Seedlings_:From Humus
by Qianxun Chen & Mariana Roa Oliva

Seedlings_:From Humus is a digital ecosystem within which human and machine writers can collaborate in caring for linguistic seedlings. Initial word-seeds are randomly chosen from six short texts written in response to the conceptual core of this project—human and non-human collaboration. They provide the diverse and rich mix of semantic soils where the seedlings will grow according to algorithms that find semantically and phonologically related words. Roots, which propagate from the word-seeds, also interact with the text-soils. Once a root reaches a soil-word, it uses it as the new context of the word-finding algorithm—adjacent linguistic biosystems thus enter into communication with one another.

Natural Language Processing algorithms learn from language that has been in constant use and change for thousands of years. In contrast to the one-way flows of other generative literary artworks where algorithms take source text as input and use it to generate new content, this piece proposes a cycle of collaboration as the creative process. Inspired by both initial and generated texts, participants act as semantic fertilizers, dislocating and recontextualizing concepts, and transforming the browser into a landscaped garden for the creation and symbiotic development of linguistic migrants.

[Press spacebar] to export the current view as an image.
[Double click] a word to plant a random type of plant.
[Right click] a word to plant a chosen type of plant.
[Right click] on a plant to remove it.
[Drag] a "soil" word to reposition it to different positions of the canvas.

Special Thanks to Sawako Nakayasu, John Cayley and Daniel C. Howe.