The archaeology of religious practice at the time of the Druids - Druids: A Very Short Introduction - Barry W. Cunliffe

Druids: A Very Short Introduction - Barry W. Cunliffe (2010)

Chapter 3. The archaeology of religious practice at the time of the Druids

Evidence for religious belief and practice in the second half of the 1st millennium - the time when the Druids are known to have been practising - is both extensive and varied. In this chapter, we will examine something of the range of the data available for study, but we will resist the temptation to interpret it in the light of what the Classical sources tell us: the archaeological evidence must, at least for a while, be allowed to speak for itself. Its overriding message is that ritual behaviour pervaded every aspect of life.

The disposal of the dead continued to feature large, with a bewildering variety of practices varying quite markedly from place to place and time to time. At the elite level, burials were usually accompanied by a range of grave goods related to the status of the individual. In eastern Gaul and southern Germany, most elaborate burials were placed in wood-lined chambers set into the ground and were provided with a wide range of grave goods. The female burial found at Vix, in eastern France, was laid on the body of the four-wheeled vehicle that had carried her to the grave and was accompanied by a complete set of wine-drinking equipment - bronze, silver, and pottery - all imported from the Mediterranean at the end of the 6th century. A decade or so earlier, at Hochdorf near Stuttgart, a male aristocrat had been laid on a bronze couch next to his funerary cart with his bow, quiver of arrows, drinking horns, and Greek bronze cauldron close at hand. Some of his equipment was covered in gold sheeting manufactured at the time of the burial ceremony. The tradition of elite burial accompanied by vehicles, weapons, and feasting gear continued in the Marne-Moselle region into early La Tène times (5th century BC) and recurs elsewhere, in the Ardennes and Yorkshire, into the 3rd and 2nd centuries.

Individuals afforded elite burial were usually inhumed, and in Britain this tradition of inhumation continued, at least for a sector of the population, up to the time of the Roman invasion, but other burial rituals were also widely practised. Perhaps the most widespread was excarnation - the exposure of the body above ground. The principal evidence for this is the general absence of any other burial mode, the occurrence of body parts in domestic contexts, and the occasional burial of tightly wrapped bodies. The rite de passage, which this range of behaviour reflects, suggests that the first stage of disposal involved the exposure of the body, perhaps bound and wrapped in cloth, in a designated area set aside for excarnation. After a period of time, the bodies were removed, either for burial in the ground, or were brought back into the settlement where, as revered ancestors, they could be reincorporated into the existence of the living. Evidence for reburial has been found at Suddern Farm in Hampshire, where a number of tightly wrapped bundles of articulated bones were interred in a small cemetery close to the settlement. Evidence in support of reincorporation comes from the large number of disarticulated bones which are found on settlement sites in southeastern Britain and in Gaul.


1. Reconstruction of the burial chamber of a chieftain buried with his finery beneath a barrow at Hochdorf, near Stuttgart, Germany in the late 6th century BC

Cremation was also quite widely practised, particularly after the end of the 2nd century BC in northern Gaul and south-eastern Britain. The cremated remains were usually placed in urns, sometimes accompanied by other pottery vessels, presumably containing food or drink, and occasionally by items of metal wine-drinking equipment, perhaps representing the elevated status of the deceased. In the 1st century BC and early 1st century AD, more elaborate cremation burials are found in large chambers where the ashes are accompanied by offerings of food, amphorae of wine and the ceramic and metal vessels needed in their consumption, and other items of personal equipment. Superficially these elite cremations look quite simple, but something of the potential complexities of the rituals that may be involved is well demonstrated by an elite burial of the 1st century AD found at Folly Lane just outside the Roman town of Verulamium (St Albans). Here a large chamber was dug and the corpse, together with a range of grave goods, were laid out, no doubt with the intention that they should be viewed by the mourners. Then, after an appropriate period of time, the body was removed and cremated, and the grave goods were smashed, following which the grave pit was filled in. This series of events clearly represented an orderly process of departure, from the moment of the individual’s last breath to the completion of the rite and the closure of the tomb.

Although burial rites were varied both regionally and chronologically, the care with which the dead were put to rest is readily apparent. The provision of grave goods is a constant theme particularly among the elite. Taken all together, the evidence shows that normative death was a highly ritualized occasion which demanded adherence to ceremony. The process seems to have embedded within it the concept that the deceased passed on from life on earth to enjoy an afterlife appropriate to his or her status. The basic concepts in the late 1st millennium BC are not very different from those of the earlier prehistoric period.

The burial in the ground of a corpse, bundle of bones, or cremated ashes may well have been conceptualized within the broader context of consigning dedicatory deposits to the earth, thereby placing them in the realms of the chthonic deities. It may also be that cremated remains were thrown into rivers or lakes in a ritual parallel to the deposition of tools and weapons in watery places, but of this there is, unsurprisingly, no positive trace.

Deposition in water and earth - traditions deeply rooted in prehistory - continued throughout the 1st millennium. A surprisingly high percentage of the elite Iron Age metalwork found in Britain and Ireland - including swords, daggers, shields, helmets, and bowls - has come from rivers and bogs, one of the most prolific locations being the middle reaches of the River Thames. The tradition, which began in the middle of the 2nd millennium BC, reached a crescendo in the century before the Roman invasion. Spectacular pieces of craftsmanship like the Waterloo Bridge Helmet and the Battersea Shield demonstrate the extremes to which people would go to assuage the demands of the gods. Items like these, readily identifiable and no doubt redolent with history, must have been of enormous social value. The ending of their earthly life through acts of deposition would have been an occurrence of great moment in the history of the community.

That certain stretches of river seem to produce more artefacts than others suggests that there may have been specific locations from which the depositions were made. This is most clearly seen in the case of the lake (now a bog) of Llyn Cerrig Bach in Anglesey, from which a large hoard of Late Iron Age metalwork was recovered. The location of the finds suggests that they were thrown in from a rock platform on the lake edge. There are also examples of timber walkways built out into the water from which items were thrown. The most famous is at La Tène on Lake Neuchâtel in Switzerland from around which a large number of artefacts were recovered. Other walkways have been identified at Flag Fen near Peterborough and at Fiskerton in the River Witham. The Flag Fen platform was in use throughout the 1st millennium BC up to the Roman period. Fiskerton, which belongs to the 5th and 4th centuries, is particularly interesting in that dendrochronological studies have shown that it was repaired at regular intervals on a periodicity of 16-18 years, which suggests regular renewal perhaps related to the 19-year lunar cycle.

Depositions in the earth continued to be made throughout the 1st millennium to the time of the Roman conquest. They take many forms. The most spectacular are the hoards of torcs and other items made in precious metals, gold, silver, and electrum (an alloy of gold and silver) which are found in some number in East Anglia. The most extensively studied is the site of Snettisham in Norfolk, where a number of discrete hoards have been found comprising torcs, coins, and scrap items buried in small pits dug specifically for the purpose. All the hoards lie within an 8-hectare enclosure, which appears to have been of Roman date - some decades after the deposition of the hoards. One possibility is that it was dug to define the boundary of the territory known, in the local memory, to have been sacred to the gods.

Less spectacular, but no less dramatic, are depositions found in pits originally dug for storage purposes. Storage pits, or silos, are a feature of late 1st-millennium settlements throughout western Europe occurring particularly densely in south-eastern Britain and the adjacent regions of France and the Low Countries wherever soil conditions are such that stable pits can be dug. It is probable that the majority of the pits were used as underground silos for the storage of seed grain. Once the useful life of the pits was over, they were abandoned and in many cases given over to a secondary use as a repository for votive deposition comprising groups of artefacts, dumps of grain, joints of meat, animal carcasses in whole or in part, and human remains. These are the tangible items likely to survive in the archaeological record: it is quite conceivable that other deposits were also made, including furs, fleeces, bales of wool, fabrics, cheeses, and suchlike, all of which would leave no archaeological trace in normal conditions. It was usual for the initial deposits to be made on the pit bottoms soon after abandonment. Thereafter, by processes of natural erosion, pits were allowed to silt up but quite often secondary and tertiary deposits were added as erosion proceeded.

While there is much variety to be observed in these processes of deposition, the overall intention is clear, but what does it mean? One plausible explanation is that the primary deposits reflect propitiatory offerings made to the chthonic deities for having safely protected the seed corn during the liminal period of winter, while the subsequent deposits may be offerings made in anticipation of a successful harvest or as thanks after the harvest has been safely gathered in. For an agrarian community whose very livelihood was based on the continued fertility and productivity of the crops, the agricultural year would have been punctuated with rituals designed to placate the deities. Indeed, the practice of storing seed grain in pits may have been conditioned by the belief that the safest place for the vital seeds during the liminal period was in the realms of the earth deities.

The discovery of human remains in pits raises the difficult question of human sacrifice. Isolated bones or even articulated body parts could be explained in terms of their being the remains of ancestors brought in from the excarnation grounds and deposited as valued offerings to the gods. But entire skeletons are not infrequently found - some tightly bound, others splayed in a variety of poses on the pit bottoms. The bound remains could be ancestor bundles, but the others look more sinister as though the bodies, dead or alive, were thrown unceremoniously into the voids. In several instances, large, heavy flint nodules were dumped above the skeletons. It is tempting to regard such deposits as the result of sacrifice though, of course, other explanations are possible.

That said, there can be little doubt about the fate of Lindow Man - a body found in a bog in Cheshire. He had been hit violently on the head, had been garrotted, and had had his throat cut before being placed in the bog. His ‘triple death’ looks very much like ritual killing and his resting place, in a bog, would have been appropriate for a sacrificial victim.

Among the human remains found in archaeological contexts, the head has often been selected for special treatment. This is vividly demonstrated at temple sites in southern France, like Roquepertuse, where human skulls were set in niches carved into stone pillars or nailed to wooden posts. Isolated skulls are also sometimes found in pits or in defensive ditches near entrances where they might have been displayed on gates. Clearly the head was perceived to be a special body part, perhaps one that contained the power of the deceased. The longevity of the belief is demonstrated by reverence for skulls at Neolithic sites like Hambledon Hill, where human skulls were found placed at intervals along the bottom of the enclosing ditch.

While ritual behaviour seems to have pervaded all aspects of daily life, and could have been practised anywhere in the landscape, there were also specific locations set aside to serve as temples or shrines, usually defined by ditched enclosures. In the south of France, these often incorporated stone-built architecture and were provided with stone statues of gods or heroes, but more often in the north of Gaul and in Britain they were timber-built.


2. The body of a man ritually killed and buried in a bog in the 1st century AD. He was found during peat cutting at Lindow, Cheshire, England

One of the best known of the northern French temples is the multiperiod structure excavated at Gournay-sur-Aronde (Oise). The sanctuary was situated on a spur overlooking a small stream: it was first built in the 4th century BC and rebuilt on a number of occasions thereafter until the 1st century, when it was destroyed by fire and the site levelled, but its memory remained, and in the 4th century AD a Gallo-Roman shrine was built on the same spot. The central focus of the sacred site was a large oval-shaped pit containing the remains of sacrificed cattle which was set within a rectangular ditched enclosure. Over the years, the pit was associated with, and later enclosed within, a succession of timber structures, while the outer enclosure was frequently refurbished and enhanced with timber palisades. The ditch was used throughout as a place for the deposition of sacrificed materials including over 2,000 weapons and large quantities of animal remains - the result of individual acts of sacrifice spanning the life of the temple.


3. Body, possibly of a sacrificial victim, laid out on the bottom of a pit in the hillfort at Danebury, Hampshire, England. The pit dates to the 3rd or 2nd century BC

Gournay-sur-Aronde is characteristic of the Iron Age temples found scattered across northern Gaul and Britain. For the most part, the temple buildings are small, usually square but sometimes circular, and are almost invariably set within an enclosure defining the sacred temenos. In a number of cases, the temples continued in use into the Roman period, at which time they were often rebuilt in stone.

Another type of defined religious location, found extensively in western-central Europe, are the rectangular enclosures known as viereckschanzen which may contain timber-built shrines, burials and shafts, or wells, in any combination. The well-examined example from Holzhausen in Bavaria contained a small timber ‘shrine’ set in one corner together with three shafts up to 40 metres deep, in one of which was an upright wooden stake associated with materials identified as the decayed remains of flesh or blood. Another example excavated at Fellbach-Schminden in Baden-Württemberg contained a number of burials and a single shaft some 20 metres deep which had been lined with timber and had probably served as a well providing water for ritual purposes. When it was finally abandoned, two pottery vessels were placed on the bottom. The top of the shaft had been ornamented with elaborate wooden carvings of animals, including rampant stags and goats, which survived largely intact in the waterlogged fillings.


4. Two of the many wooden statues placed as votive offerings at the shrine of Sequana, guarding the source of the Seine at St Germain sur Seine. They date from the early Roman period and were probably placed by worshippers to draw the deities’ attention to themselves and their particular ailments

The viereckschanzen differed from the temple sites but were evidently intricately bound up with ritual behaviour perhaps related to celebrating the memory of ancestors, if we assume the burials to represent those of a lineage. The shafts with which many were furnished, if deep enough, may have provided water for ritual purposes and are clearly in the same tradition as the much earlier (2nd-millennium) shaft found at Wilsford near Stonehenge. That some of the shafts contained offerings is again an indication of the religious nature of the sites. In the case of shafts used as wells, the deposition of offerings is most likely associated with rites of closure.

We have already mentioned the significance attached to watery places as locations for deposition. One particular type of place was the spring where water from the underworld rose to flow into the land of humans. Such places were revered and many were believed to have curative properties. One of the best known lies at the source of the Seine, some 35 kilometres north-west of Dijon, where, in the Roman period, there was a thriving shrine to Sequana, goddess of the river. Excavations in a waterlogged area brought to light a remarkable collection of wooden ex votos in the form of human or animal figurines which are thought to have been placed along a terrace wall facing the sacred area during the 1st century AD, when the old shrine was being renovated. The human figurines are a varied collection including complete figures, some wearing hooded cloaks, heads, trunks, limbs, hands, and feet. Some of the trunks are carved to give impressions of the internal organs, one of the most vivid being a representation of a rib cage with the lungs and trachea inside. There are also models of sexual organs, breasts, and eyes. The intention of the suppliants was evidently to provide the goddess, Sequana, with an unambiguous indication on which part of their anatomy she was to concentrate her curative or rejuvenating powers.

Another curative spring, at Chamalières, near Clermont-Ferrand, produced a comparable collection of wooden ex votos from early Roman contexts dating to the century or so following Caesar’s conquest of Gaul. They represented much the same range as those from Sequana’s spring, except that indications of illness and deformity are rare. However, the fact that Chamalières is a mineral water spring suggests that pilgrims were attracted there for its curative properties. The votive offerings from both springs were found in early Roman contexts. While it is possible that some of the objects survived from earlier periods, it is more likely that the practice of presenting representations to the deities is an attribute of Mediterranean culture. Even so, the veneration of the springs must go back to a more distant period.

The little wooden figurines presented to the deities are only one facet of a rich tradition of religious representational craftsmanship practised in western Europe. Wooden carvings are preserved only rarely, where suitable conditions of waterlogging allow, but given the ease of carving, one may assume that they were once widespread.

More common are sculptures in stone which reflect a varied range of traditions. One of the best-known assemblages comes from a temple site at Roquepertuse (Bouches-du-Rhône) in southern France which dates to the 3rd century BC. What survive are the stone elements of the stone and timber architecture of a temple comprising columns, with niches carved out to contain human skulls, and a horizontal frieze carved with horses’ heads and surmounted by a fearsome bird of prey poised nearby to swoop. Part of the ensemble included a carving of two conjoined heads looking in opposite directions, much like the Roman god Janus - the god of coming and going. The heads shared what is frequently referred to as a ‘leaf crown’ - a kind of bulbous extrusion - which frames the back of the head, common among other sculptures of this period. What it means is beyond recovery, but it would appear to reflect status in some way.


5. Portico of skulls from the sanctuary of Roquepertuse, near Aix-en-Province, France, dating to the 3rd to 2nd centuries BC

Associated with the temple were a series of large figures sitting cross-legged on the ground. These are thought to be representations of heroes, either real people or mythical beings.

Roquepertuse is not alone in southern Gaul. A similar sanctuary occurred within the oppidum of Entremont, near Aix-en-Provence, and is represented now by carved stone elements, including a pillar with representations of severed heads along its length and another piece with a realistic depiction of a severed head adjacent to niches in which real severed heads would have been placed. There were also representations of seated heroes, some holding severed heads. The emphasis on the severed head is notably prevalent in southern Gaul and may reflect a belief system particular to the area of the Celto-Ligurian tribes. It is worth remembering that it was this region of Gaul with which writers like Posidonius would have been familiar - their generalizations about the Celts may therefore have been biased to the areas they knew.

Religious sculpture is known in other parts of Europe, most particularly the region that is now southern Germany. The most spectacular of the discoveries is the life-sized sculpture in the round of a bearded man carrying a shield, found in association with the 5th-century burial at Glauberg (Hesse). The figure is shown with his head framed in an embracing ‘leaf crown’ comparable to that of the ‘Janus’ heads of Roquepertuse. The interpretation of the figure is a matter of debate. It is likely to have stood on or close to the burial mound covering a princely burial and may therefore represent the deified hero lying beneath. Statues adorned with leaf crowns have been found at other sites, including Heidelberg and Holzgerlingen, and bearded heads with leaf crowns carved in relief adorn the four sides of a highly decorated pillar found at Pfalzfeld. This recurring motif clearly had a great significance reflecting exalted - perhaps even god-like - status. The presence of these statues in the landscape would have been a reminder to all of the ever-present supernatural powers.

The leaf-crowned ‘heroes’ are a manifestation of the 5th to 3rd centuries. Later religious sculpture of the 2nd to 1st centuries BC tends to be more realistic and more specific. Three examples from different parts of France will illustrate the point. All three represent male figures and all wear the neck torc as a sign of high status but display other characteristics. The stone figurine from Paule (Côtes d’Armor) holds a lyre, that from Euffigneix (Haute-Marne) is embellished with a lively rendition of a powerful boar, while the seated bronze statue from Bouray (Seine-et-Oise) has hoofed feet like a deer. It is tempting to see them each as a god identifiable to the initiated by their specific attributes. Perhaps here we are seeing the local tribal deities made manifest.

The sculptures are but one aspect of a considerable corpus of iconographic material which is generally referred to as ‘Celtic art’, best known through the medium of decorative metalwork but also represented in designs on pottery and on wooden vessels. The ‘art’ originates in the 5th century in the aristocratic households of the Marne-Moselle region. It was inspired by Greco-Etruscan motifs introduced on imported metal vessels, and quickly developed into a highly original style of curvilinear design - energetic and surprising. Within the art lay embedded references and meanings impossible now to interpret, but understandable at the time and used in such a way as to communicate meaning. To take just one example - the bronze shield boss dating to the 2nd or 1st century BC, dredged from the River Thames at Wandsworth. The boss would have occupied the centre of a large rectangular shield probably of wood or leather. It was circular, with a central protuberance around which was a flowing pattern of tendrils created in repoussé with infillings by engraving. A careful look at the pattern shows that it is composed of two heavily stylized birds with outstretched wings. A warrior facing the shield held by an opponent might suddenly have seen the curvilinear motifs shift into focus and the two great birds of prey emerge from the boss. He would have understood the message. In a story told by Livy of the Celtic invasion of Italy, there is a vignette of a combat between two Celts. As they approach, a huge bird of prey lands on the helmet of one of the contestants. The other, seeing it, is paralysed by fear, knowing that the gods are on the side of his opponent and are about to devour him. Perhaps the bird of prey on the temple architrave at Roquepertuse had a similar effect on those who approached the temple.

The story shows the power of the message - and Celtic art carried those messages. It was enigmatic and constantly shifting: images could appear and disappear. Nothing was ever as it seemed.

No convincing narrative art of the period has survived, with the possible exception of the famous silver Gundestrup cauldron found in a bog near Roevemosen in Denmark. The cauldron is a puzzling piece. Stylistically, it shares all the characteristics of Thracian workmanship, and yet it depicts material culture that is Celtic - war trumpets, rectangular shields, animal-crested helmets, and torcs. One plausible explanation is that it was made in Bulgaria by Thracian craftsmen as a gift for a chief of one of the Celtic tribes living nearby, by some entirely indecipherable process ending up in the Danish bog as an offering to the local gods. The fascination of the cauldron lies in the complex scenes it depicts on its two sets of repoussé decorated plaques, one set facing outwards, the other facing inwards. One of the inner scenes centres on a seated figure, evidently a god, who wears a torc and an antler head-dress and holds a torc in one hand and a serpent in the other while animals gather around. Another of the inner scenes shows a troop of warriors, some mounted, others on foot, approaching a large figure who is busy depositing a smaller figure head-first into a large container while an excited dog scampers around. Tempting though it is to offer explanations - and many have been offered - there is little that can be said with any degree of certainty beyond simple description. Nor can we be sure that it is ‘Celtic’ traditions that are being illustrated rather than those of the Thracian world. But that said, the scenes evidently involve ritual behaviour and remind us of the complexity of the belief systems seldom accessible through archaeological evidence alone.


6. Scene from the silver-gilt cauldron of the 2nd or 1st century BC found in a bog at Gundestrup, Denmark. The large figure appears to be depositing a sacrifice in a cauldron

We saw in the last chapter how, through accurate experiment over a period of time, the rhythms of the lunar and solar cycles had been charted and monuments constructed to ‘capture’ major celestial events. All this had been achieved by the beginning of the 3rd millennium: thereafter, the knowledge would have been passed from generation to generation through oral learning. The extent of that knowledge at the end of the 1st millennium BC is vividly demonstrated by the surviving pieces of a large bronze calendar found at Coligny (Ain) in eastern France. The calendar, thought to have dated to the late 2nd or early 1st century BC, had been broken up and buried in a Gallo-Roman temple. It records 62 months of a 5-year cycle displayed in 16 columns each of 4 consecutive months, except for columns 1 and 9, each of which comprises 2 normal months and 1 intercalary month. Each month was of 29 or 30 days, and each was identified with the name followed by MAT(U) for the 30-day months and ANM(ATU) for the 29-day months. Since matu means ‘complete’, while anmatu means ‘incomplete’, the word presumably relates to the length of each month, but the words can also mean ‘good’ and ‘bad’, and might therefore be an indicator of whether the months are propitious or not. Each month is divided into two parts after the 15th day with the word ATENOUX, which signifies the end of the light half and the beginning of the dark period. Within each month, certain days are labelled as ‘inharmonious’.

Coligny is a lunar calendar in that the passage of time is measured by nights rather than days. It was probably used to cover a 30-lunar-year cycle with the beginning of each year starting with the 6th lunar month. By manipulating the intercalary months and other adjustments, the lunar calendar could be made to coincide with the solar calendar.


7. Detail of the fragmentary bronze calendar found at Coligny, Ain, France. The calendar, probably from the 1st century AD, lists 62 months of 29 or 30 days each and indicates seasonal celebrations and propitious times

The calendar is, by any standards, a remarkable achievement representing the culmination of study going back over very many generations. Its very complexity demanded that it be committed to a more permanent form using the Latin alphabet, but this need not imply any significant input from the Mediterranean world. There is no reason to suppose other than that the calendar was the product of the indigenous inhabitants of Atlantic Europe.

We have laid out, albeit in a very summary form, a range of archaeological data relevant to the intellectual life and belief systems of the inhabitants of Gaul and Britain in the five centuries or so before the Roman invasion. The data are much richer and more varied than those of the preceding prehistoric period, and yet there are many themes in common: the importance of offerings placed in the ground or in watery contexts; the digging of deep shafts reaching towards the underworld; complex burial rites involving grave goods and the consigning of the dead to the sky and the earth; the significance of the human skull; and the careful measurement of lunar and solar time to chart the passage of the seasons and to programme ceremony. There can be little doubt that the belief systems evident in the last four centuries or so of the 1st millennium BC - the time of the historic Druids - were the result of a longue duréeof development and refinement spanning several millennia. The druidic class, then, were the inheritors of ancient wisdoms.