The nature of our information about the Celts and their religion comes from a number of different sources, all of which have to be treated with a degree of caution, for reasons which will become apparent. The evidence is composed first of contemporary literature written, however, not by the Celts themselves who had no tradition of a written language, but by their Mediterranean neighbours. Second, there is archaeological material pertaining to the pre-Roman (which I term ‘free Celtic’) and Romano-Celtic world. Third, there exist vernacular written sources in Irish and Welsh. The problem is that none of these sources comes under the category of direct information. That would only be the case if the Celts had written in detail about themselves. Every piece of Graeco-Roman and vernacular literature is in a very real sense second-hand: first, because comments were made by an alien people far removed in cultural terms from the object of their remarks; second, because the post-Roman literature is separated spatially and temporally from the Celts of the later first millennium BC and the Roman period. The evidence of archaeology is at best incomplete and ambiguous; at worst, it is misleading and confusing. The survival (or lack of it) of the evidence is one problem; its interpretation is another. As Piggott so rightly points out there is great difficulty in interpreting – especially in the area of religious beliefs – by archaeology alone; any attempt at an explanation of Celtic religion must at best be extremely speculative – a construction rather than a reconstruction.
The evidence of archaeology for the prehistoric Celtic period (roughly sixth–first century BC depending on geography and the timing of the Roman conquests) may be divided into that of cult-sites including votive/ritual deposits and evidence of sacrifice, and shrines and natural features; burial rites; and iconography, including pre-Roman coins. During the Romano-Celtic period, evidence is augmented by inscriptions, an increasing number of substantial religious structures and a vastly increased iconography – mainly in stone. Sometimes archaeological and written sources are in concert but frequently they conflict, as we will see later, and this makes for difficulties. Of the different kinds of archaeological evidence, some may be more unequivocal in terms of religious interpretation than others. Inferences may be made about cult activity on the ground in the form of suggested shrines and votive deposits – but since we have no dedications to the gods in the free Celtic period, we can only argue as to the likelihood of ritual function from the seemingly irrational nature of such material. For instance, we may infer that some later Bronze Age hoards may have a ritual purpose, perhaps because the contents have been deliberately and carefully laid out, as at Appleby, Lincs. or because there is evidence of ritual breakage. Similarly, it is possible to infer that certain structures, as occur in such pre-Roman fortified settlements as Danebury, Maiden Castle and South Cadbury, may be sanctuaries, since they do not fit into any patterns of secular activity.
Iconography in Celtic art likewise conveys ambiguity in terms of purpose. In some instances, like the few pieces of Hallstatt or La Tène-period, figure-sculpture from Württemberg and the Rhineland, it is fairly evident that religous personages are being represented. But where, as on most pieces of Celtic metalwork, human or animal figures are part of an overall decorative design, it is much less easy to be sure of anything other than ornamentation. It is worthwhile here to look in slightly more detail at Celtic art and iconography to examine the forms in which possible evidence for deity-representation are present. Figural bronzework of a possible religious character occurs, in non-Mediterranean Europe, in the later homelands of the Celts, from at least the later second millennium BC. The Danish Trundholm ‘sunchariot’ (strictly speaking outside our geographical area) with its solar disc and horse-team, dates to around 1300 BC. From the twelfth century BC, small in-the-round bronze water-birds appear in Central Europe and seem to possess some form of talismanic significance. In the later Urnfield period repeated motifs on sheet-bronze include aquatic birds, sun and ship-symbols in association. Early Celtic art really begins with the rich burials of the Hallstatt trader-knights. Birds, horses and cattle appear on bronzework; and one may point to the unique gold bowl from Altstetten, Zürich, which may be a cult-vessel decorated with sun, crescent-moon and beasts. La Tène metalwork is predominantly decorative as we have seen but it is a matter of opinion as to whether the, sometimes grotesque, faces which peer out from abstract and stylised foliage-designs on bracelets, like the gold one from Rodenbach with flowing Celtic moustache, or the bronze example from la Charne, Troyes have any religious significance. Megaw argues that La Tène art employs iconography which endows even the simplest items with symbolism (Megaw 1970, 38). Whilst Britain is less oriented towards humans and animals in its art, the later La Tène material is more definitely representative, both in insular and continental contexts. In Denmark, Celtic cauldrons from Brå and Rynkeby, were both probably votive offerings and quite possibly originally the possessions of priests. The latter dates to the first century BC and is ornamented with a human head and ox-heads. Buckets like those from Aylesford, Swarling and Marlborough bear human heads which presumably have symbolic significance. The Witham Shield depicted a boar-motif, albeit stylised but unmistakably an isolated boar-figure. The iron fire-dogs found in tombs of the immediately pre-Roman period, like those from Barton, Cambs., bear unequivocal bull-motifs. On the Continent, figural bronzework becomes relatively common in the immediate pre-Roman period: boar-figurines are plentiful, illustrated by the large example from Neuvy-en-Sullias; and human bronze figures are not unknown – as for instance the cross-legged god from Bouray. Before we leave metalwork, we should look briefly at pre-Roman Celtic coinage, since this sometimes portrays figural representations. Allen argues that such items may well have a symbolic function simply because their primary purpose was as largesse and as a gauge of wealth. He rightly points out that in terms of an art-form, coins stand apart from the mainstream because of limitations imposed both by size and mass-production; indeed, from the middle La Tène period, coins were issued by the million. What is of especial interest is the tracing of iconographic links between coins and other Celtic religious art. An example of this is the coins of the Aulerci Eburovices of the Evreux region whose motifs link closely with the sculpture of man and boar from Euffigneix. The coins show a boar-motif superimposed on the neck of an anthropomorphic representation and the stone depicts a torced human figure with a boar carved along its torso. (Allen 1976a and 1980).
Pre-Roman Celtic stone iconography is rare. Two main continental clusters exist – an early group in Germany and a somewhat later set (fourth–second century BC) far away in Provence. One of the earliest Celtic figures is from a late Hallstatt tumulus at Hirschlanden, north-west of Stuttgart, where, possibly originally positioned at the summit of the mound, is a huge sandstone figure dating to the end of the sixth century BC and wearing a helmet, torc, belt and dagger. Also from Germany are a janiform (double-faced) pillar from Holzerlingen, whose heads are horned, and stone heads from Heidelberg and in relief on the Pfalzfeld Pillar. From further east comes a moustached, very Celtic-looking head from Mšecké Žehrovice in Czechoslovakia.
The southern French material is especially interesting partly in the amount and variety of sculpture present, but also in the fact that a number of cultural/ethnic forces were at work. Fortified oppida in this area belonged to Celtic tribes having associations on the one hand with the Ligurians of south-eastern France and with Greek colonists from Marseilles, on the other. There was certainly Greek influence in the establishment of built shrines and in the use of stone depictions, but the style and content of the iconography is undeniably Celtic. For purposes of this introductory survey it will be sufficient to look briefly at two key sites – Entremont and Roquepertuse. Entremont was the capital of the Saluvii, sacked by Rome in 123 BC. The shrine stood on the highest ground and possessed limestone pillars with incised carvings of human heads, and other sculptures mostly again associated with disembodied heads. At the cliff-sanctuary of Roquepertuse, located not far away, three stone pillars formed the portal to the shrine. These columns had niches containing human skulls, and a great stone bird stood poised on the cross-beam. Squatting warriors, one bearing a Celtic torc, and a Janus-head held in the beak of a huge bird, are among the repertoire of this southern Gaulish mountain-temple. It is probable that most of the sculptures in the Provençal group date from the fourth–second centuries BC, with a floruit perhaps during the third century.
In pre-Roman Celtic iconography of the La Tène period as a whole, figure-sculpture is usually relatively simple and human and animal forms are subservient to the overall design especially in metalwork, as in the Waldalgesheim bronzes but also exemplified by the Pfalzfeld stone pillar where human heads present a fluid, stylised appearance entirely in keeping with the surrounding scrollwork pattern. A representative style did not fully develop until the period of Roman rule which stimulated figural portrayal, albeit largely ignoring the ‘heavy hand of Roman classicism’. Romano-Celtic stone iconography, influenced by Roman art-formulae but exhibiting religious forms and themes alien to the Mediterranean world, demonstrates that before the conquest there must have existed local and tribal gods, each with a name (four hundred different Celtic gods are named on Romano-Celtic inscriptions) and each with specific qualities. Indeed, some non Graeco-Roman symbols appearing in the Roman period are known before – for example the horned head exhibited on the Holzerlingen sculpture, and the emphasis on human heads and animals. We will see in later chapters that isolation of iconographic types during the Romano-Celtic period is one of the best ways of classifying cult-objects, even though this must necessarily contain an element of arbitrariness.
Before we leave iconography, the use of wooden sculptures should be mentioned. Pre-Roman figure-carving seems sparse indeed when we look at the medium of stone. But the chance survival in water-logged conditions, as at the Seine sanctuary near Dijon, shows that there was a tradition of pre-Roman figure-sculpture not confined to the few stone pieces mentioned. This is of particular note when we consider the passage of Lucan where the presence of wooden effigies to the gods in a Gaulish sanctuary is specifically mentioned.
As commented upon above, two kinds of literary record exist for the Celtic world. Graeco-Roman and vernacular Celtic. The Celtic world and its customs – including religion – was discussed and described in varying detail in writings of Mediterranean authors, the main ones being Caesar, Strabo, Pomponius Mela, Pliny, Athenaeus, Tacitus, Dio Cassius, Diodorus Siculus, Ammianus Marcellinus and Lucan. Four of these writers base their evidence (acknowledged or unacknowledged) upon earlier but lost writings of Posidonius: Strabo (63 BC–AD 21), Diodorus (writing circa 60–30 BC), Caesar (in the mid-first century BC) and Athenaeus (circa AD 200). These writers are in general more useful and detailed about Druids and ritual than about the gods themselves. For instance, Caesar speaks as if the Celts’ deities were identical with those of Rome and gives them Roman names, whilst Lucan (second century AD) talks of three Celtic gods – Esus, Teutates and Taranis – as if they were really important, a suggestion not supported by epigraphy which names these gods very infrequently. Classical writers have to be used with caution; they are biased by what interested them, by choice or chance of recording and by cultural separation, ignorance and consequential misinterpretation. Caesar on the Druids, for instance, must be seen in the light of his deliberate embellishment of an alien priesthood for politico-propaganda purposes. However, most Celtic ritual, alien though it was in detail, was explicable to the Roman mind, for Mediterranean peoples too were fettered by the concepts of correct and contractual appeasement and propitiation. Only weird and obscene rites – head-hunting, human sacrifice, divination by ritual murder – were curious and distasteful enough to be commented upon in detail. Where classical writers are particularly valuable is precisely in areas where their evidence marries with archaeological data: for instance both demonstrate the existence of human sacrifice and of head-collection. With deities themselves there is less comfort. The vast wealth of iconographic evidence for Celtic gods during the Romano-Celtic period is a subject upon which Graeco-Roman authors are vitually silent.
The other major body of literature is itself Celtic, so it does not suffer the cultural alienation of classical sources. However, it brings with it its own set of problems based partly on temporal separation and partly on its being specific to the fringes of the Celtic world during the thousand years (fifth century BC–fourth century AD) of pagan celticism. As we have seen, there are no indigenous literary references to the La Tène or even the Roman period. There is a danger even where Gaulish archaeological evidence appears to match the insular data since the two types of source are separated by at least several centuries in recording. The Irish evidence may sometimes be specific to Ireland: for example, the religious festivals of Samain and Beltine are related to stock-rearing and pastoralism, not necessarily relevant to lowland Britain and Gaul. We have to bear such constraints in mind when assessing the vernacular material. When Britain and Gaul were under Roman rule, Ireland possessed a heroic society, basically prehistoric-Celtic in terms of developmental stage, whose exploits are discussed in ballads and poems of which the earliest began to be written down sometime in the eighth century AD. For our purposes the group of prose tales known as the Ulster Cycle is of most use. These describe in epic form a series of events pertaining to a specifically Homeric-type heroic, aristocratic, warlike and hierarchical society. Jackson sees this as definitely related to Iron Age Ireland, though Champion wonders if these tales might not be conscious imitation of Homer in early Christian times. Archaism is evident both in the political status of Ulster and in the political structure, customs and material culture described, which all apparently belong to a period some centuries earlier than the time at which they were first written down. The stories centre around the King of Ulster and his followers at a time when Ulster was a large and powerful kingdom with its capital at Emain Macha near Armagh and whose over-king was Conchobar. The opponents of Ulster were the Confederacy of the rest of Ireland led by Ailill of Connaught and his warlike and dominant queen Medb. The main activities appear to have been fighting, cattle-raiding and feasting. The proof that the background to these tales was earlier than the introduction of Christianity in the fifth century AD is based on a number of arguments. By the fifth century, the whole political framework of Ireland had altered: by now Ulster was much smaller and insignificant, its greatness having been smashed by the family of Niall (who died in about AD 404). Thus events in the Ulster Cycle are arguably older than this change. Likewise, though Christianity was established in the fifth century, the heroes in the stories swear not by God but by the gods of their tribes. Jackson dates the formulation of the body of narrative material recorded around the third or fourth century AD embodying tradition going back perhaps to the second century BC; thus the Ulster Cycle would describe events in late La Tène Ireland. But doubt has recently been cast on this terminus post quem for Celtic culture to Ireland. Champion argues that the society described could date much earlier than the second century BC.
Other Irish and indeed Welsh sources must be acknowledged, though their value for the period in question must be minimised by their lateness in compilation and by their very nature. References to religion and mythology in later Celtic literature cannot safely be used to illuminate a subject which is essentially pre-Christian and a part of prehistoric Europe. But it is true, nevertheless, that however late the manuscripts (mostly eleventh–twelfth century AD) they do contain a vast amount of non- and therefore pre-Christian material. For example, many deities occur in the Irish sources – such as Nuadu and Lug – who can be traced also in the Romano-Celtic record (Nuadu may well be the Nodens of Romano-Celtic Britain, and Lug’s name is enshrined in various town-names, like Lugdunum (Lyon) and Luguvalium (Carlisle). The later Irish material exists in such works as the ‘Book of Invasions’, the ‘History of Races’, and the Fionn Cyle, all of twelfth century date. Wales also had a Celtic oral tradition which was rich but poorly documented for the early period. Of the extant tales, the four Branches of the Mabinogi, of late eleventh century date, and the Tale of Culhwch and Olwen are perhaps of greatest interest. But though, for example, the Druids of Caesar and other classical writers play a large role in these late literary works, it is evident that medieval Irish and Welsh antiquarians/historians were fairly ignorant of the actual beliefs of the people about whom they wrote. If we look closely at these works, mythology abounds but there is little tangible evidence for religion apart from the names of certain gods. The message of this medieval Celtic literary material, then, is ‘examine with interest but use with extreme caution’ any possible direct link with the later first millennium BC and the early centuries AD.
These similarities between the archaeological and literary records (both Graeco-Roman and Irish) should be highlighted, since the links do strengthen the authenticity of each, and serve also to counteract the Ireland-specific constraints mentioned earlier. Some similarities between the sources are very significant indeed. The aristocratic society itself, with its hierarchical division into nobles, free landowners and landless men; the system of clientship; the pugnacity, boldness and vanity of the champions are all recorded by classical and later Celtic writers; and the importance of feasting, of pork and otherworld banquets is borne out archaeologically as well. Comments on clothing and weaponry tally and above all the use of the war-chariot is faithfully recorded in all three sources. More important for our purposes, classical and Celtic writers agree on the three learned classes of Druids, Vates and Bards, and on the ritual of headhunting – the latter being also corroborated archaeologically.
Lastly, in assessing our sources of evidence, we should glance at epigraphy, a group of data relevant only to the Romano-Celtic phase (roughly first century AD onwards). Though god-names were inscribed on monuments through the Roman tradition, they very frequently allude to specifically Celtic god-names and it would be implausible to imagine that these names did not exist orally before Roman influence on Celtic lands.
We have looked already at the character of the Celts and their society and at the type of evidence used to construct a picture of religious beliefs and practices. In this section, I wish to touch very briefly on the essence of Celtic religion as projected by this evidence. Detail is unnecessary at this stage since succeeding chapters will examine the most prominent themes at some length. Here my aim is to introduce the kind of religion with which we shall be concerned, and to set the scene by surveying the types of context within which cults were enacted.
RELIGIOUS SITES
Let us first look at places of worship, whether built sanctuaries or natural loci consecrati. It is generally considered that the Celts did not normally construct permanent, roofed temples. Certainly, except for the curious Provençal shrines stone-built sanctuaries are rare in the pre-Roman Celtic world. However, there is a steadily increasing body of evidence for wooden temples preceding Roman examples both in Britain and on the continent. One of the most interesting insular shrines is Frilford in Oxfordshire. In the Roman period, there were two contiguous buildings here – a circular structure and a shrine of Romano-Celtic type (with rectangular cella and surrounding portico). Beneath both of these, there were Iron Age sanctuaries, a round one under the Romano-Celtic temple and a structure represented by a penannular ditch under the rotunda. The evidence for the former as a ritual site is circumstantial, but inside the penannular ditch were six postholes associated with the votive offerings of a ploughshare (perhaps a foundation-deposit) and a miniature sword and shield. Frilford is important because of its specific Iron Age evidence for cult-activity, and in the presence of two contiguous, perhaps complementary, sanctuaries in both the pre-Roman and Roman periods. That Frilford may have been a cult-centre of some significance, at least in the Roman period, is supported by recent field-surveys which have revealed the presence of a small Roman town with an amphitheatre situated close to and perhaps actually within the temple-precinct. The Romano-Celtic shrine at Worth, Kent, was probably also preceded by an Iron Age temple, for pre-Roman material beneath the cella included three Iron Age model shields. At Muntham Court, Sussex, the circular Romano-British temple, which was associated with a cult-well and with evidence of a healing cult, may have had a free Celtic precursor represented by Iron Age pits and postholes. At Haddenham, Cambs., where there was cult-activity from the Bronze Age to the Roman era, a Romano-British shrine first built in the first or second century AD overlay Bronze Age cremations in a barrow, the Roman temenos embracing the barrow itself. Adjacent to this was an Iron Age penannular ditched enclosure. The Roman shrine was associated particularly with cult-activity involving sheep or goat sacrifices. Hayling Island was a pre-Roman shrine with a sacred enclosure surrounding a circular building which may or may not – like Frilford – have been roofed. A central pit possibly held a ritual stone or post. Here the cult-activity was concentrated not on the temple itself but on the outer temenos, where votive deposits of metalwork – including martial gear and cart or chariot-equipment – were offered to the gods, much of the material being first ritually damaged. Many so-called Iron Age shrines have been defined as such simply because they cannot readily be explained in terms of secular function. Small buildings in a number of hillforts are thus interpreted: such is the case at Danebury and South Cadbury where Alcock stresses that the Iron Age town was not only important as a centre for trade and industry but that ‘the focal point of the whole settlement was religious’. At Maiden Castle the only certain religious evidence for a late Iron Age circular building is an infant-burial just outside the door, but it lay under a late Roman round building and near to a Romano-Celtic shrine, and so its context is ambiguous.
There are continental parallels to these British Celtic shrines. In Austria, St Margarethen-am-Silberberg may be compared with the Frilford sanctuaries; the Romano-Celtic temple of St Germain-les-Rocheux (Côte d’Or) had an Iron Age predecessor; and, in the Marne region, small Iron Age shrines were associated with Celtic cemeteries. The evidence for enclosed temple-buildings should not be overstressed. Many rites were performed in the open air, not necessarily involving buildings at all. It is significant that where structures are present, they rarely occupy the central position within the temenos. Thus at Hayling, though there was a building, the main rites seem to have taken place outside. It is important too to realise that in the Iron Age there were few ‘purpose-built’ shrines and Henig rightly points out that the sacred buildings at Hayling and Maiden Castle were little more than large huts, and perhaps served as no more than a focus – rather like a stone, pit or tree. One exception to this is at Heathrow where, in the fourth century BC, a building with cella and portico appears to anticipate the much later and ‘mass-produced’ Romano-Celtic shrine-type. It should be remembered also that although temples following the fairly rigid architectural schema of concentric squares, circles or polygons, to define inner cella and outer ambulatory, predominated in Romano-Celtic western Europe, yet simple circular and rectangular temples – such as those at Brigstock, Muntham and Frilford – were built and used right through the Roman period.
Thus, the important role played by open-air enclosures, as opposed to roofed structures, should not be underestimated. The Frilford shrines may have been open to the sky, and there may have been several other small British sacred enclosures. On the Continent, the pre-Roman trend was towards large communal ritual open-air sites. The enclosures at Aulnay-aux-Planches (Marne) and Libeniče (Czech) are strikingly alike though widely separated in space and time (the French site dates to the tenth century BC and Libeniče to the third century BC). Both consist of sub-rectangular enclosures three hundred feet long containing evidence for possibly sacrificial human and animal burials. The Czech site contained the remains of two burnt upright timbers once adorned with neck-rings, as if they represented images of gods. Two German enclosures, the Goldberg and Goloring, both probably of sixth century BC date, follow an essentially similar pattern: the latter contained a huge central post, evidence of a similar cult-focus to those of Libeniče and Hayling, and perhaps symbolic of a sacred tree or column.
Ritual shafts or pits (discussed in detail in Chapter Four) may likewise have represented man-made foci of worship. In Britain the emphasis seems to have been on late Iron Age south-eastern Britain, but Bronze Age examples at Wilsford and Swanwick indicate the presence of a long-standing tradition. In Germany, certain enclosures or viereckshanzen, dating to the late La Tène if not earlier, contain shafts which were foci of cult-activity: at Holzhausen an enclosure contained three shafts: one, eight metres deep, with a wooden pole at the bottom surrounded by traces of flesh and blood, echoing an almost identical occurrence at Swanwick much earlier, at about 1000 BC. The inference at both places is that a human sacrifice may have been tied to the stake and offered to a deity, perhaps of the Underworld.
Actual cult-offerings at pre-Roman shrines are sparse compared with Roman evidence, and this is what makes the interpretation of free Celtic cult-sites sometimes ambiguous. Human and animal sacrifices are sometimes suggested, as at Cadbury, Aulnay and Maiden Castle; model objects, so common in Romano-British shrines appear at Frilford and Worth; and deposits of metalwork were offered at Hayling. Celtic gold coins were consecrated to the deities of Harlow and at Hayling coins were covered with gold to present a glittering show to the god. Perhaps some of the most curious offerings are those of an ‘antiquarian’ nature: a tradition was established in pre-Roman sanctuaries in Britain and Europe where Neolithic axes were offered to Celtic divinities. An additional practice in Gaulish shrines was ritually to smash these axes, and both here and in Britain, the implements were collected for votive purposes both by free and subsequent Romano-Celtic devotees.
Apart from deliberately constructed sanctuaries, the Celts made great use of natural topographical features. The Celtic word ‘nemeton’ denoting a sacred grove may be traced in derivation form in Celtic place-names from Britain (Aquae Arnemetiae at Buxton for example) and Spain to Galatia in Asia Minor (Drunemeton). The Irish for ‘nemeton’ is fidnemed. A number of classical authors, too, refer to sacred groves. Strabo speaks of the reunification of three Galatian tribes in a grove of sacred oaks, at Drunemeton for the purpose of discussion on government matters. Tacitus speaks of the forest-clearings of Anglesey as the last Druid stronghold against Rome. Dio Cassius refers to a sacred wood where human sacrifices to a war-goddess Andraste were carried out and Lucan refers to grim sacred woods in southern Gaul, which were sprinkled with human blood. It is possible that in some instances the term ‘nemeton’ may have been used loosely as synonymous with ‘sanctuary’. Later commentaries on Lucan say that the Druids worshipped gods in woods without the use of temples. Romano-Celtic epigraphy informs us that deities dwelt in natural features such as mountains, rivers and springs. Allied to this, the archaeological evidence for votive deposits points to the offering of precious objects, frequently of a martial nature, to gods associated with the ground or underworld (Chapter Four) or with water (Chapter Five). It is interesting in the context of aquatic and warrior offerings that such practices are endorsed by Graeco-Roman authors on the Celts. Strabo mentions the treasure of gold ingots at the sacred lake belonging to the Volcae Tectosages at Toulouse in 106 BC; and Caesar remarks on dedications of weapons and booty heaped on the ground in honour of the god of the winning side; the sanctity of the hoard was such that it was left unmolested and did not need to be guarded.
THE SANCTITY OF NATURAL FEATURES
The preceding discussion of shrines and sanctuaries leads logically to examination of a very significant trait in Celtic religion, that is the endowment with sanctity of natural features – a river, spring, lake, tree, mountain or simply a particular valley or habitat. The gods were everywhere, and this is expressed during the Romano-Celtic period by god-names which betray this territorial association. The same kind of worship is found in rural Italy during this time, culminating in the relatively sophisticated Roman Genius loci, but it is in the Celtic milieu that topographical features assume particular emphasis. We have already alluded to votive deposits in or linked with water, and this is evidenced epigraphically. Water welled from springs in the ground and rivers flowed without obvious cause; many Romano-Celtic divinities were closely associated with water-veneration (Chapter Five), especially in its healing capacity. Trees and clumps of woodland too were revered. We looked earlier at the role of groves in the context of sanctuaries, and it is interesting that names of deities often reflect the sanctity of forests and trees: Nemetona (goddess of the Grove) is recorded at Altripp near Spier, and the name of the tribe in whose territory she was worshipped was the Nemetes. Single trees too were important: such tribal names as the Eburones (yew tree) and Lemovices (elm) recall the essential rapport with trees. Epigraphy in Gaul also relates to individual trees or tree-species; and dedications were made to Fagus (beech tree) in the Pyrenees, and there are many other similar examples of tree-reverence. Pliny comments on the association of oaks and Druids and it may be that the large timbers associated with such sites as Hayling and Ivy Chimneys reflect the use of trees or tree-substitutes as cult-foci.
Other topographical names demonstrate the close link between divinity and the land. The god Alisonus (Côte d’Or) recalls the town-name of Alesia. Mountain-deities such as Vosegus of the Vosges, or Pyrenean dedications simply to ‘mountain’ reflect this territorial awareness. This veneration of natural phenomenon extends, too, to animals (Chapter Six). In this context it is enough to mention epigraphic allusion where it is associated with territory: the tribe of the Taurini and the place-name Tarva (Tarbes) in southern Gaul display the importance of the bull to the Gauls. Likewise the bear-goddess Artio of Muri in the Berne area of Switzerland, and the place-name Artomagus are significant; Arduinna the boar-goddess is etymologically linked with the forests of the Ardennes.
PRIESTS AND RITUAL
In the free (pre-Roman) Celtic phases, there is little archaeological evidence specifically related to priests and associated with ritual. During the Romano-Celtic period, apart from temples themselves and inscriptions referring to priests, we have evidence for regalia and ceremonial equipment. Crowns and headdresses, as have been found at Hockwold, Norfolk and Cavenham, Suffolk, suggest that local clergy did wear liturgical garb, and it may even be that the gold chains with sun- and moon-like symbols from such sites as Backworth in Durham and Dolaucothi, Dyfed (Chapter Two) were badges of office. Objects best defined as sceptres occasionally occur, for instance at Willingham Fen and Farley Heath (Chapter Two), and sistra or rattles may be evidenced at Brigstock, Willingham and Felmingham Hall, Norfolk. The curious object from Milton, Cambs. may be a similar item of ceremonial: here a long-shafted bronze terminates in a flat oval with two perforations and may have supported bells for jangling in processions; the shaft is decorated with incised signs including St Andrew’s Crosses, which could be sacred symbols. A fragment of what may be a similar object, consisting of a bronze rod marked with diagonal crosses, has recently been found in Chelmsford. Some other regalia is of interest since particular attention is paid to the human face, and we know that the head held a religious fascination for the Celts. Metal face-masks, like those from Bath and Tarbes could have been made for the priest to hold in front of his face during religious ceremonies; perhaps, on certain occasions, it was forbidden to look on the face of one so close to the divine, or alternatively, the priest could have been shielding himself from the presence of the deity. The Tarbes mask may date as early as the third century BC. In the context of masks, a large one, reputedly from East Anglia is particularly interesting; it was made of gold and in typical Celtic style, has round eyes and prominent brows. This mask was not worn; perforations in the gold suggest that it was nailed up, perhaps on the wall of a shrine, to represent a god. All over the Roman world metal plaques, often feather or leaf-shaped in bronze or silver, were made with a similar function of adorning shrines with the image or dedication to the resident deities. Conversely, lead defixiones or curse-tablets were employed with the more sinister purpose of condemning a wrongdoer in the eyes of the gods.
Little is known about the objects used in Celtic liturgy. Perhaps the most famous category of item is the ritual vessel or cauldron. Cult-vessels were being used from the Middle/ Late Bronze Age in Europe, and it is suggested that there could be a connection with water-cults. Celtic cauldrons are ubiquitous in archaeological and historical records, but the most famous vessel is the great gilded silver bowl from Gundestrup in Denmark. This cauldron is referred to repeatedly in this book, and the mythological scenes on its silver plates are much-discussed. There is some controversy as to whether the vessel is Celtic or not and whether its origins lay in the Gaulish world. Caution must be exercised in any interpretation: certain motifs – like the carnyx or war-trumpet and shields – are definitely of La Tène type; but the silversmith had access to and used a wide range of symbols, some of which have no parallel in the western Celtic repertoire of iconography. Indeed, Collis has pointed out that the closest artistic parallels lie in Romania and argues that the cauldron may have been made there. I believe that there are Celtic elements in the religious imagery as well as in weapon-types; here the stag-horned god is surely the Celtic Cernunnos, and the ram-horned snake is too idiosyncratic a beast to belong to more than one culture. Still, the cauldron’s iconography is an amalgam and its provenance in Jutland should also not be forgotten.
The most famous (or infamous) priesthood in the Celtic world was the Druidic. A number of Graeco-Roman authors, including Strabo, Diodorus Siculus and most notably Caesar describe the Druids, but these three derive their information from a single source, Posidonius, and so the common ground between them is not itself significant. There is no need to be surprised at the existence of a powerful religious leadership within the context of Indo-European society, but there are varying views as to the real importance of the Druids in the Celtic world. Tierney and Harding argue that their significance may be more apparent than real in that Caesar confines his remarks to a small part of Book VI of De Bello Gallico. Nash points out that the influence of the Druids was perhaps largely social rather than political and that in this role they were probably very powerful indeed.
It appears that once the main aristocratic hierarchy of Celtic society disintegrated under the impact of Rome, the Druids lost their national influence, for they must have been heavily dependent upon tribal chiefs in terms of status and support and even without Roman persecution, the Druids’ power would naturally have dwindled. Historical evidence indicates that Druidism was tolerated in Augustan times as long as Roman citizens were not involved. Tiberius opposed the priesthood, and Claudius tried hard to eliminate it. The Druids enjoyed a brief comeback in the Gaulish rising under Civilis in AD 69–70, but thereafter history is silent until the third century when Druids are recorded as being associated with prophecies against Severus Alexander and Maximin. In the fourth century the Druids reappear, and the Bordeaux poet Ausonius refers to two famous Druids and the tradition of succession from father to son.
There are recurrent references to Druids, as prophets and soothsayers, in the Irish sources; perhaps the most prominent was Cathbad of Ulster. Our evidence for Druids in the free Celtic world comes from Graeco-Roman literary sources. Whatever the precise nature of their influence, there is no doubt that they were a national religious force in Gaul and Briain. Caesar informs us that their main cult-centre was in Britain and that they originated there. The Druids were certainly magicians, if nothing else, in whose power was the entire oral tradition of religion and ritual. Caesar, our most prolific source, says that the Druids officiated at the worship of the gods, regulated sacrifices and gave rulings on all religious questions. On a fixed date every year they assembled in a sacred place in the territory of the Carnutes (near Chartres); and they were no doubt responsible for the fertility cycle and all its associated cult-activity. Pliny tells of a feast prepared on the sixth day of the moon, involving the Druids who climbed a sacred oak, cut off a mistletoe bough using a ‘golden’ sickle and caught it in a white cloak; two bulls were sacrificed. The importance of the Druids in divination cannot be over-estimated; in both classical and insular sources, they are recorded as being not only priests but also prophets. The importance of omens is demonstrated by the existence both in Ireland and Gaul of unlucky and lucky days. The Coligny Calendar, found near Bourg-en- Bresse (Ain) is probably Augustan: the bronze sheet bears epigraphic evidence for the division of each month into a good and bad half, there being an appropriate time for each act. In the Ulster Cycle of prose tales, the importance of omens is shown by the episode of Queen Medb of Connaught who was prevented by the Druids from commencing battle for a fortnight in order to await an auspicious day.
The main fascination of Graeco-Roman historians for the Druids was their role in human sacrifice, an emphasis which may have been designed deliberately to disgust their readers as a practice typical of outlandish barbarians. The literature makes it clear that as far as the Druids themselves were concerned, the main aim of human sacrifice was not so much the propitiation or appeasement of the gods but divination. Diodorus and Strabo describe the custom of stabbing victims and foretelling the future by observing the death-throes. Evildoers were imprisoned for five years and then killed by impaling: prisoners of war were used as sacrificial victims together, on occasions, with their beasts; the victim could be burned alive in a huge wicker man-shaped image, or shot with arrows (interesting since archery was not a normal method of fighting at this time). British human sacrifice is attested by Tacitus on the island of Anglesey where, in a grim sacred grove, altars were drenched with human blood and entrails (a grisly reminder of the stakes at Swanwick and Holzhausen) which were consulted by the Druids for divinatory purposes. Lucan mentions that the Druids, living deep in the forests, claimed to understand the secrets of the gods. The poet comments that they ‘resumed the barbarous rites of their wicked religion’, and describes the Celtic deities Esus, Teutates and Taranis, whom the Gauls propitiated with human sacrifices. Commentators on Lucan’s text elaborate on this, describing the appeasement of Taranis by burning, Teutates by drowning and Esus by hanging from a tree (perhaps the elements of fire, air and water were deliberately represented). Later in the poem itself Lucan refers to a nemeton at Marseille where altars were heaped with hideous offerings and every tree sprinkled with human blood, and he mentions that even the priests were wary of entering the grove at certain times.
Certainly the interpretation of human sacrifice suggested by Caesar is that the power of the gods could only be neutralised or controlled if one human life were exchanged for another. Thus if Gauls were threatened by illness or battle, then the Druids organised human sacrifice; if criminals were not available, then the innocent would have to supply that life for a life.
We have seen that the Druids were fundamentally concerned with human sacrifice. According to historical testimony, they always officiated at such events. In this connection, it may be significant that the murdered bog-body, Lindow Man was found to have mistletoe in his stomach; it is not impossible that he was a Druidic sacrifice. Strabo mentions that the Cimbri likewise despatched their victims and observed their dying struggles for divinatory purposes; this Teutonic (not Celtic) tribe cut the throats of its victims, collected and then examined the blood which was caught in cauldrons. The distaste of classical authors for human sacrifice is evident from their writings (though it was not so many centuries since it was practised in the Mediterranean world). Archaeologically, there is some evidence for this grisly rite. The suggestive remains at Bronze Age Swanwick and Iron Age Holzhausen have already been referred to, and from the later free Celtic phase in Britain comes fairly clear evidence that human sacrifice was practised on occasions. The infant buried at the Maiden Castle shrine may have been deliberately killed; and at Danebury three male corpses in a pit at the bottom of a quarry hollow, buried when the rampart was built, may have been offerings to the gods of the territory on which the hillfort was constructed. Some of the bodies at Danebury were weighted down with stones perhaps to prevent their spirits rising, and dismembered skeletons suggest ritual body-exposure to allow the spirits time to depart. It is possible here that the burial evidence relates specifically to unclean or sacrificial death. There is other evidence from Iron Age Britain too: burials at, for instance, Wandlebury and Hayling Island (Chapter Four) were maybe the result of some rite involving human sacrifice, and the young man crammed face down in a pit under the rampart at South Cadbury was probably a foundation offering.
Most striking, both in terms of archaeological and written evidence, is the emphasis on human heads. At Danebury, there was a preponderance of adult male skulls, placed in the bottom of storage pits, presumably to bless the corn, unless the pits were being used secondarily as rubbish/burial pits. At the Iron Age oppida of Bredon Hill, Worcs., and Stanwick, Yorks., the position of human skulls suggests that they had been attached to poles at the fort-gates; and this evidence is echoed at the Celtiberian oppidum at Puig Castelar in northern Spain. Shrines too bear witness to the dedication of heads to the gods. A number of the pre-Roman Celto-Ligurian sanctuaries of southern Gaul display the grim tradition of offering the heads of sacrificial or battle-victims. In many of these shrines, from as early as the fourth century BC, skulls were placed in niches cut in the stone structure for their insertion; niches were present at, for instance, St Blaise, Entremont and Roquepertuse (all in Provence). The Roquepertuse skulls probably date to the third century BC, and are interesting in that they are the skulls of strong men in the prime of life, and this may suggest that they represent war-dead. One of the Entremont skulls nailed onto a wall, actually had a javelin-head embedded in it. The presence of carved stone heads as well as real skulls, implies that the head was an essential offering: perhaps if human supply dried up, then symbolic representations would do instead. A shrine of Roman date in Britain demonstrates ritual remarkably similar to those of southern Gaul: at Cosgrove, Northants., a simple rectangular shrine had a human head buried in one wall.
The tradition of head-hunting or head-collecting and ritual implied by archaeology, is mentioned again and again in both Graeco-Roman and vernacular literature. Livy remarks on the killinng of a Roman general Postumius in 216 BC by the North Italian tribe of the Boii, who cut off his head, cleaned it out, gilded it and used the skull as a cult-vessel. Several writers attest the practice among the Celts of collecting the heads of enemies in battle, fastening them to their saddles or impaling them on spears. Livy speaks of Senonian Gauls doing just this in 295 BC when they defeated a Roman legion at Clusium, and Diodorus and Strabo inform us that the heads of important victims were embalmed in cedar oil and cherished above all treasures. The placing of heads in temples, attested archaeologically as we have seen, is supported by the literature.
Both Irish and Welsh sources support the Celtic reverence for the human head, and indeed the practice of human sacrifice. The Ulster hero Cú Chulainn collected the heads of the vanquished and placed them on stones. Another Ulster warrior, Conall Cernach, boasted that he slept every night with the head of a slain Connaught man under his knee, and the Irish sources attest the making of a collection of brainballs as battle trophies (the brain was extracted, mixed with lime and allowed to harden to a cement-like substance). The magical properties of heads are referred to in Welsh and Irish literature: in the Mabinogion Bran is mortally wounded and at his request his companions cut off his head and carry it with them on their travels as a talisman; the head of the Ulsterman Conall Cernach, too, had magic powers and it was prophecied that the men of Ulster would gain strength from using this head, which was of supernatural size, as a drinking-vessel.
The importance of head-ritual is unequivocal. Human sacrifices were made and the head preserved as being of the greatest cult-significance. Battle-victims, whether killed in warfare or after capture, were decapitated and their heads enshrined. The cult-importance of heads manifests itself throughout Celtic religion; in looking at death-ritual (Chapter Four) we will see that the dead were sometimes beheaded; and in iconography (Chapter Seven) the head is frequently exaggerated or it may be used by itself to symbolise the whole body. Why the human head was so important can never be entirely understood, but it was the means of identifying an individual, and was recognised as the power-centre for human action. I refute any suggestion (Chapter Seven) that the head itself was worshipped, but it was clearly venerated as the most significant element in a human or divine image, representing the whole.
Caesar comments that the Gauls were a very religious people; and this is endorsed by the seeming presence of the gods everywhere – in rivers, mountains and in each corner of Celtic territory, as well as by taboos and ritual. The all-pervasive nature of Celtic religion is demonstrated at Camonica Valley where some of the Iron Age (and earlier) carvings occur on rocks in the middle of remote forest. But what form did these gods take? The picture we get – from pre-Roman iconography, literary sources and what we can glean from the Mediterranean gloss on Romano-Celtic evidence – is confused and muddling. Were there many localised gods or was there one main deity with a number of aspects? The evidence points both ways. A glance at the epigraphic record of the Romano-Celtic period reveals about 400 god-names, over 300 of which occur once only; this argues for disparateness, and certainly demonstrates strong localisation. One problem is in the pairing or twinning of Celtic and Roman names: for instance, Mars is linked with several different Celtic epithets. This may indicate no more than the uneasiness of any attempt to equate the strictly functional order of Roman gods with the shadowy, multi-functional character of Celtic deities. It is a mistake to place the elusive Celtic divinity into methodical Graeco-Roman boxes: this is ‘certainly to misunderstand the aggressive individuality of Celtic society. . .’. Sjoestedt would argue that Celtic tribes were too disunited to share a common pantheon. However, there are arguments for some degree of mutual belief or way of looking at the supernatural. The Druids, in matters of ritual at least, may have had a unifying role in the annual council held in Carnutian territory. This commonality is supported to an extent by the widespread distribution of some iconographic god-forms – for instance horned deities; mother-goddesses; triplism; animism and many more – during the Romano-Celtic period. Certain recurring themes, such as the veneration of natural phenomena – especially water – may be traced back into the Iron Age and beyond. In the Romano-Celtic phase some definite iconographic patterns may be established: for example, the horseman-god riding down a giant (Chapter Two) appears to have been a favourite among the tribes of North East Gaul; the hammer-god Sucellus was worshipped principally in the Rhône and Saône Valleys; the three-headed form in Belgic Gaul, especially among the Remi; Cernunnos (the stag-antlered god) in Central France; and divine couples among the Aedui. Because of the paucity of pre-Roman Celtic iconography, we have no means of knowing how far these patterns existed earlier. One should not however seek an organized pantheon for the Celts; their prime concern (as with all prehistoric non-Mediterranean European peoples) was to constrain and control supernatural powers to a beneficent end. Lambrechts argues a view opposite to that suggested by the multiplicity of god-names and forms, and postulates one great, ill-defined, multifaceted deity. All we can say with any confidence is that by the Roman period, with sharper definition and formalisation, both major and minor gods were acknowledged. Before that time the animistic aspect of cult is the most pronounced, together with certain ritual behaviour concerned particulalry with sacrificial offerings and with attitudes to the afterlife. The vernacular sources do not help us a great deal but at least two general features may be isolated. One is the tribal emphasis: Irish gods, like the Dagda, were tribal gods, and in the Ulster Cycle the heroes swear by the gods of the tribe; indeed it has been pointed out that, in general, Celtic male deities were usually tribal and females were invariably types of mother-goddess. The other feature found in the vernacular sources is shape-shifting, the ability to move at will between different human and animal forms. This may, archaeologically, account at least in part for the curious faces and pseudo-faces which leer out from Iron Age metalwork; moreover it may also explain the very fundamental importance of animal iconography in the Celtic world (Chapter Six).
Certain characteristics of physical cult-expression may be isolated as distinctive of celticism, even allowing for the iconographic influence of the Roman period. This may be done by looking at the pre-Roman phases and by examining features which are essentially alien to the Mediterranean world. By the later Bronze Age, representations of birds, bulls and wheel-like symbols imply some kind of pre-Celtic European cult-expression, quite widespread, and probably connected with fertility, power and the sun. This symbolism continues into the Hallstatt period. When we come to the full Celtic La Tène, a new, specifically decorative Celtic art, based mainly on metalwork, appears at first glance to shut out man and to be essentially aniconic. Compared to the man-obsessed Mediterranean world this may be true, but we have seen that Iron Age representations of deities do occur in this period. Lucan speaks of wooden images of deities crudely hewn on tree-trunks. Wooden figures such as that from a wicker hut at Ballachulish, Argyl, which may be of first century BC date or earlier, may testify to the truth of Lucan’s observation, though the dating of isolated images is notoriously problematical. Whilst stone Iron Age god-images are rare, we have seen that in Germany and southern Gaul they are not unknown. Figural depictions in bronze occasionally appear as early as the Hallstatt period, exemplified by the seventh–sixth century cult-vehicle from Strettweg in Austria where a wheeled platform bears a central female figure carrying a bowl above her head and surrounded by horsemen and foot-soldiers with spears and shields. But human figures per se are rare among the free Celts and here the remark of Diodorus Siculus concerning the Gaulish King Brennus in the early fourth century BC may be relevant ‘Brennus the king of the Gauls, on entering a temple (at Delphi) found no dedications of gold or silver, and when he came only upon images of stone and wood, he laughed at them (the Greeks), to think that men, believing that gods have human form, should set up their images in wood and stone’ (XXII 9, 4). There is a higher incidence of animal-figures in the free Celtic phase, especially on the Continent. In the Hallstatt metalwork repertoire, the Urnfield water-bird and associated sun-sign still survives, and horses and cattle are common motifs. In the La Tène, boar images were fairly common and in the first centuries BC and AD blacksmiths made fire-dogs with bull-head terminals, probably possessing at least some symbolic or quasi-symbolic significance.
If we look at features during the Romano-Celtic period which are alien to the Graeco-Roman world, we can see that whilst the Roman stimulus in method of cult-expression is obvious and influential, the very nature of gods represented is often entirely new. Mother-goddesses, the horse-goddess Epona, the conquering horseman, are but few examples of a new and varied repertoire. Of particular interest is that during the Romano-Celtic period, we see coming to representative fruition ideas whose roots may be traced, sometimes faintly, during the Iron Age. The wheel-god (Chapter Two) existing in full human form by the second century AD, was worshipped by the wheel-symbol alone long before; indeed on the first century BC Gundestrup Cauldron, he may appear for the first time as a human image. Horned gods, ubiquitous during the Roman period, though not Roman, may be portrayed much earlier, at Holzerlingen. The animal-emphasis, absent in that fundamental form in the Mediterranean world, occurs from the Urnfield period around 1200 BC right through to the Roman phase. Portrayals of gods as heads alone may be traced in both periods, and there are other examples of recurring themes which straddle later prehistory and the period of Roman influence on Celtic lands. Whether or not the Celts used any kind of uniform system of religious symbols the continuity is there, sometimes overt, more often implied. The Celtic reluctance to construct images of deities should not blind us to the existence of symbolism, and the Roman stimulus in cult-expression, whilst sometimes confusing, serves not only to enhance the impoverished record of the Iron Age but also to display the fertile religious conceptuality already in existence in the free Celtic phase.
The influence of Roman religious stimuli on the Celtic world took the form both of physical expression, iconography and epigraphy, and of thought-processes applied to the rigidly functional and universal character of Roman gods as exported to Celtic lands. The interaction between the shadowy, multi-functional and more localised gods of the Celts and the more formal Roman pantheon produced a hybrid religious culture which is as fascinating as it is full of problems of interpretation. The normal Roman representative guise (the naturalistic human form) may conceal a deity who is completely alien and Celtic in character. Epigraphy is equally confusing. Bober puts the problem succinctly in her comment that, on the one hand, the vacuum of Celtic aniconism produced in the Roman period cult-art which was heavily biased towards the classical (the same being true of epigraphy). On the other hand, it appears that the Romans naively assumed that Celtic gods were Roman ones. It is worthwhile to look at a few examples of this hybrid and ambiguous cult-expression. The Roman Mars was popular in Gaul and Britain, but he is known by numerous different Celtic surnames or epithets, so we do not simply have equation of a Roman deity with a Celtic god assumed similar enough to be twinned with a suitable Roman counterpart. A dedication may be to Mars, but sometimes equated with a native healer-god – like Lenus Mars at Trier; or the dedication might be to Mars, but the accompanying representation might depict a peaceful, rural divinity with cornucopiae, as at Bisley in Gloucestershire. The implication is that union between Roman gods, defined strictly by function, and the ill-defined local spirits of the Celts is an uneasy and muddling one, though such equation occurred spontaneously, for whatever reason. In the case of Mars or Mercury (where twinning most frequently occurs) the process may well have happened because Mars had an Italian aspect of agricultural/protective divinity, in addition to his war-role, which appealed to the Celts with their emphasis on territory. Likewise, the Roman fertility/prosperity facet to Mercury’s cult may have struck a chord in Celtic belief. The occasional confusion between Mars and Mercury noted in the Celtic world – as on a Uley curse-tablet – may be explained in terms of the prosperity-interpretation of both. This may also account for the apparent contradiction in the Lucan commentaries where Esus and Teutates are each linked in turn with Mars and Mercury. What is very interesting is the Celtic fondness for divine couples, alien to the Mediterranean world. Mars Loucetius (‘Brilliant’) has a wholly Celtic consort Nemetona at Bath and at Mainz. Mercury possesses a Celtic spouse, Rosmerta ‘The Good Purveyor’), who enhances Mercury’s Celtic fertility-function. It is interesting that where couples are present, the goddess is the more likely to retain a Celtic name, whilst the male deity adopts a Roman title. This may reflect the essentially territorial role of Celtic goddesses.
Very frequently, a Roman art-form is borrowed to portray deities who are totally Celtic in character and do not appear in the Mediterranean world. The form may owe much to Graeco-Roman mimesis (see Chapter Seven) but there Roman influence ends. Examples are the triple mothers, Epona, Sequana of the Seine, Artio of Berne and Arduinna of the Ardennes. The evidence, as I see it, is not that the Romans transformed Celtic gods into civilised Roman ones, nor of suppression by the conquering power. Rather there is indication of active encouragement of such Celtic divinities as Nemausus of Nîmes and Sulis of Bath whose pre-existing spring-sanctuaries were positively embellished by the Romans. Indeed the very alien nature of some Celtic iconography should not be forgotten (Chaper Seven). Many North British gods owe little to classical realism, and the three stylised mother-figures from Bath would have been curious cult-objects for a romanised Briton. Finally, in assessing romanisation, we should remember that many lower class rural Celts, at any rate in North and West Britain, were probably not Latin speakers, particularly outside military areas, and one would expect their cults and beliefs to have been little altered by the presence of Rome.