Introduction

Decapitation, or the removal of a human head from the body, is a form of violence common in historic and prehistoric societies worldwide. The act is a cultural phenomenon or a social behavior related to capital punishment, head trophies, or ritual and spiritual beheading (Hutton 1928; Aldhouse-Green 2002; Bonogofsky 2015; Chacon and Dye 2007; Heron 2020; Pearson 2005).

Some early signs of beheading during late Paleolithic and early Neolithic Ages were in ritual forms related to commemoration and sanctification (Bonogofsky 2015; Hoskins 1996). For example, two decapitation cases were found in Tell Qaramel, northern Syria, dating between 12,000 and 15,000 years ago. The heads were believed to have been excised as a burial ritual before the tissues of the corpses completely decomposed (Kanjou et al. 2015). Also, a hunter-gatherer living about 9000 years ago was beheaded at the Lagoa Santa site in east-central Brazil, reflecting the complexity of early funeral ceremonies in the region (Strauss et al. 2015). In Africa, the earliest beheading is also considered to be determined by the funeral ceremony, dating back to 8000 years ago in Algeria (Haverkort and Lubell 1999). Earlier signs of decapitation, such as those in Middle Pleistocene humans (i.e., Homo erectus) entailed not only head removal but also full-body dismemberment. This stage dated back 600,000 years, and was attributed to food collection or cannibalism (Jia 1979; Jia and Huang 1984; Saladié and Rodríguez-Hidalgo 2017; White 1986; Wu 1979).

During the Neolithic Age, the development of spiritual and religious concepts related to the relationship between the cosmos and humankind led to the phenomenon of beheading (and related headhunting) became common, in which the severed heads were given a specific ritual meaning: to steal the soul and energy from the enemy during warfare and conflict (Conlee 2007; Li 1987). In Europe, the earliest evidence of interpersonal conflict–related beheading can be traced back to at least 9000 years ago. At least 9 individuals with signs of violent conflict-related beheadings were found in a Middle Stone Age cave in southern Germany (Orschiedt 2005). In the Americas, there has been a large amount of evidence for prehistoric human decapitation dating back to about 3000 years. Cases have also been found on the central coast of Peru and the Andes Mountains, with headless remains and skulls found separately (Engel 1963; Lumbreras 1981; Nagaoka et al. 2019). In Asia, previous excavations in the Levant have demonstrated the prevalence of prehistoric human headhunting in the region (Bar-Yosef and Alon 1988; Santana et al. 2012; Simmons et al. 2007; Talalay 2007). In China, there are many threads of evidence for headhunting during the Neolithic age throughout China, including head-only burials at Banpo, Xiamengcun, Jiangou, Liuwan, Yijiacheng, Shimao, and Wangchenggang, which are closely related to prehistoric human head rituals (Archaeological Team of Qinghai Provincial Cultural Relics Management Office and Qinghai Team, Institute of Archaeology, Chinese Academy of Sciences 1976; Beijing University and Handan Archaeological Excavation Team of Hebei Provincial Cultural Bureau 1959; Chen 1989; Chen et al. 2016; Henan Provincial Institute of Cultural Relics and Archaeology, Department of Archaeology Chinese History Museum 1992; Shandong University Teaching and Research Section of Archaeology, Department of History 1990; Shen 2021; Sun et al. 2013; Wang 1982). Accordingly, there have been headless burials at Baiyangcun, Beishouling, Dawenkou, Hemudu, Qilihe, Huangniangniangtai, Liuwan, Nanbaoligaotu, and Magatai (Liu et al. 1979; Shandong Provincial Cultural Relics Management Office and Jinan Museum 1974; Ta and Ji 2008; The Archaeological Team of the Hemudu site 1980; Wang 2015; Wang and Lin 1984; Wei 1978; Zhao et al. 2016). All these findings suggest the prevalence of headhunting during the Neolithic Age.

From the Bronze Age and onwards, many burials with signs of decapitation have been found in Europe and the Levant (Anderson 2001; Ardagna et al. 2005; Carty 2015; Kanjou et al. 2015; Strauss et al. 2015; Dolce 2018; Lee et al. 2017; Loe et al. 2014; Mckinley 1993; Nagaoka et al. 2009; Philpott 1991; Pitts et al. 2002; Tracy and Massey 2012; Tucker 2015; Mikulski et al. 2021), in the Americas (Chacon and Dye 2007; Tung 2008; Verano 2001, 2003), and in China such as the Yinxu Site (Henan Provincial Institute of Cultural Relics and Archaeology 2011; Institute of Archaeology, the Chinese Academy of Social Sciences 1994; Sun 2015) and Lu’an (Zhou et al. 2020).

Different from execution or ritual sacrifice by decapitation, headhunting is a special type of organized beheading, in which trophy heads were sought in the context of warfare or raiding, as evidenced by discoveries of multiple heads or headless bodies (Browne et al. 1993; Proulx 2001; Verano 2001). Archaeological discoveries indicated that headhunting became a human behavior early in the history of violence. For example, the Narmer Palette, one of the earliest historic documents dating from about 3100 BCE, carries the signs of decapitation as a token of victory and conquest. In addition to its deterrent effect, headhunting during warfare could be a strategy to take the soul and energy away from an enemy. The trophy head could then be used as a sacrifice during the ritual or religious ceremonies of the victor (Armit 2012; Benson and Cook 2001; Bienert 1991; Bonatz 2004; Borsje 2007; Boylston et al. 2000; Buckberry 2008; Chacon and Dye 2007; Girard 1977; Goring-Morris 2005; Kuijt 2008; Moser 1973; Philpott 1991; Thomas et al. 2017; Waldron 1996). The importance and carrying convenience of the head make it an ideal trophy on the battlefield; holding it demonstrates the exploits of the victor and symbolizes the complete destruction of the enemy (Carty 2015; Goldsworthy 1996; Lambert 2007; Moser 1973; Ogburn 2007; Okumura and Siew 2013; Proulx 2001).

In this study, a mass decapitation event in China was examined and discussed within its archeological background of the Neolithic Age. The decapitation dated around 4100 years ago and occurred at the Honghe site in northeast China, adjacent to northeast Asia. Unique headless burials have been found in the Kitoi culture along Lake Baikal (6900–8800 BP) in Eastern Siberia (Weber 1994; Weber et al. 2002) (which under the same fishing and hunting lifestyle at the Honghe site in this study). The same phenomenon was observed among the headless individuals in the Shamanka II cemetery (Bahn 2012). Since the sites are geographically close and share similar lifestyles and beheading practices, evidence indicates that headhunting was popular during the prehistory period of northeast Asia. In ancient China, human heads played an important role in sacrificial ceremonies since people believed that the head was the favorite sacred object of gods with supernatural powers, such as at Shimao and other sites (Chen 1989; He 2013; Jin 2005; Qian 1994). At first, the heads were primarily used as a sacrifice to ensure a good harvest (Fang and Han 2000; Shaanxi Archaeological Institute Jingshui Team 1960; Sun et al. 1961), or buried in the foundation of new construction to pray for safety and prosperity, such as Handan, Shimao, and other sites (Beijing University and Handan Archaeological Excavation Team of Hebei Provincial Cultural Bureau 1959; Chen 1989; Chen et al. 2016; Shen 2021; Sun et al. 2013; Wang 1982). The second primary use of the head was for head worship, in which the head was removed from the body as a sacred object for worship and various ceremonies (Li 1987; Qian 1994). In this archeological and historic context, references were drawn to the history of interpersonal conflicts and the development of complex human societies in prehistoric China.

Materials and methods

The Honghe prehistoric site (123°35′56″E, 47°06′49″N; altitude 156 m) is a Neolithic Age human settlement in northeastern China, located about 1 km south of what is now Honghe Village, Qindaur Township, Duermen, Fulaerji District, Qiqihar City, Heilongjiang Province, right on the north edge of the Nenjiang River (Fig. 1). The Honghe site was first noticed in the 1990s when Neolithic remains belonging to the Angangxi Culture Group in Northeast China were found. Since 2013, the Heilongjiang Provincial Institute of Cultural Relics and Archaeology has conducted six excavations at the Honghe site. In total, 42 houses (on the same level without overlapping) and 14 tombs (some overlapping with house loci) of the Neolithic Age were found and excavated. Human remains included 68 skeletons (41 headless) and 4 skulls (bodiless). Archaeological discoveries at the Honghe site also include numerous cultural remains such as Neolithic pottery, bone and stone tools, semi-terrain houses, and defensive trenches. All items indicate that the site was a settlement with sustenance modes of fishing, hunting, and farming (Tian et al. 2019; Zhang et al. 2020).

Fig. 1
figure 1

The Honghe site. a Location. b–d Excavation of the Honghe site. The Honghe site is located next to the northwest bank of the Nenjiang River (b). Signs of decapitation were found in houses F004 and F10, and tombs M101, M103, and 2013-M103 (c, d). The three residential areas (plus some tombs) were encircled by three defensive trenches (HG3, HG4, and HG5), respectively, whose open ends connected to the Nenjiang River while they were functional (c). HG3 is about 4–5 m wide and 2.8–3.5 m depth, protecting an area of 2100 m2; HG4 is about 5–6 m wide and 3–4 m depth, protecting an area of 8000 m2; and HG5 is about 7–8 m wide and 3.5–4 m depth, protecting an area of 3500 m2

Full size image

This study focused on headless human remains found from houses F004 and F10, and tombs M103, M101, and 2013-M103, presumably victims of headhunting through decapitation. An estimate of sex was carried out through evaluating specific features on the pelvis, and the skull when it was available (Brooks and Suchey 1990; Buikstra and Ubelaker 1994; Shao 1985). Bone size and skeletal robustness were considered as complementary evidence. Age at death in adults was mainly estimated by the basis of pubic symphysis and the degree of degeneration on the auricular surface (Lovejoy et al. 1985; Shao 1985; Todd 1920, 1921). For minors, estimates of age were calculated according to epiphyseal development and fusion (Baker et al. 2005; Cunningham et al. 2016; Krogman and Iscan 1986). Sex of four skulls in the head pit was estimated by cranial features, while age was roughly estimated by the statuses of sutural fusion and teeth wear (White and Folkens 2005).

To determine evidence of decapitation, cervical vertebrae of headless skeletons were visually observed for marks from sharp-force cuts. Cut marks from sharp tools on the cervical vertebrae indicate the practice of decapitation (Roberts 2009; Roberts and Manchester 2005; Tung 2012). Upon visual examination, the cut marks were further observed and measured using the KEYENCE VHX-2000 microscopic system (KEYENCE America. Itasca, Illinois, U.S.A.).

Initial C14 dating of the Honghe site yielded an age of 4000-4500 years ago (Zhang et al. 2020). Human bones were collected during this study to confirm their age and determine if there were differences between the human remains from house sites and those from tombs. Overall, dating results for Honghe yielded to be from 4100 to 4400 year BP, collaborating with previous dating results of 4000–4500 years ago (Appendix Table 3). The age of the skeletons from house sites and tombs overlapped. The age of skeletons from burial sites ranged from 4417 to 4146 cal BP with different confidence levels, while the age of headless skeletons from house sites ranged from 4417 to 4142 cal BP with different confidence levels.

Results

Paleodemography of decapitated individuals

In total, 41 headless skeletons were excavated. Skeletons came from two houses and three tombs: houses F004 and F10, and tombs M103, M101, and 2013-M103 (2013-M103 was excavated in 2013 (Tian et al. 2019); the rest of them were excavated in 2019 (Zhang et al. 2020)) (Table 1 and Fig. 2). The skeletons were presumably Honghe residents that fell victim to headhunting of multiple events; in addition, there were four skulls found in a pit outside of house F10, also presumable victims of headhunting. The bodiless skulls could have been head trophies of enemies that were decapitated by Honghe residents. Examination of skeletal trauma from the headless individuals at the Honghe site revealed that there were no observable signs of trauma except for cut marks on the cervical vertebrae.

Table 1 Headless individuals found at house (F) and tombs (M) loci the Honghe site
Full size table
Fig. 2
figure 2

Headless and head only burial loci. a F0004; b M103; c F10; d M101; e head pit outside F10; f 2013-M103

Full size image

All 41 headless individuals were females or juveniles (Table 1). Twenty two of them were identifiable by sex, all of which were female. In terms of age distribution among the 36 individuals, 15 of them were subadult, 5 were young adult (15–23 years); 11 were middle adult A (24–35 years); 5 middle adult B (36–55 years), and no old adult (>55 years). The remains of the four skulls from the burial pit outside of house F10 all belonged to adult males (30 to 40 years).

Loci with decapitation signs

F004: House F004 is a semi-terrain house with a door facing south. The floor is flat, arranged in a rectangular shape with rounded corners. It is 17.0 m in length and 6.4 m in width, with 38 pillar holes and 2 ovens (Figs. 2a and 3). A total of 22 individuals were found placed on the ground (21 headless). All individuals were females or subadults (Table 1). The placement of human bones was in a north-south direction. Based on the distribution of human bones, 22 individuals were grouped in VII groups. Except for one individual in group II with an intact skull (a female from 30–35 years of age, perhaps a victim of the same conflict yet the head was not taken), the rest of the individuals in the house were all headless. None of these skeletons showed signs of trauma. Individuals in groups I (2 individuals), II (1), III (1), and V (4) were placed with their cranial or cervical vertebrae pointing to the north. Individuals in groups IV (8) and VII (5) were placed in a disorderly manner, with some of the bones overlapping (Fig. 3c, f). Except for a partial postcranial skeleton of a juvenile in group VI (1), all skeletons in supine extended position were articulated, without signs of postmortem disturbance. Tools made of stone, shells, and bone were found in the house as well.

Fig. 3
figure 3

House F004. There were skeletal remains of 22 individuals separated into seven groups (a). Except for one individual in Group II with an intact skull (a female of 30–35 years of age, perhaps a victim of the same conflict yet the head was not taken), the rest of the individuals in the house were all headless (b-f)

Full size image

F10: House F10 is 15.9 m in length and 7.8 m in width, and its floor is 1.23–1.8 m deep. A total of eleven headless individuals in three groups were placed on the ground in a supine position with their limbs straightened. All of them were aligned parallel with their cervical vertebrae pointing to north (Fig. 2c). Three individuals were found on the west side of the house (from west to east II-1, II-2, II-3, II-4), four individuals were found in the middle (from west to east II-1, II-2, II-3, II-4), and four individuals were found on the east side (from west to east as III-1, III-2, III-3, III-4). All were females or juveniles, including three young or middle-aged adults, and 8 juveniles aged 2 to 11 years.

M103: Burial M103 was a vertical pit tomb with four headless individuals placed from east to west (respectively named M103-1, M103-2, M103-3, and M103-4). All bodies were placed side by side in a supine position with straightened limbs; the cervical vertebrae pointed northeast and were placed close to the edge of the tomb (Figs. 2b and 4). The grave goods included some simple clam shells, clay pots, and stones. Three of the individuals were estimated to be adult females, and the other was identified as a juvenile between 12 and 14 years of age. All remaining postcranial skeletons were in an normal anatomical position, indicating no postmortem disturbance after the headless bodies were placed in the tomb.

M101: Burial M101 was made in the same shape and structure as M103. There were three poorly preserved and headless individuals placed side by side in a supine position. All cervical vertebrae pointed northeast (Fig. 2d). Estimates of sex and age were not possible, yet two of the individuals might have been juveniles.

2013-M103: Burial 2013-M103 had a burial pit 1.9 m in length, 0.9–1.1 m in breadth, and 1.0 m in depth. It was found in a house 2013-F1 excavated in 2013. There were two skeletons without skulls that belong to young females; the left one was about 20 years old, and the right one was about 17–18 years old (Fig. 2f).

Burial pit outside F10: The remains of four skulls were found in a burial pit southwest of House F10 (Fig. 2e). All skulls belonged to adult males, presumably trophies taken from enemies who were decapitated by the tribal members of this settlement.

Analysis of cut marks on cervical vertebrae

Among the 41 headless individuals, 25 had preserved cervical vertebrae. The first cervical vertebra (C1), the axis, was unobserved in 24 individuals; while the first two cervical vertebrae together (C1 and C2), the axis and atlas, were unobserved in 20 individuals (Table 1 and Figs. 3 and 4). Cut marks were found on the axis of 5 well-preserved individuals, including two individuals in the F004 group IV (IV-5 and IV-6), and three individuals in the M103 (1–3), but not in other skeletons. All cut marks on the cervical vertebrae of the five individuals had no signs of healing, indicating that they were perimortem injuries. There were multiple cut marks on the axis except for individual F004IV-5.

Fig. 4
figure 4

Burial M103. There were four headless individuals placed side by side in a supine position with straight limbs (a–b). The 2nd cervical vertebra (axis) was preserved with sharp force cut marks in M103-2 (c) and M103-3 (d)

Full size image

F004IV-5: F004IV-5 was a 30-year-old female with the 2nd to 7th cervical vertebrae preserved. The upper part of the odontoid process of the axis was truncated obliquely, leaving a nearly elliptical section with a maximum diameter of 7.1 mm (Fig. 5 (a1–a3)).

Fig. 5
figure 5

Cut marks on the axis of five individuals. a F004IV-5; b F004IV-6; c M103-1; d M103-2; e M103-3

Full size image

F004IV-6: F004IV-6 was a 35-year-old female with the 2nd–7th cervical vertebrae preserved. Two groups of cut marks were observed on the front body of the axis (ventral side) (Fig. 5 (b1–b3)), whose odontoid process was missing. The first injury group was in the middle of the body with several horizontal cut marks. The most obvious one was about 5 mm long and 0.48 mm wide. The second group was on the left side near the pedicle. The cut mark was about 3 mm long and 0.39 mm wide (Fig. 5 (b3)).

M103-1: M103-1 was a female between the ages of 18 and 22. Except for the skull and the atlas, the rest of the bones were relatively intact (Fig. 4b). In the anterior view of the axis, five cuts at different lengths were observed on the right side of the odontoid process (Fig. 5 (c1–c3)). Under the microscope, it was found that the cuts were V-shaped, linear, and inclined upward as a whole (Fig. 5 (c3)). The longest one was about 10.41 mm long, the shortest one was about 1.79 mm long, and the widest cutmark was 0.53 mm (Table 2). These cut marks were left by multiple cut strikes using the same instrument.

Table 2 Measurement of cut marks on the axis of individual M103-1 and M103-3. Unit millimeter
Full size table

M103-2: M103-2 was a 35-year-old female with intact bones except for the skull and atlas (Fig. 4a–c). The odontoid process of the axis, part of the vertebral body, and the pedicle of the vertebral arch were incised apart. Three cut marks were observed on the surface of the vertebral body in our anterior view (Fig. 5 (d1–d3)). The marks were linear and caused by repeated cutting motions. One of the two most obvious marks was located on the surface of the lower one-half of the vertebral body (Fig. 5 (d2)). The cut marks were about 2.1 mm long, 0.40 mm wide, and 0.45 mm deep. The second mark was located on the surface of the lower one-third of the vertebral body, with a length of about 1.8 mm, a width of about 0.39 mm, and a depth of about 0.28 mm (Fig. 5 (d3)).

M103-3: M103-3 was a juvenile with an estimated age at death of 12–14 years old (Fig. 4). The cervical vertebrae were all preserved (Fig. 4d). In the anterior view of the axis, there was a groove left from sharp-force marks caused by repeated cutting. These markings were observed at the point where the odontoid process and the vertebral body connected (Fig. 5 (e1–e2)). Overall, the groove was wide on the left side (about 2.14 mm in width), and narrow on the right side (about 0.64 mm in width). Our microscopic observation of the groove showed that it was formed by marks caused by the superposition of at least four linear and sharp-force cuts. The longest cutmark was about 10.28 mm in length, and the shortest one was about 4.88 mm in length. Meanwhile, the widest cut was about 0.48 mm in width, and the deepest was about 0.85 mm in depth (Table 2 and Fig. 5 (e3)). Moreover, the V-to-U-shaped cuts (Fig. 5 (e6)) were similar in shape, width, and depth and could have been caused by multiple cuts with the same cutting tool moving from left to right. This pattern indicates decapitation from the front of the neck by a right-handed person. Among all cut marks on the two individuals (M103-1, M103-3), the width was between 0.43 mm and 0.53 mm for all nine marks, indicating that they were caused by the same type of tool.

Discussion

Coupled with missing heads, cervical vertebrae at the Honghe site had cut marks from sharp tools, indicating the practice of decapitation. Moreover, all the headless bodies were females and juveniles. Such an exclusive selection indicated either a ritual of selective decapitation or the result of an interpersonal conflict with a high level of cruelty. Nonetheless, along with the four bodiless heads of adult males found in a pit, the phenomenon of decapitation at the Honghe site indicates the presence of a conscious headhunting behavior. The practice is deeply rooted in the history of violence and interpersonal conflicts in developing and complex human societies. This is the first attempt to reconstruct the human headhunting behavior of prehistoric China. It is greatly significant in decoding the complex social and cultural behaviors of the people in the Nenjiang River Basin of northeastern China. This work enriches not only knowledge of the late Neolithic period, but the history of northeast Asia in general.

Events of headhunting at the Honghe site

It is hard to speculate how many headhunting events occurred at the Honghe site based on our dating results, which roughly placed all events into one (yet not so narrow) time bracket (Appendix Table 3). If the locales of the decapitation signs are used to dictate the number of episodes, there might be up to six different headhunting events, including [1] the event in tomb M101 with 3 victims, [2] the event in tomb M103 with 3 victims, [3] the tomb 2013-M103 with two victims, [4] the event in house F0004 with at least 21 victims, [5] the event in house F10, and [6] the burial pit outside of F10 with four victims. However, determining whether the different groups in houses F0004 and F10 were from separate events or from an all-in-one event with people that belonged to different families is subject to more in-depth studies. Additional work would include precision dating and aDNA studies (i.e., to determine the relationships among victims). A parsimonious reconstruction of events would suggest that events of F004 and F10 might be a single mass headhunting event (at least 32 victims), and perhaps the settlement was totally abandoned after the remains of females and children were laid in two houses; the individuals in tombs M101, M103, and 2013-M103 might be from settlements of the Honghe site and victims of earlier headhunting events. The event leading to the separate head pit outside of the F10 was prior to the mass headhunting event too, in which four victims were probably executed by the Honghe residents.

Nonetheless, the Honghe site might have yielded the largest number of decapitations (45: 41 plus 4), or the largest single headhunting event (32: F004 and F10) not only in prehistoric China but also in prehistoric Asia.

Methods and tools of decapitation

In most headhunting victims, the first two cervical vertebrae, the axis and atlas, were absent. The 3–7th cervical vertebrae were preserved, which indicated that the headhunters implemented decapitation in a standard way, only cutting for a clearly severed head without much of the neck attached. The cut marks were exclusively located on the front surface of the axis. Therefore, the decapitation was from the ventral side — with one (left) hand tilting the head of the victim backwards, the headhunter held an implement with sharp blade in their right hand, cutting the head off (Fig. 6a). The axis is the second segment of the cervical vertebra, which is located deep in the neck and surrounded by many soft tissues, from skin to muscles, facia, ligaments, glands, nerves, and blood vessels. Between the mandible and the hyoid bone, the axis could be reached using a sharp tool that could accurately penetrate the upper part of the neck for repeated cutting. In most cases, the goal of headhunting was achieved through separating the head and the axis, and between the axis and the third cervical vertebra (Fig. 6b).

Fig. 6
figure 6

Method of decapitation. The decapitation was from the ventral side — with one (left) hand tilting the head of victim backwards, the headhunter held an implement with sharp blade in their right hand, cutting the head off (a). In most cases, the goal of headhunting was achieved through separating the head and the axis, and between axis and the third cervical vertebra (b)

Full size image

There was consistency in the placing of cuts, the direction of cuts, and perhaps the use of cutting tools. The cut marks on Honghe victims were all narrow, and V- to U-shaped, indicating the presence of a very sharp stone blade. Traumatic experimental studies show that cuts produced by stone blade implements such as flint exhibit more variable shapes along their length, creating irregular V-shaped profiles or even varying from a V to U shape due to jagged blades. In contrast, most of the cut marks produced by metal blades show a more regular shape (Boschin and Crezzini 2012; Greenfield 1999). During the Neolithic Age in northeastern China, people used polished stone implements (microlith) that were well-developed for their fishing and hunting lifestyle. For example, sharp tools with bone or stone blades were used to facilitate the dissection of fish bellies and dismember prey (Dong 2012; Liu 2021; Song 2016; Zhao 2003). Compared to tools with bone blades, tools with stone blades were stronger and more conducive to cutting and thus more widely used. The Honghe villagers performed a fishing, hunting, and farming subsistence mode using bone tools and stone tools. Among all implements, bone-handled tools with stone blades were the most distinctive. The handles were made of animal bones, and stone blades were made of agate or flint. The bone-handled cutting tools were either designated a “single-bladed knife” or “double-bladed sword.” Many such cutting tools have been found at Honghe and Neolithic Age sites in the surrounding area (Fig. 7). A double-bladed sword was buried with other bone and stone tools in tomb M104 at the Honghe site (Fig. 7 (11); Appendix Fig. 8). From the location and shape of the cut marks on the headhunting victims at the Honghe site, the decapitation could be achieved with the use of sharp implements with bone handles and stone blades as found in the Honghe area. There was consistency in the placement and direction of the cuts and in the use of weapons or tools.

Fig. 7
figure 7

Cutting tools with bone handles and stone blades from Neolithic Age sites such as those at the Honghe site and other sites in northeast China. There are “straight-blade type” (1, 3–6, 9, 11) and “arc-blade type” (2, 7, 8, 10) cutting tools according to the different curvature of the blade (Zhang 2008; Song 2016; Liang et al. 2021). Tools with single sided blades tools are knives (1–4, 7–10) and bilateral bladed tools are swords (5, 6, 11). Tools in Images 1–6 are from the Neolithic Age Nanbaoligaotu site in northeast Inner Mongolia, and tools in Images 7–11 are from the Honghe site. Images 1–8 are from Zhang (2008) and Song (2016); Images 9–10 are from Liang et al. (2021). Reuse of Images 1–10 with permission. Image 11: this study

Full size image

Compared to the headhunting methods used during the Bronze Age or Iron Age in China, the decapitation at the Honghe site was equally effective. From the morphological characteristics of cut marks during the Iron Age, heavy and sharp metal tools were used to cut the entire head off at once. There are several cases that reflect this method of chopping, such as the beheading of soldiers in Lu’an, Anhui during the Warring States Period in China, the Crusader in Sidon, Lebanon, and the Towcester in Rome (Anderson 2001; Mikulski et al. 2021; Zhou et al. 2020). During the Stone Age, sharp tools were smaller in size, and larger tools were generally blunt. In the process of decapitation, the assailant made multiple cuts in the vertebrae that had to be repeated. This pattern reflected the style of cut, such as seen in the decapitation of prehistoric humans at the Lagoa Santa karst site in Peru and the Tell Qaramel site in Syria (Kanjou et al. 2015; Strauss et al. 2015). In terms of the location, the cut marks during the Stone Age were on the front of the cervical vertebrae, reflecting the action of cutting from front to back. During the Bronze or Iron Age, cut marks were located at the back of the vertebrae, reflecting that the beheading was executed from back to front.

The sociopolitical background and consequences of the mass headhunting event at the Honghe site

The multiple headhunting events at the Honghe site are the direct evidence of violence or interpersonal conflicts that prevailed during prehistoric times in northeastern China as part of northeast Asia. It is not possible to determine the exact motivation for the phenomenon of decapitation at the Honghe site. Three possible explanations follow with increasing confidence from one to the other. Headhunting motivations could have been (1) funeral customs or sacrifices within the tribe, (2) sacrifice of war prisoners from other tribes, or (3) atrocities from the invasion of an enemy tribe.

The existence of strategic and defensive trenches at the Honghe site reflect the frequent inter-tribal conflicts in the area. The burials of headless individuals reflect the fact that the tribe was attacked and that tribal members were headhunted more than once. Moreover, among the headless individuals found from the two houses, all the identified individuals were female and subadult. Therefore, the headhunting might have occurred in the form of a sneak attack on the vulnerable group. If the headless individuals from F004 and F10 were the victims of one headhunting event, at least 32 should have died (or 33 including the adult female with the intact head in F004). This loss would have been devastating to the Honghe settlement. Estimates of Neolithic family size associated with a single dwelling vary from 3 to 8 people, with the majority between 4 and 6 (Birch-Chapman et al. 2017). Considering there were 42 houses found at the same layer without overlapping, if all were occupied, theoretically the total population size at the Honghe settlement would be between 126 and 252; there would be around 210 if there were 5 members per dwelling on average. Thus, the loss of 32 people would have been a one-time 15.2% decrease in the total population, or a 22.9% loss in women and children (if females and children accounted for two-thirds of the population). Some families might have been completely ripped of their wife and kids. The large number of unburied and headless corpses at the Honghe site reflect what was the last devastation of Honghe by an external hostile force. The sense of loss would have been overwhelming.

A scenario that happened 4100–4400 years ago could be reconstructed as follows. Tribes living in the Nenjiang valley were fishers, hunters, and farmers. Different tribes were hostile to each other, competing for resources or clashing for other reasons. Warfare was not uncommon, often leading to fatal consequences. Sometimes the Honghe tribe would return as victors with trophy heads, yet often their settlement was attacked, and females and juveniles were beheaded. Eventually, during one occasion while all male adults were away from the settlement for sustenance activities or engaging in warfare, a group of people attacked the Honghe site. Using bone-handled knives with stone blades, the attackers killed many females and juveniles in the settlement. The victims were crudely decapitated, with their heads taken away by the enemy. When the survivors gathered back and male residents returned, they moved the corpses to two houses for a simple burial and then abandoned the settlement.

The Honghe site did not have a significant cemetery, indicating that the settlement had a short history. Yet the history of violence had been well preserved through the records of multiple events of headhunting. The Honghe site provides us a rare window into many aspects of prehistoric life, and some beliefs behind the practice which were unknown, perhaps for taking souls and energies, or as a “propaganda by terror.” It is puzzling that there were no other signs of violence on the skeletons, i.e., lack of signs of perimortem trauma on the postcranial skeletons, suggesting the possibility of that these victims were beheaded alive. It is hypothesized that this phenomenon was related to certain sorcery or ritual beliefs, such as a kind of Shamanistic practice. How the headhunting activities at the Honghe site reflects the views of life and death of the people during the Neolithic Age, which awaits further comprehensive archaeological and ethnographic studies of this site and other Neolithic sites at the Nenjing River Valley area, and the greater Northeast Asia alike. The study of headhunting culture would help not only reconstruct the history of violence in Northeast Asia but also probe into the thinking and ideology of human societies with headhunting cultures.

Conclusion

Osteoarcheological studies of the Honghe site revealed that there were signs of a mass decapitation event 4100 to 4400 years ago, indicating the practice of headhunting activities during the Neolithic Age in northeastern China. Moreover, the headhunting victims were exclusively female and juvenile, reflecting the cruelty of ancient warfare at the Honghe site. Cut marks were observed on the cervical vertebrae of five individuals, indicating their heads were cut from the ventral side of the neck. The cutting tools were probably the bone-handled implements with stone blades found in the Honghe area. The Honghe settlement might have been abandoned after the mass decapitation event. This is the first attempt to reconstruct human headhunting behavior from prehistoric China, and it is also the largest headhunting activity of the Neolithic Age in Asia. The study of headhunting culture would help not only reconstruct the history of violence in Northeast Asia but also probe into the thinking and ideology of human societies of headhunting culture.