I, as well as many people in some of the groups I play with, have come to the conclusion alignment is a clunky horrible mess. I personally don't like the way it oversimplifies morality and makes everyone choose a side; even if that side is neutral. Outside of Alignment having issues with edge cases, it also doesn't easily let in morally grey settings.

I would like to remove it from the games I play, but alignment has some nasty mechanical hooks in spells like Protection from Evil or Holy Word as well as magic items, planar traits, and the like. My questions is this: How do I remove alignment from D&D while keeping alignment-dependent spells/items/effects mostly intact in a mechanical sense?


Some options I have used in various games that I’ve liked:

Have purely-nominal alignment

Characters each have an alignment, but it doesn’t mean anything except for whether detect evil pings for you. Super-simple, pretty consistent, but can lead to fairly arbitrary effects.

As one friend put it,

Alignment doesn’t affect anything but the color of your lightsaber.

When you do this, it does not make much sense to enforce alignment-based requirements and prerequisites. Then again, I’d argue that’s close to the default case anyway...

Redefine Alignment

Instead of eliminating alignment, redefine it. I’m a fan of alignment as allegiance; I’ve played a number of campaigns where it’s worked very well. Good alignment becomes allegiance to Celestia, or the kingdoms of man, or the orcish empire for that matter. Evil alignment becomes whichever side is opposed to the previous side. And the same for Law and Chaos. These sides can have as little or as much association with “good” and “evil” as you like (though it’s probably easier to have such associations).

For a game like Planescape, this is very easy. You are Good because you worship a Good god, who is a Good god because his realm is on a Good plane, even if neither you nor the god are particularly good, personally. You are on the side of the Upper Planes and that’s that.

For settings that don’t have such obvious sides, it’s worked best if “Lawful” were the “winners who wrote history.” As an example, this is an adaptation of a game I played in, using Eberron since it’s an established setting that saves me having to explain some of the details and factions of the custom setting.

The human kingdoms in this Eberron game aren’t “Lawful” because of their ethics, but because the Five Nations, and before them the Kingdom of Galifar, have dominated the culture of Khorvaire and have defined themselves as “the Law.” They also label things like cults to the Dragon Below, the attempted revival of the Dhakaani goblin empire, and the monster “nation” of Darguun as “Chaotic” – not so much because of their ethics but because of their threat to the dominance of humanity, which they have already labeled as “Lawful.” “Good” and “Evil” apply largely to individuals rather than organizations: they were whether or not you worked to the benefit of your side even if it is to your detriment (Good), or worked for yourself even if it hurt your own side (Evil). And the overwhelming majority of people are True Neutral because they live peaceably within Khorvaire’s established culture and society, neither actively reinforcing it nor acting to undo it, and are mostly out for themselves (but will stop before really harming those they consider to be on their side) and will try to help their side (so long as it doesn’t put them out too much).

Ban Alignment-based effects

Spells like detect evil and holy word no longer exist. Simple enough, but problematic if people are dead-set on playing alignment-based classes (Paladin being the biggest candidate). Then again, in most cases the class is either more than powerful enough even without these effects (Clerics can do just fine without blasphemy, which is overpowered to begin with), or so weak that opening these effects up more broadly won’t break anything (Paladins that can Smite Anything still aren’t all that impressive).

The biggest problem with this, honestly, is likely to be monsters who use these effects. A pit fiend just isn’t the same threat without at-will blasphemy. Consider changing these sorts of spell-like abilities to be racially-based: the pit fiend gets an at-will ability à la blasphemy that works on non-devils, etc. Considerable power upgrade, though it may not matter much.

  • 3
    \$\begingroup\$ I think a Q&A about redefining alignment as allegiance might be cool. It sounds like a good idea but raises a lot of questions: Is there a single detect allegiance spell which reveals what leader/group a creature is loyal to? How are conflicting allegiances handled mechanically? If you're loyal to Banjthulhu the Puppet Terror, who exactly do your smite powers work on? \$\endgroup\$ – BESW Sep 24 '13 at 23:47
  • 11
    \$\begingroup\$ @BESW In our play, we still used the alignment labels (and the associated spells), they just didn't mean “you are good” so much as “you have allied yourself with the Good side, which is maybe not all that good in reality.” Particularly in settings where “Good” could be applied to the “winners who wrote history,” so to speak. \$\endgroup\$ – KRyan Sep 25 '13 at 1:13
  • 5
    \$\begingroup\$ Allegiance (specifically, allegiance on the scale of cosmic Law/Good/Chaos/Evil) is actually what alignment meant in D&D in the first place. (Which explains why it has the weird name of "alignment".) Later editions' downplaying and deviating from that origin is the source of all problems with alignment. \$\endgroup\$ – SevenSidedDie Sep 25 '13 at 17:05
  • \$\begingroup\$ I've used the "Redefine Alignment" approach in the past with great success. \$\endgroup\$ – Cypher Nov 6 '13 at 2:04
  • \$\begingroup\$ As you write it, it looks like you consider "lawful"="respecting local law", while it should be more of a mindset, not specific. \$\endgroup\$ – o0'. Jun 23 '14 at 19:48

Here's a really easy approach.

Remove alignment for mortals. Keep alignment subtypes for planar beings.

How do you tell which is which? That's easy. Any creature with an alignment subtype in its type line still has alignment. Any creature without one is functionally unaligned.

Additionally, consider adding alignment types for undead. This'll keep a few abilities like Detect Evil (which is kinda painfully ambiguous about whether it detects evil undead or undead in general in D&D3.5 anyway) a bit more useful.

So, what changes? Well, generally speaking, abilities that hit all Evil beings were already rather overly broad in terms of their practical applications in a typical adventure; dialing down their use a bit isn't going to be a big problem.

  • Spells and items: the changes don't matter.

    A few spells and items will become a bit less powerful since they no longer hit the random evil wizard or feisty orc. That's okay! D&D3.x clerics, for example, are a powerhouse class with free access to their entire spell list; narrowing the scope of a couple of the spells in their toolbox isn't going to significantly disadvantage the class.

  • Fixing up the paladin.

    Paladins are a class written around the very idea of alignment, so they'll need some changes to avoid ripping away abilities from what's already a pretty middling class.

    • The most important thing is the paladin's Smite Evil ability, since it's a major source of damage for the class. Just make it work on any target; it's not going to overpower the paladin compared to a similarly optimized martial character of a different class.

    • All of the other class abilities are pretty much fine as is. Aura of Good and Detect Evil can stay the same. Removing alignment from random monsters weakens Detect Evil a bit, but it's still a good "radar" if you allow it to find undead.

    • The spell list may be a little trickier. In both D&D3.5 and Pathfinder, Paladins have small spell lists and a lot of their spells are focused around beating on evil things. These spells become narrower (note that they are already kinda hit-or-miss already, since a lot of D&D opponents are Neutral monsters and the like).

    The big takeaway here is that you don't have to reinvent the class, but keep an eye on paladin characters in play to make sure they've got something to do with their spells. And definitely broaden Smite Evil.

Whatever you do, keep your house rules very simple so that players can easily understand how to apply them to the game text at-a-glance.

  • 3
    \$\begingroup\$ Agree with this approach, might add undead in too though. \$\endgroup\$ – mxyzplk - SE stop being evil Sep 24 '13 at 22:52
  • 4
    \$\begingroup\$ @BESW How so? I mean, if you never face a demon (or summoned monster, I suppose), then circle of protection isn't a very good spell. That's not really a big deal, though, any more than a couple of cold-damage spells not being that great in a monsters-of-the-far-north-themed campaign. \$\endgroup\$ – Alex P Sep 24 '13 at 23:24
  • 5
    \$\begingroup\$ @AlexP Your hack makes holy word useless against a rampaging horde of goblins. Protection from evil ought to provide +2 to AC and saves against a cabal of humans intent on summoning Orcus--but it won't in your hack. If evil non-planar creatures are common in a campaign, this modification has wide-spread ramifications that I think your answer downplays a bit too casually. \$\endgroup\$ – BESW Sep 24 '13 at 23:32
  • 3
    \$\begingroup\$ @BESW All around good points, but you are stating the obvious and I don't believe all answers need to state the obvious. Players should know about this modification before play, which means spell casters will only choose to use or rely on spells that target particular alignments when those spell will actually be useful. Since protection from evil won't do anything against a cabal of Orcus worshipers, then the players won't be using it. \$\endgroup\$ – Cypher Nov 4 '13 at 21:32
  • 2
    \$\begingroup\$ @BESW It seems like you're over-complicating things. This solution is much more simple than you're letting on. \$\endgroup\$ – Cypher Nov 6 '13 at 2:07

What I have done with some success (though I have played with it more in AD&D than DnD3.5) is to have alignment reflect the characters religious alignment. The deities have all chosen to be "Good" or "Evil" and "Lawful" or "Chaotic" and a serious follower of a deity picks up their alignment for all magical purposes. Someone who follows no deity has no alignment.

I find this works reasonably well since most of the times that alignment comes up it is in some way connected to the religion of the setting.

However, this works best in a setting with clearly defined deities that are at least somewhat active, such as the Forgotten Realms (which is where most of my experience comes from). It does not work so well if you try to downplay the deities.

  • 1
    \$\begingroup\$ I usually run something similar to this. Basically, I have the gods be the arbiters of alignment, even for people who aren't their followers. People detect as 'good' because they're philosophically similar to the 'good' deities, etc. This way, alignments can be as arbitrary as I want them to be and still make sense in-game. \$\endgroup\$ – DuckTapeAl Sep 25 '13 at 6:10

It's What You're Doing, Not Who You Are

So alignment isn't a steadfast representation of who you are as a person. But individual acts, those can be assigned an alignment. Is there a ruthless villain in the room plotting the destruction of the village? Probably picks up from detect evil. If he's just there for a pint, maybe it won't. Is this goblin actively attacking the village? The paladin can smite it to her heart's content. But if the goblin is just sitting there, petting a puppy? Smite evil probably isn't going to work. There are still plenty of typed creatures that are personifications of their alignment, and undead and deathless creatures are necessarily animated by evil and good magic, respectively. But mortals? Their alignment is all about what they're doing and why.

Now there's some gray area about intent, but remember that you are the DM. You know exactly if that goblin is harmless or if it is petting a puppy as part of some nefarious purpose, so you know whether or not smite evil will work on it. Yes, it makes things murkier, but what is the point of removing alignment if not to make mortal actions and intent murkier?

Edit: Obviously, undertaking such a radical shift while relying on subjective interpretations of in-the-moment alignment choices requires a strong social contract between players and DM. The DM needs to trust that the players will not seek loopholes to abuse the system, while the players need to trust the DM to be final arbiter when it comes to alignment. After all, it is the DM and only the DM's responsibility to determine the moral boundaries of the world they have created. Take the time to work with your players to establish, as firmly as possible, where those boundaries lay, but in the places where the gray areas exist, the only judge can be yourself.

  • 4
    \$\begingroup\$ This is open to some pretty epic abuse, though. "I ready an action to think about kittens" is right up there with carrying around an inch-thick lead sheet to prevent alignment detection. \$\endgroup\$ – BESW Sep 24 '13 at 23:49
  • 1
    \$\begingroup\$ @ BESW As a DM, I would argue if someone is allowing these kind of loopholes in their system, there is something wrong. Respectful players should have little problem with dropping "kitten thoughts". Furthermore, thinking about one thing and intending another argues, IMHO, that the intent (to be evil in this case) is actually stronger because you are attempting to hide it willfully. \$\endgroup\$ – Anaksunaman Nov 8 '13 at 7:35

One way that I've seen this kind of thing run pretty successfully is that alignments are all relative to the person using the alignment-based effect. Basically, you replace Holy Word and it's ilk with Word of [Deity] that works exactly the same, but has detrimental effects against your deity's enemies rather than basing it on alignment. Detect evil will detect people just like it normally does, but it instead detects people diametrically opposed to your personal belief system.

This has the added benefit of allowing "evil" paladins. If I worship the god of murder, then I detect fellow murderers as "good", and town guards as "evil", effectively.

This requires some trust and a decent social contract with your players. Basically, they need to trust you when they use an alignment based spell that you know what the target's beliefs are, and whether or not they are opposed to the player.

  • 3
    \$\begingroup\$ Poking this with a stick to see how it jiggles: If I'm a priest of Boccob the Uncaring and I cast holy word, does it have little effect on anyone, or does it hit everybody? \$\endgroup\$ – BESW Sep 25 '13 at 6:25
  • \$\begingroup\$ @BESW: I'd guess it hits either people who aren't neutral, or people who aren't magi, or people who are both, or one of either. \$\endgroup\$ – Aesin Sep 25 '13 at 11:41
  • 4
    \$\begingroup\$ It would hit anyone that is opposing the cause of Boccob, that is, anyone who opposes the spread and research of magic. More generally, it would hit anyone who threatens the church of Boccob. The intent is to keep it roughly as broad as Holy Word is by default while keeping the logic consistent with a realistic morality. For example, if a warlord is trying to pillage and burn a countryside that contains one or more Boccobite temples, he would be affected as if he were evil by Boccobite spells. \$\endgroup\$ – DuckTapeAl Sep 26 '13 at 1:13
  • \$\begingroup\$ This is an interesting option since it reflects the reality of our various points of view. \$\endgroup\$ – Cypher Nov 5 '13 at 2:52


An option used in a few MUDs I've seen, and also in a few RPGs like Legend of the Five Rings (L5R), is to make these things numerical and make people move around. If I can suggest a mechanic, you could, for example, make Good/Evil one axis, and Lawful/Chaotic another.

If you do something Good, and you gain some Goodness score, if you do something Evil, you lose Goodness/gain Evilness. The more Good/Evil you are already, the less doing something Good/Evil moves you up in Goodness/Evilness.

You could define Lawfulness/Chaoticosity the same way -- so, for example, fulfilling a bargain gives you more Lawfulness, while going back on one, or breaking a law in a town (whether for good or evil intent) gives you Chaoticosity.

Mechanical Tweaks

Once you've done that, you've got two options.

You can define two cut-offs, above which you're Good and below which you're Evil (and between which you're Neutral), and the same for the Lawful/Chaotic axis.

Alternatively, you can decide to tweak spell effects to make them able to work off smaller differences -- so, for example, Detect Evil will give you a rough idea of how evil someone is, and spells that only have effects on Evil or Good things might have lesser effects if the targets are less Evil or less Good. Obviously this might involve a fair bit of tweaking to get right.

(This is, I guess, a numerical version of agradine's answer.)

  • 1
    \$\begingroup\$ How does discretization address removing alignment? Is the idea just that you no longer have to make judgement calls on a character's overall disposition? \$\endgroup\$ – Alex P Sep 25 '13 at 13:52
  • \$\begingroup\$ @AlexP: The idea is that you're removing the current immutable binary labels, while still keeping something you can drop into the existing mechanics that run off it. You can start with a certain position on the scales and let in-play events define your actual alignment naturally. It's not perfect, but the conceit of most of the settings that use the system (AFAIK) is that there is an external force judging your actions. I actually prefer DuckTapeal's approach, but this is a bit simpler. \$\endgroup\$ – Aesin Sep 25 '13 at 14:02

Alignment is one of my pet peeves of the D&D family. They had a workable version of alignment in 2e, unfortunately when they brought in the Barbarian and Monk classes, they totally killed the law/chaos axle. One of the iconic examples of the monk class has to be Cain from the TV Show Kung Fu. Based on my look at him, he is Chaotic Good (completely out of step with society, but works for the betterment of all humanity). Also, Barbarians are only chaotic while in rage. If he's going hunting he is likely to be just as ordered/disciplined as a Monk.

I have run in the past that if your class does not require an alignment, you need not even pick an alignment as long as your final character is willing to work with the rest of the team. I don't mind pc/pc conflict, in fact I think it makes the game more fun (although PvP fighting is frowned upon). As for the alignment-restricted players, I sat down with them and went over the deity they followed and wholeheartedly stole from the World of Darkness' hierarchy of sin which has a rating from 1 to 10 where 1 is something ANYONE with that personal moral code would consider just plain wrong and 10 is something that few would consider wrong other than the most pure. IIRC, Vampire's default for most of humanity was 7 on the scale, so I told them they had to live at or above an 8 on the scale. I think in my next game I will run something similar for all characters as it allows you to do truly horrible things, yet also gets out of the way when you are acting within your own code.