Home About EIS →

The Incoherence of Reincarnation: Story vs. Telling in Videogames

On page 141 of Noah Wardrip-Fruin’s (excellent) Expressive Processing, there’s discussion of a citation of from Jesper Juul:

Unlike most literary fictions, however, the worlds of many games are, in Juul’s terminology, “incoherent” (which is one of the things that limits Juul’s interest in discussing games in terms of narrative, as opposed to fiction). These are worlds in which significant events take place that cannot be explained without discussing the game rules, such as the many games that feature multiple and extra lives without any element of the game fiction that points towards reincarnation.

Reading this got me thinking about whether reincarnation in games is something that makes their worlds incoherent, and whether it is even a part of their worlds at all. Considering in this vein the common criticism that characters in computer RPGs repeat themselves, I saw the beginnings of a productive distinction, or rather, the application of an existing distinction: the distinction between a story and its telling (also known as that between fabula and sjuzhet). It makes more sense to think of reincarnation and dialogue repetition as extra-diegetic events, rather than as diegetic (and therefore incoherent) events, and a few examples illustrate this.

First, imagine a postmodern poem:

A: “Would you like to buy one?”
B: “No.”
A: “Would you like to buy one?”
B: “No.”
A: “Would you like to buy one?”
B: “No.”
A: “Would you like to buy one?”
B: “No.”
A: “Would you like to buy one?”
B: “Yes.”
A: “That’ll be $1.50, sir.”
B: “Oh, all right.”

This deals with concepts of repetition, refusal, and persistence, and is a bit abstract. The repeated question and answer is meaningful in light of the eventual acceptance and ultimate reluctance on the part of the consumer.

Now a conversation on a noisy bus:

A: “What kind of books do you like to read?”
B: “What?”
A: “I said, what kind of books do you like to read?”
B: “What?”
A: “What kind of books do you like to read?”
B: “WHAT?”

Some of the same repetition exists here, with a bit of variation thrown in, but contextually, it is no longer significant as a part of the story. If you were to ask what happened on the bus, the answer would be: someone asked what kind of books I like to read, and I answered them. You might mention something about how it was noisy and difficult to talk, but in most re-tellings (as opposed to reenactments), the repetition itself would be omitted.

Finally, an imagined fragment from a conversation in a game:

A: “What do you want to know about?”
B: “Tell me about the invaders.”
A: “The invaders came from the west and struck hard, …”
A: “What do you want to know about?”
B: “Tell me about the battle.”
A: “It was a grim affair. The defenders were …”
A: “What do you want to know about?”
B: “Tell me about the invaders.”
A: “The invaders came from the west and struck hard, …”
A: “What do you want to know about?”
B: “That’s all that I wanted to know.”

Here again we see elements of repetition. Some would say that this is a breakdown of the simulation: that the in-game character shouldn’t repeat the same phrase twice, just like a real person wouldn’t. But of course players don’t always pay as much attention as their characters would, so the repetition serves a purpose: to allow the player to re-explore conversational topics (which is especially relevant when looking for particular details that one didn’t know to look for earlier). Of course, this purpose could be served by other means, but that’s beside the point. Critically, when the player tells a story of the in-game events (say, to a friend), the repetition usually doesn’t feature in it. “I talked to the guard, who told me about the invaders and the battle, and then I …” When the question is “What happened?”, “The guard repeated eirself,” is not the answer.

We can explain the differences between these examples succinctly by divorcing the stories from their tellings. In the context of reading a poem (presumably in some publication that suggests the contemporary format of the work), we know that each word or phrase is meaningful, and that repetition is not something to be ignored. Even in a novel or movie, repetition has impact: often an exact turn of phrase is used as a powerful symbolic element of such works, and it can be a subtle clue that helps us understand the plot. In these contexts, the story itself hinges on the repetition of words, and so “what happened” cannot be divorced from it.

In contrast, the context of a spoken conversation on a noisy bus implies that the repetition is unimportant to the gist of the situation. The reasons for it are clear, and ultimately these reasons, not the repetition itself, are what is relevant to the story. Repetition becomes merely one of several ways to realize that story in a concrete telling.

Finally, in the context of text being produced by a game, the considerations of system design, player freedom, and experience design inform us that such repetition is actually the norm. The player is experiencing a story through a telling that comes with the peculiarities of an interactive medium, and so the story invoked by that telling is shaped by the medium. The repetition in a game setting is completely unimportant to the story, and in fact can effectively be considered extra-diegetic (i.e. it is outside the story). In the story, the player-character talks to the guard and learns about invaders and a battle. There is no obsessive repetition of exact phrases. In the telling-through-a-game, the repetition occurs as an incidental aspect: the dialogue is repeated not because it is repeated in the story, but because such repetition is the normal way to convey the underlying story in that medium.1

Once this distinction is made, we can see that all sorts of things that seem incoherent about story worlds are actually extra-diegetic. Loading saved games is one example: it’s not as if the in-game character actually died (or got into some other sticky situation) and then traveled through time and/or space back to a previous state. Instead, the player has merely explored an ultimately non-viable alternate story-line, and the “real” story is getting back on-track. The (anecdotal) fact that players’ recounting of their own exploits rarely incorporates loading saved games (at least, when they are telling a story of what happened in-game as opposed to when they are telling a story of their prowess as a player) backs this up.2

Of course, not all gameplay-related elements are extra-diegetic, and Juul’s point should be taken seriously. For example, although loading a saved game or dying and respawning are (usually) extra-diegetic, extra-life powerups that exist as items in the game world can’t be explained in this way, and in these cases, the incoherency of the game world affects the story world. But ultimately, we should keep in mind the fact that stories told through games have their own conventions for the specific composition of their tellings, and applying the conventions of some other medium, like literature or film, can make it unclear exactly which elements are actually part of the story.

  1. It has been pointed out that such extra-diegetic wrinkles of the telling are not always completely irrelevant to the story. For example, the extra-diegetic fact that the player died many times when completing a particular part of a level probably also colors that player’s perceptions of that part of the story (such frequent deaths might reinforce the fact that the hero’s ultimate diegetic actions (making it through that gauntlet) were extremely difficult). This influence doesn’t mean that the deaths themselves are diegetic events, however. For example, a very similar thing occurs in films when a slow-motion effect is used: the effect is an extra-diegetic artifact (but a crucial part of the telling of the story) which colors that telling and communicates extra information to the viewer, but it’s not as if time in the story world actually passed slower during the slowed-down moments of the film.
  2. The difference between re-telling the story of a game and telling a story about the play of a game is important here. When you’re explaining the plot of a game to a friend, you’re re-telling the story of the game, and it is in this setting that things like deaths and repeated dialogue are rarely referenced. When you’re telling a friend about your experience of playing a game, you’re instead telling a story of what happened to you while playing the game, a story in which the in-game story often features as a prominent part. But that in-game story is now hypodiegetic–an embedded story–and the re-told events of dying and respawning remain external to it.

About the author:  Peter is a 2nd year PhD student interested in most of what goes on in the lab. He's done some work with StarCraft and level generation, and is working on joint generation of levels with stories right now. Read more from this author

This entry was posted in Academics. Bookmark the permalink. Both comments and trackbacks are currently closed.


  1. Gravel
    Posted May 12, 2010 at 9:31 AM | Permalink

    Fascinating. And there are, of course, games that have persistent worlds, and a character *really dies*, with no save/reload option. Nethack’s a classic example of this, although light on actual story, but it’s clear that the death arcs are intrinsic to play, and you run into past players (and their stuff) on post-walkthroughs.
    More interesting is what’s being done with something like Dwarf Fortress, where the death/destruction/failure cycle is supposed to be part of the world building. (It’s not there yet, but it’s a goal for the game.)
    As we move away from arcades, I’m finding more and more games that don’t have classic “multiple lives” concepts, and require no more than continuing your save from the point where you exited the game. I consider it polite to allow the player to save/reload at will, but that’s about allowing the player to take risks and explore the story without being too worried about slogging through the starting chapters all over again.

  2. John Mawhorter
    Posted May 12, 2010 at 3:20 PM | Permalink

    I don’t see why any of the games rules can’t be ignored as extra-diegetic. While I do feel that incoherency is a problem, at the same time why are extra-life powerups so different from the structure of RPG conversations. It is possible to ignore all sorts of game rules and structures in considering the story… which goes hand in hand with my thinking that games “themselves” aren’t narrative (also your example of telling the story of a game is where games become narrative (in the past tense)).

    Posted May 12, 2010 at 4:21 PM | Permalink

    this is interesting, and its important to acknowledge the role of non-diagetic and extra-diagetic elements of games (especially since so many developers seem dedicated to attempting to destroy them – braid, for instance)

    ..BUT i’d really like to see a similar analysis from a more player-centric (as opposed to author-centric) perspective. for me at least, the most interesting games are not about a player exploring a story (or series of branching stories) told by an author, but about a player interacting with a system, possibly inhabited by other players as well.

  4. Posted May 12, 2010 at 5:06 PM | Permalink

    You’re right: games themselves certainly aren’t narratives, and they both tell stories and shape play, which becomes the subject of stories. I guess my point here is that the aspects of the game which are treated as diegetic or non-diegetic in terms of the story told within the game (assuming, of course, that the game is telling a story, which is not universally true) are determined via a somewhat implicit understanding between the game author and the players. Which elements are diegetic is not completely under the control of either the author or the players, and to some degree depends on things like the culture of gaming at the time when the game is played, and basic human psychology. What I’m asserting here is that this understanding includes the notion that repetition in RPG conversations (and some other things, like respawning in certain circumstances) is extra-diegetic, and that authors can/do/should take advantage of that. Of course, some authors alter the implicit understanding by making parts of these mechanisms diegetic (e.g. saving and loading in Prince of Persia: The Sands of Time). But what I’d like to avoid is the assumption that repetition in conversations is somehow unnatural and should be avoided: because this game behavior isn’t character behavior, finding ways to avoid repetitive conversations is an aesthetic choice, rather than a clear-cut case of making character behavior more consistent. In my view, to claim that eliminating repeated phrases in RPGs would make character behavior more self-consistent is simply wrong, because that’s not a change in character behavior, it’s a change in the presentation of character behavior, which is the same in either case. And whether to use a presentation model that is closer to real-world interaction is a decision that implies a different set of considerations.

  5. Posted May 12, 2010 at 5:15 PM | Permalink

    Here I think that I should make the limits of my ambitions clear: this analysis is intended to apply only to games that explicitly tell a story through the game (such as modern RPGs) rather than other sorts of games (such as games without stories, like Tetris, or games about inventing stories, like Once Upon a Time, or games that might relate to stories in other ways or not at all).

    That said, I agree with you that questions about diegetic and extra-diegetic actions/events in the context of interactive non-story-directed play would be interesting. You might want to phrase them in terms of being inside or outside of various game-worlds, though, rather than being diegetic or extra-diegetic, because those terms have very narrative-specific meaning. Certainly events from a shared play experience (for example, a game of Counter-Strike) could be analysed in terms of the narrative created by those events, and some game events (off-topic chat, for example) might be extra-diegetic (or not?) in relation to that narrative. But it seems to me that analyzing the shared play experience without treating it as a narrative of shared experience might produce a more interesting perspective, and in that case, the terms ‘diegetic’ and ‘extra-diegetic’ would not really apply, even though the very similar concepts of being within a fictional world or outside of it might. Of course, I don’t have enough knowledge of game analysis to even think of off-hand what the appropriate model for such analysis of a shared play experience might be, or whether analyzing elements of it in terms of their relation to a fictional world would even be the best approach.

  6. DF
    Posted May 12, 2010 at 8:47 PM | Permalink

    Great subject!

    I do have quite an “exterme” opinion of what a good “interactive narrative” should be – no pre-scripted lines or actions, only procedurally generated content. I much rather prefer a simple but coherent world (Ex: Pacman – he doesnt care how high are those walls – his world is 2d) rather than a “locked” complex world (Ex: Façade is awesome,but I cant throw the vase in the guys face – and I wanted to do that more than once :). So my opinion on this will probably be a bit extreme too.

    To me the dialogue repetition you mentioned (in the game example) is a problem,since it happens in the game space (as opposed to a Save Game or Pause, for example).

    I understand what you said, about the repetition not being “real” in the game – it’s only a feature to help the player -, but what about a third character (NPC or not) watching this conversation? Would he have to ignore the repetition too? What if there is someone walking by us the fifth time he repeats something, and hearing that information makes that person attack me – does that mean it was “real” then? It all begins to get very complicated, because it breaks consistency (in wich I believe good interative narratives should base themselves on).

    I hope some of that made sense (I’m not very good in expressing myself in english, sorry!). Anyways,great discussion!

  7. Posted May 12, 2010 at 10:48 PM | Permalink

    Excellent point. In a more complex simulation, as the repetition begins to really affect the world (or merely be richly embedded in it), it ceases to be purely an artifact of the story’s representation and becomes part of the in-game story itself. This is why, for example, I don’t think that extra lives in Mario are purely extra-diegetic: they leak into the story world because in many cases they exist as mushrooms embedded richly in the game world. Luckily (or perhaps unfortunately?) the repetition of dialogue in most RPGs doesn’t have this property of being an interwoven part of a rich simulation. This in turn means that it (the repetition itself) doesn’t affect the events of the story world, and so it can (and should) be considered extra-diegetic. Until something in the system is affected by the repetition, or someone observing the system chooses to interpret it as other than merely a quirky abstract representation for a single conversation (which I argue we normally do not), the repetition remains effectively extra-diegetic.

  8. Jesper Juul
    Posted May 14, 2010 at 12:51 PM | Permalink

    The idea of “incoherent worlds” is really meant to describe game events that cannot be explained exclusively with reference to the diegetic world. Lives and retries are generally a good example of that.

    As for the extra-diegetic, I do think that for something to be truly extra-diegetic (such as film music) it cannot influence the diegetic world, hence retries are generally not extra-diegetic since they do influence events in the game world.

    But if I understand you correctly, you are pointing to the possibility that retrying in Donkey Kong could be interpreted as simply rewinding a time line. I do not think there is anything in the game that signals this interpretation, but it is hard to rule out since a level in Donkey Kong completely resets (bonus items and all) when the player loses a life. I think this argument is strongest when talking about quick saves and such.

    Note that even if you go back to early arcade games such as Asteroids or Pac-Man, the levels are not reset when the player loses a life, so it retrying cannot be interpreted as time being rewound, hence lives in those games are clearly not extra-diegetic. Same thing with WoW, GTA, and so on.

    This seems to say that “incoherent worlds” in relation to retries can be understood as two different questions:

    1) Whether the player can explain all the events in the game with reference only to the diegetic world, retries and all. This is what I originally meant. (Donkey Kong: clearly no.)

    2) You are asking the slightly different question of whether retries can be understood as either completely diegetic or completely extra-diegetic. (BioShock: completely diegetic, Donkey Kong: can conceivably be understood as completely extra-diegetic, most other games: some blurriness.)

  9. Posted May 14, 2010 at 4:58 PM | Permalink


    You make good points, and I think in retrospect that my initial position wasn’t fully developed. Perhaps a good rephrasing is this: I think that despite the incoherence of the game worlds in these cases, the story worlds that are presented via the process/interface that is the game are fundamentally coherent. That, as a result of narrative conventions, developed in part through the play that was enabled by games like Asteroids and Pac-Man, things like death and respawning that happen in the game world ultimately do not happen in the story world (just like an intermission happens in the “stage world” but not in the story world).

    If respawning were going back in time and switching to a different timeline, it would still be diegetic, and the story would be one including some sort of many-worlds hypothesis. Instead, I see this game-world event as the equivalent of something like turning pages to reread part of a novel: it’s not something that can even be described relative to the universe of the story. In fact, choose-your-own-adventure novels make for a pretty good metaphor: in CYOAs, you often die, and then (sometimes as explicitly instructed by the book) you turn back to a previous page and reread, making a different decision. In most cases (with some notable exceptions) this is a completely extra-diegetic event. There’s no sense in which the story is one of multiple timelines (because, as you point out, such an idea usually isn’t implied in any way). Instead the story is the successful path through the book, and the choices that led to bad endings are merely literary digressions.[1]

    So ultimately, I’m making the (perhaps rather bold) claim that the game world is not the story world. And if this claim is accepted, then I assert that the fundamental incoherence of the game world need not affect the story being told, and in fact, through developed genre conventions, it does not in several common cases.

    [1] Of course, choose-your-own-adventure novels are really textual machines that don’t necessarily present a single story (just like games). So there are many stories being constructed, and perhaps even short deadly paths should be considered their own stories rather than the equivalent of literary digressions. But I think that that depends on how the player ultimately views them, not just on properties of the work itself. Accordingly, part of my argument here is that I at least (and, I assume, many other players) view respawning or repeated conversations as merely part of the game, and not part of the story.

  10. Jason T
    Posted May 16, 2010 at 12:13 AM | Permalink

    Interesting read. I’m not entirely clear, though, on why describing elements of a game as extra-diegetic should be mutually exclusive with describing them as incoherent. I should admit up front that I come from something of a biased perspective, having written an article over at Eludamos about this very subject, but more starting from the base position that extra-diegetic death often is incoherent, and reflecting on how some games that are more concerned withs storytelling have attempted to deal with that. I’ve come across even more games since then that offer alternatives, working death into the narrative or removing death as a failure condition (e.g., Heavy Rain, Fable II, the Prince of Persia reboot, etc.). Granted, this might just indicate that certain developers are concerned that death in games is incoherent, not that it really is. Still, I think there’s just as much anecdotal evidence that players are frequently frustrated and taken out of the narrative by the die-and-retry approach (see said article for some such examples) as there is anecdotal evidence that players often reconstruct memories having edited out the reloading of saves. How much of our willingness to edit our memories is a function of just being used to the convention, and is this willingness something that newcomers to games can be expected to exhibit?

  11. Jesper Juul
    Posted May 16, 2010 at 2:41 PM | Permalink

    “I think that despite the incoherence of the game worlds in these cases, the story worlds that are presented via the process/interface that is the game are fundamentally coherent.”

    Sure you can make that argument if you simply define the story world as “that which is coherent”, and the game world as “the rest”. But in what way can the story world be said to exist independently of the game world? What is the difference between the two?

    My proposed solution was the retelling test: If you ask a player to explain the game events, the game’s fiction is incoherent the player cannot provide an explanation without resorting to talking about game rules. To me that gives an indication of how players are thinking about a game.

    But then, I don’t think we disagree as such, but I am uncertain about what you mean when you claim that there is a story world that is independent of a game world.

  12. Posted May 19, 2010 at 11:44 AM | Permalink

    I like the idea of a retelling test as a solution, but I’m not sure that the focus should be on “game events”. I’d rather ask “What was the plot of the game?” than “What events happened in the game?” That’s because the story/game distinction that I think can usefully be drawn is analogous to the fabula/sjuzhet distinction: the story is the abstract sequence of events that the author renders into a game, which the audience then reconstructs by playing the game. Although I won’t define this as by definition coherent, I certainly think that we have a tendency towards reconstructing coherent story event sequences from incoherent game world events, and that authors usually use a coherent sequence of story events as the basis for the story aspects of their games. Of course, in standard fiction, the sjuzhet is “merely” a means of expressing the fabula (although some edge cases might include another purpose for it, for example expressing a hidden message). In games, on the other hand, the gameplay is usually at least as important a product of the game mechanisms as the story being told, if not moreso.

    I agree that asking about game events is a good way of eliciting how a player thinks about the game as such, but I think that asking about plot is a better way to get at the player’s reconstructed abstract story. And if this reconstructed story is coherent (and something the author might have expected), then the author can be said to have succeeded in communicating their story through the game. Of course, incoherencies of the game world may distract from the story: that the audience was able to reconstruct the abstract sequence of story events says very little about how smooth or efficient that process was, and a well-told story normally has quite efficient telling. But to the extent that game-world incoherencies do not distract from the telling of the story, I think that they’re unimportant. Of course, different people are distracted by different things, but my argument here is can be rephrased as: “Most genre-aware RPG players are not distracted from the story by repetition in NPC dialogue,” with the addendum that this incoherency also doesn’t make the story reconstructed by the players fundamentally inconsistent.

    On thinking further about this, I can see that my case is not as strong for player death and reincarnation, especially when death is frequent. To some degree the incoherency of this–but perhaps also the simple feeling of frustration created–distracts from the smooth telling of the story, so that even if this incoherence doesn’t make its way into the player’s reconstructed story-world, it affects the quality of the story as an experience. And of course, in some cases the incoherence inevitably does make it’s way into the story world, since things like green mushrooms are part of that world, but can’t be properly described without referencing game mechanics.

    So I’d argue that in Mario, neither the game events nor the abstract story events that a player might retell can be fully described without at least a passing reference to game mechanics. But in an RPG like Dragon Age, game events cannot, but story events can be described without reference to game mechanics, and this is an important distinction.

    Of course, another idea is that the relative importance of story and gameplay (and relative richness of story content) has a lot to do with how (and how much) we abstract from game events to derive story events. This feature of game design, then, can bias interpretation of individual game events as more or less story-related, which in turn affects the coherence of the reconstructed story.

  13. Posted May 19, 2010 at 4:24 PM | Permalink

    In the interests of fair play, I’d like to point out that I just edited the post to fix a typo and rephrase the last paragraph to be less dismissive of some of Jason’s (and possibly Jesper’s) concerns about incoherence of game worlds. Thanks for forcing me to think about this issue in greater depth and making me revise some of my less-solid positions.

  14. Posted May 21, 2010 at 3:31 PM | Permalink

3 Trackbacks