The World that Knew Too Much

It’s not the detail. We pretty soon forget all of that, and — if forced — have to go back to the manual or the text book or whoever taught us the damn thing in the first place. No, what we truly remember of things we learn consists of general principles wrapped up in phrases, or pictures. We carry them around like religious icons, or perhaps skeleton keys, trying them in every lock we find, hoping that we won’t have to force the door with our shoulder. Sometimes they work, sometimes they don’t, but they’re the first tools we reach for, an establishing shot in the narrative.

I was trying to explain the concept of recursive function calls (or rather recursive predicates — it was Prolog we were using) one time a few years ago in a tutorial to which a single student had come. And he wasn’t getting it. The only pictures in his mind were those of iteration, rather than recursion, and they just didn’t fit. Clearly he’d been struggling on this for months. In a busier class, I doubt he’d have made his bafflement clear, but it was just me and him, so I could try any way that came to mind to give him an iconic representation of the concept. Eventually, I drew a diagram on the whiteboard of boxes within boxes, and arrows connecting them to show the flow of control, and he had a light-bulb moment. He got it. The difference was simply that between sequences of boxes (iteration), and boxes embedded inside each other (recursion). I could tell it was a getting-it that would stick, too, because he now had a new icon.

Another time a lecturer gave me an icon. The class was a general discussion of knowledge-based systems; these are systems that operate not by obeying a fixed recipe of computer code, but by analysing or manipulating some body of knowledge about the world (or merely their world). Example: the browser you’re using to read this post isn’t a knowledge-based system, but head over to MapQuest and ask for directions, and underneath the fancy interface you’ll be hooking into a knowledge-based system with tons of information about roads, addresses, distances, and so on. Knowledge-based systems do route-finding, task scheduling, medical diagnosis, and a million other things.

Here’s the icon he gave me, which is heavily paraphrased: For a knowledge-based system to be effective, the richness of the knowledge is far more important than the complexity of the rules which manipulate that knowledge. With a rich knowledge-base, you can get away with (and one way of getting-away-with would be effectively convincing its user that it possesses a reasonable intelligence) a relatively weak set of rules. And, as a corollary, a complex set of rules can’t compensate for a weak knowledge-base. In a nutshell: richness of knowledge is more important for the appearance of intelligence than complexity of manipulation of that knowledge. Rich knowledge is a long lever, which can be moved with a small tug. Poor knowledge is a short lever, which even the mightiest heave might not be able to shift.

I’m sure there was a lot more in that lecture, but I probably stopped listening at that point, happy that I’d picked up one thing that I could use. Like most chunks of iconic wisdom, it’s simple, clear, and true across many contexts. It’s easy to be dismissive of such iconic wisdom — what appears simple might actually be no more than simplistic — but that’s a trap. Not everything that’s simple is wise, but a great deal of what’s wise is simple.

To Grand Theft Auto: San Andreas, which I’m having huge amounts of fun with at the moment. This might seem a dizzying change of direction, but it’s actually not, because GTA: San Andreas is in effect a vast knowledge-based system. It’s also one which supports my icon. GTA:SA is not the best game ever made, as more than one review I’ve read has claimed. I understand the sentiment, though, because in some significant respects the game is certainly beyond anything I’ve played, and in one respect in particular. The GTA:SA world — its knowledge-base — is quite astonishingly rich. A couple of nights ago I spent several hours playing pool in a bar in the game’s fictive San Francisco — quite irrelevantly to the plot of the game. The sunsets are stunning. The rainstorms make me feel cold and wet.

Distance within the game has a significant effect, too. It’s not simply a matter of the game-space being bigger and richer — though it’s certainly both of those things. It’s a little like the effect of those boring going-places chapters in long fantasy novels, which I contend are necessary and important to the narrative’s structure. A quest narrative should be long, and its reading should be long, too. I think it’s a mistake to read such a narrative too quickly. A momentous journey, long in the making, is quite properly processed over a long period. The relatively long distances between places — huge cities, tiny towns, lonely shacks in the woods — in GTA:SA, and the way in which those distances are incorporated into the narrative by way of complex, goal-oriented missions, create a sense of place that I find deeply seductive. When I read, some time ago, that the game would include representations of Los Angeles, Las Vegas, and the desert between those cities, I had visions of it actually taking five or six hours to drive from one to the other. And my very curious response to that idea was: cool. I wish that had been possible. As it happens, the game-space is compressed into a smaller virtual space, but it remains vast for a game.

The richness of the world feeds in here too. Journeys of any distance would be quite comical and perfunctory if the landscape between the game A and the game B were thinly-conceived. As compelling as the spaces of a game like Tomb Raider might be, their strength comes from a cartoony reality, and a compression of tunnels, buildings, platforms and such into a complex puzzle-box. One cannot realistically imagine Lara Croft undertaking a journey of some tens of minutes across her world in order to solve a puzzle. This is mostly because the Tomb Raider world just isn’t rich enough to afford that amount of escape from the on-going narrative. GTA:SA’s world practically begs to be explored. Being pulled back into the main narrative at times feels like a parent pushing a new toy into one’s hand, when the object of greatest desire is actually the cardboard box the toy came in. One could — and I hope that the sense here is understood — actually live in the game-space, and ignore the game’s narrative completely. There are bars and restaurants to go to, embedded video games to play. One can drive, ride, or fly any vehicle in the game. Thrill-seekers can go BASE jumping from the tallest skyscrapers, swim in the deepest seas. Or just sit on the beach and listen to the radio.

As compelling as GTA:SA’s narrative is — it’s good enough, and that’s good enough — it’s quite possible to take the perspective that it serves as a Macguffin, in an unusually specifically Hitchcockian sense. The plot lays out a trail of jelly beans which leads the player across this vast landscape — North by Northwest, one might say — from the small to the large, from city to desert to city, from set-piece to set-piece, encouraging exploration and discovery. The triumph of the game is that the narrative is almost entirely dispensible, because one doesn’t need much of a carrot. The world itself is enough encouragement.

[Update: A small and only tangentially-related addendum to the above. Reading about GTA:SA, I was taken aback to see the game’s developer described as ‘Edinburgh-based’. Surely not, I thought. Someone must have made a mistake; no way was this environment steeped in SoCal geography and culture made by people in genteel Auld Reekie. So I went and looked and, sure enough, though Rockstar has people all over the world, Rockstar North is indeed in Edinburgh, just on the other side of Calton Hill from where I lived for six years. It doesn’t make any sense.

But then last night I started seeing things I should have seen earlier. There’s at least one neighbourhood in the game that carries the name ‘Calton’, after Rockstar North’s location. The bridges east from the fictive San Francisco peninsula are the bridges across the Firth of Forth — why I hadn’t seen this sooner, I dunno. And then, while flying a stolen jetpack to my lair in a deserted aeroplane graveyard (don’t ask), I realised that I was looking down at the Salisbury Crags. So now I’m watching out for more of Edinburgh.]