Our take: What makes this article particularly compelling is how it questions whether extracting geographical information from historical texts necessarily produces an accurate representation of the past. Using Thomas Roe’s seventeenth-century travel accounts, the author highlights a crucial distinction between places a traveller actually visited and those mentioned through secondhand information. This piece moves beyond the technical capabilities of Named Entity Recognition (NER) to consider the interpretive responsibilities involved in computational historical research. Similar problems arise when mapping colonial archives, reconstructing historical networks, or interpreting geographical references in literary texts. It demonstrates why computational accuracy cannot be separated from historical interpretation, reminding us that identifying a place in a text is not the same as establishing someone’s presence there.
Editors’ Choice: Where They Actually Went: The Presence Problem in Digitised Travel Writing
