The frozen body pulled from an Alpine glacier in 1991 did not belong to a man who died of exposure, starvation, or a predator attack. Ötzi, preserved for 5,300 years in the Ötztal mountains, was killed by another person. Forensic analysis found an arrowhead embedded in his left shoulder, fired from behind, which struck the subclavian artery. He also carried a fresh defensive wound on his right hand from a struggle earlier in the same day.
DNA from four separate individuals was found on his weapons, tools, and clothing. The evidence points not to wildlife but to a deliberate, coordinated attack by other humans. Ötzi is not an isolated anomaly. In 1983, archaeologists uncovered 34 skeletons in a ditch near Talheim, Germany, dating to roughly 5,000 BCE.
Men, women, and children were killed in a single event, most struck in the skull with blunt instruments. Several victims had fractures on their forearms consistent with raising their arms to block incoming blows. Researchers concluded the attackers intended not to raid and leave but to wipe out the local community and take over their land. A similar site at Asparn Schletz, Austria, from the same period, contained over 200 individuals killed in the same pattern.
This evidence forms part of a broader archaeological record on prehistoric violence. Lawrence Keeley, an archaeologist at the University of Illinois at Chicago, published findings in 1996 concluding that prehistoric warfare was proportionally more lethal than 20th-century armed conflicts. In some populations, casualty rates from interpersonal violence ranged between 10 and 60 percent. Among the Jivaro of the upper Amazon, 59 percent of male deaths resulted from violence.
For the Yanomamo, the figure was 37 percent. Across pre-state societies, roughly 25 percent of men died not from disease, hunger, or animals, but from other humans using coordinated tools designed for killing. Predators did kill people, but the threat they posed operated differently. A lion hunts opportunistically; it does not plan campaigns or coordinate approaches.
A cave bear guards territory, not strategic objectives. Wolves generally avoid capable adult humans carrying weapons. Human attackers, by contrast, could identify patterns, wait for vulnerable moments, approach from blind angles, and use projectiles from a distance. If an initial attempt failed, they could return with a revised approach.
Projectile points designed to remain embedded in wounds, maximizing internal damage, have been found at sites thousands of years older than any formal military structure. They were engineered to stop a person, not to hunt deer. The built environment reflects this threat. Jericho in the Jordan Valley, dating to approximately 8,000 BCE, featured a stone wall at least 13 feet high and a stone tower 28 feet tall.
According to Britannica, these structures were explicitly intended to protect the settlement and its water supply from human intruders, representing the earliest technology ascribed without question to purely military purposes. At Çatalhöyük in modern Turkey, around 7,500 BCE, houses had no ground-level doors. Residents entered through rooftop hatches using ladders, and the settlement’s outer perimeter was formed by continuous house exteriors with no street-level entry. Attackers would have to climb in full view of the inhabitants.
Human biology also reflects this history of interpersonal violence. The brain contains a threat detection system for unfamiliar human faces that activates faster than responses to predator silhouettes. Social exclusion, betrayal, and rejection trigger the same neural threat circuitry as physical danger. In the resource-scarce environment where human nervous systems evolved, removal from one’s group was effectively a delayed death sentence.
A 2016 study published in Nature examined violence patterns across Pleistocene and early Holocene populations and found evidence of interpersonal violence in essentially every period studied, including healed forearm fractures, embedded projectile points, and skulls with penetrating trauma. These patterns appeared consistently across populations that had no contact with one another. The record shows that predators were real, but the threat that required stone walls, rooftop-entry architecture, and a nervous system wired to detect deception and track group loyalty was not an animal.
It was the person on the other side of the ridge.
