The Last of Us: Part II - Part II: How Narrative Design Fails
TLoUII's real failing is in its inability to tell its story through gameplay.
This is the remastered version of part two of my three-part deconstructive analysis of The Last of Us: Part II. Find part one, about Ellie’s misrepresentation, here; find part three, about how horrible the story is as a whole, here.
Games are games
Games are interactive. This is what separates the medium from books and film. Consequently, games are not meant to have writers, but narrative designers. The role of a narrative designer may look very similar to that of an author or screenwriter, depending on the type of game, but it is nevertheless his job to coordinate story with gameplay systems.
The Last of Us: Part II does not tell a good story. If you were to put it into a different medium, it would still fail. But its largest defect is not its writing per se, but rather how the narrative is designed as a whole, and how the game fails to wed its interactive and non-interactive components.
Part II, not 2
The title is not an accident. The Last of Us: Part II seeks to continue the first game. It is not a spin-off, not just a sequel, but a direct follow-up on The Last of Us’ ending.
The second part of a game cannot only continue its story. It should also continue its systems. In many ways, this game does feel like the completion of a missing half of the first game. Allowing Ellie to dodge, jump, and crawl advances the stealth mechanics logically: she is small, acrobatic, and sneaky. Why couldn’t she do these things before?
But The Last of Us on PS3 was not just a stealth-action-survival game. Fundamentally it was about Joel and Ellie and their developing relationship. The Last of Us is about this in the same way that Halo is about being a SPARTAN-II super soldier. It uses its interactivity through combat and exploration to reinforce the narrative development of the protagonists’ bond. Ellie begins the game as a mostly useless child. She quickly finds a role scavenging for supplies. Later she helps Joel in combat directly. She becomes useful, and that usefulness increases the player’s opinion of her.
This happens in parallel with Joel learning to love her. The player experiences his arc through gameplay.
When Joel is assassinated during the prologue of Part II, this link to the original game is severed. These games can no longer be systematically “about” the same thing. It is now literally impossible for Part II to be the second part of The Last of Us.
But although the narrative designers did not seem to remember anything about the first game’s characters, they did apparently understand that The Last of Us was meant to be about relationships. This is why Fake Ellie’s quest for vengeance begins with Dina tagging along.
Dina’s presence introduces a major ludonarrative problem. In a game that is supposedly the second part of a beloved story, we are now traveling with a character who does not appear in the first game and whose intimate relationship with Fake Ellie has been developed entirely off-screen.
Audiences will be naturally apprehensive about new characters. The Last of Us instantly forms a bond with Joel through the trauma of his daughter’s death, then uses the character’s wariness over going on a gigantic escort quest with a child to mirror the psychological state of the player. Joel learns to love Ellie over time, but it takes a lot of work. It takes a lot of work to get the player there, too.
Joel does not begin The Last of Us already loving Ellie because the writers have skipped three months ahead and teleported us halfway across the country. That would have ruined the game.
Yet this is literally what Part II does. We return to Fake Ellie’s body and are instantly subjected to Dina at a high level of intimacy without being given a single moment to learn to know who she is or get to like her. I will not criticize the dialogue or performance of the character; both are serviceable, and I’m sure some players find her to be charismatic. But there is nevertheless an enormous disconnect between Fake Ellie and the player from the very first moment of this game.
I know, understand, and empathize with Fake Ellie due to the last game. I can hear her voice in my head. She feels like an old friend. But I have no ability to empathize with her love for Dina. Why is Dina special? What is the basis of this relationship? How have these characters formed a bond? They seem to be basically the same. I have no idea what their connection is.
The first game works extremely hard to create psychological cohesion with Joel. The Last of Us: Part II does not even try.
This would not be that big of an issue in a film. It is reasonable to assume that Fake Ellie would have found a girlfriend in our time away from her, and that two teenage girls might smoke pot and make out while on patrol. But we are not just watching Fake Ellie. We are inhabiting her. When she becomes directly physically intimate with a character the player has no relationship with, it is jarring and repulsive. These characters start the game already in love. This love happens without—even despite—your existence.
Left Behind spends an entire DLC showing Ellie’s relationship with Riley through gameplay to earn a single kiss. Part II thinks that we can start there.
Learning to love Dina was going to take an entire game. As we all now know, that is not what Naughty Dog decided to do with her.
Dina’s relationship with Fake Ellie evolves over the course of the story, but her relationship with the player is static (or even regressive). She starts as a moderately useful companion. She becomes useless baggage and disappears. There is no mechanization of her presence beyond this. We can draw no parallel with how Ellie’s usefulness in gameplay is ramped up over the course of her arc in the first game.
Even if you do not like Ellie personally in The Last of Us, you can understand through her mechanistic usefulness why Joel would want her around. But there is no reason why anyone would want useless pregnant Dina in a theatre doing nothing.
By the time Dina leaves and Fake Ellie goes out exploring on her own, it is safe to say that The Last of Us: Part II is not part II. It is 2. Its mechanics do not work to recapture the essence of the first game for some other duo. Indeed, while there are relationships, they do not form any meaningful throughline. The plot jumps around to an extraordinary degree, and the only character we ever really get to know through gameplay is Lev, in the back third of the game, long after we have been emotionally exhausted and are ready to be euthanized. This is totally disconnected from Fake Ellie’s sections and still comes following hours of Abby being alone.
This game is not about relationships. Instead, it is about extreme violence, and revenge.
This is where Part II does succeed in aligning player and character. Abby’s assassination of Joel is so brutal and heinous that any player will completely share Fake Ellie’s motivation: Abby must be killed. That is good narrative design.
Or it would be, if the way it was done didn’t simultaneously make the player hate the designers.
Abby & empathy
The introduction of Abby is a complete blunder of narrative design that is engineered to create antipathy. At the start of the second part of a beloved franchise, the player will be desperate to catch up with old characters. Our first interest is to see Joel and Ellie as we are eased back into the world.
The Last of Us follows a clear logic with its progression through its playable characters. The game begins with Sarah and transitions to Joel when she is physically incapacitated. We do not leave Joel’s body again until he is incapacitated. We return to Joel after he has recovered. Though he is not incapacitated during the epilogue, giving us control of Ellie mirrors the prologue with Sarah, and thus still has its own logic. We never jump into the body of a character we have not met or do not know.
The snapping between PoVs at the start of Part II is totally arbitrary. It happens at the whims of the designers. Nothing in the story takes us out of Joel and then Fake Ellie’s bodies; we are simply stolen away from them without reason, and forced into Abby instead.
We have no idea who this character is. We have never met her before. We have no interest in her. Now we must play as her, at the expense of playing as Fake Ellie.
Abby feels like a distraction. It is confusing and disorientating to suddenly be playing her. We may not hate her yet, but this is the first reason we are given to dislike her. The bigger one is yet to come. After Joel and Tommy rescue Abby from the blizzard, a cutscene ensues, and we encounter what is The Last of Us: Part II’s most pervasive problem.
Death by Cutscene
Joel is an unstoppable killing machine. Under our control he has butchered hundreds of infected, hunters, and smugglers. He once massacred an entire hospital of body-armored soldiers wielding automatic weapons while armed with nothing but his fists and Molotov cocktails (in my Grounded playthrough). He was impaled through the stomach by rebar, and he survived with nothing but a shot of penicillin.
Joel is a player character.
One of the only parts of Part II that I like is the brief sequence where Abby and Joel fight infected alongside each other. Joel’s AI kicks ass in combat. We witness his power through gameplay. We do not need to be told how cool he is; it can be shown through systems. Though it is predicated on a gigantic and absurd coincidence, the irony that he rescues his own killer is thrilling. What will come next? Will Abby be able to kill the man who saved her life? Will there be a boss fight?
No. Joel dies in a cutscene.
Near the end of the Grounded documentary about the making of The Last of Us, Neil Druckmann says:
Like Bruce and I talked in the beginning is like we have these mechanics. How can we exploit them in a narrative sense?
Troy Baker replies:
[Showing Ellie’s despondency with the ladder] is one of those opportunities. I just think it's incredible that that happens in game. It's not a cinematic, you know? It's it's something that adds weight to that controller when they're playing.
After The Last of Us: Part II, it’s hard to believe these two men once said these things to each other. But they did. And they were completely right. The Last of Us uses its interactivity in subtle but incredibly effective ways to enhance its narrative. This is why it endures after so many years.
Cutscenes are not interactive. They have a place in video games when characters are doing nothing but talking or to dramatically conclude the results of a gameplay sequence. This should be their use, and it is how they are used in the first game.
The death of Joel is the most important moment that will ever happen in the series.
It happens in a cutscene.
This is why this scene makes players angry at Neil Druckmann. This is why this scene is a complete and utter failure.
Imagine a version of Joel’s murder where the player controls him. You rescue a young woman, but as you fight beside her, hordes of the WLF start to attack you. You kill a few, but it quickly becomes apparent that you do not have the resources to win. This is going to be your final battle. Your luck had to run out eventually. When you are reduced to zero hitpoints, Abby comes out, and she finishes you off in a cutscene.
This is how John Marston and Noble Six die. This is how a video game character should be killed in a story game. There is no other way to do it, because normally a video game character has to win: you simply reload if you fail. Any other course feels fake and arbitrary.
Naughty Dog did have another option. They could have done this scene from Fake Ellie’s perspective and left us in control of her. Rather than bringing us into a cutscene, Fake Ellie is restrained, handcuffed, and forced onto the ground. The player continues to control the camera; you receive a QTE prompt, your controller rumbles, but no matter what you do, you cannot break free from the three men holding you down. You are then forced to watch as Joel is tortured to death. You can look away, but if you ever push the thumbstick too far in either direction, one of the goons kicks your head, and you are made to focus back on the murder.
Notice how effortlessly this hypothetical scene slips from third-person into second. In this version of the scene, you would experience Fake Ellie’s lack of agency, just like you experience Sarah’s lack of agency in the prologue of the first game. You would be placed at its center, not Fake Ellie.
This would have been a video game. This is, in fact, the bare minimum that should be expected from Naughty Dog. They have done this before. It would have been much more interesting than just watching a cutscene.
Both of my versions of this scene should have been combined. We should have played as Joel until his death in gameplay. Then we could have watched his final murder from Fake Ellie’s perspective, also in gameplay. These changes together would have significantly diminished the controversy around the game without altering the story itself.
The actual scene is much worse than I make it out to be. It is completely broken by its own logic. Fake Ellie sees Abby standing down the hall long before control is taken away from the player. As you press L2 and raise your pistol to shoot her, the cutscene forces you to lower your gun. Fake Ellie then decides to rush forward and let herself be grabbed from behind the doorway for no reason.
Stay in the hall, Fake Ellie. I was saving a Molotov just for this situation!
Like her confrontation with Abby later in the game, the cutscene that follows actively defies the game’s mechanics. There is no reason why Fake Ellie cannot kill all of these WLF goons, except that we are not allowed to. We killed huge groups of people at the same time in the last game. We will do it again in this game. If we fail, we can use the quickload feature!
But the narrative designers have decided that this scene is special. Thus it happens without any regard for the gameplay systems.
This is why it was so important to include a scene like Joel’s Last Stand, to show us why it would have been impossible for Joel to win.
They didn’t bother.
The key point is that if any of these plot beats had occurred in gameplay rather than via cutscene, the outcome would have been different. That means this entire prologue was effectively just a movie. The player is doing nothing more than occasionally pressing forward as Fake Ellie or Abby.
The entire game is written this way. The only major character who does not die entirely by cutscene is Nora.
In the first game, there are two characters who die by cutscene: Henry and Marlene. Henry kills himself, and so there would have been no way for this to be mechanized. Marlene’s death occurs at the end of a long gameplay sequence; it is kicked into a cutscene because we have already escaped the climax, and it is really an addendum to the story.
Sam, Tess, Robert, and most importantly David all die within gameplay sequences or directly following gameplay sequences, as their immediate fallout. Sam and Tess are both infected in gameplay sequences; Robert is caught in a gameplay chase; and David is defeated and mostly killed in a boss battle.
There is no gameplay sequence in Part II that matters. I would lose track trying to count how many times the player gets grabbed by some guy hiding just behind a door or how many times we lose control of Fake Ellie or Abby so that the story can diverge from the gameplay.
Consider the deaths of Mel and Owen (in a cutscene). Fake Ellie kills both of these characters. She does so with violence. Part II's gameplay is all about violence. Why is a non-interactive sequence being used to tell a part of this story that the mechanics of the game are explicitly designed for? Why was this a cutscene? We could have been forced to kill these two in a way exactly reminiscent of the death of the surgeon from the first game. Why weren’t we?
Most of The Last of Us' cutscenes are actually just people talking. The Last of Us does not have mechanics for people talking dramatically, so they may as well be a dramatically blocked cutscene. But in Part II, there are enormous fights and entire sequences of combat that are non-interactive. These sequences would be way more interesting if they were gameplay. Why aren't they?
One sequence in Part II follows Abby and Lev on a pointless sidequest to find medical supplies for a character who will die in a cutscene five minutes later anyway. Abby, who has a fear of heights, must scale across a crane atop a skyscraper. She keeps telling us about her fear of heights. “Oh my god,” she says, “I’m so afraid of heights!”
It is good that she keeps reminding us. If she hadn’t, we never would have known.
The Last of Us would have mechanized Abby’s acrophobia. The screen would have shaken. The controller would have vibrated. The colors in the game would have desaturated, and you would have had to overcome some new challenge designed to reflect what was happening in the narrative.
This happens when Joel is impaled at the University of Eastern Colorado. The entire game changes to accommodate his new injury. It becomes very difficult to aim. The controller vibrates. The mechanics shift to represent a shift in physical wellbeing.
In Part II, you press forward. You get to the other side of the crane while Abby complains about her fear of heights. The end.
The designers have utterly failed to mechanize her psychological process. They have not used their medium to place us in her body in any meaningful way.
Player vs. Player
The Last of Us: Part II’s most interesting idea has nothing to do with violence, hatred, or vengeance. What could have worked about this game, had it been designed by completely different people, is the notion of playing two opposing forces inside of the same story. This is ultimately also the narrative designers’ largest failing.
After finishing Abby’s pointlessly overlong sections, you are forced to confront Fake Ellie as Abby in order to progress the story. This moment is what completely ruins The Last of Us: Part II, and is where many players turn off the game: pitching two playable characters against each other robs the player of all agency.
Naughty Dog games are not about decisions. But agency is still important. It is the illusion that your actions matter. It is the feeling that you inhabit a real world that is responsive to your behavior. The Last of Us: Part II is seemingly concerned with agency, which is why every NPC now has a name. Agency creates the sense that the game is not a rollercoaster but a real, living world.
Death by cutscene robs the player of agency. Being grabbed from around a corner robs the player of agency. These are the crutches that hold up the smashed legs of a badly designed narrative.
When Abby fights Fake Ellie, Abby wins because we are playing as her.
That is the only reason why. Abby has to win, because we have an infinite number of tries. What is the reason that Abby can defeat Fake Ellie? How can Fake Ellie, a person who has overcome every obstacle ever placed before her, lose to this Jet-fiend bodybuilder with no gear?
Because we’re playing as Abby. That is the only reason.
You will die again and again trying to flank around Fake Ellie in this fight. Fake Ellie loses anyway, because Abby’s deaths don’t matter; only Fake Ellie’s defeat matters. This is despite the fact that you, as a player, prefer Fake Ellie. This is despite the fact that you have been desperate for Abby to go away for the last twenty hours. Fake Ellie loses, because you are playing Abby.
If you were playing Fake Ellie, the opposite would be true. But Naughty Dog has arbitrarily—for no external reason, but simply by fiat—decided that you must play Abby in this scene.
This calls overt attention to the contrived nature of the game’s systems. We are reminded that we are not experiencing any kind of reality. Interactivity was the only way to pitch two player characters together: let the player choose who to play, or have systems determine the outcome.
The outcome of this scene actively defies the game’s mechanics. Abby has lost all of her equipment, while Fake Ellie is fully stocked and loaded, and carrying about six spare guns in her backpack. If we were in control of Fake Ellie, she would never lose this fight against one unarmed NPC. You wouldn’t even need multiple tries. It would never happen.
Fake Ellie loses anyway, because that’s the story.
Would such a scene work in a movie? It would still strain plausibility. It would seem ridiculous that Abby could win. Some reason would have to be invented for the more competent character, with more equipment, to lose—a reason in the diegesis beyond “she can savescum.” But it could work.
It does not work in a game. It does not cohere with the systems. It might as well have been a gigantic cutscene.
Like with Joel’s death, I can still imagine a version of this sequence that saves the game. If you could select who you wanted to play, it could be a naturally branching path for the third act. The game could have two endings. Perhaps it happens naturally depending on if you win or lose as Abby. In either case, interactivity would determine the outcome.
Neil Druckmann is not interested in interactivity anymore.
Notice the violence
We used to discuss how games like Tomb Raider suffered from the dreaded “ludonarrative dissonance.” This meant that their stories made no sense when combined with their systems. Many games have this problem to a varying degree, especially story-focused games, because games require artificial challenge components. It will always be unrealistic for one person to kill thousands of enemies over the course of a single adventure. Reflecting that person’s mental state authentically would not be desirable.
Pointing this dissonance out is the sort of thing that makes for a decent blog post, but we can mostly just ignore it. You get used to it. Those of us raised on games can compartmentalize our slaughter to focus on what really matters instead.
But The Last of Us did not have a ludonarrative disconnect. Its post-apocalyptic world is diegetically violent and brutal. Joel’s character is that he is a butcher who will do more than anyone else to survive. It makes sense that he would kill so many people. He is a murderer. His psychology reflects this. The mechanics are aligned with the narrative. Almost no other game does this so well.
Part II appears to double-down on this notion. Every NPC has a name, and the gore is dialed up significantly. These innovations call increasing attention to the violence and make the apocalypse seem much realer.
But the ironic consequence is that calling attention to violence in this way ultimately reminds us of how fake and trivial that violence is.
You are playing on Grounded difficulty. You can sneak past ten WLF soldiers, or you can slaughter them. Deciding it is easier to show no mercy, you sneak up to them one at a time, slipping your switchblade into their throats. It may take a few tries, but with a bit of creativity, each is taken down.
As they die, they cry out each other’s names. They beg for mercy as they are blown to pieces. The gore is realistic and glorious.
Then you move on to the next encounter, and you never think about them again. Because they were not real people.
It does not matter how many names they have or if their arterial function is realistically simulated. None of the NPCs in this game have any agency. There is no illusion that they could be human. Real humans do not die so easily. These NPCs still exist only to be killed at the press of a button. They are distinguished from the mobs in another game in only the shallowest ways.
If these were real people, there is no way Fake Ellie could mow them all down. Combat would need to be evaded. She would need to be playing a completely different game.
But she is playing The Last of Us. She follows different rules. She exists on a different plane of existence than an NPC. Calling such thorough attention to the violence she is committing only serves to remind us of this fact.
The entire story is built upon this rotten foundation. Abby’s father, Jerry, was a nameless NPC that Joel killed at the end of the first game. He is the final one of countless hundreds of NPCs that Joel butchers throughout that game.
But it turns out that he was a real character. He had agency. He had a family. Look at what Joel took away!
Except it’s a lie. Joel didn’t take anything away. He killed an NPC.
This may seem trivial, but you must consider the importance of this distinction in a video game. Why is it only Jerry’s daughter coming after Joel? What about the children of those other Fireflies, or hunters, or smugglers at Robert’s? He should have thousands of embittered orphans seeking vengeance coming after him. Where have they all gone?
We know where. They do not exist, because those people were not people. This is an artifact of the medium of gaming. We know to ignore it.
We cannot ignore it when the game keeps reminding us. You’ll keep thinking, “If Jerry had a family, doesn’t this woman I just killed? Is her daughter coming after Fake Ellie for vengeance next? How couldn’t she? Why not?”
Then you will remember that you are playing a video game in which the violence is utterly trivial.
Games are not movies
Games have narrative designers, not writers. But The Last of Us: Part II is not credited with any narrative designers. Instead, it has two “written by” credits, and a “narrative lead:”
This is not a surprise. Halley Gross, the “narrative lead,” is a screenwriter, and she co-wrote her script as though it were a movie.
But games are not movies. Games are games, and their stories must be designed for their specific interactive medium.
No one designed The Last of Us: Part II’s narrative with interactivity in mind. But Gross and Druckmann did write a lot of cutscenes.
It is through its inability to utilize game mechanics and its total misuse of interactivity that I condemn The Last of Us: Part II. That the story itself is so awful is really incidental. I’ll get to that in the next part of this series.