When we read stories, we are transported into magical worlds. That magical world may be another person’s life, a point in the distant past, or it may be a world full of magic and wizardry. One of the most remarkable things we do when we read is to imagine things we cannot ever see or experience, such as fairies, quidditch games, or super heroes.
On top of that, the worlds in books are all unique – Harry Potter wizards differ greatly from wizards in Lord of the Rings, who differ from Merlin. Reading an “impossible” story therefore requires creating a representation of that world. One open question in language comprehension research is how we can accommodate such tall tales, and what we expect to encounter when we read them.
In the fantastic stories we read and watch, almost anything can happen. So how can we handle and comprehend those worlds?
We have rich representations of characters from stories we know, the worlds they inhabit, as well as a wealth of knowledge about how the world functions and what sorts of things can be done in reality. Readers use their knowledge about the real world when reading “real” stories, and in consequence they are slower to read a story about a vegetarian eating a cheeseburger.
How much of daily life can seep into a story about super heroes (like the Incredibles), and how much magic can normal people experience (as in One Hundred Years of Solitude)?
When it comes to a very different fictional world, readers must constantly integrate new information into their memory of the story. How do readers know what to consider a plausible event within a magical world? Harry Potter riding his broomstick may appear quite plausible, but what if he were kept afloat by rolls of magic paper towels?
In a recent study published in the Psychonomic Society’s journal Memory and Cognition, researchers Foy and Gerrig tested predictions from norm theory to test readers’ expectations of what realistic and fantastic characters in the same world are able to do.
When we read about a fictional world, we may use information about that character specifically (e.g. only Peter Pan can fly). This is a character-based norm, which predicts that we do not necessarily generalize and apply magical properties to new characters in the story. Alternatively, readers may just rely on the overall similarity between characters, expecting the Lost Boys or Captain Hook to fly, because their knowledge of Peter Pan’s abilities has been activated, but not that of Wendy’s father outside Neverland.
In all of the experiments by Foy and Gerrig, the materials were stories that were entirely the same, other than setting up the context (such as Neverland or the real world) and the characters (Peter and Wendy or Bobby and Judy). The final sentences were always the same for all the versions of the story, and analyses were on the length of time it took to read that final sentence.
In Experiment 1, the participants read stories about ordinary and fantastic characters doing ordinary or fantastic things. Characters doing actions that were consistent with their backgrounds (ordinary people doing ordinary things and fantastic people doing fantastic things) should be easiest to read—and this is precisely what was observed. By contrast, participants were slowest to read about realistic characters doing fantastic things. Apparently, it’s not easy to swallow that your neighbor Bob just floated past on a broomstick.
When reading, however, does comprehension really stop with the characters? Surely when we read about wizarding school, we assume that Harry Potter is not the only student? And we would expect the other students to be similarly talented?
To explore the role of similarity-based inferences, Foy and Gerrig conducted another experiment. The authors constructed stories with a familiar fantastic character (e.g., Harry Potter), an unfamiliar but similar fantastic character (e.g., an unfamiliar student from Hogwarts), and an unfamiliar ordinary character (e.g., Hermione’s muggle neighbor). The characters could either do ordinary or fantastic things.
A similarity-based theory predicts that familiar and unfamiliar fantastic characters’ stories should be just as easy to read for both fantastic and realistic events. In confirmation, participants were found to be just as fast to read about the familiar fantastic character doing fantastic things as they were to read about the unfamiliar fantastic character doing fantastic things. Likewise, familiarity of character did not matter when the fantastic characters were doing realistic things: In both cases, readers were equally slow.
Maybe that is why the Incredibles retiring and settling down in suburbia is such a jarring and unbelievable thought.
In a further exploration of why readers were unaffected by the characters’ familiarity, Foy and Gerrig gathered ratings of the two characters’ similarity and offline judgments of how much properties like the ability to fly would apply to the new character. The rated similarity of the characters predicted how much people generalized that property or ability. The similarity between characters also predicted how easy it was to read about the unfamiliar fantastic character.
One surprising result is that unfamiliar fantastic stories were just as easy to read as familiar fantastic stories. Why might this be the case? If readers don’t know anything about a character, and have to build up a discourse, how can we do it so quickly?
The authors propose that inferring that two characters are similar activates representations of the familiar character, thereby facilitating processing of the new, unfamiliar character. One way we might make these inferences is when characters are in the same family, became magical in the same way, or stand in some other causal relationship.
Foy and Gerrig’s work shows that we make powerful inferences when we read stories, and some of those inferences come online in as little time as it takes to say “expelliarmus”.