A conversation about magic in the Marvel universe over in Google Plus today has me thinking about the greater role of magic in storytelling and games. A ton of different sources factor into my thinking here, starting with Merlin and the wizard-heroes of the Kalevala, continuing through the Wizard of Earthsea, Discworld, and Name of the Wind (note: this is one of the only times I will suffer the latter to be mentioned in the same breath as the former two), on through Mage: the Awakening and Justice League Darkand whatever else comes to mind. I recall now that I've been writing nominally pseudo-academic things about magic since college and probably even high school, so now I hope to make good on all those years of confused essays about Faerie-logic in poetry and prose that my professors (inexplicably) tolerated.
Quite some time back, Rob Donoghue asserted that the central question of any combat resolution system is "How do we keep people from killing each other?" In a similar vein, then, the question of magic is, "Why can't magic solve every problem?" Obviously, if magic can solve everything, the story ends because there's no more conflict (or because the non-magical hero can't muster a threat against the magical villain) There are fundamentally three answers to this question:
Magic can solve every problem, but it carries a (scaling) cost upon the wizard... or the world.
Magic can solve every problem, but the wizard is a sufficiently flawed individual that unintended consequences erupt, replacing the existing conflict with a whole new conflict.
Magic cannot solve every problem; instead, it does effectively solve a limited range of problems, so the wizard must figure out a way to maneuver the problems he has into being problems his spells can solve.
The first two answers are found throughout myth, all the way down to present-day fantasy. The third does appear at times, but it's much less common in fiction. There's also a huge difference in whether the wizard in question is the protagonist, wise teacher, or antagonist; to create a compelling narrative, there are almost always more restrictions placed in the hands of an ally than an enemy, and still more restrictions on magic placed in the hands of a protagonist than an ally - after all, fiction writers have to work just as hard to protect their conflict from the magical powers of the characters as GMs do.
There are those - even respected fantasy authors - who regard breaking magic systems down into game-like rules as "the stomping boot of nerdism," and maintain that the setting and by extension its magic should at all times serve the narrative and nothing more. As a professional game designer and amateur author, my disagreement on this point surprises no one. I believe, though, that the unexamined magic system is not worth writing about, not because of my fetishistic love for magic systems, but because the audience can tell when the world didn't have enough thought put into it. A well-designed magic system protects a game from becoming unbalanced just as it protects a novel from collapsing under fridge logic. (Because you might be reading this at work, I am doing you a favor and not linking TV Tropes here.) I am hereby entering Justice League Dark into the evidence - I'll come back to it.
Since legends surrounding Merlin are a little too vague on the "rules" of his magic (or even indicating that he has done something magical), I'll look to Tolkien's Gandalf as an example of the first version of restrictions. As most gamers have figured out at some point, it's damned hard to model Gandalf's power in a game, since players would want to use his power more often and solve the plot more directly - perhaps by teleporting Frodo directly to Mount Doom and calling it a day. We don't have that clear of an idea of why Gandalf can't take a more direct hand in allowing the Fellowship to travel more swiftly and safely, except that there are other powers in the world equal to or greater than Gandalf the Grey that intervene; I think the text would pretty comfortably support the idea that Gandalf wants to keep a low profile. He could use his power more, but doing so would hasten the "endgame" long before he is prepared. That, then, is a kind of cost placed on his magic, in the greater context of the world. He may also have a limited amount of juice for his spells, but that is pure speculation. More directly, Frodo has the magical power of the One Ring; its story is absolutely about the price and burden of power. What could be easier than traveling across a countryside invisibly? But Tolkien attaches an overwhelming price to even temporary uses of that power, and voila! The narrative is saved.
Terry Pratchett answered this question for Discworld in Equal Rites, as he wanted to tell world-shaking stories in a fantasy world that had a whole Unseen University of wizards who had no other stated limit on their power. If the wizards use their magic, the horrible things from the Dungeon Dimension can enter and destroy the world; by consciously not using their magic, the wizards strengthen the barrier between the worlds, and Pratchett's wizards are for the most part academics who talk a good game and can make very convincing threats. (This isn't criticism of Discworld - they work quite nicely for telling stories.)
There are a huge range of inventive costs that fantasy authors have attached to the use of magic, though. One of the fastest ways to tell the good wizards from the bad wizards is that the good wizards pay the costs themselves, and the bad wizards make someone else pay them - the more innocent, the better. All kinds of different life essence apply here - blood, aging, simple fatigue, take your pick. Then there are those who sacrifice their humanity, tangibly (by mutating into something else) or intangibly (by slowly going mad, thank you very much Mr. Lovecraft and Mr. Jordan). Call of Cthulhu, in any of its game incarnations, has of course modeled Lovecraft's sense of each spell having a permanent cost to the caster. I'd also point to Pendragon as a system in which the wizard has no absolute limits on how much magic he can use, but he does so knowing that he will sleep for the rest of the year, potentially missing out on gameplay and putting himself in danger.
Hanging a magic system on costs like this presents some problems, though. If you want your wizard protagonist to be mostly intact, physically and mentally, at the end of the story, you can't have her pay much, or often. I'm not saying that this line of thinking causes storytellers to use Option 2 of the restrictions on magic, but it doesn't hurt. The examples here are just as varied; the hero-wizards of the Kalevala have the power to do damn near anything, with smithing, singing, or sampo. The problem, though is that Väinämöinen, Ilmarinen, and Lemminkäinen are possessed of such egregious character flaws that they have plenty of problems to work through. This is the classic trickster situation, and it's why every PC in the world knows not to use a wish casually. There's a sense that the cosmic balance knows you didn't pay for what you got, so it's going to give you what you asked for, and a whole lot more on the side. (As a quick aside, this same emotional logic sits behind almost every argument about rules for death and resurrection in games. To receive something this incredibly good, there has to be a price. Good luck with that.)
Schmendrick the Magician, Jonathan Strange and Mr. Norrell, and almost every bumbling apprentice ever written operate on these principles: that magic at best a tool of last resort, and probably isn't a great idea even then. At first, I said to myself that this was a system you didn't see often in gaming, and the reasons for that are obvious enough: players don't like having no particular control over the magic that they've gained over the course of play. That's wrong, though; there are plenty of clear examples of this approach to magic, beyond the chaos of the wish and the deckof many things. "Wild magic" has a lineage going back to AD&D Second Edition at the very least.
Option 2 leads to a lot of narrative complexity, because the consequences are seldom simple, and resolving those consequences takes more time than the initial conflict itself. Other than the Kalevala, it's generally not advisable to have multiple wizard protagonists if your magic uses this model. My example here is Justice League Dark, which is a comic book line of DC's New 52 in which all of the superheroes on the team have magical powers to one degree or another, and other than Deadman, their magical powers could be used to do just about anything. John Constantine, of course, is struggling to stay one step ahead of the bill collector, so he's more Option 1. The rest of the characters - Zatanna, Shade, Madame Xanadu, and Mindwarp - were all previously solo heroes, and this comic makes it clear that they should have remained that way. The narrative complexity of inflicting consequences on each character after they use their magical powers is juggling too many balls at once, even when Zatanna's powers mostly don't work, and Shade is more in the thrall of his powers than in control of them.
But in literary terms, there's more to this than the wizard not having much control over his spells: this is about magic as shortcut, a way of solving problems without doing the heavy lifting of figuring out a solution. This is the trap of ceremonial magic in LARPs: the game presents the players with a problem or set of problems, and the best solution the players can work out (given the tools at hand) is to perform a ritual casting. Once such ceremonial magic is on the table, it can make both Plot and players lazy. In each game where this has been an option, I think those involved have edged up to the yawning chasm of solving problems with rituals morning, noon, and night, but backed away and worked on varying things up. The ceremonial magic systems of each game involved risks of failure and partial success that the players could mitigate but not eliminate. The most staggering example of the Law of Unintended Consequences making its voice known in Shattered Isles was something that my character did - returning from death after the game's rules said I should have been permanently dead. The unintended consequences tore apart the cosmology and ruined the power of the Fair Folk courts... all because the magic my character used was parsed as a ritual and my randomized result was the absolute worst outcome available. At least for me - as the protagonist in my character's story - it made for a tragic and compelling narrative to that eventually re-closed the character's career.
Having said all of this, Option 2 is still less common in gaming than in fiction. In gaming, characters are usually already engaged in one kind of conflict when the need for magic comes up; PCs generally won't want to solve their problem by deliberately accepting a new problem. PCs know that this never really works out well for the protagonists in stories, after all. No, it's Option 3 that really makes the gaming world go 'round.
Option 3 is almost completely unrelated to the other two. See, Options 1 and 2 are both about trade-offs and difficult choices - much of the narrative tension centers on whether or not the character should use magical power to solve their problem. Once the protagonist makes that decision, there are consequences, otherwise known as "falling action." The story isn't necessarily over, but that cycle of conflict certainly is. In Option 3, though, the protagonist won't be punished by cosmic forces for using magical power to solve problems. The narrative tension is how the protagonist will use that power; it is a foregone conclusion that she will use it.
This kind of magical power is thematically different in a number of important ways. Firstly, where the wizard of Option 1 sacrifices for power, the wizard of Option 3 works. It is a sacrifice of time, to be certain, and possibly enough time that the protagonist has effectively given away youth - but I'm really more interested in cases where it's just a matter of study. Magic in these settings represents specialized academic study, more akin to having a doctorate in astrophysics than a sealing a deal with the devil. To put this another way, Options 1 and 2 raise questions about magical thinking, whether the character would have been better off solving the problem through mundane hard work, and the like. If you haven't earned power through normal channels, they say, what did you pay for it? Or perhaps it's just a matter of time until cosmic justice evens out the scales again. Option 3, though, doesn't care about bargains or cosmic justice. It cares about inventiveness, and therefore it makes a great game.
Option 3 doesn't see nearly as much use in mythology, but anytime a wizard is fundamentally equal (even in the short term) to a fighter-type in terms of ability to overcome challenges, this is probably what's at stake... for simplicity's sake, let's call it "game balance," though I can readily point to fantasy fiction that obeys this model. Skeeve in the Myth Adventures novels initially knows only a very few spells. They don't look like much power, but he proves to be quite creative in applying them to avoid overwhelming circumstances. The Name of the Wind novels by Patrick Rothfuss are another example of a magic system that drives its hero to solve problems creatively. I don't care for much of Rothfuss's writing, but it does do me the favor of being an example of what I'm talking about; he lays out clear laws of magic and spends some time exploring them.
At the other extreme, Rowling doesn't lay out a clear system of magic that the wizards of Hogwarts use, but the characters do operate with a relatively limited list of spells... that just happens to increase whenever she wants to showcase Hermione's superior academic spells. The point once again is that the characters have to work and practice to gain power; it's not a coincidence that Kvothe, Harry, et al., are students in schools of magic who sit through a whole lot of studying montages. The thing about Option 3, though, is that it turns magic into a kind of science - predictable and available. Any sufficiently advanced technology is magic, but any sufficiently comprehended magic is science, too.
D&D's magic system, while more expansive than that of the Potterverse or Myth Adventures, is at its core a limited list of tools intended to be applied to a vast variety of challenges. Here too, there's a clear sense on the part of the player that the power is earned, and it is narratively just for wizards's spells to do what they want and nothing else. Even so, that spellcasting still has usage limits, with the exceptions of at-will powers and the 3.5e warlock. The LARP systems that I've played and worked with seem, on the surface, to be more like Option 1, with mana or Fatigue as the cost that is at stake. Given the finite variety of spells and the full restoration of this resource each day, with no lasting cost to the caster, I think it fits more in the puzzle-like and earned magic of Option 3.
In many of the settings I've outlined above, I think it fair to say that the magic system defines the setting - the feel of the magic does more to set that world apart than any other single element does. Further, the laws of magic are the laws of the cosmos and the laws that govern some of the setting's most significant characters - if not the protagonist, then the protagonist's mentor or greatest foe. I write about magic all the time because I want to write about fantasy that is as new, inventive, and compelling as possible; ideally I'll recombine system and setting ideas on magic in a way that no one else has tried before, that models familiar ideas better than other systems, or that just makes for a more satisfying game experience.
Oh, um, also?
I write about magic all the time because I really like wizards.