Literary history is no longer written in books alone. As literary reception thrives in blogs, Wikipedia entries, Amazon reviews, and Goodreads profiles, the Web has become a key platform for the exchange of information on literature. Although conventional printed media in the field—academic monographs, literary supplements, and magazines—may still claim the highest authority, online media presumably provide the first (and possibly the only) source for many readers casually interested in literary history. Wikipedia offers quick and free answers to readers' questions and the range of topics described in its entries dramatically exceeds the volume any printed encyclopedia could possibly cover. While an important share of this expanding knowledge base about literature is produced bottom-up (user based and crowd-sourced), search engines such as Google have become brokers in this online economy of knowledge, organizing information on the Web for its users. Similar to the printed literary histories, search engines prioritize certain information sources over others when ranking and sorting Web pages; as such, their search algorithms create hierarchies of books, authors, and periods.

This article explores these algorithmically constructed hierarchies as cultural representations of what is (and what is not) presented as important to the Google user, taking information about authors from a particular body of national literature as a case study. We examine the relations between a sample of Dutch writers through the carousels of related searches generated by Google's Knowledge Graph. The sample used for this experiment comprises all 2,287 individuals who were labeled on Dutch Wikipedia with the category "Dutch writer" (Nederlandse schrijver). On Wikipedia, this is the general category for book authors from the Netherlands and includes Dutch writers from all possible genres, such as literary prose, poetry, (literary) thrillers, fantasy, nonfiction, and cookbooks. The names of authors were fed into the search engine and, for each writer, all entities—either Dutch writers or other individuals—that Google returned under the "People also search for"-function were scraped and stored. By using methods derived from network analysis, we then compiled a "canon" of Dutch literature as it emerged through the relationships established by Google's Knowledge Graph. Furthermore, to evaluate the network constructed in this way, a comparison was made between this canon and the academic preferences concerning literary authors described in Dutch literary historiography. This comparison focuses on the gender balance and occurrences of the 3453 authors mentioned in three volumes (covering the period 1800-2005) from the recent nine-volume series on Dutch literary history Geschiedenis van de Nederlandse literatuur published between 2006 and 2017.5 The results allow an assessment of Google's possibilities for constructing alternative hierarchies of canonicity or confirming the approach to the canon that remains prevalent within the Dutch literary field.
Original languageEnglish
JournalJournal of Cultural Analytics
DOI
Publication statusPublished - 01 Oct 2019

ID: 11757975