K-means集群方法在區域性考古遺址空間分佈研究上的應用：以晚商山東地區為例

晚商時代的中國山東地區在眾多商代政權控制區域中表現出諸多不同於差異之處，與中原區域相比，山東被商王朝納入勢力版圖的時間相對較晚，但相對於真正意義上的“邊疆”又呈現出較強的控制能力，特別是彰顯出明顯的集聚與分散並存的分佈特徵，此外在山東地區的一些重要商文化遺址之上出土的器物組合也與商王朝核心區域的中原安陽一帶不盡相同。在此情況之上則因發出了許多問題，其中最重要的就是，在經過一定的歷史過程之後，最終在晚商時代所形成的這一地區商文化遺址分布狀況，在多大程度上影響到了中原商王朝上層對於山東地區的治理？他們這些遺址彼此之間的關係和分佈模式又是如何的？這將是本論文將著重探討的核心問題。前輩學者透過研究豐富的出土甲骨文資料與傳世文獻，曾對商代晚期廣大的東方區域有過一定的認識，從商夷交錯乃至分野的族群格局到商王對其他部族的征伐路線，不勝枚舉。考古學者也確實能夠在山東地區發現一些中原商文化與本土岳石文化分佈地域上的差異。這無疑為進一步探討商王朝對晚商時期山東地區的控制、組織都提供了重要材料。作為一種廣泛應用在多個領域的統計方法，K-mean Clusters 群集分析的方式可以將雜亂無序分佈的數據根據彼此之間距離和預設的群集數目劃分成數個群集單位。在不同群集總數的層次之下檢視遺址分佈的情形，則可看出不同區域尺度之下的遺址分佈與組織層次。經過K-means 群集分析演算與GIS地圖可視化操作可以看出，晚商山東地區的商文化遺址分佈的多個特點與組織情況，首先在魯北區域的渤海灣地區有大量鹽業生產相關遺址，彰顯晚商王朝對於山東地區鹽的需要，同時在魯北東部地區大量遺址密集分佈，其中還包括了高規格的青州蘇埠屯，這似乎和鹽業生產、轉運的管控，以及軍事震攝膠東半島區域的“非商”族群珍珠們文化不無關聯。魯北西部中部東西走向河流沿線的遺址群則一定程度上具備為保護鹵鹽西運水道的職責。蘇埠屯遺址被認為屬於商王朝子姓貴族族氏與文獻記載以及前人研究中所稱的“族氏政治”不謀而合；在較為不重要的區域如泰沂山脈以南地區，以及岳石文化等非商族群聚居的地區，則僅有稀疏的分佈，且在群集劃分上並不具備穩定性，顯示出其不同遺址間關聯較弱，看似屬於較為鬆散的社會組織模式。因此可以看出，商王朝在晚商時期對於山東地區的經營處於族氏政治重點經營和控制重要資源以及交通線相結合的策略，而非全面而強力的控制。

關鍵字

空間分析； K-means 集群分析；山東；商代考古；中國

並列摘要

During the Late Shang period (ca. 1300-1046 BC) in the Shandong region of China, many differences appear amongst various Shang Dynasty polities. The timing of when Shandong was incorporated into the territory controlled by the Shang Dynasty was relatively late compared to the Central Plains region, but in contrast to what the true sense of what a “frontier” is taken to be, there was a strong capability to control the region, which is particularly manifested in the coexistence of agglomeration and dispersion in the distribution of sites. Furthermore, the assemblages of artifacts unearthed from a few important Shang Culture sites in the Shandong region are not completely the same as those from the core region of the Shang Dynasty near Anyang in the Central Plains. This circumstance gives rise to many questions, among which the most important is, in the end, how greatly did the situation of the distribution of Shang Culture sites in this region, after having gone through a certain historical process and then formed during the Late Shang period, impact the upper strata of the Central Plains Shang Dynasty’s governance on the Shandong region? What is the relationship and distribution patterns between these sites? This is the core issue on which this thesis focuses. Previous generations of scholars, who studied the rich data from the unearthed oracle bone records and the received historical texts, had their own understanding of the vast region in the east during the Late Shang Dynasty, looking at the interaction between Shang and Yi and the patterns of the boundaries between ethnic groups, as well as the routes of military campaigns by the Shang kings against other tribes—these are too numberous to enumerate. Archaeologists, as well, have indeed been able to discover some differences in the Shandong region in the distribition of Central Plains Shang Culture sites and indigenous Yueshi Culture sites. These past studies undoubtedly offer us important material to advance our investigations of the Shang Dynasty's control of the Shandong region during the Late Shang period and of organization. K-means Clusters analysis is a statistical method broadly applied across many fields. It can be used to divide irregularly distributed data points into a predetermined number of Clusters based upon their distances between each other. By viewing the changing distributions of the sites under the different number of Clusters, we can see the distribution and organization of sites under different regional scales. After calculating K-means Clusters and doing GIS map visualization, we can see many characteristics and the organization of the distribution of Shang Culture sites in the Late Shang period in the Shandong region. First, there are a large number of sites related to salt production in the Bohai Bay are of northern Shandong, highlighting the needs of the Shang Dynasty kings during the Late Shang period for salt from the Shandong region. At the same time, a large number of sites in the eastern part of northern Shandong are densely distributed, including the large-scale Subutun site in Qingzhou, which seems to be in control of salt production and trans-shipment, as well as it must having been related to military matters in the watch over the “non-Shang” groups of the Zhenzhumen Culture in the eastern Shandong Peninsula region. The group of sites along the east-west riverways in the western and central parts of northern Shandong had a certain degree of responsibility for protecting the waterways used for transporting salt toward the west. The Subutun site has been considered to belong to the noble Zi lineage of the Shang Dynasty kings, a view which also coincides with the literary record and what earlier research has called “lineage politics”. In less important areas such as the region south of the Taiyi Mountains, as well as in the areas occupied by the Yueshi Culture and other non-Shang groups, there is only a sparse distribution of sites, and there is no stability in clustering as the number of Clusters changes, showing that the connectedness between different sites is relatively weak: it seems that these sites belong to a relatively looser mode of social organization. Therefore, it can be seen that during the Late Shang period, the Shang Dynasty’s management of the Shandong region was through combined strategies stressing the control and management of major resources and transportation routes rather than comprehensive and strong control over the entire region.