Language may work more like LEGO than we thought

Every time you speak, you're improvising. You string together words in combinations that have never existed before — and your brain pulls this off without consciously running through a rulebook. For decades, linguists assumed this flexibility came from deep, tree-like grammatical structures buried in your mind, hierarchies of rules stacked on top of each other. A new study suggests the reality might be simpler: your brain might rely more on familiar word patterns, the linguistic equivalent of pre-assembled LEGO pieces.

Morten Christiansen, a psychologist at Cornell University, and Yngwie Nielsen from Aarhus University analyzed how people actually process language — not just how grammar textbooks say it should work. They ran eye-tracking experiments and studied phone conversations, looking for patterns in how we recognize and repeat sequences of words. What they found challenges 70 years of linguistic theory.

The Shift From Trees to Building Blocks

The traditional view treats language like a tree with branches. Each sentence has a trunk, and words split into increasingly specific categories — nouns branch into subjects and objects, verbs into actions and states. This hierarchical structure was supposed to explain how humans generate infinite sentences from finite rules. It's elegant. It's also, Christiansen suggests, more complicated than necessary.

Instead, speakers may rely heavily on linear sequences — short chains of familiar word patterns. Think of phrases like "in the middle of the" or "wondered if you." These aren't complete grammatical units. Traditional grammar can't fully explain them. But they're everywhere in actual speech, and they matter.

When Christiansen's team showed people these sequences once, they processed them faster the next time they encountered them. The brain was priming itself on patterns that don't fit neatly into classical grammar rules. This suggests those patterns are genuinely part of how we store and use language — not just exceptions to the real system.

The implications ripple outward. If human language doesn't require the complex hierarchical machinery linguists have assumed for generations, then the gap between human communication and animal communication systems becomes smaller. A whale's song, a bird's call, a primate's gesture sequence — maybe these aren't as fundamentally different from human language as we thought. The complexity might be a choice, not a necessity.

Nielsen put it plainly: "Traditional rules of grammar cannot capture all of the mental representations of language structure." Christiansen added that if you don't need hierarchical syntax to explain how we actually speak, "it might even be possible to account for how we use language in general with flatter structure."

This doesn't mean grammar rules are wrong or useless. It means they're not the whole story. Your brain is doing something more flexible and pattern-based than the textbooks suggested — and it's doing it with less overhead. The next phase of research will test whether this simpler model can actually explain how we learn language, how we stumble over difficult sentences, and how we understand speech we've never heard before.