A Swadesh List is an index of basic words that seemingly exist across all languages. Named after American Linguist Morris Swadesh, they are compiled for the purposes of understanding linguistic similarities across cultures in an effort to create an essential human vocabulary. Basically, they are words throughout 'all' cultures that have resisted change over the course of much of human history.
There are many Swadesh lists and they have been revised by various linguists. The Long Now Foundation is keeping an archive of the Lists here.
I came across the Swadesh List today while reading John D'Agata's About a Mountain. He was writing about it in the context of Yucca Mountain and the 'warning: keep out!" signage that was being designed to stand the test of (deep) time.
Studying Heidegger in undergrad, I became deeply imprinted with the idea that technology is a force of evolution, more than a simple tool that we use at will. This idea has influenced much of the post humanist thought, Science Fiction, and contemporary Futurism that I follow . As we progress closer and closer to developing a unique singular Artificial Intelligence, how we understand ourselves and our species unravels byte by byte. Thinking about a 'homo sapien' and a 'homo technologicus' is something that is super fascinating, super terrifying, and very seductive.
To think through some of this stuff, I sketched out a Swadesh Haiku on the margin of the book. I've had a thing for Haikus for a while and I've written some amazing ones about the banalities of an office job that I had for four years. I'd give anything if I had saved those poems.
Writing out my haiku, I paired words that not only made sense, but were loaded emotionally. I tried to see if I intuitively followed any sort of structure:
adjective + noun, adjective + noun, / adjective + noun, adjective + noun as adjective/ adjective + noun, subject + verb
The resulting scribble in the margin of About a Mountain :
Haikus themselves are algorithms. According to Wikipedia, the basic requirements of haikus, are that they have an essence of 'cutting', a 17 syllable structure, and some sort of seasonal reference. The 'cutting' aspect is interesting. Machines and algorithms are really very good at parsing information, breaking up and adding together chunks to achieve a desired result. The 'cutting' aspect, I assume, is a Zen characteristic: a principle of asymmetry inherent in the world - dialectics - a great Cartesian riddle to solve...Dr. Frankenstein: or how i learned to stop worrying and love the New Aesthetic. Charles Stross dubbed this philosophical relic and existential quandry the Cartesian Theatre. Machines, now more than ever, are co-conspirators.
Anyway, Using Daniel Shiffman's rita library and this code, I took an initial stab at creating a computer generated swadesh haiku. Not surprisingly, its pretty ugly. The first one I generated was the most interesting, the rest complete nonsense.
The inherent problem with the Swadesh list is that the words in the list are either one or two syllables. The original code looked at words or word pairings up to 5 syllables. In order to make these more interesting, I'm going to have to dive into Markov Chains and be pretty heavy handed in the generation of the haikus. This list might be more suited to Truisms a la Jenny Holzer.