Advances in AI rooted in Chinese linguistic data suggest Chinese will play a significant role in this new phase of human development. Photo: TUCHONG

The 20th century witnessed an unprecedented intellectual effort to understand language—the most fundamental medium through which human beings exist, think, and develop. In philosophy, Ludwig Wittgenstein dismantled the long-standing subject–object dualism that had shaped Western thought since Plato and Aristotle. By placing philosophy onto the terrain of language, and by deploying rigorous semantic and pragmatic analysis, he sought to clear and reconstruct its foundations. Noam Chomsky’s linguistic revolution soon followed. Working within a rationalist framework and taking English as his primary object of inquiry, Chomsky developed a theory of syntactic structure, articulated the concept of universal grammar, and proposed the hypothesis of an innate language faculty.

In doing so, he deepened our understanding of the relationship between language and mind and became a pioneer in exploring the mysteries of the human mind. Chomsky’s work exerted profound influence on psychology, computer science, and artificial intelligence (AI), and his transformational-generative grammar continues to shape discussions in AI decades later. Seen from a historical perspective, the inquiries of Wittgenstein, Chomsky, John R. Searle, and Richard M. Montague—from the early 20th century through the mid-1970s—form a sustained continuum of reflection on language and mind. One of the most significant outcomes of this intellectual movement was the birth of cognitive science.

Emergence of cognitive science

Cognitive science emerged as an interdisciplinary enterprise integrating philosophy, linguistics, psychology, anthropology, computer science, and neuroscience. In the early 21st century, education was added to this constellation, producing what is often described as a “6+1” framework. This integrative momentum has only intensified. By situating language within the broader architecture of mind and cognition, cognitive science has enabled scholars to revisit fundamental questions from fresh vantage points—drawing on human evolutionary history, the history of human mental development, and the history of philosophical thought, so as to truly grasping essence of language.

The mind may be understood as the set of capacities humans have acquired over the course of evolution to comprehend both the external world and themselves. These capacities can be described, in ascending order, across five levels: brain and nervous system, psychology, language, thinking, and culture. Among these, linguistic cognition occupies a pivotal position. It distinguishes higher-order human cognition from the more limited cognitive capacities of non-human animals. More importantly, it underlies all subsequent levels: thinking and culture are built upon the foundation of language. In this sense, a people’s language determines their characteristic modes of thinking and cultural expression.

Language as the foundation

The 21st century is often described as an age of synthesis and innovation.

Synthesis occurs at two levels: thinking (logic) and culture (science, philosophy, and religion). The foundation of both is language. Without language, synthesis at the levels of thinking and culture would be impossible; only language provides the structural basis for integration. Chomsky’s generative grammar illustrates this point: the reason sentences in different languages can be translated and mutually understood lies in their shared deep structure. These structural features of language enable communication, the exchange of ideas, and cultural interaction. On this foundation, comprehensive development becomes possible.

Innovation is likewise inseparable from synthesis. All genuine innovation is, in essence, synthetic innovation, and synthetic innovation ultimately manifests as cultural innovation, since culture contains the richest connotations of human life. From the perspective of the five levels of cognition, cultural cognition encompasses and integrates the levels of thinking and language. For humanity, culture rests upon language and thought; accordingly, cultural innovation depends upon innovation in language and thinking. Language alone can account for the development and progress of humanity in the 20th century, and it alone can provide the foundation for thought and cultural innovation in the 21st century.

The history of cognitive science bears witness to a transition from thinking-centered cognition to language-centered cognition. From its advent through to the emergence of ChatGPT, AI largely sought to simulate human reasoning—thinking-type intelligence. ChatGPT, however, is a language-conversation form of AI and thus stands closer to the structure of human cognition. Since the Dartmouth Conference in 1956, AI has developed through more than seven decades of theoretical evolution. Yet in essence, ChatGPT can be understood as an extension of Chomsky’s transformational-generative grammar. Notably, Chomsky himself has sharply criticized such systems, arguing that they damage language, diminish human intelligence, and challenge human morality. The root of this dispute lies, once again, in language.

Language is the foundation of human cognitive ability and of the capacity for innovation. Human linguistic ability has two dimensions: the use of language and the creation of language. Chomsky’s notion of an innate language faculty emphasizes that language is the most important cognitive capacity acquired in the course of evolution, and that the first language—one’s mother tongue—is innately inherited as a cognitive structure. Contemporary AI, by contrast, obviously does not possess such a faculty. If AI appears to have language ability, that ability is conferred upon it by human beings.

Chinese characters as a living pictographic script

Chinese characters are the only living pictographic script still in continuous use. Each character integrates sound, form, and meaning into a single visual unit, giving rise to distinctive structural features and cognitive advantages. In written Chinese, meaning can often be accessed directly through the character’s form, without mandatory phonetic conversion. For this reason, the processing of Chinese can be exceptionally efficient compared to alphabetic scripts.

The approximately 6,000 commonly used basic characters function as independent lexical units. Through arrangement and combination, they generate an immense range of compound forms. From a combinatorial standpoint, the potential number of two-, three-, and four-character constructions expands exponentially. At the morphological level, Chinese therefore possesses extraordinary generative capacity—an attribute that distinguishes it from alphabetic systems.

These characteristics of Chinese language and script have profoundly shaped Chinese modes of thought and Chinese culture. In the era of digitization, informatization, and AI, it was once predicted that the logographic system might impede technological development and digitization. The opposite has proven true. Chinese increasingly demonstrates itself to be particularly well suited to digital and AI environments, and to serve as a powerful cognitive instrument for comprehensive innovation in the 21st century.

DeepSeek offers a concrete illustration of this claim. In terms of language processing, its system can efficiently recognize and model complex patterns in Chinese syntax and morphology, responding flexibly to the distinctive features of the script. At the level of contextual understanding, DeepSeek continually incorporates new textual data, enabling it to remain sensitive to ongoing shifts in usage and to refine its grasp of contemporary linguistic environments. It is also capable of engaging both classical and modern Chinese, exploring layered meanings and cultural references embedded in different historical registers. Beyond processing and interpretation, such technologies contribute to the broader circulation of Chinese: as the language assumes increasing importance in international affairs, platforms like DeepSeek make access to Chinese learning and communication more widely available. In this reciprocal relationship, the structural richness of Chinese supports the development of systems like DeepSeek, while those systems, in turn, extend the language’s global reach.

From the perspective of linguistic cognition, Chinese may be characterized as experiential in orientation, and its script as embodying experiential features. This characterization aligns with what has been described as the experiential turn in cognitive science. In terms of reasoning style, Chinese intellectual traditions are marked by holistic rather than fragmentary thinking, and by synthesis rather than narrow analysis. Among four commonly distinguished forms of inference—induction, analogy, abduction, and deduction—greater emphasis in this intellectual tradition has traditionally been placed on the first three. As a consequence cultural cognitive patterns shaped—indeed, in this account, determined—by linguistic and thinking structures are said to differ markedly from those characteristic of Western traditions.

In this account, deductive reasoning cannot generate new knowledge and therefore makes no contribution to scientific discovery. Rather, scientific innovation emerges from the dynamic interplay of induction, analogy, and abduction. In this sense, the experiential turn in cognitive science is not merely a methodological adjustment, but a reorientation of emphasis: it calls for modes of thought that are synthetic, generative, and oriented toward innovation.

For most of humanity’s evolutionary history—indeed, until the advent of the modern industrialization movement—conditions of food scarcity demanded analytical precision and problem-solving efficiency. Rational and sequential forms of reasoning—often associated with left-hemispheric functions and with the cognitive demands of hunting and tracking—were therefore cultivated and privileged. By contrast, holistic, intuitive, and creative capacities—frequently associated with right-hemispheric activity—remained comparatively restrained in their expression. As material conditions improve and basic survival needs are increasingly secured, these capacities are said to find renewed space for development, signaling a new mode of human existence and forming a defining characteristic of the cognitive science era.

Since the emergence of human language in deep prehistory, humanity has undergone successive revolutions grounded in linguistic development. Another such transformation is now underway. Advances in AI technologies rooted in Chinese linguistic data suggest that Chinese will play a significant role in this new phase of human development. In the age of AI, Chinese possesses distinctive structural and cognitive advantages for innovation.

Cai Shushan is a professor from the Department of Psychological and Cognitive Sciences at Tsinghua University. This article has been edited and excerpted from Philosophical Analysis, Issue 5, 2025.