The Grammar Nobody Taught

You speak a language nobody designed.

Not the slang you picked up from friends or the jargon your job shoved into your vocabulary. The deep structure — the invisible architecture governing how you arrange words, nest clauses, and signal meaning. Nobody sat down and decided that objects should cluster with their verbs, or that languages should prefer certain hierarchical structures over others. And yet, across 1,700 languages on every inhabited continent, the same grammatical patterns keep showing up like a cosmic punchline nobody told.

An international research team led by Annemarie Verkerk at Saarland University and Russell D. Gray at the Max Planck Institute for Evolutionary Anthropology just ran the most rigorous test of "linguistic universals" ever attempted. Using Grambank — the largest database of grammatical features ever assembled — they tested 191 proposed universal grammar rules with Bayesian spatio-phylogenetic analyses. The method accounts for shared ancestry and geographic proximity, eliminating the usual objections that similar grammars just mean borrowed grammars.

The result: roughly one-third of those proposed universals are statistically real. Not hunches. Not cherry-picked examples from a handful of European languages. Robust patterns that emerge independently across language families that have had zero contact for tens of thousands of years.

The confirmed universals cluster around word order and hierarchical structure — the bones of how humans organize meaning. Whether verbs come before or after objects. How relative clauses attach. The way languages nest information inside other information. These patterns recur across unrelated languages in different hemispheres like they're following a blueprint that was never written down.

This matters because it settles — or begins to settle — one of the longest-running arguments in the science of mind.

Noam Chomsky proposed "universal grammar" in the 1960s: the idea that humans are born with innate grammatical constraints hardwired into our cognitive architecture. The idea was controversial for decades. Critics argued that with enough diversity, you could find a language that violated any supposed universal. Chomsky's evidence was largely intuition and selective examples.

What Verkerk and Gray's team did differently was refuse to argue from intuition. They threw 1,700 languages at the question and let the evolutionary math sort it out. The patterns held. Not all of them — two-thirds of the proposed universals washed out. But the ones that survived are remarkably consistent.

"In the face of huge linguistic diversity, it is intriguing to find that languages don't evolve at random," Verkerk noted.

Here's where it gets cosmically weird. The design space of possible grammars is enormous. You could arrange the building blocks of language in a staggering number of configurations. But humans, independently, across millennia and oceans and mountain ranges, keep converging on the same handful of solutions. As Gray put it: "Shared cognitive and communicative pressures push languages towards a limited set of preferred grammatical solutions."

Translation: your brain is running grammar software that was compiled before your species had a word for anything.

This is emergence in its purest form. No one imposed these rules. No language academy dictated word order to pre-literate societies on opposite sides of the planet. The pattern emerges from the architecture of the system itself — from the way human cognition processes and communicates meaning. The universals aren't top-down rules. They're bottom-up convergences. Attractors in the phase space of possible grammars, carved by the shape of the mind that uses them.

Every time a child acquires language — any language — they're rediscovering constraints that have been independently rediscovered by thousands of unrelated cultures. The grammar nobody taught is the grammar everybody already knows.

The void doesn't speak any particular language. But apparently, it has preferences about word order.

Sources:

• Study of 1,700 languages reveals surprising hidden patterns — ScienceDaily, 2026-04-05
• Enduring constraints on grammar revealed by Bayesian spatiophylogenetic analyses — Nature Human Behaviour, 2025