Sometime in December, I received a very considerate e-mail from Mr. Jack Hoefnagel, who had read some of my posts and was wondering when the next might be, as (my words here) they’re always so darn long in coming.

We got to talking, and in a follow up e-mail at the end of December, Mr. Hoefnagel proposed the following considerations:



I'm curious about causality in games. If you're in a game, holding a gun in your hand and pointing it at someone's head, is there a way that a game designer can make every single player not shoot this person? The very nature of games is first that you have the ability to experiment without personal consequence, and also that the act of play itself means escaping from reality, so a player will be more tempted to try out things he or she normally isn't able to do in real life.

There's maybe two options to this, narratively (creating a sympathizing backstory) and ludologically (if you kill this character, you miss out on certain in-game benefits), but the fact remains that the consequences always remain at a certain distance from the player.

What are your thoughts on this? Are we able to create a game where every player feels the need to save this person's life, even the hardcore "ijustwannashootstuff"-type of player?

If not, does the current technology limit us in making truly engaging games? Or is it just because we haven't found a way to tap into a person's sympathy/survival instinct? Or is this medium inherently never able to accomplish this?



We carefully deliberated on this topic over the course of the next month, resulting in the following correspondence, which I reproduce verbatim (unless otherwise noted). As my initial response was so ridiculously long, this will be split into two parts, with my own first contribution uncouthly hogging the rest of this half.

Before I begin, however, I would like to acknowledge and thank Mr. Hoefnagel for thinking highly enough of me to contact me and initiate such a stimulating discussion (not to mention, for putting up with my burdensome style of writing!). His uncanny ability to ask just the right questions and elaborate on them propelled my own thinking and made possible many of the new conclusions I was able to reach throughout the course of our discussion. This investigation would not have been possible without Mr. Hoefnagel, who shares full ownership of it. My thanks, once again, goes out to him. And, of course, to Gamasutra for being a valuable conduit for such cooperative discussions (and an international one at that!).

So then, here is my initial response.



Firstly, let’s ask two questions: 1. What would keep a player from shooting, and 2. Why does it seem so natural to do so? It helps to keep these in mind in any effort to answer the problem.

And I would say it is indeed a problem in that, though such situations in games are supposed to provide for choice, often the choice seems foregone and immutable. So, in both ludic and narrative terms, I think there are two broad strategies which can be taken. 1. Explicitly tell or carrot/stick the player not to shoot or 2. Make the choice to shoot seem unnatural, or at least, less natural.

The phrasing of the above seems rather contradictory in that it espouses choice by denouncing the obvious one—it seems to be creating just another foregone conclusion, except for the flipside. But that is exactly our problem here: we need to overcompensate for the obvious in order to promote the less obvious.

That is to say, this being the nature of the argument, the player’s purpose is probably already at odds with what we are trying to accomplish. It becomes important, then, to duly avoid anything which too obviously tries to strong-arm the player, instead of causing him to come to his own conclusions. So, let’s go back to the two questions I proposed earlier and try to get to the root of the problem.

Explicit and Implicit Causality: Mechanics versus Meaning

There are, I believe, in both ludic and narrative cases, two classes of causal factors which may help guide player behavior: explicit indicators and implied directionality. These, of course, roughly correlate with the first and second broad strategies mentioned above, respectively.

For the ludic factors, we can split them into mechanics (explicit) and meaning (implicit), the latter of which I will spend a great deal of this e-mail discussing.

But first, mechanics. I would venture to say that mechanics based morality almost never works. At least, not in the sense we are trying to achieve. We can do things like making the player character more adept at survival, maneuvering and stealth the more of the game he progresses without the use of violence, but such mechanical inducement nearly always reduces the choice to a tactical decision.

For instance, light side points in KotOR are supposed to be rewarding recognition for player choices, but the actual effect, particularly on repeated playthroughs, is that they are more appreciated for their hurrying the player on towards light side mastery than they are for reaffirming the player’s moral stance. I would guess that this is part of why moral decisions are now completely removed from character build decisions in BioWare’s games (ME2, Dragon Age), beginning with a conscious effort at distancing from it in Jade Empire. (The other part of the reason being that min/max tendencies—the necessity of which is imposed by game difficulty—will overtake roleplay, causing the first decision to dictate every following one.)

Mechanical devices, then, may help the player decide how he will reach his goals, but they say nothing about what those goals should be. And rightly so. Huge generality: forcing a specific style of game to fit a specific style of play is arduous and didactic.

So, as far as ludic causality is concerned, the explicit strategy falls short to our purposes. For ludic direction on “appropriate behavior”—and I use this phrase in the sense that the seemingly natural response is natural because of an implicit logic—we need to look towards game language and meaning.

(I would like to apologize now beforehand if I happen to laboriously cover familiar material. Surely the concept of game language is not newly discussed here. I must also admit that, though I yet to read Mr. Ian Bogost’s seminal writings on the matter, I have the feeling from the wording of his phrase “procedural rhetoric” that much of what I am about to say has already been more thoroughly gone over before by Mr. Bogost. In all cases, apologies all around for the following.)

Meaning through Convention

For the first, it helps if we take a moment to consider film language. What is meant by this term is that movies have certain conventions which have gained specific meaning that are communicated solely (or almost solely) through the devices of the medium. For instance, if we have a close up of an individual’s face gazing intently at something, and an immediate cut to reverse shot, we know from convention that we are now sharing the view of the character.

This “logical” step or “natural” conclusion seems obvious to us since we are so used to the convention, but this was not necessarily true in the early days of cinema. Film continuity through editing was something which needed establishment first.

In a reverse shot, we are literally shown an image which has very little visual connection to the preceding shot. Unless we already equate the character’s gaze with the camera’s gaze in the reverse shot, the viewer must mentally reconstruct the environment and make the leap that the preceding shot is facing opposite to the reverse shot—something that is not always possible given close croppings and limited context. Convention now allows us to skip that reconstruction step—we already know the two gazes are the same.

An offshoot of this is the shaky, hand-heldish camera that implies immediacy or a first person perspective. This is quite important for our considerations here; the implication of embodiment is so strong that the establishing shot isn’t even necessary to evoke a driven gaze. These conventions are then abused by action films that try to camouflage the lack of intensity and momentum in poorly choreographed action sequences. (Hence, all the car chases/fight scenes with too close croppings and headache inducing shakiness.)

Similarly, there’s the oft parodied device of the montage sequence. This device is usually employed to indicate the passage of long or boring stretches of time, often working as a highlights reel that selects the choicest moments and bypasses all the tedium. The sense of progress, advancement, and the accompanying fun is so firmly attached to the convention that the montage is likewise abused in order to breathe excitement into an otherwise lifeless set of scenes. Nobody’s having fun, but how can you not have fun when there’s a montage?

So then, here we have examples where convention is so heavily attached to meaning that it colors the rest of the experience, to the point where it can be exploited. Now, let’s consider the example game scenario with the gun. How is this situation set up? Presumably, what we have is not only a gun pointed to a head, but also a pointer that is a crosshair as opposed to a cursor. Let’s take it a step further. What color is the crosshair when it mouses over the head? Is it red? Green? Or perhaps it turns back into a neutral white. Maybe even the pointer changes into a cursor from its default as a crosshair, or vice versa.

(I wrote the above before the tragic assassination attempt on Congresswoman Giffords in the American state of Arizona. It seems now insensitive to have written it, and to be writing the following as well, but I feel we can learn much from the debate about overly heated rhetoric and symbolic language affecting extremist behavior—we draw the boundaries on what is deemed appropriate and acceptable behavior based on our understanding of social values and cues. I would like to return to this subject after finishing the discussion on ludic causality).

Affordance and Game Language

As we can see, convention in games can likewise be exploited to influence player behavior, a similar topic to that explored by Mr. Michel McBride-Charpentier in his post Affordance Design in Half-Life 2. Briefly, affordance is about how the design of an object (shape, proportions with respect to human anatomy, apparent operable components, etc.) communicates what can be done with or to it. A chair’s seat is as tall as the bend of the knees, while the shape of the seat or padding suggests its usage as a rest for one’s bottom.

For the purposes of this discussion, then, I would like to make a rather tenuous distinction: if you call affordance in games the communication of what the player can do, game language could be called the communication of what the game might do. The latter is far more reliant on convention to establish meaning. Some of the more defined aspects of game language include level flow and pacing.

For instance, the Fatman in Fallout 3 exists to launch ridiculously powerful explosives—that’s affordance. It just looks (and sounds) like that’s what it’s supposed to do. But the placement of a Fatman at a specific spot—something that often indicates the presence of an obstacle that may be difficult enough to require its use—let’s call that game language.

Likewise, there’re no objects or environmental or even interface cues which clue the player in to when a tank might spawn during a campaign of L4D. But many players will be able to intuit the coming of that tank music relatively accurately from how fast/easily the team is progressing through the level. (Game language, then, is what makes the AI Director work. Indeed, the presence of game language in L4D is palpable enough to cause one to wonder whether the label of AI Director came after the fact, after its implementation, because it just seemed such a natural thing to call it that.)

So then, here we can begin to discuss “natural” action. Usually, both interactive affordance and game language conspire to encourage the player to shoot. Affordance: a loaded gun is pointed to a head. There’s basically only one action the player can take (click the button, shoot to kill), the other being to not take an action (something which is far less intuitive). In game language terms, a headshot means an instant kill, right? And that’s what we want, isn’t it? To get rid of obstacles as skillfully/kinetically as possible.

The headshot is further an illuminating example of mechanics crossing over into game language. It was originally a purely mechanical device—a ludic reward for more dexterous play, because it is harder to hit a smaller target. There’s no reason why a direct shot to the head should be any more deadly than a direct shot to the heart, or have any more stopping power than a full blast to the sternum, spine, diaphragm, etc. But this “logic” derived from the nature of hit boxes has now become set in stone.

Much as the need for an establishing shot was diminished to convey the meaning of an embodied gaze, we now take for granted that a headshot will automatically stop, or at least provide bonus damage to, one’s target. A headshot, more than any other type of shot, means instant kill. And just as it is difficult to extract the embodied gaze from a shot that is actually not embodied, it is similarly difficult to divest the headshot of its instantaneous meaning. Once again, it colors the rest of the experience.

“Ludic Literacy”

But there’s more. Let’s go back to the idea of inaction as action. When we pull the trigger in a game, we can be guaranteed that there will be a result—that the game will recognize that the player has taken an action and will respond in kind. On the other hand, there’s often no indication that lack of action will create equal recognition, particularly when so much of regular play consists solely of action taking, and so many scripting trigger events too. (How do you script for inaction without imposing odious timers? Once again, we have the situation of the realities of game programming determining causality.)

GTAIV, then, is a useful example to consider here. In GTAIV, the player is given several occasions where he can shoot a mission target or walk away—almost exactly the situation we are discussing. There are significant differences in that the player is given an interface prompt to decide between execution and mercy, but the basic premise is quite similar.

My own first experience of this example was with a friend after we had just brought the game back from purchasing it at the store. The choice first occurs at about an hour into the game. When presented with the scenario, the obvious and instantly arrived at conclusion for me was to walk away. Doing so would be an exercise of an external, personal sense of morality within the game—a rare bit of agency in a shoot or be shot world. Indeed, this instinct was so strong that I even failed a later mission by walking away and getting too far from my target (Vlad) where the execution was not an option.

I mention my own reaction, however, because the differences with my friend’s response are quite useful for our discussion here. Incidentally, the first time we encountered this situation, my friend was the one holding the controller. When given the prompt, I couldn’t help but comment immediately and out of hand, oh you should walk away. From habit, this just seemed the natural and self-apparent thing to do.

He was quite surprised by this—to him it just seemed entirely opposite of convention. And here’s the interesting thing: he was worried about closed opportunities and repercussions if the player were to walk away, whereas I was interested in exactly the opportunities dealing with those repercussions might present. The decision seemed more complicated and less comfortable for him than for me, the reasons for which I will now expound, being the point of this example.

Much like visual literacy, “ludic literacy” must also be accumulated through experience. Extensive play in the past with RPGs had created for me a greater confidence in moral choice selection—it was a decision making process I had exercised quite frequently before. But more pertinently, less diversity of experience on the part of my friend meant that he was less sure of the range of possible outcomes that could reasonably be expected by such choices in a game with the type of well-balanced design becoming of its high production values and polish. Indeed, he even turned to inquire, “Do you think we should?” (though, given his character, it is entirely possible he was merely being polite in his accommodation of shared play).

I believe that for him, the game language wasn’t strong enough (despite the explicit interface prompt) to convince him that going against his conventional experience (not to mention the affordance of the execution) would not be unduly punished (ed. note: just as it was not strong enough for me to indicate that walking away from Vlad would result in mission failure). The only choice seemed to be to shoot. For all he knew, maybe you would be permanently closed out from progress if you walked away, or have to redo the mission altogether.

Not pulling the trigger thus seemed like throwing away all the preceding effort of chasing the target down—especially when, throughout the whole mission, the player was under the impression (both explicitly narrated as well as afforded by the nature of a chase sequence) that the chase was solely in order to kill the guy.

It reminds me of this, more severe anecdote from Mr. Jim Rossignol:

Essentially, what we have, then, is something of the situation with film language in the early days of cinema (though we can’t dismiss the fault that lies with the designer in this). As the baseline sophistication and experience of the audience deepens, and as game designers (and their publishers) gain the confidence to allow for more sophisticated design, in some sense the situation will improve of itself. Players will become more experimental, and it will become more possible to have inaction as an action convention.

At any rate, I think herein lies the slightest hint to a possible basis for solving part our problem. We need to communicate to the player, not just through interface or mechanical devices like point penalties, but through affordance (like ambient player character animation with regards to how he holds the gun against the other, or unsexified weapons design which seems to imply an emotional toll on the player character at every usage) and game language (no loot for kills, equal experience for stays of execution, etc.), that not shooting, indeed saving a life, is just as momentous, decisive and meaningful—as agential—as shooting.

So far, I have consciously tried, for the most part, to avoid tackling the issues of emotional cues and narrational persuasion. As I mentioned, I would also like to do justice to the topic of socialization, and these I will definitely pursue with continued writing. You have very rightly brought up the question of, “how do we make play feel consequential?” and I believe much of the answer lies with these.


Part 2 of this discussion will be posted next week.