On the 2nd of April, Adventure Jam will kick off and I’ll be giving it a go: attempting to write a work of interactive fiction (IF) over the space of two weeks. Or, more accurately, attempting to write a work of IF in my lunch breaks over the space of two weeks.

I’ve been reacquainting myself with Inform 7 as a sort of warm up, and that – along with my work encoding human-written data into searchable XML – got me thinking about the representation of natural language as gameable elements.

(There follows a long rambling post with little structure wherein I waffle about stuff that’s on my mind.)

The main use I’m guessing readers will be familiar with is aspects from Fate: a simple natural-language statement that plugs into the game system in a few simple ways, with the intermediary step of interpretation performed by the group. But it pops up in a number of other places too: throughout Risus and Cortex+, and even in the skill lists of trad RPGs – “Can I use my Fishing non-weapon proficiency to reel in the cave fisher?” sort of deal.

The trick is the interpreter. A set of roleplaying game rules is a mechanical system – pretty much identical to an XML schema or the parser of an IF – and natural language, with all the idiosyncrasies which make poetry possible, doesn’t play well with mechanical systems.

(Although it occurs to me that maybe a thoroughly systemic language like Esperanto might work? But I’m not about to write any sort of game in Esperanto.)

Anyway, an analogue game gets to use the players at the table to lubricate the process of turning natural language into mechanical input, but computer-based interpreters like IF (or RPGs which seek mechanical rigour instead of player-group consensus, whether they’re computer-based or not) need some way of determining what any given natural-language phrase means which is… well, let’s just call it an open problem.

So what does all that mean, vis a vis RPGs?

Tags!

You can see tags in a lot of mechanical games. The most thorough example I can think of is D&D 4e, where every power that does fire damage gets the Fire keyword, and other powers or elements of the game can interact with that to suppress or enhance it.

This is probably what leads to a lot of (or at least some of) ‘4e = video game’ comments. The game borrowed this interpreter technology from digital media in an attempt to more rigorously define its functions, although you can see the trend starting in D&D 3.5 with monster types (construct, humanoid, outsider (demon), usw.) and the limited range of keywords like ‘mind-affecting’ and so on.

Unfortunately, in trying to make its definitions more exact 4e ran into the same problem that poorly-implemented IF does: things that you expect to work, don’t. There’s no basic rule in 4e that says “when you strike Water terrain with a Lightning power, the power attacks everyone in the Water terrain” so unless individual GMs implement their own, a 4e lightning bolt doesn’t conduct through water.

But this is only a problem for games which try to implement both rigorous mechanics and the ‘rules-as-physics’ model. When you’re writing IF – or any CRPG for that matter – you can get away with a less-than-perfect world simulation because the audience understands the limitations of the medium. With an analogue RPG people expect the world to behave like the world, which is normally achieved by using the players to paper over the cracks (and this too is accepted because the audience understands the limitation of the medium). But when this collides with a mechanistic set of rules that resist player intervention – whether by being too complex to easily muck around with, or specifically rejecting modification (I vaguely recall Burning Wheel does this?) – you get friction.

So… consequences for game design? A couple of possibilities:

Communicate expectations better. If you want a mechanically rigorous combat system a la D&D 4e then put that front and centre: “This is not intended to simulate a world, just make for a good action scene; handwave any inconsistencies.”

Stop trying to create detailed, rigorous systems for world simulation. This might not be the most satisfactory outcome – handing the consistency of the game back to the players instead of writing rules is perilously close to the Rule Zero Fallacy of bad game design – but it’s my preferred one.

Digital media has to use the expectation-communication approach because it cannot expect human input. But analogue games can’t be played at all without human input. When you’re writing an RPG you can’t assume that the group has dice, cards, a table to play on, etc. but you will always know that the group exists – so why not build that right into your design?