Both, and the bones can’t tell us which one ran daily life. Skeletons from before writing include people who survived serious injuries, sometimes in ways that researchers argue needed help from others. Other skeletons carry wounds made on purpose, in cases reaching back about 430,000 years. What bones can’t show is what anyone felt, meant, or believed. So any claim that humans are “naturally” kind or “naturally” cruel goes further than the evidence does.
How to read the labels on this page: [Documented] means the evidence shows it directly. [Strongly supported] means strong sources agree. [Plausible] means the evidence allows it but doesn’t prove it. [Disputed] means qualified experts disagree. [Unknown] means the evidence can’t answer the question.
Can one cave hold both stories?
Between 1951 and 1960, archaeologist Ralph Solecki dug a trench in Shanidar Cave, in the mountains of Iraqi Kurdistan, and found the remains of ten Neanderthal men, women and children. The team now digging there describes two of them in one sentence: the injuries of Shanidar 1 “suggest care for group members,” and the rib wound of Shanidar 3 “suggests interpersonal violence.” The team’s first new dates, which it calls “initial,” put both men at roughly 55,000–45,000 years ago. [Plausible for the dates]
Shanidar 1’s skeleton shows a healed fracture around the left eye socket, healed breaks in the right upper arm with the forearm and hand missing, changes in the legs and feet, and bony growths in both ear canals that nearly closed the right one (Trinkaus & Villotte, 2017). Bone only heals in a living body, so he survived the fractures. [Documented for the bone changes] The growths probably caused serious hearing loss, but nobody could test his hearing. [Plausible] Trinkaus and Villotte argue that a person with advanced hearing loss would have been “highly vulnerable alone,” and that, with his other impairments, “a substantial degree of social support” is indicated. Anthropologist K. A. Dettwyler warned back in 1991 that disabled people may contribute differently, “but they do contribute,” often in ways bones can’t show. She suggested that Shanidar 1 might have “gotten along quite well” gathering plants, preparing food and cooking (Dettwyler, 1991). A 2025 reanalysis by K. M. Kinkopf questions the timing of the injuries and emphasizes evidence of active use of the arm. Its interpretation allows Shanidar 1 to have given care as well as received it (pp. 10–11, 14). Some help from others: [Plausible]. That he needed care to survive: [Disputed]. Why anyone helped him: [Unknown]. For the full case, see our pages on care for the injured and on ancient trauma.
Shanidar 3 has a puncture wound in his left ninth rib with new bone starting to form around it. The authors interpreted this as survival for at least about two weeks after the injury (Churchill et al., 2009, p. 166). [Documented for the wound and healing; Plausible for the estimated interval] The researchers stabbed pig carcasses with replica stone points and concluded the wound best matches a light, thrown weapon. They also say they can’t rule out an accident or a hand-held weapon (p. 176). [Plausible for a thrown weapon; Unknown who threw it]
Analogy: a healed bone is like a receipt. It proves the body survived and repaired itself, but not who helped, how, or why.
How far back does deliberate violence go?
The oldest case we checked for this page is Cranium 17 from Sima de los Huesos, a cave site in Spain’s Atapuerca hills. The site holds fossils of at least 28 individuals, about 430,000 years old, from the lineage that led to Neanderthals (Sala et al., 2015). The skull has two fractures on the left side of the forehead. Neither shows any healing, the two are nearly the same shape, and they came from different angles. The team concludes that two blows from the same object were struck face to face, and that either blow was likely deadly on its own. Bodies entered the site down a 13-meter shaft, but the team called it “highly improbable” that a fall could have made both fractures, since the same object would have had to strike the skull twice. [Documented for the fractures; Plausible for deliberate killing, argued in detail by one research team] The authors call it the earliest clear case of deliberate lethal violence in the hominin fossil record (the fossils of humans and our extinct close relatives). That holds only for fossils found so far, so we don’t call it a “first.”
When does violence show up between groups?
Nataruk, near Lake Turkana in Kenya, dates to about 10,500–9,500 years ago by the researchers’ estimate, though they don’t rule out an older age. The authors identified major injuries in ten of twelve partly preserved skeletons and interpreted them as lethal. The injuries include wounds to the head and neck, likely from arrows, and a small obsidian blade still lodged in a skull. Two others lay with their hands in positions that suggest they may have been tied (Mirazón Lahr et al., 2016, full text, Cambridge repository, pp. 4–6). The authors read it as one group attacking another. A published comment is titled “Contesting the massacre at Nataruk”. We don’t summarize its arguments here because we could not read its full text. A massacre at Nataruk: [Disputed]
Are humans the only animals that kill their own kind?
No. A 2016 study compiled causes of death for 1,024 mammal species and about 600 human populations. It found that killing members of one’s own species is frequent among primates and uncommon in groups like bats and whales. Using the mammal family tree, the study predicted that about 2% of human deaths would come from violence between people (Gómez et al., 2016). That figure is close to the study’s estimate for prehistoric bands and tribes. Later rates changed a lot, and the study links those changes to how societies were organized. [Plausible: a model estimate from one large study] The authors also point out that skeletons could undercount deadly violence, because many fatal wounds never touch bone. Where they could compare skeletons with written death records, though, they found no sign of that undercount.
How did scientists argue about this before?
In 1888, Thomas Huxley, whom Kropotkin called one of the most prominent Darwinists, published an essay on the “struggle for existence.” Peter Kropotkin thought it gave a false picture of nature. His answer became a series of articles in the magazine The Nineteenth Century between 1890 and 1896, collected in 1902 as Mutual Aid: A Factor of Evolution. Kropotkin didn’t deny struggle. He argued that mutual aid “is as much a law of animal life as mutual struggle” and probably mattered more for evolution. He also called his book a study of one of the chief factors in evolution, not all of them. Some of his sources were his own travels in Eastern Siberia and northern Manchuria. His chapter titles label other peoples “savages” and “barbarians,” a reminder that even a book about cooperation carried its era’s rankings of humanity. [Documented for what he wrote]
Later scientists turned the question into games. In two computer tournaments, summarized in a 1981 paper with biologist W. D. Hamilton, political scientist Robert Axelrod pitted strategies for the Prisoner’s Dilemma against each other. In this game, two players each choose to cooperate or betray. Both rounds were won by the simple strategy TIT FOR TAT, submitted by Anatol Rapoport of the Institute for Advanced Study in Vienna: cooperate first, then copy your partner’s last move (Axelrod & Hamilton, 1981, pp. 1392–1393). In a 2002 study, 240 university students in Zurich played a money game. Cooperation flourished when players could pay to punish free riders (people who take the benefits without chipping in) and broke down when they couldn’t. Asked afterward about imagined free riders, most players reported strong anger toward them (Fehr & Gächter, 2002). The authors argued that older explanations, like helping relatives (kin selection) or trading favors (reciprocal altruism), can’t explain why people cooperate with strangers they’ll never meet again. A 2003 brain-scan study of 19 people found that receiving an unfair offer activated the anterior insula, a region the authors link to negative feelings like anger and disgust. People with stronger activation turned down more unfair offers (Sanfey et al., 2003). These studies tell us about people today. They’re not evidence about how anyone behaved 10,000 years ago. How far punishment explains human cooperation in general: [Plausible]
How we know
What evidence exists: skeletons, healed versus unhealed bone, stone points lodged in bodies, dating of sediments and shells, experiments on animal carcasses, modern lab games, and computer models built from death records.
Whose voices are missing: everyone in these stories. No one at Shanidar, Sima de los Huesos or Nataruk left words. We don’t know their names or languages, or how they explained a wound or a death. People who cared for the injured leave almost no trace. Fear, grief and lasting distress usually leave no trace in bone. And the modern experiments come from small, specific groups, like Zurich university students.
How power shaped the record: A bone lesion is more directly visible in this evidence than an act of everyday help. Where remains end up matters too. Shanidar 1 was examined in the Iraq Museum in Baghdad in 1976–78, and fresh digs at the cave began in 2015 after a delay caused by the threat from ISIS. The sources used here include no present-day Kenyan, Iraqi or Kurdish views on how these remains are studied or shown.
Then vs. now
- Then (what bones show): before writing, people were deliberately hurt and killed, sometimes in groups, and some people survived serious injuries. Now: we can count wounds and healing, but not how common violence or care was in any whole community. [Unknown]
- Then (what 1902 thinkers believed): Kropotkin argued that Huxley’s struggle-centered picture of nature left out mutual aid. Now: researchers study both sides with fossils, field data, models and experiments. The idea that humans are the only killers of their own kind doesn’t hold up (Gómez et al., 2016).
- Then (popular retellings): Solecki argued that Shanidar 1 showed compassion (Pomeroy et al., 2020), and a University of Cambridge feature still says “Signs of compassion.” Now: the bones can support survival and plausibly some help. They can’t show compassion. [Unknown]
What we still don’t know
- Which came first, violence or cooperation. The records we have can’t answer that, and it may be the wrong question. [Unknown]
- Whether early foragers waged organized war between groups. [Disputed]
- How often people were helped or harmed in any one ancient community. [Unknown]
- What any of these people thought or felt. [Unknown]