1. Introduction Stone adze blades are so ubiquitous in the Pacific that they have always been central to the work of archaeologists. Polynesian adze heads were often viewed as convenient “cultural fossils” displaying stylistic features that could be used as chronological markers (e.g. Davidson 1984; Duff 1977; Suggs 1961) or to infer inter-island relationships (Duff 1960; Emory 1968; Sinoto 1970). The typological approach developed throughout the 20th century thus proved useful for understanding the archaeological diversity in the Pacific. Yet, it rarely took into account the environmental or technical contexts involved in the production of these artefacts. A broader technological turn began in the 1970’s (Cleghorn 1982; 1986; Leach 1980; 1984), which has led to better understanding of the technical and economic dimensions involved in Polynesian stone tool production. In this paper, I propose an integrated method to the study of stone tool technology and socio- economic evolution in Polynesian chiefdoms. My approach draws on the ‘research program’ which was set up by a group of French anthropologists (Cresswell 1976; Lemonnier 1983; 1986) to emphasize the systemic and multi-scalar dimensions of technical activities and to identify the fundamental role of techniques and production activities in the performance and evolution of societies. The work presented here was conducted during my PhD research, and therefore represents a first step in a wider research program that will be address with the dynamics of stone tool production at the inter-site scale, in different Polynesian islands. 2. Background and Methodology As a first step, each rock material was assigned a specific geological feature using a set of different geochemical analyses (Hermann et al. 2012; 2016). I used these results, combined with macroscopic identification of each artefacts, to track the spatial distribution of production processes (“chaînes opératoires”) within two sites dating from the early 13th to the late 15th century AD and considered as part of the same “technical transect” (Coupaye 2015) on the northern coast of Tupua’i island (Austral archipelago, French Polynesia). Both sites were discovered within the main pre- Contact district of the island, named Toerauetoru (Aitken 1930: 31-32): the Tanataetea site consists of a quarry and several workshops where basanite prisms have been quarried and transformed in great quantity, and the Atiahara site is a domestic occupation involving small thatched houses known from the ethnographical period as ‘arepota’ata (Hermann et al. 2016). I propose a thorough description of technological patterns in the making of adze heads in these two sites, not only through the description of finished products, but also through the identification of other artefacts including discarded preforms, roughouts, and other flake wastes, each representing Journal of Lithic Studies (2017) vol. 4, nr. 2, p. xx-xx doi:10.2218/jls.v4i2.2546 A. Hermann 35 combined sets of gestures constitutive of each sequence in the overall process of production. For every sequence, I investigated four main parameters interacting with one another: the nature of the raw material selected and its physical properties (fine or coarse-grained, natural flaws, etc.), the individual involved (including the inherited and acquired know-how, the technical traditions, economic choices, etc.), the tools and techniques performed (physical actions, mechanical procedures, etc.), and the spatial-temporal dimension of the process (concentrated in one site or segmented in space). These intrinsic properties are subject to change in different processes, however in this case they could be described through direct macroscopic observations, and interpreted thanks to experimental tests previously performed with the help of archaeologist and experimenter Florent Le Mené. The extrinsic properties of production processes are eventually be inferred. This last step of the analysis regarding the scarcity or uneven distribution of resources, the social environment, and the social status of craftsmen, provides the greatest insight into the evolution of the traditional Polynesian chiefdoms. 3. Results In East Polynesian islands stone adzes were produced by both specialists and non-specialists; the assemblages from Tupua’i show very clear differences in terms of technical skills, production intensity, and formal standards of the finished products, as an example of this technical split. While the craftsmen working in the Tanataetea workshops were producing big, standardised adzes with fine knapping techniques and were displaying anticipation as well as good risk management skills; the stone knappers of the domestic site at Atiahara only managed to produce only small flake adzes and appear to have maintained and recycled adzes produced elsewhere (including in the Tanataetea workshops). They therefore did not display any skills involving adapted percussion techniques or reaction facing flaking accidents, such as hinging or plunging fractures. At a diachronic level, the technological investigation suggests a significant evolution in the organisation of stone adze production in Tupua’i during the 14th century AD, with an increase of specialised production, as visible in the Tanataetea workshops, towards other sites from the same local chiefdom. The limited distribution of good-quality material in non-specialised sites like the domestic occupations of Atiahara also implies an increase in political control over key resources. 4. Discussion and conclusion The general development of craft specialisation in the limited context of Tupua’i echoes previous observations from other parts of Polynesia (Cleghorn 1982; 1986; Leach 1993; Winterhoff 2007), and further reinforce the hypothesis of a strong correlation between technical specialisation and social complexity (Brun et al. 2006). This interpretation also fits the view of an increasing hierarchy among Polynesian chiefdoms where social and religious elites took control over different groups of specialists during the first half of the second millennium AD (Kirch 1984). I propose that the emerging view in the evolution of Polynesian craft production can also be explained in Sahlins’ terms of the “inflexion of the domestic mode of production” (Sahlins 1976). This model predicts that the intensification of production in traditional chiefdoms is driven by new economic choices prioritising centralisation and collaboration between households. Therefore, I propose that a heuristic approach of viewing material culture through the technological approach of chaînes opératoires is a promising method for investigating these patterns on the longue durée and across different Polynesian regions.