Consider this image. Not much to see. Just a grid of randomly oriented line segments, yes?
However, scroll down, and an animation will disclose a simple pattern that is hiding in plain sight.
Now you see that the image contains a diamond shape whose sides consist of sequences of parallel segments. As the pattern starts to move, it comes into awareness, and seems to disappear as soon the pattern returns to its rest position.
There are two things involved here. There is the image itself — an array of pixels. Then there is your perception of it. Of course these are things of different kinds — one is out there in the world and one is in your mind. The latter is called a “phenomenon” by philosophers, and the discipline of studying phenomena, “phenomenology”. The objective here is to investigate the nature of the phenomenon of image 1.
The investigation is carried out by asking ourselves: what is present to us in the phenomenon, and what is not? What questions can we answer about it? Evidently the phenomenon of this particular image is not a mental copy of the external image. It does not contain a list of which pixels are white, and which black, nor even a list of orientations of line segments. Even while looking at the image, let alone recalling it, we don’t know those orientations. We can focus attention on a few individual segments and discover their orientation for a moment, but only a few, and these are immediately forgotten if our attention shifts. And, via such changes of focus, the phenomenon changes. It seems clear that the phenomenon as it is formed from our first look at the image is summed up by this formulation: “a square grid of short white line segments of uniform length, randomly oriented, on a field of black”. Never mind about the image, this is what the phenomenon contains. Would you not agree?
This is a minor instance of what Husserl, the founder of modern phenomenology, called “bracketing”. Bracketing is a matter of focusing on what is present in a phenomenon, without consideration of the external object that produced it, nor even the assumption that such an object exists. The bracketed phenomenon could be any mental object — a coffee cup within view, a memory of a waterfall, something encountered in a dream. The thesis is that one can derive certain (apodictic) knowledge about phenomena from this disciplined variety of introspection. Husserl’s slogan was “back to the things themselves” (meaning the phenomena). When I ask myself what the phenomenon of the image contains, no doubt arises about the answer — I see a grid of line segments, and I discern no pattern in their orientations (that is what I mean by “random” in this context). No error is possible, because the question is about what is present directly to my awareness, not about anything external.
I’ll offer another demonstration that the orientations of individual segments are missing from their phenomena:
Can you spot the differences between left (which is identical to image 1), and right? There are two dozen differences. At the bottom of this note, there appears an identical image pair, except that the segments which differ are displayed in magenta.
Since a list of orientations of segments is missing from the phenomenon, simple patterns by which the orientations are correlated are sometimes missed too, as is the case for the diamond. Consider this variant:
The variant is identical to the first image except that the segments in the pattern have been rotated by 90 degrees, so that the segments making up each of the four lines of the diamond are co-linear. With this change, the pattern becomes visible — it enters into the phenomenon.
I can easily come up with a description of the content of the new phenomenon. The description runs: “ it is a square grid of short white line segments of uniform length on a field of black, randomly oriented, except those which make up a diamond with tips at the sides of the grid, which segments are co-linear along the diamond’s sides”. Again this description fully captures the content of the phenomenon (well, a few details are missing, notably the length and width of the segments compared to the grid cells, and the number of rows and columns in the grid). How the phenomenon itself, a mental object, is constituted, and how it is embedded in the physical world (if it is) via patterns of neural activity, is a separate, and unanswered, question.
This is not to say that the phenomenon is the description, only that the description formulates its content. The process by which we are able to produce the description is not available to introspection, only the apodictic knowledge that the description is correct and complete
Note that descriptions are characterized by their level of abstraction and lack of detail compared to the image itself. For any given description, countless images match it. We can say that the phenomenon has the nature of a description, in that it too is abstract (built of higher level concepts than pixels), and lacks nearly all detail. In mathematical terminology it is a predicate on images, not a copy of any particular image. Indeed, the descriptions that we have encountered could be precisely and easily formulated using geometry and a little mathematical logic, by anyone familiar with those fields.
The phenomenon does not feel like a description, however. It feels as if we see the image “as it is”. It seems that every little segment and its orientation is present to us. This, it seems, is an illusion built into visual awareness. How can we feel that we have direct perceptual contact with the external image, when its phenomenon is so distant in form and content from the image?
Note that the presence of a pattern in an image, such as the diamond, does not imply that it is present to awareness. We are aware of a pattern only if it shows up in the phenomenon, and thereby, in the description. We should think of patterns as we experience them as structures internal to ourselves. This is another sense in which the phenomenon fails to contact the image.
When do the descriptions arise ? In these simple cases, I place the task before me: “describe the image”, and the description flows easily, and captures the content of the phenomenon without remainder. But, in normal activity, there is only occasionally a call for a description of the things we see. Without such a call, no description is present to the mind: instead the illusion of seeing the image as it is takes the reins of awareness. A phenomenon is my seeing of the image, and the description occurs within the phenomenon, in an unknown coding (if that is the right phrase). So it is fair to say that the description normally lies within my seeing, rather than being something that I see.
As images get more complex, the possibility of writing down language which fully captures their phenomena withdraws. Consider the following image:
Clearly , it is not enough to characterize this as “randomly oriented green and yellow short segments placed randomly within a square”, although this description is accurate as far as it goes. The segments form swirls — nearby segments are similarly oriented, and their orientation varies smoothly across the image. Exact characterization the phenomenon is out of the reach of introspection. However, we can still say something about what is present in the phenomenon. A comparison of two images helps.
What differences do you see between the two at first sight? This is what I pick up: the images appear to be identical, except that the swirls in the lower right are a bit different. It took a moment to notice the latter fact, as my eyes swept over that part of the image. On close inspection, when you turn your attention to individual pairs of segments at corresponding positions, one on the left and one on the right, you will note that they are often of different colors and lengths, though their orientations are identical (except in the lower right). So, we see that, as in our earlier example (image 1), the individual segments are not present in the phenomenon prior to consciously directed close inspection. Thus we can already see that the gulf between image and phenomenon is wide.
The study of what visual phenomena (a.k.a. “percepts”) contain, and how they are built up, has been studied intensively for more than a hundred years, first by the Gestalt psychologists, then by cognitive scientists, neuroscientists, and computer scientists. Of particular relevance here is the work of Bela Julesz on texture discrimination. He showed that any texture (an image or portion thereof built of a large number of elements placed according statistical principles), can be ascribed a mathematical characterization (based on “textons”) such that, that characterization is the same for two images, or two portions of an image, if and only if the textures are indistinguishable by human vision (except by conscious investigation, not immediate impression). That is, textons capture the same information about a patch of texture as does human vision. A texton map of a textured image constitutes a variety of description, although it is formulated mathematically rather than with ordinary language. Thus, we see that even though a description of a visual phenomenon may not be available to introspection, nor even amenable to formulation in ordinary language, such a description can sometimes be uncovered and validated by scientific work.
Again, I would not say that phenomena are their descriptions, only that content of phenomena are characterized by their descriptions, and that the descriptions are sometimes accessible from the phenomena via conscious processes. They can also probed by experiment in more complex cases. The descriptions do not appear within the original phenomenon, rather they are a part of the mechanism from which appearance emerges.
Here is a variant of the image pair above, with differences in magenta.
