A courtroom takes months to reach a verdict. A photograph reaches one in the time it takes to look at it.
This is the actual mechanism behind a familiar complaint — that a suspect's face in the press has already been convicted before any trial starts. It's usually framed as bias, as a lapse in fairness, something a more careful media culture could correct with better disclaimers. But the mechanism doing the work isn't bias in the ordinary sense. It's a category substitution: a photograph is treated as though it were evidence, when what it actually offers is a reading.
Evidence has to survive cross-examination. It can be challenged, contextualized, thrown out on procedural grounds, weighed against a burden of proof. A photograph does none of that. It doesn't argue a case — it produces an impression, instantly, and the impression arrives already dressed as a conclusion. The viewer doesn't experience themselves as inferring guilt from an image. They experience themselves as seeing it. The photograph absorbs the labor of judgment into the labor of looking, and looking feels involuntary in a way that judging doesn't. Nobody feels responsible for what their eyes report.
The pattern is old enough to have a paper trail. Salem's spectral evidence let a witness's account of a dream stand in for testimony a defendant couldn't cross-examine, because the vision itself was treated as the proof. Victorian mugshot galleries, circulated in the penny press decades before any forensic use for them existed, existed to let readers do exactly this: look at a face and settle something. The McMartin preschool case ran largely on the strength of children's testimony shaped by therapists who were themselves convinced by the look of the accused before any of it reached a courtroom. In each case the actual evidentiary standard — cross-examinable, falsifiable, procedurally bounded — got quietly replaced by an unaccountable one: does this person look like what we already believe they are.
What makes the substitution durable rather than just common is that it's self-confirming in a specific way. Once a face has been publicly read as guilty, every subsequent appearance of that face gets processed through the reading already established. The second photograph doesn't get evaluated independently — it gets slotted into the verdict the first one produced. This is why a formal acquittal so rarely undoes the informal one. The legal process can revise a finding of fact. It has no mechanism for revising a finding that was never framed as fact-based to begin with, because there's nothing in the record to appeal — no argument was made that a counter-argument could answer. The image just sat there, being looked at, and the looking did all the deciding.
The uncomfortable part isn't that this happens to unsympathetic people, or even that it sometimes happens to correctly guilty people for the wrong reasons. It's that the mechanism has no separate setting for correctness. It doesn't know whether it's right. It runs the same way regardless of the facts, because it was never built to check them — it was built to let a viewer arrive at a feeling and mistake it for having weighed something.