Case Study 1: Journey — Multiplayer as Emotional Design

In March 2012, thatgamecompany released Journey as a PlayStation 3 exclusive. It was a short game — two to three hours long — with no dialogue, no text, no combat, and a single discernible objective (reach the mountain). It sold over a million copies in its first year. Players wept at the ending. Critics called it, variously, a masterpiece, a spiritual experience, and the most important game of its generation. Jenova Chen, the game's director, had done something that had been assumed not to be possible: he had produced strong emotional connection between total strangers using no tool except the mechanics of the game.

This case study examines how.

The Problem Chen Was Trying to Solve

Jenova Chen, born in Shanghai, had been thinking about emotional design in games since his undergraduate work at USC. His master's thesis and subsequent games (Cloud, Flow, Flower) were experiments in producing emotional states that mainstream games rarely targeted — tranquility, wonder, melancholy. By 2009, when he began work on what would become Journey, he was preoccupied with a specific question: why does online multiplayer so often produce negative emotion?

The observation was empirical. Anyone who had spent time in online games — Call of Duty, World of Warcraft, Counter-Strike — had experienced the toxicity of anonymous strangers. Trash talk. Griefing. Racism in voice chat. Ragequits. The default behavior of humans given access to anonymous interaction with other humans, in the context of a competitive or cooperative game, was often spectacularly unpleasant. Many players, by the late 2000s, had concluded that online multiplayer was simply structurally incompatible with emotional warmth.

Chen, looking at the same data, reached a different conclusion. He hypothesized that it was not anonymity that produced the toxicity — it was the affordances the games provided. If you gave players voice chat, they would use voice chat, and the percentage of humans who are cruel when given voice chat to strangers is approximately 20-30%. If you gave players the ability to kill-steal or to troll, they would. The toxicity was not a property of the players; it was a property of the tools.

The question, then, was: what would happen if you designed a multiplayer game in which the tools available did not afford toxicity? In which the only things two strangers could do to each other were things that produced connection?

Journey is the answer to that question, implemented as a game.

The Design

In Journey, you play a small, robed figure walking through a desert toward a distant mountain. At various points during the journey, you may encounter another figure. The figure looks like you. It is another human being, somewhere in the world, playing the game at the same time. You cannot see their name. You cannot speak to them. You cannot type to them. You do not know their gender, age, nationality, or skill level. The game has stripped away every signal of identity except what the other player chooses to show through their movement.

You can do exactly two things to communicate with each other: chirp (a small musical note, emitted by pressing a button) and stand near each other. When two figures chirp together, their chirps harmonize. When two figures stand close together, both figures regenerate a meter (scarf energy) that allows them to fly briefly. Flying is the game's primary pleasure — leaping and soaring through the sand and the ruins. The game makes proximity mechanically rewarding.

That is the entire multiplayer system. Two actions: chirp and stand near. No chat. No trade. No combat. No kill-stealing possible. No way to give items. No way to hurt each other. No way to identify each other across sessions.

And yet — and this is the observed empirical fact — players bond with each other.

Design Note: The reduction of affordances in Journey was not a limitation imposed by technology. The PlayStation 3 supported voice chat, text chat, friend lists, and all the normal multiplayer tooling. Chen's team actively chose to disable these. The design instinct here — that richer affordances often produce worse emotional outcomes — runs counter to the default assumption in multiplayer game design, which is that more is better.

What Happens in Play

The player boots Journey. They begin in a desert. They walk forward. Eventually, they hear a chirp. They look around. In the distance, a figure is standing on a dune. The figure chirps again. The player chirps back. The figure approaches.

Two strangers, bound by mechanics, begin to move together. One knows where the collectibles are hidden. The other does not. The experienced player leads the novice to a hidden area; the novice, uncertain, follows. They find a collectible together. They chirp in celebration. They continue.

Across the game's two-hour runtime, a narrative emerges between these two figures, a narrative that neither player wrote and that is not represented in any dialogue or cutscene. It is purely a narrative of shared movement. You helped me find this. I got lost; you waited. We flew together over the ruins. You were cold in the snow section, and we stood close, and the scarves glowed together, and we kept going.

At the end of the game — and this is the part many players have never recovered from emotionally — the two figures climb the final mountain together in increasingly cold, hostile wind. Near the summit, the wind becomes too strong, and the figures collapse. The game cuts to white. The two figures are revived, in a kind of afterlife sequence, and walk together into light.

When the credits roll, the game reveals the name of the other player. Just that. One PSN username, displayed on the screen. The credits list, in small text, every player you walked with at any point in the game. For many players, this reveal — the realization that the anonymous companion was a specific real person who had just shared something with them — is the moment when the weight of the experience becomes fully legible.

Why It Works

Constraint produces intimacy

The first and most important design insight is that Journey produces intimacy because of its constraints, not despite them. When two players can say anything to each other, the range of things they might say includes every possible thing, from beautiful to vile. When two players can only chirp and stand near each other, the expressive range is much smaller — but every chirp and every choice of proximity is fully visible and fully intentional. Communication is reduced to its purest channel, and what survives the reduction is affection.

This is a designer's lesson that transfers beyond Journey's specific content. If your game produces unwanted emotional dynamics — toxicity, boredom, anxiety — the solution is often to remove affordances, not to add them. Removing the voice chat removes the voice-chat-induced toxicity. Removing the kill-steal removes kill-steal trolling. You cannot stop players from doing things you have enabled them to do; you can only stop enabling things.

Mechanical reward for proximity

The scarf-energy system — standing near another player regenerates your flight capacity — is the design's core mechanical teeth. The game is not asking players to be nice to each other for moral reasons. The game is making them want to be near each other because it makes them more powerful. The cooperation emerges from incentives, not from instruction.

Chen later described this as "designing a game where selfishness and cooperation are the same action." If you want to fly as much as possible, you should stick with other players. The game has made the cooperative behavior the self-interested behavior.

Design Note: This is a different solution than most cooperative games use. Left 4 Dead, for example, uses negative incentives — players who leave their team die. Journey uses positive incentives — players who stick with their teammate fly. The emotional register produced is utterly different: negative-incentive cooperation feels like obligation, positive-incentive cooperation feels like connection.

The anonymity is the trick

Players who have identifying information about each other bring stereotypes, biases, and prior expectations to the interaction. Players who know only that the other is human — and who must read the other purely through movement choices — engage as if with a stranger whose character is being revealed in real time.

Chen has described this as the "moment of human connection" that most online games eliminate by default. When you see another figure in Journey, you do not know if they are a ten-year-old child, a seventy-year-old woman, a professional player, or a total novice. You can only see how they move. And because movement is the only channel, every movement is legible as character: this player is patient, this one is impulsive, this one is leading me somewhere, this one is waiting for me.

By the end of the game, many players report having formed a clear picture of their companion's personality — and being, on reveal of the companion's username, surprised to learn basic demographic facts about them. The companion, in their head, was more or less themselves. The anonymity let them project a full human into the figure.

The ending does emotional work that the gameplay earned

The final mountain sequence in Journey works because the game has spent two hours building the relationship mechanically. If the ending had been presented as a cutscene at the start of the game — two strangers climb a mountain together, collapse, and are reborn — it would have been affecting but not devastating. What makes it devastating is that the two strangers are you and the specific person you walked with, whose choices and patience and care you experienced through hours of mechanical interaction.

This is the principle from the main chapter, in its purest form: emotion in games is produced by the accumulated weight of the player's decisions. The ending of Journey cannot be borrowed from cinema, because the companion in the ending is not a character — they are a record of the shared activity the two players just performed.

Chen's Design Philosophy

Chen has spoken extensively about his design philosophy in GDC talks and interviews over the years. A few principles recur:

"Design for the emotion, not the mechanic." Chen begins his design process by specifying the emotional target, then asking what mechanical systems would produce it. Most designers do the reverse — they begin with the mechanic and hope emotion emerges. Chen argues this is why so many games end up emotionally flat despite having rich mechanical systems.

"Constraint is the designer's friend." When asked why Journey does not have the features players often request (voice chat, friend invites, customization), Chen has been consistent that adding these features would damage the specific emotional effect the game produces. The game is not a platform for generic multiplayer; it is a specific system tuned for a specific outcome.

"The player is always right — about their feelings, not their feature requests." Players often ask for features that would destroy the experience they are enjoying. The designer's job is to listen to the emotional reaction, not the surface suggestion. If players say "I wish I could talk to my companion," they are really saying "I want more of the connection I am feeling." The solution is not voice chat; the solution is to find other mechanical ways to deepen connection.

"Games are for feelings." Chen has argued, repeatedly and publicly, that the primary value of games as a medium is their capacity to produce emotions that other media cannot. The industry's focus on mechanics, graphics, and content, in his view, is a distraction from the medium's actual power. Thatgamecompany's subsequent projects — Sky: Children of the Light, Everybody's Gone to the Rapture (by other developers but in a similar tradition) — have continued to refine the same basic approach.

Lessons for Your Design Work

The transferable lessons from Journey are, in order of how often they go unused:

1. Remove affordances that enable unwanted behavior, rather than trying to prevent the behavior with rules. Do not have a moderation system for trolling; have a system in which trolling is not mechanically possible.

2. Align incentives with your emotional target. If you want cooperation, make cooperation the self-interested action. If you want tension, make tension the path of progress. Emotion does not survive contradictory incentives.

3. Strip identity to the minimum that supports your target emotion. Identifying information imports baggage. Strip it unless you need it. Most games need less identification than they have.

4. Invest heavily in the small number of actions the player can take. Journey has two multiplayer actions. Both are polished to a fine finish — the chirp has perfect timing and pitch-responsive harmonics, the proximity glow has elegant particle design. When you have few actions, each must be beautiful.

5. Let the player project meaning into the space you leave. The figure in Journey has no face, no voice, no backstory. Every player fills in those gaps with their own projection. The absence is the canvas.

Chen's work remains, a decade on, one of the clearest demonstrations of what games can do emotionally when they are designed for emotion rather than accidentally producing it. If you are serious about emotional design, play Journey. Play it with a stranger. Then play it a second time and watch what the game did to you.