Zork and the social impact of design choices

Return to Zork is a 1993 a point-and-click sequel to the text-based Zork games, notable for being one of the first video games to feature live-action (adjacent) performances by Hollywood (adjacent) actors. It was a blockbuster of sorts when it launched, but it’s relatively obscure today.

It’s also…not very good. At all. 

Like a lot of failed art, the specific ways in which it’s not good are fascinating. The problems point to potential improvements, which in turn make me think about how design choices directly influence the way people engage with art and, in the world of software, interact with systems. 

In a weird way, I think this obscure point-and-click adventure game taught me why humanity isn’t necessarily doomed. Let me explain. 

Return To Zork is a series of puzzles that are, in theory, solved by observation, experimentation, and reading a lot of files in the mayor’s office. From playing as a child, I mostly remember my siblings and I clicking everything until something happened. 

The game infamously begins with a sequence that, done incorrectly, silently makes it impossible to win later on. You’re standing at a mountain pass when a vulture flies overhead and lands on a nearby sign. Below the sign is a plant. It’s one of the few interactive objects here, so naturally you want to try to pick it up. Do that, though, and the vulture will attack you (try three times and it will kill you). There’s also a rock nearby, though, which you can pick up. Holding the rock and clicking on the vulture brings up the option to “throw rock at vulture”, which scares the vulture away. You’re now free to pick up the bonding plant.

Logical enough. But here’s where the game gets cruel: pulling up the bonding plant kills it, which makes the game unwinnable. The game doesn’t tell you this, though, and you won’t figure it out until halfway through the game, at which point you will need to start over (if you’re still interested in completing the story, at least). 

So what are you supposed to do? You can dig up the bonding plant using the knife in your inventory—do that and it won’t die. That’s assuming you, first, noticed you started the game with a knife in your inventory; second, imagined a knife is a good tool for digging up plants; third, assumed that there’s a meaningful difference between pulling up and digging up a plant; and, fourth, that you even realized the plant being dead or alive had any kind of consequence in the first place. 

It’s a lot of logical jumps to make. Most players don’t make them. That, in turn, makes the game feel unfair, which changes the play style. 

The goal for the designer, I imagine, is for the player to be immersed in the world and narrative—to interact with the story. Hostile design like the plant puzzle makes this impossible. The player stops thinking about the story, focusing instead on brute forcing their way through the arbitrary puzzle designs. 

I’m writing this during a period of underemployment. I’ve seen it coming for a while—a lot of what I write depends on search engines for traffic, and AI summaries at the top of search results means fewer people are clicking those links. The result: I have fewer paid gigs right now. A user interface change means I can’t write the kind of articles I could before (at least, not if I want an audience, let alone a paycheck). 

I’m in a Slack chatroom full of other technology service journalists. Lately the conversation is dominated by news of layoffs and requests for help finding a role, an apocalypse entirely created by Google’s product choices. The sense of mourning is palpable. 

Three years ago I wrote for The Atlantic about how Google’s AI plans essentially strip mine the web, removing the economic incentive for publications to post things online. Not to brag, but I called it—I, and people like me, are completely screwed. I feel very smart. 

There’s going to be far fewer articles published on the web that aim to help the reader solve specific problems. That, in turn, could mean AI bots—Google’s included—have fewer helpful articles to summarize. I’m confident that in thirty years, people will talk about this as a problem that was predictable. If you’re reading this then, well, know that it was predicted.

Return to Zork, thirty years ago, was released to mixed reviews. In 1993 Charles Ardai, writing for Computer Gaming World, praised the visual effects and the inclusion of actual actors, but criticized the puzzle design. He also found the conversations frustratingly limited. 

“Much of the time, one can click all one wants and still never get a character to say anything substantially different,” he wrote. “In the process of learning this, however, one ends up sitting through every conversation four or five times.” 

Interacting with people is something extremely hard to get right in video games, even today. Modern games typically use what’s called a conversation tree, where the player is presented with two or three responses to a character’s statement. The player picks one, hears the reply, and is sometimes given some sort of quest. 

Return To Zork pre-dates this modern convention—there are no conversation trees. The player, instead, navigates conversation the same way they interact with objects—by navigating visual menus. The player can “show” any inventory item, place on the map, or photo taken with the in-game camera to any character they meet by clicking around. This is, arguably, a deeper form of interaction than the modern conversation tree—you can almost approach it like a human conversation. Unfortunately, most characters have nothing to say about most things, even when they logically should. Yet there is vital information buried in this conversation system. 

Return to Zork‘s design incentivizes players to play a certain way—to methodically click every item in their inventory in combination with every item they come across, and to ask every character about every single item possible, and seeing which few trigger a useful response. 

Other systems, outside video games, also shape the way we interact with the world. If you’re talking with someone on the phone, they can’t read your facial expressions, so all the emotion you convey has to come across in your voice. If you’re texting someone it’s even harder to convey tone, which is part of why people use emoji—it fills in part of that gap. 

Use a tool long enough and navigating these rules become natural. This generally happens without you even noticing because humans are adaptable. We navigate new communication technologies the way we do riding a bike—practicing until they become second nature.

But just because we’re not consciously paying attention to the rules of a communication system doesn’t mean they’re not impacting us. Scroll through social media and it doesn’t take long to realize that certain ways of understanding the world come across more naturally than others. That’s not natural, or inevitable—it’s the result of the way these systems are designed. 

A Return to Zork player approaches conversations with in-game characters in a particular way, clicking every item in the inventory to try to get a response. That is the result of design choices. The shape of our conversations online are influenced almost as directly. These systems don’t reward things like nuance, connection, or understanding. The commercial internet is designed to capture as much attention as possible, meaning only voices who attract attention—postive or negative—are rewarded. 

Looking back at Return to Zork now—with three decades of game design history in our toolkit—it’s easy to imagine a better version of the game. The conversation system, for example, could punish people for spamming questions—characters could become uncooperative if you keep asking about the same thing, for example, which would make the conversations feel more lifelike. It would also reward the player for connecting concepts. Solutions to puzzles could be a bit more based on logic, and less on trial and error. And, most importantly, it shouldn’t be possible to break the game without the player knowing that happened. 

These solutions may seem obvious to contemporary gamers, mostly because we have the benefit of thirty years of improvements. This is how culture works: we learn from our mistakes and change the way we make things. The idea of people smoking inside a bar or restaurant is, in 2026, disgusting to most people. We now understand more about what second-hand smoke does to the body, a lesson that in turn resulted in social change. 

We’re in a similar period of realizations when it comes to the commercial social web. I feel like most people know the internet status quo is not healthy, for them or for society. And I have no doubt that, in 2056, dopamine-driven social networks will be viewed with the same level of derision that smoking is now.

Is that too big a takeaway from a game as silly as Return to Zork? Yes. But you just read 1500 words about the social impact of design choices and I haven’t made a single reference to the “want some rye” meme until now, so I think I’m okay with it. Thanks for taking the time to read all that—I’m looking forward to continuing the conversation.