So the spell Harm states that it cannot reduce a target below 1 hp. However Shield Other takes half the damage a target receives. If Fighter X Takes 140 damage of harm, and so really only takes 70, and Cleric Y takes the other 70 damage. If 70 damage would be enough to kill them, would they die? or be reduced to 1?
So this question comes down to a number of things that are not explicitly defined in the rules:
Does harm’s limit on how far someone’s hp can change actually count as changing either the damage it deals or the damage its target takes?
What is the correct order of operations here: does shield other apply first, applying half of the whole damage to each, or does the limitation on harm apply first, applying half of the damage necessary to drop the original target to 1?
Does the limitation on dropping the target below 1 transfer to the caster of shield other?
From a rules-as-written perspective (which has some issues here as I’ll get to), the answer to 1. is probably no, the limitation on harm doesn’t change the damage dealt, just the final value of the hp after deducting that damage. So, for example, if harm deals 150 damage to a 50-hp character, you end up with 50 − 150 = 1. The 150 itself isn’t actually changed; there is no notion of harm’s damage being 49 even though that is ultimately the damage dealt. And since shield other only sees damage, you would end up with 50 − 75 = 1, and 75 damage dealt to the caster of shield other. And, as noted by the dark wanderer’s answer, the limitation on harm is applied to harm’s target, and the caster of shield other is not a target of harm.
OK, so that’s our rules-as-written answer. What’s the issue? There are a few, but the big one is this weird discrepancy between the damage that harm’s target “takes” and the actual change in their hp. Is it really justified to say the target takes 150 damage when they would never actually lose more than 49 hp? The answer is no, it’s not really. The rules, in many places, ignore damage that doesn’t actually get through to hp, whether due to DR or resistance, or whatever. Those things actually modify damage—rather than final hp—but the difference seems extremely nit-picky to me. I might—maybe—buy that in a theoretical optimization exercise, but I might also buy the opposite.
So if we say that the limitation on harm actually changes its damage, our answer to 1. is “yes,” what does that mean? Well, it means we have to stop ignoring 2. and 3.
For #2, it makes the most sense to me that, generally speaking, spells are atomic. The rules as written seem to back this up, anyway. By “atomic” I mean that they happen all at once and can’t be split up (atomos is Greek for “indivisible,” which atoms were believed to be prior to the 20th century). That means that the limitation on harm is applied before shield other comes into play: the 50-hp character would take 49 damage, so shield other says they take half that (24 damage), and shield other’s caster takes the rest (25 damage). Then to 3. we have already answered: from a rules-as-written perspective, there is no justification for the limitation to apply, so if the caster of shield other has 25 hp or less, harm will do worse to them than it could have if they had been targeted directly.
But as I said, the rules-as-written perspective has a few issues. Question 1. is the most egregious, but even if we try alternate answers to 1., the answers to 2. and 3. seem... unsatisfying, to me. Out of keeping with what harm is, and is trying to do. Personally, the impression I get from harm—the thing I think the designers were going for—is that it is a massively dangerous, damaging spell, but one that cannot kill. Hence the name. Both reducing 150 damage to a measly 25 and the possibility that this can wind up killing the caster of shield other seem out of keeping with that. I can see my way to justifying both narratively, but I don’t like it.
So I think I would actually rule quite differently from the rules as written here: I would apply the 1-hp limitation of harm after shield other divides the damage, not before, and I would apply it to both the target and the caster of shield other.
Where there are gaps in the rules I tend to look to try to honour the intent of the actions. Shield Other is very similar to Fate Link and I think both attempt to have two creatures follow one thread of fate (temporarily).
I see the intent of the caster of Shield Other to be to suffer alongside the target of their spell/prayer such that their target can endure more punishment. There are perhaps two flavours to the use of Shield Other:
- when the caster has more hp, and
- when the target has more hp.
In the first flavour the intent is to keep the target up as long as possible but not kill the caster of Shield Other. (Should the target hit -CON and die they cease being a creature and the spell should end. So provided the caster has at least the target's CON in extra hp they'll survive.) This would be useful for protecting low hp targets like mages and rogues, but stop short of taking out the party's healer. So this is an attrition addition to the combat.
The second flavour is basically martyrdom, where the caster is going to risk their own life to keep the target up. This is more of an at-all-costs addition to the combat. Though assuming the caster of the Shield Other can spontaneously use cure spells they can somewhat mitigate the max hp disparity from one round to the next.
As for the caster of Harm, I see their intent to be to do so much damage to their target as to make them almost fall down unconscious regardless of their current health, but not actually kill them.
Trying to get everyone's intent to work together I would DM a more entwined interaction between the two spells. Have the Harm apply equal hp damage to each participant in the Shield Other to bring the first of the two to 1hp.
Eg. Oweran the Cleric prays Shield Other on Gregor the Barbarian, Oweran has 35hp and Gregor has 65hp. Kazan casts Harm on Gregor. 34hp would reduce Oweran to 1hp which would meet the terms of the Harm spell (shared fate reasoning that they are kind of one target for hp loss). Gregor would then end up on 31hp and likely Kazan would be all too aware of something hinky going on between the two.
This way the Shield Other had the intended effect, and the Harm merely redirected its intended effect. Gregor was Shielded, Oweran is sharing the suffering/damage, someone was Harmed (reduced to 1hp), Kazan was partially twarted but Gregor is still hurt by the attack so Kazan still comes out successful and more than likely knows Gregor and Oweran are linked now.
Trying to spread the wins all around. :)