Typoly

What happens when we liberate language from the tyranny of correctness on the web?

Typoly operates at the intersection of linguistic control and network intervention. By infiltrating the browser – our primary interface with digital text – it performs a continuous act of disruption against the hegemony of "correct" spelling.

wikipedia-front-page Wikipedia's front page under Typoly's intervention

The piece manifests as a browser extension, appropriating the very tools of digital standardization to undermine their authority. Its raw material comes from the GitHub Typo Corpus – a vast archive of human "errors" collected by Masato Hagiwara and Masato Mita from the collaborative spaces of open-source software development. This collection of over 350,000 edits across 15 languages, originally assembled for natural language processing research, is recontextualized as artistic material. Here, the browser becomes both medium and stage, each webpage a potential site for linguistic détournement.

Fragment of the GitHub Typo Corpus, transformed from NLP research material into artistic medium

What researchers saw as data for improving algorithmic text correction, Typoly reclaims as evidence of human variance – each "error" a trace of authentic human presence in digital space.

🔄

Browser Performance

The browser and the Internet as a stage for continuous intervention

🌱

Digital Materiality

Human variance as data, data as resistance

New York Times, before and after Typoly's intervention

Through programmatic intervention, Typoly reveals the fragility of digital textual authority. Each transformed word exposes the tension between algorithmic standardization and human variance, between institutional control and networked resistance.

The work positions itself within the tradition of browser-based interventions, yet moves beyond mere disruption. It questions the very infrastructure of digital text production, where AI-driven tools increasingly mediate our relationship with written language.