Good and bad faith
I’ve been thinking out loud for a while now about good faith and bad faith and the role they play in building consensus for complicated social systems. Mostly on Twitter. But there’s one small problem: I’ve…never really stated what these concepts actually mean. So. I think it’s about time I give these terms a little more definition, a little more heft. We live in a faith-based social system, sure, but how does that actually work?
Well, here’s my best shot at it: A system is faith-based whenever the members of that system must act on the presumed intent of other actors in the system.
- In a good faith system actors presume that others will strive to benefit the system at their own expense. Example: this is how cells operate as actors in the human body.
-
In a bad faith system actors presume others will strive to benefit themselves at the expense of the system. Example: this is how predators operate as actors in a survival ecosystem.
- A mixed system, of course, exhibits some combination of the two. Most systems that we encounter are mixed, with layers and layers of both good and bad faith, stacked one on top of the other.
We can see how this type of “faith” dictates behavior and outcomes if we consider the classic prisoner’s dilemma, where a criminal suspect has to weigh the odds of being betrayed by their partner before deciding whether or not to betray their partner themselves. From the prisoner’s perspective, if they presume their partner will act in good faith, then they will choose to avoid prosecution and simply refuse to betray them. If they presume otherwise, they will immediately spill the beans and bet on the reduced sentence. If the presumption is unclear or uncertain, then the choice becomes exceedingly difficult – we are probably in the territory of mixed faith.
The best way to get a feel for the mix of good and bad faith in your system(s) is to look at how openly actors move around in public and share information. In the harshest bad faith systems, it’s pretty much impossible to speak freely, or even at all. This makes sense, seeing as good and bad faith are all about others’ intent. Revealing your intentions can be fatal – especially if they are good intentions. It’s the equivalent of showing everyone your cards in a game of poker (the quintessential game of bad faith). It makes you vulnerable. That is quite the opposite in a good faith system, where actors will generally feel comfortable being open, or even cavalier, about all of their judgements, mistakes, and intentions. If you’re in a good faith system, you can relax. Take things easy. If you’re in a bad faith system, well… read up on Sun Tzu.
A point of clarification: It can be tempting to think that bad faith systems are always bad and good faith systems are always good, but drawing that conclusion would be a serious mistake. There are all kinds of reasons why you might want to institute a bad faith approach. Survival, for one, though that’s not usually a concern for us in our day-to-day lives. Mostly, human systems employ bad faith in order to help identify or manage bad behavior. Looking to the legal system as an example, there we can see that not only do actors not speak freely, but they even enlist professionals to speak with extreme caution on their behalf – a textbook example of a bad faith system in action. Of course, it also exhibits all the problems that come along with a bad faith system, but we’ll get to that momentarily.
Modulation vs Eradication
As I mentioned, bad faith approaches are recommended as mitigation strategies for bad actors, best used to limit, constrain, deter, and prevent the incidence of chronic problems that can never be ultimately resolved. This is how the judicial system operates for major crimes, where we know due to human nature that criminality will never be reduced to zero. It’s also how speed limits are enforced, where we understand that we won’t catch every speeder, aiming instead to discourage the worst offenders by catching 1 out of 500. These implementations, though imperfect, are at least valid attempts at deploying bad faith: modulation.
But while bad faith has valid uses, beware that there are some important trade-offs to consider. If you need to be able to trust particular actors, for example, engaging them in bad faith can and often does break that relationship, unless they are unreasonably patient. It can also push the good actors out of your system entirely, if they eventually decide it’s not “worth it” to constantly wade through all the distrust and suspicion. But most importantly, when over-applied bad faith can actually empower the very bad actors we are trying to manage.
I have a term for this kind of bad faith overreach: eradication. Eradication attempts occur when actors in mixed systems try to completely eliminate every single instance of a given behavior or problem – at all costs. This… pretty much never works. Eradication might be a conceivable strategy when you have a direct mechanism to achieve it, like using a vaccine to eradicate polio or small pox (two major bad faith actors from the perspective of our immune system). But remember, we’re talking about human behavior here, so eradication can only be achieved by means of coercion or penalty. In other words, enforcement. Eradication requires enforcement, which leads to escalation. Escalation then empowers the actors with the worst available combination of malice and resources. In other words, it leads to an arms race, which will almost always reward the actors who win the arms race.
This is really just a classic escalation pattern, and once this feedback loop sets in there are three possible paths:
- Victory – You out-escalated the bad actors and you have managed to eradicate your problem outright. Good job! (You’ll notice there are no positive examples of this in a social context. There are, however… negative examples.)
-
Retreat – the bad actors out-escalate you, mount a defense, hold their position, and eventually force you to completely abandon your eradication strategy. (This is effectively what happened during the American prohibition era.)
- Equilibrium – neither side out-escalates the other but settles into an enduring stalemate, with dynamic tensions maintaining the balance. In other words, you end up with modulation anyway. (This is what we see with speed limit enforcement and other typical modulation strategies.)
I already mentioned it above, but the ultimate example of a failed eradication attempt in American history is the Prohibition era in the United States. It’s an almost pristine example of how a new eradication strategy gets put into effect, which leads to harsh enforcement and the ensuing escalation. The escalation then empowered the worst actors on both sides of the enforcement, not only by supporting the growth of organized crime but also by encouraging a parallel growth in general political corruption. With neither side able to win outright, the enforcement eroded on a number of fronts, including the most important one: public support. This resulted in the ultimate retreat of the movement in the form of literal repeal. Retreat.
(Note: While there’s historical dispute whether or not Prohibition actually caused more crime than simple urbanization, there’s no argument that it clearly empowered the worst actors in the ecosystem.)
Applied to Our Partisan Moment
What does this tell us about partisan bad faith? In a certain sense, partisan bad faith is actually an eradication strategy of its own. Each of our major tribes seeks to defeat and overcome the other, framing every political battle as if they are just one step away from total victory. We already see how the worst actors are winning this particular arms race. Hyper-partisan vitriol is simple to generate and is easily amplified by bots and high-profile accounts, leading to followers, influence, and ever-increasing levels bad faith. Meanwhile, the painstaking work of good faith becomes less and less tenable, more and more relegated to the vacant middle.
But it would be a mistake to think we can simply eradicate partisanship. We can’t. And we shouldn’t. In a healthy environment, the forces of partisanship provide the dynamic tension the center needs to do its job well. They keep the center honest. The problem is simply that the dynamic is broken. But what if we tried to put an end to partisanship anyway? What if we tried to eradicate it? It would… not go well. If the center simply declared that it would not tolerate any partisanship whatsoever, it would quickly get shut out of the game. It would be so lopsided you might not even recognize it as an arms race, or really any kind of competition at all – that is a race where the tortoise loses to the hare, every time. The center cannot hold, not by force of numbers.
You could even argue that this hypothetical isn’t even really a hypothetical; one thing that partly (partly!) weakened the “establishment” over the past decade was it’s diplomatic refusal to engage in the appearance of frivolous partisanship, leaving cracks in the facade through which the fringe was able to leak into the center. I preach moderation and pragmatism a lot myself, but I still can’t help but wonder how the last ten years would have played out if only the major establishment actors used a little less diplomatic doublespeak and gave a little more voice to partisan frustration.
But let’s put counterfactuals aside. While partisans seem to be locked in an eternal stalemate, we are far from equilibrium. The situation is very much unstable, and in this case victory for either side likely means catastrophe for our democracy. But if eradication isn’t even an option, that means equilibrium is our only bet. We can’t inoculate people against partisanship, and bad faith only begets more bad faith, so if we want to solve this problem we’re going to need to modulate it to the point where the system can once again move smoothly. We’re gonna need to find a measure of balance. We have to set up a decent speed limit, and put just the right number of cops on the road. Then add in a little leeway – sometimes you have to let the driver off easy! – and we’ll see the ecosystem naturally settle into a healthy equilibrium.
Yes, we need to find that speed limit. Then, and only then, can we finally start to reconcile our many divides. Only then can we have all the great ideological battles we have been wanting to fight but cannot – trapped as we are, swinging at the empty air in our own little bubbles.
Now, what should that speed limit look like? I’ve got some ideas, but it really isn’t up to me, is it?
One quibble to an otherwise interesting post:
In regards to your opening example, predators are not a valid bad faith example, but cancer cells are.
An interesting quibble! I try to avoid the cancer analogy, but only because it’s too perfect. From my view, cancer is the single purest form of systemic bad faith we know of. But I can’t prove that… yet. Plus it gets overused sometimes.
You’re right though that competitive/survival systems aren’t entirely bad faith. They’re mixed too. Tons of cooperation going on, if you look for it. And I kinda assume all systems are mixed, at least somewhat. But I’d still argue that survival systems lean more heavily on bad faith than good, at least at the “top level.”
When it comes to ideology, a good faith is when you assume people are generally sincere that their ideas will improve the lot of everybody even if you disagree with it and aren’t being disingenuous or trolling. Bad faith is when you assume people who disagree with you in the slightest are trolling.
Interesting post.
In a round about way you have found a main flaw in social democracy. That you still cling to it instead of burying it in the ‘bone yard of bad ideas’ is interesting also.
There are examples of other systems that do not require social objectivity to be resolved to function. ‘Faith’ for whatever measure, requires ‘truth’. The reason social democracies will lose in the market place of ideas, is it requires the social truth component of social objectivity to be resolved to function.
Good faith and bad faith are also unresolved when social objectivity is unresolved.
I have not seen a good case by ‘social engineers’ that social objectivity can be resolved.
Oh? What systems do you have in mind? I may be trying to build something similar, but I never complain if someone has already done the work for me.
I’m also pretty sure social objectivity cannot be resolved, by the way. Thankfully. If society could ever be “completed,” that wouldn’t be a good thing for any of us 😉
Before we discuss building and systems, let’s look at some parameters.
Is that persons ‘truth’ the same as that other persons ‘truth’ over there?
Is that persons concept of ‘justice’ the same as that other persons concept of ‘justice’?
(where’d ya go?)
Sorry! Many distractions. But I always appreciate a good nudge 🙂
And no, I honestly don’t believe in the concept of social equivalence. So I would assert that we’re all walking around with our own versions of justice and truth, each slightly varied from the other. But they only work when they are well-aligned underneath. When we have an “understanding.”
No worries, completely understand about distractions.
So if I read you correctly, you are agreeing that we each have our own individual construct of what truth and justice is and that varies?
Looking at the people we have, and not the people we wish we have, do you think it is logical to expect ‘well-alignment’ and ‘understanding’? Is that a realistic expectation?
Logical? No. It’s more of a social assertion than a logical expectation.
Realistic? Now that’s a tougher one. I don’t think I can give it a compelling answer at the moment, but I still assert that yes, it’s realistic. But the only thing I can really point to is that we’re all still getting up in the morning to prep for our commutes. I’d argue that’s something we all take on faith which reveals an underlying alignment in some general direction. At some point though, I need to make a stronger case for this. Plus that doesn’t mean we don’t have imminent problems, either.
(Not for sure what those first three sentences reveal)
If you and I were developing a system, and the main component of the operation was: ‘expect alignment and understanding’.
Do you think that is reasonable for the people we have?
Compared to a different system that was more along the lines of ‘expect misalignment and misunderstandings’.
Because when I look at the truth component of social objectivity, it appears to lean heavily toward ‘expect misalignment and misunderstanding’.
What are your thoughts?
Ha! Fair. Those three statements were mostly just my odd way of saying that it’s not necessarily always rational to promote radical good faith. I mean, look at the poetry of Rumi. Pure good faith, nothing rational about it.
On which to expect, I think there’s plenty of both, right?
I don’t think it is binary which is a problem. Even in the side that does have alignment and understanding it is a varied mix-match.
Since we are talking about systems, we could produce software code that reached the most complex part of interfacing with other software and code:
‘expect alignment and understanding’
but I don’t think that would be very useful.
This is the “assume a can opener” part of social objectivity.
I must say, I have been greatly enjoying our discussions and hope you write more often. I think we need more of these questions raised and discussed.
Fascinating post.
One of the interesting phenomena is when people obviously have to obfuscate their wrongthink.
So allow me to obfuscate some of my wrongthink by referencing hypernormalization for a sec:
And then referencing hyperreality for a sec:
And I’d get into it more but we’re in a place where obfuscating wrongthink is more and more important every day so I’ll just link to Slate Star Codex’s essay on the importance of obfuscating wrongthink. I take some comfort in the fact that, eventually, it will have to come out that the emperor is not, in fact, wearing any clothing (at the very least, when winter shows up) but then I think about hypernormalization and wonder at how long it can be prolonged by hyperreality.
normal brain: the Emperor was the bad guy
giant brain: the crowd was the bad guy
universe brain: the kid was the bad guy
Without the kid, maybe it could have gone on indefinitely.
(Let’s face it, the opportunity for people to come up with off-the-cuff descriptions of the emperor’s finery created ample opportunities for social bonding.)
Systems brain: Everyone worked together in different ways to contribute to social failure. If you don’t study the ways, you can’t understand the failure. Also the emperor should get some damn clothes 🙂
Of course, that situation is impossible to fully understand on good faith alone. You have to consider how social faith and material honesty work in tandem to triangulate truth. That system would be… messy. To say the least.
By the way, if you can tolerate a bit of self-promotion, I have a short story that is basically a rewriting of the emperor has no clothes for a more real/contextualized problem (historically accurate too!), if anyone is interested: https://lazyaphorist.com/2019/02/02/jannaeus/
Interesting story.
The belief that if we only held out for just a little bit longer… just a little bit longer we could have won! is one heck of a easy fallacy to sink into.
It’s certainly true that everybody who has won did so because they didn’t give up… but, dang, not giving up is necessary (not sufficient).
Depends on whether or not the nature of the constraint that forces a trial is material or social. If it’s a material constraint? Yeah, you don’t have more time. You gotta act. You’re gonna need to run some trials and take some risks. And pray you’re not building on quicksand.
But if it’s a social constraint? Forcing a trial is a huge risk that can cause systemic failure even when the underlying social context had no underlying material limitation.
The problem is, as an actor you sometimes won’t be able to tell which is which beforehand. Especially if you dismiss the good faith of actors who try to bring social constraints to your attention (as Jannaeus does to Ben Shetach).
The problem is, as an actor you sometimes won’t be able to tell which is which beforehand.
Change “sometimes” to “always” and I’m in 100% agreement.
I can be kinda allergic to absolute statements, but if you’re talking about absolute knowledge of truth, then I absolutely agree 🙂
It’s not a failure until it fails, though.
Until it fails, it’s a beautiful and lush deep blue with ermine trim and a gold chain holding the midnight blue (almost black) cape with ivory clasps at the shoulders. See how the light hits it?
That’s kinda what I try to get at. The concept of a social trial, why they happen, and when. And what it means when your system passes or fails a trial.
One flaw, which is kind of embedded in the speed limit analogy, is that while it makes sense if all motorists face the same risk of being caught or punished, if one subclass of motorists are likely to be shot by the cops when their pulled over, modulation is not going to get us anywhere. Especially if a lot of other motorists think of this as perfectly justifiable, which is easy for them to think since they can reasonably assume it would never happen to them.
Victory – You out-escalated the bad actors and you have managed to eradicate your problem outright. Good job! (You’ll notice there are no positive examples of this in a social context. There are, however… negative examples.)
Actually, this is not really true.
The abolition of the trans-atlantic slave trade and of slavery itself is ultimately one of the success stories of eradication. Same with the Nazis. Almost no one today is openly pro-slavery. And for a good long while, Nazis had to go into the closet.
Political corruption, communism and violent crime in Singapore are also more recent instances of success stories of eradication. All three were much more common when Singapore first became independent in the 60s, but are now all but absent.
The issue is not whether coercion works, but whether the cost is worth the results.
@murali Eradication means eradication, not all-but eradication.
Well, isn’t getting rid of 99% of a problem a resounding success on any realistic account of policy success? It’s very hard to imagine an actual situation where some bit of coercion would be worth it if it would eradicate the problem completely but not if it only eradicated 99% of the problem.
Our objections to coercive policies do not rest on the failure to get rid of the remaining 1%. Or at least, the objection would have to be rather bizarre. For instance, in Germany, they ban Holocaust denial. There are still holocaust deniers and Nazi sympathisers in Germany, but presumably much fewer than without said coercion. If banning holocaust denial is wrong, it would not become right if it managed to completely eliminate for all time all Nazi sympathisers.
@murali “It’s very hard to imagine an actual situation where some bit of coercion would be worth it if it would eradicate the problem completely but not if it only eradicated 99% of the problem.”
I would argue that smallpox in the wild is an excellent example from the OP. I’m not sure how *I* feel about 99 percent vs 100 percent eradication of smallpox, but I know how governments felt about it. Strongly enough to be more coercive than they would be otherwise.
And there are many many many people who would trade all kinds of civil liberties for the sake of 100 percent eradication of people who hate other groups of people.
If I’m not one of them I’m mostly not because I think that (unlike smallpox) it wouldn’t last. If it would last, there are a lot of things – not all things – that I am ashamed to admit I would give up for 100 percent that I wouldn’t give up for 99. Because the 1 can grow exponentially at any time, and 0 – theoretically – cannot.
In the real world, of course, I don’t think it’s possible to do that so it’s not practical to spend much time moralizing about it. But there are plenty of governments who aim for “zero tolerance”, and I think that’s what the OP is talking about.
I could be misunderstanding it of course.
A strategy to reduce speeding by 50% will probably look different from one to reduce speeding by 90%. So while this post speaks in black/white terms, the debate generalizes in important ways.
I’m all for banning legally recognized slavery by 100%, even though I know that somewhere in this country, right now, someone is living under duress.
We’ll never eliminate all transphobia, but dammit I want to see the cost of being openly transphobic much higher than it is now. It’s not a debate between 99% and 100%, because I avoid rigid black/white thinking. It’s a debate between 50% and 75%. That 25% will make a material difference.
(Note, I don’t think you can really measure this stuff this way. Social metrics are really fuzzy. It’s a mistake to view them otherwise. That said, I think these kinds of numbers can communicate general notions.)
Right. Like DUI enforcement. There’s a difference between increased patrols on weekends and holidays, aggressive use of DUI checkpoints, and posting a cop in the parking lot of every bar.
@veronica it’s not even the case, last I checked, that legally recognized slavery *is* banned by 100 percent, Colorado only got rid of it in this past election year and I’m pretty sure there are still other states where it’s legal to work prisoners without paying them.
(I, like you, am 100 percent in favor of banning it by 100 percent.)
Interesting discussion! Keep it going, by all means.
If I can chime in though (with apologies for the wall of text),
I’d suggest that just because eradication of behaviors is ultimately impossible does not mean it isn’t worth trying in the right circumstances! Sometimes you do want eradication, not modulation. And for sure, I’d agree that eradicating slavery is an obviously worthwhile goal, even if it could never be accomplished beyond 99.99%. (I don’t even think we’ve gotten to 99%, but that’s a different debate.)
But in adopting this goal, we still need to understand that if we want to enforce good behavior, bad behavior will move to all the spaces that enforcement cannot reach. Our ocean’s would be one of those places: https://www.nytimes.com/2016/02/21/opinion/sunday/slave-labor-on-the-high-seas.html. But you could probably argue the oceans aren’t “within” our society, only external to it. I’d dispute that, but even without disputing it you could make the case that there are also places within the continental US where slavery is still happening. Right under our noses. Again, everywhere that enforcement cannot reach or extend. In my home state, for example: https://en.wikipedia.org/wiki/Human_trafficking_in_Georgia_(U.S._state). And these are the “known” areas where enforcement does not extend.
Lastly, slavery is not the only kind of bad behavior we might want to eradicate. And every in every domain where we try, we must not forget there will be areas where our knowledge and enforcement cannot reach. I might have a name fore that: https://lazyaphorist.com/2018/07/31/the-veil-of-misery/. (Hint: We should be focusing on all areas of custody.)
“Our objections to coercive policies do not rest on the failure to get rid of the remaining 1%.”
apparently you haven’t been paying attention to the all the arguments about why the shutdown was a dumb waste of time because the wall wouldn’t stop ALL the people coming through the border
I haven’t processed this essay completely, perhaps because I am so bewitched by the badass cover art.
I am so ready to ride that train, and I don’t even care where.
Train looks cool, that wheel/rail layout….that’s some sketchy stuff.