That is a scientifically grounded claim.
Cognitive psychologist Dan Johnson and I make a version of it in our paper “The Genre Effect: A Science Fiction (vs. Realism) Manipulation Decreases Inference Effort, Reading Comprehension, and Perceptions of Literary Merit,” forthcoming from Scientific Study of Literature.
Dan and I are both professors at Washington and Lee University, and our collaboration grew out of my annoyance at another study, “Reading Literary Fiction Improves Theory of Mind,” published in Science in 2013. Boiled down, the authors claimed reading literary fiction makes you smart. And, who knows, maybe it does, but if so, their study gets no closer to understanding why–or even what anyone means by the term “literary fiction” as opposed to, say, “science fiction.”
Our study defines those terms, creates two texts that differ accordingly, and then studies how readers respond to them. The results surprised us. Readers read science fiction badly. If you’d like all the details why, head over to Scientific Study of Literature. Meanwhile, here’s a preview, beginning with the study set-up:
Rather than selecting different texts based on expert but unquantifiable impressions of literariness or nonliterariness, our study uses a single, short text that we manipulate to produce isolated and controlled differences. The text is less than a thousand words and depicts a main character entering a public eating area and interacting with acquaintances including a server after his negative opinion of the community has been made public. We designate the first version as “Narrative Realism,” the common designation for literary fiction that takes place in a contemporary setting but does not fit another subgenre, such as romance or mystery, that also takes place in contemporary settings. In the narrative realism version, the main character enters a diner after his letter to the editor has been published in the town newspaper. Rather than attempting to study multiple subgenres, we select one, designating the second version “Science Fiction,” the most common term for fiction that includes such “accessible, well known features” as “interplanetary travel and aliens,” “hypothetical advances in technology and science,” and being “set in the future” (Dixon & Bortolussi, 2005, p. 15). In the Science Fiction version, the main character enters a galley in a distant space station populated by humans, aliens, and androids. The Narrative Realism and Science Fiction versions are identical except for setting-creating words, such as “door” and “airlock.” Both versions, therefore, should promote identical levels of theory of mind, requiring a reader to draw inferences about the main character’s and other characters’ unstated thoughts and feelings.
Because Kidd and Castano (2013) identified theory of mind as the distinguishing quality of literary fiction, we also created two versions of Narrative Realism and Science Fiction. The first version of each included statements that directly state a character’s thoughts and feelings, for example, “Jim knows everyone in the diner will be angry at him.” The second version of each includes no theory of mind explaining statements. The versions of the texts that include theory of mind explaining statements should have lower theory of mind demands than the versions that do not include them, because the explanations state the inferences the text would otherwise only imply. If theory of mind is the defining quality of literary fiction, then the texts with theory of mind explanations would be comparatively nonliterary.
In addition to theory of mind, we address an additional form of inference, which we call theory of world. Where theory of mind requires the inference and representation of a character’s implicit thoughts and emotions, theory of world requires the inference and representation of a world’s implicit laws and systems, potentially including such things as laws of physics, systems of social organization, and public history. Both texts, for example, include a sentence that begins: “He was awake in his bunk just a few hours ago, staring at …”; the narrative realism version then continues: “… the shadows of his ceiling slowly ebbing to pink, when the delivery kid’s bicycle rattled onto the gravel of his driveway,” while the science fiction version continues: “… the gray of his sky-replicating ceiling slowly ebbing to pink, when the satellite dish mounted above his quarters started grinding into position to receive the day’s messages relayed from Earth.” Although theory of world would be present in both Narrative Realism and Science Fiction, because Narrative Realism’s world is a representation of the reader’s world, theory of world demands are minimal. Because science fiction often depicts worlds that differ significantly from a reader’s world, theory of world demands would be higher. The narrative-realism text then should promote theory of mind but not theory of world, and the science-fiction text should promote theory of mind to the same degree as the narrative-realism text and theory of world to a greater degree.
Such an understanding, however, treats both theory of mind and theory of world as intrinsic qualities, while ignoring the role of extrinsic influences. The term “narrative realism” is sometimes conflated with the term “literary fiction” because narrative realism is a genre distinct from “genre fiction” and exists only in the sometimes mislabeled category of “literary fiction.” But if literary fiction is defined by theory of mind, a story’s setting, whether realistic or fantastical, indicates nothing about its literariness. However, while neither narrative realism nor science fiction then are more likely to be literary in terms of intrinsic qualities, we hypothesize that the narrative-realism text is more likely to be extrinsically identified as literary and that the science-fiction text is more likely to be extrinsically identified as nonliterary. Because theory of world is more prevalent in science fiction than in narrative realism, the promotion of theory of world processing is also more likely to be extrinsically identified as nonliterary.
And here are some of our results:
Addressing the effects of genre first, in comparison to Narrative Realism readers, Science Fiction readers reported lower transportation, experience taking, and empathy. Science Fiction readers also reported exerting greater effort to understand the world of the story, but less effort to understand the minds of the characters. Science Fiction readers scored lower in comprehension, generally and in the subcategories of theory of mind, world, and plot. The last finding is striking because Science Fiction readers reported exerting the same level of effort for understanding plot as Narrative Realism readers, but their actual comprehension of plot was weaker. Science Fiction readers reported exerting a lower level of effort for understanding theory of mind than Narrative Realism readers, and scored comparatively lower in theory of mind comprehension. Science Fiction readers even scored lower in theory of world comprehension, the one area they reported higher inference effort than for Narrative Realism readers.
Comparatively higher theory of world effort and lower theory of world comprehension, however, should be expected because a narratively realistic setting is understood to be a representation of the reader’s own world, allowing high comprehension with little effort. The science-fiction setting demanded far more inference and so greater effort to achieve comprehension. As discussed, we hypothesized this difference in theory of world to be a defining difference between science fiction and narrative realism.
The Science Fiction’s lower plot and theory of mind scores, however, are not a result of intrinsic qualities, unless the theory of world features influenced theory of mind processes. Because the science-fiction and narrative-realism texts differ according to theory of world but are essentially identical for plot and theory of mind, effort reports and comprehension of plot and theory of mind should be statistically the same. Therefore we conclude that the difference is a product of the readers’ prior social constructs regarding texts like Science Fiction and Narrative Realism. Since science fiction is “characterized as being focused on settings and content, with comparatively less emphasis on interpersonal relationships” (Fong, Mullin, & Mar, 2013, p. 371), that expectation may produce an assumption of nonliterariness for readers who also experience theory of mind-promotion as a primary quality of literariness. Science fiction story details would therefore produce a lower perception of literary quality. Based on their low theory of mind effort scores, the Science Fiction readers expected a story that involved less theory of mind. This expectation, or a subsequent exertion of less theory of mind effort, would also account for the low theory of mind comprehension. Though readers were neutral regarding plot effort, lower plot comprehension suggests a generally lower exertion in reading effort. The Science Fiction readers appear to have expected an overall simpler story to comprehend, an expectation that overrode the actual qualities of the story itself. The science fiction setting triggered poorer overall reading.
And if you’d like to see the actual texts we used in the experiment, they’re now here.