How Readers Empathize with Characters: Theory and Example
How do readers empathize with characters? Empathy vs sympathy, Keen's and Smith's theories, what research has found, and an original example scene.
Summary and Conceptual Framework
For a reader to empathize with a character is to share, to some degree, the character's feeling and outlook. The article first separates empathy from sympathy, then introduces two theories: Suzanne Keen's theory of narrative empathy and Murray Smith's levels of recognition, alignment and allegiance. It then reports what research has found on whether reading fiction increases empathy, and which findings have failed to replicate. In an original example the same scene is written in two ways: by naming the character's emotion, and through actions the reader can infer from.
For a reader to empathize with a character is for the reader to share, to some degree and within herself, a fictional person's feeling and way of seeing events. The reader knows she is not in the character's place, yet grasps from the inside what the character is going through. Literary theory calls this "narrative empathy". The question is what starts this sharing: being told the character's emotion, or the reader working that emotion out for herself?
At a glance: kinds of bond with a character
| Response | What happens? | Example |
|---|---|---|
| Empathy | The reader shares, to some degree, the character's feeling and outlook | The reader's breath catches as the character knocks on the door |
| Sympathy | The reader feels something for the character, but does not share what the character feels | The reader pities the character or wishes her well |
| Emotional contagion | The reader is caught up in a feeling without separating whose feeling it is | The general air of the scene takes hold of the reader |
| Identification | The reader puts herself in the character's place; the term is vague and disputed among theorists | The sense that "that could have been me" |
The difference between empathy and sympathy
In everyday language the two words are often confused. In empathy the reader looks from inside the character: whatever the character is wary of losing, the reader too sees for a moment as something that could be lost. In sympathy the reader stays outside: she finds the character's situation sad and wants a good ending for her, but may not share the character's feeling at that moment. The two responses can also occur together.
In a 2004 article the philosopher Amy Coplan argues that trying to explain the relationship between reader and character with a single concept causes problems, because concepts such as "identification" and "empathy" are used vaguely. Drawing on text-processing research, Coplan reviews studies indicating that readers often adopt the perspective of one or more characters, and she argues that engagement with characters takes several forms, of which empathy is only one. An important condition in Coplan's definition of empathy is that the reader does not lose the distinction between herself and the character: the reader shares the character's feeling but remains herself.
Keen's theory of narrative empathy
One of the foundational works in this area is by Suzanne Keen. Keen developed her theory first in the article "A Theory of Narrative Empathy", published in Narrative in 2006, and then in the book Empathy and the Novel, published by Oxford University Press in 2007. Keen brings together work in psychology, narrative theory and neuroscience, and proposes a series of hypotheses about which narrative techniques may invite readers to empathize. Chief among these techniques are naming and describing the character and opening the character's inner life to the reader.
One of the most quoted sides of Keen's theory is its caution. Keen argues that no single narrative reliably evokes empathy in all its readers. She also finds the claim that reading novels makes readers more altruistic or better citizens unpersuasive. For Keen, then, a writer can invite empathy but cannot decide whether the reader accepts the invitation.
Smith's three levels: recognition, alignment, allegiance
In Engaging Characters, published in 1995 (expanded second edition 2022), the film theorist Murray Smith proposes, in place of the notion of "identifying with a character", a model made of three distinct levels, which together form what he calls the "structure of sympathy". Smith writes about cinema, but the model also applies to characters in prose:
- Recognition: The reader perceives the character as a continuing person. A name, a body, a recurring habit is enough.
- Alignment: The narrative places the reader beside the character; the reader sees what she sees and knows what she knows. This is directly tied to the choice of point of view.
- Allegiance: The reader judges the character's actions and attitudes morally, and takes or does not take her side.
The model offers an important distinction: a reader can be aligned with a character without giving allegiance to her. In a scene written from a villain's point of view the reader follows every step but does not take the villain's side. Or, conversely, the reader keeps following a character she does not approve of morally; Dexter's voice-over is one well-known example of this pairing. Keeping alignment and allegiance apart makes it plain that following a character and approving of her are not the same thing.
Does reading fiction increase empathy? What research says
In a 2008 article the psychologists Raymond Mar and Keith Oatley argue that fiction is a simulation of social experience: the reader takes part from the inside in the characters' relationships and conflicts, and can in this way learn to understand others through experience. This is a theory, not a direct measurement. A 2006 study by Mar and colleagues reports that people who read more fiction score higher on empathy measures; but this is a correlational finding and the direction of cause has not been established: empathic people may simply like fiction more.
The experimental study that drew the most attention is the article David Kidd and Emanuele Castano published in Science in 2013. The study reported that participants who read a short passage of literary fiction did better on a test of recognising others' emotions than those who read popular fiction or non-fiction. A replication carried out in three separate laboratories with the same texts (Panero et al., 2016) found no significant advantage for literary fiction. The FORRT Replication Atlas, which records replication studies, lists six replications of this study: four failed, one mixed, one successful.
Interpretation: These findings do not show that fiction has nothing to do with empathy. What they show is narrower: the claim that a single short reading leaves a measurable effect has not held up in repeated attempts. Panero and colleagues also write that their aim is not to show that reading does not benefit social cognition, but to draw attention to the difficulty of measuring such effects. Keen's caution connects here: the writer invites empathy, but the reader's response varies from reader to reader.
Example: the same evening, in two ways
The example below was written for this article. Hasan Usta is closing his watch repair shop in the back street of the market. The scene was written twice.
Version A: the emotion is named
That evening Hasan Usta was very sad. Closing the shop he had kept for forty years had crushed him. There was a deep loneliness inside him and he was afraid of the future. Still, he was kind to his last customer.
Version B: the emotion is left to the reader
White letters on the glass of the door: "Watch Repairs – 1984." Hasan Usta lowered the shutter halfway, then raised it again. On the counter lay twelve watches wrapped in paper; on each one a name was written in pencil. Most of the clocks on the wall had stopped; only the pendulum clock above the door still ticked. The door opened. A child in a school smock put a plastic wristwatch on the counter. Hasan Usta fitted the loupe to his eye, opened the case back with his thumbnail and changed the battery. The child took coins out of a pocket; Hasan Usta closed the child's fingers back over the coins. After the child had gone he took the pendulum clock down from the wall, stopped the pendulum with two fingers and wrapped the clock in newspaper. This time he lowered the shutter all the way.
Version A gives the reader four emotion words (sad, crushed, loneliness, afraid) and a judgement as well (kind). The reader receives this information; she knows what the character feels. Version B contains no emotion word. The reader works out for herself what Hasan Usta feels: from the shutter that goes halfway down and back up, from the watches with names on them, from the pendulum that is stopped.
Smith's levels can be traced in version B too. Recognition is built through the lettering on the glass, the loupe and the case back opened with a thumbnail: the reader sees a particular person with particular habits. Alignment is built by the narrative never leaving Hasan Usta's side through the scene: the reader sees only what he sees. For allegiance the narrator states no judgement; the coins are closed back into the child's hand and the reader does the judging.
Does version B evoke more empathy? This has not been measured. Some readers may find the directness of version A more moving; Keen's theory says exactly that such differences should be expected. All the example shows is that the two versions ask different work of the reader: in A the reader accepts a statement, in B she makes an inference.
From the viewpoint of Objective Projection: naming, or making room?
The Emotion Embargo, one of the two constitutional rules of Objective Projection, asks the narrator not to state an emotion by name in a scene. The rule's relation to empathy can be put this way: when the narrator says "Hasan Usta was very sad", the reader is handed a conclusion. The reader accepts it or does not. When the narrator gives only seen and heard details, the reader has to build the character's state within herself. If empathy is the reader grasping the character's outlook from the inside, it can be suggested that this work of building makes room for empathy. This is a suggestion and has not been tested. Objective description is one way of making this choice, not the only way.
The suggestion has limits. The Emotion Embargo is not a universal rule. A character naming her own feeling in her own voice, for example in interior monologue or first-person narration, can bring the reader even closer to the character, because there it is not the narrator who states the feeling but the character herself. The distance between reader and character is also a writing decision: as narrative distance shortens, the reader comes nearer the character's consciousness; as it lengthens, she watches from outside. Filter words such as "saw, heard, felt" are among the tools that increase this distance.
Removing the emotion's name is not enough on its own. The name can also be erased by putting the same bodily pattern in its place in every scene: trembling hands, a lump in the throat. Such patterns still hand the reader a ready-made conclusion; only the word has changed. How these patterns multiply in AI-generated text is taken up in the article on bodily clichés.
How to try it on your own text
- Mark the emotion names. Find the sentences in which an emotion name appears in the narrator's voice. The Scene Checker can flag them; the tool only finds the words and does not decide whether a sentence belongs where it is.
- Build recognition. Does the character have a habit or an object the reader will remember? A movement of the hand or a thing she always carries is enough.
- Choose the alignment. Whose side does the scene stand on? Whose seeing does the reader see? If this decision does not change through the scene, the reader stays with that character.
- Leave allegiance to the reader. Do not say that the character is good or bad; give an action the reader will judge.
- Ask a reader. Ask someone who has read the scene what the character feels. If the answer does not match what you had in mind, the details may be saying something else. One reader's answer is not enough for generalisation, but it is a sign.
What the doctrine cannot say here
- It cannot say that not naming the emotion increases reader empathy. Version B in this article shows a writing choice, not a measured result.
- It cannot measure whether a reader empathizes with a character. The six dimensions of Objective Projection are scene-planning dimensions; they are not a measurement of the reader's feeling.
- It cannot say that reading literature makes people more empathic, or that it does not. That question belongs to psychological research, and the findings are mixed.
- It cannot say in advance that a particular character will evoke empathy in readers. Readers respond differently to the same text.
- It cannot say that empathy is always a desired effect. Some texts keep the reader at a distance on purpose.
Limitations
This article summarises the work of Keen, Smith, Coplan, and Mar and Oatley; none of them is quoted, and not every detail of their theories is reported. The Turkish version of this article chose its own Turkish equivalents for Smith's terms; other sources may use different ones. The information about replication studies rests on the abstracts of published articles and on the FORRT Replication Atlas entry; the studies' methods were not examined in detail. The example scene was written for a single situation and is not enough for generalisation. The definitions of Objective Projection and the relations between them are in the knowledge graph of Levent Bulut, and the full list of published work is on the corpus page of Levent Bulut.
Frequently asked questions
What does it mean for a reader to empathize with a character?
It means the reader shares, to some degree and within herself, a fictional character's feeling and way of seeing events. The reader knows she is not in the character's place, yet grasps from the inside what the character is going through. Literary theory calls this narrative empathy.
What is the difference between empathy and sympathy?
In empathy the reader looks from inside the character and shares, to some degree, the character's feeling. In sympathy the reader stays outside; she feels something for the character, for example pity or a wish for her well-being, but may not share the character's feeling at that moment.
Can a reader empathize with a character she dislikes?
According to Murray Smith's model, a reader can be aligned with a character, that is, see what the character sees and know what she knows, without taking her side morally. So a reader can follow the outlook of a character she does not approve of and grasp some of that character's moments from the inside.
Does reading literature increase empathy?
The findings are mixed. Kidd and Castano's 2013 study reported that reading a short passage of literary fiction led to better results on a test of recognising others' emotions, but most later replications did not find this effect. A link between reading a lot of fiction and empathy has been reported, but the direction of cause has not been established.
How can a character's emotion reach the reader without naming it?
One way is for the narrator to give the character's actions and the seen and heard details around her instead of stating the emotion; the reader infers the emotion from these details. This is a writing suggestion within the Objective Projection framework, and its effect on reader empathy has not been tested.
References
- Keen, S. (2006). "A Theory of Narrative Empathy." Narrative, 14(3), 207–236. doi:10.1353/nar.2006.0015
- Keen, S. (2007). Empathy and the Novel. New York: Oxford University Press.
- Smith, M. (1995). Engaging Characters: Fiction, Emotion, and the Cinema. Oxford: Clarendon Press. (Expanded 2nd edition: Oxford University Press, 2022.)
- Coplan, A. (2004). "Empathic Engagement with Narrative Fictions." The Journal of Aesthetics and Art Criticism, 62(2), 141–152.
- Mar, R. A. and Oatley, K. (2008). "The Function of Fiction is the Abstraction and Simulation of Social Experience." Perspectives on Psychological Science, 3(3), 173–192. doi:10.1111/j.1745-6924.2008.00073.x
- Mar, R. A., Oatley, K., Hirsh, J., dela Paz, J. and Peterson, J. B. (2006). "Bookworms versus nerds: Exposure to fiction versus non-fiction, divergent associations with social ability, and the simulation of fictional social worlds." Journal of Research in Personality, 40, 694–712.
- Kidd, D. C. and Castano, E. (2013). "Reading Literary Fiction Improves Theory of Mind." Science, 342(6156), 377–380. doi:10.1126/science.1239918
- Panero, M. E., Weisberg, D. S., Black, J., Goldstein, T. R., Barnes, J. L., Brownell, H. and Winner, E. (2016). "Does reading a single passage of literary fiction really improve theory of mind? An attempt at replication." Journal of Personality and Social Psychology, 111(5), e46–e54. doi:10.1037/pspa0000064
- "Reading Literary Fiction Improves Theory of Mind." FORRT Replication Atlas. forrt.org. Accessed 19 November 2026.
Citation
@misc{bulut2026empathy,
author = {Bulut, Levent},
title = {How Readers Empathize with Characters: Recognition, Alignment and Allegiance},
year = {2026},
howpublished = {\url{https://leventbulut.com/how-readers-empathize-with-characters/}},
note = {leventbulut.com}
}