Structure and Style

First up, a little business. Through StoryOrigin, I’ve been asked to do another newsletter swap for Isabelle Quilty’s Siphon Break, which I reviewed a while ago.

Getting on to analytical matters… STRUCTURE.

Lawrence Block has commented on the “formula” used by Donald Westlake when (under the pseudonym Richard Stark) he wrote the Parker novels. Block, a close friend of Westlake’s for many years, notes that anybody who thinks writing formulaic fiction is easy hasn’t tried it. And in the course of developing my fantasy homage to Parker, I’ve spent a lot of time thinking about how the formula works.

A Parker novel has four parts, distinguished most obviously by point of view. Parts 1 and 4 are always from Parker’s perspective, in extremely narrow third-person. One of the others, usually Part 3, moves through numerous characters’ points of view, one per chapter, in equally narrow third-person mode. Wikipedia (incorrectly) explains: “The first and second sections are written in a limited third-person perspective focused entirely on Parker as he plans and undertakes a robbery or heist with colleagues. The second section ends on a cliffhanger, as Parker is betrayed—often injured and left for dead. Section three shifts to the perspective of Parker’s opponents, usually in flashback as they plan and execute their double-cross. Section four returns to Parker’s perspective as he survives the plot against him and sets out for revenge.”

Well, that’s one version, but there are others.

From a distance, we have the general impression that the four parts go planning, setting up, heist, aftermath, or alternatively planning, heist, double-cross, revenge. For simplicity’s sake, consider one instance when the former is followed strictly: Backflash (1998), the 18th of the series, the second after Parker’s long absence. (There are 16 novels from The Hunter [1962] to Butcher’s Moon [1974], and then Parker vanishes for twenty-three years, only to turn up again in Comeback [1997].) Roughly:

  1. (Parker): As one job ends, Parker is brought in on a new one, and tries to figure out whether it’s doable

  2. (Parker): Parker assembles his string, acquires guns and other accoutrements, and gets ready

  3. (Various): The gang performs the heist, but there are other people around who are going to be a problem

  4. (Parker): Parker deals with loose ends

Notice that the heist itself is narrated indirectly, through others’ eyes. The intended plot is already clear to us, because we’ve been through the planning. So in Part 3 we get, for example, a chapter from the perspective of George the money-counter, a guy whose job it is to sit in a locked room and tally up cash, and how his life changes when he realizes that those three official-seeming fellows are holding guns and telling him to shovel money into bags.

Importantly, George does not appear elsewhere. From a plot perspective, he’s incidental. If you were filming the heist sequence, George would barely get a line, and might even be an extra. But Stark gives him a whole chapter, in which we learn his story, how he happened to be in this place at this time, doing this job.

This raises two questions. First, why tell the heist indirectly? Second, why give George such centrality at this most crucial moment of the book?

The answer is simple: because the heist is not the most crucial moment.

Suppose you’re playing a role-playing game with friends, and you decide to undertake a full-blown heist. You research it, come up with intricate details and a timetable, and after many hours of enjoyable work embark on the job. What happens? If your game-master is good, the heist goes pretty smoothly, and you may even find it a bit of a letdown. From the GM’s perspective, if you start the job and suddenly there’s a bunch of complications you didn’t plan for, so that it becomes impossible to pull off, you’re going to feel cheated. You went to all this trouble, why didn’t your research turn up these problems? All that time you put into planning now turns out to be useless. You went from feeling clever to feeling frustrated. Since your GM wants everyone to have a good time, this is a bad result—thus the heist goes fine. If on the other hand you actually play out every detail of the job, there’s little drama to sustain the focus. You know what’s going to happen at each stage, because you planned it out and it’s not going to fail drastically, so where’s the element of surprise?

Now in a caper-heist film (Ocean’s Eleven, The Sting, etc.), the audience does not actually know the plan. We think we do, and the film tricks us into misreading detail X as meaning something other than it does. As the thing plays out, we keep thinking everything’s going horribly wrong when in fact that’s all part of the plan. When we get to the end, the magic is the revelation: “wait, they did that on purpose? But then… oh, I get it now!”

A common alternative solution, used by Stark in many Parker novels, is to have the heist go wrong. Just when it was going well, something unexpected happens to screw it up, at which point things get ugly fast.

The moral of the story is a simple but hard truth: the heist is not the crux. The heist is the situation, the setup, the frame—but it is not, in itself, all that important.

Shifting back to fantasy, this lesson has a strong implication. If the basic plotline is that our heroes go to the Lost Land of Nog to steal the Jeweled Scepter from the Black Temple, the actual going-into-the-temple-and-stealing-the-scepter bit is necessary but not important.

Without going into details, we see this in many of Howard’s Conan and Leiber’s Fafhrd-Grey Mouser stories. Sure, the point is to steal the treasure from that misshapen black temple, but what makes it interesting is that (a) there are freaky cultists trying to repel invaders, and (b) the temple is misshapen because it is the god, which now gets up on its hind legs and starts killing people.

In Backflash, as noted, the solution to the problem is to narrate the heist indirectly, through many people’s perspectives, so that our sense of the total world is considerably expanded—and we have a good sense of why Part 4 can’t just be the gang leaving with the money in a van. Furthermore, a lot of that lead-out into the aftermath/cleanup phase depends on stuff we read in the earlier phases. For example, in setting up the job, Parker had to deal with all kinds of people, and if some of them guess the job and decide to heist the heisters, then during the heist (Part 3) they’re going to be setting up for their ambush (for Part 4). If there’s enough messiness about, we could push the heist section back to Part 2 and spend 3-4 on fallout. This is how the film Rififi works; even more extreme, there’s The Seventh (1966) or Comeback (1997), in which the heist is in Part 1.

The glory of this formula for the fantasy writer is that the backstories, however incidental to the plot, allow worldbuilding exposition. POV backstories populate the world with a different sorts of people with varying perspectives, assumptions, backgrounds, and lives. If that stuff is imaginative and variable, the world is greatly expanded. Plot-wise, this section gets something crucial done—the temple does get robbed—but it achieves something much more useful.

Last time I talked about Le Guin’s “From Elfland to Poughkeepsie,” where she criticizes a lot of contemporary fantasy for its drab, workaday style. In part, the set-dressing—dragons, elves, goblins—is like the heist in Backflash: you’ve got to have it, but it’s not intrinsically important. Worldbuilding—languages, histories, governments—is like the heist-planning: we only care because it grounds us. What actually matters is people and characters. If their lives and perspectives, the way they speak and act, express all that worldbuilding and set-dressing, then we’re getting close to Elfland. If they’re just ordinary plain folks like us, only there happen to be dragons around and the king’s courtiers collect the taxes, then no amount of linguistic brilliance or alternative-physics analysis is going to keep the story out of Poughkeepsie.

My conclusion is that the Parker structure-formula is brilliant for what I’m trying to do. The chapter-length vignettes allow us to see and understand this world from a new, unexpected perspective, achieving that crucial distancing (Elfland), but they’re narratively justified because we’re telling what seems like the most important part of the plot. At the same time, I don’t have to retain these characters: I can do a chapter from a goblin’s POV, and if he dies or scurries away, I don’t have to justify his inclusion. Extras become useful.

Now that I know how to retool my novel, I’m off to do some restructuring.