Logic Does Not Ride on Words

The Invisible System

Language feels like thinking. You can hear yourself narrate a problem, talk through the steps, arrive at an answer in words. For thousands of years, philosophers have suspected the two are the same machine.

Cognitive neuroscientists at MIT's McGovern Institute put that suspicion to the test — and found a separation. Their work was published in the Proceedings of the National Academy of Sciences.

You can lose language and keep the logic underneath.

What They Measured

The researchers worked with two people who had strokes that damaged the language areas of their brains, leaving them with severe impairments in both understanding and producing speech. The team gave them language-free logic games: infer the hidden rule that turns one list of numbers into another, or pick the shape that completes a pattern.

The language-impaired participants solved the puzzles as well as a control group. They even communicated the rules they found using gestures and sketches.

• Inductive reasoning: finding a hidden rule.
• Deductive reasoning: judging whether a conclusion follows.
• Neither task, the scans showed, called on the brain's language network.

The Surprise

There was a second finding, and it was not what the team expected. The "multiple demand network" — a distributed brain system long suspected of powering hard thinking — did light up during inductive reasoning. But it stayed quiet during deductive reasoning.

That means "reasoning" is not one thing either. Finding a rule and checking a conclusion may run on different machinery, and neither is the machinery of words.

Why It Matters

The result is more than philosophy. People with aphasia keep their intelligence, yet others often confuse their difficulty speaking with difficulty thinking. The authors argue the distinction should reshape how severe language loss is understood — and how artificial minds are compared with human ones.

The voice narrating the problem is not the one solving it.