Photography has always had a ghostly character. Contrary to the early and dominant theorists, the connection or ontology of photography with reality was not actually related to the camera or camera obscura. Here, the dominant discourse that concerns the majority and draws photography to itself, the theories created by the concepts of reference, trace, indexicality and causality, have become so entrenched that photography education, as in most geographies, has not been able to go beyond this theory in our country. The influences of theorists such as Andre Bazin (1960), Charles Sanders Peirce (Lefebvre, 2007), Roland Barthes (1981), Rosalind Krauss (1977a; 1977b) and Walter Benjamin (1972), while confronting photography only with the question of whether it is real or not, have also attracted postmodern discourse and have brought the automatic nature of photography and its indexical relationship with reality to the forefront. Based on this interpretation, Toister adopts the concept of similarity in reassessing the ontology of photography – arguing that photography should be understood as a set of diverse practices with overlapping characteristics rather than as a medium with a single defining feature. The author ultimately argues that photography is not a unified medium with a fixed essence, but rather as a set of media with diverse characteristics. Drawing on the philosophy of Vilém Flusser, Toister (2020) reinterprets the term “photography” not as “light” but as “appearance”, allowing photography to contain illusions and possibilities beyond fixed realities. In parallel, Fontcuberta’s argument regarding the literal definition of photography is related to “ghost” and “imagination (or dream, imagination, fiction)” in relation to appearance, rather than “brightness (light)” or “spark”. Here, Fontcuberta unwittingly refers to the classical again, to Walter Benjamin’s (1999) ability to see and commission photography as a savior of history, to moments of gleaming. Fontcuberta (1997, p. 166) argues that photography is more concerned with appearances - illusions or phenomena - rather than an objective reality derived from light. He criticizes the historical role of the medium as a means of representing reality and instead argues that photography should be understood as "the writing of appearances" and is inherently linked to illusion and constructed meaning.

Therefore, photography is not a camera obscura first of all [1] . Because before it became a camera obscura, the idea of photography always existed in the Platonic and Neoplatonian period. According to Mikuriya, the invention of photography is only a material manifestation of what has always existed. The existence of photography at the root of this emergence can be revealed in the roots of Western metaphysics, in the discourses of magic, mysticism and spiritual practice (2017,11). Slater (2017) explains these images by arguing that the scenario revealed by the diaroma exceeds the objective goal of modernity to explain the mysteries of the world. Photography, on the contrary, revives enchantment. Paradoxically, it aims to re-enchant the world with natural magic rather than rational analysis. In this sense, the diaromas created by Daguerre were not just representations; they were simulations full of immersive virtual experiences. Photography is, in a sense, a rhizomatic event: a fluid, cyclical and ghostly image. Therefore, photography has never been post-. The aftereffect is the state of the image disintegrating, becoming fluid, waiting for a new rethinking process that needs time to emerge, just to overcome the dominant discourse. In other words, it is the field of image study that necessitates the end of the discourses of the first theorists who created dichotomies such as reality and truth. Likewise, as Fontcuberta also states, post-photography is not a new technique, style or movement in the digital age, but a complete rethinking of the image. In a time when the internet and social media are proliferating, it uses new media in both its theory and practice as a way of investigating what constitutes the post-photographic image (Moreiras, 2017).

Now, since 2014, we are immersed in the images produced by the machine learning model Generative Adversarial Networks (GAN; Goodfellow et al, 2014), which consists of two neural networks working in opposite directions to produce new data. The first topic of discussion was, of course, whether these images were photographs or real. In her definition of “mean images,” Hito Steyerl (2023) redefines these images in the context of machine learning (ML) and artificial intelligence (AI), arguing that images generated by AI are no longer bound to the indexicality of traditional photography, but instead represent statistical representations, that is, images generated according to probabilities rather than real-world references. Steyerl compares this change to historical traditions of photography, which relied on the direct effect of light on film or sensors. Rather than recording reality, AI-generated images average out large data sets, producing “mean images” in both a statistical sense (mathematical average) and a socio-political sense (pejorative, reductionist, and shaped by societal biases), she says. [2]

We can reconsider this dual approach through Robert Capa’s famous photograph, The Falling Soldier. Following its capture, the image became the center of intense debate, with critics questioning whether it was an authentic moment or a staged composition. Richard Whelan (2017) explains the inextricable situation of the photograph as follows: “I struggled with the dilemma of how to deal with a photograph that was believed to be real, but was not a genuine document. With a title that spoke of the doubts raised about its authenticity? Did the stain of doubt render it permanently ineffective? Should Capa’s photograph be consigned to the dustbin of history? As I will try to show here, the truth about the “Falling Soldier” is neither black nor white. It is neither a photograph of a man pretending to be shot nor an image taken at a moment we would normally recognize as the heat of battle.” Leaving us in a ghostly space, this photograph reveals the photograph’s intention to document something that does not exist. We may define this as fiction or propaganda. On the other hand, the definition of photographs produced by artificial intelligence is that they are not real because they are data belonging to another data set. The photographs produced with this data have been called non-human photography (Zylinska, 2017), neural photography (Zylinska, 2023) or hyperobject [3] . From here, let’s consider Capa’s photo again with a GAN-based text-to-image model, which is a type of artificial intelligence that uses machine learning algorithms to understand the meaning of a text prompt and generate a corresponding image, that is, it generates images based on a textual description provided as input. Let’s assume that Capa actually saw the moment of the incident and was unable to photograph it even though he had a camera in his hand. To obtain the “Falling Soldier” photo, let’s use the following text about the moment of the incident: “A Republican soldier at the moment of death. The soldier, shot in the head, falls backwards with his rifle flying out of his right hand. The soldier is wearing civilian clothes, trousers and a white shirt with the sleeves rolled up, but he is wearing a leather cartridge belt around his waist. The photo is set from a low angle in a rural area with the sky as the background.”