Play, Curiosity, and the Social Brain: Why Language Classrooms still matter in the Age of AI
By: Lynsey Mori
In my university classes I sometimes watch two students learn how to care for each other without meaning to. It begins with a mispronounced word, a shared laugh, an awkward repair. Nothing dramatic, just the slow choreography of attention as one person tries again and another waits long enough to understand. Language classrooms are unusual spaces in modern education because they require this kind of visible risk. Students must speak before they are ready, misunderstand one another, and repair those misunderstandings in front of others.
Of course, life is full of such moments. What makes the classroom distinctive is that these encounters happen repeatedly and deliberately. Mistakes play an important role in the process where playful repairs become part of the rhythm of learning and students learn that communication is about responding to one another.
This is not only linguistic practice. It is rehearsal for intimacy. In moments like these, students learn how to adjust to one another through laughter, hesitation, and repair. Such playful interaction makes learning possible.
Play as structured freedom
Play is often imagined as rulelessness, yet every game depends on constraint. A ball must stay inside the lines; a joke must read the room; and a conversation must wait its turn. Neuroscience has long shown that playful interaction builds executive function and emotional regulation. Jaak Panksepp’s work on affective systems describes how play activates curiosity without triggering the chemistry of threat. Jean Decety’s research on empathy demonstrates that perspective taking grows through cycles of misunderstanding and repair as people adjust to one another in real time. These moments often appear in playful interaction when a joke lands slightly wrong, a correction is offered, or a speaker tries again. Such moments are not obstacles to social development. They are a part of play, and one of the ways the social brain learns how to live with others.
Romance as adult play
Alfred North Whitehead called the first phase of learning Romance, meaning the stage when curiosity and attraction invite a learner towards the unknown. The word may sound sentimental today, but Whitehead used it to describe the excitement of discovery. At this stage learners approach the unfamiliar with curiosity, testing possibilities and responding to what emerges. In social settings, this early phase of learning often resembles a kind of intellectual flirtation. People approach cautiously, test responses, misread signals, laugh, and try again. These playful exchanges allow participants to explore together. In this sense, romance works the same way: reading subtle cues, tolerating ambiguity, risking embarrassment, trying again. Romance can therefore be understood as a form of adult play.
Language classrooms often reproduce this pattern. Students move quickly from partner to partner, searching for connection and discovering that communication grows through playful interaction. Activities such as rapid partner exchanges or “speed conversation” are used precisely because they encourage experimentation. The goal is not perfect performance but discovery: who responds, who listens, and how meaning begins to form between people.
Why relationships take work
To care for someone is to accept inconvenience. The ethics of care reminds us that relationships involve effort: listening when tired, adjusting plans, tolerating misunderstanding, repairing harm. Yet many students tell me they would rather avoid the risk of relationships altogether. They speak of the emotional labour as exhausting and the possibility of hurt as too visible. Digital life offers an appealing alternative in which connections can be paused, filtered, or left unanswered.
Education has a quiet responsibility here. Classrooms are rehearsal spaces for the kind of social adulthood a society hopes to grow. If students never practise the small work of caring, of waiting, negotiating, misreading, repairing, then where will they learn it?
In childhood, play provides these experiences naturally. Children move between roles, voices, and identities, testing possibilities without much concern for how they appear. As people grow older these boundaries tend to harden. Personal identities stabilise, social expectations become clearer, and vulnerability carries greater risk. Adults often protect the voices they have already established rather than experimenting with new ones. Opportunities for playful exploration become rarer with age.
The language classroom briefly reopens that space. Speaking another language places even confident adults back into uncertainty where mistakes are unavoidable and identity feels less fixed. Students must listen more carefully, adjust their timing, and tolerate small misunderstandings. What appears to be linguistic practice is also a temporary return to the exploratory conditions of play.
Learning a new language gently destabilises the self. Students must rely on one another as they search for words, interpret tone, and repair misunderstandings. Now that AI has entered the classroom, there is a promise of quicker, cleaner, and more comfortable learning. Some of the tools can be genuinely helpful, especially for language learners seeking practice and confidence. I use them in my own language learning, too. Yet most educational AI systems are designed to remove uncertainty. The student interacts privately with a system that predicts needs and corrects instantly. What is optimized cognitively may be thinned relationally. When answers arrive before questions fully form, there is less need to turn toward a peer, less reason to sit with confusion, and fewer occasions to discover that another person can help. The risk is subtle: a generation fluent in prompts yet less practised in presence. Many forms of learning depend on small negotiations that occur when people work things out together.
A classroom experiment
Last year I removed grades from several classes and replaced them with short recordings of feedback in my own voice. The response was initially chaotic. Students searched for orientation: How much is enough? What counts? They compared tone and length, unsure whether a longer message meant success, concern, or simply more to work on. In the absence of numbers, they began reading these signals as clues about what mattered and where they stood. Their questions were not only about performance but about interpretation. Matthew Lieberman describes this as the social brain seeking regulation; students were asking not only “Am I good?” but “Where am I, and who is with me?”
Gradually something shifted and the conversations grew longer. They began explaining to each other what they had heard in the recordings and how they understood the suggestions. Without numerical grades to settle the question quickly, they had to talk together about meaning and progress.
Designing conditions for learning
In an AI age, teachers become designers of conditions rather than distributors of content. In language classrooms this means beginning with people before tools, letting stories and dilemmas open the room before solutions arrive. It means creating routines where disagreement can be voiced without humiliation, and where apology and working things through are ordinary parts of learning. Feedback becomes a form of relationship rather than a score, and AI, when it is used, serves as rehearsal space while interpretation remains social. These choices keep the conditions in which the social brain can develop.
Reframing the question
The question is not whether AI belongs in schools, the question is what kinds of learning we want to preserve. Smooth and individualised systems may increase efficiency, they also reduce the shared uncertainty through which people learn to coordinate with one another. If we want students to fall in love with learning, we must let them experience what Whitehead called Romance: the uncertainty of approach, the embarrassment of misunderstanding, and the relief of being understood. Classrooms can feel a little like courtship; they’re structured, and respectful, yet alive with possibility. AI may support learning practice, but it cannot replace the playful interaction through which people learn with one another.
Lynsey Mori is an educator-researcher based in Japan whose work explores social-emotional learning, relational pedagogy, and planetary well-being. She teaches at Japanese universities and advocates for culturally responsive, care-centered approaches to language education that honour emotional depth, justice, and our interconnected futures.