Populations in the Graeco-Roman world were always attached to gods that embodied specific and remarkable characteristics of nature. So, gods and goddesses linked to water and springs raised substantial interest for the dedicants all over the Empire during the Roman period, and the provinces of Gallia Belgica and the Germaniae were no different. Indeed, worshippers were attracted by the fact that such deities represented the specific attributes of water – the heat and/or the striking colour of the water, its ability to heal, and so on – or those of the landscape such gods and goddesses contributed to creating (Scheid, 2007–8). Although gods and goddesses of water and springs were numerous during this period only a few were believed to possess the specific ability to heal, performing such healing actions either alone or with a consort. In the latter case, the divine couple personified the entire process of healing, with the god acting as the healer and the goddess representing either the recovered health or the embodiment of the water or the spring.

In this article I will focus on the feminine consorts of the healing gods whose water or spring was used by the god in the healing process and/or who were considered as embodying the Salus, i.e. recovered health. Before any further remarks or investigation, it is worth warning readers that not all the gods and goddesses considered to be living close to water or to a spring had the ability to heal and not all the sanctuaries that were close to water were in fact sanctuaries dedicated to water and to healing gods. This remark has already been made by J. Scheid (Sheid, 2007–8) and S. Deyts (Deyts, 2003: 19) who suggests that: ‘un sanctuaire, tout comme un village ou un établissement agricole, ne peut s’installer qu’à proximité d’un point d’eau. De ce fait, la liaison entre vital et sacramental ne peut être pris comme un postulat’ [‘A sanctuary, as a village or a villa (i.e. a large country house belonging to the elite who owned the estate composed of lands, a farm and a housing arranged around a courtyard), cannot be settled far from a natural water supply. So, liaising vital and sacramental cannot be used as a postulate.’]. Moreover, being a deity who represented water or a natural spring did not imply that this goddess had the ability to heal, to be a part of the healing process, or to be the consort of a healing god. All of this reveals the need to define precise and specific criteria to identify the consorts of the healing gods, such as the architectural pattern of the sanctuary, the recovered offerings uncovered during archeological campaigns, or the characteristics associated with the divinity (as, for example, revealed in altar inscriptions). With regards to the latter, we also need to make a preliminary remark: the feminine consorts of the healing gods were not properly healers: the god was the medicus, but the female consorts played a key role along with the god.1 Few articles have focused on these goddesses in the past 50 years, but their presence was central to many dedicants’ practices, especially in the Northern provinces of the Roman Empire. These provinces have various similarities. Conquered and organised at the time of Caesar or Augustus, they were all merely a single province until the Flavian times when they were separated in Gallia Belgica and the two Germaniae. They had also a Gaulish and/or a Germanic background that maybe could have had an influence on the names, practices and geographical spread of both the deities and the dedicants.

In order to categorise the deities, the ritual practices and the dedicants, we need to rely on sources that we can be certain about, i.e. epigraphy and sculptures when they accompany inscriptions, archaeological remains of offerings, and the sanctuaries themselves. All these materials can be dated mainly from the 2nd and 3rd centuries AD, and sometimes from the 1st to the 4th centuries AD in the case of the sanctuaries. This does not mean that traces of the cult cannot be found after those periods, but the process of confirming materials after this time is far more complex. This is because inscriptions and sculptures were no longer available after the 3rd century AD and archaeological traces of offerings are very difficult to find in sanctuaries that were investigated decades ago when archaeology of ritual practices was not a crucial question.

These preliminary remarks and the analysis of the sources raise several questions: who were the goddesses clearly identifiable as consorts of healing gods? What kind of ritual practices was performed by worshippers? And what were the identities and the actions of such worshippers?2

I. Identifying the goddesses as consorts of the healing gods

In Gallia Belgica and the Germaniae, thirteen goddesses of water and springs can be identified but only a few were consorts of healing gods and even fewer represented the Salus. In order to identify them, several criteria must be taken into account. First, their masculine consort must be a healing god and clearly recognised as such. Second, the temple in which the goddess was worshipped must have a clear bipartition of its inner space between a sacred space in which water gushed out and a profane space in which the humans could use water. Third, some attributes in the goddess’ representations had to highlight her role, such as the caduceus found on sculptures figuring Maia or Rosmerta (see inscription nr. 15 in Appendix 1). Fourth, the inscriptions must mention the existence of a request to protect someone or to make him recover his health and some offerings must be monetary and/or anatomic, symbolic of real healings (Scheid, 1992: 31). Other offerings may be present, and I will consider some in this article, but their presence must function as the key to identify a sanctuary where a god heals. Finally, some goddesses could have thermal facilities in which their water can have a specific characteristic that the god can utilize to heal. But this last criterion has to act in conjunction with some of the previous criteria, since having thermal facilities with specific water characteristics did not always mean that the goddess was the consort of a healing god, as Brixta in Luxueil-les-Bains proves. Indeed, Brixta was a goddess of water, Luxouius’ consort but, so far, it has been impossible to find clues of worship dedicated to his ability to heal.

None of the criteria is sufficient alone but in combination they allow us to assert that a goddess can be established as the consort of a healing god.

Six goddesses match the criteria for the feminine consorts of the healing gods I have outlined above (Table 1). They were consorts of Apollo, Boruo – probably a Gaulish interpretation of ApolloMercury, Asclepius and Veraudnus. Thus, not many healing gods in Gallia Belgica and the Germaniae had a feminine consort and Mercury and Apollo were the gods who had the largest diversity of consorts. However, many historians have proved the topic characters of the Gallo-Roman gods and the transformation of their role over time and across the Tres Galliae and the Germaniae (Duval, 1993; Van Andringa, 2002).4 Mercury, for example, had three goddesses that helped him to heal: Maia, Rosmerta and Visucia. If we consider the first two of these, Maia and Rosmerta, all the evidence shows us that the divine couples – Maia/Mercury and Rosmerta/Mercury – were worshipped in many temples and sanctuaries all over the provinces. Maia and Rosmerta had several functions across the Empire, especially as goddesses of fecundity and goddesses of spring and water (Ferlut, 2011). But, when these goddesses became associated with Mercury, it was not the fact that they provided fecundity that was central to the dedicants’ demands. Rather, it was their link to water and to the use of that water by the god which was most important to the dedicants (Ferlut, 2011). Maia, in some cases, was represented with the caduceus (Espérandieu, nr. 5977 and inscription nr. 15 in Appendix 1) which is an attribute that clearly demonstrates Maia’s connection to the healing process. In the sculpture, however, she did not hold the caduceus in her hand; the caduceus was engraved in the lateral side of the altar which indicates that she was associated with the healing process, rather than functioning as a medica. As part of the healing process, Maia was understood to be the nymph of the spring that Mercury utilized to heal, but the presence or the offering of the caduceus also meant that Maia participated in the act of healing and represented recovered health for the dedicant. As for Rosmerta, her association with Mercury made her not simply a goddess who helped the god but suggests that she embodied the representation of the result of healing, personifying recovered health, i.e. the Salus. Indeed, a caduceus was consecrated to her by a dedicant (see inscriptions in Appendix 1) which proves that she was a part of the medical process. She was also represented a few times with a caduceus in her hands (see Appendix 1). These kinds of representations are similar to the depictions of the divine couple Asclepius/Hygia about whom we know that the god healed and the goddess was the representation of recovered health.

Table 1

Goddesses identifiable as consorts of healing gods.3

Goddess Consort Characteristics Sanctuary
Damona Apollo Boruo Boruo was a healing god Sanctuary with a thermal facility in Bourbonne les Bains where inscriptions for the Salus of a member of a family and where monetary offerings were found.
The sanctuary was bipartitioned between sacred and profane spaces.
Hygia Asclepius Aesclepios was a healing god One inscription in a sanctuary in Bonn
Inciona Veraudnus Veraudnus was a healing god Sanctuary in Widdenbierg – the exploitation of a stone quarry completely destroyed it
Maia Mercury Goddess protecting from the diseases as Mercury.
On sculpture, she is holding a caduceus
Visible in several temples all over the three provinces even if they were not dedicated to her.
Rosmerta Mercury Goddess protecting from the diseases like Mercury.
On sculpture, she is holding a caduceus
Goddess who was worshipped in the sanctuary of Deneuvre in Mosel, even if it was not dedicated to her.
Some inscriptions were recovered with the mention of the expression pro salute.
Dedication of a caduceus by a worshipper in Bad Kreuznach.
Sirona Apollo Goddess of purification, capable of divination.
Healing goddess as Apollo
Visible in many sanctuaries, especially in Valenborn and Niedaltdorf where monetary offerings were found. In Niedaltdorf, the sanctuary was bipartitioned between sacred and profane spaces.

Another couple had its cult spread across the provinces during this period: Apollo/Sirona. Sirona was rarely worshipped without Apollo but this association cannot be linked only to the healing function of the god. The representations we have of the goddess are too damaged to be a real help. However, several inscriptions were discovered in Niedaltdorf, which we know to be a water and spring sanctuary. In this case, we can assert that the goddess participated in the healing process, probably as the embodiment of the spring and of its characteristics utilized by the god to heal.

The other consorts of the healing gods had a much more localised cult, such as the cults to Inciona and Damona. The latter goddess was the Boruo’s consort, a Gaulish god for whom some inscriptions were recovered in two sanctuaries: Bourbonne-les-Bains in Belgica and Bourbon-Lancy in Lugdunensis – this last sanctuary is beyond the range of my study but it will help us to narrow down some conclusions. In Bourbonne-les-Bains, the god had the epithet Boruo and formed a topic couple with his consort Damona. Boruo has already been defined by historians as a healing god and the aspect of the sanctuary and the offering uncovered there – which I will examine more closely later in this article – proved that real healings were made by the gods. In Gallia Belgica, Boruo was a Gaulish interpretation of Apollo as the inscription nr. 1 (Appendix 1) proved (Vaillat, 1932: 27, 95–111). Indeed, in this particular case, Boruo became the epithet of Apollo proving that the dedicant clearly associated both of them within the same divine representation. This connection between Boruo and Apollo was probably different from the case of Lugdunensis where no traces to the link with Apollo seemed to be made, and maybe it was not the same deity considering the fact that several inscriptions named him Bormo (CIL XIII, nr. 2805 = D 4659) and that the location was called Aquae Bormonis. Even if he was associated with Damona, we must be cautious about comparisons between the two deities and we must assume that both divine entities are commensurate in their functions and ability to heal, but remain different as a result of their local particularities. Damona was the embodiment of the source that helped in the healing process. We know that the water in Bourbonne-les-Bains had a characteristic that was utilized in the thermal facility close to the ‘puisard’ to heal. In Bourbonne-les-Bains we have the proof that Apollo was acting as the healer and Damona was providing him with water.

Apart from Mercury’s and Apollo’s consorts, two other healing gods had feminine consorts in the three provinces of my study, Gallia Belgica and the two Germaniae: Asclepius, from a Graeco-Roman origin and Veraudnus, from a Gaulish origin. In the Roman and Greek world, Asclepius was a healer and Hygia was always presented as his companion. Only one inscription made in the Roman colony of Bonn mentions the divine couple (see Appendix 1). For such a powerful couple, this dearth of inscription seems unusual but it could be attributed to the fact that populations in those provinces preferred Celtic and Germanic goddesses rather than the Roman and Greek healing gods and their consorts (Ferlut, 2011). The last god relevant to my discussion is Veraudnus, a Celtic healing god associated with Inciona, a goddess of water. In the case of Inciona, it is not certain that she represented recovered health. More probably, she was just a goddess of the spring whose power Veraudnus utilized to heal (Kuhnen, 1996).

Looking at the origin of the goddesses, it is worth pointing out that just one inscription concerns a non-Celtic goddess, Hygia who had a Graeco-Roman origin, whereas all the other goddesses were of Celtic origin. An explanation emerges when we look at the geographical spread of the goddesses (Table 2).5

Table 2

Feminine consorts of the healing gods in the provinces.

Provinces List of goddesses
Gallia Belgica Damona, Inciona, Maia, Rosmerta and Sirona.
Germania Superior Maia, Rosmerta, Sirona and Visucia.
Germania Inferior Hygia

The only case of a Graeco-Roman goddess was in Germania inferior in a Roman colony. By contrast, all of the other deities’ inscriptions are spread all over the provinces in both rural and urban centers, although rarely in the largest cities or in capitals of the provinces, apart from a few inscriptions in Metz. This is likely attributable to the fact that the sanctuaries dedicated to water and springs in the provinces we study were generally in rural areas. Gallia Belgica was also known to privilege Celtic goddesses over Roman or Greek deities (Ferlut, 2014) and the analysis of the corpus (Appendix 1) proves that the province hosted the largest number of inscriptions dedicated to those feminine consorts of healing gods – 33 in Belgica and 21 in Germania Superior. In the majority of inscriptions the gods had Roman names, as in the case of Mercury and Apollo, even with a topic epithet, but their consorts were Celtic goddesses so far, including Maia, whatever her Roman name may be: Populations from these provinces decided to use names from the province, mainly because these goddesses were primarily goddesses of springs and water, like Damona. So naming them with references to the local topography seems obvious. But some goddesses may have undergone interpretation, through a process in which people from the region interpreted goddesses from Rome or Greece into local deities, so they perhaps made interpretatio indigena. Tacitus was the first to talk about interpretatio romana (Tacitus, Germania, LIII, 3). The notion of interpretatio, as interpretatio indigena, caused fierce debates among historians: firstly, in proving the existence of such a phenomenon and, secondly, in determining the deities that were subjected to such interpretationes (Dunand and Lévêque, 1973; Lévêque, 1973; Scheid, 2003; Haeussler, 2008 and 2012; Charles-Laforge, 2014); but as this is tangential to my subject in this article, I will not sum up the debate here. For the Romans, interpertatio romana consisted in giving Latin names to exogenous gods. Tacitus and Caesar (Caesar, De Bello Gallico, VI, 16–18 and Tacitus, Germania, LIII, 3) revealed that the Romans often practised this mode of interpretation. Meanwhile, for local populations of Roman provinces, interpretatio indigena consisted in giving a Gaulish name to a divinity. Three conditions were necessary to talk about this scheme of interpretatio: the dedicant should have a Gaulish name, the deity’s name should be Gaulish as well, and the god should have the same characteristics as a Roman god, characteristics identified by the gods’ attributes. The analysis of the seven goddesses proved that Roman goddesses were hardly interpreted (Clifford, 2012). Among almost eight hundred inscriptions discovered for the Celtic goddesses, only eighteen (Ferlut, 2014) – fourteen with certainty – were manifestations from interpretatio indigena but it seems that the consorts of the healing gods were those who were the most susceptible to be interpreted with six inscriptions submitted to interpretatio indigena (Table 3).

Table 3

Traces of interpretatio gallica.

Goddess Date Evidence of interpretatio indigena Reference in Appendix 1
Maia c.100 Goddess with a Celtic name.
Dedicant with a Celtic cognomen.
Evidence of interpretatio indigena: Caduceus as her consort Mercurius.
Nr. 18
c.101–250 Goddess with a Celtic name.
Dedicant with a Celtic cognomen.
Evidence of interpretatio indigena: Caduceus as her consort Mercurius.
Nr. 19
142 Goddess with a Celtic name.
Dedicant with a Celtic cognomen.
Evidence of interpretatio indigena: Caduceus as her consort Mercurius.
Nr. 13
c.150–250 Goddess with a Celtic name.
Dedicant with a Celtic cognomen.
Evidence of interpretatio indigena: Caduceus as her consort Mercurius.
Nr. 15
Rosmerta c.101–250 Goddess with a Celtic name.
Dedicant with a Celtic cognomen.
Evidence of interpretatio indigena: Roman attributes mainly used for spring and water goddesses.
Nr. 26, nr. 31 and nr. 34.
c.201–250 Goddess with a Celtic name.
Dedicant with a Celtic cognomen.
Evidence of interpretatio indigena: Roman attributes mainly used for spring and water goddesses.
Nr. 35

In this case, people may have preferred to worship goddesses with a local origin or to make an interpretation in order to be better healed by the local water or spring used by the god, or to have a better chance to find the Salus. We can also assume that water and spring, and their divine impersonations, were named and utilized by the Celts long before the Roman conquest, which could arguably offer an explanation for the persistence of Celtic names – even if such an interpretation cannot be proven. However, interpretatio was rare – six inscriptions among fifty four – which proves that people from Gallia Belgica and the Germaniae decided to worship goddesses according to their original nature. This was not uncommon at the time since the Romans generally adopted foreign pantheons. Moreover, these Celtic goddesses became more visible with Roman modes of worship, but they kept their Celtic nature and the worshippers made vows to them for their specific functions.

Identifying the gods’ feminine consorts and their functions is therefore a challenge but it is an important scholarly undertaking since it allows us to focus on the different kinds of rituals practised at the time, in the deities’ sanctuaries.

II. Roman rituals and temples dominate

According to analysis of the inscriptions, the uotum was the main ritual – thirty seven inscriptions among fifty four. As in any other provinces of the Empire, in Gallia Belgica and the Germaniae dedicants had no obligation of faith but to adhere to a strict practice of the rites with an extreme respect for religious prescriptions. I therefore do not accept the idea sometimes evoked – even during the Roman period – that the Romans, and more broadly the inhabitants of the Roman Empire, did not believe in their gods and only performed rituals as part of a mechanical ritual routine for social or political purposes.

The uotum, a Roman ritual commonly used

The uotum was largely accomplished by people believing in its power and practising it according to strict ritualistic procedure. It was a full rite followed by worshippers in order to fulfill the vow once granted, i.e. libation, sacrifice of animals, plant offering and ex-uoto of any kinds.

Let’s begin with some general explanations about what the uotum was. It was divided into three different steps: the dedicant announced the vow, then the goddess granted it and, finally, the dedicant fulfilled it. The vow was based on a contract. If the goddess did not grant the vow, the dedicant was freed from his promise. The vow had to be granted within a limited period of time. The uotum was generally practised in a sanctuary, within the enclosure of a temple, after the publication of a libellus; quoting Horatius and Juvenal, B. de Sury has shown that, in the case of an emergency, this could be done immediately (de Sury, 1994: 169–70). It was comparatively rare for anyone other than the dedicant to know the nature of the sacrifice or the offering, except in cases where the dedicant represented it on the lateral faces of the altar he offered when the vow was fulfilled. With regards to the goddesses I am concerned with here, we can state that knowledge about the nature of the sacrifice or offering was quite rare because only one altar to Maia (inscription 15 in Appendix 1) has a representation of laurel branches. The fulfillment, and the inscription associated with it, were thus the main means to show the use of the uotum. It was identifiable by the expression uotum soluit libens merito,6 an expression found many times in inscriptions dedicated to the feminine consorts of the healing gods. The uotum therefore had a public function: its objective was to prove that the dedicant fulfilled the vow. This act was addressed both to the goddess and to the entire ciuitas. It also proved that the goddess was kind to the worshipper: it proclaimed her powers but also demonstrated to the rest of the community that the dedicant had the ear of the gods. Finally, it aimed at perpetuating the transient offering, by naming or representing it. In our epigraphic corpus (Appendix 1), only one inscription represents the offerings as I have described them and few explicitly name them (Table 4).

Table 4

Offerings mentioned in inscriptions.

Deities Offering References in Appendix 1
Mercury and Maia a caduceus and an altar Nr. 15
Mercury and Rosmerta an hospitilia and an aedes with statues Nr. 24
an altar with statues Nr. 40
an aedes Nr. 41
Apollo and Sirona an aedes with statues Nr. 52
Mercury and Visucia a statue Nr. 56

In the case of Celtic goddesses, the uotum was particularly significant for the worshippers. When the goddess granted the vow, the dedicant fulfilled it with an ex-voto, as evidence of the fulfillment. He took his part of the covenant. The fulfillment had several forms: the sacrifice of an animal, a plant offering – the most common case – libations, monetary offerings, and offerings made by the building of a temple, or statue. Historians have been able to determine the various kinds of ex-voto in many ways: the formula of the inscription, sculptures on the altars, and some discoveries which were made during excavations (J. Scheid, 2000; Lepetz and Van Andringa, 2008; Ph. Méniel, 2008). Most of the time, the sculptures on the lateral faces and the top of the altars are an asset in discovering the real nature of the fulfillment; an asset we do not have in the present study of the feminine consorts of healing gods, compared with the study of other goddesses such as the Matronae (Ferlut, 2011). Moreover, many of the sanctuaries were investigated long ago so the remains of animal sacrifices disappeared in previous archaeological campaigns, reinforcing the difficulty of establishing a clear survey of sacrificial rituals for the goddesses with which I am concerned here.

The sacrifice of animals seemed to be one of the most common ex-voto. Studies of Greek and Roman religions, as in recent studies on Roman Gaul and Germanies (Lepetz and Van Andringa, 2008), have proved the high use of animal sacrifice. This prevalence of animal sacrifices might have a central meaning in the societies of the imperial period. But, with reference to the consorts of the healing gods, the recovered traces of such offerings are quite rare. The use of such sacrifices is also difficult to prove. Even in Bourbonne-les-Bains, where the sanctuary was thoroughly excavated, traces of such sacrifices are not obvious, especially because when the thermal complex and the sanctuary were investigated for the first time, animal remains were not considered to be precious archaeological material, and archaeologists privileged stone, ceramic and monetary remains. Moreover, destructions in the 1970s irremediably destroyed part of the complex so evidence would be never recovered (Maligorne, 2009 and 2011). Despite its efficiency as well as its remarkable nature, the sacrifice of animals must have been quite rare due to the price of animals at the time. The majority of the population probably did not have enough money to buy such an animal. Plant offerings – in the forms of fruits or laurel branches – or alimentary offerings such as and cakes – would probably have been the most common. Incense and wine seem to have been, in the imperial provinces such as Rome, the most widespread way to express religious reverence, and the most ordinary ex-voto. Many ancient authors confirmed this matter of fact and described it as a sufficient religious act (Ovid, Pontiques, 4, 8, 39; Seneca, De Ben. I, 6, 3; Stace, Thébaïde, 2, 247; Horatius, Odes, 3, 23). The representation of a tree and laurel branches on the lateral sides of the altar nr. 24 in Appendix 1 seems to confirm the use of such an ex-voto. The ex-voto could also take the form of a libation. The dedicant let the liquid contained in a patera fall onto the floor (for a chthonian deity) or poured it in the lighted hearth of an altar (for an ouranian deity). But no evidence has so far been recovered in our case of the consorts of the healing gods neither in the archaeological digs nor in the engraved representations upon altars, but this does not mean that libations were not used since traces of such a ritual are almost impossible to find.

Two other kinds of offering have, however, been discovered: monetary and anatomic offerings. In Bourbonne-les-Bains for example, at the location where the natural spring releases water from underground – which was called the puisard by those who discovered the site in the nineteenth century – 4900 coins were recovered (Maligorne, 2009: 227). If we take into account the high frequency of theft at such sites, this cache could have been as high as 10,000 coins. In the case of anatomic offerings, these were only present in Luxueil-les-Bains, in the sanctuary of Brixta and Voroius, where some inscriptions to the consorts of healing gods appeared. But no evidence can tell us if Vorious and Brixta belonged to that group because the presence of anatomic offerings in itself does not prove, with any degree of certainty, the existence of a healing god. In the case of Maia (inscription nr. 15 in Appendix 1), a particular ex-voto appeared, a caduceus, confirming that the offering was made because of a recovery after having begged Mercury and Maia. This is a unique case, as far as we can tell with our present documentation. Finally, as Table 4 shows us, in order to fulfill their vow some worshippers made evergetism by building chapels – aedes – with statues and, in one case, a hospitalia to host people traveling to the sanctuary. Such ex-voto were very expensive and became part of evergetism when the dedicants were a soldier and his family or a tabularius, sevir augustal and his family who were people highly marked by Romanity and involved in the life of the ciuitas and their sanctuaries.

The rituals I have discussed in this section were performed in those sanctuaries offered by the city or by the dedicants. In the case of consorts of the healing gods, however, the temples and sanctuaries had a specific form and use since the goddesses were goddesses of water and spring.

Sanctuaries with bipartition of the space

J. Scheid (2007–8) proved that sanctuaries dedicated to water and springs are identifiable because of a bipartition between the sacred and the profane spaces as well as by the way in which water was used in each part of the sanctuary. In the sacred space water gushes out and no human uses of water should be made in this part of the sanctuary – except for monetary offerings – which was clearly and strictly separated from the profane space where water was utilized for human purposes. This bipartition is applicable to some of the sanctuaries we have been able to identify among our survey (Table 5).

Table 5

Sanctuaries dedicated to consorts of the healing gods in inscriptions.

Goddess Date Location Monument Dedicant References in Appendix 1
Maia c. 201–250 Ettlingen, Morsch Aedes with statues Lucius Cornelius Augurinus, decurio ciuitates Aquensis. Nr. 17
Rosmerta c. 201–250 Uess Aedes Caius Saturninius Viriaucus Nr. 41
232 Wasserbillig ciuitas Treuerorum. Aedes with statues and ornaments plus an hospitalia Acceptus, auxiliarus and seuir augustales. Nr. 24
Sirona 201 Grossbottwar Aedes with statues Caius Longinus Speratus, veteranus XXII legio primigenia with his wife and son. Nr. 52

Presently, as most of the aedes have not been recovered but are only mentioned in inscriptions, it is difficult to be sure that all the sanctuaries were marked by the spatial bipartition J. Scheid mentioned. Only two sanctuaries can provide us with sufficient information to conduct an analysis, because we have architectural plans Studies of them available: Bourbonne-les-Bains (Maligorne, 2011) and Niedaltdorf-Ihn (Nr. 36 in Appendix 1). I propose to study these two ‘sanctuaries,’ even if the use of the term can be questioned for those two locations. Inscriptions in Bourbonne-les-Bains were discovered in a large thermal complex but the water springs out in a ‘puisard’ separated from the rest of the buildings by peristyles. In this particular area, only coins were thrown into the spring (Février and Maligorne, 2009), proving that the area where the spring appeared was clearly separated from the rest of the area where the water was used, and thus creating some sort of sacred space. The inscriptions were also recovered close to the ‘puisards’ (Maligorne, 2011), thus confirming the sacred aspect of this particular space. In the rest of the thermal complex, water was used for the thermae and for the healing process offered by Apollo Boruo at this site. However, this analysis is largely based on hypothesis because, as Y. Maligorne suggests, previous destructions and the inability to investigate the entire complex impede the detailed and precise analysis of the site by historians and archaeologists. The second ‘sanctuary’ in Niedaltdorf is also quite difficult to analyse, since researchers were unable to ascertain detailed information about the architectural aspects of the basin and the rest of the area where water was utilized. Let’s see, however, what we can say about the Niedaltdorf-Ihn ‘sanctuary,’ considering the archaeological information we have (my subsequent analysis will be based on the plan nr. 36 provided in Appendix 1). At the site an octagonal basin, encircled by a small wall, was discovered close to a villa. Inside the basin, a spring or, more precisely, a water resurgence, gushed out in what can be considered the equivalent of cella: the inner chamber of a temple in which some reconstructions imagine the possibility of the presence of a statue of Sirona, the goddess of the spring. The basin at Niedaltdorf was encircled by a colonnade supporting a roof. Archaeologists discovered the inscriptions outside the perimeter within which the water appeared, proving that the area delimited within the walls was sacred and no one could enter apart from the goddess or priest. As in the case of Bourbonne-les-Bains, bipartition is clear in the case of the Niedaltdorf-Ihn ‘sanctuary’ as well. What we do not know is if there existed a way to export water out of the sacred area to be used to heal, because the investigation cannot provide us with evidence of any thermal facility, i.e. thermae, or even with the canalization to transport water.

Bipartition of the spaces can therefore be proved in two ‘sanctuaries’ dedicated to Damona and Sirona. Moreover, this bipartition of the sanctuary into a space for water that humans can use and a natural spring where water can flow uncontaminated by human hands is not specific to Gallia Belgica and the Germaniae: as J. Scheid has demonstrated (Scheid, 2000 and 2007–8), the same partitioning practice was used in sanctuaries all over the Roman world, such as in the Clitumne spring in Rome. Roman rituals and Roman spatial bipartition of such sanctuaries can thus be understood as forms of Romanity applied to Celtic goddesses in the provinces. But a final question remains about the men and women who performed the rituals to the female consorts of the healing gods with which I am concerned.

III. Dedicants, mainly men touched by romanity

Many assumptions and hypotheses can be made concerning the kind of people who revered these deities. As they were mainly Celtic and feminine divinities, this raises the question: were the dedicants also primarily women and Celts? Although we may anticipate that the dedicants of such Celtic goddesses were Celtic women, the results of my survey would suggest otherwise.

Women were few

W. Spickermann (1994) has already revealed that women were not very visible among the dedicants in the Galliae, the Germaniae and in Retia. But, as I extend the study of dedicants to include the dedicants to goddesses, we might be justified in wondering whether women were more likely to be present among the dedicants to female deities than in any other cults. Only three of the seven consorts of healing gods were concerned with women’s dedications (Table 6). This is a confirmation of what I have already explained in previous published work examining other deities (Ferlut, 2014).

Table 6

Dedications made by women.

Goddess Number of dedications per goddess
Sirona and Rosmerta. 2
Damona 5
Total 9

In the case of the consort of the healing gods, women made 17.5% of the dedications which is a higher proportion than for the Celtic goddesses in general – in my previous research (Ferlut, 2014) I have proved that women made up only 9.5% of the total number of dedications to Celtic goddesses. Maybe this is due to the fact that they participated in the healing process and/or represented recovered health so that women might be more inclined to participate in the cult. In any case, women did not constitute the majority of dedicants, even for goddesses. It also seems that when they offered a dedication women mainly performed this act alone. Furthermore, analysis of their names reveals that most of the time these women belonged to the families of Roman citizens – 90% of the women. We can take this to mean that women who participated in the cult belonged to families where romanitas was deeply integrated. So, the women who made these vows came from families that were largely educated in Roman manners and habits. This is worth noting because many of the sanctuaries were situated in rural or suburban areas where people only came into contact with Romanity later on in the period of the Roman Empire. It can also be assumed that those women educated in Roman habits were aware of the powers of the healing gods and chose to enter the sanctuaries in order to access these divinities. Finally, dedicating an altar and performing a sacrifice were very expensive undertakings, which means that the women with whom we are concerned enjoyed a high standard of living, and certainly high enough to be financially independent in some ways.

However, we know that dedications in the form of altars are only one part of the picture when it comes to understanding dedicants. Indeed, only the richest could afford to pay for such a monument. The fact that only a few altars were made by women does not mean that women did not participate in the cult and its rituals. Most of the time, in a Roman family – and there is no clue that it was any different in Gallia Belgica and the Germaniae – it was the pater familias who performed any actions linked to religion, as his role of leader of the family obliged him to do. If we try to give a more complete picture of the way women practised a cult to the goddesses, we see that only three female deities attracted women dedicants. Damona is a particularly instructive goddess in this regard, perhaps as a result of the goddess’ specific attributes that we are not able to highlight for the moment.

Dedicants were touched by Romanity

As I have been able to demonstrate, dedicants were mainly men. But were they marked by Romanity or were they Celts? A survey of the inscriptions helps us find fourteen dedicated by people with a Celtic origin in their names – most of the time, the nomen or the cognomen – i.e. 25% of the dedications which, although it may appear only to be a few, is an amount higher than usual for Celtic goddesses (Ferlut, 2014). Finding an explanation seems pretty hazardous because even if they still had a Celtic name, usually the nomen or the cognomen, most of the time such dedicants had the tria nomina, so they were Roman citizens. Only five of the men in the survey had a single Celtic praenomen. This is not a question of province because we have the same number of examples in Belgica and in Germania superior. It is hard to find a clear answer as to the ethnic origin of the dedicants but we can reject one: the survey does not offer us evidence of a Celtic resistance to Roman influence. However, even if the number of people with Celtic names was higher for those goddesses, as compared with other deities, Celtic names were quite rare. The comparative rarity of people with Celtic praenomina or cognomina at this time thus allows us to assume that the dedicants of these deities were mainly people who had the skills to practise this Roman form of cults and who decided to make a vow and to fulfill it in the Roman manner. As ‘Romanisation’ became more widespread during the Roman occupation of the provinces of Gallia Belgica and the Germaniae, these skills touched more and more people. Because our archaeological material dates mainly from the 2nd and 3rd centuries AD, it is not surprising to see fewer and fewer people with Celtic origin among the dedicants, since many people became Roman citizens in the later years prior to the Antonine constitution.7

Insofar as most of the dedicants had Roman names and were Roman citizens, I need to look deeper into this question to ascertain whether the consort goddesses of the healing gods had dedications made by single individuals or whether ciuitates, uici and pagi also offered dedications. The participation of the priests, the municipal elite, the men from the Equestrian order, and senators should be studied as well (Tables 7 and 8).

Table 7

Goddesses worshipped by municipal elite, men from the equestrian order and senators.

Municipal magistrates Men from the equestrian order and senators
Rosmerta (one decurio and one priest, seuir augustal) and Visucia (one decurio). None
Table 8

Goddesses worshipped by uici.

Goddess Number of uicus
Rosmerta 2

Dedications to the healing gods and their consorts from the higher class and from the orders of Roman imperial society were non-existent and those of the high Roman ciuitas society were rare. This is the conclusion we must draw for the uicus as well. We have no sign of a dedication made by a colony or any other kind of ciuitas in the three provinces of Gallia Belgica and the Germaniae that I have studied. What are the reasons for the dearth of inscriptions to the consorts of the healing gods among people and communities in Roman imperial society? The first hypothesis could be that people and communities of the imperial centres were more attracted by Roman goddesses. However, as far as we know, this was not the case since Hygia, a Graeco-Roman goddess, had only one inscription, proving that the dedicants did not privilege Graeco-Roman deities. The second hypothesis is that the number of senators and members of the equestrian order were not numerous enough in the provinces to be major actors into the goddesses’ cult. But, if this explanation is valid for such men, it is not the case for the municipal elite which was far more widespread across the three provinces. It is worth noting that they concentrated their dedications upon Graeco-Roman goddesses and on tutelary goddesses from their ciuitas as Auentia (Ferlut, 2012). As for the priests and cult servants, only two seuiri augustales were mentioned in the inscriptions and dedications were clearly attached to the imperial cult, as the inscription contained the expression In honorem domus diuinae. Moreover, in both cases of civitates’ elite and priests, only the goddess Rosmerta was worshipped, proving either that her impact was higher than that of the other consorts of healing gods, or that some suburban sanctuaries were more largely integrated in the civic cult. It is, however, complicated to assert this hypothesis with certainty. As we study the provinces of Germania superior and Germania inferior, in which soldiers had a significant impact upon religion, we need to look at the dedications they made. Surprisingly, only one appears: an inscription from a veteran and his wife. So we must conclude that consorts of the healing gods were not attractive to soldiers who preferred goddesses that protected them in their duty and provided supplies and the necessary strength and courage in the battlefield (Ferlut, 2012).

In conclusion, we can assert that consorts of the healing gods in Gallia Belgica and in the Germaniae were largely Celtic and, for a few, sometimes interpreted by the means of the interpretatio indigena. These six goddesses mainly represented the Salus, i.e. the recovered health, but some were just the divine incarnation of the water whose characteristic formed part of the healing process, as perhaps in the case of Inciona. Nevertheless, although these goddesses were primarily Celtic they were worshipped using Roman rituals, principally the uotum, in sanctuaries or complexes that adopted a Roman structure of spatial bipartition. The dedicants also match the general pattern of social and gender distribution I have found in my research into goddesses, i.e. men with a Roman citizenship, well aware of Roman habits.

Competing Interests

The author declares that they have no competing interests.

Appendix 1

The appendix for this article can be found at: http://dx.doi.org/10.16995/olh.43.appendix1.


  1. As this article focuses on the feminine consorts of the healing gods, I have based my survey of the healing gods on previous works and lists made by historians (Duval, 1956; Hatt, 1967; Van Andringa, 2001). [^]
  2. This article is developed from my PhD dissertation, which is about to be published as a monograph. [^]
  3. I am basing my list of healing gods on those gods who have scholarly works already dedicated to them. [^]
  4. For a detailed historiography examining the question of Gallo-roman religious studies, see Ferlut (2011). [^]
  5. For the inscriptions, see the epigraphic corpus in (Ferlut, 2011), available online (https://scd-resnum.univ-lyon3.fr/out/theses/2011_out_ferlut_a_annexes.pdf). [^]
  6. The expression had several forms that could be used in the singular or plural: uotum soluit, libens merito, uotum soluit libens laetus merito, uotum libens, soluit libens, uotum merito, uotum fecit laetus, uotum, libens merito posuit, posuit laetus libens merito, uotum reddidit libens merito. [^]
  7. The Antonine constitution was enacted in AD212 by the Roman emperor Caracalla, who granted Roman citizenship to almost all of the inhabitants of the Roman Empire. [^]


G Alföldy, (1968).  Epigraphische aus dem Rheinland III.  Epigraphischen Studien 5 : 33.

R Brandt, J Slofstra, (1998).  Roman and Natives in the Low Countries: Spheres of Interaction. Oxford: Oxford University Press.

J-L Brunaux, (2000).  Les religions gauloises, nouvelles approches sur les rituels celtiques de la Gaule indépendante. Paris: Errance.

M-O Charles-Laforge, (2014).  Les Religions dans le monde romain. Cultes locaux et dieux romains en Gaule de la fin de la République au IIIe siècle après J.-C.. Arras: Presses de l’Université d’Arras.

A Clifford, (2012).  Die Riten der Anderen.  Mediterraneo Antico 15 (1–2) : 31.

X Delamarre, (2003).  Dictionnaire de la langue gauloise. 2nd ed. Paris: Errance.

B de Sury, (1994).  Goudineau, Ch, Fauduet, I; I and Coulon, G G (eds.),   L’ex-voto d’après l’épigraphie, contribution à l’étude des sanctuaires.  les sanctuaires de tradition indigène en Gaule romaine, actes du colloque d’Argentomagus, octobre 1992, : 165.

S Deyts, (1986).  Cultes et sanctuaires des eaux en Gaule.  Archeologia 37 : 9.

S-A Deyts, (1992).  R Chevalier,   Sources sacrées, stations thermales et ex-voto de guérison en Gaule romaine.  Les eaux thermales et le culte des eaux en Gaule et dans les provinces voisines, Actes du colloque du. 28–30 septembre 1990, Aix-les-Bains Caesarodunum XXVI : 55.

P F Dorcey, (1992).  The Cult of Silvanus: A Study in Roman Folk Religion. Leyde: Brill.

F Dunand, P Lévêque, (1973).  Les syncrétismes dans les religions grecque et romaine, Colloque de Strasbourg. 9–11 juin 1971, : 123.

P-M Duval, (1993).  Les dieux de la Gaule. 4th ed. Paris: Belles Lettres.

I Fauduet, (2014).  Les temples de tradition celtique. 2nd ed. Paris: Errance.

A Ferlut, (2011).  Le culte des divinités féminines en Gaule Belgique et dans les Germanies sous le Haut Empire romain, PhD Dissertation, (published online on theses.fr), Lyon:

A Ferlut, (2012).  B Cabouret, A Groslambert, C Wolff,   Le culte de Dea Auentia, déesse tutélaire de la Colonia Pia Flavia Constans Emerita Heluetiorum Foederata (Aventicum).  Visions de l’Occident romain, hommages à Yann Le Bohec 2 : 793.

A Ferlut, (2012). Les soldats et les divinités féminines sous le Haut-Empire romain In:  Le métier de soldat, Actes du 5ème congrès international sur l’armée romaine. Lyon: CERGR-Lyon 3, pp. 203.

A Ferlut, (2014).  Ritual Practices for Celtic Goddesses from Gallia Belgica and the Germaniae.  FERCAN Workshop. October 2014, Lampeter Lampeter University (under contract and forthcoming).

R Haeussler, (2008).  J d’Encarnação,   How to Identify Celtic Religion(s) in Roman Britain and Gaul.  Divinidades indígenas em análise. Divinités pré-romaines – bilan et perspectives d’une recherche, Actas do VII workshop F.E.R.C.AN Cascais. 25-25.05.2006, Cascais Coimbra/Porto: : 13.

R Haeussler, (2012).  Interpretatio Indigena. Re-inventing Local Cults in a Global World.  Mediterraneo Antico 15 (1–2) : 143.

J-J Hatt, (1985). Apollon guérisseur en Gaule. Ses origines, son caractère et les divinités qui lui sont associées In:  A Pelletier,   La médecine en Gaule. Villes d’eaux et sanctuaires des eaux. Paris: Picard, pp. 205.

F M Heichelheim, (1963).  s. v. Quadruviae, RE 24, : 711. col.

R Hingley, (2005).  Globalising Roman Culture: Unity, Diversity and Empire. London and New York: Cambridge University Press.

I Kajanto, (1965).  The Latin Cognomina. Helsinki: Societas Scientarum Fennica.

H P Kuhnen, (1994).  Religio Romana. Wege zu den Göttern in antiken Trier. Trier: Rheinisches Landesmuseum Trier.

P-Y Lambert, (2003).  La Langue gauloise. 2nd ed. Paris: Errance.

Y Le Bohec, (2008).  Romanisation ou romanité au temps du principat: question de méthodologie.  REL 86 : 127.

S Lepetz, W Van Andringa, (2008).  Archéologie du sacrifice animal en Gaule Romaine, Rituels et pratiques alimentaires. Montagnac: Errance.

P Le Roux, (2004).  La Romanisation en question.  Annales. Histoires Sciences Sociales, 59° année 2 : 287.

P Lévêque, (1973).  Essai de typologie des syncrétismes.  Les Syncrétismes dans les religions grecque et romaine, Actes du Colloque de Strasbourg. 9–11 juin 1971, : 179.

E Mattingly, (1997).  Dialogues in Roman Imperialism: Power, Discourses and Discrepant Experiences in the Roman Empire. Portsmouth: JRA Supplementary.

P Méniel, (2008). Sacrifices d’animaux, traditions gauloises et influences romaines In:  S Lepetz, W Van Andringa,   Archéologie du sacrifice animal en Gaule romaine. Montagnac: Éditions Mergoil, pp. 147.

M Millett, (1990).  The Romanization of Britain: An Essay in Archaeological Interpretation. Cambridge: Cambridge University Press.

S Price, (2012).  Religious Mobility.  Journal of Roman Studies 102 : 1. DOI: http://dx.doi.org/10.1017/S0075435812000056

N Roymans, (1990).  Tribal Societies in Northern Gaul: An Anthropological Perspective. Amsterdam: Albert Egges van Giffen Instituut voor Prae- en Protohistorie.

C B Rüger, (1970).  Gallish-Germanischen Kurien.  ES 9 : 251.

J Scheid, (1992).  Epigraphie et sanctuaires guérisseurs en Gaule.  Mélanges de l’Ecole Française de Rome 104 (1) : 25. DOI: http://dx.doi.org/10.3406/mefr.1992.1745

J Scheid, (1998).  La religion des Romains. Paris: Cursus.

J Scheid, (2000).  W Van Andringa,   Réflexions sur la notion de lieu de culte dans les Gaules romaines.  Archéologie des sanctuaires en Gaule romaine, : 19.

J Scheid, (2007–8).  Le culte des sources et des eaux dans le monde romain.  Lectures at college de France, Available at: http://www.college-de-france.fr/site/john-scheid/course-2007-10-11-14h30.htm.

W Schulze, (1966).  Zur Geschichte Lateinischer Eigennamen (1904). Zürich: Weidmann.

H Solin, O Salomies, (1994).  Repertorium nominum gentilium et cognominum Latinorum. Hildesheim, Zürich and New-York: Olms-Weidmann.

W Spickermann, (2003).  Germania superior. Tübingen: Mohr Siebeck.

W Spickermann, (2008).  Germania inferior. Tübingen: Mohr Siebeck.

C Vaillat, (1932).  Le culte des sources dans la Gaule antique. Paris: Aubin.

W Van Andringa, (2002).  La religion en Gaule. Piété et politique (Ier–IIIème siècle après J.-C.). Paris: Errance.

J Webster, N Cooper, (1996).  Roman Imperialism: Post Colonial Perspective. Leicester: University of Leicester.

E M Wightman, (1986).  Pagan Cults in the Province of BelgicaANRW 11 (18) : 542.

G Woolf, (1998).  Becoming Roman: The Origin of Provincial Civilization in Gaul. Cambridge: Cambridge University Press, DOI: http://dx.doi.org/10.1017/CBO9780511518614