Key Takeaways
- Ljiljana Progovac’s analysis proposes that human language evolved partly due to sexual selection, emphasizing wit and verbal competition.
- Early language primarily involved simple verb-noun combinations, suitable for insults and humor, rather than complex communication.
- Progovac argues that ‘living fossils’ in grammar provide insight into early human communication and social dynamics.
- Neuroimaging studies suggest that processing ancient grammatical structures shares neural pathways with face recognition and humor.
- The link between linguistic ability and reproductive success indicates that cleverness and humor played crucial roles in human evolution.
The silverback doesn’t need words. He displays, he charges, he thumps his chest; the competition, if they have any sense, backs away. For most of evolutionary history, that kind of physical theatre settled the question of who got to mate and who didn’t. Then something changed, somewhere in the deep human past, and the contest began to shift. Not to friendship, exactly. To wit.
A new analysis from Wayne State University linguist Ljiljana Progovac, published this week in PNAS Nexus, makes the case that human language didn’t evolve primarily to help us communicate complex thoughts or forge cooperative alliances. It evolved, at least partly, because clever people got more sex.
The proposal sounds like something you might hear at a dinner party rather than in a peer-reviewed journal. But Progovac’s argument is technically precise, grounded in syntactic theory, neuroimaging data, and a rather remarkable archive of filthy medieval compound words. The core claim is this: from the very first moment that our ancestors started combining two words together, those combinations opened up a new kind of competition. Not physical. Verbal. The ability to coin a devastating nickname, to skewer a rival with a phrase, to make the tribe laugh at someone else’s expense; these skills, she argues, were subject to sexual selection just as surely as the peacock’s tail or the bowerbird’s elaborate stick construction.
That’s the standard view, and it’s not entirely wrong. But Progovac’s argument is that complex information-sharing was probably not what early language was mainly for, since the earliest grammar was too rudimentary to express much beyond simple verb-noun combinations. What those early combinations were well-suited for was nicknaming, insult, and humor, which suggests sexual selection may have driven the initial takeoff before language became elaborate enough for richer communication.
A grammatical structure that survives in modern languages but reflects a much older, simpler stage of grammar. Words like “killjoy,” “pickpocket,” or “scatterbrain” follow a verb-noun pattern that doesn’t fully obey the rules of modern English syntax (a killjoy isn’t a kind of joy), suggesting they predate the more complex grammatical machinery we use today. Progovac argues these compounds are our best window into what the earliest human grammar actually looked like.
Sexual selection acts on any trait that is variable within a population, heritable, and reliably correlated with status or reproductive access. Eloquence and verbal agility meet all three criteria. The trait also has honest-signal properties: you can’t convincingly fake superior linguistic ability in real time the way you might fake other qualities, which is part of what makes it a reliable marker of underlying fitness.
Not wrong, but perhaps incomplete. Progovac accepts the evidence for human self-domestication and reduced reactive aggression, and her model actually predicts it: as verbal competition replaced physical fighting, genetic pressures favoring lower aggression would have intensified. But she argues that “friendliness” alone doesn’t explain why language specifically emerged, or why humans are uniquely preoccupied with verbal cleverness and humor as markers of worth.
The “survival of the friendliest,” as researchers have called it, gets something right but also misses something important. Brian Hare at Duke University has argued influentially that Homo sapiens evolved through selection for prosociality, a kind of self-domestication that reduced reactive aggression and made us more cooperative. Progovac doesn’t dispute the self-domestication evidence, but she does push back on the framing. “Friendly” types, she points out, still had to compete ruthlessly against “unfriendly” ones, and that competition sometimes included ganging up to kill dominant males. What we really selected for, she suggests, was groupishness: the ability to cooperate within a group in order to out-compete other groups. Friendly, maybe. Innocuous, not really.
More importantly for her argument, none of this gets at what makes humans specifically human. Bonobos are friendly. Dogs are friendly. Neither of them tells jokes.
The linguistic heart of Progovac’s theory involves what she calls “living fossils”: ancient grammatical structures that survive, like coelacanths, in the corners of modern languages. The very earliest stage of grammar, she argues, would have been nothing more than a two-slot combination, one verb and one noun. No tense, no subject-object distinction, just two words smashed together. And remarkably, this kind of protogrammar is still visible in languages today. In English: killjoy, pickpocket, turncoat, crybaby, scatterbrain. In Serbian: a whole menagerie including “spin-butt” for a fidgety person and compound place names for steep hills that translate, roughly, as “rip-balls.” Thousands of these existed in medieval times. Most are now obsolete, often because, as one historical source puts it, they showed “unquotable coarseness.”
That coarseness is the point, Progovac argues. These compounds specialize for insult. They attach memorable, derogatory, body-focused nicknames to people. The ability to coin one on the spot, with whatever meager vocabulary and grammar you had available, would have been an unprecedented cognitive leap, and an immediately useful one. Across traditional societies worldwide, the most skilled speakers consistently hold higher social status (which in turn correlates with greater reproductive success). As one anthropological account of South American indigenous societies puts it, describing the status of oratory: it can be said not that the chief is a man who speaks, but that he who speaks is a chief.
The cross-cultural parallels are striking. English “tumble-dung,” meaning a dung beetle, has an almost identical structural equivalent in Twi-speaking Ghana: “kukru-bin,” literally “roll feces.” Old English “burst-cow” (an insect) maps to a Tashelhit Berber compound with the same image. These aren’t borrowings. They’re independent arrivals at the same metaphorical solution, which suggests the underlying cognitive mechanism goes very deep. Humor, it turns out, depends on exactly the same grammatical trick: two words collide in unexpected ways, incongruity arises, and the brain scrambles to resolve it. Metaphor and humor, in other words, may have the same evolutionary root.
Neuroimaging adds a layer of plausibility. In an fMRI study, Progovac and colleagues found that processing these flat, ancient verb-noun compounds activates the fusiform gyrus and inferior temporal regions of the brain, specifically in the right hemisphere, in ways that more complex grammatical constructions don’t. The fusiform gyrus handles face recognition; it also handles the kind of non-compositional semantic processing involved in metaphor. That early language structures seem to recruit the same neural real estate as face perception opens an intriguing possibility: perhaps the human capacity for naming faces (associating a vivid linguistic label with a specific person) and the evolution of early grammar are part of the same story. The “scatter-brain” and the “cry-baby” aren’t just insults. They’re faces, made of words.
The sexual selection dimension has some empirical support, though it’s necessarily indirect. An analysis of roughly 3,745 personal advertisements found that women sought a mate who could make them laugh at about twice the rate they offered to return the favor, consistent with sexual dimorphism in humor production. Separate fMRI work found that women show greater emotional activation than men during humor perception. And studies have found that males increase their linguistic creativity both in the presence of attractive females and in the presence of male competitors. About as clean an experimental proxy for intrasexual selection as you’re likely to get in a linguistics lab.
Progovac is careful about what her framing does and doesn’t claim. It isn’t that the cleverest people always outcompeted the dullest. It’s that the trait was variable, heritable, and correlated with status and reproductive success: the minimum conditions for selection to operate. She also concedes something a little deflating. Wittiness, as a form of fitness, is not the same as wisdom. We elect politicians based on facility with a sharp phrase, not their capacity for nuanced analysis. Quick-wittedness is immediately, viscerally observable. Good judgment takes much longer to evaluate.
The genetics hint at how deep the connection runs. FOXP2, one of the few genes with a direct known role in language and speech, is implicated in the enhanced connectivity of the Broca’s area-basal ganglia network that distinguishes humans (and Neanderthals) from other primates. This same network underpins humor appreciation and the regulation of aggression. Verbal wit, self-domestication, and the gradual displacement of physical fighting by cognitive contest may all be part of one feedback loop, each amplifying the others across hundreds of thousands of years. We became friendlier, in part, because being funnier was a better strategy than being stronger.
The bower bird builds its elaborate stick stage and decorates it with whatever shiny objects it can find, all in the service of impressing a mate. Magnificent waste of effort, by any functional measure. We do something roughly similar every time someone delivers a perfect one-liner at exactly the right moment. The difference is that our version, somewhere along the way, also gave us grammar.
DOI / Source: https://doi.org/10.1093/pnasnexus/pgag052
ScienceBlog.com has no paywalls, no sponsored content, and no agenda beyond getting the science right. Every story here is written to inform, not to impress an advertiser or push a point of view.
Good science journalism takes time — reading the papers, checking the claims, finding researchers who can put findings in context. We do that work because we think it matters.
If you find this site useful, consider supporting it with a donation. Even a few dollars a month helps keep the coverage independent and free for everyone.