Archaeology Magazine Archive

A publication of the Archaeological Institute of America

Special Introductory Offer!
Celtic Sacrifice Volume 55 Number 1, January/February 2002
by Jeremiah R. Dandoy, Page Selinsky, and Mary M. Voigt

Grim deposits of butchered bones attest ritual slaughter by Galatians at Gordion.

Following his death, Alexander's empire broke up into smaller, competing states whose rulers sometimes hired mercenaries to supplement their own armies. In 278 B.C., King Nicomedes I of Bithynia welcomed as allies 20,000 European Celts, veterans who had successfully invaded Macedonia two years earlier. These warriors, who called themselves the Galatai, marched into northwestern Anatolia with 2,000 baggage wagons and 10,000 noncombatants: provisioners and merchants as well as wives and children. Ancient texts tell us that some of these immigrants settled at Gordion, the old Phrygian capital of King Midas, about 60 miles southwest of modern Ankara. Exactly when Galatians took over the town is unknown, but archaeological evidence suggests they were there soon after 270 B.C., the time when documentary sources tell us that Celts began raiding in central Anatolia.


A young woman's skull and neck vertebrae are juxtaposed with an older woman's upper back vertebrae, which lie above an animal's jaw. The younger woman's hip bones lie atop the older woman's ribs, also shown in photos 8 and 9 further down the page. (Mary M. Voigt/Gordion Project)


This skull of a teenager (12-17 years old) was carefully placed alongside the skull, pelvis, and upper leg bone of a dog. Beneath these was the pelvic bone of another person. (Mary M. Voigt/Gordion Project)

Earlier excavations at Gordion recovered coins of the sort used to pay Celtic mercenaries, a few artifacts with parallels in Celtic Europe (a helmet flap, sheep shears, and pin), and a sherd inscribed with a clearly Celtic name, Kant[x]uix. Nevertheless, the archaeological evidence for a large Galatian presence at the site was not overwhelming until our discovery of grisly evidence of rituals involving humans. The broken-necked bodies and decapitated heads at Gordion cannot be attributed to any local Anatolian group, but are characteristic of European Celts.

Between 1950 and 1973, Rodney S. Young and G. Roger Edwards, archaeologists from the University of Pennsylvania Museum, excavated much of the eastern part of Gordion's citadel mound. Young, who was mostly interested in a burned level that he identified with King Midas (see "Celebrating Midas," July/August 2001), was unenthusiastic about later occupation levels. In 1964, influenced by the prevailing views of historians, he characterized the Galatian settlement as follows: "The buildings of [these] levels are mostly the light structures of a farming village with their outbuildings and byres, usually in poor preservation." By 1990, the picture had changed. In reassessing what was known of Galatian Gordion, Keith DeVries--Young's successor as director of the Gordion Project--pointed out that some houses contained evidence of considerable wealth, including gold coins and stone sculptures. Moreover, some of these people could at least read Greek, the language used to inscribe some of their possessions, and favored items made in Greek style. Roman sources, DeVries noted, referred to Celts in Anatolia as "Gallograeci" because of their adoption of Greek ways. DeVries' interpretation is more in line with that of Livy, who described Galatian Gordion at the beginning of the second century in his History of Rome as an emporium, or trading center, and an oppidum, or fortified settlement.


Gordion excavation areas (Sondra Jarvis and Carrie Alblinger/Gordion Project)

The University Museum began new excavations at Gordion in the late 1980s. Investigation of the northwestern part of the citadel mound, carried out in cooperation with the Royal Ontario Museum, was the key to a new understanding of the Hellenistic period. First of all, we now know that the Galatian occupation lasted at least 100 years and was marked by two destructions or abandonments followed by a brief third occupation. Second, our soundings show that the entire top of the citadel mound was occupied in the third and second centuries B.C., confirming Livy's description of the place as a substantial town. Third, the arrival of the Celts at Gordion is marked by a change in the use of space within the settlement as well as in architectural forms and materials: above ordinary mud-brick houses, the Galatians constructed a monumental public building of cut-stone blocks that was surounded by a massive stone wall.


Double-headed figure from Gordion resembles more sophisticated ones from Europe. (Mary M. Voigt/Gordion Project)

Adjacent to the building and within the stone wall, we found a potter's house and workshop dating to the site's initial Galatian settlement. Inside this building, constructed of wattle and daub on a stone foundation, were clay loom weights, paint pots, storage jars, and a stone mortar, all still in place. On a hard-packed, well-trodden area between the workshop and the monumental building we found a toppled stone sculpture, a schematic human with faces on two opposing sides. The Gordion head is crudely carved, but its form replicates more sophisticated double-faced or "Janus" figures found at Celtic sites in Europe. The meaning of the Janus figures is not known, but the Gordion faces have the T-shaped brow and nose common on such heads. A silver coin found near it dates to the third century B.C. The construction of a monumental stone building within the earliest Galatian town indicates that Gordion's new Celtic rulers were ambitious, and able to mobilize a significant labor force. We do not know the cause or the date of abandonment of this first settlement (live by the sword?), but during the late third or early second century Galatian Gordion was rebuilt.

In the second settlement, new houses were erected on the old foundations. The monumental building was still impressive, but the nearby wall had been dismantled. A huge trench left when its stone blocks were removed was filled with earth and refuse. DeVries identified this second settlement as the one destroyed by a Roman army led by Consul Manlius Vulso in 189 B.C., as recounted by Livy. During the destruction, the large stone building burned fiercely, bringing down its tiled roof. The nearby, rebuilt workshop was thoroughly ransacked. Pots, clay figurines, and even figurine molds were smashed into small pieces and scattered, presumably by soldiers looking for valuables. The Romans' success perhaps accounts for the scarcity of metal finds in this building. Preserved in a small pit within the structure were a few more valuable items, including a small bone lion with a flat back that was probably used as an inlay.

[image] [image]
In 189 B.C., Roman troops pillaging Gordion left smashed pottery strewn across floors, including a painted bowl reconstructed from the debris. (Mary M. Voigt/Gordion Project)

These finds confirm and strengthen the picture of Galatian Gordion suggested by DeVries. What is striking is the juxtaposition of Greek and Celtic customs illustrated by new evidence for Galatian religious practices. On the mound, the workshop next to the monumental building produced figurines totally Greek in style that were presumably used in household ritual; figures of Greek deities such as Nike and Kybele were also found by Rodney Young. But off the mound we found remains left by very different rituals--chilling evidence of strangulation, decapitation, and bizarre arrangements of human and animal bones. Such practices are well known from Celtic sites in Europe and are now documented for Anatolian Celts as well.

In 1993, we began excavating in two parts of Gordion's lower town. South of the citadel mound, this part of the site had been heavily fortified beginning in the eighth century B.C. By the third century, the houses that had covered it had been abandoned, but heavy mud-brick walls with stone foundations were still standing. Within our trenches in area A, in the eastern part of the lower town, we found five bodies strewn across an outside ground surface that was dotted with small pits and trash, including rare sherds dated to the third century B.C. The bodies had been left exposed, and eventually were covered by a thin layer of silt eroded from the fortification walls.

We found the first body only inches below the surface. There was no evidence of damage to the skull or neck of this 20-35-year-old male, who lay sprawled face down. A second male, 30 to 45 years old, lay on his right side, his head twisted back and away from his torso, and his spinal column clearly broken as a result of a neck injury that presumably caused his death. To the north of the two men, a 15-20-year-old female, who also had a broken neck, lay on her right side. A fourth individual, a woman over 50, shows no signs of violence, but the position of her body suggests that she was tossed (rather than carefully laid) in a convenient pit.

We do not know the precise date of area A's grimmest deposit, the remains of four people thrown into a deep pit. Even if this burial group were later, their treatment is undoubtedly linked to ritual practices that began in third-century Gordion and would represent continuity of Celtic traditions after the town became part of the Roman province of Galatia. The uppermost body was that of a 30-45-year-old female who had been struck by two blows, which fractured her skull. Perhaps these blows did not cause her death, since she was also strangled, as a catastrophic angle in mid-neck attests. Beneath this body was that of a younger woman, aged 18-23 years. She shows no skeletal damage, but two heavy grinding stones weighing down her upper body do not suggest a peaceful interment. Resting to the west of the two women were the bones of a child aged 2-4, its preserved leg detached and reversed so that the knee rests where the hip should be. Because the entire body was disturbed and only partly present, we initially thought that much of the disturbance of small, light bones could be attributed to rodent activities. This relatively rosy picture disappeared when we found that the jaw, which appeared to belong to the cranium of the 2-4 year old, actually came from a 4-8-year-old child. Two neck vertebrae and a single foot bone also appear to be from this older child. It seems most likely that the children had died before the two women and that their bodies had decomposed, perhaps lying on the surface. We do not know why they were later deposited with the women.

[image] [image]
The uppermost of these two women had suffered blows to the head and a broken neck; large grinding stones had been placed on top of the lower woman. (Mary M. Voigt / Sondra Jarvis and Carrie Alblinger, Gordion Project)

It is clear that several of the people whose remains we found in area A died violently, with strangulation the most common cause of death, whether by hanging or garotting. All of these people were presumably "sacrificed," but we cannot determine the exact circumstances. One possibility is that they were killed as part of Celtic divination rituals. Greco-Roman sources report that the Celtic religious leaders, or Druids, were prophets who killed humans in order to discern the future as revealed by the dying victims' movements. The Gordion victims could have been war captives--a category of people used in divination, but sometimes simply slaughtered.

Excavation in area B, in the western part of the lower town, revealed clusters of human bones from bodies that had been dismembered. The remains, co-mingled with animal bones, were then carefully rearranged, sometimes in symmetrical patterns, on an outside ground surface with shallow depressions. A small number of sherds on this surface indicate that the area was in use during Hellenistic times, and two distinctive ones place it within the third century B.C. Again, only silt eroded from the fortification walls covers the bone deposits. Bone cluster 1 is the most complex. At the top was the skeleton of a young woman aged 16-21, but where her skull should be there was instead a lower jaw of an adult male over 50, the only bone of this person in the deposit. Beneath the young woman was a 35-45-year-old female whose legs had been detached and placed on either side of her torso. When we excavated bone cluster 1 we thought that the lower woman had been strangled because of the distorted angle within the spinal column. Analysis of the bones showed that we were dead wrong. Instead, the skull and first five vertebrae of the young woman had been placed at the top of the older woman's spinal column. Decapitation is obvious. There are no cut marks on the human bone, but placed around the young woman's feet were animal bones bearing cut marks from butchery. The skull of a 20-35-year-old male was found in bone cluster 2. Decayed wood in the opening at the skull's base through which the spinal cord passes suggests that this individual's severed head had been mounted on a wooden stake for display, a practice documented in Celtic Europe. Scattered around the skull were fragments of an ass's lower jaw, a pig's lower jaw, a cow's upper jaw, two cow pelvic bones, and the foreleg of a dog.

[image] [image]
Mixing of bones from different individuals was common in the Galatian burials at Gordion. Here, the jaw of one individual was placed at the top of another's spine. (Mary M. Voigt / Sondra Jarvis and Carrie Alblinger, Gordion Project)

In bone cluster 4, the skull of a teenager 12-17 years old was carefully placed above a dog skull, pelvic bone, and leg bones. At the bottom of this pile, which rested within a shallow depression, was a human pelvic bone from a 20-35-year-old male. Heavily weathered, it probably lay on the surface for some time before the rest of the bones were placed above it. The teenager, whose sex could not be determined with certainty, had been decapitated; this is clear not only from the fact that the first two vertebrae were still in place beneath the skull, but from damage to the vertebrae consistent with a butchery pattern found in animals, in which the neck is weakened by cutting to the point where it can be forcibly snapped to crack through the bone and remove the head.


A large deposit of animal bones with a few human remains mixed in, bone cluster 3 may be the remains of a Celtic autumn feast. (Mary M. Voigt/Gordion Project)

The largest deposit, bone cluster 3, consists of a few human bones, and over 2,100 animal bones and fragments. Three humans are present. A 4-8-year-old child is represented by a lower jaw and some cranial fragments. A fragmentary right pelvic bone came from an adult female aged 35-39, and a pair of pelvic bones represent an adult male aged 40-44. A sacrum (the fused vertebrae forming the back of the pelvis) and several long bones could belong to either adult. The human bones are cracked from weathering and the pelvic bones were gnawed by carnivores--signs that they were exposed on the surface for some time. More startling is a distinctive "spiral fracture" on a femur (probably from the male), which can only happen if the bone is fresh when broken. The shaft of the femur is one of the strongest parts of the skeleton, and to fracture it requires great force; today this type of fracture in an adult most commonly occurs from high-energy collisions such as a car crash. In this case, there is little that could have generated the twisting force great enough to cause such a spiral fracture except a fall from a great height or a direct blow near the time of death. Given that the bone is separated from the rest of the skeleton, a blow seems more likely, and this may be evidence of the offering of marrow to the spirits, a Celtic practice documented by textual evidence from Europe.

The distribution of body parts represented in bone cluster 3 is revealing. For humans, cranial fragments and pelvic bones predominate. Marks produced by carnivore teeth suggest that the pelvic bones bore meat when they were thrown into this bone pile. A similar pattern is found for horse and pig, both of which had symbolic value for Celtic groups, and in this case they are treated like people rather than like other animals. Bones of domestic cattle, sheep, goats, and ass are far more common, and the distribution of their body parts is more "normal" for animals used as food. If bone cluster 3 represents the remains of a feast, what happened to the three humans in this deposit? A well-known image from the Gundestrop cauldron, a large silver vessel found in Denmark that depicts deities and people and is usually attributed to the Celts, may provide an answer. One of the figures, twice the size of the horsemen and foot soldiers arrayed before him, is dunking a human into a large pot!

[image] [image]
The Gundestrop cauldron; detail shows a man being put into a cauldron (Malene Thyssen, Wikimedia)

We might also use European parallels to speculate about the timing of the feast. Based on their age at death, the animals in this large deposit were slaughtered in the fall. And it was in the fall that Celtic groups in Europe celebrated Samhain. Around November 1 each year, herds of domestic animals were brought from their summer pasture and culled, the herdsmen slaughtering weak animals that could not survive the winter. Celts believed that during Samhain the barriers between the natural world and the spirits broke down, and the veil between the present and the future was most transparent. Rituals were performed to foretell future events through various forms of divination, and it may not be too far a stretch to associate bone cluster 3 with this Celtic festival, which we still celebrate as Halloween.

The human and animal remains in area B document the Galatians' display and manipulation of human heads, a trait well known from texts and archaeological finds in Europe. While these heads are presumably trophies collected by the warrior elite, the importance of another male role, herding, is suggested by the animal remains in bone cluster 3. Both male and female victims, however, played an important role in the rituals conducted.


Stone foundations from the Galatian settlement at Gordion belie earlier interpretations of the site as a poor farming community. (Mary M. Voigt/Gordion Project)

Our view of Galatian Gordion has changed considerably over the past few decades. The picture of a simple farming community has been replaced by one of a prosperous town. But when the Galatians settled permanently in central Anatolia, they did not simply shed their old ways and adopt those of the native peoples (presumably Hellenized Phrygians in the case of Gordion), as our discoveries of sacrificial ritual involving humans as well as animals have shown. Major questions remain. For example, was the Galatian presence limited to a religious and military elite or did they form a larger segment of the population that gradually integrated with the local peoples? Our next step will be to compare the material culture of the pre-Galatian and Galatian settlements. If we can then distinguish immigrant from indigenous households, we should be able to discuss issues of ethnicity as well as the ways in which a farming and herding people adapt and prosper in an environment very different from that of their homeland.

Jeremiah R. Dandoy has been the Gordion Project's zooarchaeologist since 1994. Page Selinsky is a doctoral candidate in anthropology at the University of Pennsylvania, focusing on skeletal biology and molecular anthropology. Mary M. Voigt is Chancellor Professor of Anthro-pology at the College of William and Mary. In 1987 she became associate director of the Gordion Project in charge of excavation and survey. The authors would like to thank G. Kenneth Sams, director of the Gordion Project; T. Cuyler Young, Jr., head of the Royal Ontario Museum team; and Keith DeVries.