The Lydians and their World (2010)

by Christopher Roosevelt

Lydia Before the Lydians


Long before the rise of Sardis, the capital of the Iron Age Kingdom of Lydia, the broad region of historical Lydia had been inhabited for millennia. Yet precisely when Lydian-speaking populations first migrated into this area of central western Anatolia is inexactly known, and the regional definition of “Lydia” itself, as known best from later Classical sources, may have had little significance before the reigns of the historically attested kings of Iron Age Sardis. Thus, a review of “Lydia before the Lydians” must be defined by somewhat arbitrary geographical and chronological limits; we are still very ignorant of the cultural composition and regional definition of the area’s earliest inhabitants. Our focus here, then, will be on the area of inland central western Anatolia that overlaps the historical kingdom of Lydia, defined at its core by the broad valleys of the Gediz (Classical Hermus) and Küçük Menderes (Classical Cayster) Rivers and their larger tributaries in the modern Turkish provinces of Manisa, İzmir, and Uşak. Chronologically, coverage will conclude with the end of the Late Bronze Age, ca. 1200 BCE, and the still little-explored transition from the Late Bronze Age to the Early Iron Age, which led to the gradual rise of a centralized power at Sardis, as chronicled in Greek sources.

Research into the prehistoric archaeology of central western Anatolia has been conducted now for more than a century, but only recently intensively. That it has been defined primarily by surface surveys rather than by excavations has resulted in a general understanding of the distribution of major sites throughout the region, but the lack of stratigraphic control leaves something to be desired concerning the correlation of local chronological phases to broader developments in Anatolian and Aegean spheres. Only stratigraphic excavations will provide more conclusive understandings of all periods of Lydian prehistory. Gaps in knowledge result also from both the targeted nature of most surveys, which have aimed to locate and chronologically characterize conspicuous occupation mounds at the expense of other types and sizes of sites, and from still-active alluvial and erosional processes1 that have obscured many early sites in and along the edges of the main river valleys. Nevertheless, the prehistoric archaeology of this region seems to be particularly rich, spanning Paleolithic through Late Bronze Age times, and ongoing and future intensive surveys and excavations promise to produce increasingly important results.2 In addition to the significance of ephemeral traces of Paleolithic through Mesolithic activities and the cultural affiliations and settlement patterns of the sedentary populations living in central western Anatolia between the Neolithic and Late Bronze Age periods, key remaining questions concerning a “Lydia before the Lydians” include the nature of the transition from the Late Bronze to the Early Iron Age and the date and circumstances of the arrival of Lydian-speaking populations.

The Paleolithic and Epi-Paleolithic Periods

Paleolithic remains in inland central western Anatolia are few and far between, yet they show evidence of early activities in the area that were probably concentrated in both cave and open-air sites located along what must have been common foraging grounds and routes of travel from early times. That Anatolia in general was significant in the migration of technologies and early human ancestors between Africa, southwest Asia, and Europe is a common assumption,3 and central western Anatolia in particular may have participated in these broader processes. Aside from possibly Paleolithic tools found in caves on the slopes of Manisa Dağ (Classical Mount Sipylus) and rock-art from the southern side of the Küçük Menderes River valley,4 the most securely dated evidence for Paleolithic activities comes from the open-air site of Bozyer in Bin Tepe, an area better known for its Iron Age burial mounds, just south of Marmara Gölü (the Classical Gygaean Lake) (Figs. 1, 2, 3, 4, 5, 6).5 Found distributed across the low mound and its accompanying gully, chert cores and flakes of Levallois type and quartzite cores of possibly Abbevillian type prove the presence of early human foraging groups already in Middle Paleolithic times (ca. 250,000–45,000 years ago), and possibly even earlier.6 Local tool types reflect well-known Levantine and European technologies, demonstrating the broad affinities of cultural manifestations during these eras and highlighting the role of central western Anatolia as an intermediary between these better-known regions.

Evidence from the Upper Paleolithic and Epi-Paleolithic periods (ca. 45,000–15,000 years ago), characterized in Europe by “revolutionary” increases in the abundance and types of symbolic representation usually associated with the appearance of fully modern humans, remains relatively unknown in central western Anatolia, as in many other areas of Anatolia.7 A few possible examples of geometric microliths and other stone tools known from central Lydia alone may represent local Epi-Paleolithic activities contemporary with the period known as the Mesolithic in Aegean spheres.8 The scarcity of Paleolithic–Epi-Paleolithic sites elsewhere in the area probably results more from the methods and interests of surveys, usually extensive and focusing on later prehistoric periods, than from a real scarcity of early prehistoric activities. Recent work on other open-air and cave sites in northwestern, southwestern, and central Anatolia confirms this point and suggests promise for future discoveries in the area.9

Long revered as the most spectacular evidence of Paleolithic people in central western Anatolia is a set of footprints located near the village of Sindel in the Gediz River valley, between Adala-Karataş and Köprübaşı (Fig. 7). The footprints were originally made in wet sediments, were coincidentally preserved by basaltic lava flows from one of the Pleistocene–Holocene-epoch cinder cones of the Kula Volcanic zone known as Çakallar Tepe, and are reminiscent of the later-discovered and far better known Australopithecus afarensis footprints from Laetoli, Tanzania.10 Despite earlier attempts to date the footprints to much earlier human ancestors (around 250,000 years ago), recent radiometric dates suggest they were made during the Holocene epoch11—that is, during the last 10,000 years or so, a date that might have been suggested otherwise by the fact that the human prints were accompanied by dog prints, and dogs had not been domesticated until around 15,000 years ago. Given the paucity of pre-Neolithic finds in Lydia to date, the prints were very likely made during somewhat later Holocene times, perhaps by local inhabitants of the Neolithic period, and well before the area was well known to Classical historians who might have chronicled the volcanic events had they occurred in contemporary or recent times.

  • Fig. 1

    Map of Paleolithic, Neolithic, and Chalcolithic sites in central western Anatolia (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 2

    View of the environs of Bozyer in Bin Tepe, on the northern margin of the Gediz River Valley (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 3

    View of the environs of Bozyer in Bin Tepe, on the northern margin of the Gediz River Valley (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 4

    View of the environs of Bozyer in Bin Tepe, on the northern margin of the Gediz River Valley (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 5

    View of the environs of Bozyer in Bin Tepe, on the northern margin of the Gediz River Valley (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 6

    View of the environs of Bozyer in Bin Tepe, on the northern margin of the Gediz River Valley (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 7

    View of a footprint preserved by the Çakallar lava flows near Sindel (Dedeoğlu 2003, 18)

The Neolithic and Chalcolithic Periods

The transition from the Epi-Paleolithic (or Mesolithic) to the Neolithic period in central western Anatolia is not well understood, but at present it seems that new populations may have produced the changes witnessed in the archaeological record.12 Current research suggests that the simultaneous appearance of sedentism and agricultural lifeways in small-scale village society defines the beginning of the Neolithic in this area by the mid- to late-seventh millennium BCE, here called the “Late Neolithic” because these changes are witnessed elsewhere in Anatolia at much earlier dates.13 Where Neolithic settlers came from, either the Lake District or elsewhere, is an issue yet to be resolved with any clarity, and ongoing research may eventually show that Late Neolithic cultures developed out of still unknown and earlier local Neolithic traditions. This Late Neolithic period, extending into what is considered to be the Early Chalcolithic period of the early sixth millennium BCE elsewhere in Anatolia, is contemporary with the earliest Neolithic cultures of southeastern Europe and the Balkans. Western Anatolia, in fact, is sometimes viewed as the land bridge over which agricultural peoples passed from the Near East to Europe,14 but more research in western Anatolia is needed to support this idea. Nonetheless, we have much fuller data available for the Late Neolithic period than we do for earlier periods.

The type-site for the central western Anatolian Late Neolithic is now Ulucak Höyük, in the Kemalpaşa Valley, where excavations began in 1995 and where the Late Neolithic of the late seventh millennium and the early sixth millennium BCE is well represented by at least two primary building levels.15 Of the cultural and economic connections among the inhabitants of the site, interpretations based on a variety of features—architecture and site organization (waddle-and-daub, or “post-wall,” agglomerated units and later mudbrick, pisé, and timber architecture built along streets), ceramics (red-slipped and burnished pottery of diagnostic forms) (Fig. 8, 9), and other small finds (e.g. figurines, bone spoons)—suggest affinities with, but local distinctions from, areas to the north, especially the “Fikirtepe Culture” around the Sea of Marmara and Thrace, and to the east-southeast, around the Lake District.16 Connections with the Aegean and central Anatolia are indicated also by the presence of both Melian and central Anatolian obsidian at Ulucak and elsewhere in the region.17 Figurines and patterned-textile production at Ulucak no doubt reflect well-developed ritual and symbolic practices and beliefs as well.18

While Ulucak remains the only excavated Neolithic site in inland central western Anatolia, at least eleven other sites exhibiting similar surface assemblages, especially ceramics, are known elsewhere in the area: two more in the Kemalpaşa Valley itself, six sites in the Manisa and Akhisar Valleys, north of Manisa Dağ, and three others in the Alaşehir Valley, to the southeast.19 All these Late Neolithic sites are found on mounds raised from valley floors or on the natural heights of foothills extending from valley-defining mountain ranges; all are surrounded by or in close proximity to well-watered, agriculturally rich land, suggesting, in part, that site locations were chosen to suit communities with agricultural ways of life.

Following the local Late Neolithic, the Middle Chalcolithic period of the second half of the sixth millennium and the first half of the fifth millennium BCE appears to be defined by a widespread abandonment. Witnessed both by surface surveys and in the stratigraphy of Ulucak, this apparent gap in settlement is attested also across much of western Anatolia. When settlement resumes in the beginning of the Late Chalcolithic period of the second half of the fifth millennium and the fourth millennium BCE, sites are still located in agriculturally productive areas. Elsewhere in western Anatolia, upsurges in metallurgical and stone-working technology, long-distance exchange, and social complexity are contemporary trends. In central western Anatolia these trends are best illustrated by the site of Kulaksızlar in the Akhisar Valley, a mid-fifth millennium BCE production center of marble vessels and the marble Kilia-type figurines that are found distributed throughout western Anatolia and also in the Aegean (Fig. 10).20 Aside from Kulaksızlar, a general burgeoning of settlement is indicated at this time by the appearance of at least twenty-four sites dating primarily to the fourth millennium BCE.21 As in the Late Neolithic, the ceramic repertoire again indicates relations with sites to the north, in the Troad and around the Sea of Marmara, and to the east-southeast in the Lake District, but now also with coastal and east Aegean sites; yet in its somewhat different quality and proportions, the pottery suggests the region may have been culturally distinct, as in the Late Neolithic period.22 By the late fourth millennium BCE, ceramic connections with southwest Anatolia appear to have been severed, and more distinct traditions common to central western Anatolia and areas to the north became prevalent.23

With apparent continuities in both site preferences and interregional connections between Late Neolithic and Late Chalcolithic central western Anatolia, one wonders about the significance of the Middle Chalcolithic gap. While some scholars have argued that the Late Chalcolithic resurgence is one of immigrant populations from southeastern Europe, such continuities argue for caution in invoking migrations at this time.24 Nonetheless, increasingly strong connections between southeastern Europe and western Anatolia in the late fifth millennium BCE and, especially, in the fourth millennium BCE are indicated by ring-idols, ceramics, and figurines.25 The nature of these interconnections, however, has yet to be firmly established, and they cannot be attributed confidently to massive migrations; a broad cultural interaction sphere may be the better model.26 The participation of central western Anatolia in such an interaction sphere would be unsurprising in light of its status as a nexus of communication corridors, a function that is well attested by the distribution of sites at regular intervals along the courses of major river valleys connecting both coastal and inland and northwestern and southwestern Anatolia.27

  • Fig. 8

    Selection of Neolithic pottery sherds from Ulucak Höyük (Courtesy of the Ulucak Excavations)

  • Fig. 9

    Neolithic cup from Ulucak Höyük (Courtesy of the Ulucak Excavations)

  • Fig. 10

    Composite illustration of a Kilia-type figurine from working fragments found at Kulaksızlar (Courtesy of T. Takaoğlu)

The Early Bronze Age

The broad cultural interaction sphere that defined the Late Chalcolithic cultural affinities of central western Anatolia continued into the Early Bronze Age (EBA) of the third millennium BCE with evidence of increasing social complexity and regional interconnectivity, and with little to no evidence of cultural discontinuity. In general, the western Anatolian EBA is characterized by a number of gradual changes: population increases marked by the appearance of many new sites; changes in settlement size and structure, including the appearance of monumental buildings and fortifications; increases in the sophistication of craft-production technologies, especially with respect to metallurgy and ceramics; and increasing evidence of long-distance exchange, probably practiced most frequently between elite members of increasingly centralized communities, perhaps chiefdoms, who signified their status in one way with exotic goods, as indicated by mortuary evidence. These trends emerge in the EB I phase (ca. 3000–2700 BCE) but become emphasized especially in the EB II (ca. 2700–2400 BCE) and EB III (ca. 2400–2000 BCE) phases of the broader period. These phenomena can be identified only tentatively in central western Anatolia, and their relationship to developments and chronological phases more clearly identified elsewhere remains obscure; yet some broad patterns emerge from the available evidence.

Limited excavations—in the early twentieth century at Yortan (modern Bostancı), in the upper Bakır River valley just south of Gelembe; in the 1960s at Ahlatlı Tepecik and Eski Balıkhane, on the southern shore of Marmara Gölü; and more recently at Gavurtepe, near Alaşehir, and Ulucak, in the Kemalpaşa Valley—provide the best contextual evidence for the period, revealing broad similarities of material from in both burial and habitation contexts.28 At the same time, surface surveys have revealed a widespread and more regular distribution of over 100 sites of EBA date throughout the region, the majority of them apparently newly settled in this period (Fig. 11).29 As in earlier periods, most settlements are on mounds that rise from valley floors, while some are located on higher ground, on the natural ridges and plateaus at valley edges.

For the EBA, as for the Neolithic and Chalcolithic periods before it, the paucity of thoroughly excavated sites in the area limits cultural and economic interpretations to those that can be derived from pottery and other small finds, frequently found in burials of distinctive traditions. While details still need to be worked out, based on the discovery of fuller assemblages of material, the EB I and EB II phases may generally be characterized by the solidification of broad pottery zones within western Anatolia, so that Lydia falls within what has been identified as the Troy I-Yortan-İznik zone,30 with pottery like that recovered from the pithos cemetery of Yortan predominating. The appearance of the same style pottery at sites located throughout the Gediz River valley as far east as Kula shows the broad extent of the style, which may have penetrated the uplands and valleys to the east and south as well, probably indicating a shared cultural milieu within this area.31 Ceramic connections to sites to the east (e.g. Kusura and Beycesultan), to the west (e.g. coastal and Aegean island sites such as Iasos and sites on Lesbos, Chios, and Samos), and to the south (e.g. Aphrodisias in Caria and Karataş-Semayük in Lycia) are apparent as well, showing the broader connections of the EBA interaction spheres of western Anatolia.32

Settlement sites have yet to be excavated fully enough to reveal details of local settlement structure, yet mortuary remains from the excavated sites mentioned above provide good information about the range of burial types and grave goods found in both intramural burials (usually associated with children) and extramural cemeteries. As elsewhere in western Anatolia, the most frequently attested type of burial is in a pithos, with the deceased accompanied by grave goods that almost always include ceramics, and, in particular examples, a rich assortment of other items: stoneware vessels; animal and anthropomorphic figurines, or “idols,” in stone or metal; polished stone axes; copper-alloy weapons; and silver, gold, or electrum personal adornments (Figs. 12, 13, 14, 15, 16, 17, 18, 19).33 As elsewhere in regions peripheral to Aegean coastal territories, burials in cists are found as well, sometimes alongside pithos burials, and from these are usually inferred influences from Cycladic traditions.34 Additional Cycladic connections can be found in two stone bowls of Cycladic manufacture and EB II date found in a pithos burial in the Akhisar Valley.35 These Cycladic interconnections, among others suggested by mortuary and settlement evidence, need not represent the migration of new peoples with their particular mortuary traditions; rather, they probably resulted from the same type of broad interaction sphere that defined Late Chalcolithic cultural affinities, if not from entirely internal Anatolian developments.36

In the late EB II phase and, following destructions common to many sites at the end of EB II, the EB III phase, such broad interaction spheres resulted in significant increases in the exchange of materials across Anatolia between elites now associated with chiefdoms at such sites as Troy, Liman Tepe, and Karataş-Semayük in western Anatolia, Demircihöyük and Küllüoba in central Anatolia, and Tarsus in Cilicia, among many other sites. Known casually as the “Anatolian Trade Network” or “Great Caravan Route,” this interaction sphere of the second half of the third millennium BCE resulted in the interregionality of material assemblages, including characteristically western features such as stone idols (Fig. 17), and ceramic forms, including depata, tankards, and jugs with cutaway spouts (Fig. 15), and characteristically eastern features such as wheel-made pottery, Syrian bottles, and a general increase in metal goods now made from real tin-bronze (e.g. seals, weapons, and pins).37 While many such finds can be located among the assemblages of central western Anatolia, the most ubiquitous manifestation of its participation in this network of the late third millennium BCE is the widespread occurrence of wheel-made pottery known as “West Anatolian Red Slipped Ware.”38These and other pieces of evidence point towards a culmination of the trend of centralization, especially of production, that began earlier in the millennium and that led to a crescendo of social and political complexity in the subsequent period.

  • Fig. 11

    Map of Early, Middle, Middle/Late, and Late Bronze Age sites in central western Anatolia (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 12

    Copper (alloy?) dagger, from an Early Bronze Age pithos burial at Eski Balıkhane. (EB 69.3), No. 4 (Courtesy of the Vedat Nedim Tör Museum, Istanbul)

  • Fig. 13

    Silver ram “pendant,” from an Early Bronze Age pithos burial at Eski Balıkhane (EB 69.3), No. 1. (©Archaeological Exploration of Sardis/President and Fellows of Harvard College)

  • Fig. 14

    Gold earplugs, from an Early Bronze Age pithos burial at Eski Balıkhane (EB 69.3), Nos. 2-3 (©Archaeological Exploration of Sardis/President and Fellows of Harvard College)

  • Fig. 15

    Cutaway-spouted jug, from an Early Bronze Age pithos burial at Eski Balıkhane (EB 69.3), No. 5 (Courtesy of the Vedat Nedim Tör Museum, Istanbul)

  • Fig. 16

    Small tankard, from an Early Bronze Age pithos burial at Eski Balıkhane (EB 69.3), No. 6 (Courtesy of the Vedat Nedim Tör Museum, Istanbul)

  • Fig. 17

    Marble idol, of the Early Bronze Age from Sancaklı Bozköy, No. 7 (Courtesy of the Vedat Nedim Tör Museum, Istanbul)

  • Fig. 18

    Polished stone handaxe, of the Early Bronze Age from Turgutlu, No. 8 (Courtesy of the Vedat Nedim Tör Museum, Istanbul)

  • Fig. 19

    Spearhead of the Early Bronze Age from Turgutlu, No. 9 (Courtesy of the Vedat Nedim Tör Museum, Istanbul)

The Middle and Late Bronze Ages

The long-term results of the “Anatolian Trade Network” elsewhere in the Middle Bronze Age (MBA, ca. 2000–1600 BCE) included the rise of city-state systems best known from sites like Kültepe, Alişar, Acemhöyük, Karahöyük-Konya, and Beycesultan, where large site sizes, palaces, massive fortifications, and evidence of long-distance commercial interactions are characteristic. Gradually developing and strongly centralized state systems appear in central western Anatolia, too, but perhaps not until slightly later, after several centuries sometimes thought of as transitional to the western Anatolian MBA.39 While the earlier part of the MBA is still not well understood locally, by the latter part, ca. 1700 BCE, large fortified citadels with lower towns had begun to develop in the middle Gediz River valley, along the lines of similarly pronounced and roughly contemporary developments in the Troad (e.g. Troy VI). The centralization of production and power associated with this period might be reflected also in settlement patterns, wherein a general reduction in the number of sites spread throughout inland central western Anatolia—from over 100 sites in the third millennium BCE to roughly 70 sites in the second millennium BCE—may suggest local abandonments in favor of nucleation at larger population centers.

Identifying changes within second-millennium BCE settlement patterns, however, is complicated by the durability of local pottery traditions throughout the millennium, so that it is often impossible to date sites more closely than to the millennium. This problem arises from an inability, in many cases, to separate the Middle Bronze Age from the Late Bronze Age (LBA, ca. 1600–1200 BCE) based on surface pottery alone. Despite the general contraction in settlement numbers throughout the second millennium BCE, the same general areas occupied in the third millennium BCE remained populated, with site preferences appearing to follow the pattern set millennia earlier. A reduced number of sites can be found in areas such as the valleys between Akhisar and Manisa; yet, significantly, an increase in the number of sites and a distinct change in settlement patterns occur in the middle Gediz River valley, the heartland of the later Iron Age capital at Sardis.

Here, in the foothills and along the lakeshore surrounding Marmara Gölü (Figs. 20, 21, 22, 23, 24), a recently discovered network of at least four fortified citadels—two small and two large—developed by ca. 1700 BCE, and each citadel was used until the end of the LBA, after which none was significantly reoccupied (Fig. 25).40 All show some evidence of burnt destructions on their surfaces, but it is impossible to determine from surface materials alone whether such destructions caused their abandonment or came somewhat earlier in their histories. The two smaller citadels, Kızbacı Tepesi and Gedevre Tepesi, enclose roughly 1 ha of fortified space each. The two others enclose areas that are substantially larger than any known contemporary citadel in all of western Anatolia, with Asartepe enclosing 3.8 ha and Kaymakçı a total of 8.6 ha. The citadels of Gavurtepe, Bademgediği Tepe, and Troy, for example, are much smaller, each enclosing between 1 ha and 2.2 ha. In addition, the two larger citadels of the network surrounding Marmara Gölü are each associated with large lower settlements. Because of its dominant position and size and its internal complexity, as determined from microtopographic and geophysical survey, Kaymakçı is best understood as the capital of the local network, and probably also as the capital of the entire region (Figs. 26, 27, 28).

Preserved in ruined condition at the highest elevation of the site is an inner circuit of fortifications that defines a central sector or inner citadel characterized by concentric terraces surrounding an acropolis-like platform. The general organization of this inner citadel at Kaymakçı bears some similarity to the citadel of Troy VI–VII, yet at Kaymakçı it composes only one part of the larger well-fortified space. An outer circuit of fortifications at Kaymakçı helps to define lower terraces in a broad southeastern sector that seems to be characterized by buildings arranged in a roughly grid-like organization, as well as a western sector characterized by large open spaces and monumental buildings, at least one of them apparently built in basic megaron form. Upper slopes within the outer circuit of fortifications at Kaymakçı were also built up, but the definition of their use remains obscure at this time. Outside the citadel, a large lower settlement at Kaymakçı adds tens of hectares to the total site area and indicates its potential to host a large community. This lower settlement is evidenced primarily by surface pottery that extends away from the highpoint of the site to limits that have yet to be identified.

Pottery from these second-millennium BCE sites in central western Anatolia bears closest affinities to assemblages from sites in coastal western Anatolia, from Panaztepe to Troy, as well as from the inland sites of Kusura, Beycesultan, and Aphrodisias (Fig. 29).41 Contacts with regions further afield, such as Hittite central Anatolia and the Mycenaean sphere of the Aegean, are apparent in imported wares and their local imitations, and perhaps also in burial customs.42 That central western Anatolia was part of the larger Hittite world has been assumed ever since the early discovery of rock-cut monuments in the area: one on a cliff at the northern terminus of Manisa Dağ at Akpınar; the other on a south-facing cliff in the Karabel Pass through the Boz Dağ range, at the southern edge of the Kemalpaşa Valley (see Fig. 11).43 Both monuments present relief sculptures accompanied by inscriptions in hieroglyphic Luwian, a language of the Anatolian sub-branch of the Indo-European language family (along with Nesite (Hittite), Palaic, etc.). Both are understood as monuments established by local rulers rendered in styles, traditionally considered Hittite,44 that had become common throughout much of central, south-central, and western Anatolia by the thirteenth century BCE. Associated with a nearby spring, the monument at Akpınar has been identified traditionally by Classical geographers and modern scholars alike as a seated image of Cybele or another "mother goddess" figure, yet an alternative reading sees it as a bearded mountain god (Figs. 30, 31).45 The Karabel relief is more overtly geopolitical in nature and provides evidence for the identity of the LBA kingdoms that occupied the area of early Lydia (Fig. 32).

The hieroglyphic Luwian inscription of the Karabel monument records that it was established by Tarkasnawa, a late-thirteenth-century king of Mira, a territory known from archives recovered from the Hittite capital at Hattusa (modern Boğazköy).46 Earlier in the second millennium BCE, Mira and its neighbors had been part of a regional complex known to the Hittites as the Arzawa Lands. Comprising all of central western Anatolia, and probably greater territories around it, Arzawa was a significant power in LBA Anatolia, whose kings corresponded diplomatically with Hattusa and Egypt alike.47 Frequent sympathizing with the Hittite adversary state known as Ahhiyawa (probably an east Aegean Mycenaean state or group of kingdoms known to Homer as the Achaeans), and Ahhiyawan territorial conquests during the mid- and late fourteenth century BCE, led to the conquest and dissolution of Arzawa into component vassal kingdoms by the Hittite king Mursili II.48 Mira, with its capital at Apasa (Classical Ephesus) and with territories extending at least over the Büyük Menderes (Classical Meander) and Küçük Menderes valleys, coincided with the southern areas of Classical Lydia.49 The Seha River Land, another member of Arzawa forced into vassalage by Mursili II, shared a border with Mira and is now identified as the Gediz River valley and areas farther north, coinciding with the heartland of Classical Lydia.50 Thus the largest of the citadels around Marmara Gölü, Kaymakçı, might be identified as the regional capital of the Seha River Land. We know from scarce mentions in the archives from Hattusa that the vassal Seha River Land remained a strategic intermediary between Hittite and Ahhiyawan lands and interests at least until the late thirteenth century. Thereafter, all historical traces of it were lost in the tumultuous decades of destructions and migrations associated with the end of the Late Bronze Age throughout the eastern Mediterranean.

  • Fig. 20

    View of Marmara Gölü, the ancient Gygaean Lake, in central Lydia (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 21

    View of Marmara Gölü, the ancient Gygaean Lake, in central Lydia (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 22

    View of Marmara Gölü, the ancient Gygaean Lake, in central Lydia (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 23

    View of Marmara Gölü, the ancient Gygaean Lake, in central Lydia (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 24

    View of Marmara Gölü, the ancient Gygaean Lake, in central Lydia (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 25

    Map of Chalcolithic and Bronze Age sites around Marmara Gölü (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 26

    Sheep grazing on the lower southeastern terrace of the citadel at Kaymakçı (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 27

    Sheep grazing on the lower southeastern terrace of the citadel at Kaymakçı (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 28

    Three-dimensional rendering of a QuickBird satellite image of the citadel of Kaymakçı (view to WNW) (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 29

    Selection of Middle and Late Bronze Age pottery from sites around Marmara Gölü (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 11

    Map of Early, Middle, Middle/Late, and Late Bronze Age sites in central western Anatolia (Courtesy of Christopher Roosevelt and the Central Lydia Archaeological Survey)

  • Fig. 30

    View of the Akpınar Monument, just east of Manisa, probably representing a bearded mountain god (Photograph by Crawford H. Greenewalt, jr.)

  • Fig. 31

    View of the Akpınar Monument, just east of Manisa, probably representing a bearded mountain god (Photograph by Crawford H. Greenewalt, jr.)

  • Fig. 32

    View of the Karabel Monument, just southeast of Kemalpaşa, showing Tarkasnawa, king of Mira, as identified in the accompanying inscription (©Archaeological Exploration of Sardis/President and Fellows of Harvard College)

The Transition to the Iron Age and the Arrival of the Lydians

Even though strictly historical evidence for Lydia is lacking for the period between the fall of the local Late Bronze Age kingdoms of Mira and the Seha River Land and the rise of the Lydian kingdom at Sardis, later Classical authors cited the existence of at least two dynasties of Lydian kings ruling before the Mermnad Dynasty initiated by Gyges, who is historically attested in Neo-Assyrian archives of the early to mid-seventh century BCE. Among many seemingly irreconcilable references to early Lydian dynasties are the stories of Herodotus, perhaps the simplest, from which we learn that the first Lydian dynasty was founded by Lydus, son of Atys, and that twenty-two generations of a second dynasty, known as the Heraclids, ruled for some 505 years until Gyges’ overthrow of the last Heraclid king, Candaules.51 Archaeological corroboration of a Heraclid conquest of Sardis in the early twelfth century BCE has been somewhat presumptuously cited in a level of destruction and imported Mycenaean pottery found at Sardis roughly dating to this time.52 The restricted exposure of these levels at Sardis makes it impossible to speak of site-wide events, however, and Sardis is only one of many sites in central western Anatolia at which have been found Mycenaean pottery of the fourteenth through twelfth centuries BCE. Further complicating the matter of these fabled early Lydian dynasties are the “Maeonians,” the name by which inhabitants of the middle Gediz River valley were known to Homer in the late eighth or early seventh century BCE.53 The Maeonians have been traditionally understood as early Lydians and Maeonia as a name for early Lydia. While largely credible, these understandings are based on little actual fact, bringing us to the question of what evidence there is for the arrival of Lydian-speaking populations in the area. Did the Maeonians speak Lydian, or did Lydian-speaking populations arrive with Gyges and the Mermnad Dynasty? Were Lydians already present in the MBA and LBA, or did they arrive in the aftermath of the LBA collapse?

The excavations at Sardis have provided only limited help in addressing these questions, yet they attest occupation by the LBA, if not earlier, some level of cultural continuity into the Early Iron Age, and high levels of cultural continuity between the Early Iron Age and the seventh- and sixth-century Mermnad period, when we can be assured of the presence of Lydian speakers.54 The continuity of early levels at Sardis, then, supports the general assumption that Lydian speakers arrived well before the Mermnads.

Archaeological evidence from central Lydia around Marmara Gölü, however, points to significant disruptions in local traditions during the transitional period following the LBA. The citadels associated with the previous ruling power(s) were all abandoned, yet a wider phenomenon of regional abandonment cannot be demonstrated with current evidence. When Iron Age sites appear in the early first millennium BCE, at any rate, they are associated with different settlement patterns, avoiding earlier mounds and citadels, and different production traditions, especially those associated with new pottery types and clay recipes matching later “Lydian” ceramic traditions, best known from Sardis, and also with different mudbrick manufacture. These changes might easily be explained by internal developments arising after the fall of the local LBA kingdom. Alternatively, they might support a view adhered to by most historical linguists, that Lydian speakers, having entered Anatolia at the latest by the early second millennium BCE (along with the speakers of other languages of the Anatolian sub-branch of Indo-European), migrated into Classical Lydia, perhaps from northwestern Anatolia in the Early Iron Age.55 Lydian speakers are presumed, in this view, to have replaced the previous Luwian-speaking populations associated with monuments like Akpınar and Karabel, and presumably also with the ruling kingdom of the Seha River Land.

Other linguistic evidence, however, highlights continuities between Bronze and Iron Age toponyms in the Seha River Land and Lydia,56 especially, and there remains the possibility that Lydian speakers were already present in central western Anatolia in the second millennium BCE. In such a scenario, the Luwian that appears in hieroglyphic form at Akpınar and Karabel may have been used exclusively for diplomatic correspondence and public monuments in at least a partially bilingual Luwian and Lydian environment. The territorial uniformity of the kingdoms of the Seha River Land and Lydia may also be sought in support of cultural, if not also political, continuity across the Bronze and Iron Ages.57

We are left with the unfortunate, but all too common, predicament that current evidence can be interpreted in multiple ways. The distinct changes in the archaeological record of central Lydia across the transition from the Late Bronze to Iron Age might be explained by the arrival of the Lydians from elsewhere in western Anatolia, just as they might be explained by local, social, economic, and political developments arising from the LBA collapse and the subsequent redevelopment of complex society. Exactly when Lydian-speaking populations arrived in Lydia, then, is an enigma that Lydian archaeology, or at least the archaeological record of Lydia as it is known today, seems unable to resolve. Again, despite the successes of surface surveys in illuminating the richness of Lydian prehistory, stratigraphic excavations are greatly needed to clarify further such cultural developments, both within and across prehistoric phases and periods, and especially across the transition from the LBA to the Iron Age. Whatever may have happened in the period between the abandonment of the LBA citadels and the appearance of Iron Age sites in central Lydia, and regardless of when Lydian speakers did actually arrive, one thing is clear: When state-level society did reappear, the central node of power had shifted from the area around Marmara Gölü, the probable core of the kingdom of the Seha River Land, to the south, across the Gediz River valley, and to Sardis, where culturally and linguistically distinct Lydian populations living under the historical Mermnad Dynasty thrived in their florescence of the middle of the first millennium BCE.