There is a category error embedded in the way most people think about reading fiction. It gets classified as a leisure activity — something you do in the space left over after obligations are met, a pleasant way to disengage, a retreat from the demands of the real world into one that doesn’t exist. Even people who read seriously and widely tend to defend the practice in these terms, apologetically: it’s my way of relaxing, it’s just something I enjoy. The apology acknowledges an implied critique. Fiction, the logic goes, is not productive. It is not exercise or language learning or professional development. It is consumption, dressed up in cultural clothing.
The science on this is substantially at odds with that framing, and has been for more than a decade. What fiction reading does — specifically, reliably, and in ways that distinguish it from comparable activities — is build the capacity to model other minds. To understand, with some accuracy, that other people hold beliefs, desires, fears, and intentions that differ from your own, and to navigate the world in light of that understanding. This capacity is called theory of mind, and it is among the most consequential cognitive abilities humans possess. It underpins most of what makes complex human cooperation possible: relationships, negotiation, parenting, management, medicine, diplomacy. The suggestion that reading fiction is leisure, given what it exercises, is a category error of some magnitude.
The research
The foundational study in this area was published in Science in 2013 by David Kidd and Emanuele Castano. Across five experiments, they found that reading literary fiction — as opposed to popular fiction, nonfiction, or nothing at all — produced measurable improvements in performance on theory-of-mind tests, specifically on the Reading the Mind in the Eyes test, which assesses the ability to infer emotional states from facial expressions alone. The key variable was the type of fiction: literary fiction, which tends to portray inner lives with complexity and ambiguity, improved theory-of-mind performance in ways that other reading did not.
The study attracted significant attention and, in subsequent years, a significant replication effort. The results of that effort are mixed in a way that has been used to dismiss the original findings, but which deserves more careful reading. Three preregistered replication attempts by Kidd and Castano themselves in 2019 found two uninformative failures and one successful replication — not a clean vindication, but not a refutation either. The more important finding from the replication literature is that the association between habitual fiction reading over a lifetime and theory-of-mind ability is considerably more robust than the effect of reading a single passage in an experimental setting. The short-term effect is fragile. The long-term effect is not.
Research on lifetime reading exposure has consistently found that people who read more fiction across their lives perform better on theory-of-mind measures, even after controlling for personality variables like openness to experience that might independently predict both reading habits and social cognition. The association runs specifically through cognitive empathy — the ability to understand what someone else is thinking and feeling — more than through affective empathy, which is the tendency to feel what others feel. Fiction reading appears to train the model-building capacity specifically: the ability to hold another person’s perspective in mind and reason from within it.
What theory of mind actually is
Theory of mind is not a single skill. It is a family of capacities that together make social life legible. It includes the ability to understand that other people have beliefs different from your own and from reality — the classic test of this, the Sally-Anne task used with children, asks whether a child understands that Sally will look for a marble where she left it, not where Anne has moved it while Sally was out of the room. It includes the ability to model not just beliefs but intentions, desires, and emotional states. And at more sophisticated levels, it includes the ability to reason about what someone else believes about what you believe — the nested, recursive quality of human social cognition that makes possible everything from bluffing in negotiation to understanding irony in conversation.
Deficits in theory of mind are associated with severe social difficulties; autism spectrum disorder, which is among other things characterized by atypical social cognition, is often described partly in terms of theory-of-mind impairment. But within the neurotypical range, theory-of-mind ability varies considerably, and those variations have real consequences for the quality of relationships, professional performance, and the capacity for ethical reasoning — which depends, among other things, on being able to genuinely imagine the perspective of someone who is not you.
Fiction is, structurally, a theory-of-mind exercise. To read a novel is to inhabit, for sustained periods, the perspective of a character whose situation, history, desires, and beliefs differ from your own. Literary fiction in particular tends to make this exercise difficult in productive ways: the characters are not simple, their motivations are not transparent, the narrator may be unreliable. The reader has to do real work to understand what is happening in another mind. That work, repeated across hundreds of hours of reading over years, appears to build the underlying cognitive capacity in ways that transfer to real social situations.
Why the leisure framing matters
If fiction reading builds a capacity that is central to human social function, then the cultural framing of fiction as leisure — as optional, as a reward for productivity rather than a component of it — has costs that are difficult to measure but plausibly significant. The practical consequence of the framing is that fiction reading is among the first things shed when time pressure increases. It is what adults stop doing when careers and children and screens compete for attention. It is what schools cut when budgets tighten and standardized testing crowds the curriculum. It is what is treated, across a wide range of institutional contexts, as the least defensible allocation of limited time.
Meanwhile, the activities that have replaced leisure reading — social media, streaming, short-form video — are not theory-of-mind exercises in any comparable sense. They are, for the most part, passive consumption of pre-packaged perspectives. The social media feed delivers takes; the algorithm selects what it believes you already want; the experience rarely requires genuinely inhabiting a perspective that is not your own. There is no research suggesting these activities build theory of mind in the way fiction reading does. There is some research suggesting they may erode the attentional capacity that sustained fiction reading requires.
This is not a nostalgic argument about screens bad, books good. It is a more specific claim: that one particular activity — reading fiction seriously, for sustained periods, in ways that require active engagement with characters whose inner lives are complex — does something to the brain that most other activities don’t do, and that the cultural downgrading of that activity has consequences for the cognitive capacities of people who have abandoned it.
The counterargument, fairly stated
The replication problems with the Kidd and Castano study are real, and intellectual honesty requires acknowledging them. Short-term experimental effects from reading a single passage of fiction are not consistent across all studies. The theory-of-mind improvements may be smaller, more conditional, or more dependent on individual difference variables than the original findings suggested.
The more durable finding — that lifetime fiction reading is associated with better social cognition, across multiple studies, in multiple countries, controlling for relevant confounds — is harder to dismiss but also harder to interpret causally. It is possible, for example, that people with naturally stronger theory-of-mind ability are drawn to fiction rather than fiction developing that ability in them. The causal direction is genuinely uncertain.
But causation is not required for the practical argument. If habitual fiction readers consistently demonstrate stronger theory-of-mind capacity, then fiction reading is at minimum a correlate of a highly valuable cognitive trait — one worth cultivating, or at least worth not systematically abandoning. The uncertainty about mechanism does not change the association. And the mechanism proposed — that sustained engagement with complex fictional minds builds the cognitive infrastructure for modeling real ones — is plausible enough that treating it as confirmed for practical purposes is a reasonable bet.
What this asks of us
Reframing fiction reading as cognitive maintenance rather than leisure does not mean approaching it as homework. The pleasure of fiction is not incidental to its effect — there is evidence that narrative transportation, the state of being absorbed in a story, is specifically associated with the affective empathy benefits of reading. The engagement has to be genuine. You can’t read a novel under duress and expect the same cognitive returns as reading one you’re drawn into.
What the reframe asks is something more modest: that we stop categorizing fiction reading as a luxury to be rationalized and start treating it as a practice to be protected — the way people protect exercise time, or sleep, or anything else that has a demonstrated relationship with functioning well. The time fiction reading takes is not wasted. It is invested in a capacity without which the rest of human social life — negotiation, care, collaboration, ethical reasoning, the ability to disagree without dehumanizing — becomes more difficult.
That is not a small thing to give up in exchange for another hour of the feed.
