EMERGENT NARRATIVE: A BEGINNER’S LOOK

 
 

A Potential Magic Bullet

“Keep it together! Ellie”, a murmured voice came out from a middle-aged man. He looks towards a young teenage girl ducking under a shabby table. “I’m trying! Joel”, the girl answered with a trembling voice. On the other side of the table, a pack of armed survivors are slowly approaching their direction. But make no mistake, those survivors are there only to kill and scavenge the trapped couple. Taken from the game The Last of Us, the scene tries to build up that ever-lasting tension to survive in a post-apocalyptic world where the player, taking control of Joel and Ellie, not only has to fight the human-turned zombies but other survivors who live by hunting and scavenging other human beings. Or at least that is what the narrative layer of the game tries to convey: all of a sudden, Ellie vaulted over the table straight to the nearest enemy and blew his head off with a sawed-off shotgun. Before the next survivor could react, his head was met by a baseball bat. Ellie then proceeded to finish off the rest of them with Molotov cocktails. Joel stared at the seemingly vulnerable little girl standing in the middle of scattered body parts. Despite his mouth wide open, Joel failed to make any sound. Of course, this feat could not be done by the teenage girl alone: behind her is the player holding the controller. With the player intervention, a scene supposed to be narratively anxiety and fear-inducing was turned into a killing rampage.

This dissonance between the idea advocated by the narrative and the one carried out by the player action is not uncommon in the game world. In fact, “ludonarrative dissonance”, a term describing this phenomenon, was coined already back in 2007 by Hocking (2007) to criticize the same dissonance in Bioshock. Since then, the ludonarrative dissonance has become one of the long-running debates as recurring as the debates about “What is game” and “Is game art”. At first glance, the ludonarrative dissonance seems to result from the ill-conceived dynamics between the narrative and gameplay. The issue, however, points to a deeper divergence that if not more, is at least as ancient as ludonarrative dissonance itself: should the game as a medium be treated more as a storytelling device or a place where the player agency and game mechanics trump. There is also where the discussion about emergent narratives comes in, a concept that may be the magic bullet to marry narrative and gameplay to provide a coherent experience.

 

The Definition

Despite having the potential to resolve one of the fundamental conflicts in games, emergent narrative seems to be a quite fuzzy concept, having summoned as many definitions as “what is game” have done. In fact, even inside the category of emergent narrative games there exist different shades of how strong their emergence is. Discussions about the definition of emergent narrative can become quite theoretical such as Walsh (2011) did in investigating the concept within the frame of simulation. But it suffices to open up the discussion with the following intuitive definition by Jenkins (2004):

Emergent narratives are not prestructured or preprogrammed, taking shape through the game play, yet they are not as unstruc- tured, chaotic, and frustrating as life itself.

As the quote shows, it is emergent in that it is not “prestructured” or “pre-programmed”. It is also narrative because it is not as disorganized as life. Following this definition, we can categorize all narrative single-player games into two types: linear/branching games with pre-authored narratives such as the Uncharted series or Detroit: Become Human and games with emergent narratives such as Crusader King: II and Rimworld. One distinct quality of the second type is its spontaneous narrative possibilities coming from the player’s interaction with NPCs that ensues a substantial amount of player stories. The resulting immense replayability has granted several emergent narrative games great commercial success.

 

The Approaches

Now that we know what emergent narrative is, how do we achieve them? Chauvin et al. (2015) made a comprehensive summary of the current approaches to implementing emergent narrative separating them into two groups: the bottom-up approach and the opposite top-down approach. In the bottom-up approach, the narrative mainly comes from the interaction between the player and the autonomous agents whereas in the top-down it involves a drama manager system that imposes some level of authorial control on the narrative to ensure that it meets certain narrative beats or goals. Table 1 catalogs the characteristics of the two approaches.

Table 1. The characteristics of the two approaches

Another way to look at the two approaches is where the author resides: in the top-down approach, there exists an omniscient author towering over the virtual world where the player and other agents reside, always nudging the un- folding narrative towards a direction that meets general dramatic requirements. Thus, the narrative fostered would still have emergence but cannot go wild as to escape the invisible authorial hand. On the other hand, the player and the agents in the bottom-up method are free to act and react according to their agency. In a way, they are the authors of their own stories. This produces much more possibilities but as the authorial intentions are distributed they also suffer from the risk of chaotic and conflicting agencies and subsequently unsatisfying drama experience.

In reality, however, the implementation in the emergent narrative games is not always as clear-cut as the dichotomy. As the developers are also aware of the ups and downs of both approaches, they usually take elements from both sides while emphasizing one of them. We will discuss two relevant cases in the next section

 

Two Cases of Emergent Narrative Games

Before diving into the cases, it is helpful to make a distinction between the concepts of story and plot as examining the cases through these two dimensions proved to be effective for bringing out the peculiarities of the narratives emerging from the different approaches. Discussing the difference between the story and the plot in films, Bordwell et al. (2020) commented that “The story is the chain of events in chronological order (p. 75)” and the same story can be presented in different ways by employing techniques such as “flashbacks instead of linear time (p. 75)”, “organize events around one character rather than another (p. 75)”, or other choices about the presentation. Each different way of presenting the same story then becomes a different plot with a different impact on the audience. To put it more succinctly, the same story can be told with different plots resulting in different dramatic experiences.

The Uncharted series games are a good example to demonstrate the distinction. The story of the series tells the global fortune-seeking adventures of the treasure hunter Nathan Drake, with the journey punctuated by friendship, romance, as well as violent conflicts with rivals. One way to plot out the story is through the eyes of Nathan Drake, which would involve his excitement of discovering the forgotten treasures and the struggles of relationships and conflicts. But from the perspective of an insignificant mercenary Nathan Drake runs into, he is just trying to bring home the bacon for his family from a life-threatening job but is forced to go through the psychological trauma of escaping death by the skin of his teeth all the time as well as witnessing his mercenary friends being brutally executed by Nathan Drake. These two plots from different viewpoints will undoubtedly make a huge difference to the audience. The plot difference also highlighted that compared to story, plot is more tightly connected to the dramatic experience of the audience.

Crusader King II

Crusader King II is a grand strategy game developed by Paradox Interactive that epitomizes the story richness emergent narrative could generate. Players take control not of a single character but of a bloodline aspiring to become the ultimate ruler of Europe through diplomacy, war, and assassination with other powers, from which the narrative emerges. It took a hybrid approach towards generating the narratives with great emphasis placed on the bottom-up part generation. Particularly responsible for engendering the narratives from the bottom up, the opinion system in the game materializes an NPC’s attitude towards all other characters including players on a scale from -100 to 100. The attitude in turn would guide their behavior towards each other. Coupled with a multitude of available actions in the game, they form the basis for the enormous amount of possible events. At the same time, a scripted vignettes system was also put in the game that intersperses snippets of dramatic events into the narrative, spicing it up with some pre-authored content.

Figure 1. Crusader King II

What kind of narrative experience does it produce then? A good place for answers is looking at the stories player shares in the community. Below is one player story of common flavor happening in Crusader King II :

Started in Ireland my second or third game still learning the ropes. Had some early success, controlled all but two Irish counties and manage to declare myself King Harold of Ireland. It was not to last. The evil bitch Queen Maude of Scotland presses some random courtier’s claim on Ireland and I end up a duke.

Thus the young king Harold begins his life long revenge. I mar- ried my daughters and sons into the powerful families in England and Scotland. Meanwhile I began snapping up pieces Wales. At the same time I begin pruning the family tree of the evil queen of Scot- land. After a few years of murder and sowing dissent Harold raises his army and overthrows the idiot who took his throne. I don’t re- member his name. He remained locked in Harolds dungeon for the rest of his life miserable.

By this time England has unified and is fighting repeated wars with Scotland over northern England. Because of marriages to both royal families I have truces with both England and Scotland. This allows me to gobble up the rest of Wales that is independent. Harold manages to murder all of Queen Maude’s immediate family. So I pull up the dynasty tree and begin murdering everyone else. Cousins in Danmark: dead. Nieces in France: dead. Random family member married into the HRE: dead. Harold’s son in law in Scotland and his son: dead.

Scotland eventually starts to collapse and I realize ”Holy crap! Harold’s grandson will inherit all of England!” I just have to get rid of Harold’s son. So I make the craven jerk a commander but I can’t get him killed. Eventually I catch him in plot to assassinate Harold (told you he was a jerk) and try to imprison him. He gets away. I grant him titles and revoke till he rebels. Old Harold Jr thinks he is going to kick dad’s but. Except Harold Sr. dials up his favorite son in law, the King of England. Harold Jr. ends up in prison and bound for the headsman.

About this time the King of England dies. Harold’s grandson inherits and I realize all I have to do is die and I get to play as the Queen of Ireland and regent of England. Harold joins the crusades at head of his army, joins all his allies wars. Unfortunately Harold is badass who just wont die. Harold cheats on his third wife (and spymaster) repeatedly but she wont murder him.

Finally Harold Tyrant dies in his 80s. He outlived all of his legitimate children, and all of his enemies. He died a a hated man, a known murder, excommunicated by the Pope, and maimed from from years of war.

I had to play through 6 years of regency as his grandson. It sucked. (HungryGeneral567, 2017)

This story possesses many common traits of a bottom-up emergent narrative: an event-rich experience resulting from the various and complex interactions between agents while lacking a tight and detailed drama, requiring players to fill in the motives for the agents. It requires a decent amount of effort from players to have their own narrative goal and co-author the story. Indeed, in the same post where this story is shared, another story that wins the most upvotes goes like this:

I fucked a horse. (JamesBCrazy, 2017)

It is true that it is a quite dramatic experience but it is a kind that deviates too far. When the lack of player co-authoring meets the lack of external author intent, the possibility of unsatisfying stories is high. In short, because of the focus on the bottom-up approach, Crusader King II can produce an abundance of stories but fails to keep the plot drama on the same par.

Shadow of Mordor/Shadow of War

Shadow of Mordor and its sequel Shadow of War explored the top-down approach to emergent narrative. The narratives in the games unfold from the dynamics between the player-controlling character Talion and an army of orcs through the Nemesis System. Unlike the usual NPCs that will always give the same response to the same prompt, the orcs in the Shadow of Mordor series are given the power to remember by the Nemesis System: if you failed to finish them, they will remember what you did and come back for revenge. The same goes for you that when you are killed by them, they will humiliate you on a very personal level when you meet them next time. Besides the Nemesis System, there is also a Player Interaction Score (PIS) for every orc that goes up each time you interact with them. The higher an orc’s PIS is, the more likely you will meet that orc again in the future. In addition, orcs with higher PIS will have a higher chance to cheat death from you. All just to make sure that your destiny is going to cross again and again in the future. The Nemesis System and the PIS work in tandem from a top-down level to foster a very personal attachment between you and the orcs that even though the story possibilities are nowhere near the scale of Crusader King II, the resulting narratives is nevertheless full of memorable moments. The top-down narrative control has its downsides though: the encounter between the player and an orc is a positive feedback loop. As the player runs into the same orc, again and again, the orc’s PIS will go up quickly and they will cheat death over and over, leaving the player wondering why their archenemy just cannot die. The themes of the stories are almost always vengeance for players. The Shadow of Mordor series offers a narrative experience with sensational plot drama but little story variety.

 

What’s Next?

There have been numerous explorations of emergent narrative in games with each of their own nuanced understandings and implementations. To have a better grasp of the whole picture, we mapped out the notable emergent narrative games in a two-dimensional space marked by the story possibilities and the plot drama (See Fig 2). The vertical axis denotes the number of possible stories the game can generate. The horizontal one shows how dramatic the plot is perceived to be. On the top left, we have a group of games that excel at generating a large number of stories but are lacking in drama development. Unsurprisingly, these are the same games that utilize the top-down approach for generating narratives. On the bottom right are games with a decent amount of dramatic interest but heterogeneous stories. Their narrative generation also focuses on the bottom-up method.

Figure 2. The emergent narrative games by story possibilites and plot drama

By mapping out the games on the dimensions of story and plot, we notice that there is an uncharted area on the top right where the narrative would be both rich in story possibilities and retain the drama in the narrative. It is understandably vacant, however, considering the challenges to meet these two demands at the same time. As Ryan et al. (2015) discussed in the paper, a mechanism capable of creating a large number of stories need also to be able to represent these stories through appropriate channels such as text or animation, otherwise, these rich possibilities will only stay in the low level as different system states and are unable to be perceived by players. This representation very likely requires procedurally generated dialogues and animations since the underlying states are also not pre-constructed. How to procedurally generate them alone is challenging enough not to mention the pacing of the events, another equally important factor to the dramatic experience. Nevertheless, we believe that exploring that empty region would open the door to a new area of interactive narrative experience and a new understanding of narrative in games.

 

References

Bordwell, D., Thompson, K., & Smith, J. C. (2020). Film art : An introduction (Twelfth edition, International student edition.). McGraw-Hill Educa- tion. https://libraryproxy.his.se/login?url=https://search.ebscohost. com / login . aspx ? direct =true & db =cat09042a & AN =his . oai . his . se . 50522&site=eds-live

Chauvin, S., Levieux, G., Donnart, J., & Natkin, S. (2015). Making sense of emergent narratives: An architecture supporting player-triggered nar- rative processes. 2015 IEEE Conference on Computational Intelligence and Games (CIG), 91–98.

Hocking, C. (2007). https://clicknothing.typepad.com/click nothing/2007/10/ ludonarrative-d.html

HungryGeneral567. (2017). https://www.reddit.com/r/CrusaderKings/comments/ 6egkss/the best stories of crusader kings 2/

JamesBCrazy. (2017). https://www.reddit.com/r/CrusaderKings/comments/ 6egkss/the best stories of crusader kings 2/

Jenkins, H. (2004). Game design as narrative architecture. Computer, 44 (3), 118–130.

Ryan, J. O., Mateas, M., & Wardrip-Fruin, N. (2015). Open design challenges for interactive emergent narrative, 14–26.

Walsh, R. (2011). Emergent narrative in interactive media. Narrative, 19 (1), 72–85. http://www.jstor.org/stable/41289287

Previous
Previous

SPELKOLLEKTIVET: A COMMUNIST MANIFESTO