Being Borges proposes a new form of literary translation, begging the question: What’s at stake when language becomes literal via the visual?

In this ongoing series, Ana María Caballero takes Jorge Luis Borges and Margarita Guerrero’s The Book of Imaginary Beings (a vast compendium of humanity’s imagined creatures) and its 1970 English translation by Norman Thomas di Giovanni as points of departure from which to explore how AI interprets Spanish versus English text, unmasking biases ingrained in large data sets. This collection also delves into the impossibility of translation–AI cannot “read” Spanish and English in the same way because they are different sign systems, with nuances and meanings that exist beyond their constructed signifiers, their words.

Caballero’s process was three-fold. She first used Borges and Guerrero’s Spanish descriptions of imaginary beings as prompts to create a large corpus of images from which two were selected. The process was repeated using di Giovanni’s English translations.

Thirdly, Caballero wrote a new poem, an inspired, compressed recasting of the original Spanish text, and used this poem to create an additional compendium of images. Her poems delve into the poetics of prompts, incorporating text-to-image generation semantics in their construction.