Category: Digital Poetry Network (page 2 of 2)

Automation of “Twenty Little Poetry Projects” by Jim Simmerman

Twenty Little Poetry Projects is a prompt created by Jim Simmerman published in The Practice of Poetry:

“Give each project at least one line. You should open the poem with the first project, and close it with the last, but otherwise use the projects in whatever order you like. Do all twenty. Let different ones be in different voices. Don’t take things too seriously.

  1. Begin the poem with a metaphor.
  2. Say something specific but utterly preposterous.
  3. Use at least one image for each of the five senses, either in succession or scattered randomly throughout the poem.
  4. Use one example of synaesthesia (mixing the senses).
  5. Use the proper name of a person and the proper name of a place.
  6. Contradict something you said earlier in the poem.
  7. Change direction or digress from the last thing you said.
  8. Use a word (slang?) you’ve never seen in a poem.
  9. Use a piece of false cause-and-effect logic.
  10. Use a piece of “talk” you’ve actually heard (preferably in dialect and/or which you don’t understand).
  11. Create a metaphor using the following construction: “The (adjective) (concrete noun) of (abstract noun)…”
  12. Use an image in such a way as to reverse its usual associative qualities.
  13. Make the persona or character in the poem do something he/she could not do in “real life.”
  14. Refer to yourself by nickname and in the third person.
  15. Write in the future tense, such that part of the poem seems to be a prediction.
  16. Modify a noun with an unlikely adjective.
  17. Make a declarative assertion that sounds convincing but finally makes no sense.
  18. Use a phrase from a language other than English.
  19. Make a nonhuman object say or do something human (personification).
  20. Close the poem with a vivid image that makes no statement but that “echoes” an image from earlier in the poem.”

It is possible to automate much of the process of writing a poem using this prompt. Some of the twenty “projects” could be completed using an algorithm and others using both an algorithm and a poet.

The order of the twenty projects (“open the poem with the first project, and close it with the last, but otherwise use the projects in whatever order you like”) can easily be generated by a computer.

 

Response to “How to Write Poems With a Computer” by John Morris

From How to Write Poems With a Computer by John Morris: “Perhaps the difficulty of writing poems with a computer may discourage all computer programmers from ever trying seriously to do it. If it does, then poems may remain, as they are now, one last refuge for human beings, for their personal communications (not to be confused with what the schools call “communication science”), for the strange, non-random phenomenon that we call love.”

Morris’ work developing algorithms that write haiku’s led to the conclusion that computer-generated poetry that considers syntax, semantics, and surprise will inevitably lack human genius. For now, and at least until artificial intelligence catches up to humanity, individuality is at the core of any answer to the question “what is a poem?”

Morris’ work confirms my belief that all computer-generated poetry will ultimately fail. But I believe that computer-assisted poetry is possible, especially if the algorithm is designed by the poet to produce a nontraditional form and if user input (choice; “human genius”) is required throughout the process. Both the poet and the computer would be making decisions and therefore the resulting poem would be a hybrid.

The process of “writing” a poem would actually become an objective, predictable process. As an example of what watching someone “write” a computer-assisted poem might look like, consider an artist that is paid to draw caricatures. They draw a different portrait for each person that pays them to do so, but they have simplified and made their process as predictable as possible in order to decrease the amount of time that it takes to complete the portrait. But a poet that uses a computer to assist in the generation of a poem has more of an advantage over other poets than a caricature artist has over other portrait artists. A computer-assisted poet can simplify the process of writing a poem without simplifying the poem itself. Caricature artists get paid to produces portraits in a short amount of time, but the portraits are of less detail than portraits produced by traditional portrait artists.

Hypothesizing “What is a poem?”

Cover of the book Anarchism is Not Enough

In the piece “What is a poem?” from Anarchism is Not Enough, Laura Riding writes: “A poem is nothing. By persistence the poem can be made something; but then it is something, not a poem. Why is it nothing? Because it cannot be looked at, heard, touched or read (what can be read is prose).” According to Riding, a poem becomes prose the moment that it is measured by any form of observation.

Perhaps an experiment can be conducted that will test Riding’s hypothesis; apparently, a poem is “nothing” for the same reason that, according to Quantum Theory, an electron will behave like a wave until it is observed and then it will behave like a particle. That was proven using the Two Slit Experiment. I am researching digital poetics and experimenting with algorithms to generate poetry to explore the possibilities of another hypothesis for “what is a poem?” that better parallels Quantum Theory. Discovering parallels to Quantum Theory may illuminate new aspects of the theory therefore progressing electronic communication systems and technologies (Tomorrow’s Technologies).

Justifying Free-Form Algorithmic Poetry

Instead of using commas to signify a pause, the spacing between one fragment of language and another naturally creates a pause when read. My assumption is that all punctuation and grammar is unnecessary for an algorithm that generates poetry because punctuation and grammar are unnecessary for all free-form poetry.

Newer posts

© 2024 Tref's Travels

Theme by Anders NorenUp ↑

Skip to toolbar