The Phenomenology of Abstract Images

6 min readJan 17, 2020

By “phenomenology” I mean the philosophical discipline, founded in its modern form by Edmund Husserl, which studies appearances.

Consider an image, and the situation of a person (say Albert) attending to it. The image itself, as a physical object, might be a photograph, or pixels on a screen. The appearance of the image within Albert’s mind (that is, the phenomenon to which the image gives rise) is not itself an image, but a mental object. This note concerns the ways in which the structures of image and phenomenon differ.

There are two perspectives from which the phenomenon can be considered. The first is Albert’s own. The second is that of an outside observer — for example, a philosopher or neuroscientist trying to make sense of the situation.

When I am considering the matter from the outside, I favor emergentism, a form of non-reductive physicalism which holds that phenomena are structures built out of the activity of neurons, but are not reducible to that activity (see this earlier note for details). But, for the most part, the observations in this note are made from the internal, phenomenological standpoint, and as such do not depend on any particular theory of about what phenomena might be as viewed from the outside.

You yourself know the internal perspective as well as anyone, since you are a human looking at the image too. You see crossing diagonals against an otherwise random field, don’t you?

The pattern of the crossing diagonals is an aspect of Albert’s phenomenon of the image. Consider the following alternative image:

Image 2

This image would seem to be the same as the first one, with the difference that it is missing the crossing diagonals, and consists entirely of a random field. However, if you look closely, you will see that the little lines on the diagonals in image 2 are not random at all. Instead of running along the direction of the diagonals as they did in the first image, they are at right angles to them. The following image was constructed in the same way, but with the diagonal elements colored red.

Let us call the phenomenon of the first crossing-diagonals image P1, and of the second image P2. P1 contains a pattern and P2 does not. We know this because we can ask Albert (or ourselves) what we see — are aware of — when we encounter the images (that is, before the pattern in P2 is explained to us). However, the two images contain patterns to an equal degree.

We can deduce that whatever P1 and P2 are, they are surely not just internal copies of their images. It seems reasonable to say that P1 contains the first image in a certain way, at least when Albert is actually looking at the image, rather than remembering it. It is hard to doubt that those little lines are present to Albert’s awareness in this case. But awareness of the little lines is not enough to ensure awareness of a pattern in them, as is shown by P2. Rather, P1 seems to contain annotations which assign to the lines on the diagonals additional explicit notes that they are aligned along the diagonals to form the crossing-diagonal pattern. These explicit notes are necessary for the awareness of the pattern. No such notes are present in P2 for the equally systematic orientations of the lines on the diagonals.

This at least is my simple phenomenological analysis of this situation — my analysis of what appears to me in the first image. Namely, I am aware of several things: first of a collection of short lines arrayed in a particular way across the image, second of the presence of a particular pattern, which in this case, I can describe easily with the phrase “crossing diagonals”, and finally an assignment of the little lines either to the pattern, or the random field.

This is the view of the situation from the inside. What about the external view — the view, say of the neuroscientist? A long history of work in neuroscience correlates with this picture, in that the coding of simple patterns via the activity of neurons has been uncovered in many contexts. Phenomena, viewed from the outside, might be information structures coded, at some level of abstraction, in neuronal activity (though phenomena in their entirety cannot be accounted for in this way, due to the problem of qualia). The explicit notes, in some situations, might find a particularly simple encoding: the firings of individual neurons. However, such neurological correlations are known only for the simplest of situations. The relationship between phenomena and brain activity remains, as a general matter, mysterious (the study of this falls under the heading “Neural Correlates of Consciousness”). Note that, other than this paragraph, observations made here do not depend on neuroscience.

Presence in awareness of a pattern brings with it causal power — it allows Albert to answer out loud when asked about the content of the image: “I see crossing diagonals”. One could train a dog to press a lever only when the pattern is present, as well. Awareness is constituted by this sort of causal power, according to one theory. Namely, according to the global workspace theory of consciousness, bringing something into awareness amounts to cataloging it in a workspace to which the various mechanisms of the mind have common access. The contents of the workspace are all those things that can be brought to attention, and which can thereby play a role in conscious thought and behavior.

The conclusion is that patterns within an image sometimes show up in phenomena, and sometimes not. Note that such patterns, in the absence of their observation, have absolutely no macroscopic causal power (that is they might affect some photons, but never have any effect at the macroscopic scale). Patterns are macroscopically impotent unless they are observed. They are like George Berkeley’s tree, only more so. At least a tree has interactions with the other trees in the forest, even if there is no observer to behold it. An unobserved pattern in an image, however, is a completely isolated thing in the universe, with no macroscopic effect on anything whatever. It has the lowest possible grade of existence. Note that unheard music has the same low grade of existence as an unobserved visual pattern.

We now know how to build machines which behave as observers in the above sense. Does consciousness of some sort arise in this situation too? I have no idea. How could one know that an entity is conscious without being that entity?

Consider this image in which the patterning is more ornate.

The complexity and nature of the annotations which accompany your phenomenon of this image can only be guessed at. What is the extent of the formal language of patterns (if that is the right way of thinking of it) in which annotations are expressed? It would include, for example, ways of formulating the repetition of visual elements in geometries such as lines, arcs, and grids. Such visual elements would include primitives such as line segments, but also patterns at a lower level, rendering the formulations recursive. But this is only the beginning.

In any case, it is the structure of the annotations that constitute your seeing of the image above, much more so than the pixels. It is inevitable that there are many patterns in the image of which you are unaware. In this sense, you cannot see anything of this complexity in its entirety.

Note that any image of a tree or sidewalk includes patterns of even higher intricacy, only some of which find their way into phenomena. You will never see a tree “as it is”.