Ten years ago, the psychologist Steven Pinker published The Better Angels of Our Nature, in which he argued that violence in almost all its forms – including war – was declining. The book was ecstatically received in many quarters, but then came the backlash, which shows no signs of abating. In September, 17 historians published a riposte to Pinker, suitably entitled The Darker Angels of Our Nature, in which they attacked his “fake history” to “debunk the myth of non-violent modernity”. Some may see this as a storm in an intellectual teacup, but the central question – can we learn anything about the future of warfare from the ancient past? – remains an important one.

Pinker thought we could and he supported his claim of a long decline with data stretching thousands of years back into prehistory. But among his critics are those who say that warfare between modern nation states, which are only a few hundred years old, has nothing in common with conflict before that time, and therefore it’s too soon to say if the supposed “long peace” we’ve been enjoying since the end of the second world war is a blip or a sustained trend.

In 2018, for example, computer scientist Aaron Clauset of the University of Colorado Boulder crunched data on wars fought between 1823 and 2003 and concluded that we’d have to wait at least another century to find out. Clauset doesn’t think it would help to add older data into the mix; indeed, he thinks it would muddy the picture.

“It’s up to researchers who study sub-state-level violence to substantiate their claims that the dynamics of such violence are relevant to the dynamics of war and, in my view, they haven’t done a great job there,” he says.

Most researchers accept that there is a difference between war and interpersonal violence – and that these two things are governed by different forces – but there is disagreement over where to draw the line between them. Historian and archaeologist Ian Morris of Stanford University, author of War! What Is it Good For? (2014), is among those who say that the nature of collective violence hasn’t changed much in millennia, it’s just that human groups were smaller in the past. For him, therefore, a massacre of a couple of dozen of hunter-gatherers in Sudan around about 13,000 years ago, the earliest known example of collective violence, is relevant to a discussion of modern warfare.

Archaeologist Detlef Gronenborn of the Römisch-Germanisches Zentralmuseum in Mainz, Germany, agrees. In 2015, he and others described a massacre among Europe’s earliest farmers at a place called Schöneck-Kilianstädten in Germany, about 7,000 years ago. More than two dozen individuals were killed by blunt force instruments or arrows and dumped in a mass grave, their lower legs having been systematically broken either just before or just after death. The absence of young women from the group suggested that the attackers may have kidnapped them. Gronenborn says that massacres of entire communities were frequent occurrences in Europe at that time and that one of their hallmarks, judging by the human remains, was the desire to erase the victims’ identity. “The only difference between then and now is that of scale,” he says.

But while some researchers may agree with Pinker that prehistoric and modern warfare are essentially the same phenomenon, they don’t necessarily agree with him that the evidence points to a long-term decline. Pinker based his claim that prehistory was extremely violent on around 20 archaeological sites spanning 14,000 years. Those sites unequivocally attest to ancient violence, says historian Dag Lindström of Uppsala University in Sweden, “but they cannot be used for quantitative comparative conclusions”. We simply have no way of knowing how representative they were.

“The further you go back in time, the more difficult it becomes to have an accurate assessment of how many people died in battle,” says historian Philip Dwyer of the University of Newcastle in Australia, who co-edited The Darker Angels of Our Nature. Civilian death counts are even less reliable, he says, and have likely been significantly underestimated throughout history. In Dwyer’s view, all war-related statistics are suspect, undermining attempts to identify long-term trends.