The Future of Poetry in the Digital Age

hcarp

I was slightly taken aback at Rettberg’s suggestion that a new role for the human poet could be editing “poetry entirely generated by machine intelligence based on a process that remains both invisible and largely incomprehensible for the human poet” (53). Typically, I feel that we steer clear from or at the very least caution against giving artificial intelligence space with which it can overtake humankind (Hello Disney’s Smart House and Wall-E), but here Rettberg, by impartiality, seems open and even acceptive of the notion that artificial intelligence could further develop and become sophisticated enough to write poetry for us.

It should be emphasized that Rettberg is not, indeed, for or against the processes which he describes, as his book explores electronic literature rather than advocating for it (although the exposure garnered from writing about it could very well be considered advocacy), I began to think about the process of getting to the point of wanting to develop some sort of computer software that could write something we consider as intellectually engaging as poetry. More specifically, I want to know why. Are these people ruining the integrity of human-made writing? Computer-made poems aren’t thought to be the most profound, but Rupi Kaur’s Milk and Honey, a book of poetry mimicking the most famous six-second vine video scripts, isn’t either. Computers aren’t that smart when it comes to syntax, but neither are we nowadays; who is go say what has or hasn’t been written by a computer nowadays. Or are these people just interested in exploring the limitations of computers and thought that poetry would be the easiest thing to start with? After all, it isn’t that hard to find an assortment of words to go together and make a haiku and many of the computer programs that create poems hardly make sense. Or are they exploring a rebranding or redefining of poetry, one without preconceived constraints placed on it from practically the beginning of time? Up until this new digital era, poetry was permanent once it was printed on the page; now, it can be changed at the drop of a hat with a delete key or, in the case of Abra, a swish of a finger. It will certainly be interesting to see where poetry goes with this added technical nuance.

Combinatory Poetics: Chance, Randomness, and Surprise

maahari

Combinatory poetics is an intriguing genre due to its chance and randomness. In Electronic Literature, Scott Rettberg explains that “poetry and story generators and other combinatory writing systems are characterized by structured complexity and procedurality” (37). The theme of chance and randomness is on display in Jason Nelson’s this is how you will die and Nick Montfort’s Taroko Gorge. Nelson’s work functions like a slot machine in that each spin results in a different death. Each death scenario is composed of multiple poems and short videos that appear on the side. Each spin costs a certain number of credits and if you’re lucky, the machine will give you more credits. After several spins, I began to see the same poems in different death scenarios. Taroko Gorge is an endless poem with many possible combinations. Nature is a common theme throughout the poem, giving Montfort’s work a certain peacefulness. Rettberg explains that Taroko Gorge’s “code is open and accessible” (47). This allows anyone interested in combinatory poetics to create their own version of Montfort’s work, with many having done so already. I found this particularly interesting because, as Rettberg explains, it serves as an opportunity for people to become authors of electronic literature (48).

I think it’s interesting to compare combinatory poetics with standard poetry. While standard poetry is more cohesive and contains more meaning, the chance and randomness present in combinatory poetics does not allow for as much cohesiveness but does give the reader an element of surprise, something that is evident in both Nelson’s and Montfort’s works.

Generating New Appreciations for Electronic Literature

viboe

Of all the tracery narratives we looked at in class, I felt most drawn to the Missed Encounters generator. I think this narrative outline stuck with me in particular because of both its comedic potential and its meta-like qualities. The interactions and observations described in the randomly generated encounters seem to mimic the random-ness of the generated narrative system itself. One missed encounter may seem so important and singularly impactful, but on tomorrow’s walk a new story may present itself—the generator goes to work again.

In contrast, I did not find either “This is How You Will Die” or “Taroko Gorge” as engaging. In the case of “This is How You Will Die,” I think I became distracted with the seemingly dated interface that did not appeal aesthetically. “Taroko Gorge” also presented a challenge in that I found the scrolling text made it difficult to focus on any narrative structure or flow. In Electronic Literature, Rettberg points out that electronic literature now focuses on the “creation of more consciously aesthetic poetic works that explore different modes of procedural and aleatory structure” (38). In consideration of this manifesto, and my above case-study of sorts, I would determine that when in search of a Tracery narrative, I value an actual narrative of sorts–as displayed in Missed Encounters—and a visual display that does not distract from the text itself. 

I am also reminded of Peter Rabinowitz’s “Rules of Reading” that we discussed in class. The Rules of Signification and Coherence seem particularly relevant here. For example, with “Taroko Gorge,” the Rule of Signification prompts us to interpret what we notice, such as the moving text and the patterns in the generated words. The Rule of Coherence reminds us to be generous readers, assuming that the text is supposed to make sense and that deviations are intentional. Keeping these Rules of Reading in our back pockets, we can explore new imaginings of literature, that often at first appear only loosely related to literature we’re used to. However, we can also put these Rules to the side to allow for variables such as the creator’s intent. A conscious decision to either play by these rules or to suspend them empowers readers to bring their reading skills to the table, while also allowing for a new mode of reading and exploring a text. 

Combinatory Digital Poetics, CYOA Books and Their Limitations

asip

As I read through Scott Rettburg’s description of procedural and syntactic poetry in Electronic Literature, I found myself comparing a lot of similarities between this genre and CYOA books. 

Rettburg points out that with procedure and syntactic poetry, “they typically involve elements of chance and surprise, constraint and variability, and explore the poetic and narrative potentialities of text machines” (38). The chance and surprise that Rettburg notes here is evident in the two pieces we looked at today – Jason Nelson’s This is how you will die and Nick Montfrot’s Taroko Gorge. In Nelson’s piece, each spin produces a new narrative of various death scenarios. The randomness of the stories one reads based off of the slot machine and the surprise of various videos, and short poems that Nelson implements in his game showcases the elements of chance and surprise. Not only that in Montfort’s piece, there is an endlessly scrolling poem and when I refreshed it each time, it started out with a different line. There are many different possibilities and surprises embedded within Montfort’s work, especially because there are so many different combinations. The surprise and chance that is shown in Nelson’s and Montfrot’s piece reminds me of the CYOA maps and books we explored. Depending on which choice a reader made, the storyline could also alter. Rettburg also mentions that “running the program multiple times and analyzing output from the same generator allows us to consider and decipher the program’s structure” (42). This is exactly what we did when we mapped out our CYOA books – we saw which choices were “prioritized” and how the choices of the reader dictated the endings. Both of these genres show that language and storytelling does not have to be stagnant and restrictive, rather it can be used as a tool to generate different possibilities. 

As I was going through the two pieces assigned for class today, I was also thinking about the question that was posed in class: What is critical to understanding these pieces? I discussed with my classmates that the source material was vital, especially because it dictates the narrative that is told. The user, whether they know it or not, is limited by the choices of the author. This was seen in the game Travesty which is described by Rettburg as “a text generator in the form of procedural operation by the program on any given text” (38). The user is able to select the source material and Travesty generates text using the source vocabulary and substituting elements of the text (38). In the example that is given, this generator preserves the style of the original text and that not only speaks to the sourcing material, but also the programming itself. Because the text generator uses the original source to create its new piece and the way this piece is programmed, it makes sense that the new text is similar to the source. When we looked at 100 Billion Poems and created sonnets from the base poems, the “new” poems could differ vastly from the base poems, but we were still limited to those poems.

Apophenia: a Comfort or a Fear?

ellilly

Incorporating chance and randomness into poetry creates an entirely separate genre named “combinatory poetics.” This medium, as articulated in Electronic Literature, poses several challenges that are foreign to regular poetry. Cohesion between the various words as they shift is the key feature that authors must learn to navigate. Within the trillions of combination, each one must result in a coherent line. 

The lack of a rigid direction in which the reader must go allows a much greater reader involvement than print poetry. The randomness opens the work up to the apophenia humans so greatly covet. By filling in the interpretive space with her own ideas, the reader is able to find her own different meaning every time the work reloads. 

The two games assigned for today pressed against the limits of that covetousness. “This Is How You Die” was engaging for the first few spins. After only about four turns, I started to see repetitions.  Despite the game’s lack of intriguing aesthetics and much interesting content, the topic of death was enough to keep spinning until the end. The second game, “Taroko Gorge”, in contrast, had no concrete them as interesting as death to incentivize user engagement. Thus I found myself quickly losing focus. 

Where the textbook reading and the two games interact, I believe, appears in the capacity of the reader/player to engage in the game. Some topics, such as mortality, strike a nerve so central to the human identity that the reader is unable to resist reloading the game. While the nature aspect in “Taroko Gorge ” initially piques the reader’s curiosity, it fails to maintain it throughout the various editions. Neither of the two games have any plot, per se, but the subject matter of the first separates it from the second. Does user involvement require a certain level of substantive plot? Or, do topics alone separate the games we do engage with from those we do not. Does mankind find comfort in structure or ambiguity? What happens when a game enlists both?

Combinatory Tweets: “Slotted” Technique vs. Markov Chain

anonymous

I’ve enjoyed learning about combinatory writing more than previous topics in class because I have a better understanding of the code that produces them and because they are somewhat familiar to me. In a previous class I learned about artificial intelligence and natural language processing which are intertwined in combinatory poetics and have had major influences on them. As we dove into discussion in class this week and viewed Don’t Drink the Water by Dr. Sample, I was immediately reminded of a possible combinatory writing example that one of my friends showed me last year. The piece my friend showed me is called Automatic Donald Trump created by Filip Hráček. Famous for his twitter presence among other things, this site allows you to generate a realistically fake tweet by Donald Trump with just the click of a button. As I read further into the code and explanation, I found that this generator was created using a Markov chain. The author explained a Markov chain in simple terms. He comparing it to how an autosuggesting keyboard works on an iPhone when you’re texting. After you text one word, the keyboard supplies you with options based on common sentence structure and based on what you’ve texted before. This specific Donald Trump tweet generator used all of Donald Trump’s previous tweets as the information bank. Therefore every word you generate has already been tweeted by Donald Trump, just most likely not in the same combination.

I compared the Markov chain style to the random “slotting” technique that we are using for our tracery projects. Both of them have random aspects but the tracery project is more random since the next generated word does not depend on the previous one. It would be interesting to see if there was code that allowed us create a dependence for some slots on our projects.

Continuous Content in Combinatory Poems

owcorkery

The first thing I noticed and I think everyone noticed when first reading through these types of poems was how a new stanza, or line in the poem would appear every couple seconds and that each combination would seem random as if it was generated by some sort of algorithm within the computer. This allows for a seemingly endless amount of possible combinations for the poems. However, this did in turn create a lot of stanzas or lines that just seemed to not make much sense. However, I think that the stanzas or lines that seemed to not make much sense to me, might actually have a meaning to a different reader and vice versa.

The Love Letters would produce a new stanza every few seconds, while Scholastic Texts would produce a new line every few seconds. Personally, I sometimes felt as I did not have enough time to read each line or stanza in order to fully comprehend it. However, with A House of Dust, it would produce a new line every couple seconds and slowly build each stanza to give off the feeling that the reader is actually keeping up and reading along with the computer. I much preferred the experience given by A House of Dust.

It was interesting to me the amount of influence Dadaist poems had on combinatory poems, especially since the instructions on how to make a Dadaist poems with a newspaper is how the algorithm for combinatory poems works. Tzara even said within the instructions, “You will now become ‘an infinitely original writer with a charming sensitivity, although still misunderstood by the common people'”. This line is interesting to me because these were the thoughts I first had when reading through the poems. I noticed how combinatory poems are seemingly endless with original content and that they produce many poems that seem very confusing and dumb to me but might actually have meaning to other readers.

Understanding Combinatory Poems & Their Unique Form

brfoster

While looking through the poems/games, I was quite intrigued by the rapid and spontaneous generation of text. Initially, I assumed that these poems were very random. However, Scott Rettberg explains that the use of combinatory poems expresses an unconscious collectiveness and helps inspire other digital media works. Another point made about the use of combinatory poems was the idea that these works are trying to fool people into believing that the computer/game/poem is actually communicating with them and generating works that cater to them.

I saw many similarities between “Love Letters” and “House of Dust” because they produced new stanzas of poetry about every 2 seconds. Sometimes the poems would be very moving, other times it was hard to comprehend these works. I honestly enjoyed these works because they allowed me to break free from the traditional poetic style, which is full of hidden meanings and deep emotions. These works allowed me to appreciate poetry and electronic literature in a different light. These combinatory poems forced you to understand these random lines and dive deeper than the surface level. They make you really try to find deeper meanings within the text.

The work, “Scholastic Texts” stood out because instead of providing a whole new stanza, it only generated a new line about every 2 seconds, to generate a poem that goes on forever. It is constantly changing the theme of the poem while also sticking to a strict pattern. I still see more similarities between “Scholastic Texts” and “House of Dust” because both works do not feel personal and are following a systematic pattern. On the other hand, “Love Letter” felt as if I was reading different versions of letters to give to someone or that someone gave to me.

In relation to the Twine games we focused on last week, these works are only similar to the forms of using a dark or colorless background. There is a clear difference of the lack of making your own choices and choosing your own path and the form of interactivity. The Twine games seemed to be internal exploratory while these texts are completely external exploratory. Twine games seemed more personal and allowed you to form quick connections with the text but these combinatory poems force you to make connections and aren’t personal at all. Despite all of this, I can appreciate combinatory poems as being a unique genre of poetry that allows the reader to be fully engaged because of the quick generation of new text while simultaneously losing the attention of the reader with the random pairings and spontaneous text.

Creativity and Originality in Combinatory Poetics

Ethan Evans

Scott Rettburg discusses the history of combinatory poetics by analyzing some noteworthy works like “Love Letter Generator,” “Stochastic Texts,” and “A House of Dust.” These pieces of literature all produce their content through a programmatic procedure where lists of words are inserted into a standard sentence structure. High quality writing was never the goal of these early combinatory literature producers; instead, these works were meant to highlight how computing could produce literature with meaning and diverse variable outcomes.

Combinatory literature is not always as poetic or even as coherent as the programmer would hope. Lutz’s “Stochastic Texts” often produce sentences that make little sense at surface level, such as “every laborer is narrow.” This is because the combinatory algorithm is imperfect and does not account for linguistic trends and meanings. However, I think producers exploit this element of combinatory literature; ambiguous or ‘meaningless’ sentences result in a greater variance in reader interpretations of the text. Text that provokes thought in the reader is characteristic of poetry.

Some of the earliest origins of combinatory poetry come from the Dadaist movement and their methods of poetry creation. Words were cut out of a newspaper and picked at random to be used in a poem. I have personally used this technique several times, either for poetry subjects in English classes or for subjects in magazine collages. In both cases, picking words at random reveals unique outcomes that might not have been formed by human thought; I call it artificial creativity. However, artificial creativity is still creativity.

Falling close to the idea of creativity is originality. The randomized combinations of text allow for completely unique sentences, at least in relatively complex combinatory algorithms. In a way, this mirrors the creative processing system of a human writer, who is sifting through the text possibilities in their own mind. For this reason, it can be difficult to differentiate a work created by a computer and one created by human hands. I especially see this in “Love Letter Generator” where I could be convinced that any of the love letters were written by a human.

Combinatory Poetics Then and Now

sakline

The genesis of electronic literature began with the kind of combinatory poetics seen in “House of Dust” by Allison Knowles and James Tenney, “Stochastic Texts” by Theo Lutz, and “Love Letters” generator by Christopher Strachey. All of these works were released as early as 1952 but no later than 1967, and by looking at the extent to which understanding the historical context of combinatory poetics facilitates a greater appreciation for the works, we being to understand how that type of electronic literature might not be as applicable or effective as it was at the time of its creation.

What struck me most about the historical context was the way artistic movements such as the Dadaists heavily influenced creativity taking place. The cut-up technique adopted by many within this movement rejected notions of common sense and unoriginality by taking an article, cutting up all of the words, and then drawing randomly from the pot to create something new which became a common practice amongst creative writers (Rettberg 21). Thus, digital combinatory poetics’ connection to print forms of poetry lies within the same principle—feed the computer random words and have it arrange them in a random order to create a distinctly original poem. Rettberg highlights the way as well in which “computers provide a variety of ways to easily select at random” material that can be used for randomization within poetry (23). This aspect of randomness, however, as a reader located well within the 21st century, doesn’t as easily trigger my sympathies considering my externality from the original culture the text was placed in. Consequently, the randomness’ interjection in my ability to easily understand the text proves rather frustrating.

As a result, I am moved to ask: does the importance of combinatory poetic works lie within the text itself or rather within the process by which the text was created?

Simultaneously, the search for originality in the mid-20th century not only gave rise to the cut-up technique but also the distinct form of the works. The text becomes more indecipherable, I argue, due to the way in which electronic combinatory poetics escape formally from the confines of the screen to never return except in a reincarnated fashion. The Nick Montfort’s implementation of Strachey’s “Love Letters” flashes the letters across the screen without proper time to read them, and if you can’t read it in time, it vanishes. This movement provides a reason for the why “Stochastic Texts” and “Love Letters” remain fundamentally poetic. Their representational meanings in addition to evocative meanings, as we discussed in class, remain under the control of the works’ forms. With regards to these two combinatory poetic pieces, their representational meanings are hardly the only meanings present, for the evocative meanings are heavily privileged when the content is difficult to decipher, disappears, and is recycled.

I remain intrigued by Rettberg’s claim that “if these letters are not great literature, they are literary writing produced by the procedural operations of the computer” and if this is the way we choose to think about combinatory poetics, then to what extent have modern computer science advancements also advanced (or diminished?) the ability for writers and scientists alike to be better storytellers (32). Particularly, what is the purpose of poetry if not to somehow come to understand it? Am I unable to understand the gravity of these works because I could not relate to the contextual culture?