Expressive Input

Changeful Tales

“Changeful Tales” is a blog series where I rework my dissertation into more bite-sized, readable, and visible ideas.

I have defined storygames as artifacts that require an understanding of both a narrative and a system to make meaningful progress. It follows that a way to interrogate that narrative and those systems, such that both may be better understood, is vital for players. “Expressive input” is one approach towards enabling that interrogation.

My use of “expressive” here comes from Noah Wardrip-Fruin’s 2009 book Expressive Processing. Imagine considering a complex system (such as a digital game) from two perspectives: a surface which the player sees and interacts with, and the processes and data underneath that respond to those interactions and send new content to the surface. We can call a system’s processes expressive if authorial intent can be recognized in them. Such recognition might come from studying their output (interacting with the surface of the work) or through direct study of the processes themselves. An expressive process might be an intentional statement about the way the world works, as in Vi Hart and Nicky Case’s Parable of the Polygons, or it might be unintentionally revealing, as when a designer of a social simulation leaves out elements that others find essential. Regardless, we learn something about the designer through interacting with the processes and data they leave behind.

If we apply this term to the surface of an interactive work, we can likewise say the work offers expressive input if the user can express something through the affordances provided by its surface. “To express” in both case signifies something more individualistic and distinctive than voting or choosing off a menu. An expressive process demonstrates distinct authorial intent (not one of three or four possible intents), and expressive input likewise implies the space of possible interactions allows for unique expression. Its potentialities are not easily enumerable, and its possibility space large enough that a player might rightly feel a sense of discovery, surprise, or even ownership over their input.

To illustrate, imagine two character creation systems for a role-playing game. The first, Class Picker, offers players a choice between five character types, including Rogue and Wizard. The second, Skill Creator, lets players assign a pool of stat points among five possible skills, including Sneaking and Magic. I would posit that Class Picker does not offer expressive input, while Skill Creator does. We can say the following things about Creator’s expressive input that don’t apply to Picker:

  1. The range of outcomes implied by the possible inputs may not be immediately obvious to the player, allowing for “discovery” of options not at first considered: for instance, equally splitting points between Magic and Sneaking to make a mystical assassin.
  2. Some possible inputs might even surprise the system’s designer (such as splitting all points equally to make a generalist).
  3. While the set of possible inputs is technically finite and enumerable (each possible distribution of points), it would not be practical or useful to present those possible inputs to the player as a comprehensive list of options.

We might therefore say that systems like Photoshop, Adventure, or the Spore Creature Creator, about which players might proudly tell stories of their unique interactions, have expressive input. Systems like Pong, a Choose Your Own Adventure book, or your bank’s phone tree, all with easily enumerable inputs and little support for surprising or discovered actions, do not.

The character creation in “Mass Effect: Andromeda” offers expressive input, even if the game has few expressive processes responding to that input.

This is, of course, a subjective distinction. Two people might disagree about whether Myst has expressive input. But generally speaking, this term can be helpful to distinguish systems that tend towards offering the potential for discovery and surprise from those that do not.

It’s important to note that expressive input does not require a corresponding expressive process under the hood. The RPG attached to Skill Creator might not actually implement all of the listed skills; or perhaps no matter what skills the player picks, all outcomes are determined at random, because of a bug or even intentional design. This does not change the input’s perceived expressiveness.

Of course, the player is likely to be disappointed if their expressive input is ignored, one of several reasons why expressive input is not always a good idea nor inherently superior to alternatives. Overwhelming the player with perceived agency can lead to frustration, especially if there are actually only a few — or one! — correct courses of action. This is the classic pit trap of the adventure game.

Assuming you do want expressive input in your game, what kinds of interfaces enable it? Multiplicative input is an obvious example, either simultaneously (as with a verb and a noun, or a skill and a number) or cumulatively (as with a series of choices that build up into something unique). Single-input expressiveness is also possible, as in a color palette selector, or Scribblenauts in which thousands of items can be created by typing in different words, or an open world adventure game where the number of potential objects to interact with across a set of linked screens feels too large to easily enumerate.

“Scribblenauts Unlimited” and other games in the franchise offer both single-input expressivity (individual words) and multiplicative expressivity (adding adjectives or combining multiple objects together).

In all these cases, the interface provides room for possibilities: a space in which discoveries can be made. Most of the storygames we’ll consider as this series moves forward, as well as my work with Character Engine at Spirit AI, have expressive input: the changeful aesthetic implies a degree of control over the story world that’s hard to achieve otherwise, a way to enable genuine conversation between the expressive processes of a game’s designers and the expressive input of its players.

In the next post, I’ll talk about the problem of storygame genre: why designers must make baffling decisions such as whether to classify their narrative game as “Adventure” or “RPG” on Steam, and how we can build a more useful framework for talking about what makes storygames similar to and distinct from each other.