliefs regarding the nature of divinities and their capacity to possess multiple forms existed. The problematic nature of classical and vernacular texts relating to Iron Age beliefs, coupled with the absence of written evidence by Iron Age populations means that we may never truly understand the complex cosmologies that were present. Nevertheless, the archaeological record and particularly the art of the Iron Age may provide a tangible glimpse into what was undoubtedly a complex set of world views and cosmologies (Cunliffe 2004: 183). Against a background of constant transformation and regional variation, one aspect of Iron Age art that remained consistent was its complex and ambiguous forms, where abstract designs depicted human faces, vegetation, or animals (Green 1998: 23; Garrow and Gosden 2012: 5, 33). An intentional act of design, this ambiguity was argued to be indicative of the role of art in the negotiation and expression of relationships between humans and the supernatural world (Megaw and Megaw 2005: 22; Garrow and Gosden 2012: 33). Zoomorphic imagery featured on an extensive range of Iron Age artefacts. The presence of animals on ritual paraphernalia and the deposition of animal remains in ritual contexts have been drawn upon to suggest the divine or celestial role of some animals in the Iron Age. Whilst the exact nature of the animal imagery is likely to remain elusive, it is certainly clear that animals played an important role in what was likely to be a complex and multi-faceted belief system. Although relatively rare in Britain, there are instances whereby human and animal features are drawn together to depict one single being. Indicative of this phenomenon is a sheet bronze from Aylesford which features two horses with human feet and hind legs. The depiction of the hind legs is a subtle yet important distinction, as the forward projection of the joint is anatomically incorrect for horses, yet mirrors the human construction of the joint (Ross 1992: 407; Aldhouse-Green 2012: 165). Examples of semi-zoomorphic imagery are more numerous on the continent, and include representations of a horned anthropomorphic being known as Cernunnos (depicted for example at Val Camonica, see Bober 1951: 18), human headed dogs and horses from Reinheim (Aldhouse-Green 2001: 213) and a bronze figure of an anthropomorphic male with hooves from Bouray, Essonne (Aldhouse-Green 2001: 213). This cross-species imagery is perhaps indicative of beliefs regarding the nature of the divine or supernatural, and, as proposed by Webster (1986: 36) and Aldhouse-Green (2001: 231) may symbolise the ability of supernatural beings, religious specialists and animals to cross otherworldly thresholds or to shape shift.

The ability of divine beings to shape shift is perhaps just one component of a complex Iron Age belief system whereby form, species and gender were fluid and permeable. Concepts of gender fluidity may find expression in the imagery of the British Iron Age, with a heritage that stretches back to the Neolithic period (Coles 1990). Whilst gender duality may find expression in hermaphroditic imagery, it may also be expressed through asexual representations. A collection of chalk figurines from sites across Yorkshire, including Garton Slack, which has been dated to the middle Iron Age, features a number of individual lacking any representation of gender (Cunliffe 2004: 521; Megaw and Megaw 2005: 25; Giles 2013: 61). This is particularly striking due to the presence of figures within the collection depicted with phallic imagery, and suggests that the lack of gender-specific features on the remaining figurines was deliberate. Similarities in somatic representation have led to parallels being drawn between the Yorkshire figurines and an isolated find in Deal from an underground chamber (Parfitt and Green 1987: 295; Halkon 2013: 56). The Deal figurine displays an absence of any attempt to model the body and whilst more attention has been paid to the face, it too is of a schematic nature. Although of a Roman date in terms of deposition, the Deal figurine betrays no classical influence, adhering instead to Iron Age norms of representation. The martial aspect of the accoutrements depicted on the chalk figurines has widely led to their ascription as warriors, yet whether these were representations of divine, living, or ancestral warriors, or perhaps toys or talismanic charms is unknown (Dent 1983; 1984; Halkon 2013: 56). Irrespective of whether the figurines represented mortal beings or deities, the ambiguity in gender ascription may suggest that gender was not conceptualised in the Iron Age as a simple binary opposition between masculinity and femininity, but operated on a spectrum. Inferences from the archaeological record of the Iron Age suggest a degree of correspondence between Santerian beliefs regarding the complex, dual identities of the divine, and Iron Age concepts of divine expression and gender perception. It stands to reason that, without drawing direct parallels, Murphy’s insights can therefore provide important new ways to consider how native beliefs came to be reflected in the archaeological record of the provinces following the introduction of classical deities, iconography and mythologies.

Divine Marriage: Expressing ‘Problematic Translations’

Whilst not denying the existence of equivalencies between Roman and Iron Age belief systems, it has been increasingly acknowledged that key differences in the perceived nature of the divine existed alongside considerable variation in cult practice and expression. Noting, for example, the shared tradition of specific divine ‘types’ such as mother goddesses and weather gods across many cultures, Häussler (2012) emphasised the variable functions, attributes, mythologies and rituals associated with each across varying cultures. Whilst scholars such as Derks (1998) have argued that Roman deities had more than one function, it is important to acknowledge precisely how Roman concepts of divine function and identity compare with its Iron Age counterparts. The existence of Graeco-Roman divine epithets has provoked considerable debate, centred upon the conflicting notions of the epithet as signalling distinct or separate deities, or signalling one form of a particular, singular, divine identity (Wallensten 2008). King (2003: 291), arguing for the latter, suggests that the Romans believed that divinities had multiple aspects and possessed more than one form — operating in particular contexts under particular names. Thus the god worshipped in one particular context under a particular name, was the same god that could be worshipped in a different context using a different name. Juno Lucina, the goddess of childbirth, was thus the same as Juno Sospita, who operated in a martial capacity (Rives 2006: 15). To some extent, this may be considered indicative of an overlap between Iron Age and Roman beliefs, where a single deity encapsulates multiple roles and aspects. Where these belief systems appear to diverge, concerns the extension of multiple divine roles and aspects across gender and species boundaries. Iron Age conceptualisations of the divine, as speculated, appear to have been based upon the fluidity of the divine who was able to transcend gender and species boundaries, whilst — in spite of rare occurrences such as the myth of Zeus taking the form of a swan — such conceptualisations were broadly absent in Roman conceptualisation of the divine, who were almost universally conceived of in terms of a single sex.

The existence of differing divine conceptualisations most likely made straightforward, linear identifications between a native and classical deity incredibly difficult. Illustrative of this difficulty is Lucian’s description in De dea Syriae (31–32) of an image of Hera from the temple of Atargatis in Hieropolis, which drew upon iconographic references of a number of deities including Athena, Aphrodite, Rhea and Artemis (Ando 2005: 45; Häussler 2012: 154). This phenomenon was mirrored across the provinces, as for example in the case of the statue of Brigantia recovered from Dumfriesshire — a local goddess whose image was created by drawing upon elements from multiple classical deities including Minerva and Victory (Keppie and Arnold 1984: figure 12). For Häussler (2012: 154), these sculptures marked the culmination of attempts to represent a local deity with their own functions and mythology, using an established visual vocabulary of divine iconography from the Roman world.

Divine marriage, reinterpreted according to Murphy’s insights, may too reflect this process of ‘problematic translation’, in that there was no direct, simplistic parallel between a gender transcendent local deity and a classical counterpart. The Roman deity Mercury, for example, frequently depicted in divine pairings could have been considered as an aspect of Rosmerta — the male element of this divinity that operated in specific, local contexts. It may have been that both Mercury and Rosmerta provided an overlapping function, the protection to flocks for example — yet aside from this, operated in divergent fields of influence and power. By juxtaposing the two deities, indigenous populations were perhaps able to draw new deities into native belief systems and to express them in the context of these systems, as in the case of Santerian religion. In this way, the worship of Mercury alongside Rosmerta could have provided new ways to access a particular power that he shared with other divinities such as Rosmerta, as well new powers that were not possessed by these other divinities, as discussed by Murphy (2011). The presence of animal attendants in the reliefs depicting divine marriage (Henig et al. 1993: 26) may also reflect the polyvocality of this imagery and the various ways in which classical deities may have been reinterpreted in the context of local belief systems. Whilst classical divine imagery often featured animal attendants which were perceived as a distinct entity from the deity in question, in the context of Iron Age belief systems, it may have been that some members of the provincial population read the animal attendant as reflective of the shape-shifting abilities of a deity and its corresponding animal form; an additional expression of the fluid gender and species boundaries of the divine.

New Perspectives and Layers of Understanding

That divine marriage reflected a complex translation of differing belief systems is not a new concept. Admittedly, early accounts of divine marriage dismissed the independence and importance of the female divinities, designating them as consorts who did little more than to reflect the female personification of the male deity (Watson 2005: 89–92). From the archaeological evidence however, it is irrefutable that the female partners had their own identity outside of the confines of the divine partnership, and enjoyed a level of importance and popularity of notable significance (Green 1989: 60; 1995: 126; Watson 2005: 92; Ferlut 2016). In more recent accounts, the consistent mirroring of attributes in the male-female counterparts in some cases of divine marriage was said to have operated as a deliberate expression of balance and equality between the respective gods and goddesses. Expressions of equality were also proposed to have been manifest in the lack of gender dimorphism; the comparable size and stature of the god and goddess is a notable feature of a number of reliefs depicting divine marriages (Aldhouse-Green 2003; 2012).

Further refuting accounts of the low status ascribed to the female deities, Watson (2005: 20), noted that both Rosmerta and Sirona, two of the most prominent deities in divine partnerships, are depicted with attributes foreign to Mercury, negating the argument that they operated simply as a straight-forward female personification of their male partner. Sirona, like Rosmerta, appears to have been associated with fecundity, regeneration and healing — as inferred by attributes such as fruit. Drawing upon early insular myth, Aldhouse-Green (2012) has suggested that a female rather than male deity occupied a position of supremacy in Iron Age Britain — a personification of the land, who entered into a ritual marriage with an inferior being, divine or mortal, to ensure its continued prosperity. Foreground in this knowledge, the consistent imagery of fertility and abundance attributed to the female partner of the divine couple is particularly noteworthy. Viewed through the prism of indigenous belief systems, it was argued that divine marriage may reflect the subversion of the dominant/subject status of male/female divinities traditionally conceptualised in acculturative readings. Whilst Aldhouse-Green (2012) acknowledges that it is problematic to make direct inferences from medieval narratives, they may provide important insights into early belief systems and how these came to be reflected in associated imagery.

Further challenges to the notion of divine marriage as a simplistic tool of interaction can be found in Jane Webster’s studies on the nature religious syncretism in the provinces (1997; 2001). Far from a straight forward, simplistic adoption of Roman divinities and artistic expression, divine marriage was argued to represent an ambivalent act that permitted not only the maintenance of indigenous religious beliefs, but operated in a subversive capacity to ‘submit a Roman deity to the power of an important local goddess’ (Webster 1997: 327). This use of creolisation theory was subsequently critiqued for maintaining binary analyses of religious change framed in absolute terms of native versus Roman and acceptance versus rejection. Murphy’s analysis of Santerian religion has arguably transformed the way in which Creole religions can be understood. Rather than operating as a tool of subversion, it was shown to operate as a powerful means through which complex new identities were forged. Through combining native traditions with adopted non-native traditions, new forms and religious traditions were born. Divine marriage, as reconceptualised through Murphy’s work, may therefore have offered a way for indigenous populations to both maintain and enhance their divine world by drawing upon the power of the classical gods and incorporating it into local belief systems, reframing Roman deities in terms of indigenous belief systems. These divine interactions arguably contributed to the creation of a new religion that was neither Roman nor indigenous, but drew upon these two elements to varying degrees dependent upon the individuals that produced it. As elements of classical mythology and iconography were adopted and adapted by local populations, local cults and belief systems were read anew, resulting in new religious expressions.

This new reading of divine marriage has provided a unique insight into how native populations engaged with non-native deities, iconographies and mythologies. Employed as a tool through which individuals negotiated and constructed the world around them, the incorporation of Roman deities or symbols was arguably a means through which indigenous populations were able to contest and construct their own religious identity in a greatly changed world. It reveals a careful, thoughtful and deliberate adaptation by local peoples, who drew upon new deities, concepts and forms of representation and expressed these in the context of local belief systems. As such, it provides insights that alternative forms of divine pairings arguably cannot. The interpretation of foreign divinities, or Interpretatio Romana by classical authors, including Caesar’s equation of Gallic gods to their Roman counterparts, is notably problematic (De Bello Gallico 6.17), representing superficial and simplified understandings of Iron Age religions and consequently considered of very little use (Green 1995; Häussler 2012). The value of epigraphic name pairing outside of revealing religious change amongst the elite has similarly been called into question. Often the preserve of high ranking military officials and indigenous elites, over half of the cases of epigraphic name pairings in Roman Britain were dedicated by individuals associated with the army (Zoll 1995). Drawing upon Hartog’s (1988) analysis of Herodotus, whereby the process of divine translation or equation was considered a means of superimposing one belief system over another, Webster (1995: 156) argued that equivalences drawn between an indigenous and classical deity resulted from a limited and superficial Roman understanding of indigenous belief systems, and as such, indigenous deities were said to have been ‘squeezed, with varying degrees of discomfort, into imported conceptual moulds’ (Derks 1991; Webster 1995).


This new reading of divine marriage sits within a wider discourse on the nature of provincial religion, the role of non-Romans in its creation, and the acknowledgement that native religion was not entirely abandoned. Provincial religions were conventionally interpreted as reflective of a neutral, laissez-faire syncretism whereby polytheistic peoples willingly welcomed new divinities and/or drew equivalences between indigenous and non-indigenous deities (Haverfield 1923; Webster 1995; 2001; Green 1995; Aldhouse-Green 2003, 2012). A large body of work from scholars such as Woolf (1997; 2000; 2014), Hingley (2000, 2003), van Dommelen and Lopez-Bertran (2003), Mattingly (2010), and Stek (2014) sought to challenge the dominant acculturative discourse in Roman archaeology by focusing on the range of responses elicited, from collaboration and adaptation to resistance. The diverse interactions that took place in the provinces created a new culture and new religion that was neither Roman nor indigenous, but drew upon these two elements to varying degrees dependent upon the individuals that produced it (Aldhouse-Green 2003, 2012; Watson 2005; Revell 2007; Häussler 2012).

This interpretation of divine marriage is not intended to supplant previous understandings of this process, but to add to them, emphasising that it was a political, social and cultural act: a negotiated encounter with Rome. The result was a dynamic religion, with roots in traditional and non-traditional elements — continuously adopted and adapted to meet the needs of its diverse adherents. Divine marriage reflected multiple strategies and interpretations each valid according to the viewer. Some inhabitants may have engaged with and drawn upon Roman deities not in an acculturative capacity to replace indigenous divinities or to ‘Romanise’, but as a means of divine and spiritual expansion, so that indigenous populations could share in and access the power of Roman divinities that were likely to have been considered highly powerful and effective, whilst expressing these new powers in local terms. Others may have interpreted divine marriage in terms of the Roman power dynamic and the superiority of Roman divine world. We should not confine our conceptualisations of divine marriage to a discrete or distinct category — but as a multivocal phenomenon that operated as an integral component of various strategies of cultural and religious interaction.

The alternative reading of divine marriage provided in this paper has drawn upon a postcolonial interpretation of religious transformation. Whilst postcolonial approaches have been subject to increased scrutiny from advocates of globalisation theory, arguably, these two approaches are not incompatible. In championing the use of globalisation perspectives, Pitts and Versluys (2014: 7) have argued that there ‘is an urgent need to transcend post colonial approaches’, which have served to strengthen and maintain the binary oppositions of Native and Roman. However, in developing new readings of divine marriage, this paper has endeavoured to demonstrate how postcolonial approaches can decentre the notion of fixed, stable identities and extend beyond simply identifying an object as the passive result of cross-cultural interaction. In this revised conceptualisation of divine marriage, the phenomenon is seen to reflect the creation of flexible identities, where the global and local was reinvented in the context of local belief systems; a process of divine expansion that drew upon local and non-local traditions. As such, divine marriage has been positioned as a culturally creative process where diverging traditions were adopted, adapted and read anew. As a process of divine interaction which deliberately referenced the traditional and non-traditional, this reading of divine marriage subverts concepts of fixed, monolithic identities. As such, it is hoped that this mitigates some of the concerns levelled at postcolonial accounts and bridges the divide between postcolonial theory and globalisation.