engleski

Portraits of people that do not exist. Generative Adversarial Networks (GANs) and the problem of existence in portraiture

Although not being trustworthy in terms of qualities, the portrait genre is strict in terms of existence. It represents people that do/did exist. Jean Luc Nancy thus wrote that portrait represents someone who necessary exists/or existed defining that it: 'paints a subject only by setting itself within a subject-relation ; as such, it sees a putative subject (me, you, the painter) within relation to the subject that is being exposed. It sets a subject within a subject- relation and so within relation to self'. (Nancy, 2018, 19). Portrait establishes a triple relation ; 'portrait resembles (me), the portrait recalls (me), the portrait looks (at me)' (Nancy 2018, 20). None of Nancy's relations works with portraits of people generated by Generative Adversarial Networks. For example, in This Person does not Exist, as no comparing subject to any of these images. At the same time, a creator is a machine (Helfand 2019). Yet, this problem can be solved with object-oriented ontologies. According to these ontologies, still, even objects that are not strictly "things" (or physical) exist, in a certain way. Or as Graham Harman, the author of the theory of OOO flamboyantly describes ; "along with diamonds, rope and neutrons, objects may include armies, monsters, square circles and leagues of real and fictitious nations." (Harman 2011, 66-67). So is true with "people who do not exist" - they do exist for everyone who is incapable of distinguishing their physical presence (or can experience individual existence remotely), and that is especially convincing in photographic medium as we are used that – whatever is photographed necessarily exists. Such objects, which do not refer to physical things, but are rather ideas Francois Laruelle names a photo- fiction, defining "a new type of object, about adding fiction to the photo according to a precise logic, without imitation or dialectics, and then elucidating its structure" (Laruelle 2012, 12). Such objects are not made by adding photographs but by "generating fictions that are alike "theoretical captions" (ibid). Both types of objects construct the world, which, according to Laruelle, is named a photosphere (Laruelle, 2011). Thus, both real portraits and generated ones create the sphere of portraiture.

artificial people ; Generative Adversarial Networks ; generated photographs ; photo collage ; photographyOn

nije evidentirano