Disabling Pawn Stance and Enforcing Character Beliefs on Action

[Daniel B]

The only other option I can see is to just encourage them to play as their characters' roles, i.e. nurture any interest they may have in adopting their characters' roles. However, for this to work, the players themselves must be willing to go at least half way.

Reliance upon the player to do what they're supposed to do is not a thing that I want to do. While most players will not try to game the system, I tend to envision the worst possible group ever conceived when it comes to that sort of thing, and try to envision ways to stop them from doing things "wrong," as it were.

I hate to sound authoritative, because I actually know very little :-D but..  you're asking the impossible. This is how I imagine things playing out if your mechanics could speak:

"
Mechanics: Okay, GM, ensure your players follow my rules. Players, here are your personalities.
GM & Players: Okay
Player 1: Cool, I get to play a <certain imagining of character>
Mechanics: NO! You're not playing the character correctly.
Player 1: Sure I am, that's what the character sheet says.
Mechanics: No! You're interpreting the numbers wrong. GM, you tell him.
GM: Actually .. he's not wrong. I don't see anything in the rules that says he can't play those numbers with that interpretation.
Mechanics: But .. but it's in the numbers. It's obvious!
GM: Not really. The way he's playing the numbers is legal by the rules.
Mechanics: But .. that's not how it was originally intended.
"

Obviously, this is a little silly but I hope you see my point. You can't imbue your mechanics with your intentions. You can only write up your mechanics and hope for the best. Even if you COULD imbue your true intentions, you can't force people to follow them. The "worst group possible" that I can envision has already dropped your book and gone to play a different game. What's left is people who aren't the worst group possible, and of these, again, you must hope they come at least half way.

Dan

[Callan S.]

Hi GW,

This threads getting a little frazzled. When I last checked it sounded like that you'd grant that players change character persona, but you just wanted them to adjust the numbers to follow what they change too. I'm assuming either they may spend points, potentially having to save up to change persona, or they can just change the values and this says to you as GM "Yes, I grant I have changed my characters personality, by having changed these numbers. And these numbers help you get how he's changed in personality". Which seemed pretty reasonable and useful. If it is what I thought, good, but the thread seems to have meandered from that? Perhaps start a fresh one based on that? If it wasn't, just making a quick post in case it helped your project

[JoyWriter]

Hey Contracycle, I'm not really sure how to do this post-splitting reply business compactly, but I'll give it a go:

Ok, Gnomeworks, you want to watch interesting characters. Someone else starts playing in pawn stance. It's not about whether they are saying "my guy says he wants this" rather than "I want this", because you can cope with that kind of narration for physical actions, so what stops you thinking about it as a proper character?

Umm, maybe becuase its not a proper character?

Ok, either we're at cross purposes or you think that talking in the third person is universally bad. I'm guessing it's the former; the point I was hoping to make is that we tolerate 3rd person narration for times when the characters voice is not how they are expressing themselves, such as "he leans on the bar, almost creaking from exhaustion" or "he thrusts his sword quickly into the beasts belly" or something. Now some people prefer it if character actions are expressed as if the character was saying it as much as possible, even down to narrating action intentions in character as some form of internal monologue, but I wanted to suggest that the things he was interested in could be seen when ignoring this level.

So if that is not the issue, what is? As I understood it GnomeWorks is interested in well rounded characters with proper motivations, and it is characters without proper in-fiction motivations that he specifically wants to ban, preferably by creating guides that help people make better ones.

Basically, you've said you don't care how the player sees it, so you don't really care what stance they are in, only that when watching them you can pretend that they are following a specific character. So what happens if they do something their character "would never do".

Why would they do that.  Explain.

Can't you just adjust your expectations to what the character actually does, and have people start to treat him differently? If his background and behaviour don't match, then anyone who knows his background will say things like "but he was such a good trainee in the monastery??", or something like that.

Yes; but that leads to a game-breaking conclusion: all PC's are mad.

So if player expectation doesn't matter, only that you see a world of real characters, then that one player always plays crazy characters.

Worse, from my perspective: ALL players play crazy characters.  Thus actual play devolves to: a day out for the insane asylum.

So if it's all about motivation and character realism, how do you deal with a player who is not participating?
Well 1, can you work out their motivations anyway? In other words, can you act as if they are a well played character by looking at their actual actions, just like you would with a character you read about in a book? I suspect such a success would end up with a mix between who they are supposed to be playing and their own personality, which has happened with me in many a game with certain players; I went full immersion and just mixed what I knew of my friend into what I knew of the character they were playing, producing a composite motivation that allowed my character to work with his.

As for 2, well then we get into what they "would never do" when the actions of the player characters go outside the bounds of making things fit, they just don't make sense, because the contradict things that have either been stated to be important to the character in fiction or on a sheet. Now one approach is deception; assuming the character is lying or being self-deceptive. This produces the most amazing conversations, where one character takes another up on who they say they are, and in the course of it the players learn something about roleplay psychology, and the character goes through one of those self-discovery moments. Now I'm not talking stupid "one tree hill" "what'cha gotta realise is" dialog, all ultimatums and platitudes, but situations that build character, in the character if you see what I mean. Taking 1d characters through the required journey to make them substantial.

The other is insanity/intentional incomprehensibility, of the worst sides of chaotic neutral or "mad character" play. I say worst, because in my more chilled out moods mental swings can be pretty fun to play through, as part of playing surreal consequence free games. But in a game of proper well drawn characters, that's anathema.

Now both of these are extremes, and I suspect it can be too easy to declare that complex motivations are simply incomprehensible and random. I wouldn't say that the PCs are mad, I've already suggested that a character can be found to be consistent even when a player plays them badly; if your good enough at finding motivation, you can turn the lightswitch paladin into someone who uses it as a cloak for his own ends, or someone who believes themselves to be justified in doing a little evil for their greater good. Then you can create situations to tease out the difference. I used the extreme example because I wanted to show how even severely disruptive play can be handled from an in-fiction perspective.

Do you agree with this, or do you feel that there is something about PCs as characters that can never be explained using only in-fiction tools? Something that makes them inherently mad from an NPC perspective? (feel free to defy the question if you can think of a better one!)

Ones that have sudden breaks in behaviour because of inexplicable reasons (read out of game events). You can tell people that as far as all the characters are concerned, his guy will appear mental, and instead of doing it GM "voice from on-high", do it through NPCs.

..which therefore becomes GM constraint.

This is control by feedback, and it is just as hardcore as rules based control, but more flexible.

It is not more flexible it is crude.  If the player wanted to be "a traitorous bastard", why did they not take that as  a trait at the outset and thus allow th GM to plan accordingly?

You get the idea. In-character based feedback encourages in-character based decision making, or something that looks like it.

And in-characer definition allows the GM to plan for the practical eventualities.  I fail to see how any of this justifies blind-siding the GM.

Oh yeah, this is serious GM constraint! That's why I said it was as hardcore as the "roll for motivation" idea. It's a fair point to ask why the player did not define their character from the start as someone who will betray close friends, but to be honest many people just don't think about characters that way, they might just do it without considering the ripple effects in their relationships, and this is supposed to be able to cope with such people, even be useful to them. The idea is that instead of backing up and saying "no, your supposed to be the hero", you take the character as played, and have everyone act towards them appropriately. This creates an in-fiction expression of what was un-heroic about their action.

The trick of such a system is supposed to be that people, as learners, will be able to smell out the feedback loops of the situation, and understand when things aren't working, so instead of automating the very bit that the player is supposed to be learning, you put a context around them that deals with them as if they were doing it intentionally. If constructivist learning theory is correct, when dealing with a system that provides consistent response/action loops, people can actually internalise those patterns and get used to working with them. This is in contrast to learning to copy a dice roll system, which although valuable, may not express the multilayered motivational elements of a single action. And more broadly it's like the difference between learning by rote and learning by experiment.

My idea is that if you want to create a framework for teaching character identity, the best framework is a social one, or one that responds to identity concerns as if it were social (ie a magic system would also work). It's not really my idea, it's to do with the idea of the social construction of identity, and I have my issues with it, but because it comes from a very external and mechanistic perspective it seems a perfect fit to the mechanical and statistics inspired perspective already evident in the system.

What you call in-character definition I'm guessing means "on the character sheet" definition, whereas in-character definition as in "through play" definition is exactly the kind of thing I'm talking about. It could make planning difficult, and unpredictable players always do, but it might be eased if you have a social system that already gives you the general parameters of response, perhaps even based on the MtG colour wheel, except as virtues you "exhibit" rather than ones you hold to.

Personally I think it is very valuable to consider what the player thinks about stuff, as a tool to help you make games smoother and better for everyone. But given the constraints of the original posters objectives, which I paraphrased as;
"Create a system that stops people breaking immersion with characters that are rubbish without chucking them out of the game, in some way that helps them learn to stop leaning on the rules system over time and do it themselves. But this rule system must be compatible with the preferences of my co-writers not to directly interfere with character portrayal decisions, although apart from that player motivations are irrelevant.",
given those objectives, I thought the fit was an external behaviourist/constructivist approach that ignores player motivation and treats the character as a black-box that is revealed through play, and interacts with it purely on revealed information regardless of "author intent".

Now you know what I mean, what do you think?

[GnomeWorks]

Player 1: Cool, I get to play a <certain imagining of character>
Mechanics: NO! You're not playing the character correctly.
Player 1: Sure I am, that's what the character sheet says.
Mechanics: No! You're interpreting the numbers wrong. GM, you tell him.

It's not like the game would tell you what a character does in a given situation, nor how those virtues came to be valued by the character, nor - necessarily - how they are executed.

Obviously some are a little more grey than others. Not much that can be done about that.

The reason we use direct opposites - or at least as direct as possible - is to avoid the situation you are talking about. It is much harder to say that a given interpretation is right or wrong when you're dealing with "white and black," as it were. Where you fall in the white or black is a shade of grey.

The "worst group possible" that I can envision has already dropped your book and gone to play a different game. What's left is people who aren't the worst group possible, and of these, again, you must hope they come at least half way.

There are varying kinds of "worst groups possible."

Regardless, though, what is perhaps more important is not the worst group possible, but the newest group possible. In addition to the worst group possible, I try to envision how a group of 16-year-olds, with no prior experience with tabletop gaming, sitting down with these rules would handle them.

For them, the point of the virtues is to help solidify the character concept, to establish that the character and the player are two different people.

When I last checked it sounded like that you'd grant that players change character persona, but you just wanted them to adjust the numbers to follow what they change too. ... Which seemed pretty reasonable and useful. If it is what I thought, good, but the thread seems to have meandered from that?

Yes, that is currently the general idea. There have been some useful suggestions made upthread, which I haven't had the time to digest fully, which may lead to a change in how this subsystem works.

And yes, the thread has gotten slightly derailed, but is still generally dealing with the question at hand and still relevant to resolving it.

[contracycle]

So if that is not the issue, what is? As I understood it GnomeWorks is interested in well rounded characters with proper motivations, and it is characters without proper in-fiction motivations that he specifically wants to ban, preferably by creating guides that help people make better ones.

The issue is characters that are divorced from their setting, that act as if in ignorance of the setting they inhabit.  The Pendragon traits were not simply any view of morality, they were an articulation of the specific morality of the chivalric romance; their appearence in the system thus serves as prompt and guide for playing within that setting.  Characters are not "proper" if they ignore such influences and behave in a manner that disregards the fiction.

So if it's all about motivation and character realism, how do you deal with a player who is not participating?

As in a player that is not present?  That's not a solvable problem IMO.

Now both of these are extremes, and I suspect it can be too easy to declare that complex motivations are simply incomprehensible and random. I wouldn't say that the PCs are mad, I've already suggested that a character can be found to be consistent even when a player plays them badly; if your good enough at finding motivation, you can turn the lightswitch paladin into someone who uses it as a cloak for his own ends, or someone who believes themselves to be justified in doing a little evil for their greater good. Then you can create situations to tease out the difference. I used the extreme example because I wanted to show how even severely disruptive play can be handled from an in-fiction perspective.

Tihs makes little sense to me for two reasons; firstly becuase playing in pawn stance essentially precludes complex motivations, because the character is just a token to be moved around, and secondly because you have only discussed consistency in terms of the characters previously established behaviour, not in terms of the behaviour expected in the setting.  Playing a complex paladin first requires that you be aware of the behaviours espected of a paladin; if you simply ignore those expectations, then violations of those expectations are not complex at all.  It just appears to be crazy.

Do you agree with this, or do you feel that there is something about PCs as characters that can never be explained using only in-fiction tools? Something that makes them inherently mad from an NPC perspective? (feel free to defy the question if you can think of a better one!)

No, what I am saying is that when characters are portrayed in ways that violate the social expectations of the setting, then they behave in a way that must necessarily appear to be crazy to NPC's.  And that leavesa the GM with an unenvious choice between ignoring their antisocial behaviours, or turning all disciplinarian - neither of which is, IMO, a healthy form of play.  Behavioural mechanics, serving as prompt and guide to players, are expressions of the fiction which are pertinent to player choices, and are set out for them on the character sheet.

Basically, prevention is better than cure.  Establishing contextual mores explicitly so that significant lines are crossed accidentally or in ignorance obviates having to take corrective action once they are crossed - or makes that corrective action something the player has bought into themselves and has chosen deliberately.

Oh yeah, this is serious GM constraint! That's why I said it was as hardcore as the "roll for motivation" idea. It's a fair point to ask why the player did not define their character from the start as someone who will betray close friends, but to be honest many people just don't think about characters that way, they might just do it without considering the ripple effects in their relationships, and this is supposed to be able to cope with such people, even be useful to them. The idea is that instead of backing up and saying "no, your supposed to be the hero", you take the character as played, and have everyone act towards them appropriately. This creates an in-fiction expression of what was un-heroic about their action.

Yes, but it also obviates everything else.  The game turns into a sort of "stranger in a strange land" kind of thing, in which the the in-fiction action is derailed toward setting exposition alone.  And especially within a group, if one player is violating expactations, and thus causing consequences for the others as well, then those other players have ceased to be relevant - they have become hangers on to one characters act of self-discovery.  That's a full blown breakdown IMO, a game effectively hijacked by one player.  And if the whole group is doing it, then you end up with a situation in which the GM is continually rescuing characters from the implicit consequences of their actions - which effectively means, discounting their contribution to the fiction - or applying ever escalating coercion that can reach the point of a TPK.

I mean you seem to assume that such GM coercion can be soft and gradual, but this is not necessarily the case for in-fiction reasons.  Committing a serious social blunder might poison an NPC's responses toward that character forever - or oblige the GM to make excuses for why it does not.  Sufficiently egregious behaviour could provoke social sanctions up to an including the ultimate penalty, or something else that renders the PC unplayable or requires them to be separated from the other characters.  So yes, this sort of in-fiction GM constraint is certainly "hard core", becomes the dominating concern of the game - to the detriment of what the game was originally conceived of being about.

The trick of such a system is supposed to be that people, as learners, will be able to smell out the feedback loops of the situation, and understand when things aren't working, so instead of automating the very bit that the player is supposed to be learning, you put a context around them that deals with them as if they were doing it intentionally. If constructivist learning theory is correct, when dealing with a system that provides consistent response/action loops, people can actually internalise those patterns and get used to working with them. This is in contrast to learning to copy a dice roll system, which although valuable, may not express the multilayered motivational elements of a single action. And more broadly it's like the difference between learning by rote and learning by experiment.

I disagree - just because a mechanic appears on the character sheet doesn't mean that players simply copy it.  All it means is that the mechanic is ever present and is considered a fundamental part of what the player is concerned about.  Which seems only right and proper to me - learning can occur through seeing trigger conditions before you trip them, not only by tripping them and dealing with the consequences.  For players to learn by "smelling out" the implicit system first requires that they actually survive long enough and keep playing long enough to deduce the underlying principles; if your GM is compelled to take in fiction corrective action, characters may suffer consequences that prevent that longevity.  It's just another form of "guess what the GM wants you to do", instead of laying out in advance what sort of things are expected.  The mechanic serves that laying out function.

What you call in-character definition I'm guessing means "on the character sheet" definition, whereas in-character definition as in "through play" definition is exactly the kind of thing I'm talking about. It could make planning difficult, and unpredictable players always do, but it might be eased if you have a social system that already gives you the general parameters of response, perhaps even based on the MtG colour wheel, except as virtues you "exhibit" rather than ones you hold to.

Yes exactly, a social system, visible to the player, presented to them before they take actions that bump heads.  I did not mean in-character definition to apply opnly to pre-play creation, I specifically intended to include violations of a characters previously established behaviour.  If the character is a pawn, then the player is not going to concern themslves with that much.  But if you control the moves which are valid for the pawn, then it may not become too disruptive.

given those objectives, I thought the fit was an external behaviourist/constructivist approach that ignores player motivation and treats the character as a black-box that is revealed through play, and interacts with it purely on revealed information regardless of "author intent".

I wouldn't say "ignores" so much as "informs", and authorial intent is dubious unless and until the socially expressed mores are fully internalised.  But that requires a method for them to become internalised, and that method is system.

[JoyWriter]

Hey contracycle, first of all I get the idea of metal stats, such as motivations, relationships or something like the what I call the "moodswing" system from Runeslayers, they're really valuable. The question is how they interact with player choice. Now I like incentive systems that use experience or localised bonuses, they're really cool, but what does it mean to use such stats as cast iron limits on player action? Do you generate the character opinions and reactions from the system, and have the player act them out? This was GnomeWorks first idea, but still produces the problem of interpretation "was that really in his interests given what he is supposedly motivated by in this instant?", which is not conducive to cast iron limits, it's too floppy.
So instead, I suggested that you look at the character from the outside: You say that pawn stance precludes complex motivation, and that might be true if you consider the player and character separately, but what if you just look at the two as a unit?
I mentioned before that players can build a characterisation from their natural actions, and you can then impute a character on them, saying "the world sees you like this" and having NPCs react appropriately. You're right that prevention is better than cure, and that is why I suggested that you have an explicit reaction system; it tells players how their character will be reacted to if he behaves in a certain way, rather than telling him how his character behaves.

Your dead right about the social expectations and crazyness, but not all pawn stance play will be outside of those limits. In fact, if you make those limits explicit; just like fire will burn their character or a wall is impassible, then most players, pawn stance or not, will stay inside them, thus making characters appropriate to the setting.

So just like players can learn the combat system and so fit into combat roles over time, they could learn the social system and avoid the problems of being treated as crazy! Actually you say the same yourself, so looks like we have been in agreement all this time, at least in part.

The same consequences you mention, and I see the problem, occur when dealing with life and death combat; we had one player once run off into a dangerous area, and almost cause TPK. You're dead right about indirect consequences, I didn't consider that. The thing is, we except that someone may cause problems for the group in a combat situation, why not in a social situation too? Can't we use the same methods to deal with both?

There is a general problem of insuring that players have information appropriate to the consequences, avoiding a sort of injustice that can occur when people randomly insult a king and get their heads chopped off before they know it! One method is revealing the consequences "the character would know" during resolution, another is adding hints in framing narration, and possibly expecting people to use knowledge skills to scope this stuff out.

One method I have used before, which I probably should have mentioned explicitly, is when the GM sets up conflicts as a way of teaching players the ground rules of the setting, especially useful for those who learn by doing or can't be bothered to read the rule book! This is the kind of soft consequence I meant, and your right, sometimes it's better just to read the rules, but after a few semi-dangerous encounters people quickly understand how things work, providing you make them the right way. I hope the same could apply to character consistency

I can't decide whether such information should be on the character sheet or not; I suspect by analogy with combat systems it should be split between some kind of GM record, and something on the players sheet that relates to change. I can imagine in some games instead of a combat map you would have a relationship map and causing alterations to it is recorded there, whereas the player sheet has some more generalised information about what sort of effect their skills can do in terms of social and physical effects. But in other systems yeah, I can see it all being on the character sheet.

As a general principle, based on taking physical considerations out of gameplay, I prefer it if information that should be publicly available is in the centre of the table, just to stop people having to check each other's character sheets all the time to find out their defences or negative traits excreta. It allows you to work the game with less cooperative players (read under 15s) without the game stalling because someone is being stupid.

So I hope you can see, I'm not suggesting that we remove explicit, designed, system from the equation, just that we put it on the outside of player character decisions, influencing but not fully determining them, as a "roll to see what you decide" system would. The point I was trying to make is that this form of control can be just as specific in it's in-character results as the rolled system, but hopefully more player friendly.

Regardless, though, what is perhaps more important is not the worst group possible, but the newest group possible. In addition to the worst group possible, I try to envision how a group of 16-year-olds, with no prior experience with tabletop gaming, sitting down with these rules would handle them.

For them, the point of the virtues is to help solidify the character concept, to establish that the character and the player are two different people.

In my outside-in version, that could be done by specifying the background of the character, and some of the cultural features that come with that. So it would tell you the kind of virtues that kind of person normally has, perhaps without completely setting up their opinions on stuff, unless they added more background elements. But for those who just don't get their character yet, and would prefer to fall into a groove over time, well they could just start with nothing and build up a reputation. Hopefully reflecting their actions back at them would help them explore things they hadn't considered.
Having said that, I'm quite keen on the idea of tying power during character creation into backgrounds, so that a player will have stuff to play off from the start, as it comes with what he is interested in doing. I'm not sure how well that would fit into a stricter system though.

[contracycle]

So I hope you can see, I'm not suggesting that we remove explicit, designed, system from the equation, just that we put it on the outside of player character decisions, influencing but not fully determining them, as a "roll to see what you decide" system would. The point I was trying to make is that this form of control can be just as specific in it's in-character results as the rolled system, but hopefully more player friendly.

Which is something of a straw man argument, because at no point did anyone argue for that; GW's proposition specifically contained the caveat that the player can overule the character.

The same consequences you mention, and I see the problem, occur when dealing with life and death combat; we had one player once run off into a dangerous area, and almost cause TPK. You're dead right about indirect consequences, I didn't consider that. The thing is, we except that someone may cause problems for the group in a combat situation, why not in a social situation too? Can't we use the same methods to deal with both?

The difference between the two is that in physical combat you know that you are in a fight, and know this when it begins.  The thing that makes social issues different is that they are pervasive and invisible, and if you don't already know what to expect you may find yourself locked into something before you even knew the issue had arisen.  Thus, it is something that benefits from being articulated as a mechanical system even more than depictions of physical interaction do.   Even if you try to go the laborious-GM-explanation-plus-trial-and-error, there is no guarantee that players will necessarily learn the right rule.  It seems to me more useful, in terms of outcomes, and more fair, in terms of expactations, to make the thing as explicit and functional as possible.

[JoyWriter]

So I hope you can see, I'm not suggesting that we remove explicit, designed, system from the equation, just that we put it on the outside of player character decisions, influencing but not fully determining them, as a "roll to see what you decide" system would. The point I was trying to make is that this form of control can be just as specific in it's in-character results as the rolled system, but hopefully more player friendly.

Which is something of a straw man argument, because at no point did anyone argue for that; GW's proposition specifically contained the caveat that the player can overule the character.

Seriously? Then I've been creating some obscure design challenge for myself then! Well in that case there are pre-existing incentive type methods that should do the job just as well. I still like the idea I came up with though. Perhaps I could exaggerate the feature you mentioned, the "stranger in a strange land" one, and make a game specifically about teaching separate character identity. Could be cool, providing it ramps up "social difficulty" slowly. Anyway, I thought that that game example was an expression of the design intent I mentioned and was trying to find something to fit it, if it doesn't fit it doesn't fit!

The same consequences you mention, and I see the problem, occur when dealing with life and death combat; we had one player once run off into a dangerous area, and almost cause TPK. You're dead right about indirect consequences, I didn't consider that. The thing is, we except that someone may cause problems for the group in a combat situation, why not in a social situation too? Can't we use the same methods to deal with both?

The difference between the two is that in physical combat you know that you are in a fight, and know this when it begins.  The thing that makes social issues different is that they are pervasive and invisible, and if you don't already know what to expect you may find yourself locked into something before you even knew the issue had arisen.  Thus, it is something that benefits from being articulated as a mechanical system even more than depictions of physical interaction do.   Even if you try to go the laborious-GM-explanation-plus-trial-and-error, there is no guarantee that players will necessarily learn the right rule.  It seems to me more useful, in terms of outcomes, and more fair, in terms of expactations, to make the thing as explicit and functional as possible.

Fair point. I imagine that that the same criticism applies to pure dungeon crawl rpgs; the only way to take a break from complex challenges is to take a break from the game! I wonder if there is a way to make breather sections, where people can rest their social skills? Actually, perhaps the dungeon itself provides the clue; the only way to take a break from the town is to enter the wilderness, the only way to take a break from the wilderness is to enter the town, that kind of idea.
But there is a big difference in that the existence of any kind of party structure brings society with you, and so presumably the social mechanics, unless the reaction component is restricted to NPCs (which is a distortion of consistency that I don't like much, but is inevitable given the freedom of players I assumed). To be honest by inclination I would rather make such a system magic based, so that it follows the characters around, and use it as a feature of how the game plays.

There is also the problem of hidden social/identity traps, which would be situations that effect their moral standing without the player realising: Traps of some form can be inevitable whenever there is hidden information, so actually the reputation/identity system would not be as susceptible to this as offending people would be. The reason for this is that there should be quite a bit of feedback as I said before there is quite a lot of leeway before a complex character becomes effectively insane. You can build up little impressions about how they behave, perhaps using a fuzzy set approach; where they slowly increase their membership of the "principled" club. This is the scrappy game-version of Bayesian case-reasoning, where you compare an unknown situation to one you already recognise, and base your reactions on that. It's like having WOD's morality system, but instead of x levels, you have membership of the two extremes, and the more strongly you fit either decides peoples opinion of you, and in more multi-polar systems the more you strongly fit two "conflicting" categories is the degree to which people consider you contradictory. That's a simplistic version, and more subtlety could happen, such as creating a new case just for them because of how they mix two cases, but even then you'd draw the line at a certain level of tweaking to stop recording concerns getting out of hand, or making it another obscure minigame like "confuse the 20 questions machine"!
Honestly, there's loads you can do on this subject, I've barely scratched the surface. The rules could be both rigorous and flexible, although the combination naturally produces complexity, but particularly with encounter design, I think you could use rules like this to "reveal" characters from uncooperative play in quite a satisfying way!

I'm beginning to like the idea of a case-based system that explicitly records the revealed "nature" of the PC on their sheet, so they know when they tripped a thematic trap, but I can see less-recorded alternatives working providing NPC reactions are frequent and integrated.

Tell you what, if you don't want to work on this Contracycle, or if it doesn't fit your criteria GomeWorks, then I'll just pop it in my back pocket and bring it up a few years from now as a crunchy identity based rpg of it's own, but if you want to keep talking to me about it, feel free.