Digital history uses computational methods—textual analysis, data visualization, mapping, network analysis—to ask new questions about historical evidence at scale. Digital historians create databases of historical documents, analyze patterns in large textual corpora, and visualize networks of relationships across time. These methods complement traditional close reading while raising new questions about reproducibility, interpretation, and what large-scale patterns reveal about historical meaning.
From your introduction to historiography, you know that every generation of historians develops new methods for engaging with the past — from the documentary positivism of the 19th century, to the social science turn of the Annales school, to the cultural and linguistic turns of the late 20th century. Digital history is the most recent of these methodological revolutions, and it is distinctive in an important way: it changes not just what questions historians ask but the *scale* at which they can ask them. Where a traditional historian might closely read hundreds of documents in a career, digital methods allow analysis of millions of documents in a research project. This quantitative shift creates qualitative changes in what kinds of historical questions become answerable.
The term that captures the core tension is distant reading, coined by Franco Moretti in contrast to the "close reading" that has been the core practice of humanistic scholarship. Close reading means sustained, attentive engagement with a single text — reading every word, noticing ambiguity, attending to rhetoric. Distant reading means processing thousands of texts computationally and analyzing the patterns that emerge: word frequency over time, genre distributions, the rise and fall of particular concepts. The literary critic reading one novel closely and the digital humanist analyzing 50,000 novels statistically are doing fundamentally different things. Neither is simply better — they answer different questions. Close reading reveals the specific texture of a particular historical moment; distant reading reveals trends invisible to any individual reader because no individual can read at that scale.
The major digital methods each open specific kinds of historical questions. Topic modeling identifies clusters of co-occurring words in a large corpus and surfaces recurring themes — a historian can use it to trace how a concept like "liberty" or "contagion" changes across decades of newspaper coverage. Named entity recognition extracts people, places, and organizations from unstructured text and enables analysis of networks of mention. Network analysis maps relationships — correspondence networks, trade connections, kinship systems — and identifies patterns of centrality, clustering, and brokerage that are invisible in individual documents. GIS and spatial analysis allow historians to map where things happened and how spatial patterns relate to other variables. Each method involves tradeoffs: topic modeling requires decisions about the number of topics and the cleaning of text; network analysis requires a theory of what counts as a relationship.
The deeper historiographical question is what computational patterns can tell us about historical meaning. A topic model can show that certain themes cluster together in 18th-century political pamphlets, but it cannot tell you whether the authors chose those combinations strategically, whether readers interpreted them as the algorithm groups them, or whether the pattern is an artifact of the corpus selection. This is the irreducible hermeneutic gap between pattern and meaning. Digital historians argue — correctly — that these methods generate hypotheses and reveal patterns that demand explanation; critics argue — also correctly — that the risk is mistaking computational artifacts for historical facts, or substituting the appearance of rigor for genuine interpretation. The sophisticated practitioner holds both: uses computational methods as a powerful discovery tool, then returns to close reading and contextual knowledge to interpret what was found.
Topics in reflective domains aren't scored by quiz answers. Read, reflect, and mark when you've thought it through.