52

Since the 2000s, new techniques for text analysis have appeared at the intersection of computer science, artificial intelligence, and natural language processing. While these technologies were designed independently of sociological concerns, they are now being used by researchers—sociologists and non-sociologists alike—with a view to renewing our knowledge of the social by drawing on the considerable volume of textual materials now available. By providing an overview of the sociological investigations that are based on the “datafication” and quantitative analysis of text corpora, this article identifies the conditions necessary for these approaches to constitute a resource for sociological inquiry. Three conditions emerge from our analysis: 1) knowledge of the context in which textual traces are produced; 2) the incorporation of extra-textual data into the inquiry; and 3) the adjustment of algorithms to sociological reasoning.