PB1 B: A Crack at Genre Generators

In full disclosure, I have history when it comes to genre generators. In seventh grade english class I used one as my accomplice to produce a haiku for my English class because I thought I had lacked a creative je ne sais quoi.

My teacher quickly spotted the skunk and much to my fortune allowed me to write my own haiku with no penalty. But how had my teacher been able to spot my loophole? Conventions. While it is fairly simple to identify what makes a certain thing a thing, it takes another level of thinking to go back and analyze, correct, or think twice about something put down on paper.

The first genre generator I looked at came way of the Massachusetts Institute of Technology. They created a generator that could pump out research papers dedicated to the field of computer science. The concept came off genial, however, theory and implementation are two separate junctures.

SCIgen's lack of basic english conventions

SCIgen asks the user to input only one field; the author of the piece to be written. The algorithm would then go to work, tailoring an empirical research paper on a certain aspect of computer science. I then tested the generator’s randomness by having it create three fictitious papers for different individuals. The titles for the three papers all dealt with human behavior concepts that could be solved with developments in computer science. Deeper in the text, the paper begins to unravel itself rhetorically, with the use of improper grammar and repetitiveness among the three randomly generated papers, this is just one of many examples how the generator occasionally falters when it comes to formatting the document.  

On the other hand, SCIgen does a laudable job organizing the papers in a scientifically professional way. All of the papers contained numerous conventions of an empirical research paper, such as an abstract, a conclusion, and various results and references. The papers were typically went about six pages in length and contained a strong collection of sources. Similarly, the papers included a fair bit of technical jargon, that could be understood by someone with a background in the field of computer science. 

From SCIgen, I then went to take a look at other, more whimsical, genre generators. Meme Generator does exactly what the title insinuates; it yields amusing memes that are generated from a certain set of conventions. Unfortunately, the generator was unable to function on any of my devices, however it is still able to make assumptions about the service provided.

Meme Generator's Most Popular page

By looking at the home screen, a section is displayed followed by a list of the website’s most popular memes; another way to describe this could be the most frequent input of certain conventions. These pictures with captions commonly share the same humorous overtones, they too also awaken an inner feeling of nostalgia.

The captions, like their corresponding images, also abided by certain conventions that many of its neighbors on the popular page shared.  Themes or genres that were ubiquitous throughout many of the pictures included topics such as religion, politics and human behavior. By steadfastly sticking to a set of conventions, memes are able to live up to their definition by being widely disseminated and enjoyed by the population. 

The website lures in users by touting it can "Create a Haiku in Seconds"

Finally, in hopes for some emotional closure, I came back to the website that got me in trouble back in middle school. A haiku is a very specific type of poetry, being that it follows a rigid pattern of seventeen syllables in three lines of five, seven and five. haiku’s also tend to deal with characteristics of nature, and this is reflected in the input that the generator requires the user to fill in. The conventions, in this case, take on a bit of a predetermined nature, meaning that even the user must abide by a certain set of conventions, not just the generator. The inputs that the algorithm require help to produce a more poetic or flowery haiku in order to make the poem flow better. This led to haiku’s being produced that followed a very similar tone, taking away from the uniqueness that helps to identify a haiku. 

However what the haiku generator, and the other generators lack, is something humans do naturally, revise. For instance, embedded in mounds of fictitious SCIgen research laid dozens of grammatical errors and incoherent passages. Revising our work requires another order of thinking, and does not only help us to identify mistakes in our writing, but by going back and reviewing what we write, we are able to dig deeper into metaphors and create more ardent thoughts. These websites know only what the algorithm that runs them tells them to know, they have no freedom to deviate from the conventions embedded within their structure.