It’s not the theft of 2 GP that turns the fighter’s alignment in the example; it’s the betrayal of a trust or a vow. (I think they were referring to her Lawful alignment, not her Good alignment.) And I think Troy made the right call about a ‘pattern of behavior’ vs. a single act, but this is why these things should be talked about before the game starts.

Killing someone trying to steal from you is a world away from killing a random bum on a street corner (which is why it’s legal to shoot an intruder in your house in most states).

A soldier typically engages military targets; an assassin typically engages political or civilian targets (he says, having worn a uniform). Aside from that, the “evil” of an assassin is a value judgment, not an absolute. If an assassin punched Hitler’s ticket in 1942, would that be evil?

Alignment was borrowed from a number of sources; Michael Moorcock’s books in particular had law and chaos themes (and are the source for the symbol of chaos).

When it comes to alignment, the mistake I see gamers make over and over is to use “evil” to describe “different”. There’s a big difference between the “us and them” of the real world and the “objective evil” of D&D. Native Americans might be considered evil by pioneers (and vice-versa), but both would agree that the demon who tears the skin off of women and children just to hear their screams, is Evil. And if you read through the various editions of the game, that’s what they’re talking about.

Alignment in D&D can be a pain, if you let it. It should be a standard of measurement, nothing more. It definitely should not be a blunt object with which to beat your players (or, as in the example, to allow them to beat each other with). If you are going to use some kind of alignment system in your game, talk about it before you do, so you can avoid these nasty surprises.

So you can do whatever you want. What are the consequences? D&D has none. You could be excomunicated by your deity or by some guildmaster for not following doctrine, but that’s hardly alignment. The only penalty for not following your alignment is that it might change. And that you wouldn’t like being of the new alignment because of the way alignments and magic interact. Your best magical goodies require you to be good, and good people who you can trust won’t deal with someone who pings as Chaotic Evil.

I use the “Multiply it by a Thousand” rule for a lot of stuff. What happens when you multiply the effect of something by a thousand? What if the player casts this spell ten times a day? Etc.

In this case, if a breach of trust and theft of 2 GP would cause an alignment change, then stealing 20 GP should put a saint straight to hell. And that’s just ridiculous.

So let’s use the example of a murder. A True Neutral PC kills some bum in town. Most of you might be apalled and demand some penalty. But what about when that same character kills a pickpocket trying to steal from him? Is that somehow less evil?

If an assassin (a man paid to kill) is evil, why is a soldier (a man paid to kill) any different? Is killing for personal pleasure better or worse?

We’ve probably all heard the argument of two Good or two Evil nations fighting. England vs. France, anyone? I think that’s a bit more complicated and it’s not possible to say that one side is more righteous than the other.

The D&D alignment system doesn’t work because even the simplified environment of the fantasy game is more complex than that. If you let alignment boil down to “Good is us and everything else in our way is Evil” then you’re left with a valueless rule.

The alignment system appears to have been taken from the works of Poul Anderson, where the Chaotic side represented inhumanity and the Faerie, and paganism and magic and the Old Ways. The Lawful side was taken by humanity, Christianity, and a dispersal of the veil of Faerie. In that regard alignment was less a personal choice as an inherent nature – something that would be impossible to change. You either are a Troll or you are a Human.

But again, that reduces alignment to “we are good and they are evil, let’s kill them all and take their stuff” and “if they live in a town they’re off-limits but if they live underground we can kill them”. These are not logical ethical positions, they’re unthinking dogma.

By the time a game master is mature enough to ask such questions, he’s mature enough to discard the alignment crutch and play a more nuanced game, filled with characters driven by complex and interesting motives, where no one is pure good, and evil is a description rather than a motivation.

And believe it or not, when I was a kid, I actually _was_ called out by the DM once for a simple act of kindness as a druidic “alignment violation”. That was the day I started hating alignment systems.

I for one will be happy when alignment goes the way of the dodo. I seem to recall a lot of people grumbling about various alignment related shifts from 2e to 3e as well.

4e thankfully makes alignment talk about as relevant as your character’s name, just something you put on a sheet. Given a few years when 3.5 finally starts to fade out, only the grognards will need to worry about it.

@valadil – I think your point that evil characters shouldn’t necessarily realize that they’re evil is an excellent one, and one that doesn’t get made nearly frequently enough. As somebody who sees the alignment system as a rough approximation of people’s actual moral codes, and not as the determining factor of every one of a character’s actions, I tire very quickly of seeing people making decisions that are the moral equivalent of burning down an orphanage with the excuse of: “I’m evil”.

I actually recently threw a scale system for alignments into my game, to throw a little bit more customization into things — and hopefully get players a little bit more into the mindset that an alignment is something that you pick to briefly describe your character’s moral code, not something that you pick to determine your character’s moral code.

I’d like to weigh in on the idea of a sliding scale, though. I absolutely agree that a character’s alignment is something that should be free to shift as the game progresses, but I think that the idea of assigning point values to specific acts is coming at the issue from the wrong side. If a character suddenly steals 2 gold from the party (or lets an innocent bystander get killed in order to defeat a greater for, or whatever), that’s not a sign that their alignment should suddenly change on that scale. That’s a sign that their alignment has been shifting for a while. I tend to think that it’s not so much the average of a character’s actions that determines their alignment. A person who occasionally likes to go out and , but otherwise is a pretty good guy would still be “evil”. Those are the sorts of people that parents warn their kids about. I think on a more precise alignment scale (say, 3 to 18) the fighter’s alignment would have changed slightly (from 16 to 15, perhaps), but probably not enough to be noticed on a Lawful / Neutral / Chaotic scale — but I don’t think it’s the act of stealing the gold that would have changed her alignment; the years of poverty were what changed her. That being said, the player might not necessarily be aware that there was that attitude shift until the event happened, in which case you’re back to having to look at things at the time of the incident. I still think, though, that it’s important to keep in mind that it’s not the act of stealing itself that determines alignment: it’s a willingness to steal.

With this outlook, perhaps the character’s ethical limits had begun to shift without her knowledge, and finding that she had succumbed to a temptation that she would’ve been able to walk away from years earlier is what brings her attention to that fact (or perhaps it’s something that she’s been aware of for a while). I definitely agree that the final call on motives should come from the player, though. If the player wants to play her character as feeling remorse for the act, there’s no reason that the act can’t serve as nothing more than a warning to the character that she’s begun to lose some of her values, and needs to be more strict with herself.

In the end, though, it should primarily be left to the player to decide what their alignment is, and when it’s changing. Some players will need more help with this than others, but unless the player shows an unwillingness to look at their character’s alignment subjectively, I don’t really see any reason to make it more than just advice. Obviously no two people at the table will ever fully agree on alignment issues (which is probably because no two people around any given table have exactly the same alignment themselves), but that’s what makes it so much fun.

If something of this type has to be tracked, I’d prefer it to be reputation, not alignment. Alignment is implied to be observed and measured by an infallible, omniscient source… and real people and characters don’t conform to alignments in any way that is easily tracked by an x-y graph. Reputation, on the other hand, is only affected by what a character is observed doing, and then only depending on how the observer views the action. Average that out over the general population, and account for rumors and word of mouth, and you’ve got an uncomplicated metric that can believably be defined by a series of two-dimensional points.

Of course, if that’s not what you want, yeah, 3.5 and earlier’s alignment system is a frustrating obstacle. And if you just jettison it, you also lose a chunk of spells (making clerics and wizards less powerful) and magical items (making PCs less powerful).

If you do want to keep an alignment system, you probably need to discuss the general scope of things up front. What makes someone good versus evil? What makes someone chaotic versus lawful? What might cause an enforced alignment change? The interpretations of the rules are so varied that the likelihood for misunderstanding is high.

Of course, off the top of my head, I can’t think of another class that had similar alignment restrictions. I have a vague thought that maybe the Druid used to require the character to be neutral on the law/chaos scale, but I’m not sure I’m right.

Okay, okay, I guess there is still a few things to consider. In general, as a GM, I tend to treat all characters as generally neutral with tendencies as listed on their character sheet which is a role playing effect. The special powers/defenses/etc. only really kick in when someone is obviously good (Paladin swearing an oath to the Angelic Kingdom) or evil (Necromancer summoning demons to kill people for mayhem and power). So, in a sense, for the spells really to matter, the players have to be ‘touched’ by the forces of evil, good, law, chaos, etc. Trust me, when I have a player using ancient demonic devices, they’ll know when they’ve a member of the dark side. Until then, its all meaningless shades of grey.

One other thing — if the player stole the money AND got away with AND kept stealing in the future, the player has changed and their role playing alignment changed itself — all without of the need of the GM to do, well, anything.

So… Does that help?

You have to ask yourself(ves): So what if her alignment changes? Does it alter gameplay? Does it alter what she did? Does it alter how she’ll act in the future? I don’t see the answer to any of those questions as being “yes.” She swiped 2 gold and is feeling remorse. Perhaps she’ll portray her character as slowly sliding down the slippery slope of rationalized morality… or perhaps she’ll be extra righteous and seek to redeem what the player sees as a shameful act for her character. Perhaps the other players will use their characters to tempt her further or seek to help her redemption. Awesome!

Alignment X or Y doesn’t really matter. How the player has fun with it and how the game flows from it is what matters.

I don’t think alignment can or should be mechanized to rules, points, rolls, whatever, if you are truly running a role-playing game (not a roll-playing game). I think of alignment as a general trend in the character’s actions – a way for the player to be guided when decisions need to be made about actions.

If nothing else, think about it from the paperwork aspect. If one (or even a couple, widely-spread) action is sufficient to drop one from being lawful good to, say, neutral good, at what stage can the PC turn back to lawful good? Six months without slipping? A year? Never? Three Hail Bahamuts?

Better to let her be internally torn over her actions (assuming she hadn’t been caught). She might have chosen to confess anyway and take punishment, preserving her lawfulness. Or perhaps she might get a thrill from the success and try it again and again. After a couple times, you privately warn her that she is risking an alignment shift (along with any negatives that brings) and give her the opportunity to mend her ways (or choose to continue).

On the other hand, someone professed as being lawful good shouldn’t be permitted to even consider something like random murder – it is just too alien to that character’s psyche (barring a triggering event). The player would have to gradually move the character from lawful good to something closer to evil by steps until it reached a point of likelihood.

The thing is, I haven’t played D&D, as such, since 2nd edition–I largely skipped 3 and 3.5, spending my RPG time on other systems. My only real exposure to D&D during that time was Neverwinter Nights, which uses a version of that Hallowed Might subsystem you describe (their scale is measured with 100 points, not 10, but otherwise fairly similar), and I had always assumed, therefore, that was a part of the default system for 3rd edition. Knowing now that it isn’t, I’ll need to revise a couple of the things I said to my group, since one of them is a veteran player, and knows 3rd edition better than I do, so he probably assumed I was talking about the default, static system.

I don’t care for the idea of a set of discrete alignments, whether there are 2, 6, 9, or 20–a spectrum makes much more sense to me, precisely for the reason of situations like the one you describe–no single act should transform a person’s alignment completely, even something as dramatic as murder: a strongly good character that commits murder would still feel a great deal of guilt about their act at first, even if they begin to rationalize it in their own minds. The idea of assigning certain acts a point value on a scale, and allowing that to determine where someone stands on the two axes makes a lot of sense to me. I can see how simplifying it into a few discrete segments along each axis might streamline things for the GM without oversimplifying, too.