The Computational Stylistics Group exemplifies a quiet revolution in digital scholarship: the rise of informal, bottom-up, lightweight, ad hoc, and self-governed collaborative research networks. Loosely affiliated across a few academic institutions and sustained by scholarly curiosity rather than grant-oriented research agenda, the Group demonstrates how agile, low-resource collaboration can produce high-impact work.

Unlike experimental sciences, which often depend on expensive, centralized infrastructure that only top universities can afford and sustain, Digital Humanities thrives on open, community-built tools and shared standards. Such a dispersed research infrastructure includes not only hardware, but also datasets, conversion schemas, repositories, tools, and methodological frameworks, which all shape our scholarly lingua franca. TEI, Voyant Tools, NLP, Python, R, Topic Modeling, Network Analysis, LOD, IIIF – these and similar concepts have become the invisible scaffolding of our field. The R package ‘stylo’, developed by the Computational Stylistics Group, operates within this ecosystem not as a competing giant, but as a vital, accessible node – democratizing computational text analysis for scholars without advanced programming training.

Text analysis has a particular potential to serve as a starting point for newcomers to DH. The package ‘stylo’ offers a gentle introduction to this area of study, being an entry-level software to conduct text classification in no time. Taught at DHSI, ESU, and other international summer schools, it has helped embed computational text analysis into DH pedagogy.

This talk will explore how informal, bottom-up networks like the Computational Stylistics Group are redefining what research infrastructure can be in DH. Rather than awaiting top-down solutions, we can cultivate a more equitable, sustainable, and pedagogically integrated DH practice by promoting open tools, shared standards, and community-driven training. The future of DH infrastructure is not just in servers and grants – it’s in the collaborative, lightweight spaces where scholars make analysis possible for everyone.