This is a transcript of the original talk given at the 2012 Game Developer's Conference. Click here to view the original recorded talk (GDCVault Access required).

I studied abroad in Japan 4 years ago.

It was my first morning there and I was at the Tsukiji Fish Market searching for inspiration for some collage work that I was creating for a gallery show. Tsukiji is the largest wholesale fish market in the world. There are about 1000 vendors, each selling live fish just as it comes in off the boats. Fish of every color and shape. Some were familar, but I was especially fascinated by the ones that seemed alien.

It was incredibly immersive—so much so that I lost all sense of time and place.

Earlier that day, my professors, classmates, and I passed by a nearby shrine and chose it as our rendezvous point. At a very specific time, a bus was going to show up and we would depart. The bus was on a schedule, and so we couldn’t be late.

Well, when I pulled myself away from the fish, I realized that that time was only a few minutes away. And here I was, my 2nd day in Japan...I didn’t have a phone, I barely knew a word of Japanese...I didn’t even know the name of the hotel we were staying at. I was about to be stranded with very little means to get me home.

I began to panic.

Fortunately, I had taken a photograph of the shrine—and so I pulled out my camera and called out to the closest person to me. I said “Sumimasen”, one of the only words of Japanese I knew, which means “Excuse me!”.

That person happened to be a woman, about 60 or so years old, speaking to a man in a truck. Upon hearing me she turned, and I pointed to the display and threw my hands up as if to question “Where is this?”. She immediately ended her conversation, grabbed my hand and took off running. We ran about four blocks to the shrine, just as my classmates were boarding the bus to leave. She bowed, smiled, and disappeared forever.

Now, this event took place in the infinitely complex system we call reality.

To recreate it and find meaning, as designers we have to focus in and choose the parts relevant to the scope of my perspective and my narrative as a player, and the woman’s perspective and her narrative as a player.

And so we pick out factors like the stimulating environment that led to me feeling overwhelmed, my goal of getting to the rendezvous point, the amount of time before the bus leaves, the high stakes of getting left behind, the distance of the shrine, the camera with the picture of the shrine which served as my only available tool...

...and because its multiplayer, we simultaneously need to be doing the same for the woman. Her old age, her investment in her conversation, her knowledge of where the shrine was.

And then there’s the relationship between us, the communication barrier that separates us, and the empathy that allows us to understand each other in spite of that.

All of these relate to create a dynamic that leads to players feeling what we want. And then we have to take a deep breath, sit back, and give them the freedom to act.

Both games I’ve helped design, "Journey" and "WAY", attempt to herd two strangers toward friendship. And both do it in similar and different ways.

But how do we do that? How do we design so friendship will emerge? And what is friendship really? It’s impossible to completely define or quantify. Nor can it be forced between players. Their emotional connection has to happen as a spontaneous creation of the participants.

I’ve been asking these questions for a while now and I’d like to start a discussion about ways in which we might be able to build systems that trend toward such a relationship. Now I can’t possibly cover all aspects or types of friendship in this talk...nor do I claim to have any absolute if you disagree or have a different methodology, i’d be happy to chat afterward. I don’t expect there to be much time for questions, so please, come grab me outside after the talk...or over in the IGF pavilion.

So what type of friendship am I speaking about?

What I’m interested in, is that spontaneous bond between strangers. I want to focus on online multiplayer that emphasizes shared goals, freedom of choice, anonymity, vulnerability, and communication.

Online Multiplayer provides an incredible opportunity for people around the world to play together, exempt of prejudices or stereotypes. If we can get two people to have a connection in a game, before their prejudices get in the way, then perhaps we can challenge those prejudices. Get the two people to empathize before they draw lines between each other.

Play is a language all humans speak—it’s fundamental, universal, neurobiological, it’s how we’re wired—and so we can use play to build an understanding between people around the world.

Because of this potential, I’m advocating that we treat Play with all the seriousness it deserves. A lack of seriousness can lead to oversight, misuse or abuse...and so we need to put care into our designs to ensure players are acting in the ways we value. And value is obviously subjective to the designer.

So I just recounted a story about spontaneous face-to-face interaction... Well let’s take that face-to-face interaction and add a few different rules. By rules I mean all the things we put into our games to define the space. Instead of being face-to-face in the same physical space, we’re connected over the internet by cameras—randomly—each of us in our own home.

Chat Roulette is an amazing story. What could have been an incredible portal to engage and empathize with people around the world was instead used for other things. I’m sure we all know what those were.

Now, why did this happen?

In Chat Roulette, players are behind a screen without any rules to guide their behavior. Without rules, and secured in privacy, problematic players were able pervert the space and claim control, forcing others out.

Now compare that to the London, New York Telectroscope.

Designed by Paul St. George for the Brooklyn Bridge’s 25th Anniversary, the Telectroscope was a 2-way camera linking London and New York, themed as a giant tunnel burrowing through the Earth. Unlike the crude interface of Chat Roulette—the Telectroscope is a spectacle—a work of art. There’s an immediate, aesthetic beauty. And so when persons interact with it, they treat it with a greater deal of respect.

More importantly, it’s in public—players, if we may call them that, interact with one another knowing they will be seen by more than just their counterpart. Or in other words, they’re operating within the rules enforced by a public system, which have understood behavioral codes that are not regulated in Chat Roulette.

On top of that, the designer adds another restriction—no voice communication. And so players create their own playful ways to engage each other and it becomes a toy.

The chance for abuse in the manner of Chat Roulette, well, we can’t call it zero, but I’m fairly certain you’re going to be arrested if you try.

Seeing how just those few additional rules affected player behavior—let’s go one level deeper, playing behind digital avatars in an online game.

My first experience playing an MMO was in 2004. I was 16 and the game was Final Fantasy XI. Like the Fish Market story earlier, when I entered Final Fantasy I was surrounded by all sorts of information I couldn’t understand. I wandered outside of town, where I saw a Mountain. I approached the mountain—and at the base, I met up with another player, and we starting fighting monsters together. And then we died...

When we died, a prompt appeared on my screen which read “You can wait to be revived, or you can click to teleport and return to your Home Point.” accompanied by a timer counting down toward Zero. This was my first time dying, so I didn’t know what that meant.

Where’s my home point?

Is that where I started?

Is my buddy’s home point the same as mine?

If I clicked the button, would we stay together?

What happened when the timer reached zero?

None of this was explained by the game. And so again I’m nervous. I quickly typed to my buddy “What should we do?”. But the response didn’t come from him, it came from the game:

“You are not allowed to type when dead”.

Here I was, my corpse laid out next to his, and we couldn’t communicate. And then something phenomenal happened. I realized that if I was here, at my corpse, and my buddy’s corpse is still there, than he must still be there too.

I suddenly felt more aware that there was a person behind that avatar than ever before.

A human being.

I could see him sitting at his computer, staring at my corpse, waiting for me as I was waiting for him. We were having a conversation without any actions at all.

Now, it’s worth pointing out that this was all totally in my head. For all I know, he could have been away heating up a Hot Pocket knowing he had 8 minutes to spare, but that’s not what I felt.

Just as the counter neared Zero, my anxiety growing, a White Mage appeared out of nowhere and revived both of us.

My voice returned, I sprang up and thanked the mage. This time my response did come from the player, but instead of a “You’re Welcome”, it was a single character of Kanji and a question mark.

When we realized we couldn’t speak to each other, the Mage began using the in-game gesture system to communicate. I wasn’t even aware yet that there was such a thing, and so now the Mage was teaching me, building on top of the positive relationship we already established.

Over the next couple of years, I would wave to that Japanese Mage whenever we crossed paths. Unfortunately, that was not the case with many other Japanese players, and I’ll explain why.

In Final Fantasy XI, all of the things that are valued—obtaining new armor, earning new abilities, fighting new monsters, even seeing the world—each of these is predicated on the need to Level Up. Well, leveling up meant obtaining lots of Experience Points...

...and experience points were obtained through Combat.

If you died, you lost Experience Points, and so Combat was important. In order to get the most Experience Points, you needed to partake in Team Combat to kill monsters stronger than yourself.

In order to perform good Team Combat, you needed to know how to link moves with your teammates—You do a move....then I do a move...and those moves combine to perform a super move. In order to coordinate these moves, we had to Communicate.

More specifically, we had to communicate through Text.

And so it became difficult to succeed with the Japanese players. And so you’d see this...

and this...

Soon enough the entire world was segregated into English and Japanese-speaking players— all because the best way to achieve Experience Points relied on strong textual communication. It’s not that players didn’t want to play together. Of course they did! It’s that the design discouraged it. It didn’t have value in the game world. Certainly not if you wanted to get those precious Experience Points.

And so it went unconsidered.

"Oh that’s just how you play the game."

As a player I felt alienated...unwanted...disconnected... Because that was my experience. As a player, I only knew how I felt. I was not able to see the root of the problem which was occurring because something may have gone unconsidered or undervalued in the design.

A single rule or value can pollute an entire system.

But what was it that got me so hooked in the first place? What were the seeds of my connections?

There was investment...I had made it to the Mountain and met up with a partner... There were serious consequences...we died and were about to be separated, possibly forever... There was empathy...myself projecting what I thought he was thinking and feeling... There was him honoring my vulnerability—at least as I inferred it—his choice to stay with me, even when he didn’t have to... There was the Mage, freely choosing to help... And there was the moment of teaching and communication without words.

But how do you design for that?

This all took place within an enormous MMO. For the large majority of players, it’s safe to say they didn’t have this specific experience. Where the system is so big, it’s unlikely this kind of moment was the sole intent of the designers...or at the very least, I doubt it was their primary focus. So how do we create a game where similar emotions happen more often? Where they are our focus?

These are the types of questions and problems that influenced our thought processes for Journey and WAY.

You may have heard of Journey. For those that don’t know, Journey is an online game where anonymous players walk, surf, and fly toward a Mountain—and can connect with 1 player at a time in the process. Connections occur when two players are in the same part of the world. If the two players move far enough apart, they are disconnected. Players can’t talk or type, instead they communicate through a musical call and the simple motions of their characters.

I’d like to briefly mention that I wasn’t on Journey for the first half of its development. That time was spent developing WAY. And so I’ll speak specifically on Journey as we developed it in the last 15 months.

One of the focal questions of Journey is "What a game about discovery and transformation, you experience that transformation alongside another?"

What will you feel?

Will you feel connected?

Like many linear games and other classic stories, Journey borrows the structure of the monomyth—a transformation story. Emotions arc as the protagonist passes from experience to experience. Because you engage in these experiences with another player, there is the potential to share a wide range of emotions with them.

Now because there’s no textual or verbal communication in Journey, it’s impossible to know exactly what the other player is feeling or thinking. Personally, I find that beautiful. After all, no two persons in this room could ever know exactly how each other feels—we are creatures of interpretation—each with our own perspective. Like the dead corpse in Final Fantasy XI, without words, the player is able to project their own thoughts onto the other person and what they think the other person is thinking...generating perceived empathy.

This is true not only of a moment of discovery, but also when players choose to walk away from each other. When two players separate, they disconnect. Why did that person leave? Did they not want to travel with me? Were they distracted by something? Should I follow? These are all questions. You wonder about them because you don’t know for sure.

When players disconnect, they disappear from the world. Because the world of Journey has many large and barren vistas, players feel small and lonely. Because it’s barren, this also means that when another player appears and moves through the landscape they become a focus.

Journey also maintains the important 4th wall. Though one might argue text and speech allow players to more clearly discuss or rejoice over a particular event...text and speech also introduce a number of unwanted verbs that could break the experience. Players could be pulled out, the game could become an object for discussion versus the world that simply is, etc.

Personally, I prefer that players communicate through nonverbal actions. This places everyone on a more even ground—any two players have the same actionable verbs.

In Journey, one of the clearest examples of these is the players’ musical call. The call serves a number of functions—one of which is to get the attention of the other player. Though it’s a single button, the call has a range of expression—and this is important. Tap the call in quick succession and it may resemble excitement. Hold the call button down for a longer period, and the character releases a more powerful shout. How players call helps bring identity to their person.

I spent many hours playtesting with my teammates at thatgamecompany, and I could often tell one person from another just by how they used their calls to express themselves. Like any strong relationship—it requires the other person has a unique, personal, human identity for you to resonate with.

We can remove text and speech, but we must not remove the player’s voice— their ability to express emotions, feelings, and opinions.

The call provides other functions. Players can “harmonize” with other players in the world— glowing the players and giving each other expendable energy with which they fly. Players can obtain this resource by physically touching each other—encouraging them to stay close, again bonding them—or by calling to each other within the radius of the bubble created by the call. The shared dynamic this creates is that of a Call & Response.

I call to you, you fly and expel your resource, you call to me, and I fly and expel mine. This creates a rhythm between the players. I personally like to think of it like a dance, or perhaps a version of Leapfrog. This is an example of Reciprocity, equal give and take between both parties. Reciprocity is a fundamental characteristic of friendship.

Without the other person around, players are unable to continually fly. And so when they’re gone...or when they choose not to can feel that. You feel it in the slow walk, your progress hampered, though not obstructed. This is us designing for a feeling of Longing.

Together we fly toward a single shared goal—the Mountaintop—which funnels the players together.

Replayability often comes up in critiques of Journey. One of the questions we find interesting is, once the endpoint has been reached and the story completed—once the discoveries of the world have been revealed to the player—what becomes important then? What does it mean to play the game a 2nd, 3rd, 5th or hundredth time? Where will players find value then? And what about years from now, when the deserts are barren and there’s rarely a journeyer walking to the Mountain? What will it mean to suddenly find another person in the dunes?

Before working on Journey, I spent 12 weeks designing the shortform version of WAY with a small team. It was made without any preexisting knowledge of Journey, mostly designed before Journey was even announced—and yet there are quite a few similarities. In our attempts to form connections between strangers, both designs arrived at similar conclusions—and yet they are two very different games. Each with a very different set of aesthetics and design philosophy.

WAY is a communication puzzle-platformer. In WAY, you and an anonymous player venture toward each other from opposite ends of the world, solving the puzzles between you. The game is played in split-screen, and each player sees the world differently. One player will have information the other does not—such as the location of an invisible platform where they need to jump, or perhaps a trap they need to avoid—and so you need to communicate this information with each other to progress. Like Journey, there is no text or speech communication. Instead, players have to share information by puppeteering their avatars. You can point, wave, mime a circle, nod your head—whatever works for you in the system. All of these are player controlled—ensuring the puppetry is a manifestation of your specific voice.

Because play is anonymous, players cannot use third-party chat clients like Ventrilo or IM to break the 4th wall.

When the game starts you are alone. It’s not until you proceed a bit further that your world splits in two and the other person appears. This creates a sense of surprise. The other person, like in Journey, is a epiphany—and it’s this realization of the Other that grabs your focus.

Little is explained within the game world, empowering the players to become teachers unto each other. Rather than teach the players ourselves, we instead empart as much of that responsibility to the player. This encourages players to speak to each other and creates opportunities for a stronger connection. When 1 player understands and does not teach it to the other—it can lead to failure and frustration—but this too can be positive! We’ve seen players suddenly realize they had information the other did not, and then feel shame for not helping each other. If this occurred outside the game, we would hope that person reflects on not helping when they could have.

That is something we value.

Because neither player has all of the information, players are interdependent. This is a big difference from Journey. Each player needs the other. In order to finish the game, you must invest yourself and commit to playing with this single other person.

Portal 2 multiplayer has something similar, but it's not quite the same for an important reason. I am a fan of Portal 2, but there are certain rules in the puzzle design that WAY approaches differently.

In Portal 2 multiplayer, it’s common for one player to understand the full solution of a puzzle. Once that is understood, the player simply need tell their partner what to do. This can lead to a sort of single-player domination effect where one player is solving the puzzle and the other is simply following orders. Because orders in Portal 2 are easily understood, because orders are not designed to be puzzling—I tag a wall with a cursor of a portal, you place a portal on that cursor—the puzzle need only be solved by one person. This can lead to a scramble effect, where players are trying to solve the puzzle faster than each other—and that can result in a feeling of spite if your partner solves it before you. I haven’t asked, but I wonder if this was the feeling they were going for. Another result of that same scenario is one player stops trying to solve puzzles entirely, and simply waits for the other to figure it out.

Because communication in WAY is interdependent, one player communicates while the other player actively listens, and vice versa—both are engaged and equally important. This creates a shared Call & Response dynamic—somewhat comparable to the call & response in Journey where players take turns sharing the resource to fly.

In WAY, meaning is constructed from a freeform puppet system. The players create their language, freely choosing what means what. This ensures that whatever players come up with— however they solve the puzzle—is their own. We’ve actually seen players discover entirely valid solutions that we the designers did not even consider.

And there’s another win here too. Because puzzles are open-ended—because players create their own language to solve each problem—solutions can feel both smooth or clunky. This too is part of WAY’s philosophy. Group A might arrive at a transparent solution to a puzzle and it feels graceful, while Group B comes up with a convoluted one and it fails and fails until it finally works. Group B might walk away from that experience saying “oh that was horrible” or “how sloppy”—but I find that to be honest design. That’s how the world works.

They didn’t consider an alternative.

And then we’ve seen these players revisit those puzzles with a new partner who then shows them another way. And suddenly their perception grows.

As a metaphor, consider talking to your friend and describing something they have never seen. How you explain what you mean will determine whether or not your friend understands you. Depending on the approach, coming to an understanding could be simple or difficult. So is true in WAY.

And in a game about communication, should we or shouldn’t we be teaching players to value how they communicate?

In the pamphlet I mention that I’d also talk about how we “maintain” these friendships. Both of these games attempt to establish a bond, but what it takes to carry a friendship transcend and become something bigger than the game...requires growing beyond the game itself.

Within these controlled systems like WAY and Journey, it’s impossible to include so many of the things that might strengthen the bond between two persons. And so instead, we focus on establishing the connection. Then, upon completion, each game provides an invitation to carry that bond forward. The design for these invitations is very different in each game.

If it’s your intention to allow players to grow their bond beyond the game, you’ll need to eventually release them from the cave and relinquish your control as designer.

I’m going to reserve discussing the ending of WAY specifically. Instead, I’d like to share one last personal story that speaks to its spirit.

For those of you that aren’t familiar, the iPhone app Ocarina transforms the iPhone into the flute-like instrument from The Legend of Zelda: Ocarina of Time. I’m a big fan of this app, and I use it in a very specific way. There is a feature in Ocarina that allows you to listen to people play as they stream their music live. Whenever there is a big event in the world, I like to open the app and listen to the songs being played there.

The day after the Tsunami hit Japan, I touched my finger over the eastern coast and waited for notes. I wasn’t sure any would come.

The song that came was “Ode to Joy”.

Those notes stunned me, and they stun me still. I immediately envisioned that person sitting amidst a horizon of rubble. I couldn’t process what they must be feeling, but I desperately wanted to send notes back.

I couldn’t.

That wasn’t part of the design.

As architects of games, we have full control over what is valued. If the world isn’t valuing what we consider significant, we have the responsibility to create worlds that do.

Our medium is Play, an instinctual language of action that reaches across all cultures. And aren’t actions what’s important?

What I will say concerning the ending of WAY is that players have the choice to make something.

It’s what you choose to make that reveals who you are...