Linearity is a bad word.
When it comes to games, linearity is one of those dreaded concepts no one dares say aloud, for fear of angering the internet. It's not the worst concept though. That terrible title probably belongs to "free-to-play", or maybe "full reactive eyes entertainment”. In fact, if you just go back a few years, linearity wasn't the hot buzzword everyone loved to hate, it wasn't even talked about. Certain games were set aside as being open world games, because linearity was the standard. Now, the script's been flipped. The last major AAA release I can think of that didn't feature an open world was Call of Duty: Ghosts, and its single-player campaign is by no means the "point" of that game.
Meanwhile, every multi-million dollar series worth its salt features an open world, each one claiming to be bigger, open-er, and world-ier than the last. Assassin's Creed IV, Batman: Arkham Knight, and of course Grand Theft Auto V lead the charge, coming from a relatively long line of open world predecessors, but even brand new IPs, like Watch Dogs and Sunset Overdrive are launching as big, open world games. Not that we're seeing much in the way of new IPs these days. Metal Gear Solid V: Ground Zeroes is taking the Metal Gear series into an open world for the first time, and the recent Tomb Raider reboot combined its spectacular Uncharted impression with its best shot at placing Lara in an open world that didn't really matter.
Open worlds are basically the MSG of game mechanics. You just add a little, and it makes everything seem to taste better. Or, the way publishers see it, adding an open world is a guarantee that your game will sell better than if it didn't. Because open worlds sell, you see.
And why wouldn't they? Video games are, at their core, about interacting with a world and having agency over it. A big open world that doesn't lead your anywhere by the nose is pretty close to the ultimate expression of that concept. Some people see video games' endpoint as the holodeck from Star Trek, a fully immersive, totally realistic simulation of whatever you want to experience. They aren't terribly far off, at least if the progression of open world games is anything to work off of. Add a health bar and a wanted meter, and Picard is basically playing Grand Theft Auto: XXVII.
But there's another school of thought here, the idea that games are a focused exploration of a specific set of concepts and mechanics, that they shouldn't try to be everything, because you'll never perfect that. Games didn't start as trying to replicate the holodeck experience. At first they were trying to replicate ping-pong, to be fair, but then we got into focused looks at more fantastical mechanics. Mario let you jump twice your own height, and asked you to figure out how to best control that. Mega Man put you through a gauntlet of mazes and traps, then asked you to learn from the environment and figure out how to best use the tools you'd acquired from bosses. Even Adventure wasn't trying to simulate something so much as it was trying to help you learn that dragons could look like ducks too.
But, in order to properly explore those mechanics, those games had to be linear. Directed. Focused. It wasn't a technical limitation either, considering that River City Ransom came out in 1989. To be fair, RCR was a pretty small, simple take on the open world concept, but it does show that the idea not only existed, but was possible, even in the early days of game design. But, RCR was big and spread thin. There wasn't a ton of complexity to it, most of the fun was had in seeing how the open world and mechanics could be abused. It introduced the world, or at least the few people who had played it at the time, to emergent gameplay, which would go on to become one of open world design's strongest selling points.
The point is, games were linear for a long time for a reason, and it wasn't technology. A linear, heavily directed experience is the best way to show players how to best take advantage of deeper mechanics, the easiest way to adjust the difficulty curve, and the easiest way to tell a story. Look at Ocarina of Time's Shadow Link battle. Link enters an empty room with an island in the middle. You walk over to the island, nothing's there, but you see a door on the other side of the room. You check out the door and it's locked, so you turn back the way you came and suddenly Shadow Link is waiting on the island. It's a simple experience, but distinctly affecting, and one of the most memorable parts of that game. Obviously it's just a tiny moment in one room, but it works as a microcosm of Ocarina of Time's design philosophy as a whole:
Make the player think they have agency.
The best linear design is all about illusion- tricking the player into assuming they have choices over how events transpire. In the locked room, Link can go anywhere, but reaching the other door will trigger the encounter. So the developers put the island in the middle of the room, giving the player an extra stop on their journey, a point of interest that delays the inevitable. You have a choice of stopping at the island, but you have to go to the locked door, no choices there. Of course, you don't really choose to go to the island either, since it was put there specifically so you'd notice it and go there first. It's a magic trick of game design, perfect direction that wouldn't be possible in a nonlinear experience. Ocarina of Time does it the whole way through. Hyrule Field opens up to four or five areas, but you can only get so far into each before having to turn around due to missing equipment. You can choose to hit up the Zora river before Death Mountain, you just won't get very far. In the end, Death Mountain has to come first.
It's a great way to tell a story, and a great way to prey on player expectations and surprise them. Of course, open world mechanics do worm themselves in everywhere, because more choice is actually a pretty good selling point. If games really are power fantasies, then choice is what makes us feel powerful. The more choices, the more power, and the more those choices affect the world the better they become.
Of course, when you get to a situation like Telltale's The Walking Dead games, which feature choices with real consequences to them, most of the power is stripped away. Power doesn't react well to real consequences, it just wants immediate gratification. If I choose to ramp this car over a bridge, I want to see a sweet jump and maybe an explosion at the end, not get a ticket and file paperwork for the damages I caused.
Which is why everything is open world now. Power fantasies are not only in right now, they've defined games for a very long time. It may suck for some, but it is true. All those big, AAA open world titles I mentioned up top are all power fantasies. The bigger a world is, the more thinly spread it is as well. Choices don't really have consequences in an open world because they can't. The whole world can't shift so easily, that would require a massive amount of assets that most developers just don't have. It's why Mass Effect's and Infamous's moral choice choice systems are all smoke, mirrors and fluff, and why Telltale's the Walking Dead is such a small, linear experience. An open world is better suited to power fantasy, because it inspires choice without consequence, while a linear experience is better used when trying to tell a cohesive story.
It's why linear shouldn't be a bad word either. Linearity lets you focus on mechanics and refine them to perfection, lets you get players caught up in a focused narrative, lets you construct a difficulty curve that makes sense. Opening up the world and letting the player mess around as they choose throws those things out of balance. Which isn't to say open worlds are bad. They aren't just mindless power fantasies, they can only be huge worlds to explore, or have rich histories to discover, like Skyrim, Shadow of the Colossus, or Wind Waker.
So, no, linearity isn't a bad word. Neither is open world. Instead of thinking of them as positive and negative concepts, maybe we should start thinking of the two like we do writing perspectives. A novel can be written in first person or third person, with an omniscient narrator, or an unreliable one. They're different tools, with different uses, and each one is best suited to a different kind of job.
Except for the free to play tool, that tool broke years ago, no way we're going to fix it now.