@mastersthesis{bal2025literary,
  author = {Bal, Onur},
  title = {Literary Work Readership and Its Relationship with Mentalisation and Empathy},
  school = {Yeditepe University},
  year = {2025},
  type = {Master's thesis},
  url = {https://onurbal.com/en/academics/thesis/}
}

2.3. Narrative, Fiction, and the Transformative Power of Literary Fiction

Like a triplet of nested Russian dolls, narrative contains fiction, which in turn encompasses literary fiction (Koopman & Hakemulder, 2015). According to Bruner (1990), what sets narratives apart from logical discourse is that the former deal with specific accounts of human experience, while the latter deals with generalisable truths. Narratives are universal in human culture and at the centre of human communication and social interaction, serving as tools for sharing knowledge, values, and emotions. Recent research suggests that exposure to narratives triggers a simulation-like process in the mind, allowing individuals to rehearse social experiences and derive meaning from them (Mar & Oatley, 2008; Oatley, 1999).

Fiction involves characters, events, and worlds that do not necessarily correspond to reality, while maintaining coherence and plausibility within an imagined setting (Oatley, 1999; Ryan, 1980). Readers treat plausibility as a separate dimension from factual truth; even when they know a story is fictional, perceived realism strongly predicts narrative engagement (Busselle & Bilandzic, 2008). According to the poet and philosopher Samuel Taylor Coleridge, engaging with fiction requires a “willing suspension of disbelief” (Coleridge, 1817/2008; Oatley, 1999), accepting the narrative as real within the story’s context. Not merely a means for entertainment, fiction offers opportunities for perspective-taking and emotional simulation (Mar et al., 2006).

Literary fiction represents a distinctive form within fictional narratives. Though some argue that literary fiction arises from genre conventions and culturally shaped canons (Pavel, 2003), according to Miall and Kuiken (1999), its distinguishing qualities include foregrounding, defamiliarisation, and potential modification of personal meaning. Foregrounding, as detailed by Miall and Kuiken (1999), refers to unique stylistic and narrative features, such as deviations from everyday language, unusual metaphors, and complex plot structures. These features lead to defamiliarisation, a cognitive and emotional effect that prompts readers to re-evaluate familiar concepts or emotions through new, unfamiliar lenses, ultimately transforming the reader’s feelings or beliefs.

Koopman & Hakemulder (2015) tackle a fundamental issue in the field: the inclination in research on the psychological impact of literature to confuse narrativity (story structure), fictionality (referential status), and literariness (stylistic foregrounding). This concern echoes past criticisms by Hakemulder (2000) and Mar and Oatley (2008), who cautioned against treating narrative exposure as a monolithic phenomenon. Koopman & Hakemulder (2015) build a more exact framework for analysing which textual dimensions (i.e. plot architecture, fictional status, or stylistic deviation) drive empathy and reflection related outcomes by separating these constructs. They argue that, rather than narrative structure alone, literariness, characterised by foregrounding, ambiguity, and stylistic complexity, may be especially potent in eliciting perspective-taking and reflective processing. This point is crucial for the present study, which asks whether exposure to literary fiction predicts gains in mentalisation and empathy as complementary capacities (see also Dixon et al., 1993; Miall & Kuiken, 2002).

Through the fundamental processes of role-taking and defamiliarisation, Koopman & Hakemulder suggest that literary reading activates psychological change. Role-taking aligns with simulation-based Theory-of-Mind accounts in which readers internally recreate others’ mental states (Gallese & Goldman, 1998; Mar & Oatley, 2008; Oatley, 1999). Defamiliarisation, a term from Shklovsky (1917/1965), empirically studied by Miall & Kuiken (1994, 2002), describes how stylistic foregrounding and linguistic deviation slow reading and deepen processing. While defamiliarisation can spur metacognition and ethical self-reflection, role-taking promotes cognitive empathy and social reasoning. Because literary fiction combines complex characters with unconventional stylistic choices, it can uniquely engage both empathy and reflective (explicit) and pre-reflective (implicit) mentalising (Hakemulder, 2000; Miall, 2006).

2.3.1. Short-Term Experimental Evidence

Kidd and Castano (2013) conducted a series of studies showing that exposure to literary fiction, but not popular fiction or nonfiction, enhances mind-reading. In five experiments, participants who read short literary excerpts outperformed those in popular-fiction, nonfiction, or no-reading conditions on affective (Reading the Mind in the Eyes Test [RMET]; Diagnostic Analysis of Nonverbal Accuracy 2 – Adult Faces [DANVA2–AF]) and cognitive (false-belief, Yogi) ToM tasks; lifetime print exposure to fiction, indexed by the Author Recognition Test (ART), also predicted RMET scores (Kidd & Castano, 2013). Using an expanded ART that distinguished literary, genre, and nonfiction authors, they found that only familiarity with literary fiction correlated positively with RMET, whereas genre and nonfiction did not; this relation held after controlling for Interpersonal Reactivity Index subscales such as Empathic Concern and Personal Distress (Kidd & Castano, 2017). Three preregistered replications retained the original methodology and again showed that literary fiction, but not genre fiction, familiarity predicted RMET scores, with highly exposed readers rating popular-fiction characters as flatter, suggesting long-term literary reading sharpens sensitivity to character complexity (Kidd & Castano, 2018). Although later multi-lab efforts report smaller or null effects, the overall pattern supports the view that stylistic and psychological depth selectively strengthen the Online Simulation and Perspective-Taking facets of empathy central to mentalisation. These findings laid important groundwork for exploring how the features of literary fiction bring about such effects. One proposed mechanism is foregrounding, the stylistic defamiliarisation that may enhance both aesthetic and emotional processing.

In 2016, Koopman conducted a controlled experiment to investigate how foregrounding might affect affective responses, empathy, and reflection. Participants (N = 142), comprising parents as well as students, were randomly assigned to read one of three versions of the same literary work: (1) the original text high in phonetic, semantic, and grammatical foregrounding, (2) a version without imagery (semantic foregrounding removed), and (3) a version without any foregrounding at all. Foregrounding was linked to more complex emotional reactions including ambivalence and aesthetic appreciation and found to dramatically increase perceived stylistic originality. Even after controlling for variables including trait empathy, personal experience with grief, sex, and literary exposure, quantitative results revealed that only the original version significantly raised empathic understanding for real-world others. Although the original condition had initially greater general emotional response and empathic distress, these effects vanished once SES variables were accounted for. For the readers of the original version, though foregrounding had no significant effect on reflection, the qualitative analysis showed that they were more likely to report ambivalent emotional states and general reflections on human suffering and resilience. Koopman (2016) concludes that, especially when stylistic complexity aligns with emotional themes, literariness can actively deepen affective and empathic engagement.

While Koopman emphasised style at the narrative level, other studies have examined whether specific linguistic devices, such as metaphors, may influence ToM more directly. Bowes and Katz (2015) investigated whether processing metaphorical language improves ToM. Participants completed the RMET after reading or writing literal or metaphorical sentences in three different experiments. Participants exposed to metaphor performed better than those in the literal condition in each of the three experiments. This effect was seen not only when metaphors were presented as part of text (Experiment 1), but also when the participants created imaginary contexts for the metaphors (Experiment 2). and even when the participants read metaphors in isolation (Experiment 3). The authors suggest that as metaphors force the reader to think about the speaker’s intention, a process strongly tied to ToM, metaphoric language can enhance affective ToM capacity in the short-term by eliciting social-cognitive inferences.

Other researchers have emphasised the role of individual differences. In 2013, Djikic and colleagues looked at whether reading literary fiction as opposed to nonfiction essays leads to increase in empathy. In their study, participants were randomly assigned to read either a literary short story or a nonfiction essay matched for length and literary quality, before and after which they completed cognitive and affective empathy measures (Davis, 1980, 1983), and the RMET (Baron-Cohen et al., 2001). Although no general effect of text type was found, participants low in Openness showed increased cognitive empathy after reading fiction, whereas people low in Openness did not. Another finding was that higher lifetime exposure to fiction, as measured by ART, predicted improved performance on the RMET test across both conditions. These results confirm that fictional stories can improve social cognition, particularly in those who might be less naturally inclined to perspective-taking, and that the cumulative effects of exposure to literary fiction on ToM develop over time.

Finally, two experimental studies by Bal and Veltkamp (2013) tested the hypothesis whether reading fiction increases empathy and whether emotional transportation mediated this effect. Participants were randomly assigned to read either a fictional narrative (excerpt from Saramago’s Blindness or a Sherlock Holmes story), or a nonfiction news article. Empathy was assessed at baseline, immediately after reading, and one week later. Their results revealed that reading fiction increased empathy only for readers who had a high degree of narrative transportation into the story. On the other hand, fiction readers who experienced low transportation showed a decline in empathy over time. Furthermore, the empathic gains among highly transported fiction readers became more evident following a one-week delay, indicating that the psychological impact of fiction may develop gradually. This delayed effect supports earlier research demonstrating fiction simulates complex social experiences and affects social cognition over time (Mar et al., 2006). No relationship was observed for the nonfiction condition regarding empathy. Hence, the authors provide experimental evidence for the causal influence of fiction in promoting empathy, particularly in emotionally invested readers over time.

2.3.2. Long-Term Exposure Evidence

Castano and colleagues (2020) investigated whether long-term exposure to literary as opposed to popular fiction predicts differences in social cognition in three areas: attributional complexity, egocentric bias, and interpersonal accuracy. The authors discovered that reading literary fiction was positively correlated with more sophisticated causal reasoning about other people (attributional complexity), improved performance on affective ToM tasks (RMET), and increased accuracy in estimating the beliefs and attitudes of others. They did this by using a modified ART that distinguished between authors of literary and popular fiction. These findings support earlier findings that the psychological depth and structural ambiguity of literary fiction foster reflective and socially attuned cognition (Kidd & Castano, 2013; Mar & Oatley, 2008). On the other hand, exposure to popular fiction was unrelated to both RMET performance and interpersonal accuracy and negatively correlated with attributional complexity, indicating that it may reinforce simplistic views of human behaviour. After controlling for gender and education, the link between literary fiction and reduced egocentric bias was significant The findings suggest that popular fiction does not seem to provide the same cognitive or empathic affordances as literary fiction, which seems to promote nuanced, accurate, and less self-centred social reasoning (Castano et al., 2020).

Expanding earlier studies linking general fiction exposure to social-cognitive skills, Fong and colleagues (2013) looked at whether various genres of fiction are differently correlated with interpersonal sensitivity. The authors studied the effects of lifetime print exposure to Domestic Fiction, Romance, Science Fiction/Fantasy, and Suspense/Thriller genres using the revised Author Recognition Test (ART-R). They found that performance on the RMET was positively correlated with overall fiction exposure while nonfiction exposure was not. Where genres were concerned, only Romance fiction consistently predicted higher interpersonal sensitivity when controlling for personality traits (Big Five), sex, English fluency, and other genre exposures, with the Suspense/Thriller and Domestic Fiction genres showing weaker, less robust associations. Interpersonal sensitivity was unrelated to reading science fiction or fantasy. These results imply that genres stressing interpersonal relationships and emotional complexity may offer richer simulations of social experience, hence improving readers’ affective ToM (Mar & Oatley, 2008; Zunshine, 2006).

Finally, Appel and Richter (2007) studied whether fiction narratives could modify readers’ real-world beliefs, even in the case when they included false information. Participants read a short story that included subtle accurate or inaccurate information about the world, and their beliefs were assessed either immediately after reading or two weeks later. Their findings were that not only their beliefs were influenced, but that the effect of the inaccurate information in the text increased over time, a phenomenon known as the absolute sleeper effect. The results highlight the persuasive potential of fiction in the long-term.

2.3.3. Moderators and Mechanisms

This section surveys theoretical and empirical models explaining how literary fiction exerts its effects on social cognition. These include mechanisms of embodied simulation, stylistic complexity, emotional immersion, and motivational alignment. Each pathway may differentially influence affective and cognitive empathy, self-mentalisation and other-mentalisation, and broader existential reflection.

Gallese (2001) puts forward the Shared Manifold Hypothesis, arguing that there are three levels to embodied simulation which is at the core of empathy: subpersonal (the mirror neuron system, whereby observing someone else’s behaviours, feelings, or emotions triggers the same brain regions in the observer as if they were experiencing these themselves), functional (the ability to take on the other’s perspective as if it were one’s own), and phenomenological (the sense of similarity between self and other). Importantly, this model views empathy as a pre-reflective, embodied understanding of others that facilitates goal attribution and emotion recognition rather than as a detach process of inference. By establishing simulation as the main mechanism for social cognition and interpersonal resonance, Gallese (2001) provides a convincing neurocognitive substitute for theory-theory explanations.

A complementary model is Castano’s (2024) “less-is-more” hypothesis, suggesting that fiction improves emotion recognition not by means of frequent use of emotional words, but by means of inviting readers to infer emotions from indirect cues, training their inferential and simulated social cognition. Based on correlational and experimental data (e.g., Kidd & Castano, 2013; Schwering et al., 2021), Castano (2024) argues that literary fiction, with its complex syntax, implicit emotion language, and concrete descriptions helps to develop emotional awareness and ToM. This contrasts with popular fiction, which offers less chance for inference and embodied simulation by using simpler syntax and more explicit emotional labels. Based on this, Castano argues literary fiction’s tendency to “show but not tell” elicits deeper mentalising, drawing on corpus evidence of linguistic (Castano et al., 2023) and on work linking familiarity with literary authors to higher emotion recognition scores (Castano, 2024; Schwering et al., 2021). Developmental work (e.g., Peskin & Astington, 2004) supports this assertion by demonstrating that emotionally implicit stories may better train ToM even in children and extends it to adult readers by linking literary style to cognitive engagement.

Among the most original contributions of Koopman and Hakemulder (2015) is their conception of “stillness,” a condition of cognitive openness and slowed processing made possible by literary engagement. Conceptually derived from Martel’s (2009) contemplative view of literature and developed empirically by Koopman and Hakemulder (2015), who relate “stillness” to aesthetic distance and reflection (Kuiken et al., 2004; Miall, 2006). Stillness, in contrast with transportation which emotionally immerses the reader, gives space to the reader to be reflective in an engaged but not absorbed manner. Koopman and Hakemulder (2015) contend that this reflective stillness might be what helps readers to face existential concerns and ambiguity. Within the framework of the present study, this model presents a convincing explanation of how literary fiction might influence not only ToM and empathy but also existential issues, which remain underexplored in most empirical research on reading. Koopman and Hakemulder (2015) argue that literary fiction, through round characters and foregrounded stylistic features which lead to defamiliarisation (Miall & Kuiken, 1994), encourages reflective engagement through aesthetic distance (Cupchik et al., 1998), which together give rise to stillness, a contemplative state that enables deeper self-understanding and other-understanding. They distinguish between cognitive and affective empathy, trait vs. state empathy, and narrative vs. real-world empathy, emphasizing that literary fiction uniquely fosters narrative empathy and reflective thinking through this interplay of narrativity, fictionality, and literariness. Although empirical evidence is still developing, they suggest that reader engagement, prior reading experience, and emotional transportation moderate these effects. Their model helps to explain why studies like that of Johnson (2012) and Bal and Veltkamp (2013) revealed empathy increases only among participants who demonstrated high transportation. The authors emphasise the difference between affective and cognitive empathy, referring to evidence from the study of Shamay-Tsoory and colleagues (2009) who found the two types of empathy rely on different neural systems.

Bal and colleagues (2011) in their conceptual analysis article propose a conceptual framework explaining how fictional narrative exposure influences behaviour through the processes of transportation and transformation. The authors differentiate between long-term cognitive and behavioural changes, such as creativity, interpersonal behaviour, and empathy, from short-term affective benefits, such as detachment from work stress. Building upon Oatley’s transportation-transformation model, they argue that fictional narratives expose readers to emotionally compelling and cognitively dissonant viewpoints, which through introspection challenges their mental schemas and leads to behavioural changes. They identify key moderators as frame of reference, identification with characters, and perceived realism. Though the work is theoretical, preliminary empirical research by Bal and Veltkamp (2013) supports their assertions that over time emotional transportation and transformative reflection after reading fiction can increase empathy.

Appel and Richter (2010) investigated how readers’ emotional disposition affects the persuasiveness of fictional stories. In two experiments, they tested whether need for affect, the tendency to approach or avoid emotional experiences, predicts narrative transportation, which in turn predicts belief change. Participants in the first experiment read either a fictional story including a psychiatric patient who commits a violent crime, or a control story on an unrelated topic. The emotionally intense story led to increased belief in the dangerousness of psychiatric patients, but only among participants high in need for affect who also reported high transportation. In the second experiment, the subject was organ donation for both conditions, though the first story had high emotional content whereas the second had low emotional content. Again, only the highly affect-motivated and transported readers showed belief change, but crucially, only after reading the high-emotion version. In both experiments, belief change only happened among participants high in need for affect who also reported high transportation into the story. These findings imply that emotional motivation impacts not only the degree of readers’ engagement with a story but also their internalisation of its implied worldview.

Djikic and Oatley (2014) show how literary fiction causes self-change through a process called indirect communication. Based on previous experimental data (e.g., Djikic et al., 2009; Djikic et al., 2013), they argue that literary fiction serves as a simulation of selves interacting in the social world, allowing readers to experience other mental and emotional configurations without any instructions. Literary fiction, they suggest, causes temporary personality fluctuations, especially when the story moves readers emotionally, and that these fluctuations might be precursors for more permanent transformation. Most importantly, these changes are unique to each reader and not determined by authorly intent, unfolding in ways moulded by personal resonance and emotional response. Rather than providing answers, fiction presents the reader with questions and ambiguity, which create a cognitive and emotional environment where they may step outside of their usual personality structures and imagine alternative aspects of their selves. Through literary fiction’s use of metaphor, foregrounding, and defamiliarisation, fiction engages readers in indirect communication which encourages interpretation and self-reflection (Miall & Kuiken, 2002; Zunshine, 2006).

Finally, Coplan (2004) argues that readers connect with fictional characters primarily through empathy. She shows that readers frequently take on the spatial-temporal and emotional viewpoints of protagonists, encoding events from the character’s point of view, by drawing on empirical research from text processing and narrative comprehension (e.g. Gernsbacher et al., 1992; Rinck & Bower, 2000). Coplan (2004) makes an important distinction between empathy, emotional contagion, and sympathy, emphasising that empathy requires perspective taking and self-other differentiation simultaneously to imagine the thoughts and feelings of characters while remaining conscious of one’s own thoughts and feelings. Opposing the view that readers usually feel for rather than as characters Coplan (2004) argues that empathy can accommodate emotional congruence and divergence in knowledge, preference, and identity at the same time, that is, they can both feel for a character, while also feeling like them, both of which could be different, even opposing emotions (e.g. feeling happy for a villain’s demise while also empathising with their despair and sadness, imagining what it must be like to be in their shoes). Her account reinforces the view that literary fiction promotes social understanding by allowing readers to imagine themselves in another person’s subjective world, which is a mechanism central to ToM.

Taken together, these models offer a multi-layered understanding of how literary fiction may influence the full range of mentalisation and empathy processes from implicit resonance and emotional tracking to explicit reflection and existential self-inquiry. Mechanisms such as embodied simulation and emotional inference foster affective empathy and other-mentalisation, while stylistic foregrounding, narrative ambiguity, and aesthetic distance support self-mentalisation, Perspective-Taking, and engagement with existential concerns. These processes are not uniformly activated, but depend on readers’ motivational traits, prior experience, and emotional transportation, highlighting the complex, moderated nature of fiction’s psychological effects.

2.3.4. Beyond Print

While literary fiction remains the focus of most research on narrative and social cognition, recent studies suggest that other narrative media, such as television, video games, and film, may similarly engage mentalising processes when they exhibit sufficient psychological depth and narrative complexity. If this is the case, the mechanisms through which fiction affects empathy and mentalisation may be medium-independent, hinging instead on features like character depth and interpretive demand.

Black and Barnes (2015), drawing from previous research linking literary fiction exposure and ToM capacities, investigated whether award-winning television dramas had similar effects. Participants were randomly assigned to view either an award-winning TV drama or a documentary matched in quality and production value. After viewing, participants’ affective ToM by was evaluated through RMET. Even after controlling for prior fiction exposure and sex, those who watched fictional dramas scored higher on RMET. In a second experiment, a control group viewing nothing scored the lowest on RMET. As award winning drama has a similar impact to literary fiction on affective ToM, the authors argue that rather than the medium of the content being consumed, it is the narrative complexity and character depth that facilitates mentalising.

Extending fiction exposure and social cognition search into the domain of film, Castano investigated whether art films, likened to literary fiction, enhance ToM more than Hollywood films, which are likened to popular fiction (2021). In a randomised experiment participants watched the first 20 minutes of either an art film or a Hollywood film and then completed two ToM measures: RMET for affective ToM and the Moral Judgment Task (MJS) for cognitive ToM. Viewers of art films outperformed those who viewed Hollywood films on both measures. Mediation analyses revealed that this effect was mediated by participants’ perceptions of the character’s complexity and unpredictability, supporting the idea that cognitively demanding characters prompt greater mentalising. Castano argues that, as with literary fiction, art films foster ToM by withholding schematic cues and demanding interpretive effort, as opposed to the formulaic storytelling typical of Hollywood cinema, starring stereotypical characters. This study strengthens the claim that narrative complexity and psychological depth, rather than medium, are key to cultivating social-cognitive skills.

Narrative-driven video games represent an even more interactive form of storytelling. Bormann and Greitemeyer (2015) investigated whether video games could facilitate similar effects on immersion, psychological need satisfaction, and affective ToM. In the two conditions out of three, the participants played a narrative-driven game, with the narrative group receiving instructions to focus on the narrative, and the second group to focus on the technical and gameplay aspects (hence avoiding the narrative). In the third condition, they played a simple platforming game with almost no narrative elements. Those in the narrative condition who actively interacted with the in-game story showed notably higher immersion, more satisfaction of autonomy and relatedness needs, and better affective ToM performance (Bormann & Greitemeyer, 2015), measured with RMET. Using video games in their methodology gave the authors the chance to test whether narrative processing, rather than mere exposure, drives these effects, which was found to be the case, and that literary fiction’s ability to enhance ToM is not unique to the medium.

These findings collectively underscore that the mentalising effects observed in literary fiction are not exclusive to the printed word. Rather, they extend across media that maintain literariness through psychological complexity, narrative ambiguity, and character depth. Whether conveyed through books, screens, or controllers, it is narrative richness, and not format, that emerges as the key driver of enhanced empathy, perspective-taking, and mentalisation.