Do we think because we speak? The complex relationship between language and thought

MRI brain scan

Scholars across different disciplines – linguistics, psychology, philosophy, and anthropology – have long researched the connection between language and the way we think. The debate about whether language completely shapes thought or merely reflects it has sparked various theories and extensive research based on empirical evidence. Classic linguistic theories, such as the Sapir–Whorf hypothesis and Wittgenstein’s philosophical insights, have laid the foundation for understanding this relationship.

However, a comprehensive review of research spanning many years, recently published in Nature, challenges these traditional views. It provides neuroimaging evidence suggesting that language is primarily a tool for communication rather than thought.

My intention is to take a look at these classic theories and contrast them with the latest findings, providing an overview of the evolving understanding of the relationship between language and thought.

The Sapir–Whorf hypothesis

Linguistic relativity

Probably the most famous theory about language influencing our thought is named after the linguists Edward Sapir and Benjamin Lee Whorf – the Sapir–Whorf hypothesis. This theory is also known as linguistic relativity, and it argues that the structure of a language affects its speakers’ cognition and worldview. This means that the grammatical and lexical categories of a language influence how its speakers perceive and think about the world.

Here are some key examples illustrating this concept.

Example 1: Hopi language and time perception

Whorf’s studies on the Hopi language, a Native American language spoken by 5,000 Hopi people in Arizona, led him to argue that the Hopi’s perception of time fundamentally differs from that of English speakers because of linguistic differences. While English uses tenses to denote time (past, present, and future), the Hopi language emphasises the validity of an event (manifested or unmanifested). This linguistic distinction, according to Whorf, shapes how Hopi speakers conceptualise time, influencing but not completely dictating their perception of temporal events.

Example 2: Colour perception

Research supporting linguistic relativity has found that language can influence various cognitive domains, such as colour perception. A notable study by Berlin and Kay (1969) on colour terminology demonstrated that the specific colour terms available in a language can affect how speakers categorise and remember colours. For instance, speakers of languages that differentiate between light and dark blue as separate colours (e.g., goluboy and siniy in Russian) are better at distinguishing between shades of blue than speakers of languages that do not make this distinction.

Example 3: Spatial orientation

Similarly, studies on spatial orientation have shown that languages that use absolute directions (e.g., north, south) instead of relative terms (e.g., left, right) influence how speakers navigate their environment. For example, speakers of Guugu Yimithirr, an Australian Aboriginal language, use cardinal directions for all spatial relations (“north of the house”), which influences their navigational skills and spatial memory, making them more adept at orienting themselves in their environment.

Linguistic determinism

Linguistic determinism is the stronger version of the Sapir–Whorf hypothesis, asserting that language determines thought and that linguistic categories limit and shape cognitive categories. According to this view, the language one speaks constrains the kinds of thoughts one can have. Put simply, determinism implies that when a language lacks words for a thing, its speakers simply cannot think about that thing. Let’s look at some examples that illustrate the concept of linguistic determinism.

Example 1: Eskimo words for snow

Whorf famously illustrated this idea with the example of Eskimo languages, which he claimed have unusually numerous words for snow. According to Whorf, this linguistic richness means that speakers of these languages perceive and conceptualise snow in ways that are inaccessible to speakers of languages with fewer snow-related terms. In 1940, he wrote: “We have the same word for falling snow, snow on the ground, snow packed hard like ice, slushy snow, wind-driven flying snow…. To an [Inuit person], this all-inclusive word would be almost unthinkable; he would say that falling snow, slushy snow, and so on, are sensuously and operationally different.” Although this clichéd example has been criticised and debunked in later studies, it underscores the determinist perspective that language defines the boundaries of cognition.

Example 2: Pirahã language and numbers

A more contemporary example comes from the Pirahã language, spoken by an Indigenous group in the Brazilian Amazon. This language lacks words for specific numbers beyond “one,” “two,” and “many.” Researchers like Daniel Everett argue that this linguistic limitation impacts the Pirahã’s ability to perform tasks involving precise numerical concepts, suggesting that the lack of numerical vocabulary constrains their numerical cognition, aligning with the determinist view. In very simple terms, we could say that those Indigenous people do not count because they are lost for words. Everett’s conclusions were met with scepticism from universalists who claimed that the linguistic deficit is explained by the lack of need for such concepts.

Are these theories accurate?

Linguistic relativity and linguistic determinism are sometimes conflated, but they are distinct concepts. Linguistic determinism, the strong version that suggests that language strictly dictates and limits thought, has been rejected by scholars. Instead, linguistic relativity, the weaker version, postulates that language influences patterns of thought, especially at the conceptual level. Many scholars support this view, indicating that while language can shape thought, it does not rigidly determine it. Thus, linguistic relativity remains a significant area of study, acknowledging the nuanced ways in which language impacts cognition.

Wittgenstein’s philosophical insights

Ludwig Wittgenstein, a prominent 20th-century philosopher, contributed significantly to the philosophy of language and thought. In his early work, Tractatus logico-philosophicus, Wittgenstein argued that the structure of language mirrors the structure of reality. He famously stated, “The limits of my language mean the limits of my world,” suggesting that language constrains what can be thought and expressed.

In his later work, Philosophical Investigations, Wittgenstein shifted his perspective, focusing on the practical use of language in everyday contexts. He introduced the concept of “language games” (German: Sprachspiele), emphasising that meaning arises from the use of words in specific social activities. Wittgenstein argued that the diversity of language games reflects the multiplicity of ways in which language is woven into the fabric of life. For example, he provides the example of “Water!” which can be used as an exclamation (expressing relief or urgency in a desert), an order (demanding water from someone), a request (politely asking for water), or an answer to the question “What would you like to drink?” (water, not juice). Words and sentences only acquire significance if we fix them within some context of use. This later view acknowledges the flexibility and variability of language, suggesting that language is a tool for various forms of social interaction rather than a rigid framework for thought.

Language is not for thinking – new research from the Nature article

The new perspective published in Nature examines the interplay between language and thought, questioning if language is truly a prerequisite for thinking or if it is a separate cognitive process. According to the researchers, language primarily serves as a tool for communication and is a sophisticated way in which humans differ from other primates. However, this doesn’t mean that language is the gateway to cognition.

This presents a paradigm shift in understanding the relationship between language and thought. The study, titled “Language is primarily a tool for communication rather than thought,” argues that language’s primary function is communication, not the mediation of thought (the language-for-thought hypothesis).

The researchers reviewed evidence from neuroscience and cognitive science, demonstrating that complex thought processes can occur independently of language, as seen in individuals with severe linguistic impairments. Functional MRI (fMRI) studies further revealed that non-linguistic cognitive tasks do not activate the brain regions associated with language processing.

The human brain supports various cognitive functions through different networks. The language network, encompassing Broca’s and Wernicke’s areas, supports language comprehension and production. Other networks, such as the multiple demand network and the theory of mind network, support executive functions, novel problem-solving, mathematics, social reasoning, and mentalising. This diversity in neural networks indicates that language is neither necessary nor sufficient for thought.

Individuals with aphasia, despite significant language deficits, can still solve mathematical problems, perform executive planning, follow non-verbal instructions, and engage in formal logical reasoning.

This challenges the language-for-thought hypothesis and suggests that language is not essential for these cognitive processes.

Double dissociation between language and thought

One of the key findings of the new research is the concept of double dissociation between language and thought. This refers to the observation that certain cognitive abilities can remain intact even when linguistic abilities are severely impaired, and vice versa.

As previously noted, individuals with aphasia can still engage in diverse forms of reasoning, including formal logical reasoning and causal reasoning about the world.

Conversely, individuals with schizophrenia often have impaired thought processes while their linguistic abilities remain relatively intact.

This provides a compelling counterpoint to the idea that language is necessary for thought.

Comparisons with animal cognition provide additional insights. Non-human animals, such as primates, exhibit sophisticated problem-solving and social reasoning abilities without possessing human-like language. For example, studies on chimpanzees have shown that they can use tools, engage in complex social interactions, and solve problems that require a deep understanding of their environment. Similarly, dolphins have demonstrated the ability to understand symbolic representations and perform tasks that require advanced cognitive skills. This evidence suggests that complex cognition can exist independently of language.

Evidence from fMRI scans supports this concept by showing that non-linguistic cognitive tasks do not engage the language network, a set of brain areas responsible for processing word meanings and syntactic structures. Instead, these tasks activate different neural networks, indicating that the brain has distinct systems for handling language and non-linguistic thought processes.

Language as a tool for communication

The new research emphasises that language has evolved primarily as a tool for communication. This has facilitated the transmission of cultural knowledge across generations. From the evolutionary fitness standpoint, this could provide an adaptive advantage to humans.

The study highlights several properties of human languages that make them efficient for communication: ease of production, ease of learning and understanding, conciseness, and robustness to noise.

For example, languages tend to minimise dependency lengths, preferring word orders that reduce the cognitive load on speakers and listeners. This feature makes sentences easier to produce and understand. Additionally, languages often exhibit preferences for particular word orders that align with cognitive processing patterns, further enhancing communication efficiency.

One example of this is the subject–object–verb (SOV) order, which is the most common across the world’s languages, including Japanese, Persian, and Hindi (approximately 47% of languages). Another common word order is subject–verb–object (SVO), used by approximately 41% of languages, such as English, Ukrainian, and Mandarin. The SVO order, in particular, allows listeners to use positional cues – whether a given noun appears before or after the verb – to reconstruct who is doing what to whom, even when some information is lost during communication.

Cross-linguistic tendencies, such as the prevalence of ambiguity and the use of context to resolve it, also support the idea that language is optimised for communication. Ambiguity allows for flexibility and economy in language use, while context helps disambiguate meaning, making communication more effective.

These features enhance the efficiency and effectiveness of language in conveying information and facilitating social interaction. By optimising for ease of production, learning, and use, languages are better suited to their primary function of communication rather than serving as a framework for thought.

Additional insights from the research

The impact of language deprivation on cognition provides further evidence for the dissociation between language and thought. Individuals who experience language deprivation, such as those who are born deaf and do not learn sign language until later in life, demonstrate advanced cognitive skills.

Comparisons between cases of schizophrenia and aphasia also illustrate this dissociation. In schizophrenia, thought processes are impaired while language abilities may remain intact. This condition often involves disturbances in thinking patterns, such as disorganised thoughts and delusions, which occur independently of linguistic abilities.

In contrast, individuals with aphasia have impaired language abilities but can still engage in complex thought processes.

These cases emphasise that language and thought operate through distinct neural mechanisms.

The cognitive abilities of pre-linguistic infants further support this view. Infants exhibit sophisticated problem-solving and reasoning abilities before they acquire language. For example, studies have shown that infants can understand basic principles of physics, such as object permanence and causality, and can perform tasks that require logical reasoning. These findings suggest that the foundations of thought are present even before language development begins.

Contrasting classic theories and new research

Comparing perspectives

Classic theories of linguistic relativity, linguistic determinism, and Wittgenstein’s early views emphasise a deep interconnection between language and thought, suggesting that language shapes, constrains, or reflects thought.

In contrast, the new research from Nature argues that language primarily serves a communicative function and that thought processes can occur independently of linguistic representations.

The concept of double dissociation challenges the classic view that language is necessary or sufficient for thought.

While the Sapir–Whorf Hypothesis and Wittgenstein’s early work highlight the role of language in shaping cognition, the new findings suggest a more nuanced relationship where language facilitates communication and cultural transmission without being essential for individual cognitive processes.

Broader implications

The new research has significant implications for linguistics, cognitive science, and our broader understanding of human cognition and culture. It suggests a need to reconsider the extent to which language influences thought and to explore the distinct neural and cognitive mechanisms underlying language and thought.

This paradigm shift emphasises the role of language in communication, social interaction, and cultural transmission rather than its function as a mediator of thought. Language, as a tool for communication, has been crucial in the development of human societies by enabling the transmission of accumulated knowledge across generations.

While language is not necessary for individual cognitive processes, it supports and enhances the unique human ability to express ideas and has been instrumental in shaping human civilisation.

Future research

Future research could investigate how different languages optimise for communicative efficiency and how these optimisations impact social and cultural practices.

Exploring the neural and cognitive mechanisms underlying language use in social contexts can provide deeper insights into the interplay between language, thought, and culture. For example, examining how language influences social behaviours, such as cooperation, conflict resolution, and group decision-making, can highlight its contribution to social cohesion and collective problem-solving.

Moreover, studying the cognitive abilities of individuals with linguistic impairments and those deprived of language can further explain the boundaries and capacities of human cognition independent of language.

Longitudinal studies tracking cognitive development in these populations can offer valuable insights into the interplay between language and thought.

This requires a multidisciplinary approach that integrates insights from neuroscience, psychology, anthropology, and linguistics.


The relationship between language and thought is a complex and evolving topic. Classic theories like the Sapir–Whorf Hypothesis and Wittgenstein’s philosophical insights have long suggested a deep interconnection between the two, focusing on how language shapes our thinking and perceptions of the world. However, these theories do not necessarily claim that language is the only way we can process cognition.

New research challenges the weight given to these classic views by supporting the language-for-communication hypothesis. It presents compelling evidence that language primarily serves as a tool for communication, thereby refuting the idea that language is essential for thought.

The concept of double dissociation between language and thought strongly demonstrates the independence of cognitive processes from linguistic representations. This has profound implications for linguistics, cognitive science, and our understanding of human cognition and culture.

Leave a comment

Your email address will not be published. Required fields are marked *