1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
JðMELJAN HAKEMULDER/EMY KOOPMAN
Readers Closing in on Immoral Characters’ Consciousness.
Effects of Free Indirect Discourse on Response to Literary Narratives
1. Introduction
Much work in literary studies is based on the assumption that style matters. Narratology, for instance, constantly refines its set of analytical instruments, claiming
that every new distinction that is made is an essential one, with significant implications for our understanding of a particular story, or narratives in general. One
widely debated aspect regards the difference in modes of representing consciousness, such as free indirect discourse (FID). Scholarly disputes pertain to its origin,
history and distribution, its proper name (e. g., style indirect libre, erlebte Rede, free
indirect speech, veiled or disguised speech), and nature (e. g., as a purely linguistic
phenomenon, or as context-dependent reader inferences). Given the extensive theoretical attention the phenomenon enjoys, we wondered whether empirical research supports claims about the impact of this stylistic device on reader responses.
In the present article we will investigate whether FID matters, and if so, under
which conditions.
Besides the prominence of FID in narratology, there are two more reasons it may
make an interesting case in point. Hypotheses about the effects of FID are sometimes contradictory, which makes them suitable candidates for empirical testing.
Moreover, the relevance of these issues goes beyond the study of literary texts.
FID occurs in various text types (Semino/Short 2004; Fludernik 1993) as well
as in speech (Zyngier/Van Peer/Hakemulder 2007). We would like to argue, however, that the relevance of a phenomenon does not only depend on the frequency of
occurrences or their distribution. Relevance also depends on effects. Knowing more
about the validity of narratologists’ claims would deepen our understanding of the
workings of narrative in literary and other modes of communication.
To enable exact predictions, we first need to define the phenomenon and explore
the hypotheses formulated by narratologists. Consensus is that FID represents
characters’ thoughts and feelings as well as the narrator’s voice. As Leech and
Short indicate, free indirect speech has the »ability to give the flavour of the character’s words but also to keep the narrator in an intervening position between character and reader« (1981, 326). Similarly, in a linguistic definition »FID give[s] the
JLT 4:1 (2010), 41 – 62.
DOI 10.1515/JLT.2010.004
42
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Jmeljan Hakemulder/Emy Koopman
impression of combining direct discourse with indirect discourse« (RimmonKenan 1983, 112). As opposed to direct discourse, FID does not use quotation
marks, and as opposed to regular indirect discourse, FID does not use a reporting
clause (»s/he said…«). Moreover, a verb that would be present tense in direct discourse becomes past tense in FID. To take an example from the literary text we used
for the studies reported below, in the sentence ›I don’t know‹, the young officer answered, the first clause is direct discourse, while He did not know! is FID. While the
use of the third person indicates that a narrator is ›reporting‹, the thoughts and feelings that are expressed seem to originate in the character’s consciousness, which is
emphasized by the use of the exclamation point (cf. Rimmon-Kenan 1983, 115).
Thus, linguistic markers for FID are typically the presence of third person pronouns and past tense, and the absence of a reporting verb or quotation marks.
Leech and Short, however, argue that free indirect discourse does not depend
upon specific linguistic features. According to them, an exclamation point (as illustrated above) or another phrasing which signals the character’s manner of expression can make the difference between simple indirect representation of a character’s
speech and free indirect discourse (1981, 332). What primarily defines the free indirect mode is context, or more specifically, its intermediary position between characters’ and narrator’s speech. Compare the examples given by Leech and Short: »He
said that the bloody train had been late« versus »He told her to leave him alone!«
(1981, 331). While these sentences incorporate a reporting clause, the swearword
in the first and the exclamation point in the second indicate the plausible presence
of the character’s voice. In the studies presented below we examined the effect FID
has on reader responses by transposing narrator text and direct speech into FID. To
maximize the percentage of text that might be interpreted as FID by our readers, we
made use of all possible means suggested by narratologists, that is, both the linguistic markers as well as the context-based approach.
Speculations about the effects of FID typically include both irony and empathy.
According to Leech and Short free indirect speech distances the reader from the
character’s words, since »the authorial voice is interposed between the reader and
what the character says« (1981, 334). The simultaneous occurrence of two points
of view can have an ironic effect when the narrator and the character do not share
the same point of view, but irony does not necessarily occur (ibid., 336). Moreover,
as Leech and Short also indicate, compared to more indirect modes of speech representation, free indirect speech can be perceived as rather direct (ibid.). In the context of the present study, we will use free indirect speech and FID interchangeably. It
seems to us the differences are small and unperceivable for a general reading audience, although, of course this remains an empirical question.
Like Leech and Short, Rimmon-Kenan recognizes the possible effect of ironic
distancing, but suggests that FID could also promote empathy: »the tinting of the
narrator’s speech with the character’s language or mode of experience may promote
an empathic identification on the part of the reader« (1983, 115). Similarly, Wal-
Readers Closing in on Immoral Characters’ Consciousness
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
ravens (1980) suggests that the representation of thoughts through FID is paradoxically the most direct form. It creates a connection between narrator and character
and therefore between reader and character. Bringing readers closer to characters,
Walravens proposes, may well lead to identification. For our studies we selected a
text, Borowski’s short story ›The Record‹ that pertains exactly to this difference of
opinion. This short story has an ironic narrator, and a morally dubious main character with whom readers might not easily identify with. We were interested to see
whether higher levels of FID might facilitate identification, or, on the contrary, create an even larger distance between readers and the character.
Related to the issue of empathy versus irony is the debate about the polyphony or
dual voice (Pascal 1977) that is assumed to resound in FID. Again, predictions are
contradictory. Some suggest that FID can function as the masked expression of the
opinion of the author, smuggled in, in order to ingratiate itself with the innocent
reader; hence veiled or disguised speech (Kalepky, quoted in Pascal 1977, 13). A
more complex prediction can be found in Rimmon-Kenan’s account of the
FID-Hypothesis (1983, 115). On the one hand, she says, it can be argued that
the polyphony of voices hinders readers’ interpretation of the story’s moral. On
the other hand, FID may help readers make sense of immoral attitudes and behavior of characters and help readers reconstruct the implied author’s attitude toward
the character (ibid., 113–15). As we will explain below, our story presents a sinister
outlook on the main character. However, we are uncertain as to what implied readers ›should‹ feel: either straightforward moral resentment, or a milder combination
of disgust and understanding. It is not clear whether the character is a moral degenerate or just pathetic. Nor does the story hint at what makes him that way, circumstances or personality. Hence, we were interested to find out whether FID would
clarify the point of the story, the opinion of the (implied) author and narrator, or
make them even more indistinct, and finally, whether it affects readers’ judgments.
In sum, multiple contrasting hypotheses can be deduced from theory: FID either makes the attitude of the author and/or narrator toward the character less clear
or more clear, thus either blurring or clarifying the point of a story. In addition, FID
either increases readers’ empathy and sympathy for a character, or it enhances the
distance between readers and character (through the effect of irony). FID may make
the thoughts and feelings of the character seem more visible to readers. If readers
feel they have increased access to characters’ stream of consciousness, their understanding of the character may increase too. But does it? And, if so, under which
circumstances?
Before launching into empirical research attempting to answer so many questions, let us refine our hypotheses in a brief overview of earlier experimental studies.
Schram (1985) appears to have been the first to examine the effect of FID on identification with a literary character. Participants in his study – high school students –
read a version of a literary text (›Queen’s Day‹ by the Dutch author Jan Arends,
about a mentally confused but kind man who is arrested for wanting to knight a
44
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Jmeljan Hakemulder/Emy Koopman
member of the royal family) and a version of a popular narrative on a particular
disease (an anonymous account entitled ›She will never be cured‹). Schram varied
the amount of narrator’s speech and free indirect discourse in both texts. Moreover,
the gender of the main character was varied in order to find out whether a larger
amount of FID has a different effect for identification when the character is of the
same sex as the reader.
Results indicated that, as Schram predicted, readers reported higher levels of
identification when the story contained a high level of FID (1985, 226). This
was the case for the literary as well as for the popular text, but only when the protagonist was male. Confusingly, FID had the opposite effect for identification with
a female character: both male and female readers identified less with the female
character when the story contained more FID. Interestingly, in the case of the literary narrative, male as well as female readers reported more willingness to share the
experience of the different-sex character when encountering the version with a high
occurrence of FID (ibid., 229). Maybe FID increases readers’ sympathy for those
who are different from them. Schram emphasizes, however, that effect sizes were
not large enough to be convincing evidence for his central hypothesis (241).
Indirect support for Schram’s conclusions regarding the effects of FID on sympathy can be found in a study by Van Peer and Pander Maat (1996). Van Peer and
Pander Maat examined the effect of internal focalization in a story about a husband
and wife arguing about washing the dishes. Internal focalization is related to FID in
the sense that FID can be considered a specific narrative technique to ›internally‹
represent the character’s thoughts or speech. In their manipulation, they mainly
added thoughts of the characters in the direct speech mode. They found that internal focalization significantly increases readers’ sympathy with a character.
This effect only applied to a group of 15-year-olds. No effect was found for a sample
of 17-year-olds. This suggests that reader variables can override the effect of textual
features, as might also have been the case with Schram’s study.
In a study by Hakemulder (2000) two literary stories were rewritten, ›Learning
to fall‹ by Ann Beattie and ›The Butterfly‹ by Anton Chekhov. Both stories concern
a woman who commits adultery. The text manipulation resulted in a version dominated by focalizations of the main character and a version in which focalizations
were removed as much as possible, as the examples below illustrate:
Original: Focalization main character
I stand beside him on the curving walkway.
»Don’t throw coins from up here, Andrew«, I
say. »You might hurt somebody.« »Just a
penny«, he says. He holds it up to show me. A
penny: no tricks.
Reduced focalization main character
Joan stands beside him on the curving walkway.
»Don’t throw coins from up here, Andrew«, she
says. »You might hurt somebody.« »Just a
penny«, he says. He holds it up to show her.
»A penny: no tricks« should probably be interpreted as a focalization of the Inarrator. She fantasizes about what goes on in Andrew’s mind. Hence it was left out
Readers Closing in on Immoral Characters’ Consciousness
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
45
in the manipulated version. However, what readers make of the final sentence on
the right side of the table is uncertain. Do they conclude that this is what the narrator sees, or Joan? As the example shows, (character) focalization can only be reduced, not eliminated.
One of Hakemulder’s dependent variables relevant in the present context is perceived morality – the degree to which participants considered the main female character ›moral‹, ›sincere‹, ›selfish‹ and ›promiscuous‹. In both stories, focalization by
the I-narrator caused higher perceived morality than the manipulated reduced-focalization version did. The effect was significant for the Beattie story only. In all
likelihood, the rewriting was not powerful enough to counterweight the ironic narrator in the Chekhov story. Even though these results do not pertain to effects of
FID only but to a broader category of focalization, it seems this study does
make it more likely that small changes in style affect readers’ moral judgment.
Again, we see that readers are biased in favor of the agent that narrates and perceives; see for similar results also Andringa (1986).
A more recent study on FID is that of Bortolussi and Dixon, published in their
imperative publication Psychonarratology (2003). They examined the effect of FID
on character perception. In their experiment they used the story ›Rope‹ by Ann Porter. The plot is similar to that of the story used in Van Peer and Pander Maat (1996)
in that it concerns a domestic quarrel between a husband and wife. The original is
predominantly written in FID. Using this text the researchers produced eight new
versions: two in which thoughts and feelings of the male character were represented
in either FID or direct discourse, two in which thoughts and feelings of the female
character were represented in either FID or direct discourse; to enable researchers to
exclude the possibility that the differences between the groups of readers could be
attributed to the role and/or the sex of the characters, four more versions were created, using the first four and switching the roles. The resulting materials allowed for
a full-factorial analysis of the effect of FID on character perception. Assessment of
the effects focused on the degree to which readers considered the characters to be
reasonable. The prediction was that perceived rationality would increase for the
character to who FID would be attributed, irrespective of character’s gender and
role in the story.
Results indicate that for the versions in which the wife’s thoughts are represented, FID makes her more rational in the eyes of the readers. The same effect is found
for those versions in which the same thoughts are presented as the husband’s. Again
FID made the character seem more rational. These findings comply with a second
study by Van Peer and Pander Maat (2001), in which they found that readers who,
because of focalization, sympathized with a certain character were also significantly
more likely to explain this character’s behavior in situational terms rather than dispositional terms. Since people are also likely to explain their own behavior in situational terms, this means that focalization as well as the more specific narratological
device of FID make a character’s behavior more understandable.
46
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Jmeljan Hakemulder/Emy Koopman
Besides readers’ perception of the characters, Bortolussi and Dixon were also
interested in their perception of the narrator. Participants were asked whether
they thought the narrator was a man or a woman. Results revealed that indirect representation of the male character’s consciousness through FID led readers to suspect
the narrator to be a man. A similar effect was found for the female character’s FID:
readers of those versions considered it more likely that the narrator was a woman.
The researchers conclude that in readers’ perception FID strengthened the association between narrator and character. To us it seems that these results reveal that
readers used stereotyped expectations, with a male narrator assumed to be biased
toward male characters, and a female narrator toward female characters.
In sum, the available research suggests that FID might indeed affect reader responses to stories. The previous studies suggest that FID stimulates readers to ›take
sides‹, or perceive the events from the perspective of the character. The present
study will attempt to take these findings one step further. Where Schram examined
the effects of a confused but harmless character, and Bortolussi and Dixon showed
effects on perception of characters involved in a rather mundane argument about
shopping, we wanted to find out whether FID affects the perception of characters
with an immoral attitude. In our view, Rimmon-Kenan’s (1983) FID Hypothesis
seems to be more complex than the conceptions of FID we find represented in the
research designs of earlier experiments. This complexity of the effects of FID might
explain why Schram’s results were difficult to interpret. Rimmon-Kenan’s FID Hypothesis suggests that this style leads readers to assume that they are dealing with a
stream of consciousness of the character, which may bias readers’ judgment of the
characters, or at least stimulate readers’ assumption that the narrator or (implied)
author is biased. But on the other hand, FID is also assumed to enhance a polyvocality in the text, making it difficult for readers to determine the origin of the utterance or attitude. FID can help readers to reconstruct the attitude of the implied
author, but does not need to. As noted in the introduction, polyvocality may
lead to an ironic distance of the narrator toward the character (Leech/Short
1981, 336). In an attempt to add to the relevance of empirical work within the
field of narratology, we used a text in which this ambiguity, this double-edged effect, might be expected to cause more (moral) tension in reader responses. We will
attempt to assess readers’ responses at various levels: at that of the perception of the
implied author, the narrator, and character. Given the assumptions of narratologists
and the research conducted so far, the present article will examine whether FID
makes the moral of the story less clear or more clear; whether FID enhances or reduces the visibility of characters’ thoughts and feelings, and how this might be interrelated with the narrator’s perceived presence; and whether FID affects readers’
sympathy, understanding, and/or judgment of an immoral character.
Readers Closing in on Immoral Characters’ Consciousness
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
47
2. Method
2.1. Materials
Following the procedures used in previous studies, we manipulated a literary text so as
to enable testing our hypotheses. The four studies we will present here are all based on
Tadeusz Borowski’s ›The Record‹. It is set in WWII and told by an omniscient narrator, who introduces the narrative by saying that the story will be about the ambition
of one man to set a unique record in life, but will also illustrate the transient nature of
records in general. Then a young American officer enters the story. Soon we learn that
it is his ›ambition‹ to sleep with a German prisoner, Frau Elsa. What makes this a
particularly special event to him is that Frau Elsa collects lampshades made of
human skin – preferably tattooed – and, moreover that her husband is about to
be hanged for war crimes. Frau Elsa consents, but, as the narrator concludes the
story, soon after the officer has set this ›record‹, it is broken again by someone else.
Borowski’s short story (just under a 1000 words) makes use of irony through
contrasts and hyperboles, which makes the text’s meaning ambiguous. The narrator’s introduction, stressing the fragility of records, makes the aim of the officer
sound very serious, while in the context of post-WWII Germany and particularly
given the hideous atrocity of the lampshades of human skin, the aim of the officer
appears ridiculous and immoral. However, we maintained the narrator’s introduction in all versions of the text, since we wanted to find out how FID would moderate
the effects of this ironic narrative on the elements mentioned earlier: the visibility of
characters’ thoughts and feelings, readers’ sympathy, understanding and/or judgment of the character, and the clarity of the moral of the story.
Our manipulations were limited to the way speech and thoughts were represented. No information was added nor removed. Thus would enable us to attribute any
difference in responses to the story to the level of FID in the text. Most occurrences
of direct discourse in the original version were replaced with FID. The table below
shows examples of the adjustments that were made (translated from Dutch into
English).
Direct speech was replaced with free indirect discourse, removing quotation
marks and reporting clauses, changing first person pronouns into third person pronouns and present tense into past tense. The change from direct discourse to free
indirect discourse makes it unclear, to us, whether the officer was saying something
out loud or simply thinking it. The last section of the table illustrates that the text
did not allow for a full transfer of direct discourse to FID. At some points direct
discourse had to be maintained, as in the last sentence. Frau Elsa responds, of
course, to what the officer says, not to what he thinks. All these and other changes
from direct discourse to FID were expected to give readers the idea that they gained
access to the inner world of the character, accordingly influencing their sympathy
and understanding.
48
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Jmeljan Hakemulder/Emy Koopman
Table 1. Manipulations to increase FID
Original version – Direct Discourse
›You will be hanged for this, Frau Elsa‹, said the
officer, and looked
contentedly at the woman.
›He will be hanged!‹, the young officer said, and
looked sternly into the half-closed
eyes of the woman.
›I would like to set a record, an inviolable record‹,
said the young officer. For a short moment he was
lost in thought, then he said: ›Some have gathered
war trophies, but of those there are so many, they
are not exceptional. I have seen the camps, but
thousands have seen the camps. I have taken my
picture with a Russian officer, but there are
hundreds of those photographs in the Army. I have
slept with girls from liberated nations, but everyone has slept with them. I want to be the only
American who has slept with a woman who collected lampshades of human skin.‹
Manipulated Version – FID
She had to be hanged for this. He looked at the
woman contentedly.
That man would be hanged! The young officer
looked sternly into the half-closed eyes of the
woman.
He would like to set a record, an inviolable record.
For a short moment he was lost in thought. Some
gathered war trophies, but of those there are so
many, they are not exceptional. He had seen the
camps, but thousands have seen the camps. He
had taken his picture with a Russian officer, but
there are hundreds of those photographs in the
Army. He had slept with girls from liberated
nations, but everyone has slept with them. ›I want
to be the only American who has slept with a
woman who collected lampshades of human skin.‹
2.2 Procedure and Measures
All participants were undergraduate students and the experiments were presented
to them as an introduction to empirical studies of literature. After the studies were
completed the participants were debriefed about the true purpose and background
of the experiments. During the experiment, they were randomly assigned to read
one of the versions of Borowski’s story and subsequently asked to complete a brief
questionnaire containing ten items. Each item had two statements on opposite
sides of a six-point scale. Participants were requested to indicate on those scales
to what degree they agreed to two (opposing) statements. For example, item 1
read »The moral of the story is very unclear« on the left side, and on the right
»The moral of the story is overly clear«. It may be argued that FID makes it harder
for readers to determine a story point. Its polyvocality undermines the position of
the narrator as sole evaluating agent. The statements were formulated in a way that
it was either the one or the other. We chose an even number of response options (six)
to force participants to make a decision rather than opt for a neutral position which
would have been difficult to interpret, given the nature of the choice they were presented with.
Besides the item just mentioned pertaining to the ›visibility‹ of the moral of the
story, two other items assessed the perceived attitude of the (implied) author. Participants had to choose between either »The author rejects immoral aspects in
human beings« or »The author tries to increase understanding for immoral aspects
of human beings«, and between either »The author emphasizes the naivety of the
Readers Closing in on Immoral Characters’ Consciousness
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
49
officer« or »The author emphasizes the malice of the officer«. One item was meant
to determine the visibility of the narrator’s stance, with one pole labeled »It was unclear what the opinion of the narrator was« and the other »The opinion of the narrator in this story was very clear«.
Participants’ response to the character was assessed at three levels: their personal
sympathy, understanding, and judgment. They had to choose between one of the
following pairs of statements: »The officer seemed very unpleasant to me« versus
»The officer seemed very pleasant to me«, »I felt no understanding for the ambition
of the officer« versus »I felt understanding for the ambition of the officer«, and finally »It felt just right to me that the officer’s record was broken« versus »I felt sorry
for the officer when his record was broken«. We expected FID to enhance sympathy
and understanding, and hence moderate their judgments.
On top of these character perception measures we wanted to assess participants’
identification with the character. Because of his morally suspect nature we decided
to use an indirect measure. We reasoned that if FID allows readers to imagine the
position of the character a little better, they would show a tendency to explain character behavior in terms of situational factors (what is known as the fundamental
attribution error, Ross, 1977). However, if FID distances readers from the character
rather than bringing them closer to him, they may be expected to reveal a bias toward internal attribution in their understanding of the behavior of the character,
that is, in terms of his (immoral) personality. The items offered participants the
choice between »The story emphasizes how circumstances bring people to behave
immorally« (external attribution) versus »The story emphasizes the inclination of
people toward immoral behavior« (internal attribution), and between »The officer
is represented as a disturbed person« (internal attribution) versus »The officer is
represented as a victim of the war« (external attribution). The study of Van Peer
and Pander Maat (1996) showed that a story character’s focalization stimulates
readers to consider situational factors in their judgment of that character’s position
in a domestic quarrel. We were curious whether these effects would be strong
enough to influence more grave moral judgments as well.
Finally, and perhaps most importantly for the present enterprise, we wanted to
know whether FID increases or reduces the visibility of the thoughts and feelings of
the character. The relevant item therefore read on the left side »The feelings and
thoughts of the character were invisible« and on the right »The feelings and
thoughts were clearly visible«.
3. Results
3.1. Study 1
59 students of Theatre, Film, and Television Studies participated in the first experiment. The sample consisted of 44 women and 15 men. We expected that FID
50
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Jmeljan Hakemulder/Emy Koopman
would make the text more polyphone, and make the moral of the story less overt.
FID might betray the (implied) author’s sympathy for the (immoral) character, but
in combination with the ironic undertone we expected readers to be more confused
about a point or moral of the story. FID offers the illusion of an inside look into the
character’s mind and feelings, hence we expected readers to believe that the (implied) author aimed at increasing their understanding and softening their moral
judgment. Considering the topic of the story we expected that gender might
play a role in the responses to the questionnaire. However, because only few
men participated in this study (as well as the following studies) the cells were
often too small to allow a meaningful interpretation of mean differences between
men and women. Moreover, because our theoretical premises did not generate directional predictions about what differences we might expect, we decided to enter
gender as a control variable in our analyses, that is, keeping its possible influence
constant.
The results of our univariate analyses of variance confirmed the polyphony hypothesis: the FID version made it less clear to participants what the opinion of the
narrator was (p < .028; all p’s reported as two-tailed). Responses to the original
narrator version yielded an average of 3.1 (SD=1.2) and for the FID manipulation
2.4 (SD=1.1). Furthermore, we found a tendency toward an effect of FID on the
perception of the author’s aim. The FID version made readers suspect that the author wanted to enhance readers’ understanding of the immorality of human kind
(M=4.2, SD=1.0), and more so than the (original) narrator version did (M=3.8,
SD=1.0); (p <. 094). See for an overview of all the results the appendix to this
article.
Prominent in our conclusion should be the observation that very few of our hypotheses were confirmed (with only one effect at the .05 level of significance). Most
expectations did not hold. It may be that the participants who were selected for this
experiment were less likely to pick up on the subtleties of FID. We therefore decided to rerun the same study with students of literature, in particular, participants
of a course in which FID was one of the topics taught.
3.2. Study 2
For Study 2 we used exactly the same procedure and materials. We assumed that the
48 students of literature (40 women and 8 men) in this experiment would be more
sensitive to the manipulation than those of Study 1.
The results, however, indicated only a tendency toward an effect of the text manipulation on one variable: FID led to higher levels of sympathy (M=3.9,
SD=1.1) than the original narrator version of the text (M=3.3, SD=1.4); (p <
.065).
Readers Closing in on Immoral Characters’ Consciousness
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
51
Since the contexts of Study 1 and 2 differed considerably they are presented separately. One difference is that in Study 1, the first author of the present paper was
also the main instructor, while in the second he only gave a guest lecture. Another
difference is that in the context of the second course students read about narratology and were familiar with FID, where in the first participants were unlikely to have
heard of the phenomenon. Finally, only in Study 1 did the students participate for
partial course credit. However, because of the relative lack of results, we were curious to find out whether readers’ background made a larger difference to readers’
responses than text factors. It did. Conclusions need to be drawn with caution because of the limited control. We do not know exactly which participant variables
might be relevant. On four of our variables we registered near-significant differences between the two groups. These differences were consistent. The univariate analyses show that students of literature were more convinced that the story emphasizes
how humans are inclined to immoral behavior rather than showing how circumstances can are to blame (M=3.24, SD=1.28 for the literature students, M=3.73,
SD=1.32 for the film students, p < .057); how the officer is represented as a disturbed person rather than a victim of the war (M=2.56, SD=1.15 for the literature
students and M=3.02, SD=1.22, p < .074); and finally how the author condemns
the immoral in human beings rather than tries to make readers understand
(M=3.62, SD=1.10 for the literature students, and M=4.02, SD=1.02 for the
film students, p < .058). Also, as compared to the participants of Study 1, and irrespective of our text manipulation, they thought the story showed less of the inner
world of the main character (M=2.94, SD=1.24, versus M=3.41, SD=1.22 for
the film students, p < .053).
Again, what is striking about these results is the lack of effects for our text variable. We do find that the manipulation affected responses, but not to the extent
that narratology would lead us to expect. One possible explanation is that a manipulation of FID is hardly strong enough to balance readers’ objections against the
immorality of the main character. To test this assumption, we rewrote the texts,
both the original and FID version, so as to reduce the offensiveness of the characterization. Also, re-evaluating our first manipulation, we discovered more ways to increase the level of FID. Thus we ended up with a new set of four texts that would
allow us to examine the role of FID and character valence, as well as possible interactions between these two factors. In the original version of the story, Frau
Elsa was said to collect lampshades of human skin and the officer wants to sleep
with her exactly for that reason: he wants to be the only American to have slept
with such a woman. We hypothesized that the element of the lampshades of
human skin was rather distracting, and made the officer look immoral beyond understanding. To control for this, a version was created without lampshades, reducing the immorality of the officer. In Table 2, the main changes are indicated. Every
time that the original version referred to human lampshades, we changed this into
the open description of an unspecified crime.
52
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Jmeljan Hakemulder/Emy Koopman
Table 2. Manipulations to decrease immorality
Original version – Lampshades
›Which ones have they found?‹, Frau Elsa asked
peacefully while taking a sip of coffee from her tin
cup.
›I want to be the only American who has slept with
a woman who collected lampshades of human
skin.‹
After this Frau Elsa’s husband was hanged. Then
the young officer was the only man alive who had
slept this many times with a woman who collected
lampshades of human skin.
Manipulated Version – No Lampshades
›You know of my crime?‹, Frau Elsa asked peacefully while taking a sip of coffee from her tin
cup.
›I want to be the only American who has slept
with a female German criminal.‹
Then the young officer was the only man alive
who had slept this many times with a
female German criminal, whose husband had
been hanged.
As to the maximization of the level of FID-text, we noticed that in the first two
studies only direct discourse had been replaced with FID, while at one important
moment, namely the last section of the story, the narrator’s text was left as it was.
Again the narrator’s introduction was kept the same, since that introduction could
not be changed without also changing the information given. Following Leech and
Short’s (1981) remarks on the relative directness of FID in comparison to indirect
speech, it was hypothesized that replacing the narrator’s text by FID could promote
readers’ understanding of the character. Thus, in the four versions created for the
third and fourth study, the following addition of FID was made:
Table 3. Addition in Manipulation to increase FID
Original Version – Narrator’s text
Manipulated Version – FID
After this Frau Elsa’s husband was hanged. Then After this Frau Elsa’s husband was hanged. Now
the young officer was the only man alive who had he was the only man alive who has slept this many
slept this many times with a woman who collected times with a woman who collected lampshades of
lampshades of human skin.
human skin!
We expected the change from the narrator’s text to free indirect discourse to give
the readers’the impression that they gained more access to the character’s consciousness. The use of ›now‹ and the exclamation point clearly indicate the presence of the
character’s feelings, as opposed to the neutral observation in the narrator’s text.
3.3. Study 3
The new set of materials tested in the third experiment allowed us to investigate the
effects of FID on two levels of character valence. We expected that FID might be
effective in changing perceptions of a character that was only mildly objectionable,
whereas the negative valence of the original story character seemed to have overruled such effects.
Readers Closing in on Immoral Characters’ Consciousness
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
53
The results showed a main effect of our manipulation on perceived visibility of
the character’s feelings and thoughts, with FID leading to higher visibility (M=3.8,
SD=1.4) than the narrator version (M=3.13, SD=1.2, p < .003). However, this
effect is mediated by our manipulation of character valence (p < .002). The effect of
FID depends on the level of immorality. Contrary to our expectations, however, we
found that the manipulation of FID made a difference only among the readers of
the original moral depravity versions, with the FID version resulting in a higher
average score on visibility (M=4.2, SD=1.33) than the narrator version
(M=2.7, SD=1.15). In the new versions where we reduced the negative valence
of the character, it did not matter whether the text was dominated by FID
(M=3.5, SD=1.39) or not (M=3.5, SD=1.15).
One possible interpretation of these results is that negative character valence
triggers readers to focus on information that might help them to understand the
character. Where the narrator stands in between readers and the character and hinders the perception of his feelings and thoughts, FID gives them the illusion of having a privileged inside look. It is important to reemphasize that readers basically
read the same text and received the same information about the character’s thoughts
and feelings; only the style of representation differed. Why this did not turn up in
the two earlier experiments, we suggest, may be explained by the increase in the
percentage of FID text realized in our further manipulation of the story.
An alternative explanation is that the differences in responses are caused by some
reader factor that we did not assess, and, hence, could not control for. As a study by
Sopčk, Kuiken, and Miall (in press) has demonstrated, for example, readers with
different reading orientations experience FID differently. In their study, participants scoring lower on a measure of the capacity to vividly imagine a narrative
world had to be gradually convinced of the co-presence of the narrator and character in a literary text with FID, while participants scoring high on imagery vividness experienced the co-presence from the start. In a somewhat similar move, we
wanted to determine whether reading experience would make a significant difference in reactions to FID. Hence we decided to add one more experiment.
First of all, we wanted to see whether using the same set of texts would reveal
similar results. Second, we wanted to make a start in tracking background variables
that interact with FID. Comparison of the results of Study 1 and 2 did suggest that,
for instance, readers’ interests and experience with literary texts might play a role.
However, conclusions need to be extremely cautious because we have no evidence
that the participants of Study 1, the students of theatre, film and television, were
not exposed to literary texts as much as the students of literature participating in
Study 2. Finally, we wanted to investigate to what degree participants were
aware of the significance of FID in their perception of the visibility of character
thoughts and emotions.
54
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Jmeljan Hakemulder/Emy Koopman
3.4. Study 4
This final study aimed to both replicate and extend the findings of Study 3. 135
undergraduates (113 women and 22 men) enrolled in a course on literary studies
participated in the experiment. The procedure was the same as for the previous
studies, except that we asked participants to estimate the average amount of reading
for pleasure per week. In previous studies (e. g., Dixon et al. 1993) such a simple
measure was sufficient to distinguish those sensitive to particular literary text qualities from those who were not. Due to practical considerations (i. e., the time allotted to this experiment in the context of the courses) we were unable to use a longer
questionnaire – compare the use of the Literary Response Questionnaire by
Sopčk, Kuiken, and Miall (in press). For our purposes we added a (time consuming) task. As in previous experiments we asked participants to fill out the questionnaire after they read the story and put the text back in an envelope. In this final
experiment we asked participants to take out the text again and mark those
words or sentences in the text that indicated to them the thoughts and/or emotions
of the main character. We expected that readers of the FID-versions would underline words and sentences more frequently, and especially text parts that were in FID.
On the other hand, it may also be that FID affect readers’ responses on a subconscious level.
As in Study 3 the results indicate that FID significantly affects estimates of visibility of character thoughts and emotions. In the narrator versions those estimates
were significantly lower (M=2.76; SD=1.28) than in the FID versions (M=3.95;
SD=1.14, p < .000). Readers had the distinct impression that they learned more
about the inner life of the character, while they in fact received exactly the same
information. One important difference with the results of the previous study is
that this effect was independent of character valence. Thus, both in the extremely
immoral version with the human lampshades and in the moderately immoral version without human lampshades, FID had a significant effect of increasing readers’
perception of character consciousness. We are unsure about the reasons for this,
apart from the fact that the study was better controlled for possible intervening variables like reading for pleasure.
As expected we found that the narrator versions stimulated participants to consider the behavior of the character in terms of internal factors, while the FID versions made them focus more on external factors. On the scale assessing internal attribution the narrator versions scored an average of 3.93 (SD=1.39), while for the
FID versions we registered a mean of 3.36 (SD=1.44), with higher scores indicating a stronger inclination toward internal attribution (p < .027).
A tendency toward significance (p < .081) was found for the manipulation of
FID on the clarity of the opinion of the narrator. Higher scores for FID (M=3.04,
SD=1.35) than for the narrator version (2.72, SD=1.39). So the version with less
text attributed to the narrator gave readers the impression that his opinion is better
Readers Closing in on Immoral Characters’ Consciousness
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
55
articulated; maybe they considered FID to make his opinion more clear. The effect
is not significant however.
Looking at our additional factor, using a median-split for self-reported reading
frequency, we did not see any interpretable interactions with our central variable,
the manipulated level of FID. However, we did find indications that this variable is
an important one to consider in this type of studies. For instance, high reading frequency readers showed less understanding for the main character (M=2.67,
SD=1.5) than low reading frequency readers (M=3.67, SD=1.6, p <.054). It
seems to us they were more susceptible to the sinister or cynical timbre of the
story. Maybe the irony of the narrator was less obvious to the low reading frequency
participants.
Analyzing the texts participants underlined we hoped to determine whether they
were aware of the role of FID in representing characters’ thoughts and feelings.
Analyses of the frequencies with which particular sentences were underlined demonstrate readers were probably not consciously aware of the manipulation of FID as
the determining factor in increasing their perception of character consciousness.
The words that were underlined most often, in all of the four versions of the
story, were mainly emotion words (e. g. ›contented‹, ›bitter‹, ›somber‹) and phrases
characterizing the officer (e. g. ›simple boy‹, ›kind face‹). There were a few sentences
that were put in free indirect discourse which were underlined more frequently,
namely two sentences pertaining to the officer explaining his ambition to set his
record, but these differences were small and were only found in the version without
the human lampshades. In the negative valence version, that is with human lampshades, no remarkable differences between the narrator-version and the FID-version were found. It can thus be concluded that FID exerts an unconscious influence
of increasing the awareness of the character’s feelings and thoughts.
4. Conclusion and Discussion
The four studies reported in the present article pertained to the alleged effects of free
indirect discourse (FID) on readers’ responses. We focused on aspects assumed relevant in theories concerning this stylistic device: readers’ perceptions of the visibility
of an (immoral) character’s consciousness, readers’ empathy, sympathy, and understanding for this character, and readers’ perceptions of the author’s attitude towards
the character as well as the clarity of the moral of the story. Rewriting an existing literary text, Borowski’s ›The Record‹, we increased the percentage of text represented in
FID. We found that FID increases the perceived visibility of character feelings and
thoughts. Readers of the FID version of our story had the distinct impression that
they learned more about the inner life of the character than readers of the original,
even though all participants were exposed to the same information. This particular
aspect of response was found to be independent of reading frequency, that is, the
56
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Jmeljan Hakemulder/Emy Koopman
number of hours participants report spending on reading for pleaACHTUNGREsure. On the other
hand, we did find that the illusion of having access to a character’s consciousness may
also depend on other variables; independent of the version read we found that participants of Study 1 (students of theatre, film, and television studies) reported better
visibility than those of Study 2 (students of literature).
While some of our results suggest that reader variables are more important than
text factors, we maintain that these are likely to interact. For instance, it was found
that FID caused confusion about the attitude of the narrator among the film students
of Study 1, whereas it did not make a difference to the literature students of Study 2.
Moreover, in post hoc t-tests of the data of Study 4 we found that the moral of the text
was clearer to participants who reported high reading frequencies than to those with
lower reading frequencies. But this only held for the FID versions. The results do not
clarify the mechanisms of these interactions between text and reader, but it seems
probable that they do play a role.
Still, an important conclusion of our four experiments is that the effects of FID
did not bear as far as theorists suggest (cf. Ludwig/Faulstich 1985 and Andringa
1996). While the general effect on visibility of characters’ consciousness appears robust, we did not register consistent effects on character, narrator, implied author perception, nor on perceived transparency of the moral of the story.
Rather than asking the all-encompassing question ›Does style matter?‹, we posed
an even more complex question: ›To whom does style matter?‹. An answer obviously
requires much more experimentation. The current contribution looked at one aspect
of the recipient, their self-reported reading frequency, but it may be that we need
more reliable and specific assessment of reader variables than is provided in participants’ self-reported reading for pleasure. The first candidate, according to us, is the
Literary Response Questionnaire (cf. Sopčk, Kuiken, and Miall (in press)). It is important to note that, although students of film may be less experienced in processing
literary texts than the students of literature, conclusions are necessarily restricted to
responses of relatively skilled readers. Moreover, we cannot be sure about the role that
literary competence plays. The lack of strong differences between reactions of film
and literature students may suggest that this factor is irrelevant, but since it was
not actually assessed we need to be cautious in our conclusions.
An additional problem, not addressed in the present studies, is the question ›At
what particular moment does style matter?‹. It may be that it is rather during the reading of a text (›online‹) that readers’ reactions are directed by textual qualities such as
FID, and that these fade away easily, leaving no trace in responses to simple pen-andpencil questionnaires afterwards (›offline‹). Future research may attempt online assessment of what may well be dominantly subconscious reactions to FID.
To increase the relevance of empirical work to narratology it is essential to be particular about the manipulations. In stories written for research purposes, textual conditions can be easier kept exactly the same, making it possible to vary the presence of
FID exclusively. Our decision to use a literary text rather than such so-called experi-
Readers Closing in on Immoral Characters’ Consciousness
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
57
ment-generated texts has important implications for our conclusions. Variations in a
simple story are easier to control; restricting changes in literary texts to just one aspect
is much harder, if not impossible. Whereas the level of control in experimenter-generated texts allows for more precise testing of predictions, conclusions might be less
relevant to our understanding of response to literary texts. In other words, the lack of
control goes hand in hand with higher ecological validity. However, it may also be one
of the reasons we found very few effects.
Where does this bring us? Did we falsify the FID Hypothesis? It is tempting to
seek fault in the experimental design (e. g., the texts used, or the exact choices in rephrasing), rather than in the theory (see Duhem-Quine thesis, Quine 1951). On the
other hand, considering the progress we made between the first two and last studies, it
seems reasonable to suggest that this type of study needs to be carried out with some
patience, allowing for time to develop instruments and procedures. In fact, comparing the four studies, we can conclude that to register an effect of FID may require a
critical amount of text represented in FID. Some would say we confuse claims about
›effect‹ and ›function‹ (cf. Schram 1985), where the former pertains to influences of
clearly defined text qualities on responses of readers, and the latter to expert interpretation of those qualities resulting from a close reading of the text – one that is hardly
expected to follow from a general audience reading a text only once. In that sense
narratology does not differ from the study of visual arts or musicology. Also, we
are not sure that in this particular discourse talking about ›the readers‹ should be interpreted as a reference to actual readers. Rimmon-Kenan, for instance, clearly states
that she wants to »present a system governing all fictional narratives« and »indicate the
way in which individual narratives can be studied as unique realizations of the general
system« (1983, 4). With her we believe that the study of poetics does not necessarily
need to be concerned with actual readers. Similarly, analyses of the art of contrapuntal
composition do not need to concern listeners’ ability to distinguish different fugue
forms. However, when claims pertain to the effects of a descending chromatic scale on
listeners’ emotions, it seems proper to test such assumptions; not to falsify interpretations of musical scores, but to add to our understanding of music. Since the present
study illustrates how reader responses may show few traces of what narrative analysis
suggests a text can do, it seems that narratology and psychonarratology are from two
cultures. However, if we want to consider narration as communication things change:
then poetics functions as a rich and irreplaceable source for hypotheses, hypotheses
about how actual addressers’ (real authors) choices in the art of storytelling affect actual addressees (real readers).
58
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Jmeljan Hakemulder/Emy Koopman
Jmeljan Hakemulder
Institute for Media and Culture Studies, Department of Media Theory
Utrecht University
Emy Koopman
Department of Literary Studies
Utrecht University
The authors would like to thank Peter Goldie, Paul Sopčk, as well as the anonymous
reviewers of this journal for their insightful comments on an earlier version of this article.
References
Andringa, Els, Perspektivierung und Perspektivenbernahme. Zur Wahrnehmung literarischer Figuren, SPIEL 5 (1986), 135–146.
–, Effects of ›Narrative Distance‹ on Readers’ Emotional Involvement and Response, Poetics 23
(1996), 431–452.
Borowski, Tadeusz, Het Record, in: Arnon Grunberg (ed.), Angst Overwint Alles: De Beste
Verhalen uit het Nieuwe Europa, Amsterdam 2005, 38–41.
Bortolussi, Marisa/Peter Dixon, Psychonarratology: Foundations for the Empirical Study of Literary Response, Cambridge 2003.
Dixon, Peter/Marisa Bortolussi/Lesley C. Twilley/Alice Leung, Literary Processing and Interpretation: Towards Empirical Foundations, Poetics 22 (1993), 5–33.
Fludernik, Monika, The Fictions of Language and the Languages of Fiction: The Linguistic Representation of Speech and Consciousness, London 1993.
Leech, Geoffrey/Mick Short, Style in Fiction: A Linguistic Introduction to English Fictional
Prose, London 1981.
Ludwig, H.-W./W. Faulstich, Erzhlperspektive empirisch. Untersuchungen zur Rezeptionsrelevanz narrativer Strukturen, Tbingen 1985.
Pascal, Roy, The Dual Voice: Free Indirect Speech and its Functioning in the Nineteenth-Century
European Novel, Manchester 1977.
Peer, Willie Van/Henk Pander Maat, Perspectivation and Sympathy. Effects of Narrative Point
of View, in: R.J. Kreuz/M.-S. MacNeally (eds.), Empirical Approaches to Literature and
Aesthetics, New York 1996, 143–156.
–, Narrative Perspective and the Interpretation of Characters’ Motives, Language and Literature 10:3 (2001), 229–241.
Quine, Willard, Two Dogmas of Empiricism, in: Williard Quine, From a Logical Point of
View, New York 1951.
Rimmon-Kenan, Shlomith, Narrative Fiction, New York 1983.
Ross, L., The Intuitive Psychologist and his Shortcomings: Distortions in the Attribution
Process, in: L. Berkowitz (ed.), Advances in Experimental Social Psychology, New York
1977, 173–220.
Readers Closing in on Immoral Characters’ Consciousness
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
59
Schram, Dick, Norm en Normdoorbreking, Amsterdam 1985.
Semino, Elena/Michael H. Short, Corpus Stylistics: Speech, Writing and Thought Presentation in
a Corpus of English Writing, London 2004.
Sopčk, Paul/Don Kuiken/David S. Miall, Beyond ›Voices‹: Empirical Studies of Free Indirect
Discourse in Middlemarch, [in press].
Walravens, F., ‘T is net of je het zelf beleeft. Identificatieprocede’s in jeugdliteratuur, Levende
Talen, 1980, 178–191.
Zyngier, Sonia/Van Peer, Willie/Jmeljan Hakemulder, Complexity and Foregrounding: In
the Eye of the Beholder? Poetics Today 28:4 (2007), 653–682.
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Jmeljan Hakemulder/Emy Koopman
Appendix
Table 1. Participants were asked to indicate to what extent they agreed with sets of two opposing
statements using a six-point scale. In the following tables higher scores indicate more agreement
with the second statement, on the right-hand side. The lower the scores, the higher participants
agree with the first statement.
1. Visibility moral The moral of the story is very
The moral of the story is overly clear.
of the story
unclear.
The story emphasizes how
2. Internal
The story emphasizes the inclination
circumstances bring people to
attribution
of people toward immoral behavior.
behave immorally.
3. External
The officer is represented as a
The officer is represented as a victim
attribution
disturbed person.
of the war.
4. Judgment of
It felt just right to me that the
I felt sorry for the officer when his
character
officer’s record was broken
record was broken.
The author tries to increase
5. Attitude implied The author rejects immoral
understanding for immoral aspects
author (1)
aspects in human beings.
of human beings.
6. Attitude implied The author emphasizes the naivety The author emphasizes the malice of
author (2)
of the officer.
the officer.
The officer seemed very
The officer seemed very pleasant to
7. Sympathy
unpleasant to me.
me.
I felt no understanding for the
I felt understanding for the ambition
8. Understanding
ambition of the officer.
of the officer.
9. Visibility
It was unclear what the opinion of The opinion of the narrator in this
narrator’s stance
the narrator was.
story was very clear.
10. Visibility
The feelings and thoughts of the The feelings and thoughts were
character thoughts
character were invisible.
clearly visible.
and feelings
Readers Closing in on Immoral Characters’ Consciousness
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
61
Table 2. Results of Studies 1 and 2. Means marked with an asterisk differ at the .05 level of
significance.
Study 1 (Film Students) Study 2 (Literature
Students)
Items
Narrator
FID
Narrator FID
M
SD M
SD M SD M SD
1. Visibility moral of the story
3.1 (1.3) 2.9 (1.4) 2.9 (1.2) 3.2 1.1
2. Internal attribution
3.0 (1.3) 3.4 (1.2) 3.7 (1.5) 3.6 1.1
3. External attribution
3.0 (1.1) 3.1 (1.4) 2.3 (.98) 2.7 1.3
4. Judgment of character
3.6 (1.2) 3.7 (.95) 3.4 (1.2) 3.6 1.1
5. Attitude implied author (1)
3.8 (1.0) 4.2 (1.0) 3.6 (1.1) 3.6 1.1
6. Attitude implied author (2)
3.9 (1.2) 3.9 (1.2) 4.0 (1.3) 4.4 1.1
7. Sympathy
3.7 (1.1) 3.8 (1.4) 3.3 (1.4) 3.9 1.1
8. Understanding
2.7 (1.4) 2.6 (1.4) 2.9 (1.4) 3.0 1.6
9. Visibility narrator’s stance
3.1* (1.2) 2.4* (1.1) 2.4 (1.1) 2.8 1.3
10. Visibility character thoughts and feelings 3.4 (1.3) 3.4 (1.2) 2.9 (1.2) 3.0 1.3
Table 3. Results of Study 3.
Morally Objectionable
Mildly (»no lamp
shades«)
Items
Narrator FID
M SD M SD
1. Visibility moral of the story
3.4 (1.4) 3.6 (1.5)
2. Internal attribution
3.2 (1.3) 3.1 (1.1)
3. External attribution
2.6 (1.2) 2.8 (1.0)
4. Judgment of character
3.4 (1.4) 2.9 (1.2)
5. Attitude implied author (1)
3.5 (1.2) 3.3 (1.2)
6. Attitude implied author (2)
3.9 (1.3) 4.2 (1.1)
7. Sympathy
3.9 (1.4) 3.6 (1.1)
8. Understanding
3.2 (1.7) 2.6 (1.2)
9. Visibility narrator’s stance
2.8 (1.5) 2.8 (1.2)
10. Visibility character thoughts and feelings 3.5 (1.1) 3.5 (1.4)
Highly (»with lamp
shades«)
Narrator
FID
M
SD M
3.7 (1.4) 3.5
3.4 (1.2) 3.5
2.7 (1.1) 3.0
3.0 (1.4) 3.0
3.5 (1.1) 3.7
4.2 (1.5) 3.9
3.8 (1.2) 3.8
2.9 (1.4) 3.0
2.7 (1.2) 2.7
2.7* (1.1) 4.2*
SD
1.1
1.4
.92
1.3
1.2
1.2
1.3
1.5
1.3
1.3
62
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Jmeljan Hakemulder/Emy Koopman
Table 4. Results of Study 4.
Items
1. Visibility moral of the story
2. Internal attribution
3. External attribution
4. Judgment of character
5. Attitude implied author (1)
6. Attitude implied author (2)
7. Sympathy
8. Understanding
9. Visibility narrator’s stance
10. Visibility character thoughts and feelings
Morally Objectionable
Mildly (»no lamp
shades«)
Narrator
FID
M SD M SD
3.8 (1.3) 3.7 (1.3)
3.7* (1.4) 3.2* (1.3)
2.5 (1.2) 2.6 (1.1)
3.6 (1.3) 3.1 (1.6)
3.9 (1.3) 3.5 (1.2)
2.9 (1.2) 2.8 (1.4)
3.9 (1.4) 3.9 (1.5)
3.0 (1.6) 2.9 (1.7)
2.8 (1.4) 3.1 (1.4)
2.8* (1.3) 4.1* (1.2)
Highly (»with lamp
shades«)
Narrator
FID
M SD M SD
3.5 (1.5) 3.6 1.3
4.2* (1.3) 3.5* 1.5
2.8 (1.2) 3.1 1.3
3.4 (1.5) 3.7 1.6
3.7 (1.2) 3.8 1.1
3.2 (1.6) 3.1 1.5
3.5 (1.3) 3.9 1.3
2.5 (1.5) 3.3 1.4
2.7 (1.3) 3.0 1.3
2.7* (1.3) 3.8* 1.1