Internally consistent magical paradigms: A paean to good design

I talk a lot about verisimilitude in fantasy settings, both here and in real life (whenever a friend doesn't manage to escape). One thing I often think about is the metaphysical nature of magic, and the competing desires in fantasy TTRPGs to have player-graspable/predictable magic, mysterious/fantastic magic, and sensible/plausible world design.

I think it's both possible and desirable to have an internally consistent magical metaphysics ('thaumaphysics'?). By which I mean: a framework of magic that is based on inviolable rules (not necessarily complex ones) that are theoretically discoverable by people within the setting. At this point, some readers will clamour that I intend to "Conquer all mysteries by rule and line, / Empty the haunted air, and gnomed mine— / Unweave a rainbow", but my definition makes for a broader umbrella than you might think:

• The kind of spellcasting where you learn secret words and gestures which affect reality can be internally consistent.
• Magic where psychic phenomena are measurable in the universe, and spells are entities which occupy some analogue of physical space in your head, can be done with internal consistentency.
• If every magical effect is always completely random, in the sense of 'drawn from an enormous table or procedurally generated', that's internally consistent as long as the methods of invoking magic are.
• Magic that is fueled by spell points / mana / energy which the body recovers naturally over time can be internally consistent.
• Some creatures being innately magical by their nature, and defined methods existing to borrow those beasts' power, is internally consistent.
  
• If near-omnipotent gods are locked in conflict over the world, with factions and/or individual minor power variations, and all magic is that which they pass down, then with sufficiently delineated gods the magical framework is internally consistent.
• High fantasy magic aligned along ideas like "magic always requires a sacrifice" or "magic comes with a commensurate cost which can only be delayed or passed on to others" can easily be internally consistent, being based on a few simple rules.
• Frameworks where multiple types of magic and spellcasting co-exist (D&D-style) can be internally consistent.

What magic isn't internally consistent?

In most fairytales, magic is presented as something of a fait accompli. Sometimes there are rules based on the implied rules of storytelling, but not usually.

• The tree can be cut down with a single chop by a princess's hair pressed into the edge of an axe made of wax, and this idea goes unexamined. It says nothing about whether the same axe can break through a stone wall, or whether a prince's hair would work, or if a regular axe with the hair glued on would still have done the job in one chop.
• The wicked sorcerer tricks the king into agreeing to give up his firstborn child, but it's not clear whether the king follows through on this because of an ensorcelment, or because of the king's honest nature, or the king's fear of the sorcerer's unspecified magic.

Some TTRPG systems and settings aren't sufficiently well-described to have internally consistent magic. What Justin Alexander called 'dissociated mechanics' are sometimes an indicator that magic isn't internally consistent: characters, monsters, or environmental features might be able to do things or have in-game effects that "exist outside of the character’s world and they are only rough approximations of that world." They are insufficiently described in the rules for a player or GM to work out cohesive underlying principles (and use that knowledge to infer capabilities, make predictions, judge interactions and edge cases, etc).

Now, just because the metaphysics is insufficiently developed for the magic to be recognisably consistent doesn't mean the magic is inconsistent. But it does mean that the onus of filling in the gaps is on the GM, and usually the GM won't even think to try to do that ahead of time (which would require, at a minimum, reading and internalising all the official rulebooks and adventure scenarios of the TTRPG, and then coming up with something that is consistent - i.e., doing the rest of the game designer's job for them).

No, it will tend to only come up when some interaction in the game needs a ruling, and the GM has no (or too few) established 'first principles' of magic to fall back on. With every such ruling about the way the world is, the implied underlying rules of magic get squirrelier and squigglier, until it's nearly impossible to come up with an internally consistent framework. At some point, the GM makes a ruling which contradicts an earlier one or something else in the books, and all of a sudden magic is inconsistent.

(Those who went before warn those who go ahead; it's happened to me.)

Why care?

An absolutely fundamental aspect of a roleplaying game is that a player can get a measure of what is going on around them and make choices based on that. If you could boil down roleplaying to three words, "make reasonable choices" would be a decent first draft. That requires a character's surroundings to be (mostly) intelligible. As they travel, interact with things, look at stuff, learn and explore, 'a character's surroundings' expands to 'the observed world', and not just objects within it, but processes too. Simple logic has to work in the metaphysics, too, or it's impossible to do basic roleplaying.

If absolutely anything in the world can change at any time for any reason or no reason — if it's a world where things can stop being true, or retroactively not have been the case — then it's a world you can't roleplay in, let alone get invested in. You certainly can't identify character goals and plan to achieve them. In a world like that, when the GM asks "what do you do next?", the appropriate answer is "who cares?" and to start packing up your things.

If a game rule says that a spell of type X does Y, and another rule says that whenever a spell does Y it's [verb]able by any magical effect of type Z, then that tells the wizard's player something about how magic works. If there's a monster which accomplishes Y as an effect of its X spell, then the wizard's Z magic better bloody be able to [verb] it, and if not (because the game was poorly designed), the onus is on the GM to find a theoretical underpinning for that. They don't necessarily have to tell the player why not, but they do need to assure the player that there's a reason, and a theoretically discoverable one in-world.

If the GM has three options and they are 'sacrifice magical consistency to create this exception' or 'change an established rule of magic to support this monster design' or 'change this monster so it follows established rules of magic', they should understand those options to be terrible, mediocre, or necessary, respectively.

This is all such pain in the neck for the GM that a well-designed game goes out of its way to avoid it, by making its magic internally consistent and spelling out at least the more major of the rules on which it rests.

One particular example: Magical realism

Take the magic of artsy literature, 'magic realism'. Here's the kind of magic you would expect from that metaphysical framework:

The plants in your grandfather's garden all grow three metres tall and are well-stocked with flowers and berries. You have a photographs with you as a kid, overshadowed by them fivefold. You move for college, then for work, and when you come back to visit fifteen years later the plants are all in the same places you remember, as vivid as ever, but they only reach to your knee.

A knight in shining armour loses her child; her sword and armour break into splinters, leaving her able to be harmed by the worst of the world and unable to lead the charge against it. After a long period of character growth and psychological healing, an ogre appears on the horizon and, looking about for her sword and armour for the first time in a long time, finds them to be intact once more.

A stressed businessman goes home tired every night and must then spend hours shoveling food into the dragon's head that grows out of his bedroom wall, tongue lolling over his bed. He is completely exhausted by this and then has to sleep on the floor because his bed is taken up by the horrid mass. One day he sees an advertisement to work at a charity. He applies, gets it, and quits his former job, then goes home and finds the dragon's head has turned into the softest blanket in the world. He gets the best night's sleep he's ever had.

Despite its seemingly vague nature, magical realism can be made an internally consistent framework, by creating rules like:

• The world is divided into important characters and background characters by some criterion
• When an important character experiences a powerful or persistent emotion, their surroundings change to elicit that emotion in a hypothetical viewer
• The challenges an important character faces in their daily life give rise to the introduction of new fantastic objects into their surroundings representing those challenges
• When an important character develops as a person or solves a major problem, they undergo a corresponding symbolic transformation in their material possessions, environment, or body
• Any politically, economically, militarily, socially, (etc) powerful person, if they are a background character, has an appearance or key trait that is fantastic or phantasmagorical, and allegorical to the nature of their power

Not all internally consistent magical paradigms are equal (when it comes to TTRPGs)

Magic realism can be made internally consistent. But it still doesn't work in a TTRPG because of the interactive nature of the medium.

Readers can enjoy magic realism because they can, in the worst case, write to the author to say "when Jorris had his nightmare he should have woken up to find all the curtains in his room had turned into red-hot serpents, why didn't you turn them into red-hot serpents you hack". Most won't bother.

Players in a TTRPG tradition can't enjoy magic realism because not only are they able to say things like that to the game master, but they should. By  which I mean it's in their best interests to engage directly with, and explore, the magic system — as we've covered, magic is a useful technology and players would be idiots not to try to make use of it. It's not an 'exploit', it's the human condition.

A TTRPG world based on magical realism would be one of constant argument, or one where magic is proudly (unpleasantly) inconsistent, or one where the players have the (unpleasant) false belief that magic isn't internally consistent.

Other magical paradigms have other problems for TTRPG usage.

• Classic high fantasy magic in the vein of Lord of the Rings is typically too subtle for PC spellcasting. Also, the source media usually tries to hide its more detailed rules, if not its broad-stroke ones, in a way that's incompatible with a player wanting to know how many orcs their Concussive Wave is likely to knock down and how often they can cast it.
• If magic is (very nearly) random, drawn from huge tables or procedures, a spellcaster can't reliably do much other than cause a distraction, and (given the nature of many magical effects) can't do that safely.
• If magic is all in the hands of powerful entities, and people can use it to the extent that they can convince an entity to grant it, then you have too much 'roleplaying' and not enough 'game'. At the very least, you need additional game structures preventing degenerate conditions where the players talk to too many entities or win them over too completely. Along the way you need to solve the old dichotomy of player skill (at persuasion, in this case) versus character skill.
• If magic comes entirely from objects and materials, then the game must be very precisely designed to preclude too strong of a positive feedback loop whereby getting your hands on too much treasure lets you acquire too much magic which lets you seize more treasure which lets you acquire more magic which [...]

Considerations like these act as design constraints on fantasy TTRPG elements.

Design tensions in fantasy TTRPGs

I quoted Keats' Lamia above to set up a juxtaposition.

Yes, I think it's best that magic be discoverable, explicable, rules-based, and consistent.

No, I don't think that magic should be all discovered by NPCs in the setting, that it should be explained to the players, that its rules should be in a player-facing rulebook, or that its applications must all be obvious outcomes of a single consistent pattern.

Magic should be internally consistent (a) as a matter of elegance, (b) so the GM doesn't have to do a ton of extra work to avoid making bad rulings, (c) so that the world has verisimilitude, (d) as evidence that the designers knew what they were doing, and (e) so players can adequately roleplay in a world where they will interact with magic.

But the players shouldn't be sat down with a complete overview of the exact details of magical metaphysics and all the implications, any more than a fantasy book should open with a dry five-page essay on underlying rules and cosmology.

Graspable, yet mysterious. Sensible, yet compelling. Consistent, yet fantastic.