Schneier on Security

antibozo • July 24, 2007 9:30 PM

Knowler Longcloak> Again it is a statement about the photograph that is true or false. The statement you bring up is “This photo is an accurate representation of what was there.”

Knowler Longcloak> The photograph itself is neither true or false.

Uh, no.

Truth is generally the degree to which a model resembles the thing modeled. The metric by which we measure truth varies with context. A “true” statement is an accurate semantic model of its subject (which may be real or metaphysical). A “true” wheel accurately models an idealized wheel geometrically. A “true” photograph might be one in which parallel lines on the subject are rendered in parallel on the photograph, one that accurately reproduces color for the human eye, one that accurately reproduces tone for the human eye, or any number of other things, depending on the properties of the model that are sought after for the purpose of the photograph–the chosen metric. All of that is independent of anything written nearby; information is not conveyed solely by captions, after all.

Ceci n’est pas une pipe. Everything is a model, including this.

antibozo • July 25, 2007 3:36 AM

someone> a friend of mine differentiates between images and photos. a photo is not allowed to be modified, an image is.
someone> i think this language should be introduced in web and print.

There is already an established distinction: a photo is an image that is produced by a photographic process, i.e. a process where electromagnetic radiation from the subject is projected through a lens onto a recording surface, generating a chemical or electrical image of the subject.

antibozo • July 25, 2007 1:00 PM

bako> So we’re all in agreement. You need context for a photo to be labelled true or false.

I don’t think we agree at all. As I said, you need context for anything to be labeled true of false. There is no distinction between a photo and a statement in a given language as to its objective truth–to assess truth of any representation of information, you always need to decide what the representation is modeling.

One might argue, in fact, that truth of a photo is easier to assess objectively than that of a statement. If a sighted extraterrestrial visits Yosemite holding a calendar with an Ansel Adams photograph of Half Dome, it is going to be able to say the photograph is true by a number of measures, while completely unable to assess the truth of the photo’s caption without doing considerable research into human language, the history of names of North American features (not everyone has called it “Half Dome”), the history of photography (to determine whether use of the name “Half Dome” is synchronic with the representation), Ansel Adams, etc.

antibozo • July 25, 2007 1:56 PM

bako> I think we agree on the objectivity of truth re a photo. If we don’t, I”m not sure where.

Perhaps we agree on that, but citing that alone as you did above seems dismissive of our apparent differences.

bako> As to the Half Dome photo, by what measures could it be called “true”?

1. By a monotonic mapping between the energy of reflected light from Half Dome at specific wavelengths and the tone represented in the photograph, for an observer situated somewhere near where Adams set up his view camera.
2. By a spatial mapping between the morphology of a projection of Half Dome through an idealized lens onto a planar surface and the morphology of the photograph.
3. Adding some context about Adams’s particular processes, by the manner in which the photograph models his strategy for mapping measured light energy into printed tone values.
4. Adding some historical and psychological context, by the manner in which the relative effort, care, and quality of the processes Adams used to photograph and print the image of Half Dome reflect a human reverence for the processes which produced Half Dome itself, and for the effect of its sight on the human mind.

One can continue evaluating the photograph with increasing knowledge of its context to understand additional truth. To get to the point of understanding the truth of its caption, however, the extraterrestrial must do a huge amount of work, while some truth is readily evident given the photograph alone.

For comparison, consider this scenario. In a cave in Yosemite, someone discovers what appears to be a prehistoric painting of Half Dome, but with the English words “Half Dome” inscribed using the same process beneath it. We show this representation to the extraterrestrial. How much context will the extraterrestrial need to understand that there is something false about this representation, even though the painting itself represents the morphology of Half Dome accurately?

bako> Maybe it’s a photo of a very good model of the dome. True or no?

True if by the chosen metrics it is evaluated thus. A scale model of the Titanic may be very true in terms of linear proportions, while false in terms of the material it is made out of, its location somewhere other than the bottom of the North Atlantic ocean, the intactness of its hull, the function of its propulsion system, etc.

antibozo • July 26, 2007 3:48 AM

bako> I originally quoted you because you negated someone else when they said that a photo in itself cannot be true or false.

What I disagreed with is the notion that a photo by itself has any more or less inherent truth than any other representation of information.

The essay, while moderately interesting (mostly because of the small history lesson it contains), is facile and fallacious. Here’s the argument in a nutshell:

1. The author writes: “Look at the photograph below. Is it true or false?” and shows us a captionless image of a ship. Then he comments, “I find the question [is it true or false] ridiculous: ‘True or false in regard to what?'”

Indeed, the question is ridiculous, as we haven’t been given a metric by which to measure truth. That’s the author’s fault, though.

Some readers–maritime and war historians, for example–will recognize the ship without a caption. These people will say, “That looks like a pretty good photograph of the Lusitania.” For these people, the photograph is already true, in that it accurately represents what the Lusitania looked like. For the rest of us, the photograph is still “true” if it is an image of some ship, even though we might not know what ship it is. The image is also somewhat false in that the highlights and shadows don’t accurately represent the reality that was photographed. But even without knowing what ship it was, we may still assess the truth of the photograph on a number of levels.

1. The author shows us the photograph again, this time captioned “The Lusitania”, and asserts that “Only now can we ask questions that have true or false answers.”

This is simply wrong. The caption doesn’t change the truth of the photograph. What the caption does is name a referent, and now we may assess, after some research, whether the referent named by the caption is in fact the photograph’s referent, and thus whether the caption is true. Obviously, before we had the caption, we couldn’t assess the truth of the caption, so the author has actually demonstrated nothing at all.

1. The author tells us a little about WWI and the context of the Lusitania. Then he shows us the image once more and claims that “the image remains the same, but clearly we look at the image in a different way.”

This is all completely unrelated to the truth of the photograph. It’s just more warm fuzzy writing to distract the critical mind. And of course the WWI historians already knew all that and aren’t looking at the image “in a different way” at all.

1. Finally, the author claims that “photographs are neither true nor false in and of themselves. They are only true or false with respect to statements that we make about them or the questions that we might ask of them.”

This conclusion is not supported by the essay. Ironically, the author apparently hasn’t even thought much about what “truth” or “meaning” are.

Photographs are models. The truth of any model is its success at recording certain desired properties of its referent. Obviously if we don’t know what the referent is, we can’t say how true the model is. I might assert that “AB = 6″, but no one can assess the truth of that statement without knowing more. But the truth of “AB = 6″ is not determined by anyone’s knowledge of A and B. It is determined by the semantic values of the various symbols in the formula, and the particular values of A and B. To assess the truth of any model, we always must first know what the model’s referent is and what metric of truth we are to apply.

antibozo • July 28, 2007 6:21 PM

X the Unknown> I think that the statement “Truth or falsity “adheres” not to the photograph itself but to the statements we make about a photograph” is insightful, and is basically a concise way of saying the context in which a photograph (or anything else, really) is presented affects our judgment of veracity.

But isn’t it obvious that context influences everything? And “veracity” itself has many layers.

X the Unknown> A very common example of this would be those pictures of people standing with their arm around some celebrity or another – when the celebrity was really a photorealistic cardboard cutout. Since the photo of me standing with Gretta Garbo (or whomever) is itself two-dimensional, if it was taken from the correct angle it is very difficult to tell that the celebrity isn’t “real”.

But here’s what really happens: the photograph is a model of a scene. The viewer sees the image and autonomically builds a spatial model of the scene in his or her head, then compares the spatial model with the image. If the lighting or scale is inconsistent, the viewer may detect this and reanalyze the image (again autonomically) to find a more plausible scene. If there’s no cue that the referent scene isn’t actually you with Greta Garbo, the viewer may assume that you were standing next to Greta Garbo–what you mean by “default context”.

But the viewer’s assumption, right or wrong, isn’t about the truth of the image; it’s a result of a complex mental process that factors in context and knowledge. If the image is next to a newspaper article about Greta Garbo, the natural assumption is that the image represents what it appears on the surface–a scene that really happened. If, however, the photograph is on page sixteen of the book “Pictures of Me with Cardboard Cutouts”, then the viewer’s mentally constructed scene will tend toward the cardboard cutout version–even if the referent scene did not use a cardboard cutout. It’s not the photograph itself that does this; it’s a matter of how the photo’s context–caption, juxtaposition with other images, enclosing medium, etc–influences how the viewer imagines the referent scene of the photo, combined with the viewer’s knowledge. After all, if you already know it’s a cardboard cutout, that’s the scene you’ll build in your mind, no matter what the photo looks like.

It’s no different with text, audio, film, or any other medium. The medium itself presents a physical model of some referent. The viewer uses that physical model to build a conceptual model of the referent. That referent may be part of a long chain of reference linked to real, historical, or metaphysical concepts, and the viewer will perform varying consistency checks on many layers in that chain.

X the Unknown> Similarly, certain optical illusions could be said to be “inherently false” images. They are constructed in such a way as to deliberately abuse the limitations and mechanisms of vision and/or recognition. Of course, we don’t generally think of them as True or False, but rather as interesting. But, the fact that a normal person cannot accurately see the totality of the image would certainly seem to make it a “false image”.

Well, consider special effects in film–in fact, consider this scene in particular, from “Postcards from the Edge”. We see Meryl Streep hanging from the side of a building, the camera looking down over the edge to see the street below, clearly a fatal fall if she should let go. She cries “Help! Help!” Is the image false? No, the image is an accurate representation of what the camera sees. Is our conceptual model of the scene false? Yes, but we don’t know until Meryl’s cries become less convincing and she then lifts her hands from the edge of the building, at which point we perceive that she is simply lying horizontally on top of a false building with forced perspective and background painting–filming a movie scene–and we revise our conceptual model of the scene suitably. The camera photographs an image which our minds plausibly reconstruct as a life-threatening situation, reinforced by her cries for help. Before the reveal we may imagine her state of mind, and sympathize with her fear of falling to certain death and feeling of desparation and despair. All of these things are true–these are things people feel in such a situation. After the reveal, we understand different true things–Meryl’s character is an actress, her desperation is false or perhaps self-inflicted, she’s feeling depressed and uncertain about her career. All of this is, of course, in the context of a movie, so none of the surface referents are real, though we may understand, if we know enough about the film, that the story is from a semi-autobiographical book by Carrie Fisher.

Despite all these layers of reference, what most of us understand theoretically about human behavior and experience from watching this scene is true in many ways. One of those is that our conceptual model of Meryl’s character in the scene matches the model the scene’s author had in mind.

I could go on, but I’ve bored everyone to death already.

[X the Unknown–fun film, BTW. ;^) ]