Celtic Sacrifice: Pre-Christian Ritual and Religion

By Marion Pearce

Welcome to the world of the Celts, a world of magic, shape-shifting, sorcery and divination, a world where legend and fact are mixed, where the physical world and the Otherworld blend in heroism, rituals and rites.

Combining sources from mythology and archaeology with eye-witness accounts from the period, this book presents a fascinating p

• Language: English
• Publisher: Fenix Flames Publishing Ltd
• Release date: Dec 11, 2022
• ISBN: 9781913768270

Marion Pearce

­­Marion Pearce was the publisher of Pentacle magazine, the largest independent Pagan Magazine in the UK with a large circulation in the the US also. She is also a previous editor of Pagan Dawn, the magazine of the Pagan Federation. She is author of "Gods of the Vikings" (Avalonia) and author of "The Roman Calendar", and "Celts: Masters of Fire" both of which are from the republishing series by Fenix Flames.

Book preview

WORSHIP

Until recently the ancient Celts were thought to have only worshipped in the open air in groves. In fact as recently as 1966 the archaeologist Lewis was giving the traditional view of Celtic worship:

The Celts worshipped the ubiquitous invisible forces of mature in the open air, venerating such objects as trees and standing stones, mountains and streams, not as divine in themselves but as the abode of deities. They neither needed nor desired houses for their gods or images of them, apart from occasional carvings on menhirs (standing stones). The two concepts of temple and image do not seem to appear in Celtic sanctuaries until Greek, and later Roman, religious practice came close enough to exert noticeable influence.

Then it began to be realised that Celtic worship might be a little bit more complicated than first thought. Already the discovery of an early Celtic Iron Age religious building at Heathrow had knocked holes in that theory. In fact there are now twenty four known religious buildings on sixteen various sites from this period in history. Although I must say here that there were open spaces in these buildings, likely to have been used for religious ceremonies in the open air.

Nowadays it is thought that they had wider forms and ways of paying homage to their gods and goddesses. Here I quote Anne Ross in The Pagan Celts:

The evidence, then serves to indicate that, far from worshipping their gods in groves and other features of the countryside alone, the Celts did in fact have a variety of built structures of some sort, in which their rites were performed; and there is no doubt that future archaeological researches on the Continent and in the British Isles will bring many more to light. There is also quite a body of evidence for wooden temples, constructed within these banked, earthen enclosures.

Certainly the early Celts would have worshipped in groves or nemeton. Gradually these would have developed into more permanent structures, although still retaining a sacred space open to the elements in which to conduct rituals, with perhaps trees, altars and icons. Perhaps there were relatively crude wooden buildings. In 40% of these early Iron Age shrines the walls were constructed by a system of wooden posts. The spaces between these supports would have been filled with any available material. Over time these structures could have developed into stone buildings after the Roman conquest.

It is difficult to say what these early Celtic shrines were like. Interestingly 95% of these shrines had their doors facing either directly east or with some variation to north east and south east. In fact the only exception to this is the one at Frilford. This would have enabled the Celts to perform rituals facing the rising sun and be protected from the weather in the shrine. Even today in Christian churches the altar is traditionally placed east. Islamic mosques also face east.

As to the question of whether shrines were roofed or open to the elements, Professor Wain in Ritual and Religion in Iron Age Britain claims that all the present shrines found with the exception again of the one at Frilford, could have been easily roofed. If roofed they would have been thatched. Of course this material is easily decomposed and not much thatch would be left for todays archaeologists to study. It is likely that the wall would have been made from wickerwork. In fact traces of wickerwork have been found with a wooden idol unearthed at Ballachulish, Argyllshire, which supports this theory. Anne Ross discusses a small stone shrine still in existence in Glenlyon, Perthshire. This shrine has its roof rethatched biannually, a custom I would imagine dating back many years.

This theory of thatched roofs is backed by classical sources. Strabo speaks of a tribe of Celtic women on an island in the Loire. These women, known as Bacchantes as they followed the god Bacchus, lived in a female only community. One of their rituals was the annual unroofing and then re-roofing of their temple. Again this echoes the ritual at Glenlyon. In Irish mythology in the tale, Echtrae Cormaic, roofs are mentioned, although this time made of feathers.

It is difficult to come to any agreement as to the architecture of Romano-Celtic temples. It is generally assumed that the majority were roofed buildings with tall tower-like cella. A cella was the inner sanctum of the temple, a room situated in a holy part of the building where the sacred memorial idol was situated. There would also have been a roofed ambulatory, which would have been a passage running around one or more sides of the temple. Whether this ambulatory had solid walls or evenly spaced columns placed to let in the daylight is debated. There have been temples found which have both walls closed to the elements and some where the walls consist of columns. There is this variety in cella roofs. Some would have been open to the sky, others such as the temple at Great Chesterford were adorned with elaborate mosaics which have never been exposed to the elements. Most reconstructions of Romano-Celtic temples are roofed.

Now what were these roofs like? It is difficult to say after all these years. Excavations of a temple at Chelmsford have shown that there were pairs of buttresses equally spaced around each of the eight sides of the ambulatory. These buttresses do not appear to have been used to strengthen the ambulatory walls but were placed exactly opposite the cella wall. This leads to the thought that they were positioned to carry vaulting, the arched structure that forms a roof, across the width of the ambulatory. This would have resulted in 16 vaulting ribs coming in pairs from splayed corbels (supporting brackets) backed with buttresses. The ribs were thought to be made from either wedge shaped stone or rings of Roman brick, with the intermediate spaces filled with tufa, a soft sandy stone deposited by springs and streams.

Now what sort of vaulting would have been used to construct roofs? Plain barrel vaulting is the simplest type of vaulting, this is semi-cylindrical. But at some temples such as the site at Pagans Hill, the design of the temple would have made this type of vaulting as a roof very unlikely. Here it is felt that the roof would have been a more advanced structure. It is thought that a cross-vault was used. This consisted of an intersection of two or more vaults and would have had a peak. In the late Roman period this was a common type of roof on such buildings as baths, temples and churches. Pagans Hill is thought to have been designed by a Roman architect who was familiar with the more complex continental buildings. Certainly its roof would have been very advanced for this period in British architecture. In fact it is considered that this building in the Mendips, Somerset, would have been closer to buildings in Rome than most other areas of rural Britain.

It is interesting that in Celtic religion worship was considered to be open to the elements. This causes a problem when you come to closed Romano-Celtic temples. We have seen how they were roofed, the surrounding ambulatory walls being either closed or colonnaded. Once in the ambulatory there would have been unrestricted access to the cella. Due to the Celts’ practice of worshipping and carrying out their rituals under the sky it is unlikely that the outer doors would have ever been closed. This of course leaves a security risk. Valuable and sacred items were inside, and they had to be protected from enemy tribes. The archaeologist, Professor Warwick Rodwell, suggests a solution: The main entrance could have had a gate or grille which would protect the temple but still leave it open, even when the grille was closed.

On a stele (stone column) in Titelberg there is a carved relief showing a Romano-Celtic temple. This is an unusual temple and does not conform to the normal archaeology of the period. Professor Rodwell, however, says that this temple is very similar to the temple in Uley. That is interesting in itself as Uley is situated on an Iron Age Celtic ritual complex. In fact it is thought that Uley was originally a sacred grove site. The early Celtic shrine was replaced, probably in the early second century ad, by a square stone Romano-Celtic temple. There were also later additions to the building.

There is no front ambulatory, just those on the sides in the Titelberg relief. This corresponds with the earlier temple site at Uley. It is these ambulatories that are interesting. They had gabled roofs, in other words they peaked at the centre of the roof. Most roofs at this period were flatter. The centre cella is also raised and peaked. Now this is important as it would enable rainwater to be caught. Professor Rodwell beieves there would have been a gutter running between the inner slope of the ambulatory roof and upper cella wall. This gutter would have caught the rainwater which would then have flowed into two waiting tanks in the front of the building where there was the missing ambulatory. This would have been revolutionary. The Celts venerated water, there were sacred sites in many rivers, springs and wells. Rainwater, falling from the sky, would have been considered divine. A method of collecting it at a holy temple site would have been very important. This could have been a site of major devout religious importance.

Rodwell also considers the similarities between Titelberg and Uley striking. He believes the ambulatory of Uley to be gabled and therefore the roof to be ridged, allowing rainwater to be collected in gutters. In fact Uley’s cella has a large hole which strongly suggests there was once a receptacle of 2m³ volume, which the professor thinks maky have been a lead tank holding sacred rainwater.

Rodwell has an interesting definition of the various Roman religious sites that would have been in operation during this period. Now many Roman religious sites were situated on the earlier Celtic ones. I have added my own observations on the examples Rodwell has quoted. He has defined twelve distinct types of religious structure:

Private household shrine, or lararium

A single room, a cellar or just an alcove. For example the corridor shrine at Verulanium, the deep room at Lullingstone villa, and perhaps the cellar in the villa podium at Rivenhall.

Verulanium is the Roman name for the town of St Albans, Hertfordshire. This was a centre for the Celtic worship of Cernunnos, the horned god. A mosaic pavement has been found in the town with a bearded and antlered head on it. At first the head was thought to have had lobster claws identifying it as Oceanus, but now it is considered to be an antlered head which of course distinguishes it as Cernunnos. Interestingly small bronze serpents have also been found at a temple in the town. Serpents were often associated with Cernunnos.

House temple (or church)

A purpose-built room or suite of rooms, usually lending a particular architectural expression to the building. Presumably for family use only. For example the square temple-like structure in the north-east corner of the villa at Lullingstone, and the octagon at Great Witcombe villa.

Proprietary shrine, temple or church

A distinct and usually detached (apart from boundary walls) building in a villa or farmyard complex, with access from the yard. Presumably used communally by the resident family and retainers, as at the Chedworth nymphaeum (a well-shrine), the polygonal buildings at Stroud and Petersfield villas and the circular temple at Lullingstone.

Chedworth in Gloucestershire was a centre for Celtic worship, an altar has been found there depicting Olloudius. Olloudius, whose name comes from ollo-vidios (Great Tree) was a war god. Another altar was found at Chedworth, again with a god depicted on it, this deity is armed with an axe and spear and is undoubtedly another war god. Interestingly Anne Ross considers both the altars could be presumed to have come from the context of the prosperous villa and surrounding agricultural region. Of course well shrines such as that at Chedworth, were a common Celtic religious site.

Estate temple

Located a little distance from a villa but presumably still on its estate or perhaps on the boundary. This type must have been for communal use and may also have been proprietary, perhaps being shared by the retainers of more than one estate or a group of farmsteads. For instance the Romano-Celtic temples at Chedworth and Titsey. Many of the small, seemingly isolated, temples probably belong to this group, including Lumyatt Beacon and Great Chesterford.

Local cult centre

A rural or semi-rural temple complex where there may be particular evidence for the prominence of one deity (such as Mercury at Uley) and where there were ancillary buildings such as a guest house or baths. Such sites will vary in size, importance and in the provision of facilities. At the top of the scale would come Lydney, and lower down Pagans Hill, Coleshill, Harlow and perhaps also temples with large temenoi (sacred buildings) such as Colchester, Farley and Woodeaton. One wonders whether the function of this group was to provide communal places for religious activities and fairs for a sector of a tribal area, for a pagus (latin for a village or province). Prof Rivet has suggested that we may be able to identify pagus-deities such as Abandinus at Godmanchester and Viridios at Ancaster. Both of these were located in small towns but could still have served an essentially rural population. Springhead is a similar case in point.

Lydney Park, Gloucestershire is an ancient Celtic site. The temple stands on the River Severn and is dedicated to the Celtic god, Nodons. It is likely that this god and the association with water would have made it a healing centre.

Major rural sanctuary

A very extensive religious site which could have served for tribal gatherings; sometimes there was a theatre on the site. These centres are well known on the Continent, but in Britain only Gospecks falls into this category for certain. Gospecks is a Romano-British temple site near Colchester in Essex. It is thought to have also been a major trading centre. Sapperton, Gloucestershire, is a potential candidate for this type of structure.

Natural shrine, such as a grove, a spring or a river

The essence of this type is that it has remained basically in its natural state with-out the addition of religious buildings. Many of the holy wells, so prolific in western Britain, should be included here. The fact that these were shrines in antiquity (and some are to this day) is shown by the ex votos which are found in and around them. Many of the sites in this category were upgraded in the Roman period by the erection of buildings to house and service the presiding deities. Holy wells were well known Celtic religious sites. The Romans would have used the existing Celtic holy