Django Wexler writes all manner of awesome things. Many of you might know him from his flintlock fantasy Shadow Campaigns series from Roc Books. Younger fans or fans with kids might be familiar with his Forbidden Library series for middle grade kids. Here, Django talks about tension in your novel and how killing off your main character might not be the solution to the problem of a story falling flat.
(Spoiler warning: the last paragraph of this post contains spoilers for A Game of Thrones book/season one.)
A criticism you sometimes hear about a book is that it feels "safe" or characters have "plot armor" -- that is, the readers don't feel like they could be hurt or die. This, the critic alleges, saps the tension from the book, because we know our hero is going to be fine. The thing to do (this hypothetical composite critic says) is kill a few major characters unexpectedly, so the reader understands that "anyone can die" and thus will be appropriately tense for the rest of the story.
This seems like a reasonable argument, but I feel like it's a misattribution of the cause of the problem. While it's definitely possible for a story to lack tension, the safety of the main character is generally not the reason. This kind of argument based on meta-story logic is usually an indicator of a weak story to begin with.
First, it should be clear that it's possible for a story to have tension even when the readers know that our protagonists will succeed and be fine. Most obviously, you can reread a book, and still feel the tension, even if you know with a hundred percent certainty how it will come out. You can read a book in a continuing series -- James Bond, say -- and be reasonably certain that whatever happens, Bond is not going to die, since there are fifteen more books to come. You can read a book in a genre, like most children's fiction, where some kind of happy ending is pretty much guaranteed, but in a well-written work it doesn't remove the tension. That this is possible should be obvious by the fact that books and movies have been very successful in all those situations.
I think of this as the "rollercoaster effect" -- like a rollercoaster, where there's no real danger but you get excited or scared anyway -- but it's really just a part of the suspension of disbelief. The reader suspends disbelief in the world of the narrative, buying into the premise and the characters, and at the same time at some level agrees to temporarily forget about the meta-story stuff, based on their knowledge of factors outside the story itself -- that this is book one of a trilogy, for example.
When readers start giving the critiques I mentioned -- that characters feel like they have "plot armor" -- what has really happened is that suspension of disbelief has been lost. It's not the fact that meta-story logic leads gives information about the outcome that's the problem, but rather that people are thinking in meta-story terms at all. If the readers are analyzing your story in the context of its place in a multimedia franchise, you've already lost them; they should be caught up in it, willingly going along with your characters and world. This is why adding more death, essentially another meta-story option, doesn't really help -- you may be able to convince readers that anyone can die, but you haven't fixed the suspension of disbelief problem that got them there in the first place.
So what causes suspension of disbelief to fail? Any number of things. Excessive coincidence is an obvious culprit -- almost all stories rely on random chance to some extent, but when a character pulls a lucky break too often readers may balk. Another common problem is characters losing track of their own motivations, and acting in ways that serve only to further the plot or make no sense given what's been established about them. Main characters who enjoy the obvious favoritism of the universe are disruptive, too -- protagonists should catch both good luck and bad luck and have to deal with both. In all cases, though, the result is that the "authorial hand" becomes too obvious, making it impossible for the reader to stay in the fictional world as presented and forcing them to think about the motivations of the author instead of the characters.
But, we might ask, what about the times when an author does unexpectedly kill off a main character? Those moments are often famous, so doesn't that show this technique can contribute to the tension? I would say, examine those moments more closely. To take a well-known example, George R. R. Martin famously kills off protagonist Ned Stark at the end of A Game of Thrones. But there is so much more going on here than Martin simply wanting to indicate to readers that anyone can die. Ned is the classic fantasy hero -- honorable, bold, kind-hearted. But in Martin's world, all those characteristics lead to his demise. He's not struck down by a bolt from the blue -- his death is a key part of the theme of the books, the subversion of the traditional mode of heroic fantasy.
Thus, it's not that one should never kill off major characters. There are many good reasons to do so, when it fits the plot, theme, and tone of the story. But doing so just to establish "grit", the idea that "anyone can die" in order to get the readers nervous, is not a good enough reason. If you get critiques along these lines, rather than dealing out death, probe a little deeper and try to find out where the readers lost faith in your narrative.
Django Wexler graduated from Carnegie Mellon University in Pittsburgh with degrees in creative writing and computer science, and worked for the university in artificial intelligence research. Eventually he migrated to Microsoft in Seattle, where he now lives with two cats and a teetering mountain of books. When not writing, he wrangles computers, paints tiny soldiers, and plays games of all sorts.