It’s an article of faith that all-chat is a cesspool. That reputation is richly deserved. However, it’s not a given that channels for communicating with players on other teams will only ever be used for flinging insults. Global chat channels can work in games designed around them.

Let’s start by laying out the problem to be solved. As a rule, all-chat—that is, a communication mechanism in online games that allows every player in a game or match to talk to each other—is silent at best and hurtful at worst. It says something that one of the first things League of Legends did to curb unpleasantness in its playing community was to set all-chat to “off” by default. Perhaps more remarkable, MMOs now allow players to opt out of their global chat channels. That’s how bad the situation is: an entire genre built on the social aspects of gaming has to let players shut down a primary means of socializing because it’s so awful.

What would it take to make all-chat good? There are two things I can think of:

1. A good all-chat has a gameplay purpose. Everything in a game should have a gameplay purpose. Social features used to get a pass on that, on the theory that more ways for players to talk to each other automatically made for a better overall experience; time has put the lie to that belief. If all-chat is going to be rescued it will have to earn its place.
2. All-chat needs players to be reasonable when using it. Making all-chat in its current form central to a game would make that game the least pleasant thing on the internet. For it to be beneficial the messages that go through all-chat must be free of the lowest-common-denominator vitriol so common today.

We can discuss each of those in turn.

The simple part: making all-chat important to the game

The former problem is relatively easy. Opposing parties talk with each other all the time, and there are plenty of ways to bring that into a game. Negotiation, for example, can be a centerpiece of strategic play; Diplomacy is a sufficient proof of that. For a sneakier version of communication, a wargame might include the concept of sending false messages to the enemy, or an economic game could involve market manipulation. Co-ops and team games often demand synchronized effort. Semi co-ops involve lots of talking as players try to balance their personal goals with the group’s needs. There’s no kind of game that can’t be built so as to encourage the players to talk to each other.

The hard part: kinder communication

It’s the latter issue, that of achieving good behavior, that’s the tricky one.

Solution 1: Put the players in an environment where dominating others isn’t the goal.

Keith Burgun recently presented an interesting argument that a game’s thematic elements affect how players view what they’re doing, and by extension how they interact with each other. When players are told in advance that the goal is to have fun together, he explains, they generally act in ways that are consistent with everyone having fun. He cites as an example his very different experiences in games with different art styles; players were nicer to each other in Team Fortress 2 than in Counter-Strike, even though they’re both violent games, because TF2’s cartoonish visuals emphasized that everyone was there to have a good time.

It’s when players are told that the goal is to dominate and harm others, Mr. Burgun argues, that they adopt language to suit. “[W]hile a player is operating in a world of violence, he is more likely to think violently.” (emphasis omitted) Players naturally respond to a game that tells them to hurt the enemy by trying to do so in every way they can, cruel words included.

Mr. Burgun’s theory points toward games that are built from the ground up to send specific messages: that winning doesn’t require achieving power over the other players, that the overall project is fun rather than in-game success, that other players are co-participants in the overall project and should be treated as valued teammates rather than as obstacles. Global chat could work fine in such a context. Without the nudge toward unpleasantness that comes from a violent theme, most players will default to a reasonable mode of conversation. Outliers will hopefully be few, and easily dealt with.

Solution 2: Effective deterrence.

There are games that don’t look at all like Mr. Burgun’s ideal, and yet the conversation manages to be civil. Diplomacy is again my go-to example. It’s a wargame that’s expressly about conquering Europe and eliminating players, but it’s unusual to run into someone who’s openly nasty. By and large people are cordial, even when they’re stabbing each other in the back and overrunning each other’s territories. Why does Diplomacy work?

Here’s my theory: Diplomacy, along with Twilight Imperium, the Game of Thrones board game, and others of their ilk, has the most effective deterrence around. In fact, Diplomacy has a level of deterrence that the criminal law envies! The structure of the game ensures that players who want to be mean are powerfully and reliably discouraged from doing so.

I recognize that that’s a pretty bold claim, so let me back up and discuss this more fully. Deterrence requires at least three things: (1) there is a rule you want people to follow, (2) people know about the rule, and (3) people are more afraid of the consequences of violating the rule than they are eager for the rewards to be had from doing so.

(1) is trivial. (2) is very much not trivial. New laws, highly technical laws, laws about unusual issues–all of these can have a weak deterrent effect simply because people don’t understand what’s forbidden or don’t think to ask whether there’s a law on point. Still, for our purposes we can assume that (2) is easily achieved in the context of rules about “don’t be a jerk on the internet;” everyone’s been told not to be unkind at some point.

(3) is the hard one. This is for a couple of reasons. First, humans discount the threat of punishment by the chance that it won’t happen. Put simply, people aren’t afraid of violating rules when they think they can get away with it. The greater the odds of getting away with it, the weaker the deterrence.

Second, humans aren’t very good at weighing future events against current ones; we tend to discount future harms based on how far away they are. The longer it will take for punishment to happen, the less we tend to care about it.

These foibles make it harder for the criminal law to achieve its deterrent purpose. Every time somebody goes to break a law, they implicitly weigh the consequences against the ideas that (a) they might not get caught and (b) the price of getting caught will be paid at an indeterminate point in the future, whereas the rewards will be here promptly. As the continued existence of crime demonstrates, some people do that calculus and come to a regrettable conclusion.

Diplomacy, on the other hand, creates an environment where those human failings aren’t given much room. The negative consequences of being nasty to other players happen right away and are extremely predictable. Negotiations break off; other players won’t provide the assistance necessary to progress; the game ends in swift defeat. The whole process takes a few hours at most.

As the theory of deterrence predicts, that leads to most Diplomacy players being polite. Tempers can flare and the gameplay is often vicious, but the kind of hateful, profanity-laden speech one finds in online games is absent. It’s remarkable: Diplomacy is basically built around all-chat, but it doesn’t sound like the all-chat we’ve come to know and disdain.

Compare this to games that try to achieve deterrence by having rules in the Terms of Service and banning users who break them. They suffer from the very problems of uncertain and distant punishment that the criminal law does, with the added weakness that banning isn’t nearly as severe as what the criminal law can impose. The sad reputation of all-chat is in part due to the fact that the deterrent effect in these games is very weak indeed.

From Diplomacy and similar examples I think that deterrence can be an effective mechanism for promoting good communication behavior in games. However, strong deterrence isn’t achieved simply by hiring some mods. It requires that the game be designed from the ground up to have a short feedback loop that consistently discourages unkindness.

Build from the right foundation

We’ve learned from sad experience that all-chat isn’t something that can be tossed on top of a game. The results are unsatisfactory, to say the least. However, global chat could be a valuable, positive thing. A game designed with the needs of all-chat in mind from the beginning, tuned in such a way as to bring about friendly communication, could elevate the global channel from cesspool to centerpiece.