Resolving the complex mixing history of ancient Chinese bronzes by Manifold Learning and a Bayesian Mixing Model. (March 2023)

• Record Type: Journal Article
• Authors: Sun, Zhenfei; Liu, Siran; Zhang, Ji; Chen, Kunlong; Kaufman, Brett
• Is Part Of: Journal of archaeological science. Volume 151(2023)
• Publication Date: 2023-03
• Subjects: Metal provenance -- Mixing -- Lead isotope -- Bayesian mixing model -- Zhou period China; Archaeology -- Periodicals; Archéologie -- Périodiques; 930.1
• DOI: 10.1016/j.jas.2023.105728
• Languages: English
• ISSNs: 0305-4403

Abstract

Abstract: Provenance of metals is a major theme in Bronze Age archaeology since it can help to reveal complex cultural and economic entanglements in ancient times. However, where complex societies with diversified trading relationships are concerned, identifying metal provenance has often proved to be challenging due to the frequent mixing of metals from different sources in antiquity. This research addresses this question by developing an innovative method for interpreting lead isotope data of bronze artefacts. Manifold learning and a Bayesian mixing model are combined to reconstruct quantitatively the contribution of metal sources to ancient bronzes. The methodology is employed to resolve the complex metal circulation system in the Zhou period (11th-3rd century BC) of China, and reveals a significant diachronic change of metal resources from North, Central, and South China. The North China metal sources were mainly employed in the Early Western Zhou period (1046–950 BC). In the following ages, the Yangtze River Valley and Qinling Mountains became the major metal sources for Zhou people. The Middle Spring and Autumn period (660 BC-560 BC) witnessed a major shift of dependence between these two sources, demonstrating a fundamental transformation in the metal circulation system. The South China metal sources were exploited throughout the entire Zhou period and probably associated with polymetallic deposits in the Nanling area. This research reveals the long-term patterns of metal exploration and mixing in the Zhou period of China, and also demonstrates the great potential this new methodology promises in addressing the complex metal mixing history in other cultural contexts. Highlights: Resolving the isotopic mixing issue in the provenance research of archaeological bronzes. A new methodology combining Manifold Learning and Bayesian Mixing Model to reconstruct the contribution of ore sources to bronze assemblages. Revealing the long-term pattern of metal exploration and mixing in the Zhou period of China.