In a previous post, I discussed the field of interactive storytelling and where I believe that it should go from here. In that post, I touched on the concept of action-reaction vs. interaction in interactive mediums, but didn’t elaborate much on it, intending to reserve that content for another post. Now, two months later, I am following up on those statements with this post.
Before we dive too deep, it’s important to establish the difference between action-reaction and interaction. Perhaps the most simple example of action-reaction mechanics is what can be seen in practically any shooter. When you fire a weapon and it hits an AI or another player, it deals damage—in most modern shooters, the amount of damage is dependent upon where you hit the AI or player, introducing additional reactions. If it hits a wall or the ground instead, it leaves a bullet hole. In some games, if it hits any of a number of other objects, those objects break. You perform an action, shooting the gun, and the game presents a reaction based on where the bullet that you fire hits.
The most simple example of an interaction that I can think of is the same kind of interaction that we see in many RPGs. You press the action key to talk to someone and they respond with one of a small number of predetermined lines. If you keep pressing the action key, you will keep getting the same responses in a cycle or in random order. You press a button and a predetermined action occurs every single time in response. This is an interaction, an action-reaction pair that is inseparable.
These are only extremely simple examples of action-reaction and interaction, but they have hopefully informed you about how the two differ. It’s important to note that, due to their nature, most games intermix the two different paradigms of interactivity to varying degrees. However, I have found that narrative-only experiences (many of which are often colloquially referred to as “walking simulators”) often employ a high percentage of interactions when compared to other gaming genres. I’ll get to why this observation matters later.
I would also like to note that, just as it is important to distinguish between action-reaction and interaction, it’s important to not try to break down actions too far in an attempt to convey everything as action-reaction. In theory, yes, in the example above, pressing the action key to talk to a different NPC could produce a different result, but that doesn’t make that an example of action-reaction interactivity. In this situation, the action is interacting with the NPC and the reaction is a rotation of the same set of dialogue. The difference is that you can shoot a bullet and it can hit practically anything and the game will respond, whereas you have to directly interact with that NPC. In most interactive experiences, you can’t just make some gesture near them and have them respond accordingly.
It’s also important to note that a system may be designed in a way that is intended to promote action-reaction interactivity, but that doesn’t mean that works created using said system make use of action-reaction interactivity. A prime example is Storytron, which I discussed in length in the post linked above. Storytron has a lot of numbers that are hidden from the end user that drive the narratives that authors create. In theory, these numbers are meant to be used in a strictly action-reaction setting. However, there is one thing that makes it so that most storyworlds are instead a series of interactions: its choice-based nature. By creating explicit choices, you are taking the numbers and advanced systems underneath and distilling them into very distinct choices that will produce the same results every time. Most Storytron worlds are a series of distinct scenarios in which readers are given distinct choices. You lock in a choice and then move on to a new scenario. Due to being choice-based in nature, the system inherently encourages this design philosophy.
Important note: When I say “choice-based,” I mean choice-based narratives, which is a term that is primarily used in the interactive fiction community to refer to systems that rely on explicit choices, such as ChoiceScript, Undum, and, you guessed it, Storytron. I do not, as one comment on the previous post implied, “reject choice.” I simply believe that implicit choice is key.
So what, exactly, is it that makes most Storytron storyworlds interaction-oriented, but not all? In order to truly have an action-reaction system, actions must be repeatable. You must be able to continuously perform the same action and get different reactions depending on the exact situation. As mentioned above, you are able to continuously fire your gun and, depending on where the bullets hit, the game reacts differently. However, in most choice-based systems, you are only able to perform each action once and, each time that you do, the numbers behind the scenes change in the same way. To use an example that ties in, but isn’t particularly common, in a choice-based system, you may be offered the choice to fire a gun, but you would likely only do so once—and most definitely only once in that situation. Thus, choice-based systems primarily result in works in which players are forced to choose between a series of predefined interactions.
Side-note: To elaborate on the “most Storytron worlds” statement, I do want to note that the version of Balance of Power that was created to showcase Storytron does trend more towards action-reaction, as it places players in the same environment for making choices each “turn,” but most storyworlds that I’ve seen don’t do anything like that.
However, in order to ensure that the difference is clear, I want to explicitly state that repeatability is not the sole factor. You have to be able to repeat an action and get a different result. To tie it back into the discussion of interactive storytelling and experiences that challenge users’ social intelligence, you ask someone why they’re crying and they say “I don’t want to talk about it,” but then you don’t accept that answer and ask again, which results in an answer similar to an aggressive “just leave it alone, okay?” In fact, the ideal situation is one in which you can repeat the exact same action in relation to the exact same character or object and get a different result, depending on the number of times you’ve performed said action, but that isn’t always a feasible way of doing things. If you shoot a vase, it breaks and…that’s about it.
Now, here’s why this distinction is important. As we enter the end of yet another generation of games, one of the larger topics across the gaming world is that of player agency. Players want, at the very least, the illusion that they have free will within a game. The illusion that they have choice. Despite being bug-addled messes with subpar storytelling, player agency is what keeps bringing people back to Bethesda games and, if you just stopped and realized that I’m right about that, you have just realized the massive impact that player agency can have.
Here’s the thing about player agency, though: It is impossible without action-reaction systems. By definition, player agency requires players to be able to explore and tinker with systems. They have to be able to smack weird glowing rocks, have them explode and kill them, and learn from that mistake. They have to be able to kill NPCs and have guards kill them in return. A lot of what defines player agency is inherently bad for the player, sure, but it’s also what’s memorable and keeps players coming back.
Many criticize “walking simulators” for being boring despite having great storytelling and the reason is that simply having great storytelling doesn’t make them great experiences. To quote Chris Crawford, “choice is central to interactivity” and, when you set that aside just to tell the story you want, it leaves people wondering why they’re actively going through the motions in the first place.
As we enter the complex, algorithmic world of interactive storytelling, one which aims to challenge our social reasoning rather than our spatial reasoning, I believe that no factor will be more important than action-reaction systems that allow for player (reader?) agency. These new experiences need systems that allow players to interact with the AI in the ways that they want to interact with them. They need the choice to repeatedly ask the same question and just truly annoy the AI if they want. They need the choice to flip them off, even if they’ll get punched for it. They need to be able to do unreasonable things in an effort to learn how the systems work. That is how you encourage true interaction with the systems, rather than simply making a choice based on how you feel at that exact moment and moving on.