ArticlePDF Available

Abstract and Figures

The potential of literature to increase empathy was investigated in an experi-ment. Participants (N = 100, 69 women) completed a package of questionnaires that measured lifelong exposure to fiction and nonfiction, personality traits, and affective and cognitive empathy. They read either an essay or a short story that were equivalent in length and complexity, were tested again for cognitive and affective empathy, and were finally given a non-self-report measure of empathy. Participants who read a short story who were also low in Openness experienced significant increases in self-reported cognitive empathy (p < .05). No increases in affective empathy were found. Participants who were frequent fiction-readers had higher scores on the non-self-report measure of empathy. Our results sug-gest a role for fictional literature in facilitating development of empathy.
Content may be subject to copyright.
is is a contribution from Scientic Study of Literature 3:1
© 2013. John Benjamins Publishing Company
is electronic le may not be altered in any way.
e author(s) of this article is/are permitted to use this PDF le to generate printed copies to
be used by way of oprints, for their personal use only.
Permission is granted by the publishers to post this le on a closed server which is accessible
to members (students and sta) only of the author’s/s’ institute, it is not permitted to post
this PDF on the open internet.
For any other use of this material prior written permission should be obtained from the
publishers or through the Copyright Clearance Center (for USA:
Please contact or consult our website:
Tables of Contents, abstracts and guidelines are available at
John Benjamins Publishing Company
Reading other minds
Eects of literature on empathy
Maja Djikic, Keith Oatley and Mihnea C. Moldoveanu
University of Toronto
e potential of literature to increase empathy was investigated in an experi-
ment. Participants (N = 100, 69 women) completed a package of questionnaires
that measured lifelong exposure to ction and nonction, personality traits, and
aective and cognitive empathy. ey read either an essay or a short story that
were equivalent in length and complexity, were tested again for cognitive and
aective empathy, and were nally given a non-self-report measure of empathy.
Participants who read a short story who were also low in Openness experienced
signicant increases in self-reported cognitive empathy (p < .05). No increases
in aective empathy were found. Participants who were frequent ction-readers
had higher scores on the non-self-report measure of empathy. Our results sug-
gest a role for ctional literature in facilitating development of empathy.
literature, art, empathy, perspective taking, theory of mind
Is ction capable of prompting empathy in readers? In this paper we hope to take
a step towards answering this question with an experiment in which we measured
changes of empathy in people who were asked to read a literary text that was ei-
ther a ctional short-story or a non-ctional essay. At the same time we measured
readers’ personalities, to see whether particular traits were associated with any
changes that occurred.
Our study concerns the function of ction (Mar & Oatley, 2008), so to intro-
duce it we need rst to discuss what ction is, and how it diers from non-ction.
We suggest that four principles characterize ction, as follows.
Scientic Study of Literature : (), –.  ./ssol...dji
 – / - – © John Benjamins Publishing Company
© 2013. John Benjamins Publishing Company
All rights reserved
Reading other minds 
Subject matter
Fiction is oen taken to be description that has been made-up (the etymology of
the word “ction” is “something made”). It is generally distinguished from non-
ction, which implies a subject matter of fact. For psychology, the distinction made
in this way is not very helpful. More helpful is a study by Appel and Malakar (2012)
who asked people to read a piece of text, which they were randomly assigned to be
told was ction, non-ction, or fake (a story purporting to be true, but with facts
that had been fabricated). Engagement in reading, measured by a scale of trans-
portation, was lower for the text presented as fake than for texts presented as either
ction or non-ction, and readers were more critical of the text they were told was
fake. So readers were well aware of properties of ction, and dierentiated it from
fake. An issue studied by Prentice, Gerrig, and Bailis (1997) is whether ction en-
courages people to believe things that are not true. ey found that, as compared
with people who read about circumstances with which they were familiar, people
who read about circumstances with which they were not familiar were more liable
to believe assertions that were weak or unsupported. Fiction writers are, however,
usually careful to ensure factual material in their writings is correct, and they
should be. So despite ction tending to encourage belief in its imagined worlds, it
is dierent from texts in which facts have been falsied, or simply made up.
Non-ction has become closer to ction in recent times because techniques
of ction have been introduced into journalism to make it more engaging. Wolfe
(1975) has identied four such techniques: scene-by-scene construction, point
of view, concentration on dialogue, and depiction of what Wolfe calls status life,
meaning the ways in which individuals express their status in the various hierar-
chies in which they live.
We propose that rather than thinking of it as made-up, ction is better char-
acterized in terms of subject matter: the social world. In an experimental study of
this subject matter Mar (2007) compared eects of reading a short story and an
essay from the New Yorker. Participants who read the story, though not those who
read the essay, improved on a test of social reasoning. All participants performed
the same on a test of analytical reasoning.
e typical subject matter of ction is selves and their interactions in the so-
cial world. By contrast non-ction can be about many things: about the history
of warfare, about genetics, about the relation between economics and justice in
society, and so on. Published stories of either a ctional or non-ctional kind that
are based on deliberate falsication are a matter for the police rather than for the
© 2013. John Benjamins Publishing Company
All rights reserved
 Maja Djikic, Keith Oatley and Mihnea C. Moldoveanu
Bruner (1986) proposed that narrative is a distinctive mode of thinking about
agents, their intentions, and the vicissitudes these intentions meet. It contrasts with
the mode he calls paradigmatic, which is about explanations of how mechanisms
and processes work.
Mar, Oatley, and Eng (2003) tested dierences between narrative and exposi-
tory prose, by randomly assigning people to read a piece of prose text in narrative
form or a piece that was of the same length, semantic content, and reading di-
culty in expository form. Participants were asked to notice when memories came
to mind during their reading, and to mark the margin when this occurred. Aer
reading, they wrote brief summaries of these memories. For those who read the
narrative, as compared with the exposition, the memories that came to mind were
signicantly more vivid, and more frequently involved the participant as an actor
or observer in a detailed scene.
Except for some lyric poetry, ction tends to be written in the narrative mode.
Paradigmatic issues can be introduced, for instance in science ction, but when
this happens it is within a narrative framework. By contrast, though non-ction
can sometimes be written in narrative mode (as discussed above), it is frequently
paradigmatic. Most writings in science, for instance, are paradigmatic.
Emotion and identication
Successful ction is engaging and of emotional interest (Bal & Veltkamp, 2013).
It is capable of prompting emotions in the reader (Oatley, 2012). By contrast, al-
though it is good if non-ction is also engaging, emotions are not necessary to it.
Non-ction is primarily informational.
A frequent feature of ction is that it enables readers to identify with a protago-
nist (Oatley & Gholamain, 1997), perhaps to sympathize with that character, and
perhaps also to sympathize with other characters, and this is part of the emotional
appeal. is feature is suciently frequent in ction to make it typical of this form.
It is also possible to identify with protagonists in some kinds of non-ction, such
as memoir, biography, and social history, as well as news and magazine stories
about individuals.
Kaufman and Libby (2012) reported six experiments on identication. ey
wrote short pieces of ction for their student participants, in which the protagonist
was a college student whose thoughts and feelings were depicted in the story. ey
coined the term “experience-taking,” which they prefer to “identication” because
they want to contrast it with “perspective-taking.” Identication is, however, the
© 2013. John Benjamins Publishing Company
All rights reserved
Reading other minds 
more usual literary term for this kind of eect, which includes lessening of the
distinction between self and other, as readers take on the experience of a character
in a story. As Kaufman and Libby explain, rather than evaluating the events of the
story from an external point of view, readers who are high in experience-taking
relinquish some of their own individuality, and take on a character’s mindset and
point of view. In their experiments they found that the more aware participants
were of their own individuality as they read the story, the lower were their scores on
experience-taking and, conversely, when readers were asked to think of themselves
not as individuals but as average students, the higher were their scores on experi-
ence taking. In one experiment, experience taking was found to be less when the
readers had a mirror in the cubicle where they had been asked to read. In other
experiments Kaufman and Libby found that rst-person as compared with third-
person narratives increased experience taking, and also that a later rather than an
earlier introduction into a story of a protagonist’s characteristic of race or sexual
preference, which was dierent from the readers’ own, increased experience taking.
To sum up, successful ction moves one emotionally, and it oen enables
readers to take on the mindset, goals, and intentions of a protagonist, in a mode of
identication or experience-taking. e concerns and circumstances of characters
prompt emotions in the reader, but it’s not the emotions of characters one feels. e
emotions are ones own. Although some kinds of non-ction such as biography
enable it, identication is far less characteristic of non-ction than it is of ction.
e nature of ction
Fiction is oen taken to be a description of some kind. A far better characteriza-
tion is that it is a model. is idea was discussed by Aristotle (330 BCE/1970) in
Poetics, in which the principal theoretical term is mimesis. is term refers to the
relation of a piece of art to the world. Halliwell (2002) has shown that the term
has two families of meanings. English translators of Poetics indicate only one of
these, which they render as “copying,” “imitation,” “representation,” and the like.
e second and arguably more important meaning— the one on which Aristotle
concentrated— is “world-making” or “modeling.” Between the Renaissance and
the Nineteenth century, various writers referred to this sense as “dream.” e mod-
ern term is “simulation.” As Oatley (1992; 1999) has proposed, ctional stories are
simulations designed to run not on computers but on minds. ey were, arguably,
the very rst kinds of simulations.
One sense of simulation as it relates to ction is of complexes of several
processes. is is like the simulations from which weather forecasts and longer-
term predictions of climate change are made. e reason for simulations in these
© 2013. John Benjamins Publishing Company
All rights reserved
 Maja Djikic, Keith Oatley and Mihnea C. Moldoveanu
functions is that although human beings are good at understanding things one at
a time— for instance, that when a mass of cold air meets a mass of warm air, the
cold air cools the warm air so that the water vapor in the warm air condenses to
make rain or snow— we are less good at thinking of interactions among multiple
processes. Simulations take into account multiple processes. In those designed to
produce weather forecasts, these processes include changes of temperature, baro-
metric pressure, winds, topography of the earths surface, and so on, all of which in-
teract. Simulations based on such complexes give better weather forecasts than do
considerations of any single factor. Similarly if we read in a story-simulation that
Abigail is angry with Beatrice, we can understand and anticipate what is likely to
happen. If we add other factors, the situation becomes complex. Perhaps Beatrice
is Abigails three-year-old daughter, who has been behaving badly recently in a way
Abigail cant understand. Or perhaps, in a dierent scenario, Beatrice is Abigails
lover, but doesn’t want their relationship to be known to other people. e social
world is almost invariably complex in ways of these kinds, and because of this our
understanding of it can generally improve. Its at this point that the simulations
of which we are capable, in our private consciousness (Baumeister & Masicampo,
2010), in conversation (Rimé, 2009), and in ction, become helpful. e function
of ctional simulations is to enable us to imagine possible worlds and possible
outcomes, and that is why the idea that ction is a merely a description of some
kind is not helpful.
A second sense of simulation is empathy, coming to understand emotions of
others by feeling them in oneself (de Vignemont & Singer, 2006). An inuential
account of the mechanism of empathy is by Goldman (2009) who bases his expla-
nation of the process on simulation.
A third sense of simulation is also relevant to ction. We can understand
what other people are thinking, and this ability is known as theory-of-mind. One
of the two theories of how we do this is by simulation of the others thoughts in
ourselves (Harris, 1992). (e other theory is that we form a theory of what the
other person is thinking.) Arguably, as Zunshine (2006) has proposed, ction is
largely about theory-of-mind: working out what ctional characters are thinking
and feeling. Zunshine says that it is this that makes ction enjoyable. Identication,
as discussed above, can be thought of as using inner simulation processes for such
purposes. In a large meta-analysis of fMRI studies Mar (2011) found substantial
overlap between areas of the brain concerned with theory-of-mind and areas con-
cerned with understanding stories.
e conception of ction as a kind of simulation is becoming accepted. Speer,
Reynolds, Swallow, and Zacks (2009) had people in an fMRI machine read a short
story, with words displayed one-by-one on a screen so that readers did not have
to move their eyes (which adversely aects imaging). Speer et al. found that if in a
© 2013. John Benjamins Publishing Company
All rights reserved
Reading other minds 
story, a protagonist pulled a cord to turn on a light, the part of the reader’s brain as-
sociated with grasping was activated. When a protagonist entered a room, the part
of the brain associated with analyzing a scene was activated. So reading the simu-
lations of a story involves the same brain structures as those used for comparable
actions and perceptions in real life. Kaufman and Libby (2012), mentioned above,
also describe the process of experience taking as one that involves simulation.
Mar, Oatley, Hirsh, dela Paz and Peterson (2006) based their study, in which
they found that reading ction was associated with increased empathy and theory-
of-mind, on this idea of simulation. ey argued that when people learn to y
an airplane, they can improve their ying abilities in a ight simulator. Similarly,
when people engage in the simulations of ction they should become better in the
domain with which ction is concerned, including empathy and understanding
of others. is prediction about ction readers was borne out. By contrast, people
who predominantly read non-ction were found to have less good empathy and
theory-of-mind. e amount of life-time reading that people have done can be
measured accurately by Stanovich and West’s (1989) Author Recognition Test,
which is a list of names of writers and of people who are not writers (foils). Apar-
ticipant checks all the names he or she recognizes as writers. Mar et al. (2006)
modied the test to estimate ction and non-ction reading by including in the list
a set of authors of ction, such as P. D. James and Toni Morrison, a set of authors
of non-ction such as Richard Dawkins and Bob Woodward, as well as foils. In
a replication, Mar, Oatley, and Peterson (2009) again found ction reading to be
associated with higher empathy, this time aer controlling for individual dier-
ences. In a study of what genres of ction might be most eective, Fong, Mullin
and Mar (2012) found that the reading of some genres such as romance stories
had positive associations with empathy, whereas reading of science ction had
a (non-signicant) negative correlation. Mar, Tackett, and Moore (2010) found
that the amount of ction to which preschool children were exposed in terms of
number of stories they had read to them, and the number of ctional lms they
watched, predicted their performance on ve theory-of-mind tasks. e amount
of childrens watching of television (a much more variegated source) had no such
predictive eect.
In a study that is very useful in this line of thinking Johnson (2012) asked
participants to read a short story written to promote empathy and to exemplify
pro-social behavior. Measurements taken aer reading included transportation
and empathy. Soon aer they had nished reading and completing a set of ques-
tionnaires, participants saw an experimenter drop six pens, apparently by acci-
dent. ose who were more transported into the story were more likely to help
the experimenter pick up the pens, and this behavior was partly mediated by the
increase in empathy that readers experienced as a result of reading. A second study
© 2013. John Benjamins Publishing Company
All rights reserved
 Maja Djikic, Keith Oatley and Mihnea C. Moldoveanu
replicated the rst, and also showed that participants who were more transported
into the story were more likely to see photographs of faces as anxious. In another
very useful study, Bal and Veltkamp (2013) did two experiments. In the rst they
asked people to read either a Sherlock Holmes story by Arthur Conan-Doyle, or
a non-ction control piece of the same length taken from newspaper reports. e
second experiment had a similar design but with the ction piece being an excerpt
from José Saramagos novel, Blindness. ey found that readers who were highly
transported into the Conan-Doyle story became more empathetic, but readers of
the Conan-Doyle or the Saramago pieces who were less transported became less
empathetic. e eects were not found in the control condition.
e theory of ction as simulation of selves and their interactions in the so-
cial world is supported by several lines of evidence. For psychology this theory is
superior to the idea that ction is a description of some kind.
Conclusion from four considerations relating to the nature of ction
e centers-of-gravity of ction and non-ction are separate. Fiction has the gen-
eral subject matter of selves in the social world. It is in the narrative mode, and is
about intentions and the vicissitudes they encounter. It is emotionally engaging
and encourages identication or experience-taking. It is based on a simulation that
the reader runs in his or her mind. ese characteristics do not oer a denition
of ction, but they do oer a prototype. With these characteristics in mind, the
dierence between ction and non-ction can be meaningfully used as an inde-
pendent variable to investigate whether ction has distinctive eects on empathy.
e current study
Highly relevant recent studies on eects of ction, notably those of Johnson (2012)
and Kaufman and Libby (2012) have used stories that were written specially for the
studies. e whole issue of the nature and eects of ction, however, only arises
from stories that have been published by skilled novelists and short-story writers.
We thought it important to enquire about eects of ctional short stories and non-
ctional essays of a literary kind, published by known writers, which were seen as
accomplished enough to appear in anthologies.
Our aim was to enquire, in an experiment, about eects of such reading on
empathy. Two of our outcome measures were from Daviss (1983) Interpersonal
Reactivity Index: these were the scale of Empathetic Concern which Davis calls
Aective Empathy, and the scale of Perspective Taking which he calls Cognitive
© 2013. John Benjamins Publishing Company
All rights reserved
Reading other minds 
Empathy. In addition we used Baron-Cohen, Wheelwright, Hill, Raste, and Plumbs
(2001) Mind in the Eyes Test, a test of empathy that has been found to be associated
with lifetime ctional reading by Mar et al. (2006) and Mar et al. (2009). We were
also concerned to see whether readers’ personality aected the results, so for this
purpose we measured readers’ Big Five Personality Traits.
Participants were 100 university students from University of Toronto community
(average age 21.7, 69 women), who were recruited through posters distributed
across campus. Participants had spent on average 17.8 years speaking English
in English-speaking environment. All participants were treated according to
American Psychological Association and Canadian Psychological Association
ethical standard for treatment of human participants.
Participants were seated at a desk in a cubicle and given a package to complete
First they completed seven questionnaires, which included a Demographics
Questionnaire, the Big-Five Inventory (John, Donahue, & Kentle, 1991) (measur-
ing personality traits), Author Recognition Test-Revised (ART-R; Mar, Oatley,
Hirsh, de la Paz, & Peterson, 2006) (measuring life-long print exposure to c-
tion and non-ction), and Empathic Concern and Perspective Taking scales of
the Interpersonal Reactivity Index (IRI; Davis, 1980, 1983, and 1994) (measuring
emotional and cognitive empathy, respectively).
Participants were then randomly assigned to read either an essay or a short
story. Aer answering content questions about the text they had read, and rating
it on how artistic and interesting they found it, participants were given another set
of eight questionnaires, which included another administration of the Empathic
Concern and Perspective Taking scales of the Interpersonal Reactivity Index, and
one-time-only administration of a non-self-report test of empathy: Baron-Cohen
et al.s (2001) Mind in the Eyes Test.
We hoped that the multiple questionnaires, administered before and aer read-
ing the text, would disguise the purpose of the experiment and thus reduce demand
characteristics. During the debrieng, aer prompting, participants reported that
they recognized some questions as recurring, but given how many questions were
© 2013. John Benjamins Publishing Company
All rights reserved
 Maja Djikic, Keith Oatley and Mihnea C. Moldoveanu
asked, they could not recall their previous responses, and just completed question-
naires as was asked. With regards to the possibility that, due to the multiplicity
of questionnaires, participants have rushed through or randomly answered the
questions, that would only increase the random noise in the experiment, thus mak-
ing it more dicult to nd the eect if it were there. Whatever eects we found,
therefore, were strong enough (as can be seen from eect sizes), to emerge despite
the possibility of large random error.
e people were then debriefed and received a payment of $20 for their
Demographics Questionnaire. Participants were asked for their age, gender, and
number of years they had spent speaking English in English-speaking environment.
Author Recognition Test-Revised (Mar, Oatley, Hirsh, de la Paz, & Peterson,
2006). e Author Recognition Test-Fiction measures lifelong reading of ction,
and the Author Recognition Test-Nonction measures lifelong reading of nonc-
tion. e original Author Recognition Test was designed by Stanovich and West
(1989), who reported it to oer a good measure of exposure to print during a
participants lifetime. Test results correlate strongly with diary-based and other
measures of the amount of reading people do (West, Stanovich, & Mitchell, 1993).
Respondents are asked to check o from a list of names those they recognize
as authors. Guessing and social desirability eects are discouraged by letting the
respondents know that some names are not authors (they are foils). Mar et al.
(2006) revised this test to include 50 writers of ction only, 50 writers of nonc-
tion only, and 40 foils. In our experiment participants were indeed discouraged
from guessing: out of 140 names, no-one marked more than 8 foils. One of our
participants was, however, an extreme outlier, ve standard deviations above the
median on both the Author Recognition Test-Fiction and the Author Recognition
Test-Nonction, and was excluded from analyses.
Big Five Inventory (John, Donahue, & Kentle, 1991). is is a 44-item scale that
measures the Big-Five dimensions of personality (Extraversion, Conscientiousness,
Openness, Agreeableness, and Emotional Stability). It uses short phrases that are
prototypical of each dimension (John & Srivastava, 1999). Participants are asked
whether they see themselves as someone who, for example, “can be moody,” or
tends to be quiet.” Responses are scored on a 5-item Likert scale (1 = strongly
disagree, 5 = strongly agree). e scales test-retest correlations (over a 6-week
interval) are .65−.83 (John et al., 1991).
© 2013. John Benjamins Publishing Company
All rights reserved
Reading other minds 
Empathy Measures. Interpersonal Reactivity Index (Davis, 1980, 1983, 1994) is
a 28-item self-report measure with four subscales: Perspective Taking, Empathic
Concern, Personal Distress, and Fantasy. Since, in this experiment, we were in-
terested in cognitive and aective empathy, we used just the Perspective Taking
subscale and the Empathic Concern subscale, each of which has 7 items. e
Perspective Taking subscale measures tendency to assume the psychological view-
point of others (cognitive empathy) e.g. “I try to look at everybody’s side of a dis-
agreement before I make a decision.” e Empathic Concern subscale measures
the feelings of compassion for unfortunate others (emotional empathy) e.g. “I am
oen quite touched by things that I see happen.” Participants are asked to rate how
well the statements they read describe them on an 11-point Likert scale (0 = “does
not describe me at all” 10 = “describes me very well”). Davis (1983) reported good
internal consistency, with alpha coecients ranging from .68 to .79.
Mind in the Eyes Test – revised (Baron-Cohen, Wheelwright, Hill, Raste, &
Plumb, 2001). e Mind in the Eyes Test is a non-self-report measure of empathy
that consists of 36 still pictures of actors’ eye-regions (as if seen through a letter
box). For each item, respondents are asked to choose one of four possible mental
or emotional states that the photographed person might be experiencing. is test
requires an understanding of others’ mental states and their translation into an ap-
propriate emotion word, based on exposure to visual cues, so it can be considered
to test one aspect of empathy. Originally, Baron-Cohen and colleagues devised it
as a test of ‘mentalizing’ to be able to dierentiate individuals with deciencies in
cognitive empathy, such as those having Asperger Syndrome or high-functioning
autism from normal controls (Baron-Cohen et al., 2001).
Essays and Short Stories. We used eight essays and eight short stories from an-
thologies that were for the most part written in the rst half of 20th century. ey
were by well-known authors, and covered a variety of subjects, see Appendix A.
Unlike studies in which non-literary texts are constructed with particular empa-
thy-inducing themes in mind (e.g., Johnson, 2012), the present experiment tested
the eect on empathy of literary texts that have no particular agenda. For that rea-
son, the standards that were applied were literary excellence (as judged by critical
reviews and literary awards for authors in general, or texts in particular), a variety
of themes (as can be seen from text titles), and length that was appropriate for
experimental setting.
Texts were chosen to be around 6,000 words (about 10 pages). e complexity
(the level of reading diculty) of texts was measured by the Flesch-Kincaid Grade
Level score, which is calculated by the following formula: (.39 X ASL) + (11.8 X
ASW) – 15.59, where ASL is the average sentence length (the number of words in
© 2013. John Benjamins Publishing Company
All rights reserved
 Maja Djikic, Keith Oatley and Mihnea C. Moldoveanu
the whole text divided by the number of sentences), and ASW is the average num-
ber of syllables per word (the number of syllables divided by the number of words).
Since in their original form, the essays had longer sentences and more polysyllabic
and rare words, we modied all of them to reduce their complexity. We did this
in three ways: long sentences were divided, rare words were replaced with more
common synonyms, and complex syntax was simplied. We also shortened some
of the essays. All the short stories were retained in their original state.
Since 100 participants read one of 16 texts, each text was read by approximately
six participants. Given that the objective of the experiment was not to test whether
a particular text will have a particular eect, but rather whether a category (of c-
tional vs. non-ctional literary text) would have such an eect, we tried to include
as many texts as experimentally feasible. e objective for the multiplicity of texts
was to be able to generalize our conclusions to literary ction and non-ction, and
thus avoid the potential problem of making generalization about literature from
experimenting with a single text.
Level of Artistic Merit and Level of Interest. Aer they read the text to which they
had been assigned, participants were asked to rate it on Likert scales from 0 to 10
(0 = Not at all, 10 = Extremely) as to how artistic, and how interesting they found
it. is was done to ensure that essays and short stories were not systematically
more artistic or interesting in a way that might confound results.
Dependent measures. Since the subscales of the Interpersonal Reactivity Index
were administered both before and aer the experimental manipulation, Change
of Cognitive Empathy (based on the Perspective Taking Subscale), and Change
of Aective Empathy (based on the Empathic Concern subscale) were created by
regressing each of the variables at Time 2 on the respective variables at Time 1, and
using standardized residuals. Scores on the Mind in the Eyes Test were measured
just once, aer reading, and were used in their raw form.
Manipulation Checks. Aer reading the text to which they had been assigned,
participants were given ve short multiple-choice questions of fact (rather than
interpretation) to verify that they had read and understood the text. Six partici-
pants got three or more answers of the ve incorrect, and were therefore considered
not to have read or understood the text in its entirety. ey were excluded from
statistical analyses. For the other participants, the number of questions that they
got correct gave a measure of Comprehension Level of the text.
© 2013. John Benjamins Publishing Company
All rights reserved
Reading other minds 
e dierence in complexity between the essays and the short stories was rst
tested, to ensure that a confounding variable has not been introduced. No dif-
ferences were found between the essays and short stories in complexity level as
measured by Flesch-Kincaid Grade Level, t(14) = −.04, p = .97, or in length, t(14)=
.79, p = .45. Furthermore, there were indeed no signicant dierences between the
essays and short stories either on Artistic Merit, F(1,91) = .69, p = .41, or Level of
Interest, F(1,91) = .52, p = .47.
e next section presents three tables that show descriptive statistics of covari-
ates and the dependent variables. e means, standard deviations, minimum and
maximum scores of covariates (ART-Fiction/Nonction, Big-Five, Comprehension
Level, Artistic Merit, and Level of Interest) are presented in Table 1.
Table 1. Means, standard deviations (SD), minimum and maximum values of covariates.
Covariate Mean SD Min. Max.
Author Recognition-Fiction 6.2 6.50 0 31
Author Recognition-Nonction 4.71 4.07 0 21
Big-Five: Extraversion 3.34 .76 1.4 5.0
Big-Five: Agreeableness 3.48 .67 2.1 4.8
Big-Five: Conscientiousness 3.35 .66 1.4 5.0
Big-Five: Neuroticism 3.06 .67 1.5 4.25
Big-Five: Openness 3.52 .68 2.1 5.0
Comprehension Level 4.51 .72 3 5
Artistic Merit 6.02 2.24 1 10
Level of Interest 5.90 2.49 1 10
e means, standard deviations, and alphas for the Perspective Taking subscale
were M = 6.1 (SD = 1.63), α = .78 at Time 1, and M = 5.98 (SD = 1.76), α = .86 at
Time 2. e means, standard deviations, and alphas for the Empathic Concern
subscale were M = 6.72 (SD = 1.74), α = .84 at Time 1, and M = 6.63 (SD = 1.92),
α = .89 at Time 2. e correlations between Time 1 and Time 2 for Perspective
Taking and Empathic Concern subscales were .90 and .93, respectively.
Descriptive statistics for Change of Cognitive Empathy, Change of Aective
Empathy, and Mind in the Eyes Test, and correlations among them, are presented
in Table 2.
To test the central hypothesis, we ran a multivariate analysis, with Change
of Cognitive Empathy, Change of Aective Empathy, and the Mind in the Eyes
Test, as dependent variables; Condition (Essay versus Short Story) as the main
independent variable (xed-factor); with Author Recognition Test-Fiction, Author
© 2013. John Benjamins Publishing Company
All rights reserved
 Maja Djikic, Keith Oatley and Mihnea C. Moldoveanu
Recognition Test-Nonction, the Big-Five traits, Comprehension Level, Artistic
Merit, and Level of Interest, as covariates. e results showed signicant eects
for Change of Cognitive Empathy, F(11,81) = 2.09, p < .05, R
= .221, and the Mind
in the Eyes Test, F(11,81) = 1.93, p < .05, R
= .208. e eect, however, was non-
signicant for Change in Aective Empathy, F(11,81) = 1.23, p = .28. It is impor-
tant to note that the signicant eect (for both Change of Cognitive Empathy and
Mind in the Eyes Test) was driven by the covariates, and that the overall eect of
Condition (short stories vs. essays) was not signicant.
Signicant covariates of Change of Cognitive Empathy were Openness on
the Big Five personality traits, F(1,81) = 8.29, p < .01, R
= .093 , and participants
judgments of Level of Interest of the texts (1,81) = 5.38, p < .05, R
= .062.
e eect of Condition (Essay versus Short Story) on Change of Cognitive
Empathy, for high and low Openness individuals, is presented in Figure 1. Although
there was no overall eect of reading an essay as compared with a short story, you
may see from this gure that there was a strong eect for participants who were
low in Openness; the low-Openness participants who read the short story rather
than the essay underwent a positive Change in Cognitive Empathy (M = .32, SD =
.86) and this was signicantly dierent from the Change of those who were high
in Openness (M = −.28, SD = .95), t(46) = −2.32, p < .05, Cohens d = .68.
As you may also see from Figure 1, participants who were high in Openness
were generally lower on the scale of Change of Cognitive Empathy than were those
who were low in Openness, and this dierence was signicant, β = .277, p < . 01.
is is an interesting result in conjunction with our nding that Openness cor-
related positively with the Perspective Taking Scale at Time 1, r = .24, p < .05. Level
of Interest in the text had an almost signicant simple positive relationship with
Change of Cognitive Empathy, β = .198, p = .057.
Table 2. Descriptive statistics and correlations among the dependent variables.
Mean SD Min. Max. Change
of Cognitive
of Aective
Mind in
the Eyes Test
of Cognitive
.002 .78 −2.54 2.13 1.00 .21* −.04
of Aective
−.001 .71 −2.79 1.44 1.00 .04
Mind in
the Eyes Test
25.37 3.84 12 33 1.00
Note. * Correlation is signicant at the .05 level (2-tailed).
© 2013. John Benjamins Publishing Company
All rights reserved
Reading other minds 
In terms of the Mind in the Eyes Test, signicant covariates were Author Recognition
Test-Fiction, F(1,81) = 4.64, p < .05, R
= .054 and Comprehension Level, F(1,76) =
6.33, p < .05, R
= .073. e relationship of exposure to ction (Author Recognition
Test-Fiction) and the Mind in the Eyes Test was positive, β = .250, p < .05, such
that an increase in Author Recognition Test-Fiction predicted increased scores on
the Mind in the Eyes Test. Higher Comprehension Levels for the texts also led to
higher scores in the Mind in the Eyes Test, β = .266, p = .01.
As an exploratory analysis, we treated each text as a category and attempted to
predict whether some were signicantly more likely to predict Change in Cognitive
Aect. ere was no signicant dierences between texts in terms of their eect
on Change in Cognitive Aect, F(15,77) = 1.51, p = .12.
We asked how a piece of literary ction in the form of a short story, as compared
with a non-ctional essay, aects empathy. We did not nd that the type of writ-
ing (literary ction vs. literary non-ction) made a signicant dierence for our
outcome measures, except through interaction with personality variable Openness,
Essay Short story
Cognitive empathy change
Low openness
High openness
Figure 1. Perspective Taking Change across Condition (Essay vs. Short Story)
and across two levels of Openness: Low Openness (below the mean), High Openness
(above the mean).
© 2013. John Benjamins Publishing Company
All rights reserved
 Maja Djikic, Keith Oatley and Mihnea C. Moldoveanu
and the eect was limited to Change of Cognitive Empathy. For participants who
were low in Openness, change in self-reported Cognitive Empathy was signi-
cantly increased in the Short Story condition. People who were high in Openness
had their scores of Change of Cognitive Empathy lowered by reading either an
essay or a short story.
Our result is surprising on two accounts. First, given the ndings of previous
studies, such as that of Johnson (2012), we had anticipated that reading-induced
empathy to be increased by exposure to ction. One reason for the discrepancy
could be that although we designed experiment to measure dierence between
ction and non-ction, all of our texts were literary (as can be seen from the lack
of dierence in artistic merit between the two sets). If literariness is associated with
multiplicity of perspectives, perhaps the dierence between the short stories and
the essays was lost on high Openness individuals. In the ordinary course of events,
outside our study, many of the non-ction texts that people read are of a non-
literary kind, for instance in newspapers, textbooks, and reports. ese need not
invite creative interpretations, even by highly open individuals. In future it would
be worth comparing eects of non-literary texts with literary ones. Furthermore,
we did not measure transportation, which was an important mediator of the ef-
fect of reading on empathy in Johnsons (2012) study, and this measure should be
included in our future studies.
e second surprise was the dierence between people who were low and
high on Openness, with those low in Openness reporting higher Cognitive
Empathy aer reading the Short Story, and those high in Openness reporting
lower Cognitive Empathy. For those high in Openness, there may be a ceiling eect
such that they were already high in Cognitive Empathy; Openness was positively
correlated with the Perspective Taking scale of the Interpersonal Reactivity Index
at Time 1. As such, there may not have been much room for positive change on
this scale. Since these people were already cognitively empathetic, they may have
paid attention to aspects of the text other than those that might have further in-
creased their cognitive empathy. Another explanation might be that participants
who were high in Openness were already responsive to others’ opinions, so that
reading the text may have made them aware of the limitations of their empathy
with the result that aer reading they reported that they had less empathic ac-
curacy. Participants who were low in Openness reacted in the opposite direction.
Aer being immersed in another’s way of thinking in their simulation of a short
story, as compared with the more impersonal style of an essay, their self-reported
Cognitive Empathy improved.
Since the Mind in the Eyes Test measures just one aspect of cognitive em-
pathy— a visually based inference— the lack of correlation between this non-
self-report measure of empathy and the self-reported Perspective Taking scale of
© 2013. John Benjamins Publishing Company
All rights reserved
Reading other minds 
the Interpersonal Reactivity Index needs to be researched further. If self-reported
empathy is at variance with actual skills in empathic inference, than a change in
self-report in cognitive empathy, for example, may indicate a greater willingness
to understand others, rather than an actual improvement in understanding others.
Whether this increased willingness leads to actual greater empathic skills needs
to be further researched.
We also found the greater the lifelong reading of ction as measured by the
Author Recognition Test-Fiction, the better participants were in their identica-
tion of the mental states of individuals in the still pictures of the Mind in the Eyes
Test, regardless of the condition to which they were assigned. is is a useful repli-
cation of the ndings of Mar et al., (2006) and Mar et al., (2009). Mar et al., (2009)
ruled out the kind of explanation for this result in which people who were better
at understanding social cues would be more likely to read ction. ey found a
strong positive relationship between scores on the Author Recognition Test-Fiction
and scores in the Mind in the Eyes Test even when individual dierences had been
controlled for. Our result also adds to the conclusions of a recent meta-analysis
by Mol and Bus (2011) of helpful cognitive eects of reading as measured by the
Author Recognition Test. e improved scores on the Mind in the Eyes Test that
have been found to occur with ctional reading, using Mar et al.s modied Author
Recognition Test, seem likely to derive from coming to understand the emotional
lives of many kinds of literary character in many kinds of situation. is is an ef-
fect that cumulates over a lifetime. It is dierent from and complementary to the
kind of eect one can see in the short moment of an experiment, using measures
such as Cognitive Empathy and Aective Empathy derived from Daviss scales.
e set of results on increased empathy associated with higher lifetime levels
of reading ction also ts well with the ndings of Taylor, Hodges, and Kohanyi
(2003), who found that people who had been writing ction for at least ve years
scored higher on Interpersonal Reactivity Index than a normative population.
Writers are dedicated readers of their own work. To write a story they need to
enter the minds of others with greater persistence than someone reading the
story. In their work, therefore, they model a way of being that develops greater
cognitive empathy.
Another result of our experiment is that the participants who had more life-
long exposure to ction were better at correctly answering questions about the text
they have read. Comparable results have been found by Mar, Babyuk, Valenzano,
and Peterson (2008), who found that the higher peoples scores were on the Author
Recognition Test-Fiction , the larger were their vocabularies (although the general
vocabulary of ction is not larger than that of non-ction). e results also t in
with nding by Mar (2010) who found, alongside an eect of increased vocabulary
among ction readers, eects of several positive aspects of verbal reasoning.
© 2013. John Benjamins Publishing Company
All rights reserved
 Maja Djikic, Keith Oatley and Mihnea C. Moldoveanu
Given that, in our study, we found a causal eect of reading ctional literature
on empathy only in individuals who were low on Openness, we must be careful
about generalizations. We wondered whether those who were low in Openness
merely believed themselves to have increased in Cognitive Empathy when they
read a short story. More likely, we believe, is that these individuals did benet more
from the ctional story they read (or beneted from it more quickly) because their
lack was the greater. Since humans, as a species, are not born with cognitive em-
pathy but develop it in middle childhood, it seems reasonable that there could be
a potential of continuing to develop it throughout ones lifetime, and that ctional
literature could be one means of doing this. While we have obtained some evidence
for this relationship, in order to answer questions about the quality, speed, and
mechanism of this development, it is necessary to conduct further experiments.
Many people consider reading ction merely a leisure activity. e labels we
place on ction, however, do not negate its contribution to cognitive development.
e world of literature encourages us to become others in imagination, and this
may be one of most benign means of improving ones abilities in the social domain.
Of course, we can understand others by interacting with them, but in real life
misunderstanding oen causes severe upsets. Fictional literature, in which we can
misunderstand without suering negative consequences, may be a gentler teacher.
We would like to thank the Desautels Centre for Integrative inking, Rotman School of
Management, University of Toronto, for funding the research presented in this paper.
. Please note that the data collected here are a part of an omnibus experiment that addressed
several dierent dependent variables (please see Djikic, Oatley, & Carland, 2012; and Djikic,
Oatley, & Moldoveanu, in press). Here we will discuss only the questionnaires and the dependent
variables that were directly relevant to this particular experiment.
Appel, M. & Maleckar, B. (2012). e inuence of paratext on narrative persuasion: Fact, ction,
or fake? Human Communication Research, 38, 459–484.
Aristotle. (c. 330 BC/1970). Poetics (G. E. Else, Trans.). Ann Arbor, MI: University of Michigan
© 2013. John Benjamins Publishing Company
All rights reserved
Reading other minds 
Bal, P. M. & Veltkamp, M. (2013). How does ction reading inuence empathy? An experimen-
tal investigation on the role of emotional transportation. PLoS ONE doi: 10.1371/journal.
Baron-Cohen, S., Weelwright, S., Hill, J., Raste, Y., & Plumb I. (2001). e “Reading the Mind in
the Eyes” Test revised version: A study with normal adults and adults with Asperger’s syn-
drome or high-functioning autism. Journal of Child Psychology and Psychiatry, 42, 241–251.
Baumeister, R. F., & Masicampo, E. J. (2010). Conscious thought is for facilitating social and cul-
tural interactions: How mental simulations serve the animal-culture interface. Psychological
Review, 117, 945–971.
Bruner, J. (1986). Actual minds, possible worlds. Cambridge, MA: Harvard University Press.
Davis, M. (1980). A multidimensional approach to individual dierences in empathy. Catalog
of Selected Documents in Psychology, 10, 85.
Davis, M. (1983). Measuring individual dierences in empathy: Evidence for a multidimensional
approach. Journal of Personality and Social Psychology, 44, 113–126.
Davis, M. (1994). Empathy: A social psychological approach. Madison, WI: Brown & Benchmark.
De Vignemont, F., & Singer, T. (2006). e empathetic brain: How, when, and why. Trends in
Cognitive Sciences, 10, 435–441.
Fong, K., Mullin, J., & Mar, R. (2012). Fiction and interpersonal sensitivity: Exploring the role
of ction genres. Paper presented at the International Society for Empirical Research on
Literature, Montreal, July 7–10.
Goldman, A. (2009). Mirroring, simulating, and mindreading. Mind and Language, 24, 235–252.
Halliwell, S. (2002). e aesthetics of mimesis: Ancient texts and modern problems. New Jersey:
Princeton University Press.
Harris, P. L. (1992). From simulation to folk psychology: e case for development. Mind and
Language, 7, 120–144.
John, O. P., Donahue, E. M., & Kentle, R. (1991). e Big Five Inventory. Technical Report,
University of California, Berkley.
John, O. P. & Srivastava, S. (1999). Big Five Trait taxonomy: History, measurement and theoreti-
cal perspectives. In L. A. Pervin & O. P. John (Eds.) Handbook of Personality: eory and
research (2nd ed.) (pp. 102–138). New York, NY, US: Guilford Press.
Johnson, D. R. (2012). Transportation into a story increases empathy, prosocial behavior, and
perceptual bias toward fearful expressions. Personality and Individual Dierences, 52,
Kaufman, G. K. & Libby, L. K. (2012). Changing beliefs and behavior through experience-taking.
Journal of Personality and Social Psychology, 103, 1–19.
Mar, R. A. (2007). Simulation-based theories of narrative comprehension: Evidence and implica-
tions. PhD thesis. University of Toronto.
Mar, R. A. (2010). Exposure to Narrative Fiction versus Expository Nonction and the Prediction
of Vocabulary Ability. Talk presented at the Biennial Meeting of the International Society
for the Empirical Study of Literature. Utrecht, e Netherlands.
Mar, R. A. (2011). e neural bases of social cognition and story comprehension. Annual Review
of Psychology, 62, 103–134.
Mar, R. A., Babyuk, T., Valenzano, J. & Peterson, J. B. (2008). Vocabulary and lifetime exposure to
narrative ction versus expository non-ction. Poster presented at the 9th Annual Meeting
of the Society for Personality and Social Psychology, Albuquerque, NM, USA.
Mar, R. A., & Oatley, K. (2008). e function of ction is the abstraction and simulation of social
experience. Perspectives on Psychological Science, 3, 173–192.
© 2013. John Benjamins Publishing Company
All rights reserved
 Maja Djikic, Keith Oatley and Mihnea C. Moldoveanu
Mar, R. A., Oatley, K., De la Paz, J., Hirsh, J., & Peterson, J. B. (2006). Bookworms vs. nerds:
Exposure to ction versus non-ction, divergent associations with social ability, and simula-
tion of ctional social worlds. Journal of Research in Personality, 40(5), 694–712.
Mar, R., Oatley, K. & Eng, A. (2003). Abstraction and the vividness of details in ction. In
Symposium on Models and Mechanisms of Narrative Persuasion, American Psychological
Association Annual Convention, Toronto, Canada.
Mar, R., Oatley, K., & Peterson, J. B. (2009). Exploring the link between reading ction and
empathy: Ruling out individual dierences and examining outcomes. Communications,
34(4), 407–428.
Mar, R. A., Tackett, J. L., & Moore, C. (2010). Exposure to media and theory-of-mind develop-
ment in preschoolers. Cognitive Development, 25, 69–7.
Mol, S., & Bus, A. G. (2011). To read or not to read: A meta-analysis of print exposure from
infancy to early adulthood. Psychological Bulletin, 137(2), 267–296.
Oatley, K. (1992). Best laid schemes: e psychology of emotions. New York, NY: Cambridge
University Press.
Oatley, K. (1999). Why ction may be twice as true as fact: Fiction as cognitive and emotional
simulation. Review of General Psychology, 3(2), 101–117.
Oatley, K. (2012). e passionate muse: Exploring emotion in stories. New York: Oxford University
Oatley, K., & Gholamain, M. (1997). Emotions and identication: Connections between readers
and ction. In M. Hjort & S. Laver (Eds.), Emotion and the arts (pp. 163–281). New York:
Oxford University Press.
Prentice, D., Gerrig, R. J., & Bailis, D. S. (1997). What readers bring to the processing of ctional
texts. Psychonomic Bulletin and Review, 4, 416–420.
Rimé, B. (2009). Emotion elicits social sharing of emotion: eory and empirical review. Emotion
Review, 1, 60–85.
Speer, N., K., Reynolds, J. R., Swallow, K., & Zacks, J. M. (2009). Reading stories activates neural
representatins of visual and motor experience. Psychological Science, 20, 989–999.
Stanovich, K. E., & West, R. F. (1989). Exposure to print and orthographic processing. Reading
Resarch Quarterly, 24, 402–433.
Taylor, M., Hodges, S. D., & Kohanyi, A. (2003). e illusion of independent agency: Do adult
ction writers experience their characters as having minds of their own? Imagination,
Cognition and Personality, 22, 361–380.
West, R. F., Stanovich, K. E., & Mitchell, H. R. (1993). Reading in the real world and its correlates.
Reading Research Quarterly, 28, 34–50.
Wolfe, T. (1975). e new journalism. In T. Wolfe & E. W. Johnson (Eds.), e new journalism
(pp. 13–68). London: Picador.
Zunshine, L. (2006). Richardsons Clarissa and a eory of Mind. In A. Richardson and E.Spolsky
(Eds.) e Work of Fiction: Cognition, culture and complexity (pp. 127–146). Burlginton,
VT: Ashgate.
© 2013. John Benjamins Publishing Company
All rights reserved
Reading other minds 
Appendix A
Essays and short stories used in the experimental procedure
Essays Short stories
Henri Bergson: Why Do We Laugh? Paul Bowles: e Echo
John Burroughs: Science and Literature Katherine Brush: Night Club
Havelock Ellis: What Makes a Woman Beautiful? Frank O’Connor: My Oedipus Complex
Sigmund Freud: Dreams of the Death
of Beloved Persons
Jean Staord: A Country Love Story
John Galsworthy: Castles in Spain Jean Staord: In the Zoo
Stephen Jay Gould: Nonmoral Nature Wallace Stegner: Beyond a Glass Mountain
George Bernard Shaw: Killing for Sport Clark van Tilburg: e Wind and Snow
of Winter
Rabindranath Tagore: East and West Glenway Wescott: Prohibition
Corresponding author
Maja Djikic
Desautels Centre for Integrative inking
Rotman School of Management
University of Toronto
105 St. George St.
Toronto, Ontario,
Canada, M5S 3E6
Keith Oatley
252 Bloor Street
West Toronto, Ontario, Canada, M5S 1V6
Mihnea C. Moldoveanu
Desautels Centre for Integrative inking
Rotman School of Management
University of Toronto
105 St. George St.
Toronto, Ontario, Canada, M5S 3E6
... The long-term association between exposure to fiction and social-cognitive skills in adults has since been observed in multiple other studies, using a variety of measures (e.g., Black & Barnes, 2015b;Djikic et al., 2013;Fong et al., 2013;Mar et al., 2009;Schwering et al., 2021; for an overview, see . Moreover, in an fMRI study Tamir et al. (2016) found that the positive relationship between fiction exposure (ART) and performance on mindreading tasks was mediated by the degree to which the brain regions related to theory of mind were activated when participants read social narratives, providing support for the idea that social cognition develops through repeated activation of social-cognitive processes elicited by narratives. ...
... In these studies, the social-cognitive abilities of a group of participants who have been exposed to one particular kind of narrative are compared with the socialcognitive abilities of other groups that have been exposed to other types of texts (e.g., an expository text) or nothing at all. Using this approach, Djikic et al. (2013) found that participants who scored low on the personality trait "openness" experienced an increase in self-reported cognitive empathy (as measured with the self-report Perspective Taking scale of the Interpersonal Reactivity Index, IRI; Davis, 1983) after reading a literary story, but not after reading an expository text that was matched in terms of content, complexity, and length. The authors suggest that individuals who are generally not as open to new experiences benefit especially from the exposure to others' perspectives that literary narratives offer, increasing their self-reported empathic abilities (see Djikic et al., 2009b). ...
... All in all, then, the best evidence in favor of a causal effect of reading narratives on social cognition comes from the intervention studies (Kumschick et al., 2014;Montgomery & Maunders, 2015) and a handful of experiments that have not solely relied on the RMET to measure socialcognitive abilities (i.e., Bal & Veltkamp, 2013;Djikic et al., 2013;Pino & Mazza, 2016). However, even studies that have employed other measures than the RMET have not always replicated the positive effect of a single case of exposure of narratives on social cognition (e.g., De Mulder et al., 2017; see also Dodell-Feder & Tamir, 2018). ...
Full-text available
It is often argued that narratives improve social cognition, either by appealing to social-cognitive abilities as we engage with the story world and its characters, or by conveying social knowledge. Empirical studies have found support for both a correlational and a causal link between exposure to (literary, fictional) narratives and social cognition. However, a series of failed replications has cast doubt on the robustness of these claims. Here, we review the existing empirical literature and identify open questions and challenges. An important conclusion of the review is that previous research has given too little consideration to the diversity of narratives, readers, and social-cognitive processes involved in the social-cognitive potential of narratives. We therefore establish a research agenda, proposing that future research should focus on (1) the specific text characteristics that drive the social-cognitive potential of narratives, (2) the individual differences between readers with respect to their sensitivity to this potential, and (3) the various aspects of social cognition that are potentially affected by reading narratives. Our recommendations can guide the design of future studies that will help us understand how, for whom, and in what respect exposure to narratives can advantage social cognition.
... Psychologists who defend a positive relationship between fiction and social cognition typically reject the idea that fiction and non-fiction are distinguished by the former being "entertainment, with no connection to empirical validity" (Mar & Oatley, 2008, p. 173). Despite this framing in terms of the distinction between fiction and non-fiction, advocates of the value of fiction for improving social cognition generally concern themselves with features associated not only with fictionality but also with narrativity, a broad category that includes non-fiction as well (e.g., Djikic et al., 2013;Mar et al., 2006;Mar & Oatley, 2008). For example, according to Mar and Oatley's (2008) Simulation Model, which provided the theoretical framework for most of the studies included in the above meta-analysis, the primary function of narrative fiction is to simulate the social world. ...
... This, it is said, is apt to lead to increases in empathy and social knowledge, as well as differentiation and consolidation of readers' emotional repertoire. The Simulation Model does not make explicit assumptions regarding the amount of fiction exposure needed for improvement of social cognition; however, many studies that deployed the model (e.g., Bal & Veltkamp, 2013;Chlebuch et al., 2020;Djikic et al., 2013) sought and claimed to find effects that became manifest after reading a single text. Some researchers, while agreeing that narrative fiction enhances social cognition more strongly than expository non-fiction does, trace the effects to specific text features such as literariness, social content, or stylistics/structure (e.g., Chlebuch et al., 2020;Johnson et al., 2013b;Małecki et al., 2016;Panero et al., 2016). ...
... Furthermore, the expository non-fiction pieces lack the literary qualities of any of the narrative works; however, this is a natural, inevitable confound if typical narratives are contrasted with typical expository texts since literariness is a quality usually valued in narrative but not in expository texts (Wimmer, 2015). Finally, although every effort was made to justify the choice of stimulus texts, texts were still selected by the researchers, which entails the possibility of an experimenter bias even if it is the common approach in the field (Bal & Veltkamp, 2013;Black & Barnes, 2015;Djikic et al., 2013;Johnson et al., 2013aJohnson et al., , 2013bKidd et al., 2016;Koopman, 2015;Małecki et al., 2016;Tamir et al., 2016: all used reading stimuli selected by the study authors). ...
We present two experiments examining the effects of reading narrative fiction ( vs. narrative non-fiction vs. expository non-fiction) on social and moral cognition, using a battery of self-report, explicit and implicit indicators. Experiment 1 ( N = 340) implemented a pre-registered, randomized between-groups design, and assessed multiple outcomes after a short reading assignment. Results failed to reveal any differences between the three reading conditions on either social or moral cognition. Experiment 2 employed a longitudinal design. N = 104 participants were randomly assigned to read an entire book over seven days. Outcome variables were assessed before and after the reading assignment as well as at a one-week follow-up. Results did not show any differential development between the three reading conditions over time. The present results do not support the claim that reading narrative fiction is apt to improve our general social and moral cognition.
... As rationales for teaching YAL continue to proliferate-to meet reading standards; as model texts to have students meet writing standards; to have students explore the adolescent experience; to build social and emotional skills (Djikic et al., 2013;Malo-Juvera & Hill, 2020;Mar et al., 2009)-so too does the necessity to interrogate how these texts are socializing young readers to particular ways of being, especially if texts are perpetuating dominant ideologies (Sarigianides, 2012;Thein et al., 2013). Critical content analysis, such as the study described in this article, reveals the narrative structures and messages inherent in YAL that bear identity-formation implications for readers (Trites, 2000). ...
Full-text available
Numerous scholars have called for young adult literature (YAL) to be a pedagogical avenue for educating secondary and postsecondary students about sexual violence, who are often socialized into harmful beliefs about victims. In this study, we draw on Manne's theorizing of "himpathy" and "herasure" to explore the ways in which YAL considers the ideological and systemic dimensions of misogyny leading up, during, and after incidents of sexual assault. The results of our critical content analysis of eight contemporary novels reveal several themes that offer insight and implications for English educators who want to use YAL to unpack misconceptions about sexual violence.
... Frequent reading of fiction and the ability to become transported into a story correlated with higher scores on the empathy tests (Mar, Oatley, & Peterson, 2009). These results are supported by a similar study in which test results from 69 participants showed a correspondence between frequent fiction-reading and empathy scores (Djikic, Oatley, & Moldoveanu, 2013). ...
Full-text available
Fiction is used for educational purposes in various fields such as engineering, law and higher education. In medical education, the theories of medical humanities and narrative medicine propose that engaging with fictional works can provide opportuni-ties for medical students and physicians to develop central professional skills such as emotional awareness and empathy, which can help mitigate the documented empa-thy decline which occurs during medical training. Despite extensive research about fiction inclusion in medical education, there is very limited research on the in-situ processes of fiction use, and how emotional aspects of the medical profession can be conceptualised in fiction discussions. This thesis investigates the intersection of fic-tion, interaction and emotion in medical education. The purpose of the study is to contribute knowledge on the ways emotion can be constructed in medical education as affective stances and as professional emotions that medical students need to learn how to manage. Discursive psychology (DP) forms the theoretical and method-ological framework to analyse 58 hours of video- and audio-recorded fiction semi-nars from two Swedish medical schools. Emotion from a DP perspective is under-stood not as a reflection of an inner experience, but as an interactionally achieved phenomenon, deployed in the formation of social action. Results show that reflection and reflective practices are imbedded in fiction seminars, and that affective stances are constructed as a student resource to manage intensity, control, assessment and accountability. Furthermore, the professional emotions of physicians are constructed as emotional labour which students need to prepare for, and the feeling rules of the medical profession are constructed as calibrations of emotions which are both fluid and changeable. The task of learning professional emotion is oriented to as trouble-some by students as it might interfere with their focus on managing student identi-ties.
Conference Paper
Full-text available
Σκοπός της παρούσας μελέτης είναι να αναδείξει τη δυναμική της θεωρίας της Μετασχηματίζουσας Μάθησης σχετικά με την κοινωνική αλλαγή. Ο στοχαστικός διάλογος, η κριτική συνειδητοποίηση και η αλλαγή των δυσλειτουργικών αντιλήψεων, είναι δυνατόν να βοηθήσουν τα άτομα να αντιμετωπίσουν τις προκλήσεις μιας μεταβαλλόμενης κοινωνίας. Από την κριτική ανάλυση των ευρημάτων φαίνεται ότι η Μετασχηματίζουσα Μάθηση μπορεί να οδηγήσει -πέρα από την κριτική συνειδητοποίηση της πραγματικότητας- στη συλλογική μάθηση, στη διαμόρφωση κοινών στόχων και στην ανάληψη δράσης για κοινωνική αλλαγή.
Full-text available
The aim of this article is to explore the ways in which the engagement with Greek tragedy may contribute fruitfully to the unfolding of empathy in medical students and practitioners. To reappraise the general view that classical texts are remote from modern experience because of the long distance between the era they represent and today, I propose an approach to Greek tragedy viewed through the lens of historical empathy, and of the association between past situations and similar contemporary experiences, in particular. After a brief examination of the concept of empathy, its links with literary reading, and the discussion of these interrelations within the training of narrative medicine, and narrative ethics in particular, the focus turns to selected parts of Sophocles’ Philoctetes, such as the disease scene—an ancient example of pathography. Here Neoptolemus’ empathy for Philoctetes’ situation and its consequences are explored with specific interest in the modern readers’ affective response in connection with their own experiences in medical practice. Neoptolemus’ ethical conflict, which is resolved by his decision to care for Philoctetes, and the problematic nature of this attitude are both indicative of the aim of Greek tragedy to problematize universal issues and thus to point towards the instability of human life and the fluidity of human nature. Realizing through historical empathy the precariousness of human existence may lead to a better understanding and hence better care for others and open new perspectives in the development of empathy within the context of contemporary medical education and practice.
Full-text available
An innovative account that brings together cognitive science, ethnography, and literary history to examine patterns of “mindreading” in a wide range of literary works. (This full text has been made available by the MIT Open Access program; for a hard copy go to:
Full-text available
The present research examined the role of personality factors and paratextual information about the reliability of a story on its persuasiveness. Study 1 (N = 135) was focused on recipients' explicit expectations about the trustworthiness/usefulness and the immersiveness/entertainment value of stories introduced as nonfiction, fiction, or fake. Study 2 (experimental, N = 186) demonstrated that a story was persuasive in all three paratext conditions (nonfiction, fiction, or fake versus belief‐unrelated control story) and that its influence increased with the recipients' need for affect. Participants' need for cognition increased the difference in persuasiveness of a nonfictional versus a fake story. Additional mediation analyses suggest that fiction is more persuasive than fake because readers of fiction get more deeply transported into the story world.
Full-text available
Fiction literature has largely been ignored by psychology researchers because its only function seems to be entertainment, with no connection to empirical validity. We argue that literary narratives have a more important purpose. They offer models or simulations of the social world via abstraction, simplification, and compression. Narrative fiction also creates a deep and immersive simulative experience of social interactions for readers. This simulation facilitates the communication and understanding of social information and makes it more compelling, achieving a form of learning through experience. Engaging in the simulative experiences of fiction literature can facilitate the understanding of others who are different from ourselves and can augment our capacity for empathy and social inference. © 2008 Association for Psychological Science.
Full-text available
Phonological processing skills have been shown to account for some, but not all, of the variance in the word recognition ability of both children and adults. In two studies with adult subjects in the United States, the authors investigated whether orthographic processing ability--the ability to form, store, and access orthographic representations--can account for some of the additional variance in word recognition and spelling skill. A new measure of individual differences in exposure to print--the Author Recognition Test--was developed and validated in the two studies. This measure was designed to be relatively free of a confound that has plagued most print exposure indicators used in studies of adults: the tendency of respondents to give socially desirable answers. The Author Recognition Test was shown to be a remarkably robust and independent predictor of word processing ability. In Study 2, subjects' performance on this measure was shown to predict variance in orthographic processing independent of phonological factors. The results of the two studies were supportive of the idea that there are individual differences in reading and spelling caused by variation in orthographic processing skills. Moreover, these orthographic processing skills appear to be linked to print exposure, and thus to be environmentally mediated, rather than being simply indirect products of differences in phonological processing ability. Both studies demonstrate the potential usefulness of the Author Recognition Test as an indicator of print exposure in research on the cognitive consequences of literacy. /// [French] Les recherches antérieures ont montré que certaines habiletés de traitement phonologique comptent pour une partie de la variance dans la reconnaissance de mots aussi bien par les adultes que par les enfants. Deux expériences ont été menées auprès d'adultes américains pour voir dans quelle mesure le traitement orthographique peut compter pour une partie de variance supplémentaire dans la reconnaissance de mots et l'épellation. Une nouvelle mesure du degré d'exposition à l'écrit a été développée et validée dans les deux expériences: Le test de reconnaissance d'auteurs (ART). Cette mesure a été développée pour contourner le problème posé par les méthodes habituellement utilisées dans les recherches auprès d'adultes, qui font qu'ils ont tendance à répondre en conformité avec ce qu'ils jugent socialement acceptable. Le test de reconnaissance d'auteurs s'est révélé un prédicteur extrêmement robuste et indépendant de l'habileté à traiter des mots. Dans la seconde expérience, les résultats des sujets à cette mesure ont permis de prédire une partie importante de la variance pour le traitement orthographique indépendamment des facteurs phonologiques. Les résultats aux deux expériences appuient l'idée qu'une partie des différences individuelles en lecture et en épellation est due aux habiletés de traitement orthographique. De plus, ces habiletés de traitement orthographique apparaissent liées au degré d'exposition à l'écrit ce qui permet de poser comme hypothèse que l'acquisition de ces habiletés est conditionnée par l'environnement et n'est pas uniquement le produit des différences individuelles dans les habiletés de traitement phonologique. Les deux expériences ont confirmé l'utilité du test de reconnaissance d'auteurs comme mesure du degré d'exposition à l'écrit dans les recherches sur les conséquences cognitives de la littéracie. /// [Spanish] Las habilidades de procesamiento fonológico han explicado algunas, pero no todas, las variaciones en la habilidad para reconocer palabras tanto en niños como en adultos. En dos experimentos con sujetos adultos en los Estados Unidos, los autores investigaron si la habilidad de procesamiento ortográfico puede explicar la varianza adicional en el reconocimiento de palabras y su habilidad ortográfica. Una nueva medida de diferencias individuales en la exposición a materiales impresos--La prueba de reconocimiento de autores (ART)--fue desarrollada y validada en dos experimentos. Esta medida fue diseñada para estar relativamente libre de un problema que han sufrido la mayoría de los indicadores de exposición a materiales impresos que se usan en estudios de adultos--la tendencia de los sujetos a dar respuestas que ellos piensan son las que se esperan de ellos; las socialmente deseables. La prueba de reconocimiento de autores demostró ser un predictor notablemente robusto e independiente para predecir la habilidad de procesar palabras. En el Experimento 2, el desempeño de los sujetos en esta medida demostró que podía predecir la variabilidad en el procesamiento ortográfico, de manera independiente de los factores fonológicos. Los resultados de dos experimentos apoyan la idea de que hay diferencias individuales en lectura y ortografía causadas por la variación en las habilidades del procesamiento ortográfico. Además, estas habilidades de procesamiento ortográfico parecen estar relacionadas con la exposición a materiales impresos, y de esta manera estar mediados ambientalmente más bien, que ser simplemente productos indirectos de diferencias en la habilidad del procesamiento fonológico. Ambos estudios demostraron la ayuda potencial que la prueba de reconocimiento de autores tiene como un indicador de exposición a materiales impresos en la investigación de las consecuencias cognitivas del alfabetismo. /// [German] In der Vergangenheit war gezeigt worden, daß die Fähigkeiten der phonologischen Verarbeitung nur für einen Teil des Unterschieds in der Fähigkeit des Worterkennens zwischen Kindern und Erwachsenen zuständig waren. In zwei Experimenten, die mit Erwachsenen in den Vereinigten Staaten durchgeführt wurden, untersuchten die Verfasser, ob die Fähigkeit der orthographischen Verarbeitung für einen Teil des weiteren Unterschieds im Worterkennen und in der Rechtschreibkenntnis verantwortlich zeichnen kann. Für das Messen individueller Unterschiede bei der Wahrnehmung gedruckter Texte wurde ein neuer Maßstab aufgestellt und in zwei Experimenten bestätigt: der Test über Erkennen des Verfassers (ART). Dieser Maßstab wurde so aufgestellt, daß er relativ frei von einem Durcheinander ist, das die meisten Studien mit Erwachsenen, in denen Anzeigen über das Wahrnehmen gedruckter Texte untersucht wurden, negativ beeinflußte: die Neigung seitens der Teilnehmer, Antworten zu geben, die sie gesellschaftlich für erwünscht heilten. Es stellte sich heraus, daß der Verfassererkennungstest eine bemerkenswert beständige und unabhängige Voraussage der Wortverarbeitungsfähigkeiten leistete. Im zweiten Experiment zeigte sich, daß die Leistungen der Teilnehmer bei diesem Maßstab Unterschiede in der orthographischen Verarbeitung, die von phonologischen Faktoren unabhängig war, voraussagte. Die Resultate beider Experimente unterstützen die Vorstellung, daß beim Lesen und Buchstabieren individuelle Unterschiede bestehen, die durch Verschiedenheiten in den Fertigkeiten der orthographischen Verarbeitung hervorgerufen werden. Zusätzlich scheinen diese Fertigkeiten in der orthographischen Verarbeitung mit einer Wahrnehmung gedruckter Texte in Verbindung zu stehen und deshalb durch die Umgebung vermittelt zu sein--und nicht etwa einfach nur Nebenprodukte der Unterschiede in der Fertigkeit der phonologischen Verarbeitung zu sein. Beide Studien bewiesen die potentielle Nützlichkeit des Verfassererkennungstests als Textwahrnehmungsmaßstab in der Forschung über kognitive Auswirkungen in der Beherrschung der Schriftsprache.
Mimesis is one of the oldest, most fundamental concepts in Western aesthetics. This book offers a new, searching treatment of its long history at the center of theories of representational art: above all, in the highly influential writings of Plato and Aristotle, but also in later Greco-Roman philosophy and criticism, and subsequently in many areas of aesthetic controversy from the Renaissance to the twentieth century. Combining classical scholarship, philosophical analysis, and the history of ideas--and ranging across discussion of poetry, painting, and music--Stephen Halliwell shows with a wealth of detail how mimesis, at all stages of its evolution, has been a more complex, variable concept than its conventional translation of "imitation" can now convey. Far from providing a static model of artistic representation, mimesis has generated many different models of art, encompassing a spectrum of positions from realism to idealism. Under the influence of Platonist and Aristotelian paradigms, mimesis has been a crux of debate between proponents of what Halliwell calls "world-reflecting" and "world-simulating" theories of representation in both the visual and musico-poetic arts. This debate is about not only the fraught relationship between art and reality but also the psychology and ethics of how we experience and are affected by mimetic art. Moving expertly between ancient and modern traditions, Halliwell contends that the history of mimesis hinges on problems that continue to be of urgent concern for contemporary aesthetics.
SOLITARY INDIVIDUALS waiting for flights in an airport departure lounge were classified as either readers (engaged in recreational reading for 10 consecutive minutes) or nonreaders by an experimenter unobtrusively observing their behavior. Of the 217 subjects, 111 were classified as readers and 106 as nonreaders. Individuals classified as readers scored higher on several recognition checklist measures of print exposure that can be administered in a matter of minutes. Individuals judged to be high in print exposure-on the basis of either an inference from their airport behavior or an inference from their responses on the checklist measures-displayed more extensive vocabularies and cultural knowledge than did individuals low in print exposure. Although engagement in literacy activities was correlated with both age and education, exposure to print was a substantial predictor of vocabulary and cultural knowledge even after differences in age and education were controlled. The results, taken in conjunction with the outcomes of several related studies, suggest a more prominent role for exposure to print in theories of individual differences in cognitive development.
List of figures and tables Acknowledgments Prologue Part I. Theory and Function: 1. The structure of emotions 2. Intuitive and empirical approaches to understanding 3. Rationality and emotions 4. Mutual plans and social emotions Part II. Conflict and Unpredictability: 5. Plans and emotions in fictional narrative 6. Stress and distress 7. Freud's cognitive psychology of intention: the case of Dora Part III. Enjoyment and Creativity: 8. Happiness 9. Putting emotions into words Epilogue Notes References Author index Subject index.
Theorists from diverse disciplines purport narrative fiction serves to foster empathic development and growth. In two studies, participants’ subjective, behavioral, and perceptual responses were observed after reading a short fictional story. In study 1, participants who were more transported into the story exhibited higher affective empathy and were more likely to engage in prosocial behavior. In study 2, reading-induced affective empathy was related to greater bias toward subtle, fearful facial expressions, decreased perceptual accuracy of fearful expressions, and a higher likelihood of engaging in prosocial behavior. These effects persisted after controlling for an individual’s dispositional empathy and general tendency to become absorbed in a story. This study provides an important initial step in empirically demonstrating the influence of reading fiction on empathy, emotional perception, and prosocial behavior.
The illusion of independent agency (IIA) occurs when a fictional character is experienced by the person who created it as having independent thoughts, words, and/or actions. Children often report this sort of independence in their descriptions of imaginary companions. This study investigated the extent that adult writers experience IIA with the characters they create for their works of fiction. Fifty fiction writers were interviewed about the development of their characters and their memories for childhood imaginary companions. Ninety-two percent of the writers reported at least some experience of IIA. The writers who had published their work had more frequent and detailed reports of IIA, suggesting that the illusion could be related to expertise. As a group, the writers scored higher than population norms in empathy, dissociation, and memories for childhood imaginary companions.