From grooming to gossip
Monkeys and apes keep their friendships through grooming: cleaning each other's fur. It builds trust and alliances, but it takes time, and a primate can groom only one partner at a time.
The British anthropologist Robin Dunbar noticed two patterns. Across primate species, those with a larger neocortex (the outer layer of the brain) tend to live in larger groups, and in larger groups they spend more of the day grooming. Applying that pattern to the human brain, he estimated a natural group size for humans of about 150 people, a figure now known as "Dunbar's number". Keeping that many relationships alive through grooming would take around 40% of the day, far more than any primate spends.
Dunbar's hypothesis, set out in his book Grooming, Gossip and the Evolution of Language (1996), is that language took over the job. He called it "vocal grooming". You can talk with several people at once, and talk lets you learn about people you never see: who is reliable, who helped whom, who broke a promise.
Everyday talk fits the idea. When Dunbar and his colleagues listened in on conversations in public places, about two-thirds of the talking time went to social topics: people, relationships, what someone said or did.
The details are debated. A 2021 reanalysis found that the primate data allow a very wide range of "natural" human group sizes, not one clear number (Lindenfors et al. 2021). And group size in primates varies a lot: chimpanzee communities usually number a few dozen, but the largest known one, at Ngogo in Uganda, has had around 150 members, at times even more.
Other answers
Gossip is one hypothesis among several. Others link the first advantages of language to:
- Cooperation: planning a hunt or sharing work. The psychologist Michael Tomasello sees the human urge to share goals and attention as the root of language.
- Teaching: in an experiment, people learned to make simple stone tools much better when a teacher could talk to them than when they could only watch (Morgan et al. 2015).
- Choosing a partner: the psychologist Geoffrey Miller has argued that skilled talk worked partly as a display, a way to impress.
These ideas do not exclude each other. Language may have paid off in several ways at once.
What syntax adds
Whatever the first reason, gossip shows what words alone cannot do. "Bob" and "meat" do not tell you whether Bob found the meat, stole the meat or cooked it, or whether the meat is Bob's. You need rules for how words relate to each other: syntax.
The classic example is "The dog bites the man" and "The man bites the dog": the same words, a different event. English signals who did what mainly through word order. Many languages use word endings instead. In Hungarian, az ember (the man) becomes az embert when he is the one being bitten, so the words can change places without changing who bit whom. (Our article Who did what to whom? looks at these systems in more detail.)
The linguist Noam Chomsky made this combining power central to the study of language. A speaker with a limited set of words and rules can produce and understand an unlimited number of sentences, most of which they have never heard before. His famous sentence "Colorless green ideas sleep furiously" (1957) makes no sense, yet any English speaker can tell that it is grammatical, while "Furiously sleep ideas green colorless" is not.
One part of this power is nesting: putting sentences inside sentences, as gossip loves to do. "Ana told me that Bob thinks that Carla took the meat." Hauser, Chomsky and Fitch (2002) suggested that this kind of nesting may be the core of what is unique to human language. Others disagree, and whether every language uses it is debated as well.
Animals are not without combinations. Japanese tits, small birds, combine two calls in a fixed order to mean "look around and come closer", and they react less when researchers play the calls in reverse order (Suzuki et al. 2016). What sets human syntax apart is its scale: an open-ended system in which any sentence can be expanded, rearranged and embedded in another.
Could Neanderthals talk?
Neanderthals lived in Europe and western Asia until about 40,000 years ago. They were strongly built, and their brains were on average about as large as ours or slightly larger. For a long time, many researchers assumed that their speech was limited, partly because of early reconstructions of their throat (the debate from part 2).
Newer evidence points to more capacity than that:
- They had the same version of FOXP2 as we do (Krause et al. 2007).
- A Neanderthal hyoid bone, a small bone in the throat that supports the tongue, found in Kebara Cave in Israel, looks much like ours (Arensburg et al. 1989).
- Models of their outer and middle ear, built from CT scans of fossil skulls, suggest that their hearing was tuned to the frequencies of speech much like ours (Conde-Valverde et al. 2021).
Anatomy shows what a body could do, not what its owner actually said, so the question remains open. So does the question of why our species spread while Neanderthals disappeared. Language may have played a part, but so may group size, climate and competition. And the two groups met: most people whose ancestors lived outside Africa carry a small share of Neanderthal DNA.
What's next
By now we have a species with symbols, a body and brain that can speak, and a grammar that can say almost anything. But if all humans share the same capacity for language, why are there thousands of different languages? In part 4, we look at how languages split apart, change and become part of who we are.