Researchers at MIT’s McGovern Institute for Brain Research, in collaboration with University College London, have published findings in the Proceedings of the National Academy of Sciences demonstrating that the human brain processes logical reasoning through a system distinct from language. The study, led by cognitive neuroscientist Evelina Fedorenko and postdoctoral researcher Hope Kean, challenges the long-standing assumption that language is a prerequisite for abstract thought and complex problem-solving. To isolate reasoning from linguistic capacity, the team administered language-free logic puzzles to two patients with severe aphasia resulting from strokes that damaged language-processing regions. Despite profound difficulties in both understanding and producing speech, participants successfully identified hidden mathematical and geometric rules, matching the performance of neurotypical controls. They also demonstrated the ability to communicate inferred patterns through gestures and sketches. Concurrent functional MRI scans of healthy participants revealed a clear neural dissociation. Brain activity associated with the language network remained inactive during both inductive reasoning, where subjects deduced unseen rules, and deductive reasoning, which involved evaluating syllogistic statements. Instead, inductive tasks activated the multiple demand network, a distributed system previously linked to complex cognition, while deductive tasks showed minimal engagement of this region. These results indicate that symbolic rule induction and logical evaluation operate through specialized circuits independent of linguistic processing. The findings carry significant implications for clinical practice and public perception. Specialists have long noted that individuals with acquired language impairments retain intact cognitive abilities, yet societal biases frequently conflate communication deficits with reduced intelligence. This research provides empirical validation that abstract logical capacity remains largely preserved in aphasia, reinforcing the need for healthcare protocols that separate language therapy from cognitive assessment. From a technology standpoint, the results offer a critical benchmark for artificial intelligence. Current large language models achieve conversational fluency and apparent reasoning through text-based training and output, lacking the modular architecture observed in human cognition. By highlighting the independence of linguistic and logical systems, the study suggests that next-generation AI architectures may benefit from decoupling language processing from symbolic reasoning modules. Researchers indicate that bridging this gap could improve model efficiency, robustness, and alignment with human-like cognitive frameworks. The work establishes a new framework for mapping cognitive functions and opens avenues for investigating how other high-level processes, such as object categorization and social reasoning, interface with linguistic networks. As the research team continues to explore the neural geography of thought, the distinction between language and logic is increasingly recognized as a foundational principle of human cognition, with broad applications ranging from neurorehabilitation to advanced machine intelligence.