“…Techniques leveraged from natural language processing (NLP) and machine learning have recently facilitated more widespread use of connected discourse (e.g., storybooks, podcasts, corpora) in studies of semantic processing (Baldassano et al, 2017;Deniz et al, 2019;Günther et al, 2019;Hartung et al, 2020;Huth et al, 2016;Jain & Huth, 2018;Kumar, 2021;Mandera et al, 2015Mandera et al, , 2017Nastase et al, 2020;Popham et al, 2021a). A key advantage of such approaches is their capacity to extract distributional statistics from vast numbers of tokens appearing in real world corpora (Baldassano et al, 2017;de Heer et al, 2017;Hartung et al, 2020;Huth et al, 2012;Jain & Huth, 2018;Naselaris et al, 2011;Popham et al, 2021b;Simony et al, 2016).…”