Graphics, Imaging, and Games Lab

A few weeks ago, Gail Carmichael linked me to a collection of articles dedicated to narrative in the digital space, including non-linear stories, and how creativity has been changed through new technologies. I plan on posting my impressions from these articles here, summarizing how they relate to the Creativity Enhancement Project.

This first article is one that reviews the state of story generation software, and their underlying engines.

Original Article

Early in the feature, we are presented with a very important point regarding the importance of narrative itself. While most studies and software focus on applications for entertainment and crafting fiction, there is a wider perspective that is often missed.

Narrative is more than a story. It is universally known and used cognitive model, whose structure allows us to understand diverse and convoluted concepts. While this fact isn't typically acknowledged, it still means that this sort of technology could have wide reaching applications.
 

The narrative structure is one that is crucial to human understanding and communication. It is the model usually used when describing and analyzing concepts in any and all topics. There may be countless other methods to structuring these ideas, narratives are usually how we choose to represent them.

Quickly, the article identifies it's guiding taxonomy; one from Boden (Source). It is one that categorizes two kinds of creativity; Exploratory Creativity and Transformational Creativity.

Exploratory Creativity is defined the process of searching a conceptual space, one that is already governed by specific rules and boundaries. Whereas, Transformational Creativity involves manipulating these rules and extending the boundaries of a conceptual space, in order to find new territory to explore.

This distinction is a very interesting one to make, and it is one that neatly identifies the focus of the Wordplay prototype.

For most existing story-planning/creativity enhancement software (including all the examples in this article), they exclusively work within a set, user-defined world state. This manner of assistance falls squarely into Boden's exploratory creativity. This is definitely a valuable area to provide assistance in, but the notion of transformational creativity is entirely ignored within the design of these programs.

This focus on exploratory rather than transformational creativity is a counter-intuitive strategy when it comes to crafting a narrative.

For one, it assumes that the process of transformational creativity is complete. It works only in the context of a concrete world state; one whose rules and boundaries are set, and do not require revision. To me, this seems like a major fallacy when looking at the nebulous nature of creating fiction.

There is always room for transformation when it comes to creating an interesting story. In fact, it is rare to see creative work follow the original conceits or goals. There are numerous examples of books, films and games departing significantly from the first envisioning.

In fact, forcing a set world state from the start doesn't seem to match well with a goal of "creativity enhancement".

Additionally, the story-engine itself is focused on "pathing" between specified world states in an "acceptable" manner. By "pathing" I am referring to determining plot and character actions which drive the world to eventually meeting a desired/inputted outcome. This method is crafting stories backwards in a sense.

The proverb "Focus on the journey, not the destination" applies very aptly to creating any fictional story. A creator should be interested in the arc of the plot and characters rather than their outcome. Ultimately, that is where the drama and moments that narratives are built on come from. Stories are in and of themselves journeys, and that is where true creative effort should be applied.

Also, the definition of what is considered "acceptable" by these programs is troublesome, from a creative perspective. While I understand the inherent limitations of working with a digital consultant, the parameters that the example story engines use aren't well suited for enhancing creativity. At multiple points in the article, they specified that the engine was committed to fulfilling the following principles:
- Applying social norms
- Applying medium principles
- Meeting audience expectations

These are fine metrics to check against in order to create a comfortable and cohesive story. However, if your goal is make a creative and original story, these parameters won't serve you well. The insistence on creating stories within specified and established tropes/expectations is entirely counter-productive. If anything, these parameters stifles creativity, rather than enhance it.

Creative and original stories, are ones that can defy and question existing expectations or norms. In many cases, the predominate genres of games (fantasy and science fiction) are ones founded on providing a thought-provoking perspective on a contemporary issue. If your main goal is conforming to set expectations and norms, it would be difficult to come up with a profound theme within your narrative.

Another major issue presented by present story-engines is that the construction and specifications required for a world state are prohibitive for most users.

Regardless of the underlying engine and algorithms, the level of minutia required to populate the world state and required parameters is a trying task to approach. This obstacle, while a necessary evil for present programs, creates a sizeable obstacle for writers.

This sort of challenge is one that hinders a creators output at one of the most crucial points in developing a story. Rather than focus on transformational creativity, a user is forced to put their attention on meeting the criteria of the software instead of cultivating their preliminary ideas.

It also doesn't allow a writer to organize and build their stories by their terms. There are countless different methods that writers use to organize their creative processes. It is important that the software be as accommodating as possible. Having an ornery requirement for meeting predefined structures and expectations is going to be restrictive to many creators.

Once again, this system is one that ultimately limits the conceptual space a creator is working within, rather than expanding it. Forcing structure on early narratives doesn't really serve as good direction for creativity. What it does is immediately stifle processes of transformational creativity and force a preliminary story into a comfortable and familiar mold.

Judging from the examples presented in the article, it seems clear to me that present story engines have made impressive achievements as well as leaving some startling oversights. The algorithms and programs shown have truly amazing capabilities. They can understand and apply some core story structures, as well as give meaningful suggestions in certain contexts.

The oversights and problems of these programs seem to be ones centered around a misunderstanding of the creative writing process and philosophy. While their toolset certainly could be helpful at specific development points, for the most part the barriers to entry and unyielding focus on structure make these story-engines useless as a tool for enhancing creativity.

Ultimately, I feel that these failings show that there is a place for a program like Wordplay. One that focuses on transformational creativity, allowing for unstructured and unfettered brainstorming.