Photography
The medium that changed what we saw
Description
Photography arrived in 1839 and changed what people meant when they said they had seen something. Before photography, the visual record was painting, drawing, and engraving all filtered through a human hand, all slow, all limited in how many copies could exist. After photography, an image of an object could be produced in minutes, multiplied indefinitely, and circulated across continents. The implications for art, journalism, science, memory, and the basic question of what counts as evidence took two centuries to work out, and we are still working some of them out now.
The medium has always had a double identity. From the start it was both a tool a faster, more accurate way of recording the visible and a candidate for art, with its own techniques and traditions. The two identities have never fully separated. A press photograph from a war zone is being judged on its accuracy as a record and its power as an image, and these are not always the same thing. A photograph hung in a gallery is being judged as an aesthetic object, but its claim still depends partly on the camera having been there.
Two technological shifts define the medium's history. The first was the original invention the daguerreotype and the calotype in the 1830s, and the half century of improvements that followed. The second is the digital transition that started in the 1990s and was completed, for practical purposes, by the 2010s. We are now in a third shift, with AI-generated images that look photographic but were produced without a camera, and the question of what a photograph is, and what it certifies, is being asked again.
The question we're asking: what photography did to vision, and what happens when the medium changes underneath it.
What we'll see: the 1839 invention, the evidence-art split, the digital flood, and the AI question.
Table of contents
01Daguerre and 1839
The official birth of photography is usually dated to 1839, when Louis Daguerre announced his process to the French Academy of Sciences. The daguerreotype produced a sharp, detailed image on a silver-coated copper plate after an exposure of several minutes. The French government bought the rights and made the process freely available, which is part of why the medium spread as quickly as it did. William Henry Fox Talbot in England announced a competing process the calotype at almost the same moment, using a paper negative from which multiple positive prints could be made. Talbot's negative-positive principle was the one the next century would actually use.
The early uses of photography were varied. Portraiture was the first commercial application a daguerreotype portrait was substantially cheaper than a painted miniature, and within a few years there were studios in every major city. Landscape and architectural photography took advantage of the medium's accuracy in ways drawing could not match. Scientific applications astronomy, microscopy, medical documentation appeared almost immediately. By the 1860s, a photographer like Mathew Brady was documenting the American Civil War, producing the first images of mass military death that civilian audiences had ever seen. The cultural shock of these images is hard to recover now, when photographic atrocity is routine.
02Evidence and art
The argument over whether photography was art started early and has never fully ended. The early objection was that the camera was a mechanical recorder rather than an instrument of artistic judgement, and that the work of pressing a shutter was not comparable to the work of making a painting. Photographers spent the late nineteenth century trying to answer this objection, often by imitating the conventions of fine art soft focus, painterly composition, allegorical subjects. The pictorialists, as the movement was called, made photographs that tried not to look photographic. Some of the work is still beautiful. The strategy of legitimising the medium by hiding its specifics was not, in retrospect, the right one.
Alfred Stieglitz in New York at the start of the twentieth century took a different position. The medium should be treated as art on its own terms, with the qualities that the camera and the print actually had sharpness, tonal range, the specific way photography registers light — as the basis of the work rather than as embarrassments to be hidden. Stieglitz ran a magazine called Camera Work and a gallery called 291 that showed photography alongside contemporary painting and sculpture, and the project of photography as serious art took its modern shape in this period.
03The digital flood
The digital camera, in its early consumer form in the late 1990s, was clearly worse than film. The resolution was lower, the colour was less accurate, and the storage was awkward. Within ten years, this had reversed. By the late 2000s, professional digital sensors had passed film for most practical purposes, and the workflow of shooting, importing, and processing on a computer had become standard. The negative the physical object that for 150 years had been the source from which photographic prints were made was no longer how photography happened. The shift was complete enough that most working photographers under forty have never used film for paid work.
The phone camera was the larger change. Camera phones in the early 2000s produced poor images. By the 2010s, a phone camera could produce results that would have been considered acceptable from a professional camera a few years earlier, and almost everyone in the developed world was carrying one. The volume of photographs taken globally increased by orders of magnitude. Estimates put the number at over a trillion images per year by the late 2010s, the great majority of them seen by their maker and almost nobody else. Photography had become a default mode of attention rather than a specialised practice.
04The AI question
Generative AI image models DALL-E, Midjourney, Stable Diffusion and the systems that came after them appeared in usable form around 2022. They produce images on the basis of text prompts, drawing on the very large body of online images they were trained on. The early outputs were obviously synthetic. Within two years, the best systems were producing photorealistic images that, in many cases, could not be reliably distinguished from photographs by an unaided human viewer. The question of what a photograph is, and what it can certify, is being asked again with new urgency.
The narrowly technical answer is that a photograph is an image produced by light from an actual scene falling on a sensor or film, while a generated image is an image produced by a model with no scene behind it. This distinction matters legally, journalistically, and forensically. It is also harder to maintain at the level of how the image looks, since contemporary AI images often look more conventionally photographic than many real photographs do. The distinction has to be supported by metadata, watermarking, or context, none of which travel reliably across platforms.
05Conclusion
Photography for almost two centuries had a special claim on the visual record. It was the medium that had been there, that had registered the light from an actual face or street or battlefield, and that could be held up as evidence in a way no other image could. The claim was always partly true and partly negotiated. It is now being renegotiated under conditions different from any earlier moment. The camera is in everyone's pocket. The synthetic image is approaching photographic plausibility.