-
Revisiting Rapa Nui Matā
Based on a use-wear and residue analysis of a collection of 12 matā in the Australian Museum, Sydney, we question the value of relying on tool shape as an adequate indication of past use. Although the tools in this collection were used for a broad range of tasks, including plant processing, wood, shell and bone working, and cutting and piercing of flesh or skin, some may have been used in interpersonal conflict. The study illustrates the value of museum ethnographic collections for understanding past tool use.
Rapa Nui, matā, use-wear, residue
Introduction
Recent scholarship on Rapa Nui has mounted an impressive set of data contesting the role that societal conflict played in cultural change prior to European contact (e.g., Lipo & Hunt 2009; Hunt & Lipo 2011). One key part of the critique of previous theories revolves around hafted obsidian artifacts known as matā, often stated to have been spear points used in internecine warfare on the island. In a recent article, Lipo et al. (2016) tested three hypotheses concerning the relationship between the shape of matā and their possible use as "weapons of war" and found they were not supported. The lack of a "spear-like shape" on the majority of the 423 artifacts in their sample combined with the considerable variability in shape led to their conclusion that "the evidence to support matā as lethal weapons of warfare does not exist" (Lipo et al. 2016:184). Based on a new study of use-wear, we identify problems with accepting that matā were never used in warfare.
There appear to be three main problems with attempts by previous scholars to understand how matā were used. Firstly, the classification of an artifact as a matā is based solely on the presence of a retouched stem. Surprisingly, the blade portion of the tool, where the working edge is located, is not part of the definition. It is not clear why all tools with a "handle" should share a function. Secondly, one can question whether stone tool morphology is an accurate guide to how the artifacts were actually used. For example, Robertson et al. (2009) demonstrate convincingly that Australian-backed artifacts with identical forms served a wide range of uses. We argue that a more appropriate way to reconstruct how matā were really used is to employ well-understood use-wear methods that combine a program of replication with high-powered microscopy. The resulting determinations inspire high levels of confidence because they are based on an extensive number of experiments using obsidian and the wear traces have been confirmed by multiple researchers (e.g., Hurcombe 1992; Kononenko 2011; Van Gijn 2014; Stemp 2016a, 2016b). Unfortunately, this technique has largely been overlooked in discussions about the function [End Page 1] of the matā, despite previous use-wear studies of 12 artifacts that identified plant processing as the primary use (Church & Rigney 1994; Church & Ellis 1996; Church 1998). Given these studies, Lipo et al. (2016:184) do conclude that there was more than one function for matā, but they disregard the possibility of personal violence. Finally, variation in shape may not necessarily be determined by intended function. Variability may actually have been a desired feature of tools if they played a role in expressing individual identity.
In this study, we begin the process of unpacking the highly variable category encompassed by Rapa Nui matā by undertaking a use-wear and residue study of 12 matā from the Australian Museum ethnographic collection. Despite the small sample size, these new use-wear findings question the value of relying on morphometric variation as a sole and accurate measure of how tools were used or not used as proposed by Lipo et al. (2016). Our study also illustrates both the values and limitations of using historic ethnographic collections for understanding ancient tool use.
Methods and Materials
The majority of the Australian Museum matā were obtained in the nineteenth century, probably as purchases from local residents on Rapa Nui, although a recent acquisition is claimed to have been recovered from a cave on the island (Table 1). The metric data in Table 2 together with Figures 1–7 demonstrate that this small sample encompasses a broad range of shapes paralleling the large morphometric variability reported by Lipo et al. (2016), although the average size is larger, as might be expected for a selection made for sale to private collectors (Figure 1).
Acquisition of Australian Museum matā.
In order to remove grease from recent handling and loosely adhering contaminants such as dust acquired from over a century of sitting on a museum shelf, the artifacts were immersed for 3–5 min in an ultrasonic bath containing warm water to which a few drops of nonabrasive liquid detergent were added, air dried, and wiped gently with diluted ethanol (30%) using a Kimwipe ©. They were examined under a digital microscope (Dino-LiteTM AM413ZT) with reflected light and magnifications from 10× to 100× and then with an Olympus BX60M metallurgical microscope fitted with vertical incident and transmitted light sources, bright and dark field illuminations, and cross-polarizing filters under magnifications ranging from 50× to 1000×. An Olympus DP72 camera with Soft Imaging System GmbH was used to make a photographic record of use-wear traces and residues. [End Page 2]
Australian Museum matā metrics.
[End Page 3]
Matā from Australian Museum: a) E30737 used for piercing and cutting soft pliable material, for example, flesh or skin; b) E30735 used for cutting a soft pliable material, for example, flesh or skin; and c) E30736 used for scraping bone or shell. Scale bars are 1cm.
The determination of tool use was based on characteristic wear attributes shown in previous experimental studies to be especially valuable for reconstructing wear traces on obsidian artifacts. These include edge scarring, edge rounding, attrition, striations, polish, and residues (e.g., Hurcombe 1992; Aoyama 1995; Kononenko 2011; Stemp 2016c). [End Page 4] A comparison of the wear attributes on the artifacts with experimental use-wear data informed by ethnographic sources also assisted identification of the mode of use and the material worked, following procedures developed for obsidian (e.g., Kamminga 1982; Fullagar 1986; Lewenstein 1987; Hurcombe 1992; Kononenko 2011; Kononenko et al. 2015, 2016; Stemp 2016a, 2016b, 2016c).
Results
Extensive taphonomic damage was identified. First, it was not possible to remove all the modern surface residues. In particular, a lacquer-like substance coated a large portion of the surface of six tools. It is not known whether it was applied in Rapa Nui or by collectors or museum curators to enhance or preserve the tool surface. Secondly, postdepositional damage by trampling and/or during transport and handling by collectors or within the museum obscured much of the tool surfaces. Despite these difficulties, it was still possible to locate a sufficient number of spots where wear traces were preserved. These enabled a reconstruction of mode of use (cutting, scraping, etc.), worked material, and presence/ absence of hafting with high confidence in many cases. Interpretations of the use-wear traces are summarized in Table 3. The variable states on which these are based and the confidence levels of our interpretations are reported in full in the Appendix.
Summary of use-wear and residue interpretations.
As expected given the previous use-wear research (Church & Rigney 1994; Church & Ellis 1996; Church 1998), some of the Australian Museum matā with wear traces were used to chop, whittle, or scrape woody plants (25%) or soft plant material (leaves, green stems, grasses) (12.5%) (Table 3: Figures 2 & 3) E30736 illustrates a new use for matā. Several spots with severe attrition and rounding as well as slightly diagonal thin striations [End Page 5] are comparable with experimentally replicated wear patterns on tools used to scrape marine shell (Kononenko 2011: Plates 96A-B, 97A-B). It therefore seems likely that this tool was used for a scraping a hard material, probably bone or shell (Figure 4).
Soft, siliceous woody plant: E65154. a) location of images, scale bar is 1cm; b) point 1, edge rounding and polish; c) point 2, edge rounding, polish, and striations; arrow indicates postdepositional scars with freshly flaked surface; d) point 3, scars, polish, striations, and surface attrition; e) point 4, attrition and striations on ridge of stem; and f) point 7, crossed striations on surface of stem. Use-wear features are indicated by arrows.
Surprisingly, given the Lipo et al. (2016) predictions based on comparisons with extensive experimental data, the most common use material identified, comprising half of [End Page 6]
Soft, siliceous, and resinous plant or grasses: E30739. a) location of images, scale bar is 1cm; b) point 1, scars, edge rounding, polish, striations, and residues; c) point 2, polish and striations on elevated points of surface topography; d) point 3, scars, resin-like residues within scar and spot of elevated surface with polish and dense striations; e) point 3, polish, striations, and residues within scar; and f) point 4, rough striations and residues on ridge of stem. Use-wear features are indicated by arrows.
the Australian Museum matā with preserved wear traces, is a soft, pliable substance most likely to have been flesh or skin (Hurcombe 1992:43–44; Aoyama 1995; Kononenko 2011:32–33; Kononenko 2012; Stemp & Awe 2014; Stemp 2016a, 2016b, 2016c). Within this group, E30734 was probably used for cutting/slicing fish (Figure 5). This interpretation is based on the continuous microscars and light to developed polish. Key [End Page 7] attributes that match experimental data used to process fish include spots of light, pronounced attrition on the edge, and relatively numerous fine, isolated, long, and short striations on the ventral face of the edge (Hurcombe 1992:44–45; Kononenko 2011:34, Plate 90).
Shell or bone: E30736. a) location of images, scale bar is 1cm; b) point 1, postdepositional damage and residues; c) point 2, edge attrition and striations; d) point 2, edge attrition and striations; e) point 3, striations on stem; and f) point 3, attrition, striations, and resin-like residues on scar ridges of stem. Use-wear features are indicated by arrows.
In considering the Lipo et al. (2016) conclusions about matā, E30737 is particularly important because the use traces show that it was used with some force to pierce a substance [End Page 8]
Soft, elastic material, possible fish: E30734. a) location of images, scale bar is 1 cm; b) point 1, edge rounding, attrition, polish, and fine striations; c) point 2, microscars, edge rounding, polish, and fine striations; d) point 3, attrition and striations on stem; and e) point 4, resinous residues on stem. Use-wear features are indicated by arrows.
that was probably flesh, and at the same time, encountered a hard material that caused impact damage (Figure 6). A small elongate and narrow burin-like spall at the tip of the point, which resembles the "diagnostic impact-fractures" defined by Hutchings (2016), [End Page 9]
Soft, elastic material such as meat, flesh, skin: E30737. a) location of images, scale bar is 1 cm; b) point 1, scars, edge rounding, fine striations, and polish on elevated point of surface topography; c) point 2, scars and fine striations within scars on tip of tool; d) point 3, scars, attrition, edge rounding, and fine striations; e) point 4, attrition, striations, and resinous residues; and f) point 4, starch grains within resinous residues. Use-wear features are indicated by arrows.
occurs in association with diagnostic wear traces that include continuous microscars and small scars with feather and bending terminations, light and some moderate edge-rounding, small pockets of well-isolated fine striations, light to developed polish. Together these traces indicate a penetrating mode of use as in the case of a spear (for further examples see articles by Fullagar, Hutchings, Iovita et al., Yaroshevich et al. in Iovita & Sano 2016). [End Page 10] The occurrence of this "snapping spall" (Kamminga 1982:45) is also commonly observed on experimental obsidian tools with very thin fragile tips used to puncture skin with a "pushing" action, as recorded by Kononenko et al. (2016) and Stemp (2016a). As E30737 had a strong thick tip, we assume the impact fracture was caused by contact with a hard substance.
Eight of the tools preserved use-wear traces on the retouched stem. Based on similarity to experimental data (Kononenko et al. 2015), they are interpreted as having been hafted on a wooden handle of some type. Key attributes observed include rough, flattened attrition and striations on the ridges located between flake scars formed by retouch on the stem, together with numerous patches of embedded resinous residues that occasionally preserved starch granules (e.g., Figures 2–6). The ubiquity of hafting for the matā is significant given the descriptions of "lances or spears" in the historic accounts quoted by Lipo et al. (2016:174).
Given the small sample size, the actual proportions of the various categories of tool use may not be representative of the general category of matā uses, but the results from the Australian Museum collection do confirm that this tool group was used in a very broad range of activities and, in particular, that a significant proportion were used to cut and pierce flesh and/or skin. These results therefore indicate that some matā may have been used in interpersonal conflict both as spears and, perhaps, as slashing tools. Others might have been used for scarification or tattooing in ritual practices or medical practices, as speculated by Lipo et al. (2016:184).
Form Versus Function
Given the high degree of variability in the morphology of matā, as documented by Lipo et al. (2016), we are not very surprised to find that use-wear analysis shows they had been employed in a wide range of tasks including plant processing, wood, shell and bone working, and cutting and piercing of flesh or skin. Following Lipo et al. (2016), many of the uses would not be subject to selective pressure, because there would have been no deleterious consequences if the tasks were not completed in a timely or efficient manner and so their shapes were not standardized. The root problem, however, is not that "there appear to have been no systematic performance requirements that influenced blade shape," as argued by Lipo et al. (2016:184), but that the definition of this group was never based on function in the first place. The category of matā as applied by archaeologists is constituted solely by the presence of a stem. The assumption that a tool type defined only by having a handle would genuinely reflect a single-purpose reflects flawed logic. In our view, the Lipo et al. (2016) test failed from the first assumptions because there is no satisfactory argument that matā ever represented a coherent functional class. Previous use-wear analyses, now supported by our studies, show that the variety of artifacts defined as matā was never a functional class and includes a mix of domestic tools and weapons.
Where did the idea that matā were a single-purpose functional group come from? Modern observations may be the source, particularly Cook's March 17, 1774 statement concerning spears "about 6 or 8 feet long which are pointed at one end with pieces of black fli[n]t" (Beaglehole 1969:352). Routledge (1919:223) collected 14 names for matā, but these refer to shape and not to how they were used. Similarly, Skinner classified matā into six types, each with a traditional Rapanui name (Métraux 1940:166–167), none of which demarcate function. [End Page 11]
Does the finding that some matā were used to pierce flesh indicate the presence of systematic warfare on Rapa Nui? Only additional use-wear studies based on large sample sizes could adequately address that question. Our data indicate that some matā might have been used in interpersonal conflict, but sample sizes are not yet adequate for determining the incidence of warfare on Rapa Nui. Some of the tools used to pierce flesh may have been spears, whereas others that cut flesh might have been used in hand-to-hand combat. Another possibility is that some matā had ritual uses, for instance in ritual scarification or mortuary rites (cf., Lipo et al. 2016:184).
There is a final question that needs to be considered: why did people on Rapa Nui give the name "matā" to such a diverse group of tools? One possibility is that each matā was owned, used, and associated with a single individual. The tool might have been primarily a deterrent used for gesturing, threatening, and shouting, as described in the historical accounts, or as an insignia, as seen in Cook's encounter with a man who "hoisted a piece of white cloth on his spear and led the way with his Ensign of Peace" (March 15, 1774; Beaglehole 1969:344). Although applied to tasks ranging from defense to food preparation, the primary function of the distinctively shaped tool with its own handle may have been to signify the identity of its owner. Possibly the possession of a hafted obsidian artifact indicated some form of status, such as adulthood, but all were slightly different as they were meaningful personal possessions.
The combined use-wear studies, together with the Lipo et al. (2016) analysis of morphology clearly demonstrate a need for new research directed at untangling the various uses and meanings of Rapa Nui obsidian stemmed artifacts. A comparison of tool uses across the various archaeological contexts where matā are found would also be valuable. For example, studies from gardening, habitation, or cave sites (Church 1998; Church & Rigney 1994; Church & Ellis 1996) have identified plants as the major use material for matā, whereas other activities including piercing and cutting flesh are recognized in the Australian Museum collection. Perhaps the latter were collected from the surface of ceremonial sites, where large samples have been recorded previously (e.g., Mulrooney et al. 2014:303). Clearly, much research is still required before the function(s) of Rapa Nui matā can be satisfactorily demonstrated.
Postscript: Biases in Museum Collections
As a postscript, it is worth considering the values and limitations of museum collections in understanding the uses of artifacts such as the matā that are iconic representations of a place or people. Lipo et al. (2016:177) note potential biases inherent in the Bishop Museum collection of matā, which formed 69% of their sample. Although they rightfully acknowledge "actions of the original collectors" (Lipo et al. 2016:177), one must also consider the agency of the islanders who sold artifacts to visitors and overseas collectors (e.g., see accounts in Simpson 2010). It therefore becomes important to ask whether the matā in museum collections are traditional items or were made specifically for sale to outsiders.
Concerning the Australian Museum sample, it is interesting that three of the four artifacts lacking wear traces (Table 3; A18926–28) were collected by J. Weisser, the paymaster on the ship Hyäne. He accompanied Lieutenant-Captain Geiseler during a 4-day visit to Rapa Nui in 1882 (Ayres & Ayres 1995:xiii). The blades he procured are the most variable in shape among the sample, clearly lack a usable edge, and were never hafted (Figure 7). It seems highly likely that local people collected substandard, rejected artifacts [End Page 12]
Matā lacking use-wear: a) A18926; b) A18927; and c) A18928. Scale bars are 1cm.
from the obsidian quarry (Stevenson et al. 1984) or even made copies themselves as described by Routledge (1919:224) (see Torrence 2000 for a similar case from Papua New Guinea). In contrast, unused E30738 was hafted in the past, but the handle has not been preserved. The presence of a handle on a matā, as observed in a number of museum specimens (e.g., Routledge 1919:Fig. 92; Simpson 2010:Fig. 10), does not therefore ensure that the tool was genuine and not made for sale. Since museum ethnographic collections largely represent the consequences of modern cross-cultural exchanges, rather than unbiased samples of past behaviors (Torrence & Clarke 2016), use-wear studies can play an [End Page 13] important role in the detection of truly historic/ancient artifacts versus those fashioned as copies for sale (cf., Kononenko et al. 2010). [End Page 14]
Archaeology, SOPHI, University of Sydney, Australia
Centre for Australian Biodiversity and Heritage, University of Wollongong, Australia
Archaeology, SOPHI, University of Sydney, Australia.
Acknowledgments
We thank Yvonne Carillo-Huffman for facilitating access to the Australian Museum collection; Angela Rosenstein for line drawings; Richard Fullagar and Elspeth Hayes for assistance with residue identification; Dale Simpson, Jill Hassall, and Kim Akermann for information; Christopher Stevenson for advice and encouragement; and two anonymous reviewers for useful suggestions.