"It makes sense to me? But maybe there should also be space for something more... magical?"

"Maybe this should be like, the mainstream way of doing magic? Like with folk medicine? It's mostly superstition but then you find out that willow bark works because it contains... something?"

"I don't know, it feels really neat to have magic as programming, but it also feels... reductive" 

"Salicin," Hammond supplies. "It's a precursor to acetylsalicylic acid, which I think people around here use for pain relief. It doesn't work very well on walruses on account of our delicate skin."

He relates this with the air of someone repeating rote facts while his attention is elsewhere.

"And I do see what you mean, I'm just trying to think how to phrase my natural follow-up question ..."

He drums his tail in thought.

"Earlier you made the comparison of spell diagrams being used as a kind of mnemonic, to guide the creation of the spell — how would you feel about a system where there are multiple competing kinds of mnemonics, such as different kinds of rituals, some of which are programming-flavored and some of which are more esoteric?"

"Yeah, I think that works. Like you have wizards and witches and sorcerers and it's all magic but with different flavors."

He nods.

"Excellent. Alright ..."

"What are some things that you think magic should definitely be able to do?"

"Fireballs! Teleportation! Flight!"

"I want to say 'control Time and Space' but I think that could create paradoxes?"

"Well, it depends on what you mean by 'control Time and Space'," Hammond replies. "For one thing, there are some models of time-travel where paradoxes — except for pre-destination paradoxes, which I don't think should count — are impossible. The 'the past was always that way, and when you go back in time you will turn out to be making it be that way' model. Or the 'traveling to the past creates an entirely independent timeline' model, which also avoids paradoxes like that."

"If you could pick a particular way for time-travel to work (which you can), would you want to extend your answer?"

"Oooh, like the Asimov story! I love that. It means that you can go back in time but only to change things you don't already know to be different. I reeeeeeeally dislike the independent timeline one. It means that in your original timeline you just disappear, and all your friends and family mourn you. Terrible. Horrible. The worst."

Hammond diligently records this.

"Alright — what should magic definitely not be able to do?"

"Make it easier to kill people, I guess. Like, obviously if you can throw fireballs those can kill people but like, no Power Word Kill or Inflict Wounds. And honestly I'm having second thoughts on the fireball. I think it's acceptable if magic is dangerous and can be used for evil. A kitchen knife can also be used for evil. But I wouldn't want more magic to bring more misery."

"A thought."

"I don't like it, I don't think it works in the real world - well, the world is weirder than I expected but... you get what I mean I hope."

"Maybe there is a way to make it work, you seem to have good ideas."

"What if you needed to be Good to use magic?"

"Oooh! Clever!"

Hammond straightens up and considers this for a moment.

"So the basic idea is definitely possible: you can have a system that judges people on their Goodness, and only gives magic to the ones that meet some threshold. The tricky part is where the definition of Good comes from. There are plenty of universes that judge Good and Evil in various ways — sometimes contradictory ways. How would you go about defining Good? Either directly, or in terms of a process by which magic, which lacks built-in mortal sensibilities, could come to know it?"

"It's something about... using magic for good? But that's not really explaining it, is it?"

"What if... my intuition is that using magic for good means that you do something which is good for people. Most of the time this is easy. If I scrape my knee, healing me is going to be good for me and neutral for everyone else. But what if I want to be very rich? That money needs to come from somewhere, so the criterion cannot simply be 'good is what a person finds valuable', although it definitely is a component of that."

"And it's not even a matter of how many people a spell is good for, because I'm sure that you can convince the majority of people to do horrible things. If you've read anything about our history for the last century you would know."

A moment of silence, to remember.

"I don't know. Good is... helping others, working together, happiness, safety, growth, discovery, freedom, but this is still not pinning it down..."

"Well, in fact, with magic money doesn't have to come from anywhere, but places that work like that tend to get a bit strange," he notes.

"Hmm. Would you say that something that has more of those traits is more likely to be good, and something with fewer of those traits is less likely to be good, but no specific trait in that list, itself, directly indicative?"

"Yes, definitely, and I can probably come up with a better list with more time. Does it make sense to do that?"

"Maybe! One thing that magic can often do is tell whether two things are 'similar' or not, because there are a lot of magic systems that depend on being able to make that kind of judgement in one way or another, when you think about it — so the capacity to judge similarity is something that the Will of Magic often has to implement. So if you want to, you could define goodness by listing a bunch of things that are good, and a bunch of things that are bad, and saying that, when determining if something is good or bad, magic should compare it to each one of those and combine those comparisons to form a final judgement."

"The risk with a system like that is that you're giving up fine control over the precise definition of 'good', in favor of a sort of vague gesture. But if you can't give a precise definition of 'good', maybe a vague gesture is better than the alternative. Does that kind of system for judging goodness sound like something magic should do? It's definitely not the only way, so if it doesn't sound right we can talk about alternatives."

"I'm not convinced? If you go back two thousand years, no one - well, almost no one - believed slavery to be particularly evil. I don't want to bind magic to unchanging values when in two thousand years we could realize that... I don't know. Pineapple on pizza is evil. Even though it clearly is."

"Well, any specific concrete decision you could make now could turn out to be wrong in the same way," Hammond points out. "So it sounds like maybe you want a procedure that magic can go through to change the definition of 'good' as people's views change, so that it grows alongside people."

"That sounds way better, yes! Although... like. We don't have world peace right now, so 'the majority view' is not necessarily... enough? I don't know. I feel very unqualified."

Hammond bites his lip. There's a limit to how much he's supposed to lead people's answers, but surely helping her brainstorm a bit isn't a problem.

"I agree that the naive majority view probably isn't the best you could come up with," he settles on. "For one thing, to use your slavery example, the slaves probably knew it was wrong far before the majority of people did, and if magic could have picked up on that it would be a good thing."

"When I have a hard problem, I sometimes find it useful to think about how I could identify a solution if I saw one," he muses. "What properties would a satisfactory solution have? I can write them down, and then we can brainstorm alternatives and see which one comes closest."

"Well. A solution would look like magic being used to make the world better and not worse. Which is kinda self-referential, we are trying to define 'better' in the first place. Some values are prooobably going to be seen as good in the future as they were in the past? It's good when people are free to make choices for themselves. It's good when people can eat, have a place to stay, friends, when people have time off work or school to pursue their hobbies and passions. Very little poverty, ideally none. I think it's fine if some people are rich, as long as the gap isn't too big? Olivetti had a rule that the CEO pay couldn't be more than ten times a blue collar worker's pay, and it seems intuitively like a good idea? I think... wait, I'm getting distracted from the question."

"I'm somewhat tempted to just crib some ideas from the Gospel? There's a lot of stuff that was meant for... a different time, but the basic message of 'love each other' is sound I think."