This idea emerged right at the end of a previous post as I was writing it, and I've had some time since then to turn it over in my head. I liked where the ideas about effect counters even better when applied to the eight spell schools - it was just a small step to go from Death Effect counters to Necromancy counters. If we're going to identify these eight schools of magic as fundamental to spellcasting (you know, in editions other than non-Essentials 4e), maybe we can treat them as flavors of mystical energy that the caster manipulates in a mechanically evident way.
This may involve more playing around with glass beads and keeping them in the correct place on your character sheet than some groups would enjoy, but given that I've just recently played another session of Warhammer Fantasy Roleplay 3rd edition, it all seems pretty reasonable to me. The core of the idea is that both being affected by and casting certain spells can cause a person to gain or lose counters of various kinds. A Necromancy spell, for example, both grants Necromancy counters to a target and deals extra damage if the target has Necromancy counters. Some healing spells might have their amount healed reduced by an value based on the number of counters. I also like the idea of Necromancy buffs (false life) granting the caster a Necromancy token because it's using the caster's soul to mitigate physical damage to the target.
Perhaps more unusual are Evocation counters, which I admit are conceptually lifted from Bright Wizards in Warhammer: Age of Reckoning. Casting the most impressive Evocation spells of any given level grants Evocation counters to both caster and target, and almost all Evocation damaging spells deal added damage based on the amount of Evocation energy already coursing through your body. Evokers, then, deal more damage with a given spell as a battle goes along, and also develop Evocation defenses just to survive an extended engagement with an enemy Evoker.
Transmutation and Abjuration work differently, as they are so heavily based on beneficial effects. The idea that multiple Abjurations in an area interfere with each other is well-supported in some editions of D&D, and I like the parallel idea that Transmutations targeting a person either encourage further change (penalty to saves against Transmutation attack spells like flesh to stone and baleful polymorph) or cause the person to resist further change (making it difficult to set up too large of a stack of Transmutation buffs) - I'm not yet sure which of those two models I like more.
Enchantment counters wind up working a lot like tangents from 2e's psionics system - I'd either make a certain number of Enchantment counters a hard requirement for the more absolute-control spells to work, or inflict a penalty to saves based on your number of Enchantment counters. All told, it's still a lot like Necromancy counters, aside from using Will instead of Fortitude, so I'm hoping another idea for Enchantment counters will come to me.
Illusion counters are a fun one, as they can track the character's growing evidence that what he's seeing is Illusion. Perhaps each attack roll that a character makes against an illusion grants that character an Illusion token, and he makes a disbelief roll every round. Once a game is using that idea, you could also have Enchantment spells that mess with the target's mind by removing a number of Illusion tokens.
That leaves just Conjuration and Divination counters. It's well within the flavor of the fiction for anti-Divination defenses to require a certain amount of work to pierce, and counters are a reasonable representation of "work." The accumulation of those counters comes not from casting a large number of Divination spells (though a few Divination spells might at least help), but from possessing objects with a strong sympathetic connection to the target, or from creating in oneself a strong connection to the target (whatever that might mean in your setting).
Conjuration is still up in the air for me, because Conjuration is such a mixed-bag school. Teleportation, healing, summoning monsters, summoning "physical objects" (like acid arrow) that deal damage but allow no spell resistance... it's all over the place and I'm not sure what I'd do with this. The main Conjuration effect that I'd like to see limited in some way by counters is summoning creatures, so maybe summoned (not called) creatures grant the caster a Conjuration token every round they are maintained in the world, and Conjuration tokens past a certain point run the risk of drawing Unwanted Attention From Beyond.
While I might need to give this more thought, I assume that Evocation and Necromancy tokens are just about always purged at the end of an encounter (though I can special cases where this does not happen, such as magic items). Transmutation and Abjuration counters stick around exactly as long as the spells that grant them. The Conjuration idea described here would probably be purged at the end of an encounter as well. Illusion might reasonably need to build up over the course of an adventure, if the whole adventure is in an illusionist's funhouse. Enchantment is tough to judge without having written the specific spells, since there are some cases where the fiction would support long-term erosion of the target's defenses, but in general it feels like it should be a short-term effect. Since there's such an uneven divide here, maybe the whole subsystem needs a concept of transitory counters and lasting counters, but that's an extra layer of bookkeeping that is probably undesirable.
This idea could certainly be extended in any number of directions - poison counters, or counter "tolerance" for specialist wizards (or just people who spend feats). A lot of what's going with this idea is that the more awesome magical effects can include tradeoffs in a variable but accumulating way. I can imagine completely reworking the idea to model swordplay, though I wouldn't want to use that system in the same game as this one if the swordplay system involved multiple flavors of counters - again, just Too Much.
I'd be interested to hear other takes on this idea, if its core is suitably intriguing.