Familiar Stories: Using CLIP to Extract and Analyze Images of the Family in Dutch-Language Children's Books, 1800-1940.

Starting from Alston's claim that the representation of families in children's literature reflects a tenacious disciplinary discourse, which is marked by stark continuity rather than change, we analyze the representation of families in the illustrations of historical Dutch-language children's literature (1800-1940). We use CLIP, a multimodal machine learning model, to extract the illustrations which depict a family from our dataset of 20.000 illustrations. In our presentation we highlight both the potential and the pitfalls of this methodology for research in the Humanities.