拼裝社群：十三行遺址墓葬空間與統計分析

Abstract

十三行遺址（ca. 1800-550 cal. B.P.）位於淡水河河口，該遺址因出土目前台灣已知最早的煉鐵證據，使之成為台灣金屬器時代（ca. 2400-450 B.P.）最為重要的聚落遺址之一。過去針對十三行遺址之研究，多半著重該聚落遺址作為一社會體系（social system）的形貌、參與煉鐵技術之社會單位，以及出土玻璃珠所指向之海上交換網絡等等。筆者奠基於以上研究，乃嘗試從「拼裝」（assemblage）的概念出發，並結合考古學中之於社群（community）的討論，針對十三行出土墓葬進行空間與統計分析。具體而言，本文藉由相對定年並佐以絕對定年資料，首先將十行墓葬劃分為三個時期，緊接著運用GIS 軟體與R 語言，針對三時期之墓葬各自以Average Nearest Neighbor、Buffering、Global Moran’s I、Spearman 相關係數（Spearman’s correlation coefficient）與叢集分析（cluster analysis）等工具分析之。筆者進而能夠擬測墓葬空間叢集之規模，並於時間之縱深與空間之橫展中，探討陪葬品之數量與種類、死者年齡、性別與病理特徵之間的相關性與差異。根據分析結果，第一期（ca. 1800-1400 cal. B.P.）的社群規模較小且維繫較為鬆散；第二期（ca. 1400-1050 cal. B.P.）則可觀察到社群數量之劇增，筆者更發現，那些含有民族誌中所記載具特殊社會意義之物品（即青銅刀柄、陶容器、大量玻璃珠等）的墓葬，分散在不同社群中，顯示出相較前一期更為劇烈與多元的社群領域化（territorialization）過程；到了第三期（ca. 1050-550 cal. BP.），各社群乃透過縮短墓葬間距來實踐與維繫內部凝聚力，而筆者認為，此與陪葬品之數量大幅下降有關。筆者進一步懷疑，第一期至第二期的葬俗變化實與煉鐵技術之引入有關；至於第二期至第三期的轉變則可歸因於，十三行人原先所參與的東南亞海上交換網絡，逐漸受到漢人影響。總體而言，筆者定調比起「體系」，「拼裝」一詞應能更精確地描述十三行人所身處的社會——蓋因十三行並非一依循功能機制運作的單一體系，而是由多個異質的社群所匯聚而成，並拼裝成一具特定認同與內部凝聚力的動態整體。

The Shihsanhang (ca. 1800-550 cal. B.P.) site, located along the Tamsui River estuary in Taipei, northern Taiwan, is one of the most significant Metal Age (ca. 2400-450 B.P.) settlement sites in Taiwanese archaeology. As the oldest site known to yield remains of in situ iron production in Taiwan, past research on Shihsanhang has focused on how the settlement functioned as a social system, the social units that carried out ferrous metallurgical techniques, and the maritime exchange networks to which the abundance of glass beads unearthed at the site point. While grounding my study in these discussions, in this thesis, I attempt to theoretically employ a new approach drawn from the anthropological concept of the “assemblage” (which differs from the traditional archaeological term of assemblage), coupled with archaeological theorizations of community, in the spatial and statistical analyses of the mortuary remains at Shihsanhang. This usage of “assemblage” draws primarily from Delanda, Latour, and Bennet’s discussions. After dividing the mortuary remains into three chronological phases, I utilize the Average Nearest Neighbor, Buffering, Global Moran’s I, Spearman’s correlation coefficient, and cluster analysis tools with GIS software and the R coding language, to delineate spatial clusters of graves and detect correlations and differences among the quantity and quality of grave goods, age, sex, and pathological traits. According to the analytical results, in Phase 1 (ca. 1800-1400 cal. B.P.), communities were comparatively small and loosely maintained. In Phase 2 (ca. 1400-1050 cal. B.P.), I observe an increase in the number of communities, as well as the sporadic distribution of bronze knife handles, pottery, and relatively large quantities of glass beads, indicating intensified but varied community territorialization processes. Finally, in Phase 3 (ca. 1050-550 cal. B.P.), the amount of all the mentioned grave goods decreases, and communities respond to this material decline by burying their dead at shorter spatial intervals to maintain the internal cohesion of communities. I attribute the changes in mortuary practices between Phase 1 and Phase 2 to the introduction of iron production, and attribute the shifts in mortuary practices from Phase 2 to Phase 3 to the Shihsanhang residents’ transition from Southeast Asian exchange networks to those linked with the Han-Chinese. Based on these results, I propose that “assemblage” more accurately characterizes Shihsanhang than “system.” This is because, rather than one systematic social entity operating according to functional mechanisms, Shihsanhang appears to be assembled, that is, to be gathered from many heterogeneous communities into a dynamic compound involving an identity and internal cohesiveness.