Occasionally engineering provides me with startling perspectives on art. This happened for me with the third book in the Harry Potter series, the Prisoner of Azkaban, generally acknowledged to be the most accomplished of the series. Critics make a compelling case that among the movies too, Alfonso Cuaron’s treatment of this book has been the best of the lot so far, and better than even the book itself. So let me offer you a perspective on why the third book and movie are great, based on an analogy to a problem in robotics.
(This is the third part in my series on the Harry Potter phenomenon. The first part analyzed the concept of magic, and the second examined the reasons behind the genre-transcending success of the series).
Let us say that storytelling is mostly about the navigation of a literary landscape littered with plot elements and moving (“growing”) characters. Robotic motion is mostly about the navigation of a physical landscape littered with physical obstacles and other moving entities. That analogy leads to some interesting insights into what makes for good story telling.
The easiest way to get a robot to navigate through terrain littered with obstacles is to have it wander around aimlessly till it gets to its destination by accident. Call this random navigation. The next easiest method is to provide it with a complete bird’s-eye-view map. Here is one such map, with the trails of 3 robots moving through it. The picture is from some navigation simulations I did a couple of years ago. In my simulations, robots always knew the whole map and could plan the best path through it.
The best known way to frame the problem is in terms of Simultaneous Localization and Mapping (SLAM). You figure out where you are and how to get where you want to go at the same time that you are building up your map. Harder, more realistic and more interesting. This post is about stories, not a robotics tutorial, so I just want to show you what a robot navigating according to a SLAM algorithm might “see” with its laser range-finder eye as it moves around.
(Irresistible urge for product placement: Don’t you think that my artwork with Corel Painter Essentials
When you watch a robot navigating under the control of a SLAM style algorithm, the trajectory simply looks far more interesting, exciting and biologically “real.” The robot behaves in ways that appear almost human (or rather, catlike): it gets into and backs out of dead ends. It can sometimes plunge ahead greedily and, at other times, explore cautiously. It encounters a good deal more of the landscape, but not because it is wandering randomly — it is trying its best to build up the most limited map it can, in order to find its way to its destination. By contrast, the random wandering and global map approaches make your robot act in pretty dull ways. It will either look like an automaton or like it is slavishly under the control of an omniscient being.
Navigating Literary Landscapes
Good stories move through their landscape like robots driven by SLAM algorithms. Everything you learn about the fictional landscape is necessary to the progress of the plot and contributes to character development and the ebb and flow of dramatic tension. There is just enough exploratory wandering to let the reader learn along with the protagonist (you can generalize the argument to multi-character and partly/wholly omniscient narrator viewpoints, but I won’t go there for now), and establish the situation to the point where more purposeful narrative progress is possible. Enough is left mysterious (my favorite example is the mystery of Tom Bombadil in the Lord of the Rings).
We all recognize that Rowling is no literary great, and are fans despite that fact. Except for Azkaban, her stories aren’t “good” in the sense above. All her books are cluttered with vast amounts of unnecessary exposition, tiresome detours and a constant stream of magical clutter being introduced for no good reason. While background development is a particular need in fantasy and science fiction, Rowling goes way beyond the necessary minimum and tells us far more than we need to know. Her robot often wanders more randomly than necessary. A great example is the entirely unnecessary putter-outer device that Dumbledore uses to put out street lights (why not just use a wand?).
A story that knows too well where it is going doesn’t work either. There is no dramatic tension, no discovery of essential information, no character-building through setbacks, and too little landscape revealed. Book 4, Goblet of Fire, has this problem. The story marches dully and efficiently without much tension through the very contrived landscape of the Tri-Wizard tournament to get to a very poor non sequitur of a climax (it is random along other dimensions, which explains the volume’s size).
Why Cuaron’s Azkaban Works
This glaring literary shortcoming though, is actually a major reason for the peculiarly modern nature of her success. I am referring here to the unusual friendliness of her books to movie adaptation. Because Rowling herself explores her world so clumsily, enough is revealed, by way of raw material, for others to navigate more artistically through it.
And that leads us to an interesting conclusion, Cuaron’s movie is good despite the fact that the third book is also good, not because of it. Why? Because Rowling herself crafts such a finely-judged trail through the landscape set up in Books 1 and 2 that it is hard to one-up her. Cuaron succeeded because he took a giant risk and made a movie that would be incomprehensible to viewers unless they had read the books or watched the previous movies. His pace is briskly impressionistic, and assumes intimate familiarity with the story on the part of the viewers. His visual style is focused on developing mood rather than plot. There is a sense of intimacy and participation, rather than observation. In fact, you might almost say, Cuaron allowed us to relive the story in a richer way than the experience of reading it allowed. In fact the movie would not have worked if it hadn’t been for the fact that the first two movies and books weren’t very good.
(If you don’t like my robot analogy, a more literary one with much the same point is to be found in a parable in The Alchemist
About Venkatesh Rao
