Toward a Computational History of Ideas

VICI-project (2017-2021)

Imagine a ‘Google Concepts’, a tool to trace how ideas such as tolerance, evolution, or science have changed through history.

This project establishes the proper methodological foundations to build such a tool.

My team and I will construct a ‘Google Concepts’-prototype methodologically appropriate to a humanities enterprise by relying on the novel ‘model approach’ to the history of ideas as its theoretical foundation. We draw on philosophy, intellectual history, computational linguistics, artificial intelligence and computer science.

New computer methods can provide large-scale evidence impossible to obtain otherwise. Our main tenet is however that ‘big data’ computer methods yield truly relevant results only if a certain amount of domain expertise is first top-down captured by humanities experts in well-designed conceptual and interpretive models, then used to guide bottom-up corpus exploration.

A team of philosophers and computational experts will carry out the project using a 1749-1940 corpus in English, German, and Polish. We will focus thematically on my subfield of expertise, the history of ideas relevant to axiomatic ideals of science; we will centre in particular on the history of logic/mathematics and biology in European thought. Computationally, we will combine an ontology approach with NLP-based machine learning and distributional semantics.

Our results will be generalisable to any field concerned with tracing the development of ideas in texts and thus create opportunities for fruitful synergies across related subfields.

By showing how we can best research centuries of scientific ideas in a cross-linguistic corpus, e-Ideas will represent a breakthrough in the study of the immense riches of Europe’s intellectual heritage. This outcome will establish late modern and contemporary history and philosophy of science, and intellectual history in the local and global landscape of the digital humanities, laying the presuppositions for a new field: computational history of ideas.