Poétique Du Mixtape and Knowledge Graphs

This really doesn't have much to do with disco, except that I usually find a way to make everything about DJing, which I find endlessly fascinating on a theoretical level. DJing is about connecting lots of things which may be quite similar to each other or may be wildly different. Stéphane Girard has theorized the different facets that make up the identify of each selection a DJ plays in his Poétique du mixtape (Les Éditions de ta Mère, 2018), a book I finally finished reading recently. They are:

  • L'identité générique (what genre or genres does the song relate to?)
  • L'identité auctoriale (here, Girard is focused on what the name of the artist connotes)
  • L'identité linguistique (what lyrical content does it carry, which includes the song title)
  • L'identité spatiale (where did it come from? What scene? What continent?)
  • L'identité temporelle (what point in time does it come from?)
  • L'identité harmonique (the musical content)
  • L'identité symbolique (think cultural capital: is it obscure or a mega hit?)

Generic, temporal and spatial facets already take you pretty far to establishing a framework for triangulating a "position" for each piece of music relative to everything else. That's one reason I always try to figure out the release year and geographic origin for every song I play on the show.

Sidenote: there's a thematic dossier related to Girard's book with more writings and some mixes. Girard really knows his way around the DJ scene, referencing everyone from minimal icon Ricardo Villalobos to SOPHIE, Detroit's Omar-S, Pete Tong, and the Discwoman collective. He is thoroughly familiar with all the major writing on DJing to date. The highlight of the book is an in-depth analysis of Montreal superstar Tiga, and the development of his persona over the course of multiple mixtapes and albums. You can tell this book was a labour of love. The only trick is that it is only available en français.

Figuring out which musicians were working with who, or in the same studio or for the same label is another way that people look to find connections (something shared in common? similarity or difference?) to explore music. One way that people have attempted to map these kinds of networks is with so-called linked data, which is, in broad terms, a schema-less data format ideal for sharing and combining data from different sources. The Linked Jazz project is an example of this approach. The idea is that you can continue to add data at any time and then use special query languages to carry out the type of network analysis that is common in various applications today.

Linked data, the semantic web and knowledge graphs are things I study in my professional capacity as a librarian and I recently published an article that proposes using the mathematical ontology of philosopher Alain Badiou to theorize knowledge graphs. You can check it out over here in the very 'cool' journal KULA along with many other impressive papers about metadata and knowledge.