Perceived realism of fictional dialogues and every-day conversations

[lillamagyari]

Speaker: Lilla Magyari @lillamagyari

Affiliation: University of Stavanger

Co-authors: Frank Hakemulder (Utrecht University), Anne Mangen (University of Stavanger)

Title: Perceived realism of fictional dialogues and every-day conversations

Abstract: Direct speech in fiction is often embedded in dialogues, i.e. in written conversational exchanges between characters. These are assumed to be a means of realistic depiction of the fictional world. However, fictional dialogues also differ in many ways from real speech and from every-day conversations. In an experimental study, we examined whether readers fall into the illusion of experiencing fictional dialogues as realistic, as if those could also have happened in real life. According to our preliminary data-analysis, adults can differentiate transcripts of real-life conversations and fictional dialogues. However, adolescents might perceive fictional dialogues to be more realistic compared to adults.

Long abstract

Introduction
Direct speech in fiction is often embedded in dialogues, i.e. in written conversational exchanges between characters. Dialogues and direct speech are often assumed to be a means of realistic depiction of the fictional world(1). However, fictional dialogues also differ in many ways from real speech and from every-day conversations. In an experimental study, we examined whether readers fall into the illusion of experiencing fictional dialogues as realistic, as if those could also have happened in real life.
Background
Direct speech might be essential for the authenticity and believability of the characters(2) and it is an effective way to enliven a narrative(1). However, direct speech in novels can be also highly schematized, stylized and might reflect literary conventions than actual speaker’s speech behaviour(3).
Dialogues are more than representation of direct speech of characters because they mimic conversations between characters in the fictional world. Dialogues might be also realistic in the sense that those could reflect the “rules” of spontaneous real-world conversation(4). Nonetheless, there might also be many differences between fictional dialogues and every-day conversations. For example, real-life conversations are often about mundane topics, while dialogues in fiction are part of a story which is a “tellable” material(4).
To study whether readers can differentiate fictional dialogues from every-day real-life conversations, we conducted and online, quasi-experimental questionnaire. Participants’ read a few lines-long excerpts of dialogues from classical realistic novels of the XIXth century and transcripts of real-life conversations. Participants were told that all excerpts are from fiction and were asked to judge how similar these excerpts are to real-life conversations.
Hypothesis
Our hypothesis was that participants might be unaware of the differences between real-life conversations and fictional dialogues, hence, there would be no differences in participants’ judgments between excerpts of fictional dialogues and real-life conversations.
Methods
366 participants filled out our online questionnaire in which they were presented with excerpts from every-day conversational transcripts and excerpts from fictional dialogues. Participants were told that all excerpts are fictional. They rated on a scale between 1 and 5 how similar the excerpts were to real-life conversations. The fictional dialogues were selected from five modern Norwegian translation of British and Russian XIXth century novels (Jane Eyre by Charlotte Brontë, Middlemarch by George Eliot, Oliver Twist by Charles Dickens, Crime and Punishment by Dostojevsky and Anna Karenina by Lev Tolstoj). The main selection criteria of the stimuli were based on formal requirements, such length (i.e., number of turns and number of words). Additionally, excerpts of dialogues were excluded if their content contained references to objects, services (e.g., transportation by horse riding), historical events, social norms or patterns of behaviour which could explicitly relate the text to the XIXth century. Real-life conversation excerpts were selected from the transcripts of the NoTa-Oslo spoken language corpus. The transcripts were slightly modified to reduce the formal differences between the written dialogues and the orthographic transcription of the live speech (i.e., punctuations were inserted into the transcripts). Conversation excerpts were matched in number of turns and (approximately) in number of words to excerpts of the fictional dialogues. 80 fictional dialogues and 80 real-life conversation excerpts were selected and were divided into 4 experimental lists. Hence, 20 fictional dialogues with their matched pair of conversational excerpts were included in each list.
Results
The data collection is still in progress. In our preliminary analysis, we analysed the data of 200 healthy participants speaking Norwegian as their first (native) language. The distribution of age was very unequal across different ages, therefore we created four age-categories: 16-years (N=15), 17-years (N=48), 18-years (N=122) and adults (N=15). For statistical analysis, we used linear mixed-effect modelling with the similarity judgements as a response variable and participants and excerpt-pairs (i.e., fictional and real-life excerpts matched in length) were random factors. There was an interaction between excerpt type (i.e., fictional or real-life) and age-category. Participants judged the real-life excerpts more similar to real-life conversation compared to the fictional excerpts in all age group, except for the 16 year-olds where there was no difference in judgments. The difference in the judgements between real-life and fictional excerpts increased across age groups.
Conclusions
The preliminary data-analysis suggest that adults can differentiate transcripts of real-life conversations and fictional dialogues. However, there might be a developmental trend in how aware readers could be about the differences between fictional dialogues and real-life conversations. Adolescents might perceive fictional dialogues to be more realistic compared to adults.

References:

1. Thomas, B. Fictional Dialogue: Speech and Conversation in the Modern and Postmodern Novel. (UNP - Nebraska, Lincoln, UNITED STATES, 2012).
2. Page, N. Speech in the English Novel. (MacMillan, 1988).
3. McHale, Brian: “Speech Representation”. In: Hühn, Peter et al. (eds.): the living handbook of narratology. Hamburg: Hamburg University. URL = http://www.lhn.uni-hamburg.de/article/speech-representation
4. Toolan, M. Analysing fictional dialogue. Lang. Commun. 5, 193–206 (1985).

[Danial (Tohidi)]

Thank you so much for sharing your interesting work, dear Lilla. It’s really insightful, and I’m looking forward to hearing about your validation of the new ART version, which seems designed for adolescents.

During your presentation, when you mentioned your larger project, I started thinking about how your project might relate to Iser’s concept of the “fictionalizing act,” tied to reader-response theory and literary anthropology. I’m not entirely sure how closely it aligns with your current presentation, but I wanted to share the idea and get your thoughts. Iser argues that fiction transforms elements of reality into a fictive construct through what he calls the “fictionalizing act.” To me, this seems relevant to your project, particularly in how “fictional dialogue” differs from “real-world conversation.” Fictional dialogue draws on real-world conversational patterns but reshapes them to serve the story’s purpose.

In my view, fictional dialogue is a product of the fictionalizing act, turning real-world conversational elements into a fictive construct. Real-world dialogue, on the other hand, belongs to the “reality” in Iser’s framework, bound by practical constraints like communication, social norms, etc. and lacks the intentional fictionality of literary dialogue. Plus, fictional dialogue often highlights its constructed nature. I should also note that, in Iser’s model, all of this is discussed in relation to the imagination, not whether readers perceive it as real or not. -Dan

[lillamagyari]

Dear Dan, thank you for drawing my attention to Iser’s model and thank you for your comment which raises many thoughts. If I understand well, Iser’s concept of “fictionalzing act” can be understood so that the purpose of fictional dialogues is related to the narrative, and not to the direct representation of reality. And if dialogues are constructed for fiction (or for narratives in general, I do not know how much “fiction” is a general term here), it can not represent reality. I agree with this. But my question is how much the reader is aware of the constructed nature of dialogues? Such awareness of the constructed nature of fictional dialogues is especially, interesting in the light of more psychology-driven models, like e.g., Mar& Oatley (2008) which emphasise that narrative fiction creates a simulative experience of the social word for the readers. Hence, I think it is interesting to ask (although I have not directly studied this question here) if we are more aware that fictional dialogues are very different from real-life interactions, do we immerse less likely, or in terms of Iser’s model: do we use our imagination less? Best, Lilla

[fpianz (Federico Pianzola)]

Following up to the question I asked live: this spreadsheet can be useful to identify when perceived realism is considered a subdimension of narrative immersion (and related constructs).
https://osf.io/rbz8g/files/osfstorage

The analysis of several questionnaires is here:
Pianzola F. Presence, flow, and narrative absorption questionnaires: a scoping review [version 2; peer review: 2 approved, 1 approved with reservations]. Open Res Europe 2021, 1:11 (https://doi.org/10.12688/openreseurope.13277.2)

[lillamagyari]

Thank you very much for the links to the study! And yes, I can see that your analysis shows that many items measure spatial presence by comparison to reality. But the order of direction can be both ways, right? Immersion occurs due to higher perceived similarity to reality or the higher perceived similarity can create immersion. Or was your question about whether it makes sense to differentiate immersion from perceived realism, and aren’t those the same notions (constructs…etc.)? I think when we ask questions about 1) “whether something you read is similar to what could happen in real life” (how I write this sentence here, is similar to Cho et al.'s notion of perceived realism) is a different question then asking that 2) when you were reading/watching video/gaming did it feel like real? The first I think taps into a more conscious judgement about the difference between reality and fiction, while the later is more about the experience which relates to immersion. In my talk, I think I refer to perceived realism in this first sense, although I can see that the two can be related (i.e. if I perceive something very real while I am reading about it, I will also think that it could happen like that in real life), and it is a very good question how these two notion of realism relate to each other. In the experiment itself, the judgements are probably very little influenced by “how much the participants felt it being real while reading it” because the excerpts were just a few lines and were taken out of their context. Moreover, I could argue that if fictional dialogues were created to provide an immersive experience (while transcripts of conversations were not constructed for this), I would expect that if any immersion happened that would be present more for the fictional dialogues. Hence, we could actually measure higher “reality judgement” for the fictional compared to the real-life excerpts. But the results are the other-way around (i.e. conversations were judged to be more realistic). Therefore, I think that the judgements in this study rather reflect awareness of differences between fictional dialogues and real-life verbal interactions. Best, Lilla

[fpianz (Federico Pianzola)]

thanks for the clarification!
I think you have a good design but it can become tricky if you plan to measure immersion in future research, precisely for the reasons that you also mention: the direction of the reaction between perceived realism (and how it’s measured) and immersion.