To be sure, the most primitive living creatures have little or no self-perception. By analogy, we can think of a TV camera rigidly bolted on top of a TV set and facing away from the screen, like a flashlight tightly attached to a miner's helmet, always pointing away from the miner's eyes, never into them. In such a TV setup, obviously, a self-turned loop is out of the question. No matter how you turn it, the camera and the TV set turn in synchrony, preventing the closing of a loop.

We next imagine a more "evolved", hence more flexible, setup; this time the camera, rather than being bolted onto its TV set, is attached to it by a "short leash". Here, depending on the length and flexibility of the cord, it may be possible for the camera to twist around sufficiently to capture at least part of the TV screen in its viewfinder, giving rise to a truncated corridor. The biological counterpart to feedback of this level of sophistication may be the way our pet animals or even young children are slightly self-aware.

The next stage, obviously, is where the "leash" is sufficiently long and flexible that the video camera can point straight at the center of the screen. This will allow an endless corridor, which is far richer than a truncated one. Even so, the possibility of closing the self-watching loop does not pin down the system's richness, because there still are many options open. Can the camera tilt or not, and if so, by how much? Can it zoom in or out? Is its image in colour, or just black and white? Can brightness and contrast be tweaked? What degree of resolution does the image have? What percentage of time is spent in self-observation as opposed to observation of the environment? Is there some way for the video camera itself to appear on the screen? And on and on. There are still many parameters to play with, so the potential loop has many open dimensions of sophistication.

...

Despite the richness afforded by all these options, a self-watching television system will always lack one crucial aspect: the capacity of perception, as opposed to mere reception, or image-receiving. Perception takes [reception] as its starting point, but then it goes much further, eventually winding up in the selective triggering of a small subset of a large repertoire of dormant symbols - discrete structures that have a representational quality.. The missing ingredient in a video system, no matter how high its visual fidelity, is a repertoire of symbols that can be selectively triggered. Only if such a repertoire existed and were accessed could we say that the system was actually perceiving anything.

...

The depth and complexity of human memory is staggeringly rich. Little wonder, then, that when a human being, possessed of such a rich armamentarium of concepts and memories with which to work, turns its attention to itself, as it inevitably must, it produces a self-model that is extraordinarily deep and tangled. That deep and tangled self-model is what "I"-ness is all about.


from "I am a Strange Loop", by Douglas Hofstadter.