• Tech
  • Design

‘The Witness’ Creator Jonathan Blow on Science, Language and Reality

27 minute read

In Jonathan Blow and studio Thekla’s The Witness, mysteries abound on a deserted island that may or may not exist. The island is beautiful but oblique, sublime yet functionally inscrutable. Glowing screens with maze-like grids are everywhere, connective cables snake through sun-dappled underbrush or down into cavernous passages. All of it leads players to ever-more bizarre, seemingly impenetrable line puzzles. It’s weird and gorgeous and categorically defiant. The Witness was widely quantified as a game created by a genius for geniuses, though it was accessible to all.

Blow began working on The Witness in 2008 with proceeds of sales of his first game, Braid, read by many as an elliptical critique of the heuristics in side-scrollers like Nintendo’s Super Mario Bros. Critically ballyhooed, Braid helped stoke the indie games movement that emerged in the latter 2000s, when it became financially possible and counter-culturally vogue to craft disruptive to mainstays like Activision’s Call of Duty or Microsoft’s Halo.

TIME spoke with Blow recently to explore some of the broader points raised by The Witness, as well as where some of his iconoclastic ideas about game design might lead down the road. Here’s what he told us:

TIME: I don’t want to flatter you for an hour, but I’m a little in awe of what you’ve managed to do with The Witness.

Blow: Well I’m glad you’re enjoying it. Compared to previous years when I was in game development, I don’t actually talk to many people about the game or read what people say about it on the Internet and stuff. So I’m not inundated with flattery of any kind, I guess, which is good, because you don’t want to get that way. But yeah, I’m glad you’re getting a lot out of the game. That’s good.


How are you feeling four months out about the general reaction to The Witness contrasted with Braid?

It’s hard to compare, because I was paying a lot of attention to the reaction to Braid. And with The Witness, I think two things have happened. One is I don’t really want to pay attention to general reactions. And two is just that the climate has changed. So even though I think The Witness is a much bigger, more interesting game than Braid, if only by size, forgetting the quality or interestingness of any individual element, there’s just so much more in The Witness.

But I think that the climate has changed, such that there’s just a lot of things to talk about now, and I think, relatively speaking, it doesn’t get as much notice as Braid did back in 2008. Maybe that’s wrong. It might actually be just my attitude to current news, but I feel like that’s true. I feel like more people talk about Braid. I mean more people play The Witness. It’s sold more copies. Even if it had only come out on one platform like Braid did, it would have sold more copies. But I feel like the critical reception is different, though again, I don’t know because I’m not paying that much attention.

In an interview about Braid a few years ago, you steered into a weedy conversation about metaphysics and the nature of scientific inquiry. You talked about critics missing the point in some of those reactions to Braid, which focused instead on cultural readings. And I think you said something like ‘Well, there’s actually all this other stuff I was thinking about, like science and quantum theory, and Braid is really about a lot of that stuff.’

What was your process for approaching those concepts in The Witness?

Yeah, so back to 2008 when I had the initial picture for The Witness in my head of ‘Alright, this is what I’m going to make.’ That picture changed during development as I fleshed out the details, but part of the original picture was that it was a game that was very concerned with issues of what . . . I don’t want to say what it means to exist, because that’s a Western existentialist way of phrasing the idea. But concerned with understanding the truth of the world. That’s the through-line of the game.

To the extent that there are puzzle-solving activities, on the one hand you can take them literally. Like here’s a puzzle, it’s interesting in the abstract as a puzzle. But there’s also a layer at which the puzzles are at least metaphorical and about understanding the world that we’re in. Basically, before you even talk about anything that’s harder to find in the game, if you just look at the individual panel puzzles that are very obviously placed around the environment . . . I say this a lot in interviews, but they’re about ideas. They’re not just a thing you solve. There’s an idea behind each one, whether you consciously realize it or not. That’s part of what makes the puzzles compelling, that the idea goes into your head a little bit.

Are there investigations we can undergo in games that get us closer to the truth about the world we live in?

So what are the ideas? Are they anything? Not really. What they are is an exploration of the things that can happen when you’re in a simpler version of the world we live in. So you have light and shadow, and you have colors and shapes occluding other shapes, and there’s an exploration of ‘Let’s make this as simple as we can and look at it with the greatest degree of focus that we can and see what we can see, and what is that like?’ Not even necessarily in a high-minded philosophical way, but let’s experientially look with fresh eyes upon this activity of walking around in a world from day to day, before you even add in other people that send you off into a weird mental place and all that. And then some of the panels are even more primitive. The first ones are more abstract, they’re pre-spatial. So here’s the black and white spots, and you need to figure out that you need to draw a line separating them. That’s an attempt at engaging whether there’s some kind of Platonic idea of category or space that precedes what you get when you have a full 3D world-like space that you can walk around.

This is a rambling answer, but the point is that those things all work together on a few levels. On one level it’s just, ‘Hey we’re getting the player into the mindset of looking with fresh eyes upon a world.’ Even if they don’t understand what’s happening, that’s fine, that’s just what we’re doing. But then also it’s metaphorical. There’s a metaphor for being a person in the real world just trying to understand ‘What is the truth about where we are? Are there investigations we can undergo in games that get us closer to the truth about the world we live in?’


Speaking of rambling, the writer and neuroscientist Sam Harris talked recently about his reluctance to transcribe his podcasts. He said something like ‘It’s because the spoken word is its own medium. When you transcribe spoken words, what you get is really just bad writing.’

[Laughs] Yeah.

He had the astrophysicist Neil deGrasse Tyson on, and I want to frame this next point by saying I’m not probing for ultimate answers in the game, because I know that’s not what it’s about, that there’s not some hidden prize at the bottom of the Cracker Jack box.


Harris was asking deGrasse Tyson, in so many words, ‘Why are you so cautiously apolitical about everything?’ And deGrasse Tyson responds that it’s because he wants to expose more people to science, to carry scientific ideas past peoples’ ideological forcefields. And as I was playing The Witness, it felt similar, like there was this nonlinguistic colonization of my brain happening.

It’s not that it doesn’t occur in other games, but it happens so subtly and uniquely in The Witness. You’re conveying this elaborate, interconnected grammar that lets me solve real problems, and that’s then drawing together this network of deep existential and epistemological ideas. It feels like you’re really making a point about games as a way to have genuinely meaningful dialogue, or maybe games as nonverbal educational tools.

Well for sure, definitely that’s true. I’m interested in nonverbal communication broadly. However saying that, it’s also easy to take for granted how astonishing linguistic communication is, and that we developed that and use it so effectively. Without linguistic communication we wouldn’t have much of a society. Almost all of it’s built on that, and that’s good. But it’s also very easy, because we’re linguistically communicating all the time, to get confused and think that linguistic communication is what communication is and that there isn’t a different kind. And you can look at everyday examples to see that’s not true.

The idea that I'm not going to obfuscate is the same as 'I'm going to allow whatever communication is inherent in this thing to come through more clearly.'

If you’re in a room talking with someone, you can see their body language, and that’s a kind of nonverbal communication. It’s a kind of unintentional communication even. Unless people are being manipulative, they’re usually not directly shaping their body language to send you messages. It’s just the communication channel that naturally happens between human beings. It’s easy again to marginalize that or think of it as a minor thing. And linguistics is a major thing. I’ve been interested for a long time in what kinds of communication are possible. Are there other forms of communication that we can refine in the way that we’ve refined language? What would those look like? Even if they’re more intangible than language, and even if you can’t refine them as much, I’m interested in that.

I noticed it first with Braid. When I started out designing Braid, I didn’t mean for it to be a game about nonverbal communication, but a lot of nonverbal communication ended up happening, because I had these other ideas about what the design should be. I decided the levels shouldn’t have lots of red herrings, and that the puzzles should be minimal, and they’d come through more strongly because they weren’t obfuscating the solutions. When and if they’re hard, they’re hard because you genuinely don’t see the possibilities. The idea that I’m not going to obfuscate is the same as ‘I’m going to allow whatever communication is inherent in this thing to come through more clearly.’

I came to understand that over the process of designing Braid, and that’s why I decided to focus on it in The Witness. That’s why The Witness is so minimally interactive. You hardly do anything. You can’t jump. You can’t pick things up. You don’t have a sack that you put objects in and pull them out later to look at, which is contrary to the tradition of adventure games. Even in Myst, which was a similarly minimal game, you still picked up book pages. The Witness has absolutely none of that, because I wanted a clear slate for this communication to happen. It’s also why the puzzle panels are so obviously puzzle panels. It’s a hard thing to communicate in the messaging for the game when I’m trying to get people interested in it. People have this reaction, ‘Why would I be interested in a game where you just walk around and draw lines on a bunch of panels? Why is there even a world there? Why is this not just a cheap iOS game or something?’ There are very good answers to that, but you don’t want to give people those answers because you then spoil the game for them.


It makes for a really interesting situation. But the reason I’m in that situation is because I wanted extreme clarity, at least around the first layer of puzzles. ‘How do I recognize when there’s a puzzle? What do I do? What is the mechanism by which I solve it?’ You always know that you’re drawing a line on the panel. And then the question becomes ‘What is this line?’ And that’s the point at which communication starts to happen. It’s ‘How do I know what line to draw?’ We focus and we strip out all the confusing things about video games, like ‘Am I pushing this button at the right time or not? Can I take out this object or not? Should I be pulling this lever?’ And so forth. You get rid of all that stuff and it lets you focus on the communication.

So as you mentioned before, it’s not like this kind of communication doesn’t exist in other games. It totally does. But I don’t know of any other games that have focused on it exclusively. And that was my interest as a designer, to focus on the communication and really look at it. Because, and I know this is a long answer, I feel like we don’t yet understand what games are capable of as a medium. And there’s not enough genuine interest throughout the game industry in dealing with that, because people have figured out how to make money. And that’s great, at least people have figured out how to make money for now by employing old gameplay discoveries in a continuously refined way, and-or borrowing things from other media.

So you have some very successful games like Uncharted 4, and it’s very successful and well-reviewed. But it’s also trying to be a little more of a movie than a game in terms of the value that it gives you. It’s trying to give you story and characters in the way a movie does. The point being that you don’t see many people genuinely exploring the limits of what games can do right now, and I think we need more of that.

The Witness was my attempt to do that, both in the abstract sense and to show other designers ‘Look, here’s what I came up with,’ but also to do it in a commercially successful game, which is that much harder. It’s very easy to get all experimental and crazy and make something not that many people want to play. And then it maybe goes into a video game museum somewhere and that’s it. And that’s fine. But I also always want to show people that experimentation like this can result in things that are very exciting and genuinely interesting to many people, and that it can be commercially successful. So I’m glad The Witness has been commercially successful for that reason, and of course because it keeps me in business.

The thing I find admirable in that, is that you’re speaking boldly in this era of attaching self-deprecating tags to everything. That if you’re seeing something you think maybe no one else is seeing, you’re not afraid to say so.

Video games are in a weird spot now. I feel like we’ve been living through this time of anti-intellectualism across the culture—for the past few decades at least, but in video games especially. I mean crazy anti-intellectual. Part of that is because so much of the intellectualism we’ve had in video games is actually really pretentious and dumb. I feel like we’ve seen a lot of people just trying to be the person who says smart things about games, instead of doing the work to understand gaming well and discover things and then explore what those discoveries entail. And I think people have rightly reacted negatively to that sort of behavior. It doesn’t mean there aren’t people doing that work and genuinely figuring out what games can be and pushing them forward. I just hope that eventually we can get to a stage where that work’s more broadly celebrated as part of the medium, say in the way that film does.

I feel like we've been living through this time of anti-intellectualism across the culture—for the past few decades at least, but in video games especially.

There’s the stereotype of the pretentious director who makes crap that nobody wants to watch. Despite that, there persists this idea that you have directors working hard to make visionary films. They’re not the biggest films or the biggest blockbusters, but they’re deeply appreciated by a substantial number of people. I guess we’re going to have that in games, but I don’t know right now. Maybe we already do and it’s hard for me to see, because I’m too inside it.

Speaking of pretension, academic critics started talking some years ago about the way the hero story in so many games clashes with the fact that you’re really a sociopathic maniac who winds up killing a bazillion people by the end. Someone coined the phrase ‘ludonarrative dissonance’ to describe it. And Uncharted 4 developer Naughty Dog made that phrase the name of an achievement after you’ve killed so many people as a wink at some of the pretension.

I’m wondering about that tendency in academia to want to colonize a thing with jargon and obliqueness. It seems like the opposite of what you’re up to in The Witness, where you’re grappling with the nature of what we think we know, but exploring it through quotes and videos from scientists and philosophers who speak in totally straightforward and comprehensible ways.

There are a lot of directions I could go in from here, but let’s just say you could start from some of those very smart scientists who are quoted in The Witness, say Richard Feynman. Brilliant guy, discovered a lot of very smart things, very celebrated, very plain speaking. When he’s trying to communicate, he’s really trying to communicate, and you can understand what he says, even though some of it is very hard to understand, actually. And it’s usually not hard to understand because of the way he’s explaining it, it’s because the concepts are deep. Physicists generally as a group—I study physics on the side a bit just to try to broaden my understanding—are generally like that. They’re not all gifted communicators. Some are very poor communicators, just in that way that some people are very nerdy and introverted and don’t speak well.


In general, the tradition in physics, and I would say in many of the sciences, is to be very plain speaking, because you genuinely want the audience to understand what you’re saying. Because if they don’t, if you try to make things more complicated than they are or obfuscate, nobody will understand because the material is legitimately hard. It’s fortunate that that’s where the scientific culture is today. I don’t think it had to happen that way, and maybe in the future scientists will be more about obfuscation and jargon. But at least at present, to the degree that there’s jargon in scientific disciplines, it usually has good reason for being there.

I meant jargon more in the academic sense, where you sometimes see simple concepts deliberately obscured or elevated to the level of fetish.

The thing about ludonarrative dissonance specifically, if we go in that direction, is it’s actually . . . I don’t like the term, because I agree it’s a brick wall of a term, and those are not the kinds of words that I like coming out of my mouth. Usually when I see the words ludonarrative dissonance, it’s a signal to me that the person probably isn’t saying stuff I’m going to find interesting, because it’s usually followed by a lot of other jargon.

That being the case, I agree that what it’s referring to is actually a problem, and I think Uncharted specifically has this kind of problem. It’s not the game with this problem in the deepest, most self-defeating way, but it’s a problem. The thing about the term ludonarrative dissonance is it’s overly specific. It makes it seem like this is something peculiar to games, when in reality, you can more broadly look at any work of art or even things that aren’t art and see when they’re being hypocritical or self-defeating or inconsistent. These are much easier words for people to understand, that mean the same thing and that are not peculiar to games.

You can have a movie that’s supposedly about love and family or something, but the screen time the movie devotes to those subjects is 15 minutes, and the remaining hour and 20 minutes is the guy being an action hero kicking guys off the top of trucks driving down the freeway or something. And it’s like, ‘Okay, well, this movie isn’t really about love and family, because those other guys have families and stuff, right?’ And the movie in fact wants to dehumanize those guys so it’s okay to kick them off the tops of moving trucks. And the business model of the movie is to have this kind of action excitement, and that’s why you’re in the theater. So if the family part is just the thin veneer to justify the action scenes, then it’s really a deeply hypocritical movie in some way, or at least parts of it are.

It’s the same in games, it’s just more extreme, because we have this history of not knowing how to make gameplay aside from shooting stuff that goes back to video arcades. Okay, that’s not entirely true, because back in the eighties we had a lot of games about not shooting stuff, but those are in much more of a minority now, I think, unless you’re counting iOS or whatever. So when you call it ludonarrative dissonance, you make it seem like it’s this hard thing to understand, when I think it’s really easy to understand.

My favorite extreme case is [Grand Theft Auto studio Rockstar’s open-world Western game] Red Dead Redemption. It has this thing at the end that’s supposed to be the touching finale, where you go back and take care of your family, having brutally murdered some 800 people, including at one point pillaging a poor Mexican town to get in with some group. So you’ve literally firebombed a Mexican village, and now it’s supposed to be this sweet family moment. And then you go on the stat screen and it says ‘Guys killed: 860.’ It’s on the screen because you’re supposed to care how high the number is. Those things are incompatible for any reasonable human being, and when a game tries to pass it off as normative, it’s jarring.

When people say these games have good stories, I think they're only comparing them to other games. If you compare them to good movies or good novels, the stories are terrible.

At the same time, it’s not an easy problem to solve. So I know why Uncharted 4‘s developers made the game the way they did. I just don’t agree. I wouldn’t do the same thing. When people say these games have good stories, I think they’re only comparing them to other games. If you compare them to good movies or good novels, the stories are terrible. Or at best, they’re like a story from an episodic TV show where there’s some villain and someone trying to defeat the villain. It’s very simplistic. Even all these games that win awards. Where games are in terms of storytelling . . . it’s not Moby Dick, it’s not Finnegan’s Wake, it’s not Pride and Prejudice. It’s Airwolf, or Knight Rider. Uncharted is probably on the level of some of the better TV shows, but it’s not good. It doesn’t do justice to the millennia of storytelling craft, and that’s not me trying to say bad things about Uncharted, it’s about all story-based games. I don’t feel like they’re there in terms of that craft of storytelling. And part of the reason, but only part of the reason, is what some have taken to calling ludonarrative dissonance.

So I don’t think it helps to thumb your nose at it as a developer, because it prevents you from getting to a more sublime place. If you really want to have a moment about family, or caring about your friends, or about the awe of discovering an ancient artifact, that’s all degraded by ‘I just shot 100 guys.’ Because that’s just ridiculous. Those things can’t live together.

I have one final point about this. The last Uncharted game I played all the way through was Uncharted 2, and it’s trying at various points to be a comedy game. At one point it tries to make a joke about getting frustrated with doing the same thing over and over. But it didn’t seem to understand that making a joke about a problem that’s true doesn’t make the problem go away. And so the ludonarrative dissonance achievement strikes me as the same thing. You can make the joke, and ha-ha, it’s a little bit funny, but it’s not really funny because the problem is there. Saying the name of the problem doesn’t banish it.


Are these fundamentally incompatible approaches to storytelling? You’ve heard the argument that games will never be like novels or movies, because delivered narrative and player agency aren’t compatible. This gets semantically tricky, because of course you can read a book or watch a movie any way you like, skip to the middle or end first…

Yeah, but nobody does.

It’s true, I’m just thinking about the exceptions to the rule, because people tend to represent the older mediums as static where they’re not.

It’s just that it’s hard. Again, I don’t want to come across as being super-negative about Uncharted, because those folks are making a very well-crafted version of a story-game that a lot of players do appreciate. And it’s the state of the art of what we know about doing a linear story in a game right now. And that’s fine, because as soon as you try to go in another direction, you’re grappling with unsolved problems. That’s really research territory, not commercial territory. I don’t mean research should only be the realm of academia, I just mean that if somebody’s goal is to make money, they’re not going to do that stuff because it’s hard.

My approach to narrative in The Witness was obviously different. With the original idea of the game, I was going to have a little more of a narrative. There were going to be two sets of audio logs, one of which was a bit more like a traditional story. I decided to get rid of that stuff in the end because I decided that sort of narrative wasn’t what mattered and what I really cared about was the ideas. I didn’t want to fictionalize the ideas any more than necessary, because I felt that that weakened them.

In what ways might we see you trying to grapple with some of these issues in future games?

There’s some stuff you see examples of in Braid and The Witness that I think I’m going to carry further in different ways in future projects. There’s this thorough way of exploring the space of possibilities given some initial ideas. In The Witness it’s probably easier to see, because there’s so much of it. You get a simple idea, like ‘What if shadows on a panel can help you determine what the answer is?’ And then there’s this play of form, like ‘What are all the variations on that? In what way does that combine with other ideas from elsewhere in the game?’ That’s one level at which things can interrelate and play. But also then within its own subject, within the idea of a shadow landing on a panel, what are the things you can do with that?

I think in that particular area The Witness ended up being mostly internal to the idea itself, rather than combining with other ideas. In other sections, we go more outward than inward. So there’s this play of form that happens. It’s almost like a celebration of possibilities. And this is part of what happens in reality. If I were to say anything about reality, it’s that things influence and interrelate and cause increasing situational complexity. You think that you have one or two or three simple things, and then you find that the way those interact creates new phenomena or sub-phenomena, and then those phenomena interact with the original things, and eventually you have a quite complex tapestry.

If I were to say anything about reality, it's that things influence and interrelate and cause increasing situational complexity.

In The Witness, I was doing that deliberately, with regard to exploring the topic of puzzles. But you could also design a game that tried to engage it in a different way. So for example, you have games that do procedural generation, sometimes to good effect, like Spelunky, where they build the levels, or Dwarf Fortress where it generates terrain. That’s a thing that’s been around in games for a long time. And it has this quality where it generates these kinds of surprises, these kinds of emergent complexities.

That said, in the same way that all games have this nonverbal communication but designers don’t ask the question ‘What would this be like if we focused on it exclusively?’, designers using random generation and recombination and emergent complexity in games are just doing so to generate funny or quirky gameplay situations.

I’m interested in making something that looks at the phenomena itself. Procedural generation feels intrinsic to reality. And in the Eastern spiritual traditions—granting that’s an overgeneralizing phrase—there are ideas both in Hinduism and Daoism about the universe being the result of a kind of play of form, of this kind of complexity that emerges from simple beginnings. I find that very interesting. I have something that I’ve prototyped—though it’s probably not the next thing you’ll see from me—that plays in that space. But it doesn’t do justice to the idea entirely, and it’s not even close to being finished.

More Must-Reads from TIME

Write to Matt Peckham at matt.peckham@time.com