I’ve been reading Dungeons & Dragons books from the end of the ’80s – meaning the Mentzer edition, to be exact. They’re full of freaky shit that makes little sense, but is certainly thought provoking. My favourite is the Crucible of the Blackflame, the halfling racial artifact that has the minor power to repeal entropy alongsides its actual function of producing cloth that flies. Apparently the racial purpose of the demi-humans in D&D is to create kites that fly without a wind and take decades or even centuries of slave labour to create. Quite random stuff, that.
Anyway, reading all this old D&D material has turned my mind upon the Vancian magic system of D&D. I like the Dying Earth stories of Jack Vance a lot, but I never really appreciated the magic system of D&D – what works in a book won’t necessarily work in a game that has quite difference concerns, anyway. When I was younger I was actually quite adamantly opposed to D&D style magic because of its utilitarian nature – in my fantasy adventure roleplaying and reading magic was a mysterious, powerful force that would basically be a big deal whenever it made an appearance. Many people also dislike the somewhat counterintuitive memorization system with spell levels and such, preferring different sorts of mana systems – this is not the case with me, though; a homogenous pool of magical energy has never provided interesting detail to my play, so while it certainly has made appearances in my games, the solution was always a cludge.
After playing the Mentzer edition of D&D and reading the related books I’m starting to think that I could perhaps make a Vancian magic system palatable for my own D&D gaming. Let’s see what such a system would look like:
Vancian magic might seem arbitrary and strange in D&D for a fantasy literature fan, but that’s to my mind largely because D&D has this basic tendency of cutting cultural ties and representing ideas in isolation. D&D magic is strange, as it presents us a world where magicians work their craft as neatly prepackaged and standardized effects that are sharply limited in what they can achieve, and apparently work with complete reliability like any engineering. The original Dying Earth representation is rather different: in it magical forces are mighty, unknown and dangerous, and spells are just a half-forgotten operation manual for forces largely beyond human ken. A spell may in this context be seen more as one specific form magic takes, rather than the basic building block it is in D&D. The necessity of working with spells is caused by a misunderstanding of magic and the consequent inability for working with the background theory.
From this viewpoint, let’s say that my Vancian magicians only use spells – which are these prepackaged things familiar from D&D – when they get out of the laboratory and need to carry magical force with them in a form that can be whipped out on a moment’s notice. This requires intent meditation and concentration as the sage forces his brains to store magic – it’s essentially like binding a demon into your own head for most purposes. The ability to use spells is something that only experienced magicians would even consider, and them only when they expect a genuine need for it. In everyday conditions wizards would not cast spells – they’d perform rituals and alchemy, which are the real building blocks of those spells: a spell is just a dangerously packaged magical ritual crammed into the cranium.
In other words: my wizards have a wide range of magical skills that are of no immediate use in action adventuring of the D&D sort, and their capability for memorizing spells arises from this expertise. The lore comes first, and the spells follow.
Spell memorization table
System aesthetics are really important for me. I’ve always tended to dislike the D&D wizard simply because he has a cludgy table to use in figuring out how many spells he knows. I know this is petty, but it’s important enough for me that I’m going to use a simplified table:
As you can see, my table is very simple and can be expressed as a rule: the wizard may always memorize spells equal to his level, and he always has to memorize more spells of the lower levels than higher ones. This resembles the old D&D table quite a bit on the lower levels, but apparently playability concerns inspired Gygax to go for a faster progression on the higher levels. These are not my concerns, though, as I design from the opposite direction: I first choose to use this simple table for spell memorization and then create the spells themselves and other wizard abilities with these capabilities in mind.
One may also easily see that the power curve of these wizards is less than that of D&D wizards – getting even fifth level spells requires 15 experience levels, which in my rough homebrew sensibilities is already a rather high level. I’d imagine that third level spells would be the highest that one might expect of a vocational or professional wizard in these rules. Fifth level spells are probably a major campaign issue if there should be wizards capable of using them. Of course, such a wizard would probably do even better with time and ritual magics.
Not all magic-users are wizards – in the context of my own homebrew rules, a character would need to learn the right feat to get access to memorized magic in the first place. And learning feats in the system requires the right fictional context and an ability check, so the aspiring wizard might not get to memorize his first spell on the first level, unless he has the suitable Ability scores to succeed in learning that feat.
Actually using spells
I don’t have many complaints about the D&D logistical rules for spell “memorization” in tactical terms. The requirement for rest, memorizing once per day and so on are pretty reasonable. They’re also not very flavourful, though, and there are stupid and unnecessary bits from my viewpoint. For instance, the requirement for spellbooks to be ultra-expensive that some D&D editions have is something I don’t like. For my purposes the spellbook is not a special tool, it’s just a book that the wizard uses to aid his memory. If a character wants to memorize the contents so as not to have to lug around the spellbook, feel free – I’d imagine that a given spell would equal anything from a couple of pages to a couple dozen pages of content, so it’s by no means an impossible feat to memorize even several spells. Memorizing the spell text itself from the book won’t mean that the wizard can use the spell more often, though – the magic has to be rebound after having been released, which is a separate concern to whether the wizard knows how to bind the magical spell in the first place.
(Compare to the D&D metaphysics, which have the wizard magically forget the spell once he has cast it, explaining why he can’t cast it again and again. My conception of a Vancian system deals with the spell text as instructions for binding the magical force, and remembering it after the force has been released is not of any more use than referring to a book or scroll with the same instructions. The text is just a text, not magic in itself.)
The actual spell-binding process would be a potential conflict for my purposes. A wizard in ideal conditions wouldn’t need to make any checks about whether he succeeds in memorizing spells, but I could imagine many situations that would require such a check – having to work the spellbinding in a dark dungeon instead of a laboratory, having to work from memory instead of written notes, trying to bind a new spell for the first time, trying to cram more spells than the wizard normally could… lots of reasons for preparing spells to be a challenge in itself. Such situations could then be resolved with the normal ability check mechanics, just like anything else the characters try. Consequences of failure could range from losing spell slots until the next rest to outright damage as the character messes up and releases the magical forces he’s trying to bind.
I probably wouldn’t enforce a need for rest before rememorizing spells that have been cast, by the by; that’s just another opportunity for ability checks and seeing if the wizard lobotomizes himself by pushing his abilities too hard. And of course a competent wizard could cast any effect he can bind as a spell as a relatively lengthy ritual to avoid having to memorize it at all. Basically I’d handle memorizing spells just like I’d handle breaking down doors: it’s tough work and the character can expect to be able to do it at a certain level pretty confidently, but still things go wrong now and then, and the GM is quite within is pregorative to limit the opportunities when the wizard pushes himself too much.
As for casting the spells, I actually like the D&D method of having the spells themselves indicate whether they require somatic, vocal or material components to release. This is because I don’t brook with spell lists, though, as I’ll explain below. The same goes for casting time: spell-by-spell determination is quite fine with me as long as it doesn’t translate into a humongous spell list.
Spells in the memory
I like the idea of the spell reserve feats in recent 3rd edition D&D: a wizard who has a spell in his head can manifest some minor power of that spell without actually casting the spell itself. I find that this captures well the idea that the wizard is constraining a great power within himself. I likely wouldn’t mechanize it, but basically this would be my replacement for the clunky cantrip system D&D uses to make wizards seem more like fantasy literature wizards. For instance, a wizard with a mighty fire spell in his head could light a candle with his magical powers as long as the spell remained, that sort of thing.
I’m also grim enough to consider making the “spells”, or at least some of them, resemble demonic entities. What if they have personal names? What if they want things? What if they drive you insane if you try to sleep with them in your head? Plenty of potential for quite original inconvencies in the Vancian magic model; it’s not all utility and laughs. The notion of spells as living magical power cuts many ways, though – I probably would exercise the idea that a wizard could memorize the same spell on different spell levels to give it more or less power, for example.
Now, this is probably my most radical departure from D&D practices. I don’t like having a long pre-built spell list that the player uses to choose what spells to learn. Older editions of D&D were smarter in this regard, as they left the choice of spells up to the GM. I’d do this a bit differently, though, by having the GM and players design the spells instead of picking them from a list.
Without a ready-made list the GM will create flavourful and strange magics that fit the conditions and metaphysics of the setting. Of course, mechanically this could be tricky if you insisted on perfectly mechanized spells – the modern D&D approach to spell creation takes a lot of work, as the rules system is a finely tuned machine in which the spells would break things were they not carefully measured.
In my approach, however, magic would be primarily defined in the fiction, and it would be mechanized using the same rules that everything else uses. So a spell that causes things to burn would work mechanically just like anything that caused things to burn. The GM could set the constraints for the spell, such as material components, casting times and limits on how the spell can be controlled at the casting time, as well as some general sense of the mechanics of how the fire is caused and how powerful it is – but these are all, still, in-setting facts that do not need mechanization before the spell is actually used.
As an example, consider a spell that allows the spellcaster to fly. We might know that it lasts until the next sunset/sunrise and that the user flies sort of like a bird in aeronautical terms. Do we really need any mechanics apart from such natural language that refers to the fiction? A simple ability check can be used to find out whether the flier manages to control his flight in a tight spot, and that’s pretty much that. Of course I would allow the spell to be used creatively within the fantasy paradigm: instead of flying himself, the wizard might direct the magic to make somebody else fly, that sort of thing.
The actual difficulty in mechanizing spells is, of course, in setting their spell level. Other things are easy to balance simply by choosing an appropriate level for the spell. So how would I go about this as the GM? My first idea is to define the spell level in pretty naturalistic terms: the weakest spells I care to create are first level and spells that are stronger than that are of higher levels. So “second level” really only means “spells that are appreciably stronger than first level spells, but weaker than other stuff”. Presuming that the GM can meddle with this stuff as he goes along, I don’t really see much need for carefully balancing spells to their levels before play. I find this stuff to be setting-dependent, anyway: if a 5-level wizard in my game is a master of the art, then I just need to know what sort of things such a master should be able to do according to my sensibilities. In fact, keeping in mind my preferred level range, I can easily say that first-level spells are minimally useful apprentice stuff, second-level spells are the stuff people would expect wizards in the setting to be able to do, third-level spells are known by the masters and fourth-level spells are only known by the archmages. With that in mind it’s not that difficult to sort each spell to its appropriate level.
I find nothing particularly abhorrent in the Vancian magic system, it seems like it’d be fun and flavorful to play with it. The D&D implementation is more of a problem with its mechanics and limitations, and a metaphysic that resembles an engineering textbook. But remove all that and it’s quite colorful, actually. “High-magic” by nature, perhaps, but even that could be toned down by making the memorizing itself more expensive, more hazardous and more of a big deal in general. At heart the Vancian conceit is really just that powerful magics are cast in advance and then stored to be released later. This is not exceptional at all in fantasy literature, it just feels weird in D&D which has completely lost touch with the ritual magic part of the equation.