I Witness
The Belief in Seeing
Analysts now say that eyewitness testimony in court cases is actually unreliable. Yet when trial is underway, most people, most of the time, will elevate the testimony of someone who says “I was there, and I saw something”…
If that testimony works against your interest, the key defensive tactic is to erode the credibility of the testifying party. “Regardless of what you think you saw, why should we believe you?”
When you think it through, the power of this question involves two things.
One, if you, as well as the witness, were at the scene yourself, you would compare your own confidence in what you saw against your confidence in what they may have seen. Did the “data” collected by their look at things match what you likely would have noticed?
And two, you would compare your idea of what importance should be attached to what they say they saw, gauging why that attachment might be too weak to withstand debate. Is it logical? Is it relevant?
Yet another thing, pointed out in the famous “Invisible Gorilla” experiment, is that what we notice when we are looking around is heavily preconditioned by what for some reason we are expecting beforehand. The key finding in this experiment is that we don’t see as well as do see, based on why we are paying attention.
(This video features co-creator Daniel Simons explaining the original experiment and includes the famous footage where you can test your own awareness.)
… The Invisible Gorilla (featuring Daniel Simons) …
And a final interesting phenomenon is that in seeing, the human brain does something automatically, which is to try to predict what’s coming before it comes, or inject what’s missing that “ought to” be there. Mostly this allows us to see a part of something and, based on that, imagine the rest of a “whole”. The question is, what do we think that whole thing is? Are we right, or wrong?
Furthermore, that applies to things that are moving as well as to things that are not.
So far, the above may not have brought up anything of news value to you. But it gives a kind of framework within which to understand how some images affect us as viewers, and highlights that when we are viewing we are interpreting.
In this chart above, we see that in step 4 there is a big opportunity to contest or confirm the credibility of a “witness’s” claimed information. We might just go make ourselves witnesses, too, assuming that the important data still exists at the scene or can be proved to have been there. Unfortunately, this sometimes can happen only as a re-enactment of conditions that will argue “it must have been this way” (or “it couldn’t have been that way”). For simplicity’s sake, let’s call that test “Plausibility” because it is not the same as “Certainty”.
And, let’s call that notion of reenactment Re-Presentation. The weakest case in our favor, then, is one where a representation is plausible — acceptable in the absence of hard core verification and certainty by contact with the original phenomenon.
All of that usually gets bundled into the category of information called “evidence”.
Getting from step 4’s evidence to step 5’s non-fiction is not as easy as it looks. Using evidence to build a story is one way that things could go. But the other way is to use a story as the selection criteria for evidence. When that second pathway is chosen, the question immediately arises as to how we decide the story is the right one to use. The story organizes information into patterns that have meaning. This is essentially what “modeling” is, and in studies of perception, the notion of a mental model is predominant. It’s how that gorilla winds up “not showing”, or why tiny amounts of some given color in a scene are suddenly vivid only after we are told to find them.
Analysts call this mental modeling cognitive bias.
When the preconceived “story” tells us how to interpret the evidence, there is no absolute version of objectivity available to say what is correct or incorrect.
Instead, there is relevance and correspondence that is tested by our own strength as being interpreters. And the more we want a certain interpretation to be accepted, the more we have to evaluate the importance of two things: why we want that to happen, and where the story is coming from. Those are bundled together in the singular descriptive term circumstantial.
How do we see that in terms of whether something is fictional or non-fictional? The real difference is not in the result of the interpretation. Instead, it is in the intent of interpreting. Non-fiction holds itself to a standard of high-fidelity to verifiable evidence as the qualifier for what gets modeled into a description or explanation. The model then is acceptable whether the evidence is circumstantial or not.
That represents the quality of the model, but it does not define the value of it.
Back when the earth was flat and the sun revolved around the earth, circumstantial evidence was so convincing that most people didn’t question the story that used it. The actual gap between description and explanation wasn’t strongly felt, because the subject matter itself (the cosmos) was not something easily accessible in most people’s “real life” experience, but the representation gave a reliable explanation of something they cared about.
History, Science, Journalism, News
Observation, Representation, Images, Photography
With 150 years of photography under our belts, we have fully adopted expectations of what it offers as virtual witnessing of the world of visible environments and events. We typically spend no time wondering whether it can do it well.
But the effort to have it done is just one of three dimensions of its effectiveness a value. The other two dimensions are our ability to interpret photographs, and the motivation for making them (instead of doing something else). So, in the above flowchart, we see at the left the mindsets that govern our attention, and at the right the reasons why there is value.
Below, there is a more specific accounting of what is happening as a credible witness generates a plausible representation.
Usually we’re okay with proceeding as if plausible is as good as “truthful”. In detail, it points out how we might get from information (at left) to “truth” (at right)… Its description of photography is included to address photography’s legacy value of being inherently the most truthful of all visual mediums. But the idea that this outcome is inherent in photography is a fallacy; only the capability is inherent. In photography, there are (as seen here) fifteen (15) different opportunities to depart from the plausibility that we want to call truth. It is, instead, only the ethic of the imagemaker that prevents those departures from occurring when they might. Otherwise, if they do occur, someone has to correct them.
That faith in the image-maker is why we caution ourselves to “consider the source” when we evaluate received information. The more we think we know about the source, the more confident we are in our ability to evaluate whether the image’s plausibility is useful as “truth”. Truth is our agreement to trust in the plausibility.
False Witness, Fake News, Artificial Intelligence
As we think about documentation, journalism, news, and history, we can note their distinguishing kinds of rigor in practice. Our precaution of considering the source is important in each of the four practices. The contemporary environment of our exposure to images is now dominated by our powerful personal ability to transport a given image from one context to another, independently of what other people intend. And there is a community-level culture of “sharing” that is more about stimulating desired interpersonal relationships than about “truthful” knowledge. This social culture, powered by digital media technology, effectively shrinks the domain of influence for objective verification, leaving it more as a special purpose effort in the broader societal-level culture, and far less as a default basis for social stability and progress at that level.
A viewer’s trust in the practices of the image-maker is part of perceived value all across documentation, journalism, news, and history. Each of those practices loses meaningfulness when the image-maker does not present verifiable evidence of a “chain of custody” of evidence connecting what is re-presented to what has actually occurred.
Pretending to be a witness, reporting something that did not happen, or fabricating evidence are three powerful corruptions of an acceptable chain of custody. The safeguards and remedies against these problems must include both behavioral and pragmatic approaches.
There may not be any “silver bullet” solution to the runaway proliferation of what we today call false witness, fake news, and AI imaging. But in an evolutionary strategy, an obvious approach it to upgrade the existing mechanism that already systematically addresses the problems: Publishing.
There are many high-impact practices in society and the environment that are too useful to diminish but too risky to leave unmanaged. The way they are dealt with — from hunting to driving to construction to law to medicine — is by granting an earned license. Licensed practitioners carry the standards for value and risk management into the public domain. Conversely, accepting content, service, or assurances from unlicensed practitioners is a kind of personal risk-taking that makes a person directly responsible for the consequences.
It will be crucial for publishing to not allow a popular belief that reputations and degrees are sufficient instead of licensing. But it is also crucial to recognize that licensing itself can cover differing levels of competency to coexist within the same set of safeguards.
