King Midas. The Phrygians. Alexander cutting the Gordian Knot. These are among the many subjects ordinarily associated with the site of Gordion (modern Yassihöyük), located along the banks of the Sakarya River approximately 95 kilometers southwest of Ankara, Turkey (Fig. I). Only rarely does one think of Romans in connection with the ancient Phrygian capital, for the city’s fascinating Iron Age and Hellenistic discoveries have tended to overshadow the remains of later periods and civilizations.
This is hardly surprising since Gordion had become a mere shadow of its former self by the Roman era. The site had lain abandoned for at least a century prior to the region’s annexation by the emperor Augustus (27 BC-AD 14), who in 25 BC transformed the former client kingdom of Galatia into a new Roman province of the same name. Both literary and archaeological sources attest to the establishment of a small settlement atop the Citadel Mound at some point not long after the annexation. The town was not among those chosen to become one of Galatia’s new administrative centers, however, and it never managed to regain its former size or political and economic importance.
In spite of its modest size and status, the Roman-period settlement at Gordion is of considerable importance for modern archaeologists and historians. It remains at present the only excavated rural site in the former territory of Roman Galatia, a province within which the vast majority of the population lived in rural surroundings. As such, the archaeological finds from Gordion offer a rare opportunity to examine the character of life in the ancient countryside, an environment about which the surviving literary sources are largely silent. The site is also an ideal location for studying local and interregional trade, since it sat alongside a major Roman highway that linked two of the province’s major cities, Ancyra and Pessinus (Fig. 2). Moreover, the community’s physical remains provide an excellent medium through which to evaluate the complex phenomenon of Romanization, the process by which non-Italic populations throughout the empire adopted Roman cultural and behavioral patterns. Although in recent years this phenomenon has become the subject of intense study and debate, the degree to which the rural population of Galatia was Romanized has rarely been examined and is not clearly understood at present. Thus while the settlement at Gordian was a modest one, its contents are precious if we are to reconstruct an accurate picture of ancient life in rural areas.
The Roman-Period Cemeteries at Gordion
Somewhat paradoxically, the best primary source that we currently possess for bringing Roman Gordian and its inhabitants back to life is the site’s cemeteries and their contents. The field of cemetery or burial archaeology has enjoyed something of a renaissance in recent years, with the result that mortuary evidence has come under increased scrutiny for its inherent historic and scientific value, particularly in the empire’s former western provinces. In contrast, scholarship in the Roman East has tended to remain narrowly focused upon urban cemeteries or necropoleis, monumental funerary architecture, and the associated inscriptional evidence. The cemeteries excavated at Gordion allow a very different phenomenon to be investigated, that of the rural necropolis, the type that must have existed outside of hundreds of similar Roman towns and villages across central Anatolia.
The Roman-period cemeteries at Gordion represent a model subject for discussion for several reasons. First, we possess stratified data from both the cemeteries and the settlement itself, portions of which were uncovered by the Körte brothers (in 19oo), Rodney Young (195o, 1969-1973), and most recently by Mary M. Voigt (1993-1997). Physical developments within the site and cemetery can be studied, compared, and correlated not only to improve our separate interpretation of each, but also to help us understand the complex relationship that existed between the spheres of the living and the dead.
Equally as important is the fact that at Gordion we possess a sufficient quantity of excavated burials for analysis. Few Roman cemeteries in any country have ever been excavated to their full extent, and in Turkey the number of recovered burials from rural cemeteries is small, most often the result of limited and hastily executed salvage operations. While the Gordion necropoleis have not been completely uncovered, the total number of excavated burials is large enough to allow us to detect general patterns in mortuary customs and burial techniques.
The Common Cemetery
Over the past half-century of excavation at Gordion by the University of Pennsylvania Museum, three separate Roman-period necropoleis have been uncovered. Here I concentrate upon the largest and best understood of the three, the cluster of 51 Roman-period graves that lies within the boundaries of the Common or Northeast Ridge Cemetery. This enormous burial ground is located slightly less than a kilometer from the Citadel Mound on a gentle, undulating ridge that forms the eastern edge of the Sakarya valley (Fig. 3). The ridge’s northern slope is now completely occupied by the modern village of Yassihöyük, while its spine, bisected by the modern road to Polatli, is crowned by a concentration of over 3o Phrygian tumuli, including the famous Midas Mound.
Among the monumental tumuli were found cremation and inhumation burials dating from the Early Bronze Age through the Roman period, a chronological span of almost three millennia. Between 1951 and 1953 and in 1962, Gordion teams excavated a total of 23o graves dating to nearly every major phase of the site’s occupation. G. Roger Edwards, in a 1963 Expedition article on recent finds from the cemetery, was quite correct to describe the necropolis as “a capsule account of Gordion’s history” (1963:43). In spite of its importance, however, much of the material from the Common Cemetery—including the Roman-period graves—has never been fully analyzed or published in detail. Although G. Anderson compiled a partial list of the Roman graves in the late 197os (198o:136-47), a full analysis of the burials and their contents was completed only recently (Goldman 2ooo:3o4-57). Fortunately, because the original field teams proceeded methodically and recorded their finds in detail, it has been possible to reconstruct the Roman-period necropolis via the wealth of information now stored in the Gordian Archives at the University of Pennsylvania Museum.
Archival research has led to the identification of 51 Roman-period inhumation burials at the approximate center of the Common Cemetery, encircled by tumuli and intermingled with the graves of the earlier periods (Fig. 3). That the Roman-period residents would select this ancient burial ground as an appropriate resting place for their own deceased is owed to a variety of factors. The ridge’s geographical situation is commendable, since it was secure from flooding in the lower valley, lay in a location unsuitable for agriculture, and possessed an unobstructed line-of-sight from the Citadel Mound, a fact that would have discouraged unwelcome attention from grave robbers. This stratagem appears to have been successful, since over three-fifths of the graves were found to be intact at the time of excavation, and at least half of the disturbed burials appear to have been disrupted by natural rather than human causes. Yet another appealing factor was the comfortable distance of the cemetery from the settlement proper, beyond the immediate environment of the living. For both ritual and hygienic reasons, the Romans invariably placed their cemeteries outside the gates or occupied limits of cities and towns, a custom that saw widespread adoption across the entire Mediterranean world by the Roman imperial era.
The presence of the Ancyra-Pessinus road is likely to have been an additional factor in the cemetery’s location (Fig. 3). Roadside cemeteries were a ubiquitous feature of the landscape during the Roman era, providing an eminently public setting within which monuments could be erected in clear view of passersby, thereby demonstrating the extent of one’s familial piety, social rank, and wealth. The Phrygian tumuli along the ridge undoubtedly served a similar purpose during the Ist millennium BC, and it would be surprising if the Roman-period residents were not cognizant of the analogous tradition embodied by these prominent monuments. Whether their own selection of the ridge was influenced by this consideration remains interesting to contemplate but impossible to prove. There is no question, however, that the inhabitants of Roman Gordion were aware of the earlier subterranean cemetery, since their own interments disturbed a number of pre-Roman burials.
Only a rough estimate of the Roman-period cemetery’s extent can be offered at this time (Fig. 3, in yellow). A series of test trenches (“TT”) was opened across the slope in I952 and 1953, and this irregular patchwork of 14 soundings (Fig. 3) has proved useful for delineating the cemetery’s original boundaries. Only 5 of the trenches contained Roman-period burials: 23 graves located near the Common Cemetery’s center. Confirmation of this pattern was obtained in 1962, when digging at the prospective site for the new Gordion Museum (hence the designation “Museum Site Trench”) led to the discovery of 28 additional Roman-period graves. Based on these findings, the necropolis of the Roman period appears to have covered an area of roughly 8o square meters, approximately one-quarter of the Common Cemetery’s full extent. If the spacing and density of the graves remain constant across this area, a total of 13o-15o Roman burials would not be an unreasonable estimate for the necropolis.
Grave Construction and Burial Offerings
Beyond the identification of grave offerings that are indisputably Roman in date, two features help to distinguish the Roman-period graves from the earlier burials. First, nearly all of the Roman burials have a roughly north-south orientation. The skeletal remains, on their backs with arms at the sides or crossed over the lower abdomen, normally lie with heads placed to the north and feet to the south. This general pattern appears to have been characteristic of and adhered to only during Roman times; pre-Roman burials rarely lie on a similar axis or were interred in a systematic manner. The careful arrangement and orientation of the graves suggest that some sort of markers were in use, possibly made of wood since no trace of them has yet been discovered. A second characteristic is the depth of the grave shafts in relation to the modern surface. The shafts proved to be quite deep, averaging 1.5o meters in depth, and contrast sharply with the less intrusive burials of pre-Roman times.
Six separate categories of burial construction may be differentiated among the 51 single and double interments. The most common type was the simple pit burial (Type I), 2o of which were discovered in the northern, western, and central sectors of the cemetery (Fig. 4, in blue). These simple pits, usually rectangular but occasionally oval in shape, held in general fewer burial offerings than the other five types, and those goods tend to be of comparatively poor quality. It is tempting to suggest on the basis of this evidence that the individuals interred in these simple graves were among the less wealthy residents of the settlement. Since several pit burials were found to contain valuable offerings like a gold earring and a silver ring, however, such generalizations must be treated with caution.
Another frequently encountered construction type is the step grave (Type 2), 15 of which were recovered in the central and southwestern portions of the cemetery (Fig. 4, in pink). These burials feature deep rectangular shafts at the bottom of which a second, narrower pit was cut to receive the skeleton (Fig. 5). The two small, longitudinal ledges or steps that were created along the sides of the upper shaft were used to support a covering of wooden logs, the fragmentary remains of which were discovered in several of the graves (Fig. 6). These graves were found to be more frequently robbed than any other type, the reason apparently lying in the method of construction itself. The weight of the shaft’s heavy earthen fill eventually crushed the cover of wooden boards perched upon the thin ledges, and the resulting collapse would have left a noticeable depression on the surface and have been easy pickings for the observant grave robber.
While fewer examples of the remaining four burial types have been unearthed, they tend to be more elaborately constructed and contain a greater number of offerings. They include graves with a stone and/or mudbrick lining (Type 3; Fig. 4, in purple), graves with wooden coffins or covers (Type 4; Fig. 4, in orange), graves constructed of reused baked bricks (Type 5; Fig. 4, in yellow), and a single example of a chamber or “catacomb” tomb (Type 6; Fig. 4, in green, and Fig. 7). One notable find was a carefully built tent of mudbricks (Fig. 8), an exact imitation of a common 3rd and 4th century AD type that normally utilized Roman rooftiles (teguloe) in their construction. Very few bricks or rooftiles have been recovered from the settlement, where it appears that most of the buildings were covered with a combination of thatch and clay. The desire for a thoroughly “Roman” style of burial was apparently not to be denied, however, and some creative individual evidently opted for the use of mudbrick as an acceptable substitute.
A majority of the intact burials contained at least one burial offering, and their quality indicates that the inhabitants of Roman Gordion possessed at least a modest level of wealth. The offerings were usually placed near the head or feet of the individual, a common pattern among graves of the Roman period. Similar objects found in stratified contexts across Turkey. the Black Sea region, and the eastern Mediterranean indicate that these offerings generally date to between the 2nd and late 4th centuries AD (Goldman woo :33(3-48). Interestingly, virtually absent from the burials are several categories of funerary offering common in nearly every quarter of the Roman world, namely coins (only one early 3rd century issue was found) and lamps.
Jewelry is the most common find, including necklaces of beads, pins of bone and iron (fibuIae), and signet rings of gold, silver, iron, and bronze. Carvings on the stone intaglios of the rings depict a broad range of subjects, from popular motifs such as goddesses (Athena; Fig. 9) and animals (eagle with a wreath; Fig. 1O) to more abstract symbols such as a hand pinching an ear (a sign of memory; Fig. 11) and a pair of clasped right hands (dextrarum iunctio, often used as a sign of matrimony; Fig. I2). One of the more unusual pieces of jewelry is a bronze necklace with a cast ornament (Fig. I3), a reused pendant of the type often found on military horse trappings of the 3rd and 4th century. Such decorative pendants were believed to have had magical or apotropaic qualities, and this particular piece was probably intended to ward off evil during the life and perhaps also the afterlife of the child who wore it.
Many of the offerings represent categories of goods commonly found in Roman graves of this period. These include ceramic vessels such as pitchers, long-necked bottles (unguentoria), two-handled jars (amphoras; Fig. 14), and tiny dishes, the latter perhaps used for cosmetic purposes. It is possible that some or all of these vessels played a role in the burial ritual itself, then were placed carefully in the grave before it was closed. Other objects seem to have been prized possessions, such as a delicate glass flask (Fig. 15), a collection of well-worn bone spindles and whorls, a mirror, a small knife, and a bell.
The most common item within the burials was pairs of hobnail boots, worn on the feet of the deceased or, in at least one case, placed carefully beside them (Fig. 16). Over one-third of the graves contained remnants of these boots, either clusters of small nails by the feet or small, omega-shaped (Q) eyelets probably used to hold the laces. While similar boots are often found in graves on the Roman frontiers and in Western Europe, usually at sites with a military connection, there are at present no published parallels from Turkey.
Who Lies Buried in the Common Cemetery?
In many ways, the Common Cemetery is typical of rural necropoleis dating to the 2nd-4th century AD and found throughout the Roman world. The location of the cemetery beside a road, the range and type of burial offerings, and the mélange of diverse construction types are by no means exceptional. but rather fit a standard profile that archaeologists have come to expect at Roman sites where long-term occupation has taken place. That the evidence from Gordion appears to fit comfortably within this tradition tallies well with our understanding of the Roman empire, within which the effects of Romanization often brought to the cultural landscape a startling level of homogeneity. The mudbrick version of the standard tegulae grave is an excellent example of such processes at work.
One active component in this complex process of cultural transmission was the movement and resettlement of people into and within the empire, an activity that appears to have left traces within the Common Cemetery. Both the step and catacomb graves (Types 2 and 6) have their closest parallels in the northern Black Sea region, where they were firmly established as the standard types of burial by the 1st century BC (Firsov 1999:1-4). The deep shafts and north-south orientation of the graves are considered characteristic of Late Sarmatian (Ist-4th century AD) mortuary traditions, as is the general range of burial goods found in the graves at Gordion. Indeed, many of the offerings, such as the small amphoras. match most closely those found at Crimean sites like Chersonesus and Tanais (Zubar’ 1982:72; Arsen’eva 1977:86).
The Sarmatians were an association of Indo-European-speaking tribes of central Asian origin. In the 6th century BC, these nomadic tribespeople began to move westwards towards the Black Sea, and by the mid-1st century AD had established themselves along the Black Sea littoral between the Danube and the Sea of Azov. They soon began to lead a more settled existence, and by the Late Sarmatian period were both trading and intermarrying with the Greek colonists and former Scythian overlords of the region (Millar 1967:281-92). The Romans were their immediate neighbors, and their general relationship is probably best described as one of a “love/hate” character, since the Sarmatians fought as often with the Romans as against them.
Exactly how a burial type characteristic of the northern Pontic region came to be used at Gordion is difficult to say at present. Although several literary and epigraphic sources tell us of Sarmatians being resettled within the nominal boundaries of the Roman empire, there is no mention of a Sarmatian presence in Anatolia. In addition, there exists no incontrovertible evidence, such as the presence of Sarmatian objects with tribal markings (taigas), to indicate that these people had relocated to Gordion. Indeed, the deceased themselves present a thoroughly “Roman” appearance, buried with their signet rings, unguentaria, and hobnailed boots. Step and catacomb graves, however, are not currently attested to anywhere else in Turkey, and the Sarmatians are known to have traveled widely and been resettled in many territories within the boundaries of the Roman empire (Sulimirski 197o:164-82). Direct cultural contact would help to explain another interesting phenomenon. the striking resemblance between the Roman ceramics from the settlement at Gordion and pottery from sites along the northern Black Sea coast. It is possible that long-distance trade or a Black Sea influence on regional ceramic production could account for these similarities. However, since the published accounts of extensive regional surveys in northern Anatolia contain no corresponding wares, this phenomenon at present appears more likely to be a localized one. Thus, while we cannot say for certain that people of northern Pontic origin were present at Gordion, evidence from both funerary and settlement contexts reveals that strong, perhaps direct links existed between the Black Sea region and this central Anatolian town.
A second, apparently unique aspect of the burials—the presence of the hobnail boots— offers a possible explanation. Historians have attempted to equate Gordion with a posting station (statio) named either Vindia or Vinda, a site along the Ancyra-Pessinus highway that appears in ancient road itineraries (French 1978:294; Mitchell 1993:55). Although no direct proof has been found to confirm this hypothesis, Gordion distance from other known points on the road indicates that the statio lay at or near the site. In 1996, the discovery of a Roman soldier’s tombstone in a field near the Citadel Mound has provided the first concrete evidence that the military were operating in some capacity in the area (Goldman I997). While the hobnail boots do not prove per se that Gordion had a minor military function, several other finds from the Common Cemetery—the pendant necklace and the intaglio with the wreath-bearing eagle—seem to bolster the idea that either soldiers or veterans were present at the settlement. If they were Sarmatian auxiliaries, who are known to have served as cavalrymen in the Roman army, the facts would dovetail neatly. Such a proposal is purely speculative at this stage of research, however, and only future excavations or other chance finds will reveal the truth, if it can be recovered at all.
Much work still needs to be done within the Common Cemetery and the other necropoleis at Gordion, and numerous issues remain to be explored. The skeletal remains, for instance, an important source of information about diet and disease in the Roman world, are currently stored in Ankara and have yet to be analyzed in any detail. The value of this material’s continued study is heightened by the fact that rural cemeteries in central Anatolia are disappearing at a rapid rate, their contents pillaged by treasure seekers or destroyed as a result of modern development. Gordion is more fortunate than many sites, enjoying not only protected status but also the presence of a local gendarme station in nearby Yassihoyiik. It is to be hoped that the study of its Roman-period cemeteries
ANDREW L. GOLDMAN was the 2ooo-2oo1 Rodney S. Young Postdoctoral Fellow at the University of Pennsylvania Museum and is currently a Research Associate of the Mediterranean Section. In May 2o0o, he received his Ph.D. in Classical Archaeology from the University of North Carolina—Chapel Hill. He has been an active member of the Gordion Excavation Team since 1992, and is currently engaged in cataloguing and publishing archaeological material related to the Roman-period occupation levels at the site. His research has focused upon the impact of Roman suzerainty and the effects of Romanization in central Turkey, primarily within the territory of the former Roman province of Galatia. He is currently at work preparing a monograph on Roman Gordion, the first comprehensive study of a Roman-period town in central Turkey.