#would this intuition have naturally developed if I just actually did enough category theory for real? probably

[as-if-and-only-if]

I have a new zest for category theory! by going through a couple of very basic proofs with attention to how I was thinking, I think I identified part of what was making it frictional in my brain.

weirdly (in part) I had the intuition for composing arrows, but not decomposing arrows. that is, my brain and intuition had internalized how you could have f : a -> b and g : b -> c and get gf, but not how you could start with h : a -> c and be on the lookout for factorizations of h through some such f, g.

which is kind of odd, right? I mean, aren't these just the same fact? well, yes, but they're not necessarily regarded as the same by your intuition! it's a reminder of how brains (and, more generally, "implementations" of math, such as in a proof assistant) need not expose the parts of the math that actually drive its use, but only "incidentally" allow the important parts to happen. I mean, I was perfectly fine seeing some such h be decomposed in a proof, and even doing it myself as an incidental step in a proof, but I realized I hadn't indexed it as a fundamental part of how categories(/categorical proofs) work for ready application.

and that's what intuition is, really: not the stuff you understand per se, but the mental processes that are "at hand", that are "active" or "triggerable" in your brain when you're thinking about something.

(by the way: the other part of re-developing fundamental parts of my intuition here was realizing that I can intuitivize(? not intuit, that means something else...) universal properties the same way I intuitivize unique factorization theorems.)

[thefifthworld]

RPG Theory July: Asking Questions

There's a hashtag on Twitter for #rpgtheoryjuly where a lot of brilliant and talented game designers are talking about what makes roleplaying games work. Here's a contribution from Jason Godesky about some of his design principles for the Fifth World.

My first attempt at a Fifth World RPG was actually a D20 system. It was a disaster, but I kept trying new things. There was a Fate version at one point, which was less of a disaster, but not quite right either. Where things really started to come together for me was when Epidiah Ravachol, in promoting Brie Sheldon's "Five or So Questions" series on Google+, said that questions offered the most innovative, powerful tool in roleplaying games today.

(To close the circle, I was interviewed by Brie for "Five or So Questions," where I mentioned how Eppy's review had influenced my design.)

It's amazing how much you can control where a conversation goes by controlling the questions asked. The Fifth World is not an intuitive setting for most people. Roleplaying games ask players to create a shared fantasy, which already poses a challenge in and of itself. Most roleplaying games rely heavily on understood genres to keep that fantasy shared, and still often stumble. RPG players often recount their horror stories of the Lord of the Rings style fantasy game where someone insisted on playing a robot or a samurai. Most games lean heavily on a genre that all the players know and understand to keep players from introducing elements that are too jarring and might break that shared fantasy. It still happens, because a shared fantasy remains a daunting undertaking, but with a well-understood genre, we at least all understand the expectations and can keep it to a minimum.

The Fifth World has no such genre, and for a long time we struggled with maintaining the shared fantasy. Many players would look to the post-apocalyptic genre and make it far darker, more violent, and brutal than it should be, which would break the shared fantasy. Some players would push it towards the fantastic, making magic explicit instead of personal, tipping from animist realist to high fantasy. Some players would turn the families of the Fifth World into silly cargo cults. Others would turn them into noble savages. Without a genre to point to, you face a bunch of bad options, primarily big, thick setting books that players have to study before they can play (meaning that most won't, either playing it without studying the big book, or just not play it at all), or move that to the table and have someone lecturing everyone at length about what the setting is and what it isn't.

I think we found a good answer — in questions. The game begins with family creation, where the game asks you questions: about your family's history, how they dealt with the changes in the world, and the customs they developed along the way. This sets up a powerful co-authorial relationship. Since the game is asking the questions at this point, it gets to lead the discussion. It sets up that the world has changed, and that your family did face these challenges. By providing the answers, you and your friends come up with a family that has a history. We often generate customs that we think are not so great, or belong to an older time and should be set aside, but because they came from our own answers to these questions, at least one of us can see why it was a good idea at one time. That makes us all feel connected to this family, and invested in its continued welfare. While I still see players regularly plotting how to overthrow families they had no hand in making and how to bring back cities and kingship and 9-to-5 jobs and poverty, I haven't seen players who want to do that to families that they helped create. They might want to overthrow a particular custom, but for the good of the family and because they're invested in their lives overall, not because they aren't.

Once you have a family, the game still centers on questions, but it opens up a bit. Now it's all about who gets to ask questions, what questions you can ask, who gets to answer, and whether or not that answer is reliable.

I think one of the biggest design triumphs in roleplaying games in the past decade is the principle Jason Morningstar used in Fiasco. Throughout that game, you can choose to establish one half of the truth or the other, but never both, and so the result is always something shared. You can choose how the scene begins, or you can choose how it ends, but not both. You can choose to establish the general category of a bond or the precise nature of it. Questions have this same quality. Asking the question holds a great deal of power, but so does answering it, and by its very nature a question breaks this power in half and gives each to a different person.

I was inspired by a story that Calvin Luther Martin tells in The Way of the Human Being about a man who unwittingly visits a beaver family. He emphasizes the theme in the story of keeping our perception open as long as possible. Each definitive answer makes the space of possibility smaller. It's a story that emphasizes the importance of not limiting that space too quickly, to give it time. This is an idea lurking at the heart of gameplay in the Fifth World. The story is a wild thing, something that exists between us and amongst us but could never exist anywhere else but here, with us, right now. It takes solid shape as we ask questions, but it also dies. Only at the end do we know what the story was, what it had to be because we were the ones telling it and sharing it at this place and at this time, but it's also at that moment that it is no longer the wild, autonomous, shape-shifting thing we set out after. We hunt the wild story. Our questions allow us to see it, but they also kill it.

You can ask any question, but unless it's on the short list of questions available to you, and you spend a resource (awareness) to back it up, the answer may not be reliable. Unreliable answers can give us sensory impressions, but they may or may not lead us to the truth. They don't necessarily limit the space of possibility. They don't necessarily bring us any closer to the wild story.

Strategy involves picking the right person to give you an answer that will shape the story the way you'd like. For example, in our playtest saga, everyone knows that if you'd like more philosophical or anthropological depth in the story, you should ask me; if you want things to get weird and dark, you should ask Wayne. Strategic gameplay is about knowing your friends.

You can learn the names of places by getting to know their story, and that unlocks new questions that you can ask and get reliable answers to, but only in that place. If you want to ask that question, you'll need to lure the person you want to ask there, to use the unique spirit (the unique atmosphere, the unique setting, the unique properties) of that place to draw out a reliable answer. If you get to know a place very well, you might know it well enough to carry that question with you and ask it wherever you go.

While listening to the audiobook for Peter Godfrey-Smith's Other Minds, I gained a deeper level of understanding of why this works so well. In that book, Peter writes about the two functions of the brain: to perceive the world, and to coordinate action. Of course, these are also deeply connected, because when we coordinate action we change the world that we're perceiving. They exist in a loop. The gameplay in the Fifth World being built around questions would seem like it only hits one half of this cycle, perception. And explicitly, that's true. The other half of the loop is what you do with it, that is, the ritual phrases, where we do things that are difficult and/or dangerous (another point where I stole shamelessly from Epidiah Ravachol, this time from Vast & Starlit) and try to understand another person well enough to appeal to her needs.

Asking questions can help build a shared fantasy even where there's no genre to rely on. It focuses gameplay on exploration, and it creates an anarchic flow where we all get to wield the supreme power of the person asking the question or answering it, but never both, and always only for a moment.