What "Fragments of Cultural Change" Means on karsdorp.io

"Fragments of Cultural Change" is the title and organizing theme of Folgert Karsdorp's personal academic site, karsdorp.io. It names his research program: using computational and quantitative methods to study how culture spreads, develops, and disappears, with special attention to the problem that the historical record survives only in fragments — partial, biased, and unevenly preserved. The phrase works on two levels at once: culture changes over time, and what we know about that change comes to us only in pieces.

Who runs the site

Folgert Karsdorp describes himself as a researcher and head of the Ethnology, Everyday Culture & Society (EECS) department at the Meertens Institute of the Royal Netherlands Academy of Arts and Sciences, and as Professor of Computational Methods for the Study of Cultural Change at Utrecht University. His stated research draws on computational and quantitative methods from ecology, population genetics, and machine learning.

That combination of fields is the key to reading the site's title. Ecology and population genetics supply models for how populations, species, and traits rise and vanish; machine learning supplies tools for working with large, messy cultural data. Applied to culture, these methods let him ask questions about diversity and loss that are hard to answer with close reading alone.

Why "fragments" is the operative word

The site's own description of his work names "methodological challenges in analyzing biased historical collections and measuring cultural diversity and loss." That is the fragment problem in plain terms:

• Survival is uneven. Some texts, languages, and traditions were copied, printed, or archived; others were not. What remains is not a random sample of what existed.
• Bias compounds over time. The further back you go, the more the record is filtered by who had the means and motive to preserve things.
• Absence is data too. Measuring what disappeared requires models that can reason about unseen material, not just count what survives.

So "fragments" is not decoration. It signals that the central methodological difficulty of this research is inferring cultural change from incomplete evidence.

What the site contains

The homepage is short and points to two main areas: a Publications page and a Selected publications list. The selected publications shown are:

Read together, these titles map directly onto the theme. "Correcting collection bias" addresses the fragment problem head-on. "Saved in Translation?" estimates diversity that is no longer directly visible, using medieval French and Dutch literature as a case. "Forgotten Books" applies unseen species models — a tool from ecology for estimating how many species exist but were not observed — to the survival of culture. A link to "All publications" leads to the fuller list.

What the site is and is not

karsdorp.io is a personal academic homepage. Based on the page evidence, it presents:

• A statement of the author's position and research interests
• A selected list of publications with venues, years, and DOIs where given
• Navigation to a fuller publications list

It is not a blog, a data repository, or a course site as far as the available material shows. If you arrive looking for datasets, code, or teaching materials, the homepage itself does not advertise them.

Who this is useful for

• Researchers in computational humanities or cultural evolution who want to see how ecological and genetic models are being adapted to cultural data.
• Students looking for a supervisor or research group in computational methods for cultural change, who need to know the author's affiliation and output.
• Anyone trying to understand the phrase — the short answer is that it is a research theme about inferring cultural spread, development, and loss from incomplete historical records, not a description of the website's format.

If you want the substance behind the title, the three selected papers are the place to start; "Forgotten Books" in Science (2022) is the most direct statement of the fragments problem, and the two 2026 papers extend it toward bias correction and latent diversity.