Introduction

Ground stone tools have been studied for a long time worldwide as they played a key role in prehistoric human life. In the last thirty years, the study of ground stone tools has remarkably increased, and the methods used have evidently diversified, bringing some level of standardization and making it a crucial branch of Neolithic archaeological research1,2. Previous studies have covered many aspects of these tools, from classification and designation to manufacture and raw material, as well as function and life history3,4,5. While in particular, for polished bevelled stone tools—stone axes, adzes, chisels, knives, etc.—which tend to be regarded as functionally fixed and associated with woodworking and grain cultivation, there is a lack of wider exploration of their function and an understanding of the dynamic nature of the tools6.

Previous research on the function of polished bevelled stone tools has been comprehensively reviewed recently by Masclans. A, who has focused specially on this field, providing both experiments of use-wear analysis and specific site studies in Europe6,7. It has been summarized in her work that research on the function of polished bevelled stone tools began roughly with experiments and ethnographic investigations at the beginning of the 20th century, with experiments gradually systematised in the 1960s and 1970s, and it laid the groundwork for other studies which continue to be pursued by scholars to this day7. Such experimental research has mainly focused on woodworking, including deforestation and construction, distributed in the regions like Europe and America, as Huang. J has also reviewed in his work that scholars from the former Soviet Union, Denmark, and the Netherlands have been conducting experiments on woodworking since the 1920s8. In recent years, a considerable number of related experiments are being carried out in Europe in particular9, while earlier scholars in Japan have also carried out an in-depth and comprehensive study focused on the axes10. Before this focus on polished bevelled tools, use-wear analysis has long been a reliable method of analysing the function of lithic artefacts, from the publication of the English translation of S. A. Semenov’s pioneering work Prehistoric Technology11 to the remarkable doctoral dissertations of G. Odell12 and L. Keeley13. Recently, there have been increasing use-wear analyses of polished stone tools2, along with focus on polished bevelled tools which reveals the multiple possibilities of these tools, including woodworking, hide processing, butchering, etc.6,14.

In China, there have already been a considerable number of multidisciplinary and integrated studies on ground stone tools especially from the beginning of 21st century15, applying analysis of use-wear, residue and technology from many perspectives, such as ethnography and cultural anthropology, as seen in many examples16,17,18. While there have been few studies describing reuse behaviour, including tool type modification and tool length shortening8,19, along with edge renewal or repair behaviour pointed out in these studies, there is a general lack of focus on research into the reuse of ground stone tools.

This paper presents a study on function and reuse behaviours of polished bevelled stone tools from Zoumaling Neolithic site. Located in Shishou city, Huber Province of China (Fig. 1) and first discovered in 198920, Zoumaling site(112°31′45.64″, 29°40′21.92”) has undergone several investigations and excavations, and numerous artefacts have been unearthed, including stone artefacts, pottery, jade, etc. In 2014, the Department of Archaeology of Wuhan University, in conjunction with the Shishou Museum, carried out excavation at the site, and in particular made staged progress in understanding the layout and nature of this Neolithic walled-town site21. With the nature of its cultural remains and the result of AMS 14C dating of charred rice or millets samples, this walled-town site is believed to have been occupied through the Qujialing (5500–4500 cal BP), Shijiahe (4500–4200 cal BP), and Meishan (4200–3900 cal BP) periods21,22,23,24,25. Nearly 200 unearthed stone artefacts have already been subjected to analyses of raw material, techno-type and starch grains, revealing that various types of raw materials, with coarse sandstones and metamorphic sandstones forming the vast majority, were selected for different purposes based on their physical properties, and that many of the tools might be related to agricultural or gathering activities26,27,28. Given the situation that the use process, predetermined and actual function of these artefacts haven’t been explored thoroughly, we here conduct a study on function and reuse of polished bevelled stone artefacts unearthed during 2014–2018 from Zoumaling site, such as axes, adzes, chisels, etc. We try to offer an answer to the questions of what happened during the process of using these tools, and further explore the human activities of Zoumaling with the approach of use-wear and techno-functional analysis.

Fig. 1
Fig. 1
Full size image

Location of Zoumaling Site.

The scope of this study focuses on polished bevelled stone tools, defined as stone tools with bevelled edges produced through grinding. These tools primarily include stone axes, adzes, chisels, knives, Yue-axes (ceremonial battle-axes), and sickles. Tools such as arrowheads and spearheads are excluded from this study because, while they may also feature bevel structures, it’s the tip rather than the edge first contacts the objects. Their fracture mechanisms differ from those of the bevelled stone tools described above, and their functions are relatively fixed. Therefore, while they might broadly fall under the category of bevelled stone tools, they are classified in this study as projectile points and, therefore, not included.

The specific materials for this study (n = 74) are selected from approximately 200 stone artefacts unearthed during the 2014–2018 excavations at the Zoumaling site. These stone artefacts were found from strata, ash-pits, walls and cemeteries in various localities with dates of approximately 5500–3900 cal BP21,22,23, and the stratigraphy is rather complexed due to uncompleted excavation of the whole site and different construction structures of the city. Among these, 74 artefacts—including both complete tools and identifiable fragments—were identified as polished bevelled stone tools, tool types including stone axes, adzes, chisels, Yue-axes, knives, sickles and tool with bevel structure but unable to be categorized (Table 1). Artefacts without identifiable traces of grinding, lacking bevel structures, classified as projectile points, or too fragmentary to identify were excluded. The selected fragments are those with identifiable features characteristic of stone axes or adzes, typically broken at the middle part of the tool, with at least two polished surfaces and a relatively flat fracture surface. The vast majority of the raw materials of these tools are coarse sandstones, with very few of slates, mudstones, tuff, metamorphic sandstones, etc, obtained from nearby mountainous areas26.

Table 1 Tool types of study materials
Full size table

The study includes all artefacts meeting the selection criteria from the approximately 200 unearthed items, ensuring the proportion is not the result of subjective selection, therefore, allows the findings to reflect certain characteristics of the site. Stone artefacts other than these 74 samples, except projectile points, are mainly sandstones used to produce other stone tools. They are excluded from this study due to the fact that the function and production of these tools are rather clear and fixed.

Methods

The methods used in this study are a combination of techno-functional and use-wear analysis. These two approaches, from different perspectives, analyse the preconceived and actual function of stone tools.

Established by Boëda. E in the 1990s, techno-functional analysis is an approach that aims to understand the intended structuration and functionalization of the tool, as well as the predetermination of each part of the tool and the intention behind each production step29,30,31. It’s neither a method to explore the actual function of the stone tools from a technological view, nor a combination of lithic technology and traceology32,33. This approach has been mostly applied to knapped stone tools rather than polished ones, since the structure and techno-functional units are usually very clear on polished stone tools, reducing the necessity for deeper study. However, modification knapping on the broken edge after initial use provides more possibilities in terms of both structure and preconceived intentions, thereby allowing the application of techno-functional analysis to these tools (Fig. 2a).

Fig. 2: Two types of reuse behaviours.
Fig. 2: Two types of reuse behaviours.
Full size image

a Edge change of a single bevelled tool in the process of use and modification. b Scar change in the process of repair

Use-wear analysis examines microscopic traces on stone tools, aided by magnifying equipment like microscopes, to identify the employed units, tool motions, and the materials they processed, based on the principle that mechanical use and sometimes physico-chemical constraints cause irreversible alterations, leaving identifiable traces such as scarring, polish, rounding and striation on the tools’ surface13,34,35,36,37,38,39,40. In addition to exploring specific functions, use-wear analysis can also reveal the repair behaviours. Traces of re-grinding on the scars caused by use are also discernable on stone tools, as evidenced by the presence of distal parts of the scars, but with the two sides worn away, and possibly forming a pocket-like shape (Fig. 2b).

The selected materials (n = 74) are here subjected to techno-functional analysis first, with the ones identified as modified being detailed studied by this approach. Then, after cleaned in an ultrasound cleaner, they are observed, analysed and photographed by a Keyence VHX-5000 microscope with magnifications between 20× and 200×. The whole process follows the widely accepted protocol, as the low-magnification method is considered to be an appropriate way to study the function of ground stone tools41. Thanks to a large number of experiments conducted by scholars all around the world, we are now able to get access to comprehensive reference data, particularly on non-flint materials and ground stone tools in China, and thus apply it on this study.

Results

Results of techno-functional analysis

As for the polished bevelled stone tools, it is rather simple to observe them at Zoumaling from a techno-functional perspective, with the edge part being the transformative techno-functional unit, the butt part or the poll part being the prehensive techno-functional unit, and the middle part being the transmitting techno-functional unit. It is the structure of the transformative techno-functional unit, or in simple words, the bevel structure, that distinguishes different tool types. In the case of the most numerous tool types at this site, axes and adzes, even though the composition of their techno-functional units is very similar, we identify them as different types because axes have double-bevelled edges that are usually slightly convex, while adzes have single-bevelled, straight-lined edges. This fact, however, can also be easily observed with naked eyes and they are classified in similar ways by different edge morphology in a traditional typological view. We therefore tried to find more about the intention of each production step, though unfortunately, we were not able to figure out the exact production procedure of these tools, since all materials unearthed are finished tools, and there’s not a single blank or semi-product. It can be inferred from the knapping scars and traces of pecking still visible on some pieces that ancient humans primarily obtained the blanks for these tools through knapping. The bevel structures were pre-determined during this process, meaning the producers had already envisioned the intended tool by shaping the morphology of the blank. Traces of pecking and grinding often overlay the knapping scars, becoming increasingly delicate as they approach the edge. It can be seen on several pieces that the butt end is covered by knapping scars, and then in the middle part of the tool appear pecking traces, and then near the edge it becomes smooth, polished surface. Pecking and grinding were applied just for better shaping the morphology or obtaining a more structured and neat cutting edge; they did not take effect in determining the bevel structure since they could only reduce little volume. There are also a few of these tools that present equal degree of grinding throughout the entire tool, perhaps indicating the producers did not mind spending more time and energy to beautify their tools under certain circumstances, but most tools show visible knapping or pecking traces. Apart from pre-designed transformative unit, it has also shown clear evidence that the prehensive unit was also determined when obtaining blanks. We have observed on several stone axes a noticeable reduction in thickness starting around one-third of the way from the edge towards the poll end (Figs. 3a, 4a, and 6). This feature may provide a clue to their hafting, as the thinner end could more easily be inserted into the middle of a stick parallel to the edge.

Fig. 3: Modified tools with zigzag-like edge in transversal view.
Fig. 3: Modified tools with zigzag-like edge in transversal view.
Full size image

a Stone axe No. 2016SZT33④:2, b Stone axe No. 2016ST25③:1.

Fig. 4: Modified stone axes.
Fig. 4: Modified stone axes.
Full size image

a Modified stone axe with wave-like edge in transversal view No. 2018SZTG9G22:2, b Modified stone axe with notched edge and convergent point No. 2018SZTG9④:4.

When it comes to the edge modification behaviour, we have observed intentional secondary change of edge structure after use, which here we call modification, on six pieces, including 5 axes and 1 flake. Due to the fact that the broken edge is barely possible to be repaired to its original form or out of other personal intentions, the behaviour of modification changes the edge structure by knapping, or creates a new edge that doesn’t exist before. That makes it different from tool repair or edge resharpening in common sense, which will be discussed in the next column.

Though not outstanding in numbers, the edge structure of modified tools varies. We have concluded five different types of modified edges in morphology: (1) zigzag-like in transversal view, produced by bifacial knapping with knapping scars forming repetitive, sharp angles (Fig. 3); (2) wave-like in transversal view, produced by unifacial concave scars (Fig. 4a); (3) notched edge with a convergent point, formed by one knapping scar and original lateral surface (Fig. 4b); (4) three-segment polyline in front view, produced by knapping on the original corner (Fig. 5a); (5) ground straight-line edge on a flake (Fig. 5b). These modified tools, technically speaking, are no longer suitable for the same use as before with such huge change of edges. However, it saves time and energy avoiding the effort of producing a brand new one. Along with bevel changes, the prehensive way and gesture of using these tools may also differ from before. Hafting would have also changed or even been unnecessary.

Fig. 5: Modified tool and flake.
Fig. 5: Modified tool and flake.
Full size image

a Modified stone axe with edge of three-segment polyline in front view No. 2015SZT3⑧:3, b ground-edged flake No. 2016SZT34H159: 2 (image adapted from27).

The key to identifying modification behaviours on these tools lies in whether we can identify the intent behind the scars, as it can be complexed to distinguish scars produced by use and those by intentional knapping when they are overlapped. Nevertheless, we can still recognize intentional ones when the goal of knapping is to create a new edge different from the original one on the basis of a damaged one, so it often takes no more than a few knaps to form a relatively neat edge to meet the producers’ demand, while scars produced by heavy utilisation are normally morphologically disordered and different-sized, some with higher degree of invasion but evidently flatter and shallow (Fig. 6). As long as the purpose and consequence of each knapping scar is deduced—principally by integrating the evidences including but not limited to the position of striking point, absence or presence of counter-bulb, size, morphology, distribution, profundity, direction, chronological sequence of the scars, which has been maturely applied to knapped stone artifacts29—it is possible to determine if the scars (traces) are more closely related to edge modification or use damage or repair (Fig. 7). It is common in this site that the scars (traces) produced by use are evidently seen and considerable in number, yet it’s evident that the majority of them are not related to shaping a neat cutting edge but are the results of gradual edge rounding and dulling, until sometimes, relatively large flakes are chipped off and form a random “edge” not suitable for further use. Another supportive clue is the distance between the point of knapping and the original edge. There is technically a distance of several millimetres existing between the point of intentional knapping and the original edge, using one surface as striking platform. However, deviation of the edge position can also be a consequence of one or several heavy shocks during utilisation, so this criteria was applied in this study in a way that combined with other evidence.

Fig. 6
Fig. 6
Full size image

Unmodified tool with damaged edge and various scars No. 2016SZT29H161:1.

Fig. 7: Diagram of comparisons between tools modified and unmodified.
Fig. 7: Diagram of comparisons between tools modified and unmodified.
Full size image

a Comparison of edge structures in transversal view, b Comparison of scar morphology and distribution.

Result of use-wear analysis

Forty-nine selected samples are here subjected to use-wear analysis, including 31 axes (edge flakes included), 12 adzes, 2 Yue-axes (ceremonial battle-axes), 1 chisel, 1 knife, 1 sickle and 1 unclassified tool. The rest of them are excluded due to poor condition of preservation: being either encrusted with adhering sediment that could not be removed, too heavily rounded to discern any edges, or obscured by recent damage scars. None of the selected tools exhibits evenly distributed medium- to-heavy-degree rounding over their entire surface. Most retain at least part of a clearly defined edge, as well as discernible pecking or polishing traces, with minimal evidence of post-depositional alteration. Based on experimental data of previous studies, especially those with non-flint ground stone axes and adzes, we have identified the quantity, size, morphology and distribution of different kinds of traces as well as degree and distribution of rounding6,7,17,18,35,41,42,43,44,45,46,47,48,49,50,51,52. The results show that polished bevelled stone tools at Zoumaling are highly related to woodworking, but they are diverse in tool motions and detailed function. Below, we present typical use-wear patterns of each tool type, from which we have inferred their functions.

Axes take up the most proportion of the tools, and we have observed many traces related to woodworking and processing other medium-hard materials. 23 of these 31 axes present clear use-wear, while the other 8 axes are either unused or unidentifiable. A typical kind of “rolled-over” scars, which are bending scars visible on the dorsal or ventral surface but initiating from the opposite side (Fig. 8), has been considered as highly pertinent to wood working34,41,53,54,55. They have been observed on at least 11 pieces (Fig. 9). Another typical type of use-wear trace is stepped scars in an overlapped distribution, possibly produced by processing of medium-hard materials (Fig. 10). Moreover, half of the axes are covered with large stepped scars in a continuous, overlapped distribution along the edge, indicating very frequent use (Fig. 6). We have also frequently observed severe damage accumulation around the two corners of the edge, which are much more rounded than the middle parts of the edge, suggesting that the corners are the first to contact the working objects, further indicating a possible chopping gesture with hafting, often used in splitting something apart vertically.

Fig. 8: Use-wear on two experimental replicas of woodworking.
Fig. 8: Use-wear on two experimental replicas of woodworking.
Full size image

a “Rolled-over” scars on an experimental stone chisel in front and transversal view, image from53, b “Rolled-over” scars in front and transversal view with a 3D model of the edge, image from41.

Fig. 9: “Rolled-over” scar on stone axe No. 2016SZT28④:1.
Fig. 9: “Rolled-over” scar on stone axe No. 2016SZT28④:1.
Full size image

a, b the same “rolled-over” scar present on both faces.

Fig. 10: Use-wear on stone axe No. 2018SZKTN6E5T1905③:1.
Fig. 10: Use-wear on stone axe No. 2018SZKTN6E5T1905③:1.
Full size image

a–c Stepped scars, d Striations and polishes by grinding.

Adzes also present use-wear traces highly relative to wood working and processing of medium-hard materials. We’ve found on an adze with a step numerous scars with stepped and feathered termination of all size in a dispersed distribution, which is a sign of frequent processing of medium or hard materials such as dry wood or bones (Fig. 11). The typical “rolled-over” scars also appear on five pieces, especially the four pieces that are of smaller size (Fig. 12). These four are all made of light-coloured sandstones or slates, with an average length of about 2–3 centimetres and an extremely high degree of grinding, which distinguishes them from other stone tools at this site. This difference between adzes of varied sizes suggests that larger ones might have been used for processing harder and larger objects, while smaller ones might have been used for finer woodworking. Unlike axes, in general we have not noticed distinction of scar numbers or degree of rounding around the corners of edge; corners of some pieces even remain intact. This may suggest a diagonal chopping gesture instead of a vertical one.

Fig. 11: Use-wear on stone adze No. 2016SZT29③:2.
Fig. 11: Use-wear on stone adze No. 2016SZT29③:2.
Full size image

a scars with stepped and feathered termination in front and transversal view, b damaged corner, c scars with stepped termination in overlapped distribution.

Fig. 12: Use-wear on stone adze No. 2016SZT25H309:4.
Fig. 12: Use-wear on stone adze No. 2016SZT25H309:4.
Full size image

a “rolled-over” scars presented on both faces and in transversal view.

Two Yue-axes (ceremonial battle-axes) present the trace pattern of wood chopping with typical “rolled-over” scars, but with lower use frequency (Fig. 13). Besides, they might also have been used to work on other materials of medium hardness. We have also noticed that the edges of these two Yue-axes are much sharper with obvious smaller thickness compared to that of most axes or adzes, while the two corners of them remain unsharpened. This reveals that for this type of tool corners are not included in utilisation when first designed, and the middle part of the edge first contacts the working object usually. In typology, a main feature of Yue-axes is a hole at the poll end, which is normally considered to be reserved for hafting. Different methods of hafting, in correspondence with different gesture, may explain the distinct characteristics observed on the edge corners of Yue-axes and axes, even though both have served as wood-chopping tools.

Fig. 13: Use-wear on stone Yue No. A⑦a:2.
Fig. 13: Use-wear on stone Yue No. A⑦a:2.
Full size image

a scars with stepped and feathered termination in front and transversal view, b scars with stepped and feathered termination, c scars with feathered termination in transversal view.

There is only one knife found at the site. We have observed relatively small scars with visible directions consistent with the edge and low degree of perpendicular invasion, indicating cutting and sawing behaviour on multiple objects (Fig. 14). The only unclassified tool also shows relevance with woodworking, with “rolled-over” scars. Apart from all aforementioned, the rest of the stone tools show no sign of use-wear: one chisel, use-wear unidentifiable because of poor state of preservation; one sickle, identified as unused.

Fig. 14: Use-wear on stone knife No. 2015SZT3⑨:1.
Fig. 14: Use-wear on stone knife No. 2015SZT3⑨:1.
Full size image

a, b scars with stepped termination in front and transversal view.

We conclude that the polished bevelled stone tools at Zoumaling site served primarily as woodworking tools, but the possibilities of processing other objects of medium to high hardness should not be excluded. Diversity in tool size, gestures and detailed processing of wood makes it possible that these three main tool types—axes, adzes, and Yue-axes—constitute an integrated woodworking tool set, with Yue-axes for chopping, axes and larger adzes for chopping and splitting, smaller adzes for finer processing. With this tool set, ancient humans could be capable to fully exploit and utilise wood resources. Due to the uneven number of each tool type, we haven’t found identifiable use-wear on chisels, which could also be part of the tool set used to split smaller wood blocks. Similar situation also appears on knives and sickles, which have often been considered as and proved to be harvesting tools highly related to rice cultivation in China43,56,57,58. We hope that subsequent excavations can provide a chance to complement these tool types.

On the other hand, tool repair, or resharpening, or regrinding, as in the cases of polished bevelled stone tool, has been commonly seen on stone tools and has been proved by experiments a way of tool maintenance47,59. Polished bevelled tools have even been considered as the category that requires most effort invested in tool maintenance and repair60. This is not hard to comprehend, as it always takes several hours and much effort to produce one of these tools starting from scratch61. Worn edges can be resharpened by grinding and this insures long-term utilisation on the same tool. We emphasize the term ‘repair’ here to distinguish it from the aforementioned ‘modification’, as the procedure aims to restore the tool to its original function rather than create a completely different one. To be more specific, the process of repair tries to make the edge as close to its original form without changing its bevel structure.

We’ve found signs of repair by use-wear analysis on 7 pieces, some even visible by naked eyes. These 7 tools include 3 axes, 2 adzes, 1 Yue-axe, and 1 unclassified piece, which indicates that repair was not preferentially applied to any specific tool type. Among them, the edges of five of them remain sharp and still suitable for further use (Fig. 15). Therefore, they are considered as the result of a successful repair action. These repaired scars usually have one or both sides of the outline blurred by grinding, while the distal end remains relatively clear because it has not been intensively ground or it is more concave and difficult to smooth out (Fig. 15b–e). Sometimes the shape of the scar seems unusual because grinding shrinks the proximal part (Fig. 15a). From a transversal view, a deviation of edge position may also be observed (Fig. 15b). The other two pieces have very clear signs of repair attempt, but the repair was interrupted before the edge was restored (Fig. 16). The edges of these two pieces have been more severely damaged, which made fully repairing them to their original form more difficult and may explain why the attempt was abandoned. Successful repairs depend highly on the low degree of edge damage. Larger scars, more concave surfaces, and more evident position deviation of edges would all make it less possible for the tools to be fully repaired.

Fig. 15: Repair traces on polished bevelled stone tools.
Fig. 15: Repair traces on polished bevelled stone tools.
Full size image

a polished scar on Yue No. 2018SZN7E5T0304F26:1, b polished scar and edge deviation transversal view of axe No. 2016SZT36H113:3, c polished scars on adze No. 2016SZT29③:2, d polished scar on adze No. 2016SZT25H309:4, e polished scar on unclassified tool No. 2015SZT12H7②:16.

Fig. 16: Attempted repairs on polished bevelled stone tools.
Fig. 16: Attempted repairs on polished bevelled stone tools.
Full size image

a polished scar on axe No. 2015SZT15H65:8, b polished scar on axe No. 2016SZT30③:1.

The repair traces share some features in common, and they can be clues for identification. The most obvious one is blurred or missing part of scar outlines caused by grinding (Fig. 17a, b), and the narrowed part of a scar often proximal (Fig. 17c). The reground area sometimes forms a boundary between the original polished surface of the tool and it can be very evident (Fig. 16b). There have even been cases in which several times of repair could be seen on the same stone tool from ridges formed by reground area62, though not seen at Zoumaling. Also, edge position deviation in transversal view is sometimes applicable, since grinding is a process of reduction. Technically speaking, the edge can never be restored to exactly the same as when newly produced regardless of how it’s ground. For most of the polished bevelled tools, the transversal view of the intact edge is a straight line, while a repaired edge can appear slightly wavy (Fig. 18). The degree of deviation is usually subtle and better observed under a microscope, as a stronger deviation would make repair more difficult or even result in failure.

Fig. 17: Diagrams of scar morphology related to repair.
Fig. 17: Diagrams of scar morphology related to repair.
Full size image

a scar with the distal part missing, b scar with the proximal part missing (c) scar with the proximal part narrowed.

Fig. 18
Fig. 18
Full size image

Edge morphology in three phases: before use, after use and after repair.

Discussion

When talking about prehistoric stone tools, a question is often proposed: what were they used for? Functional analysis provides answers to this question, and researchers are getting more support from modern methods such as use-wear and residue. The polished bevelled tools at Zoumaling were used to reach multiple goals, mainly woodworking or processing other medium to hard objects by use-wear analysis aforementioned, and processing vegetal materials by starch grain analysis28.

However, if we change a way to ask this question – how were these tools used? We may find that functional analysis cannot answer it fully. As we have done in this work, we have added importance to the whole process of use rather than focus on the goal alone. We’ve found modification and repair behaviour for reuse of tools, which reveals that there were several choices for ancient humans at Zoumaling site after the tools were damaged. If not severely damaged, they would repair it by grinding the edge again; if repair seemed impossible, they would modify the edge, usually by knapping and result in a new type of tool differently with repair. We believe that tool maintenance and reuse was a widely spread common sense at this site, though pieces with related traces constitute a small proportion of less than 20%. It cannot be ignored that traces of maintenance and use always obscure each other5. A tool without traces of maintenance doesn’t necessarily mean it hasn’t gone through any maintenance process; later use-wear would cover former traces of any kind, and it’s totally possible that regrinding has left no visible traces to be identified by researchers. This is also the reason why we cannot here provide functions of modified tools, because knapping scars of modification overlay older use-wear, and vice versa, we may not be able to identify a modified piece because use-wear has covered up knapping scars for modification. Besides, the more likely situation is that ancient humans discarded their tools after frequent use, not discarding them right after being maintained. We believe it to be a popular and stable consciousness because various specific, small-scale, localized (rather than large-scale, wholesale) modified edges and repair traces were seen on nearly all tool types.

The reason behind reuse seems very simple: it always takes less effort to maintain a tool, no matter modification or repair, than to produce a brand new one. However, it doesn’t reduce the necessity to focus on the whole use process, which is one crucial part of ancient human lives. We’ve also observed some kind of use accident, as there are four very similar edge flakes of 2-3 centimetres, and two axes with only one scar of the same size, without which the edge would be intact. Unfortunately, none of these 6 pieces is from the same tool, but we still suggest that they are the result of the same accident, as they fit together perfectly despite not from the same tool (Fig. 19). This may be the consequence of large edge damage by one single scar when first used. We call it an accident because it may result from wrong gestures, incorrect using angle or extremely hard objects that are not suitable to apply this tool on. There were attempt to repair one of the tools, but it was too damaged to repair, so it was given up eventually (Fig. 15a). This is part of how stone tools were used, and also provides evidence for the fact that modification is not the preceding step for repair, as they did not try to do any knapping even the scar of damage was so large and concave.

Fig. 19: Refitting diagram of use accident.
Fig. 19: Refitting diagram of use accident.
Full size image

a axe No. 2016SZT24⑧:2, b edge flake No. 2015SZT20⑤:1.

From what has already excavated had studied, ancient humans at Zoumaling site had preference on raw material selection, were highly skilled in stone tool design and production, and were able to reach multiple goals using these stone tools. Functional analysis suggests high demand of woodworking, and tool types concerning woodworking make up a large part of all bevelled tools. With Zoumaling being a large city with complexed structures and probable military nature, high demand of woodworking aligns with wooden building construction and shaft production for projectile points, as arrowheads have also been discovered.

The possibility of processing other hard materials, such as bones, may be interpreted as utilisation of animal resources and protein acquisition. By contrast, although we don’t exclude the possibility of vegetal processing on these tools, it is still hard to ignore the lack of certain tool types such as knives and sickles, which are often considered as vegetal harvesting tool, since rice cultivation has been proved at this site. We conjecture that this type of tools was probably made of vegetal raw materials such as bamboo or wood, which is less possible to preserve to this day, or it is a result of unfinished excavation with a focus on city structure of the whole site.

In general, the livelihood here may have been a combination of rice agriculture and hunting-gathering; wooden buildings may have been very common, and there was a large demand for woodworking, as well as a relatively high level of woodworking technology and a complete woodworking tool set. Bamboo and wood tools may also have been an important part of the tool assemblage, but have not been preserved to now.

This work is an attempt of the application of techno-functional analysis on polished stone tools, and also of a combination of it and use-wear analysis. Comprehensive multidisciplinary methodology has become an important trend in ground stone tool studies, and our work aligns with this approach63. Throughout this work, we have noticed that these two methods complement each other, providing a more dynamic perspective to the entire process of stone tool use.

However, the implement of these methods still faces difficulty. Techno-functional analysis is still in very limited scope when it comes to most ground stone tools with clear techno-functional units. How this method can be applied depends on the diversity and particularity of study materials. Use-wear can be affected by numerous variables, such as raw materials, hardness of working objects, edge angles, concavity or convexity of edge surfaces, frequency, hafting, gesture, etc. It is certain that we need more experiments to build a reference database, but obstacles in replica production and effort invested in variable controlling are still waiting to be overcome. It’s good to see that there have been lithic studies referring to rock mechanics64,65,66, because it explains how fractures and wear take place on different rocks, which happen throughout the life history of a stone tool. However, this is still very inadequate, and we hope more future work can fill the gap pertinent to rock physics.

To conclude, this study proposes a study of use process, including function and reuse on polished bevelled stone tools at Zoumaling site. It is also an attempt to apply techno-functional analysis to ground stone tools and a multidisciplinary combination of this method and use-wear analysis. The function of these tools is highly pertinent to woodworking, and they could form an integrated tool set as they are diverse in detail. Possibilities of processing other objects of medium to high hardness also cannot be excluded. Popular consciousness of tool maintenance widely spread at this site; ancient humans could choose, from modification to obtain a new edge structure, and repair to restore the tool to its original form for the same use. Both ways served as methods for further tool reuse. From all aforementioned, we suggest that they had clear planning, skilled production techniques, flexible use and diverse reuse behaviours for polished bevelled stone tools.

Even though we tried to analyze hafting behaviours from both techno-functional and use-wear perspectives, we haven’t found enough evidence to support a certain way of hafting. We hope future work will do more in this aspect, as well as experiments on use-wear and rock mechanics.

This work also attaches more importance to a more dynamic view of the entire use process, and it is also the result of multidisciplinary trend in the research of ground stone tools. Future research may further explore the combination of other methods to provide a more comprehensive understanding of these tools, as well as of Neolithic human lives.