Selected for a code-based poetical reading, “Sei autori in cerca di personaggi” is a poetic experiment on authorship and meaning
Coder and poet, I created and performed the piece
Two machine-generated poems, based respectively on tweets and Nobel price winners
Investigating the poetic language of machines, and the concepts of authorship and identity
Two machine-generated and human-curated poems from Twitter intelligentsia, and literature Nobel laureates
La Casa Encendida — a social and cultural centre that hosts the most groundbreaking artistic expressions in Madrid — organized a contest for a poetic reading where pieces had to be generated using computers to explore new and unconventional ways of writing.
“Sei autori in cerca di personaggi” (Six authors in search of characters) is a word pun on the famous play “Sei personaggi in cerca d’autore” (Six characters in search of an author) by Luigi Pirandello, and shares a similar absurdist approach.
The pieces are based upon respectively, the top 10 Twitter accounts and the last 10 Nobel laureates in Literature acceptance speeches. These datasets were then run through recurrent neural networks to create the output for the final poems, which I curated.
The results are both humorous and intriguing, especially if the audience starts looking for meaning. Future developments of the piece could include playing with different sizes of datasets to see if they could be more attuned to certain literary stiles (e.g. hermetic poetry vs novel), or looking at crowdsourcing as a way to affect the outputs.
Here are the two poems:
Of the mover
of the indarler
and the seriting
of the guts
that the is
the recind in the beet
of the the no gention of merent
that the ponent prowar in bear
and the lotter and reist in a and in
in the sald
in from the ires
to we she is
where the tomented
She know us to be for learned brack.
I'm love the fect you share a better
wirgon for the Chesters,
but about it!
ht@Cloletty2013 I deside to charting
Abala liw for new Sender to play the
He thinks here of #itsteenddels
you know this
You can find a step-by-step how to guide to replicate and play around with the experiment in this Github repo