“Linguistic meaning is not some ideal and bodiless essence that we arbitrarily assign to a physical sound or word. Rather, meaning sprouts in the very depths of the sensory world, in the heat of meeting, encounter, participation...We regularly talk of howling winds, and of chattering brooks. Yet these are more than mere metaphors. Our own languages are continually nourished by these other voices—by the roar of waterfalls and the thrumming of crickets...” — David Abram, Spell of the Sensuous, 75–82

What makes sense; what makes nonsense? There is a hierarchy at play within the semiotic practices of even the most mundane meaning-making. This hierarchy privileges sequentiality over synchroneity, sustained over spontaneous, social over individual. Sequential, sustained, social language renders knowledge deemed illegible comprehensible. Even then, it is put to the test of whether the meaning made is valid.

Speech-to-text is trained to parse the thrumming of crickets and howling of wind from their onomatopoeic cognates—to distinguish words from noise. It is very good at this. So good that we can phone the bank in a tempest and check our credit card balances. Or command Siri to find the nearest dollar pizza from the subway platform. Furtive Ground asks artificial intelligence to listen against itself, applying technology that assumes impersonal objectivity (though encoded with epistemic biases and skewed data sets) to divinatory practices often accused of irrationality.