Evidence from other Roman frontiers (e.g. Derks 1991, 1998) makes it clear that Roman forts had elements of a sacred landscape around them - formal architectural foci such as temples and shrines, natural 'sacred' places such as springs and bogs, elements of ritualised deposition within settlement contexts, and the landscapes of the dead, with cemeteries and tomb monuments. Yet, although religious inscriptions and tombstones have long been a focus of study in Roman Scotland (Keppie 1998), there has been very little attempt to contextualise these in terms of their landscape or social context. This is surprising, since in many cases the hooks for further research are already present; in others, future research projects, or modifications to current practice in developer-funded archaeology, can be clearly identified.
4.5.1 Religion and deities
Altar to the goddess Ricagambeda, from Birrens. She was probably a goddess from the lower Rhine homelands of the Second Cohort of Tungrians who erected this altar ©NMS
The Roman period provides easily recognisable evidence for religion, such as deities' names on inscriptions, iconography, and temples, and there is a tendency within the imperial boundaries to focus on this at the expense of the more difficult-to-interpret late prehistoric evidence. Due to the brief and sporadic nature of Roman occupation in Scotland, there is barely any sign of fusion between Roman military religion and local religious practices (usually conceived as Romano-Celtic syncretism). Some religious practices in Roman-period Scotland would have been influenced by proximity to imperial Rome (such as the inclusion of Roman material in votive hoards), but the reception of any new ideas and practices among the indigenous population would be largely dependent on the attitudes of the Iron Age societies living in Scotland (Hunter 1997; 2001). Andrén (e.g 2005) has argued that Old Norse mythology was markedly influenced by Roman religion, but there is little material sign of this in Scotland; for instance, religious statuettes are rare in indigenous contexts, in contrast to Denmark.
A wide range of deities is attested in Roman Scotland from altars and inscriptions (Keppie and Arnold 1984; RIB I; Hanson and Maxwell 1986, 182-6, 191-2): the official cults (notably the cult of the Emperor and of Jupiter); the traditional Graeco-Roman pantheon; deities introduced from elsewhere (from the homelands of the soldiers, such as Hercules Magusanus from the lower Rhine, or other popular 'exotic' cults, such as Jupiter Dolichenus from Syria); and local (pre-existing or invented) deities. This indicates a vibrancy and variety of religious experience among frontier communities, although surviving epigraphic and sculptural evidence is exclusively 2nd or early 3rd century in date. The limited length of occupation of the Antonine Wall was apparently not enough to establish local cults among the military communities, as occurred on Hadrian's Wall. This leaves us with very few local gods, although there are dedications to Brigantia and Maponus from Birrens (RIB I,2091; III,3482); other 'local' deities, such as the 'spirit of the land of Britain' from Auchendavy (RIB 2175) are Roman inventions, not indigenous perspectives.
Although inscribed monuments from Scotland are not as numerous as from England, comparison between epigraphy and patterns of worship (e.g. social rank, military position and gender of the dedicatee) on the Antonine Wall and Hadrian's Wall could be useful for providing chronological indicators and social interpretations of the larger southern corpus.
4.5.2 Temples/ritual sites
The Carrick Stane. A Roman altar near Cumbernauld, perhaps still in its original location © Fraser Hunter
There were temples in Roman Scotland, but none has seen modern study. Indeed, few have even been located: the remarkable monument of Arthur's O'on was destroyed in the 18th century, while, more shockingly, in modern times what may have been a temple podium at Easter Langlee (Borders) saw no significant investigation prior to its destruction (Steer 1966). Yet over the years stray finds of altars and sculpture should have guided research to these areas; indeed, one altar is believed still to stand in its original position, offering remarkable opportunities for research - the Carrick Stane, sitting forlornly in a housing estate at Cumbernauld (Donelly 1897). The editors of RIB I attributed altars to their nearest forts, but in many cases the original findspots offer clues to extra-mural ritual sites. The altars to Mithras and Sol discovered in 2010 probably lay in a religious structure, but the constraints of the development limited the amount of work done in their surroundings.
Temple sites are a priority for exploration to broaden the picture of frontier life. This requires a focus of enquiry away from the immediate fort and into the extra-mural areas associated with military installations. Any new finds of sculpture and altars might provide clues to the locations of cult sites (such as the recently-discovered Inveresk altars), and old finds have much still to tell. Such sites might not be in close proximity to military sites, as the example of the possible victory monument of Arthur's O'on north of Falkirk suggests (Brown and Vasey 1990).
Engagement with onomastic philology and the Roman place-names of Northern Britain might also provide clues for further research and potentially inform us about Roman Iron Age local religious sites (e.g. the prolific output of A. Breeze, and example of Coates and Breeze 2000 for England). The work of Strang (1997) on Ptolemy's Geography might also be usefully revisited by someone with a detailed knowledge of both Iron Age and Roman archaeology.
4.5.3 Ritual practice and deposition
Hoard of a bronze dipper and strainer set from Gainerhill, near Lanark.©NMS
Ritual practice need not focus on temples. With many of the metalwork hoards known from southern Scotland, which are most plausibly votive deposits, it is impossible to say if these were the outcome of the acts of Roman soldiers or other elements of the frontier community (Hunter 1997); yet the types of deposit (such as vessel hoards) are readily paralleled elsewhere in Roman Britain.
There could also have been deposition in settlement contexts. The ritual nature and religious motives of such acts have seen a long debate in Romano-British archaeology; in Scotland, this has focused on interpretations of the deposits in the pits at Newstead, with explanations varying from ritual to mundane (e.g. Ross and Feachem 1976; Clarke and Jones 1996; cf Curle 1911, 104-115; Manning 1972, 243-46; 2006b). Other examples include the well deposit at Bar Hill and what is plausibly a foundation deposit of denarii at Elginhaugh (Robertson 1975, 12-15; Bateson and Hanson 1990; Bateson 2007). The different types of pits and the range of material deposited in them at the fort at Newstead warns us that activity on Roman forts can be extremely complex, requiring sophisticated and careful examination. Newstead is the obvious place to pursue numerous themes relating to processes of deposition, due to the wealth of material from pits across the site (see Clarke 1997; 1999).
A full and integrated study of the Newstead material in the context of a fresh re-examination of the material itself, to understand better its nature and condition when buried, is required.
Perspectives on structured deposition familiar in prehistory (e.g. Hill 1995) should be applied to Roman settings to test their applicability (as, for instance, with the tool hoard from Strageath; Hunter 2006b, 85-6).
Findspots of stray finds likely to represent deliberate deposits should be excavated in an attempt to clarify their setting.
4.5.4 Landscapes of death
Evidence for Roman burials has so far proved elusive, with a scarcity of excavated burials (the few known burials are isolated) and precious few examples of tombstones or funerary sculpture (see Collard et al. 2000 for an initial summary; Davies 1976 for potential tombstone at Auchendavy). This again arises in large measure because the focus of previous research was on the fort itself rather than its wider context. Chance discoveries (such as the Cramond lioness, found in the river near the fort, or the Carberry tombstone, located two kilometres from the nearest fort) emphasise the need for a geographically broad perspective (Hunter and Collard 1997; Hunter and Keppie 2008). An obvious focus of attention should be the vicinity of road-line corridors, already known as likely locations for Roman burials. Recent work north of Inveresk fort has found a scattered cemetery (ex inf CFA Archaeology), and it is likely to be through such development control work that cemeteries are located. These need full, careful excavation and extensive post-excavation work to understand the range of rites (with cremation, supine inhumation, crouched inhumations, decapitated burials, and Iron Age style cist burials all having been recorded in the sparse record). The bodies themselves could yield information from standard physical anthropological approaches and from isotope work on bones and teeth. Examples from England provide good case studies for best practice in cemetery analysis (e.g Cool 2004).
Geophysical or aerial survey might find evidence for such landscapes of belief if focused beyond the expected features of forts. Development on the peripheries of Roman military areas should also be carefully monitored as important evidence for ceremonial/ritual activity may be under threat beyond areas recognised as likely to produce material of interest. In particular, areas adjacent to known road lines need careful monitoring, as cremation burials (the norm at the time) would be difficult to find in evaluation trenches, and would require careful monitoring and excavation.
Excavation of a Roman cemetery, would be of great benefit to understanding life and death on the frontier; such sites are rare generally on the northern frontier.