Aided by radiocarbon and other dating techniques, scholars have assiduously sought to identify archaeologically excavated sites with the Hsia and Shang capitals discussed in such late written materials as the Chu-shu Chi-nien and Shih Chi, thereby chronicling their respective histories and attesting to the antiquity and continuity of Chinese civilization.43 According to these traditional writings, the Hsia and the Shang shifted their capitals a number of times for unspecified but presumably strategic and environmental reasons.44 Purported Hsia capitals include Hsia itself, Yang-ch’eng, Yüan, Lao-ch’iu, Hsi-ho, and Chen-hsün, most of which have been speculatively matched with various sites, including two that have already been discussed, Pien-hsien-wang and P’ing-liang-t’ai.
At present only Hsiang-fen T’ao-ssu, Teng-feng Wang-ch’eng-kang, P’ing-liang-t’ai, and Erh-li-t’ou have been even minimally justified by archaeological evidence, though Erh-li-t’ou is now officially, if not universally, recognized as the last capital of Chen-hsün. However, the fortified cities of T’ao-ssu, Wang-ch’eng-kang, Hsin-chai, Ta-shih-ku, and Tung-hsia-feng, all of which fall within the Hsia’s projected dates of 2070 to 1600 BCE, seem to provide somewhat concrete form to the previously elusive dynasty. Being located in peripheral areas in proximity to antagonists such as the Yüeh-shih and Tung Yi, several of these sites may have been secondary capitals or major strongpoints rather than true Hsia capitals,45 and even Pien-hsien-wang probably functioned as a Hsia bastion, then suddenly declined, perhaps coincident with the Hsia’s extinction.
Artifacts excavated from P’ing-liang-t’ai’s five main layers are variously dated, but the walls themselves may have been constructed as early as 2370 (± 175) BCE, placing them in the putative transitional period for Yao and Shun.46 However, insofar as the enclosed area amounts to only a third of many earlier Yangshao sites and probably had a population of fewer than a thousand, despite claims that it was a predynastic or early dynastic Hsia capital, at best it probably was merely an early military fortress.
T’ao-ssu at Hsiang-fen in Shanxi, the defining site for the increasingly discussed T’ao-ssu culture, has sometimes been suggested as the locus for Yao, Shun, and Yü,47 as well as the first Lungshan city where emblematic evidence of Hsia culture can be discerned.48 Located in an area that has long been identified as core Hsia domain as well as Emperor T’ang’s legendary bastion, the overall site encompasses some three million square meters. The densely populated society was already marked by clear class distinctions, and the graves of esteemed males are distinguished by the inclusion of weapons and other symbols of martial status or achievement. For example, 111 stone arrowheads (out of 178 objects) were interred in one grave; others include not only a small number of arrowheads, but also stone axes.49 The presence of some thirty bodies whose appearance indicates they died violently, coupled with the essential abandonment of the walls even while the site flourished, has prompted the suggestion that Hsiang-fen was the site of Ch’i’s capital after their internal conflict.50
Another early fortified town that has been tentatively identified with the Hsia dynasty, the aptly named Wang-ch’eng-kang—“king’s (wang) city (ch’eng)”—was strategically situated on a well-chosen hillock (kang) just a couple of hundred meters to the west of the Wu-tu River and about 600 meters north of the Ying River. Excavated in two stages approximately two decades apart, it consists of a small citadel in the northeast corner of a much larger, roughly 348,000-square-meter enclosure. The nature of this double enclosure has prompted assertions that Wang-ch’eng-kang represents an early, even the earliest, manifestation of the Chinese royal city.51
The inner or original fortified enclosure of roughly 20,000 square meters consisted of two essentially square citadels of approximately 90 to 100 meters per side, arrayed on an east-west line with a common center wall. The corners were square, and there seems to have been a gate at the eastern corner of the western section’s southern wall, abutting the common center wall. Unfortunately, extensive flood damage, leveling, and agricultural exploitation over the millennia have almost obliterated the walls, preventing an accurate estimation of their original height.52
The walls themselves must have been fairly narrow, perhaps only 5 meters wide, because the carefully prepared, 2-meter-deep foundations have a top opening of 4.4 meters but taper down to a 2.54-meter-wide base. The fairly well-leveled tamped earth layers, generally 10 to 20 centimeters thick but as thin as 6 centimeters, are clearly demarked by an intervening 1-centimeter layer of coarse sand that was intended to prevent the tamping tools from adhering to the soil being compacted. Considerable variations in the size, shape, and depth of the pounding marks indicate that river stones may have been employed, consistent with early practice. The site’s prolific middle Lungshan artifacts, coupled with this tamping technique, suggest that the walls were built in the mid- to late Lungshan period, before the emergence of the Erh-li-t’ou phase.
The walls of the greater Lungshan enclosure, one of the largest yet discovered in Henan, nearly define a square that runs some 600 meters east to west and 580 meters north to south. The northern wall, which displays the laddered cross-section typical of the period, was constructed on an excavated foundation using the same technique as the inner citadel. From a top width of 6.8 meters it expands to 12.4 meters at the base, retains a remnant height of 1.12 meters, and consists of ten layers of tamped yellowish sand that vary from 4 to 28 centimeters in thickness after having been compacted with 5- to 7-millimeter-diameter tools.
Moat segments of approximately 620 and 600 meters just 4 meters outside the walls protected the northern and western sides, thereby completely encircling it with a water barrier in conjunction with the Wu-tu and Ying rivers. These external moats were refurbished at least twice during the city’s occupation. As originally constructed in the Lungshan period, they were about 15 meters wide and tapered down to 10.4 meters at the very significant depth of about 5.2 meters. During the Erh-li-t’ou phase they were about 15.8 meters wide at the top and an augmented 7 meters deep, but only 2 meters wide at the bottom; and in the Spring and Autumn they had an expanse of 16 meters, a base width of 10.8 meters, but an average depth of only 3.6 meters.
Radiocarbon dating shows that although the area was intermittently populated from at least the P’ei-li-kang, the earliest wall-building activity dates roughly to the first half of the twenty-first century BCE, slightly later than T’ao-ssu and the era frequently identified with the Hsia’s inception.53The smaller compound predates the larger, the latter having destroyed a portion of the former during its construction, but both probably fall into the late Lungshan era. Recovered artifacts and alterations to the moat in both the Erh-li-t’ou and Spring and Autumn periods indicate that Wang-ch’eng-kang was occupied almost continuously until the Warring States period due to its strategic location and the natural fertility of the greater surroundings.
As later historical writings associate the semilegendary progenitors of the Hsia—Yao, Shun, Yü, and Yü’s father Kun—with this area, and the site has historically been known as the “king’s city,” it has been suggested that the smaller citadel was constructed by Kun, the “first wall builder.” The greater city, which would befit the Hsia’s first ruler and have required a massive effort to construct and thus reflect the rise of centralized authority, is therefore envisioned as having been Yü’s first capital of Yang-ch’eng.54 The segmented double quarters would have delimited the royal quarters, the greater site would have encompassed a population of perhaps 50,000 people, and the fortifications coupled with the exploitation of the two nearby rivers would have provided considerable protection.
If, however nebulously, Wang-ch’eng-kang can be identified with Yü, the fortified town located at Hsin-chai in Hsin-mi, Henan merits possible consideration as his successor’s capital of Huang-t’ai.55 Recent radiocarbon dating confirms that Hsin-chai stands at a transition point between the late Lungshan and Erh-li-t’ou and might therefore be considered incipient Erh-li-t’ou or Hsia.56 Located on the north bank of the Wei River (now the Shuang-chi) and flanked by the Wu-ting and Ch’ih-chien (or Sheng-shou) rivers on the west and east respectively, it was nearly surrounded and thus naturally protected by three rivers as well as a preexisting ditch.
Nevertheless, commencing in the late Lungshan period, Hsin-chai underwent several stages of complex fortification efforts when the first walls were constructed and a moat excavated to the exterior, the size of the former being about 3 meters in height and 4.5 meters in width, with a narrow moat of about 2.7 meters that tapered to 1.3 meters at the bottom and ranged in depth from a very shallow 0.34 to a relatively functional 2.3 meters.57 The walls were improved to some degree early in the Hsin-chai period, then rebuilt with rammed earth construction techniques further to the exterior following floods that apparently filled the previous moat. These new walls reached the more significant height of just under 5 meters in some sections, though others dipped as low as perhaps 2 meters. A new moat with roughly double the width and depth of the previous one was excavated for augmented protection, and evidence of an interior moat has also been reported. However, after some minor revisions and improvements in the late period, all the walls and even the moat were inexplicably destroyed in the Erh-li-t’ou period, being replaced by a new 1,500-meter protective waterway said to average 6 to 14 meters wide and 3 to 4 meters deep.
Hsin-chai is an intriguing site that well illustrates not just the continued occupation of a strategic location, but also the deliberate process of expanding a fortified town by laboriously filling previous moats and ditches before erecting walls on them with soil excavated for a new exterior moat. The ruling authorities must have had considerable power and a strong impetus to compel the people to undertake what can only be considered a double process instead of simply rerouting a portion of the external moat and unilaterally expanding one side or adding a second section rather than uniformly reworking the entire encirclement in all directions. Excluding the area of the walls and moat, the interior at the end of the Hsin-chai phase encompassed about 700,000 square meters, the largest late Lungshan/Hsin-chai site known and certainly a reasonable candidate for a pre-Erh-li-t’ou Hsia capital.
In addition, an unusual site attributed to the Hsin-chai phase of early Hsia culture has been uncovered somewhat outside the historically identified Hsia focal area to the north above the Sung-shan Wan-an Mountains.58 Its unique defensive technology—four square, concentric ditches marked by rounded corners of somewhat varying width but fairly uniform spacing—provides incontrovertible evidence that the occupants intended to project power into the hinterland and protect the approach to the Hsia heartland. The outermost ditch spans some twelve meters, ranges from a startling five to six meters deep, and is characterized by laddered sides. A single defensible roadway penetrates the series of ditches on the southwest, away from the prime direction of threat.
Although considerable skepticism, not to mention vehement opposition and rejection, continues to beset the identification of all the previous sites with pre-Hsia culture, Erh-li-t’ou at Yen-shih has been more generally acclaimed as the Hsia’s last capital and thus presumably Chen-hsün, where T’ai Kang, Yi, and Chieh all dwelled.59 Numerous problems remain, including the apparent discontinuity between the cultural complex visible at Erh-li-t’ou and the characteristics of late Lunghsan culture, as well as the apparently heterogeneous nature of the populace, an admixture of clans and peoples (as would be expected from integrating conquered groups) rather than a single Hsia entity.60
Although backed by a small range of hills, Erh-li-t’ou was situated in the eastern part of the Loyang plains between the eastward-flowing Luo and Yi, two rivers that could provide ample water and constituted natural defensive barriers. Some 2.5 kilometers from east to west and 1.5 kilometers from north to south, the site encompassed structures that ranged from palatial to ordinary and dedicated areas such as workshops and burial grounds, all of which preserve evidence of a sophisticated culture. Surprisingly, the site lacks the massive earthen fortifications that define Lungshan and previous Erh-li-t’ou manifestations, only the royal city having been protected by a limited, double-walled compound. However, Erh-li-t’ou was located in the midst of a strategically advantageous area populated by late Lungshan, Yangshao, and pre-Erh-li-kang towns. Several capitals and otherwise important fortified cities would also be constructed over the millennia within 25 kilometers, including Han-Wei Ku-ch’eng, Ch’eng-chou, and Luo-yang.
Ongoing investigations continue to compel modifications of Erh-li-t’ou’s envisioned nature, role, and significance. From the artifacts already recovered it is clear that the inhabitants pursued all the highly skilled occupations appropriate to a royal capital and that trade, tribute, and conquest resulted in the acquisition of diverse, nonindigenous products. Agriculture flourished in the surrounding, fecund earth; domesticated animals including cattle, sheep, pigs, and dogs were all raised; and fishing augmented the basic food supply. The specialized handicraft industries fabricated sophisticated items from a wide variety of materials, including bone, stone, clay, jade, metal, lacquer, and plant fibers. Weaving was well advanced, and wheel-thrown ceramics had appeared. Liquor, whose liberal use would ironically be condemned by the Shang before reputedly coming to plague its own rulers, was also fermented and enjoyed, as attested by bronze drinking vessels.
An early bronze foundry containing copper fragments and molds has been uncovered, and numerous other metal artifacts that provide evidence of significant productive capacity have been found. Among the items produced from copper and high copper-bronze alloys were fishhooks, jewelry, knives, axes, and comparatively simple cauldrons intended for both everyday and ritual use. Weapons must have been highly valued, because many were fashioned from bronze and numerous bronze arrowheads have been recovered. (The use of bronze rather than stone for such single-use, irretrievably lost items has been interpreted as evidence of access to an ample copper supply and extensive mining activity.) Finely worked symbolic jade weapons such as axes, knives, and halberds, which must have been intended to mark the military prowess and authority of tomb and grave occupants, indicate the importance attached to martial values in the Hsia.
The segmented, clearly defined royal quarters of 108,000 square meters were not only well planned and carefully executed, but also protected by a minimal fortified enclosure in the shape of a slightly distorted rectangle.61 As projected from existing remnants, the eastern wall was originally 378 meters long, the western 359, the southern 295, and the northern 292. The eastern and northern walls were erected on untreated ground, the western and southern on shallow foundation trenches. Being a maximum of 3 meters at the base, 1.8 to 2.3 meters at the top, and 0.75 meter high, the walls were more symbolic than defensive, apparently intended to minimally demark and thus conceptually segregate the royal domain rather than protect it against incursion. Composed simply of reddish brown soil tamped into layers 4 to 12 centimeters thick, they were erected in Erh-li-t’ou’s second period and continuously maintained thereafter.62
Another similarly constructed, 2-meter-wide, tamped earth wall has recently been discovered about 18 to 19 meters beyond the southern side. Running parallel to the south wall, the 200 meters so far uncovered were constructed on a slightly wider foundation trench and may well have been part of a fully encompassing second enclosure. (A few wall fragments have recently been reported on the western side as well.) Somewhat surprisingly, the wall apparently dates to the middle or later fourth period and thus presumably to after the Shang conquest. Apparently the Shang simply occupied the old royal city and reinforced the perimeter with a typical double enclosure, creating a limited bulwark against the indigenous populace and external inhabitants.
Because capitals have traditionally been understood as defined by massive walls, segmented royal quarters, palatial structures, and ancestral temples, the absence of perimeter walls at Erh-li-t’ou has long prompted questions about the city’s role.63 However, apart from questions about the applicability of this assumption, it was not impossible to have a secure site without them, provided that the area benefited from natural strategic defenses and external strongpoints could be erected to control access routes. The recently discovered citadel at Ta-shih-ku in Hsing-yang, which was probably intended to project Hsia power at the limits of its domain, no doubt falls into this category of forward defenses.64
Located about 22 kilometers northwest of Cheng-chou and 13 kilometers south of the confluence of the Chi and Yellow rivers, Ta-shih-ku is roughly 70 kilometers east of Erh-li-t’ou. Mt. Mang lies just to the north, numerous small rivers and streams are found in the general area, and the Luo River now passes through the western part of the enclosure’s remnants, but sand deposits and the 30-degree declination of the northwest corner wall suggest it probably once flowed somewhat outside the first walls. In the late Hsia, Ta-shih-ku would have occupied a forward position along the historically projected interaction zone with the nascent Shang cultures in the region between the Yellow and Chi rivers.65 However, it was too close to Erh-li-t’ou to take seriously current claims that it was King T’ang’s preconquest capital of Po.
Excavations have found extensive evidence of four Erh-li-t’ou cultural layers at Ta-shih-ku. The second and third layers show a flourishing site but diminished occupation thereafter, as well as the sudden intrusion of lower Erk-li-kang elements. This suggests that Shang forces probably occupied the city between Erh-li-t’ou’s third and fourth periods, coincident with the widespread eclipse of Erh-li-t’ou culture in other statelets and at Teng-hsia-feng and Meng-chuang. Ta-shih-ku’s loss must have been the result of a penetrating incursion, one that would have enabled the Shang to subsequently mount a direct strike on Erh-li-t’ou by passing between the Hsia allies of Ku and K’un-wu, one to the southeast, the other somewhat to the northwest.66 The interval between the city’s conquest and the Hsia’s final subjugation would therefore have been quite short.
The strongly fortified, rectangular city encompassed roughly 510,000 square meters, about one-eighth the size of Erh-li-t’ou. It was protected by a continuous wall that was further augmented by an external moat that apparently exploited a part of the Suo River’s course. (The river flowed along the lower southwestern portion of the southern wall before turning up through the city and then wrapping around well to the outside of the city complex.) Although somewhat varying dimensions have been reported, the overall site appears to have been just under 1,000 meters from east to west and 600 north to south. (A ditch in the north presently extends about 980 meters, and the wall on the south is about 950 meters. However, the last third or so of the western part tapers down in trapezoidal fashion to just 300 meters in extent.)
Even though the walls were apparently undamaged, postconquest the Shang embarked on an extensive building program intended to ensure the bastion’s security amid enemy terrain. Key measures included a deep ditch on the interior edge of the original moat that considerably widened it and added the surprise of sudden depth. Although these renovations have obscured the exact width in many places, the original Hsia moat probably presented a watery expanse some 5 to 9 meters wide, marked by a functional depth of about 3 meters and a very uneven bottom contour about 3.25 meters wide. Subsequent Shang excavations then expanded it right up to the wall’s new foundation, for a width of between 13 and 14.55 meters, including a sharply defined innermost portion of 1.5 meters that had a formidable depth of 4 to 6.8 meters.
The original Hsia walls consisted of well-defined layers that had been carefully erected on leveled ground. Though the expanded moat cut into their exterior, overall the Shang’s layered additions still augmented them. The remnant walls presently protrude slightly less than a meter above the terrain but probably reached at least 3 meters in antiquity. Marked by the usual trapezoidal shape, they are about 7 meters wide at the top and 13 at the base. Some parts of the walls lie on alluvial sand deposits, others span refilled areas, and the entire configuration conforms to the terrain’s features.
Generally composed from seven discernible layers of fairly uniform tamped earth ranging from 4 to 10 centimeters in thickness, the walls utilized a complex mix of soils apparently excavated from different layers of the moat.67 The interior wall is somewhat unusual in having been erected by overlayering a sharp upward protuberance with a wider section that overcapped it and even extended out onto a new inner foundation in some areas before three outer sections were added. All the walls display a sloping contour indicative of a laddered construction created by staggering small boards with height increases.