These are the slides from a talk on the role of heteromorphisms (hets) in category theory given at the Category Theory Seminar at NYU on January 13, 2016.
Saunders Mac Lane famously said that Bourbaki (in the person of Pierre Samuel) “just missed” developing the notion of an adjunction in 1948 a decade before Daniel Kan. When adjoints are defined using heteromorphisms or chimera morphisms (first done by Bodo Pareigis in 1970), then indeed Samuel was only a now-simple dualization away from defining adjoints. Since the heteromorphic treatment allows adjunctions to be factored into two, in general, independent universal mapping properties (UMPs), we also pose the question as to whether adjunctions or UMPs should be considered as category theory’s most important concept. It seems Mac Lane and most orthodox category theorists would say adjunctions but Grothendieck among others would say UMPs stated using representable functors.
We give a heterodox treatment of adjoints using heteromorphisms (object-to-object morphisms between objects of different categories) that parses an adjunction into two separate parts (left and right representations of heteromorphisms). Then these separate parts can be recombined in a new way to define a cognate concept, the brain functor, to abstractly model the functions of perception and action of a brain. This is a preprint of the paper coming out in Axiomathes.