A recent study published in npj Science of Learning has shed light on how humans comprehend language, revealing that our ability to understand what is said to us is not a single skill, but rather relies on three separate mental mechanisms. Researchers analyzed language abilities in a large group of individuals with language difficulties and found that these abilities naturally group into three distinct categories. This suggests that our brains use at least three different approaches to process and make sense of language, ranging from basic commands to complex sentences.
While we know language is central to our intelligence, the specific brain processes that enable us to understand language have remained a puzzle. Previous research hinted at the existence of different language comprehension mechanisms, and a study focusing on autistic individuals identified three potential mechanisms.
The current study aimed to confirm and expand upon these earlier findings by examining a more diverse group of people struggling with language. The researchers wanted to see if the same three mechanisms would emerge in a broader population and to explore how these mechanisms might relate to known brain functions. They were particularly interested in understanding if these language mechanisms could be linked to different types of mental processes, such as imagining and combining ideas in our minds.
“Around six million years ago, humans diverged from chimpanzees and developed a communication system unlike any other in the animal kingdom. For over 50 years, linguists such as Noam Chomsky and Steven Pinker have proposed the existence of a uniquely human language comprehension mechanism, yet its neurological basis remains largely unknown,” said study author Andrey Vyshedskiy, a neuroscientist affiliated with Boston University, co-founder of ImagiRation, and author of On The Origin of the Human Mind.
“From the age of nine, I was fascinated by how the brain generates language and imagination. At 24, I set out to explore the evolution of these abilities from a neurological perspective. A decade later, I published a cognitive neuroscience model predicting three distinct language comprehension mechanisms: two uniquely human and one shared with other primates. This model diverged from those proposed by Chomsky and other linguists, as it identified two separate, uniquely human language mechanisms.”
“The model predicts that language acquisition occurs in distinct stages: while most individuals develop all three mechanisms, some acquire only the first, and others attain only two,” Vyshedskiy continued. “Those lacking the most advanced mechanism struggle to understand complex explanations, narratives, and fairy tales, while individuals who develop only the first mechanism cannot combine nouns with adjectives. This prediction could be tested by studying language abilities in individuals with language deficits, though such research would require a large sample size of thousands of participants.
“This time, I was fortunate. The cognitive neuroscience model of language I developed two decades ago predicted a novel approach to language therapy for children with language deficits. We implemented this approach in a gamified language therapy app, which quickly gained popularity. In exchange for access to the app, parents provided detailed assessments about their children’s health and language abilities. Over ten years since its launch, parents submitted more than 300,000 assessments.”
Parents or caregivers registered on the app and provided detailed information about their child’s diagnosis, age, and language abilities. Every three months, caregivers filled out a comprehensive questionnaire that included 133 items covering various aspects of the child’s language skills, as well as information about other factors such as screen time, diet, and the caregiver’s education level.
From the full set of responses, the study focused on 15 language comprehension items. These items asked caregivers to rate how well their child understood different types of language, such as following simple commands, recognizing descriptive words that indicate color or size, and understanding more elaborate instructions found in stories or explanations. For each language item, parents could choose from answers indicating that the ability was very well developed, somewhat developed, or not present at all.
In order to ensure that the analysis focused on children with the ability to speak, the study only included participants whose caregivers reported that their child was capable of using sentences with four or more words. This selection process yielded a group of nearly 18,000 children and adolescents with a wide range of language impairments but with the ability to speak normally. This group included children and adolescents diagnosed with autism, mild language delay, apraxia (a speech disorder), Specific Language Impairment, Sensory Processing Disorder, Social Communication Disorder, Down Syndrome, and Attention-Deficit/Hyperactivity Disorder. A small group of neurotypical children was also included for comparison.
Vyshedskiy and his colleagues then applied two statistical techniques that are designed to identify patterns in data without any predetermined expectations. These methods automatically organized the 15 language comprehension items into groups based on how often certain skills were found together.
The analysis revealed three distinct groups. The first group included skills related to responding to one’s name, following simple commands, and reacting to praise. The second group consisted of skills needed to understand descriptive words related to color, size, and numbers. The third group involved the more advanced skill of following instructions that require understanding spatial relations, verb tenses, and complex story elements. In other words, if a child had difficulty with one ability in a group, they tended to have trouble with the others in that group as well.
In addition to grouping the language items, the researchers also sorted the nearly 18,000 participants into three clusters that reflected their overall language comprehension profile. One group of children showed strong abilities in all three areas – the basic commands, the descriptive language, and the complex story-like instructions. These children were similar to those typically developing children, and they made up almost half of the study sample. A second group, making up about one-quarter of the participants, showed good understanding of commands and descriptive language but had trouble with more complex, syntactic instructions. The final group, also about one-quarter of the participants, demonstrated the ability to follow simple commands but struggled with both descriptive and syntactic language.
“This study has important clinical, linguistic, and philosophical implications,” Vyshedskiy told PsyPost. “Communication abilities are conventionally categorized by clinicians as nonverbal, minimally verbal, or verbal—a one-dimensional classification that fails to capture the complexity of individual language skills. The identification of three distinct language comprehension mechanisms offers a more nuanced framework for assessing communication levels. A composite approach that considers both verbal abilities and language comprehension level would provide a more precise characterization, ultimately improving language therapy by addressing both aspects of language development.”
“Furthermore, the neurological model of language comprehension could pave the way for pharmacological interventions designed to extend the critical period for language learning beyond early childhood. These advancements could help maximize language acquisition potential, ensuring more individuals attain full language comprehension.”
“Linguists will need to rethink certain aspects of terminology,” Vyshedskiy said. “Since the existence of three distinct language mechanisms was not previously anticipated, current linguistic terminology does not yet accommodate these findings. As Benjamin Lee Whorf observed, language shapes cognition—without precise terms to describe these mechanisms, fully conceptualizing and discussing them remains a challenge.”
“Philosophy faces a similar challenge, as the term imagination is used to describe both involuntary experiences, such as nightmares, and voluntary processes, such as imagining a fairytale. This study provides neurological evidence distinguishing these two forms of imagination. Additionally, the neurological model offers new insights that may reshape philosophical perspectives on human uniqueness.”
“Finally, the discovery of the three language mechanisms offers a framework for understanding the evolution of language over the past six million years,” Vyshedskiy continued. “The first mechanism is largely shared with chimpanzees, the second—uniquely human—likely emerged around two million years ago, and the third, which enables full language comprehension, likely developed just 70,000 years ago.”
Even though the study gathered a large amount of data over many years, the researchers recognize some limitations in their approach. One limitation is that the data relied on reports from parents, which can sometimes be influenced by personal perceptions or hopes about their child’s progress. The study also focused primarily on younger children, as the therapy application is most often used by families with children between the ages of four and eight. As a result, the findings may not fully capture language abilities in older children or adults.
Future research could explore these questions further. One area of interest is the genetic factors that may influence whether a child develops the more advanced language comprehension abilities. By understanding which genes affect the brain’s capacity to process complex sentences and descriptions, researchers may be able to develop better-targeted interventions for language impairments.
The study, “Three mechanisms of language comprehension are revealed through cluster analysis of individuals with language deficits,” was authored by Andrey Vyshedskiy, Rohan Venkatesh, Edward Khokhlovich, and Deniz Satik.
