Poetica Populi

Poetica Populi

Kyle Doubleday

ABSTRACT

Poetry is a powerful tool for expressing thoughts, ideas and language in a creative form. However, it is rarely used to give a voice to a collective group as opposed to a single author. If, instead, a computerized, generative method were applied to constructing poetry, it would be possible to construct poems based on large bodies of text that are the result of our online and offline communication. My goal is to create programs that will transcribe the language that surrounds us and generate engaging, meaningful poetry to shows us exactly what patterns and ideas dominate the language with which we interact.

INTRODUCTION

I often wonder what ideas move between our minds and through our communication channels. We so often hear of the “virality” of a thought or event, but what if there was a way to creatively express the language a collective group is using? By combining the processes for creating generative visual art, speech recognition software and sound recordings, this project aims to create generative poetry out of the language with which we come into contact in our daily lives. These poems then expose the patterns, biases, and prevalent thoughts that dominate both our geographic areas and the webpages that garner our views.

My motivation for this project was to combine my curiosity about how language interacts with how we think with my interest in creating artificial systems and creative expression in poetry. This is essentially an exploration to find out what words or thoughts are around me, and I want to know what it would look like to make poetry out of collective bits of language as opposed to one’s own composition.

This project will utilize sound recordings taken in particular locations to generate a body of text which a program will use to arrange and compose poetry based on guidelines I will establish.

PROJECT DESCRIPTION

This project will be organized into three stages, each building off of the one before it. The first and most basic is a program combining the Processing and Python languages that will take in bodies of text as input, either directly from a webpage or transcribed from a sound recording. This program will then algorithmically choose sets of words based on their frequency of occurrence and their ability to rhyme or the number of syllables of that word. These words will then be organized into poetic forms (namely haiku, four-stanza quatrains, and free verse) and displayed for the viewer to read.

The second stage takes this program and ties it to a location. A microphone will be used to sample the noise and speech being spoken in a given location. This location should be busy enough to have a large volume of words to be recorded; likely places include outside of a classroom or inside a restaurant. This audio signal will then be processed with Python’s Speech Recognition 3.4.6 API into text input, which will then be fed into the program already created in stage one. This will create poems specific to the time and place the recordings happen.

The third stage is an extension of this project in which the poems will also be displayed in a smartphone app. The app will use GPS to determine if the user is at a site where a poem was generated from stage 2 and then display the site-specific poem that user. This will most likely not be created during the semester-long work period in spring 2017, but this is a direction which the project could take in the long term.

HISTORY

Historically, this project has its primary roots in poetry and generative art. Written and oral poetry have long been means of self-expression for various cultures, from European aristocrats to Japanese government officials to Harlem artists. London poet Sumeet Grover describes poetic expression as “an art that weaves words together with imagination, emotion, passion, dreams, hopes and an uncontainable energy” (Grover, 2016). These elements of emotion, dreams and imagination are valued in human interaction, and serve as useful components of modern though and communication. This year, the Nobel Prize in literature was awarded to singer-songwriter Bob Dylan, primarily for the poetic lyrics in his music, so clearly this form of expression touches people in a unique way. Dylan also references other poets in those lyrics, including Arthur Rimbaud, Paul Verlain and Ezra Pound (Sisario, 2016). Thus, poetry has a tradition of reference, inspiration and borrowing from what came before, further connecting people to other times and experiences.

The other element of this project, generative art, is a relatively new occurrence but nevertheless has significant events on which one may build. One definition of generative art is “a practice where the artist uses a system, such as a set of natural language rules, which is set into motion with some degree of autonomy contributing to or resulting in a completed work of art” (Pearson, 2011). Generative art is hard to define, but this captures much of it: the artist uses a process, collaborates with the system, and helps create an unpredictable result within certain guidelines. A prominent early generative artwork could be Mozart’s musical dice game, in which he used dice to randomly select sections of music to play in a certain order. This game used chance and the artist was controlling the parameters, albeit in a relatively limited manner; after the advent and adoption of computers to handle large sets of data and repeated randomization, generative artworks began to flourish. Modern artists such as Brandon Morse and Ryoji Ikeda use computers to create artwork in this way: Morse uses complex systems of randomization within set bounds to create mesmerizing tornados or bubbles, which Ikeda takes the relatively random nature of human input and allows people to interact with the systems he enacts, such as a scanning radar that detects people on a beach. Generative art as a medium is useful particularly to this project because it uses computer programs to handle large amounts of input while also creating a work constructed by a system, without as much bias as a single poet trying to write about his or her relatively limited experience.

When it comes to combining these two disciplines and creating generative poetry, these are examples availableparticularly from the Electronic Literature Organization. Historicallythe generative method can be seen in the game of the Exquisite Corpse by Parisian surrealists in the 1920's, in which a paper is passes around and each player, without seeing the rest of the poem, writes a line. These constructions are pseudo-random, are the product of a system and, “uncover new perceptions grounded in the mechanisms of objective chance” (Adamowicz, 1998). They have the hallmarks of generative art applied to poetry, and they achieve an effect similar to what I want, which is a sample of the respective “inputs” put into an artistic form. More modern examples include those from the aforementioned Electronic Literature Organization, in which authors like Jim Andrews and online groups such as geniwate have created works likeStir Fry TextsandGenerative Poetry,respectively.Stir Fry Textscontains five assemblies of bodies of text, ranging from emails to activity logs. Any given assembly is then "stirred" by the reader with the mouse to create essentially random formations of that text. This is along the lines of the result I desire: a cohesive work that draws from text but assembles it randomly in a restricted set of forms. *InGenerative Poetry,the user’s clicks on a webpage select isolated phrases to be shown. This work uses randomized input to show the content of a poem; essentially, the poem is already written, it is just shown randomly. In both cases the method is much more computerized, and many modern cases of generative poetry build on this idea of computerizing the writing process.

SIGNIFICANCE AND DISTINCTION

The question of what language patterns a person is exposed to is an important one, especially in a digital age when increased exposure to and rate of communication shapes the words with which we interact. A direct example of this is the emergence of online “echo chambers,” in which like-minded people tend to interact with each other in isolated communities, resulting in more extreme opinions and a lack of challenges to those shared ideas (Emba, 2016). This issue particularly came to light after the recent 2016 U.S. elections, in which digital islands of voters grew increasingly distant from each other throughout the entire election cycle.This project will sample these discussions, construct a creative expression of their content and share it in order to break down these boundaries.

In terms of expressions of language, poetry has often been a useful tool to give form to scattered thoughts, but for the purpose of sampling a body of language a procedural, generative process is more appropriate. However, generative poetry is relatively new and tends to be more abstract, usually with randomly generated strings of words that allow the reader to search for meaning or appreciate the process of generative art. This project is distinct in that it aims to merge these two methodologies: use a generative process to sample a wide array of human inputs and create a structured linguistic expression that is representative of that sample.

EXPERTISE AND SKILLS

I have basic experience in programming in Java and C and none in Python, so these knowledge-bases will be improved that I mayably write a program to generate poetry.

I also have knowledge of poetry and its composition and feel confident organizing the sampled words into meaningful constructions.

Finally, I am comfortable using a recording device to record audio and manipulate audio files if need be.

APPROACH

My method for this project will be to develop a Python scriptto handle text and organize it into generated forms. Then I will utilize Python’s speech recognition API to analyze audio recordings and create usable input for the main program. Finally, I will compile a selection of these poems to display using Processingas the result of this project.

WORK PLAN AND TIMELINE

AUDIENCE

This project is intended for most audiences, since it examines communication and language. Since the recordings will be of sites around the UMD campus, students here will have an interest in the poems that are generated. This project will be based in English, so non-English speakers would not be able to interact with the work unless it is translated.

BUDGET

OUTCOMES

Once fully completed, this project should serve as a creative, easy-to-access way for people to get an idea of what words are being used around them. The parts that I plan on completing during the upcoming semester, the program and the ability to process recorded samples, will be demonstrated as a Python script ran by the user. A text file will be generated which can be displayed and animated in a Processing display window.This result will be the first half of a larger project involving a smartphone app to engage users en masse. This app will be both an extension of the work I plan to do as well as a way for my project to continue to engage people after I compile and share my work.

In order to consider my project complete, both the program that will compile the poemsand the processed voice recordings should be completed.Hopefully, others will take that program and utilize new inputs or apply it to specific situations, such as a unique way to show survey responses in research. This is the other goal of this project: give others a new tool to create useful and artistic expressions of the text that they encounter.

BIBLIOGRAPHY

Adamowicz, E. (1998).Surrealist collage in text and image: dissecting the exquisite corpse(Vol. 56). Cambridge University Press

Andrews, J. (2006, October). Stir Fry Texts. Retrieved February 10, 2017, from

Boden, M. A., & Edmonds, E. A. (2009). What is generative art?.Digital Creativity,20(1-2), 21- 46.

Colleoni, E., Rozza, A. and Arvidsson, A. (2014), Echo Chamber or Public Sphere? Predicting Political Orientation and Measuring Political Homophily in Twitter Using Big Data. J Commun, 64: 317–332. doi:10.1111/jcom.12084

Emba, C. (2016, July 14). Confirmed: Echo chambers exist on social media. So what do we do about them? Retrieved November 21, 2016, from

Funkhouser, C. (2013) Digital Poetry: A Look at Generative, Visual, and Interconnected Possibilities in its First Four Decades, in A Companion to Digital Literary Studies (eds R. Siemens and S. Schreibman), John Wiley & Sons, Ltd, Oxford, UK. doi:10.1002/9781405177504.ch17

Grover, S. (2016). Poetic thought, creativity and expression. Retrieved November 21, 2016, from

Pearson, M. (2011).Generative Art. Manning Publications Co.

Sisario, B., Alter, A., & Chan, S. (2016). Bob Dylan Wins Nobel Prize, Redefining Boundaries of Literature. Retrieved November 21, 2016, from

Voorhies, J. (2016). Surrealism. Retrieved November 21, 2016, from

Zhang, A. (2016). SpeechRecognition 3.1.0 : Python Package Index. Retrieved November 21, 2016, from

1