This is not the right question to ask. Of course we do: how could we upset about the death of a favourite character if we did not, on some level, believe that this character was real? The right question might be: What, in fiction, do we find believable, and how do we believe it?
One of our recent projects involved working on some aspects of that issue. I will give more details after its publication, but it appeared to us that when emotional reactions are involved, people who play a character in an immersive virtual world tend to react in a very similar way to what could be expected in real life. Even when controlling an avatar in a blatantly fictional world, they acted according to a complex set of motivations, involving emotional impulses, utilitarian reasoning and moral concerns, rather than the purely tactical moves of, say, a chess player. While they never expressed confusion about the unreality of their virtual world experiences, their reactions still matched what would be expected in a real-world situation. In other words, even people who are perfectly aware that they are having a fictional experience will emotionally react to it as if it was real.
Should we say, then, that violent movies, role-playing and video games are a danger to young people, who will start to treat reality like another shooting game?
A tentative answer might be: quite probably not, but it would be equally inexact to state that fiction has no impact on us. We do believe in fiction. The only thing is, that belief is not always literal, and different people may believe in different aspects of the fictions they engage with. Sometimes belief may be so insidious it takes a considerable amount of reflection to recognise it. One aspect of this phenomenon is the perceived authority of the text: we do not think that the events displayed in a piece of fiction are real, but we believe that the context mirrors the real world. Most of us have heard the sentence, ‘It’s true, they said it on TV’, or, ‘This is how things work in America/India/the ancient Roman world, I read it in so-and-so’s (fictional) book’. Depending on the generation or social class, a different level of authority will be attributed to different media; the phenomenon, however, remains essentially the same. Television, the internet or famous novelists will be assumed to know what they are talking about, even when they are blatantly inventing facts.
Philosopher Paul Grice stated that cooperation is the basis of human communication. When we talk to each other, our default setting is helpfulness: we try to help our interlocutor understand what we mean, and in normal communication, we tend to assume that the person we talk with is giving us relevant and truthful information, explicitly or implicitly. The notion of cooperative communication might explain why people tend to take for granted that information they find in fiction is truthful and relevant to the real world: the author should know (they are in a position of authority), and therefore, they have no reason to withhold the information they know about from their audience.
Indeed, the number of situations when people seek to derive a form of knowledge from what they know to be fiction is very large. When a man gets down on his knee to propose because he wants to be romantic ‘like in the movies’, he’s deriving a form of emotional knowledge from fiction. When someone states that American universities are driven crazy by political correctness because what Philip Roth describes in The Stain (a professor sacked for unwittingly using a racial slur), they are seeking to derive factual knowledge from invented events (even though, Philip Roth being a conservative white man, his perception of political correctness might not be entirely neutral). When writer Chimamanda Ngozi Adichie states that as a child, she would write tales about white children playing in the snow, eating apples and drinking ginger beer, despite having never been outside Nigeria, she refers to a deeper level of belief, the belief that some realities (like a white or male protagonist, or a Western setting) are more representative, normal, or ‘universal’, than others (for instance, a female protagonist in an action narrative will still be considered as a political statement in many respect, regardless of the frequent reality of women serving as soldiers or police officers).
All those are very different forms of belief. It is not the same to believe that a fictional fact is real, to believe that human nature as portrayed in fiction is real, or to assign different levels of normalcy, otherness and relevance to groups or events based on fictions. It might therefore be extremely difficult to predict how influential or ‘dangerous’ a given work of fiction might be, or to assume how people will react to it. While outright behaviour modelling cannot be ruled out (examples can be found throughout history, after all, from the most innocuous fashion trends to more disturbing cases like the wave of suicides following the publication of Goethe’s The Sorrows of Young Werther), the reality could often be more subtle. Playing Call of Duty might not make teenagers dangerously aggressive. It is possible, however, that it encourages many of them to believe in a world where all conflicts boils down to a binary opposition between good and bad guys, or where facing or avoiding physical violence are the only solutions available. Watching movies in which the black member of the hero’s team invariably dies before the end may not make audiences aggressively racist, but it could reinforce the belief that black people exist only in the background and can never come to positions of prominence.
Difficult as it is to predict how individuals engage with works of fiction, our default attitude is not one of perfect, lucid distance. Understanding more about it should be a fascinating pursuit.