Gary Stix's (2025) Scientific American Magazine article, “Thinking Without Words,” highlights the idea that language and thought are separable.
Research shows that many forms of thinking engage brain systems outside of language networks, and even individuals who lose language (aphasia) often retain high-level cognitive abilities (reasoning, problem-solving, decision-making, planning, abstract thinking, and social cognition).
Neuroimaging Evidence of Separate Networks
Logical Reasoning vs. Language
An fMRI study by Monti et al. (2009) compared brain activity during abstract logical inference tasks to activity during language processing tasks. The results showed non-overlapping brain regions: logical reasoning activated frontal areas (e.g., medial prefrontal cortex, BA 10p/8m), whereas language processing activated classic left-perisylvian language regions (e.g., Broca’s and Wernicke’s areas). Anterior and posterior language area graphic © Fascija/Shutterstock.com.
In other words, deductive thought did not recruit the language network, leading the authors to conclude that logical inference “is not embedded in natural language.” This supports the view that the brain’s reasoning circuits operate independently of linguistic grammar (language structural rules).
Non-Linguistic Tasks Activate Different Areas
Converging neuroimaging studies show that when people perform complex non-verbal cognitive tasks (like arithmetic calculations, working memory challenges, or music comprehension), the brain’s language regions remain largely inactive, instead engaging other neural circuits. For example, Fedorenko et al. (2011) identified each participant’s language-selective areas in an fMRI, then had them solve math problems. Language areas showed no significant activation during arithmetic – their activity dropped to baseline levels – even as fronto-parietal “multiple-demand” regions became active to support calculation.
Similarly, brain regions specialized for language respond strongly when we read or listen to sentences, but do not light up during tasks like math, memory encoding, inhibitory control, or music listening. Instead, those tasks recruit other networks (e.g. bilateral parietal cortex for math, or executive control regions in frontal cortex), indicating that high-level thought processes can proceed without involving language centers
Preserved Thought in Aphasia (Language-Independent Cognition)
Reasoning and Planning in Aphasics
Clinical studies of individuals with severe aphasia (loss of language) demonstrate that general intelligence and reasoning can be intact despite language impairment. Varley and Siegal (2000) reported a case of a man with profound agrammatic aphasia (lack of function words and impaired syntax) who nonetheless solved complex logical problems and even passed theory-of-mind tasks requiring him to infer others’ beliefs. Remarkably, some patients with global aphasia (almost no comprehensible language) can even play chess at a high level – a game demanding intense planning, working memory, and abstract reasoning – despite being unable to name the pieces or articulate the rules verbally. These cases show that domains like causal reasoning, strategy, and understanding others’ thoughts do not necessarily depend on language ability.
Mathematical Abilities without Language
Likewise, numerical cognition (the ability to understand and manipulate numbers) can remain fully functional in the absence of language. Varley et al. (2005) tested several individuals with extensive left-hemisphere damage and severe aphasia on math tasks. Despite their language loss, these patients could perform exact multi-step arithmetic calculations (including tasks like 12/(3–1) or adding large numbers) with normal accuracy. This finding provides “experimental evidence against the claim that language is necessary for mathematics.” In fact, the authors concluded that mature mathematical reasoning either taps into a non-linguistic cognitive system shared with language or is entirely independent of the language system. Consistently, other researchers have observed that even extensive damage to classical language areas leaves mathematical and logical problem-solving abilities intact, reinforcing that these cognitive functions are supported by brain regions outside of the language network.
Conclusion: Thinking Without Words
Overall, neuroscience research strongly supports the idea that thought and language are distinct components of cognition. Brain imaging shows that tasks like reasoning, calculation, music, or planning activate their own neural circuits, largely separate from language-processing areas. Meanwhile, neuropsychological evidence from aphasia cases confirms that a breakdown in language does not equate to a breakdown in thought – patients can continue to reason, solve problems, and understand the world even when they cannot put thoughts into words. These findings echo the message: we don’t need words to think. Our capacity for thought is rooted in distributed cognitive and neural systems, many of which operate independently of language.
Language primarily serves as a tool for communication, allowing us to share our thoughts and ideas with others, rather than being essential for the generation of thoughts themselves (Okrent, 2017).
Key Takeaways
1. Language and thought are processed in separate brain regions. 2. Non-linguistic cognitive tasks do not require language regions 3. Aphasia patients retain complex thought processes
4. Mathematical abilities are preserved without language
5. Human thought does not require words but communicates using them
Glossary
agrammatic aphasia: a form of aphasia characterized by difficulty in constructing grammatically correct sentences, often resulting in speech that lacks function words and proper syntax, while retaining some ability to produce content words.
aphasia: a language disorder caused by brain damage, typically affecting speech production, comprehension, reading, and writing. It results from damage to language-processing regions, usually in the left hemisphere.
Broca's area: a region in the left frontal lobe of the brain associated with speech production and syntactic processing. Damage to this area often leads to Broca's aphasia, characterized by difficulty in forming grammatically complex sentences.
fMRI (functional magnetic resonance imaging): a neuroimaging technique that measures brain activity by detecting changes in blood flow. It is widely used to identify which brain regions are involved in specific cognitive tasks, including language and thought.
global aphasia: a severe form of aphasia resulting from extensive damage to the left hemisphere, leading to significant impairments in both speech production and comprehension.
high-level cognitive abilities: advanced mental processes involved in reasoning, problem-solving, decision-making, planning, abstract thinking, and social cognition.
left-perisylvian language regions: a network of brain areas surrounding the Sylvian fissure in the left hemisphere, including Broca’s area, Wernicke’s area, and other cortical regions involved in language comprehension and production.
linguistic grammar: the structural rules governing the composition of words, phrases, and sentences in a language. It encompasses syntax, morphology, and phonology.
numerical cognition: the brain's ability to understand and manipulate numbers, including basic arithmetic, number recognition, and complex mathematical reasoning. This function is primarily associated with parietal lobe activity and does not rely solely on language.
Wernicke’s area: a region in the left temporal lobe crucial for language comprehension. Damage to this area results in Wernicke’s aphasia, characterized by fluent but nonsensical speech and difficulty understanding spoken language.
References
Fedorenko, E., Behr, M. K., & Kanwisher, N. (2011). Functional specificity for high-level linguistic processing in the human brain. Proceedings of the National Academy of Sciences, 108(39), 16428–16433. https://doi.org/10.1073/pnas.1112937108
Fedorenko, E., & Varley, R. (2016). Language and thought are not the same thing: evidence from neuroimaging and neurological patients. Annals of the New York Academy of Sciences, 1369(1), 132–153. https://doi.org/10.1111/nyas.13046
Monti, M. M., Parsons, L. M., & Osherson, D. N. (2009). The boundaries of language and thought in deductive inference. Proceedings of the National Academy of Sciences of the United States of America, 106(30), 12554–12559. https://doi.org/10.1073/pnas.0902422106
Okrent, M. (2017). Nature and normativity, 159-193. Routledge. https://doi.org/10.4324/9781315276700-6
Stix, G. (2025). Thinking without words. Scientific American Magazine, 32(3), 86. https://doi.org/10.1038/scientificamerican032025-5To12UBJzHHhjfzh8aMrHM
Varley, R. A., Klessinger, N. J. C., Romanowski, C. A. J., & Siegal, M. (2005). Agrammatic but numerate. Proceedings of the National Academy of Sciences, 102(9), 3519–3524. https://doi.org/10.1073/pnas.0407470102
Varley, R., & Siegal, M. (2000). Evidence for cognition without grammar: A case study of the language-thought interface. Current Biology, 10(12), 723–726. https://doi.org/10.1016/S0960-9822(00)00538-8
Varley, R. A., Klessinger, N. J. C., Romanowski, C. A. J., & Siegal, M. (2005). Agrammatic but numerate. Proceedings of the National Academy of Sciences, 102(9), 3519–3524. https://doi.org/10.1073/pnas.0407470102
Comments