Anthropologists generally agree that eating meat was a pivotal change in early human diets. But they are still piecing together exactly how we evolved from big-stomached plant-eaters into big-brained meat-eaters about 2 million years ago.

For example, a study published this March in the academic journal Current Anthropology, suggests that if you want to eat like our earliest human ancestors, you ought to learn how to make bone broth. And maybe collect some roadkill too. 

The paper relies on the earliest evidence of butchery to argue that eating meat may have begun about 3 million years ago, when our ancestors collected bone marrow and brains left uneaten by Pliocene predators like big cats.

This idea is part of a broader shift in anthropologists’ thinking about how our early ancestors started their carnivorous careers. For decades, anthropologists thought the earliest humans began by hunting small prey before graduating to larger animals later in our history. Then, in the 1980s, some scientists began proposing that scavenging may have predated hunting.

“The question is when early humans first started eating meat 2 to 3 million years ago, were they hunting those animals, [or] were they scavenged?” said Briana Pobiner, a paleoanthropologist at the Smithsonian’s Human Origins Program, who was not involved in the new study. “This has been a raging debate, in a sense, in the paleontological research community for decades.”

Ideas about the carnivory styles of early humans started in the 1950s with the “killer ape” hypothesis—the idea that there is something innately bloodthirsty within human nature. Over the decades, a number of other archaeological sites with early human remains were discovered resting in sediment alongside simple stone tools and butchered animal bones, bolstering the idea of “man-the-hunter.” This all led to the idea that flaked stone tools and meat eating defined our Homo genus when it emerged in the archaeological record 2.6 million years ago.

“It’s a very appealing story,” Jessica Thompson, a paleoanthropologist at Yale University and author of the recent bone marrow paper, told Sapiens. “Right around that time there appeared to be the first stone tools and butchery marks. You have the origins of our Homo genus. A lot of people like to associate that with what it means to be human.”

However, those early human sites don’t contain weapons alongside stone butchery tools—a problem with interpreting how early meat eaters got their grub. It’s possible early people pummeled their prey with rocks, threw wooden spears that later decomposed, or outran their quarry.

Henry Bunn, a paleoanthropologist at the University of Wisconsin-Madison, has tried to get around the lack of weapon evidence by comparing modern hunter-gatherer bones to fossil bones from an important Tanzanian archaeological site. In a 2014 paper, he studied the age patterns of kudu and impala that Hadza hunter-gatherers ambushed with longbows and poison-tipped arrows. He found that modern hunters’ selection of prime-age animals matches bone ages found in the Olduvai Gorge 1.8 million years ago. When combined with other evidence, this points to hunting rather than scavenging at that site.

Bunn thinks interpreting early humans as opportunistic scavengers sells our ancestors short. “If you’re investing in the labor to carry meat a significant distance before eating it and sharing it, that it implies a more human-like social pattern,” Bunn said. “For me, it ultimately then means that many anthropologists have underestimated the capabilities of early Homo.”

For others, the lack of hunting tools in that part of the archaeological record, along with specific cut mark patterns on fossil bones, point towards “man-the-scavenger.” Early humans could have collected meat by chasing big cats off their kills or waiting to pick over the scraps.

Some critics of the scavenging hypothesis think wild carnivores may not leave behind enough carrion for early humans to make a living. Pobiner tried to address this by becoming a scavenger. After spending months in the Kenyan savanna collecting and analyzing lion kills, she calculated that the leftover meat from a single zebra carcass could have provided almost 6,100 calories. That’s enough to meet the daily caloric needs of three Homo erectus adult males.  

The new bone marrow paper provides a twist on the scavenging hypothesis that intrigues Pobiner. “The assumption has always been that the primary target of animal tissue was meat,” she said. “And I think one of the most interesting things about Thompson and colleague’s paper is that they’re saying maybe the beginning of human predation was actually all about just getting access to bone marrow.”

Anthropologists have typically viewed meat and marrow as resources that are collected together, as Bunn observes with modern Hadza hunter-gatherers. Now Thompson argues that scavenging primarily for marrow and brains may have helped early hominids evolve into the big-brained creatures we are today.

Marrow could have had an advantage over scavenged flesh because harvesting it might lead to fewer encounters with predators, bones are handy packaging for delaying spoilage, and marrow doesn’t require the extensive chewing work that raw meat does. Much like modern chimps, our Australopithecus ancestors likely spent half the day chewing plants. Plus, brains and marrow could have provided fatty acids important in human brain and eye development.

As potential evidence for this, Thompson and colleagues point to the oldest fossil bones that show possible evidence of butchering. Found in 2010, the 3.4 million-year-old bones from Dikika, Ethiopia, push the earliest known date of meat-eating forward by a million years. Australopithecus afarensis was the only known human ancestor around then, so this means flaked stone tool and meat eating may not be unique to the Homo genus after all.

Some anthropologists, like Bunn, are skeptical of the butchery interpretation, thinking that trampling, rock scratches, or crocodile teeth may have marked the Dikika fossils. Thompson, however, thinks the bones show traces of human scavengers smashing them open with rocks to access bone marrow. While the bone marrow proposal needs more study, Thompson and her colleagues hope it will push anthropologists to take a fresh look at how human predation got started.

Pobiner, however, thinks figuring out how early humans obtained meat doesn’t need to be an either-or debate: “We shouldn’t assume that there was just one way to be an early human.”

What is uniquely human, she said, is the habitat of butchering multiple animals at the same place over and over, as opposed to stumbling over the occasional zebra leg every once in a while.

Pobiner calls this behavior “persistent carnivory” and it first appears at a 2 million- year-old Kenyan archaeological site called Kanjera South. There, Pobiner and her colleagues have found evidence that people transported dead animals and tools back to the same place over hundreds to thousands of years. Early humans likely hunted smaller, antelope-sized animals and scavenged for larger ones, cracking skulls open to access brains.

While the scientific debate continues over how exactly we should imagine the first time humans used animals for food, we do know that around 2 million years ago people started bringing meat back to central locations for processing. This implies the start of food sharing, food preservation and social behaviors that ultimately may have made us human. 

For modern omnivores looking to connect with our deep ancestry, hunting and scavenging alone may not singularly define our Homo genus. But sharing our spoils likely does.