Reductionist Magic 3: Souls and Illusions
Previously I announced a pretty ambitious project: applying reductionism to DnD magic in order to establish broad “laws of physics” that could be used to consistently model circumstances that aren’t defined in the rules. Previously I discussed some broad rules of how magic works, specifically when it comes to spells. Today it’s time to discuss souls, their internal modules, and their connection to the body.
As usual, if you are confused about any terms, make sure to check out the glossary, and if you are looking for other posts, check out the blog map.
What do we know about souls from our core sources? At first glance, it may appear that we know absolutely nothing. There is no passage in DnD rules, to the best of my knowledge, that talks about this topic in depth. But actually, if we reflect on what we already know about souls and laws of magic, we can see that some conclusions about their structure are simply inevitable. Furthermore, these conclusions will help us understand how certain kinds of spells have to work. Here is the general model of a soul - body pairing:
What do we have here? Well, we have the body, which can interact with the physical world. We have the soul, which can interact with the magical world (by e.g. casting spells, or receiving feedback from previously cast spells). They have to interact with one another - after all, the soul has to pilot the body around, and that involves information transfer. We of course also have the mind, which has to be split between the soul and the body - it has to have a part in the soul, otherwise spells that let you possess other people would not function on account of you having nowhere to store your mind, but it also has to have a part in the body in the form of the brain, or you wouldn’t need a brain. Naturally, the mind has to be split into at least two broad parts, for the sake of making our thinking easier - conscious part, and a subconscious, emotional part, representing the parts of the mind affected by compulsion vs charm or fear spells. Physical world interactions are split into two parts, to emphasize that the soul is what is responsible for spell creation.
Alright, so we have a picture. So what? Well, let’s play around with the model a bit. First of all, we can now draw how a spell might act on the spell-body pairing:
What could this be? Well, suppose we have a spell that finds a soul at it’s target destination, hooks itself up to their mind, and sends pain into the mind in order to make their body spasm and drop something. This is how the effect chain of that spell could be represented on this model.
This model also immediately tells us broadly how we could protect ourselves from this spell:
We could prevent the spell from coming into contact with our soul
We could prevent the spell from then affecting our mind
We could prevent the mind from affecting the body afterwards
Each of these interventions can be represented as a red “wall” on the path of the spell. Effect chain can’t go through - spell is nullified. “Walls” in a practical situation would be protective spells of some form, such as Mind Blank to prevent the spell from affecting the mind. This further means that if we can’t draw an effect chain on the model of some creature, because their soul-body model is incompatible, that spell can’t affect them. At the same time, it also implies that defensive spells that counteract a specific spell by breaking this chain in some place mey be broadly applicable against all other spells whose effect chains go through the same spot.
Let’s dive a bit deeper into what I mean by incompatible soul-body models by talking about different creature types. Let’s start with golems:
Golems are, effectively, fantasy robots. Typically, they have strong physical bodies and are used for labor and as muscle. Depending on the lore, they may be animated and controlled by spells, bound spirits, encased souls, and so on. They have a key property that they are immune to all spells that allow spell resistance (effectively, almost all direct targeted spells), which makes them good at handling spellcasters in combat.
After consideration, it becomes clear that there are two parts to this immunity. First of all, their physical body is immune to magic: for example, Call Lightning would slam the golem with a lightning bolt, which the golem would completely ignore. Second, the animating force - whatever it is that contains their mind and issues commands to the body - is immune to magic too. Discussion of how their body may be immune to magic would have to wait until a post on warding, but our body-soul diagrams already give us a simple answer to the question of how their animating force may be immune to magic: it may just be buried behind layers of armour.
I’ve already said that spells, almost always, do not go through solid objects: this is evidenced by the Line of Effect concept. Furthermore, due to locality, spells would need to come into contact with the animating force of the golem to affect it. In most creatures, their souls are distributed through their bodies and are easily accessible through any point on their skin: this is why spells don’t limit which part of the body you can target to affect the entire creature. In golems though, this does not hold true: animating force is hidden entirely within the body, and thus isn’t accessible to spells. This is a very simple design decision that doesn’t introduce any unique properties to the golem animation spells, such as one-of-a-kind complete immunity to directed magic.
One corollary of this is that if you could access this animating force, you could affect it, and thus control or disable the golem. This is presumably a part of the maintenance or modification process by the manufacturer, and also an option for anyone that could make a large enough hole in the golem’s body.
Next creature type I’d like to discuss is undead. One notable feature of undead is that their bodies aren’t really alive - depending on the type, they may not have pulse, digestion, cell division and so on (e.g. zombies, vampires), not have any organs (e.g. skeletons) or even not have any bodies at all (spirits, shadows). As a result, their entire mind has to be contained within their soul: if you don’t have a brain, you can’t use it to think.
There are three consequences to this. First of all, spells that work on the brain, or through the brain, will not work on undead for fundamental reasons. For example, I would argue that Eagle’s Splendor, as a transmutation spell, has to work on the brain, and as such should fail for undead.
Even mental spells that would not run into this limitation would have fundamental issues with affecting undead, because their minds will simply be structured in a different manner: this is why undead are immune to mind affecting spells, unless those spells are specifically tailored to work on undead.
Second, when an undead possesses another undead, it’s hard to draw a line between the original owner of the body and the thief. Both of them will be connected to the body with soul-to-body links, so it would be simply a question of whose link is stronger. This is different to the situation of when someone possesses a live creature, as in that situation the brain will only be used by one of the souls. As a result, spells that affect the soul-to-body link should, generally, work equally well on the possessor and the original owner.
Mindless undead work somewhat differently - because they don’t have souls, they should be closer to golems, but without the inherent magic immunity.
Finally, since I brought up possession, let’s draw a diagram for two souls within a single body. Typically possessing someone also gives you a telepathic link to their mind, which is not shown here.
It becomes pretty clear that in order for someone to carry out the stereotypical possession scene, several things have to happen:
Possessor has to hook up their soul to the possessie body
They have to disable or modify the mind-to-body link so that original owner couldn’t control the body
They have to modify the soul-to-body link so that original owner can’t sneak any control that way
They have to isolate the possessie soul from the environment so that they couldn’t communicate with the outside world through spells or telepathy
If any of these are not done, then possession will have unique and interesting effects based on that. For example, if they don’t isolate the original owner’s soul from the environment, the original owner may be able to cast silent still spells (spells that don’t require speaking or movement to cast) or use previously cast spells that they have a mental link to.
In terms of the spells cast on the possessed person, they may affect one, both, or neither of the people there. For example, mental spells that go through the body would most likely affect the original owner, spells that damage all souls connected to the body will damage both, and mental spells that go through the soul may affect neither (because original owner soul is isolated, while possessor soul may be incompatible with that kind of spell)
Analytical abilities of soul-body models arren’t limited to analyzing different creature types. For example, we can even derive a novel classification of illusion spells. First, let us ask a question - what is an illusion spell? At their core, illusion spells are modifications of information flows in order to confuse their subject. Because illusion spells are defined by their purpose (i.e. why the caster is casting them: to confuse the target) they can’t really form a logical group based on their function (i.e. how do they work under the hood). I’ll talk about some ways in which this may not be true later, but keep this in mind for now.
What do I mean by an information flow? You look at your shoes: light strikes them, is reflected into your eyes, gets processed into nerve signals that are delivered to your brain, where you construct an image of the shoe and make decisions based on it. This flow of information (shoe -> eye - > nerves -> brain -> consciousness) is very easy to represent on our model. Illusion spells modify some link in this chain, and make you see something else instead of a pair of shoes.
We can thus group effects of illusion spells into 4 mutually exclusive categories, corresponding to red breaks in the information flow on our diagram.
The first of them is what I call “Stimulus modification” (StM on the diagram): spells that modify a stimulus, that is to say actual physical light, sound, vibrations, or something else. Key property here is that it’s impossible to resist these spells: once light is emitted, it can’t be un-emitted. The only thing a target of these spells can do is disbelieve the newly created image within the context of the surroundings: this is a fundamentally logical, and not magical, process. If you see a dragon inside a room it couldn’t logically enter, you can deduce the dragon is a hologram; but that doesn’t make the dragon transparent in any way. There is no particular save against these spells.
Second group is “Sensor modification” (SM on the diagram): spells that modify your sensors, such as eyes, ears, or your connection to other spells. They make your eyes report different information from the light coming into them. Two key properties arise out of this. First, because these spells modify sensors that are on the “outside” of your soul and body, they are not mind affecting. In much the same way that wearing glasses would not be a mind affecting effect, having your eyes lie to you is not one either. As such, it may not be appropriate to use a will-based save for these spells: physical resistances to modification would be much more fitting. Second key property is that this modification would apply to anyone else using your eyes: a ride-along possession, someone you are sending information to over a telepathic link, whatever. Naturally, they would not receive a second save, because the initial save is for your eyes to avoid being tampered with. But once they have been tampered with, there is no recovering “true” information by an act of will.
Third group is “Qualia modification” (QM on the diagram). These are spells that modify the information flow at some point between the sensor and your conscious mind. This may mean modifying electrical signals within the nerves, or modifying the parts of your brain responsible for crude, pre-conscious processing of sensory information. By their nature, most of these spells are mind-affecting. They also would not affect anyone sharing your eyes, because they normally would have a different information pipeline from you. If you make your save on these spells, then presumably your soul was strong enough to fight the entire effect off; but if you fail, you can still choose to not trust your senses: after all, these spells just modify what information your conscious mind has access to, but can’t make you believe anything.
Fourth and final group is “Mental modification” (MM on the diagram). These spells go into your mind and actively work to confuse you. This may mean memory erasure in real time, things that screw with your conscious awareness, ability to recognize certain objects or faces, and so on. This kind of spell lies on a very smooth spectrum leading to direct mind control: the only key difference here is in the goals and ability of the caster. Illusion spells are meant to confuse and misdirect you, while mind control subverts your decision making entirely. All of these spells are mind affecting. Save against these spells is thus to avoid your mind being tampered with, but once you fail that, the attacker has almost arbitrary control over you.
It is important to mention that because of the purpose of illusion spells, effectiveness of all of them will depend on how plausible the victim finds new information being received by them. For example, while there is no save against stimulus modification spells, your character may still deduce the new stimulus is an illusion through logical thinking. Obviously, the caster would try to make the illusion as plausible as possible, while the victim would try to find anything out of place about it. To simulate this, I would suggest using some kind of intelligence roll appropriate to the situation instead of a save: perhaps an opposed roll, with DC for the “save” being equal to the roll of the caster’s intelligence, with take 10 and take 20 rules applying as normal, and with feats increasing DCs of illusion spells giving a bonus on the roll, and overall plausibility of the situation modifying the roll as appropriate. If the victim fails this check, they believe the illusion until they receive new information to the contrary.
I think this grouping of illusion spells is much more sensible compared to the classic DnD one of splitting illusion spells into figments, glamours, patterns, phantasms and shadows. Just for fun, let’s talk about it:
Glamour & Figment: The main difference between these two is that a figment “cannot make something seem to be something else”, whereas glamours can. But of course that’s a trap rule: the kind that will inevitably cause the GM to have headaches at the table. Figments can clearly produce visible images, and so at the very least can make something look like “something plus an image”. How about making a mundane piece of armour appear to be enchanted to show illusory figments - is that “something else”? It’s pretty clear to me that what is implied by the rule is a very human-centric definition of “something else” that relies, inherently, on the ability of these spells to deceive others. But that would mean that inherent ability of this school of magic depends on how smart your players are and who is observing them: this is almost inherently anti-reductionist, and will inevitably lead to confusion. My classification avoids this entire class of problems entirely, by not pretending that you can simultaneously have the ability to affect light and shadow and not have the ability to make things appear different.
Pattern: In my classification, this would be a combination spell: one part stimulus modification, one part one of the other three groups. This would mean that one part of the spell can be resisted, while the other is going to happen no matter what. This avoids a whole host of issues; for example, if someone sees this and succeeds on their save, can they still see the lights? Can you use this to send at least 1 bit of information to an accomplice (by agreeing to cast the spell if you notice something), or are you all out of luck if they have a high will save?
Phantasm: My classification splits this group of spells into three, and as such automatically avoids a whole host of issues with deciding how they work in edge cases.
Shadow: These are not illusion spells and should not be treated like they are. These are conjuration spells that work with the plane of shadow! They conjure objects or creatures made from the energies of that plane, and those conjurations should work like actual bloody conjurations. This is the only internally consistent way to treat them. Otherwise you will hit a myriad of nonsensical contradictions. This guy likes the shadow spells and even he admits they are a headache! Dealing with magical placebo/nocebo effects that can cause impossible to resolve inconsistencies whenever more than 2 people are present in the room is an entirely unnecessary headache: just say no. And don’t get me started on crunch letting you have 140% real illusions - as in, if you believe the effect is not real, you suffer more damage. This is not to say that spells can’t damage your body through your mind - just that shadow line of spells goes about this through the worst way possible.
My grouping also pretty cleanly resolves the question of what counts as “interaction” with an illusion - often, if you “interact” with an illusion in DnD, you are allowed a second (or first) save against the effect. What is meant by “interaction” is a famous source of headaches for GMs in DnD, because of how inherently vague that term is. For example, if I cast an Illusory Wall, then the spell explicitly notes that prodding at the wall reveals the true nature of the surface. Otherwise, the wall appears real, unless “interacted with”. What if I see someone go through the illusory wall - is that an interaction? I didn’t interact with it, but I should hope any reasonable person would realize the wall is fake in that situation. What if I see torches near the illusory wall moving weirdly, as if there is a draft - that doesn’t seem like an interaction, so am I cursed to believe the wall is real even then? Would trying to prod at the wall be metagaming in this situation?
Similar issues arise with other illusion spells. In my system, there is no concept of interaction - instead, any new information that tells your character the wall may be an illusion would potentially allow you to make a new intelligence check to realize this fact. And, naturally, you can always be paranoid and assume everything can be an illusion until proven otherwise - your character would hit a lot of false positives (because most walls are not, in fact, illusory), but that is an internally consistent way to behave given the information available to them. Assuming something is an illusion is only metagaming when there is no evidence to support this deduction, even weak, false, indirect or coincidental evidence, including any reasoning made by the characters based on the circumstances or motives of any of their enemies, and the assumption isn’t made as part of a sort of Standard Operating Procedure where the character accepts their judgement will be wrong 99% of the time because the 1% of the time they are right the benefit is large enough to outweigh the inconvenience, and instead it’s pretty clear the conclusion is mostly based on out-of-game information.
I think it’s pretty clear that drawing these simple body-soul diagrams allows us to clear up a lot of confusion pretty quickly. I hope this trick can be useful to other people in your thinking about DnD effects. Now that we have a structure for our souls, next time I will be talking about the structure for the spells.
Meanwhile, if you enjoy what I write, you can subscribe to receive updates by email: