(Spoiler: it has to do with belief and imagining)

A post by Neil Van Leeuwen.

Most of us don’t like having a plot spoiled. We hate spoilers. For example, if you and I were going to see Hamlet (suppose you hadn’t seen or read it yet), and if I told you before the final act (spoiler alert!) that Hamlet dies, you would likely be annoyed. But why? You were going to learn what happens anyway. Why get so worked up?

I think that in order to explain why so many people hate spoilers, we (as theorists) have to appeal both to people’s beliefs about what happens in stories and to their vivid and rich imaginings. This position, if I am right, shows that a (more or less) tacit dichotomy that’s been running through recent posts here on The Junkyard is false.

Here’s my explanation for why we hate spoilers:

We typically want to experience certain emotions, like suspense and surprise, when consuming fiction. Spoilers thwart having those emotions, so we hate them (or at least strongly dislike them), because they deprive us of something we want.

So far, so obvious. But here’s where it gets interesting. The way that spoilers ruin a fiction’s emotional impact tells us something crucial about which cognitive states characteristically cause the desired emotional experiences. For now, let’s focus on surprise, since that’s the most relevant emotion. We could just as well focus on suspense or others.

Spoilers are ruinous of surprise because they produce beliefs about what is going to happen in the story before we’re ready to have them. In particular, a spoiler (say, of Hamlet) produces a belief with this structure:

Bi <in Hamlet> Hamlet dies

That is, the spoiler (say, in the form of my telling you at intermission) is something that prematurely produces a belief. If you have that belief too early, then when you get to the end, there’s no surprise.

Thus, your overall profile of beliefs about a fiction that you have going into it (and over the course of consuming it) makes a big difference to the emotions you experience. Go into Hamlet without that belief, you get surprise (holding other things fixed). Go in there with that belief, you don’t get surprise.  

Furthermore, acquiring that belief when you get to the end is also crucial to surprise, since surprise in general comes with acquisition of a new belief. To see this, suppose that the actors’ head microphones (it’s a big theatre) all cut out in the last ten minutes of the play, so you couldn’t really make out what was happening. Here again, you would have no surprise; you would just be confused. The point is that you also need the belief for the surprise; just it can’t come too early. Just imagining Hamlet dies without feeling sure (that is, without forming the belief that he does) would never give you that pop of surprise. Rather, you would still just be wondering if what you imagined was how it went: merely imagining something doesn’t generally produce surprise. So at least with respect to surprise, which is crucial to unleashing many other emotions, it is true both that premature beliefs can ruin it and that beliefs at the right time are key to facilitating it. 

It follows from all this that anyone who says things like “it’s imagining rather than belief that gives us our emotional responses to fiction” is in error for having excluded a crucial cognitive factor that helps explain certain emotions. The “rather than” points to a false dichotomy (guilty ones: you know who you are).

At this point, Peter Langland-Hassan is probably feeling quite vindicated. In an earlier post, Peter advocated the position that one’s emotional responses to fiction depend on beliefs about the fiction. And he got a lot of pushback in the comments (“No, it’s not belief! It’s imagining!”). I appear to be one of few people who agrees with him about the importance of belief in this respect, and that importance is revealed by the spoiler phenomenon.

But Peter shouldn’t start gloating just yet.

Peter argued for the importance of belief by attempting to argue for the explanatory irrelevance of imagining. So as far as I can tell, he is tacitly guilty of the same false dichotomy as those who criticize him (belief or imagining, not both!).

More particularly, he is wrong to say that one’s imaginings do not play a role in eliciting emotional responses to fiction. Positing imaginings of how a fiction unfolds, I think, is also crucial to explaining people’s hatred of spoilers, as I’ll show in a moment. But first, let’s see where Peter’s argument against imagining goes wrong.

Peter writes (I add the underlining):

We can choose what to imagine. Therefore, if the affect we experience in response to a fiction depends on what we are imagining, we should be able to choose the affect we experience in response to a fiction. But we cannot choose the affect we experience in response to a fiction. Thus, by modus tollens: our affective responses to fiction do not depend on what we are imagining.

To see the flaw, let’s consider an analogous argument that has two substitutions in the underlined places (the structure of the argument is thus the same):

We can choose to consume water. Therefore, if living to 100 depends on consuming water, we should be able to choose to live to 100. But we cannot choose to live to 100. Thus, by modus tollens: living to 100 does not depend on consuming water.

Should we conclude that one can make it to 100 without hydrating? Obviously not. The second argument shows what’s wrong with the first: X can be voluntary, and Y can depend on X and not itself be voluntary, as long as Y also depends on some things that are not voluntary.[1] So Peter’s argument against imagining, as it stands, is invalid.

However, there is a positive moral. Properly modified, Peter’s argument would not rule out voluntary cognitive influences on emotional responses to fiction (like imaginings); it would rather rule in at least some involuntary ones, in order to explain the involuntariness of the emotional responses. And an obvious candidate for an involuntary input is belief.[2]

And now that they’re not ruled out, we can ask: why should we think imaginings are crucial to emotional experiences in response to fiction?

Here the phenomenon of spoilers is again telling. If Hamlet is spoiled, you form the belief too early that Hamlet dies; you form it, say, at time t. If Hamlet is not spoiled and you get a feeling of surprise at the end (and other powerful emotions), you form the belief at the right time that Hamlet dies; you form it at the end. So you get the belief either way, just at different times.

The question now is: what happens between t and the end that makes you have the amazing pop of emotion when (you finally form the belief that) Hamlet dies? Answer: you imagine a lot of stuff in rich and vivid detail. You have vivid imaginings of Hamlet’s mental state, Ophelia’s death (which happens offstage, so you have to imagine it), Claudius’s obtuseness, Yorick’s past playful manners, Laertes’s resentment and contempt, etc. And these imaginings are needed as set-up for the emotional impact to hit when we finally learn that Hamlet dies. Put the belief before those imaginings, and much of the emotional impact is spoiled (though of course not all of it, as Alon Chasid would point out).

Importantly, your imaginings of all these things are richer (have more fine-grained contents) than your beliefs. One’s beliefs about what happens in Hamlet could be represented in a longish Wikipedia entry. But imagining, say, what Ophelia was thinking right before her death will have a profound impact on your emotional experience of the play, even though the play doesn’t contain enough evidence to license beliefs with the correspondingly fine-grained contents about her thought. In short, Wikipedia-level knowledge (encoded in beliefs) never could give us the emotional impact that good fiction delivers, and that’s because good fiction prompts imaginings with richer contents than the Wikipedia-level knowledge (encoded in beliefs) could capture.

This has been a long blog, so let’s sum up. Why do we hate spoilers? They give us beliefs before we’ve imagined all the interesting things in all the personally powerful ways. So if we hate spoilers, it’s because we’re creatures whose beliefs and imaginings interact to produce emotions in response to fiction. They just have to be sequenced in the right order.

[1] Shen-yi Liao made this point quite clearly in his comments on Peter’s blog. The hydration example is mine.

[2] Luke Roelofs comes close to conceding this point in his blog, but he still seems to refuse to call the relevant cognitive states beliefs. And yet the states he describes appear to have the functional properties of belief, so I would encourage him to label them accordingly.