The Diagrams of AI (Image)

Abstract

A history of images about images is as mesmerising in its own right as images themselves are. This is not merely a history of the copy—the attempt to reproduce an image through an image— but also of the various guides and diagrams that tell a story of production of images. This is also the entry point for my argument about the changing ontology of the image. This argument about ontology concerns then not merely what an image is in its essence, but how images function as as operative ontologies1: described, drawn, pictured, instructed, guided, and diagrammed into existence. Such diagrams are an educational arm of knowledge about images, but obviously they are also images already in themselves. Diagrams occupy a central role as a modern form of knowledge about images. Diagrams that describe the operations and insights of image geometry are a special case in point, where the linear perspective in (and since) the Renaissance period has given rise to a long line of commentary, in the arthistorical way of tracking the changing ontologies of the image. How to calculate image surfaces, lines, and ratios becomes instrumentalised into a productive machinery and subsequently into an analytical machinery, as is the case in the various techniques of reading the geometric data packed into an image. From Johan Heinrich Lambert’s Die freye Perspective, oder Anweisung Jeden Perspektivischen Aufriß Von Freyen Stücken Und Ohne Grundriß Zu Verfertigen (1759) to Colonel Aimé Laussedat’s works on photogrammetry (or “metrophotography”) toward the latter part of the 19th century, the work of descriptive geometry becomes crucial to the diagram of the technical image and image as data.2 They are manuals of “this is how that operates” and take on a second order quality themselves: a cultural technique that recursively images an image. One can also observe a similarity with the function of the metapicture, as per WJT Mitchell’s term that refers to such images that “might be capable of reflection on themselves, capable of providing a second-order discourse that tells us—or at least shows us—something about pictures.”3 In terms of the contemporary image, the shift from questions on the ontology of digital images (do they capture reality? Do THE DIAGRAMS OF AI (IMAGE)