Douglas Hofstadter - I Am a Strange Loop

However, Roger Sperry knew very well that he wasn’t embracing dualism or mysticism of any sort, and he therefore had the courage to take the plunge and make the assertion. His position is a subtle balancing act whose insightfulness will, I am convinced, one day be recognized and celebrated, and it will be seen to be analogous to the subtle balancing act of Kurt Gödel, who demonstrated how high-level, emergent, self-referential meanings in a formal mathematical system can have a causal potency just as real as that of the system’s rigid, frozen, low-level rules of inference.

CHAPTER 15

Entwinement

Multiple Strange Loops in One Brain

TWO chapters back, I declared that there was one strange loop in each human cranium, and that this loop constituted our “I”, but I also mentioned that that was just a crude first stab. Indeed, it is a drastic oversimplification. Since we all perceive and represent hundreds of other human beings at vastly differing levels of detail and fidelity inside our cranium, and since the most important facet of all of those human beings is their own sense of self, we inevitably mirror, and thus house, a large number of other strange loops inside our head. But what exactly does it mean to say that each human head is the locus of a multiplicity of “I” ’s?

Well, I don’t know precisely what it means. I wish I did! And I reckon that if I did, I would be the world’s greatest philosopher and psychologist rolled into one. As best I can guess, from far below such a Parnassus, it means we manufacture an enormously stripped-down version of our own strange loop of selfhood and install it at the core of our symbols for other people, letting that initially crude loopy structure change and grow over time. In the case of the people we know best — our spouse, our parents and siblings, our children, our dearest friends — each of these loops grows over the years to be a very rich structure adorned with many thousands of idiosyncratic ingredients, and each one achieves a great deal of autonomy from the stripped-down “vanilla” strange loop that served as its seed.

Content-free Feedback Loops

More light can be cast on this idea of a “vanilla” strange loop through our old metaphor of the audio feedback loop. Suppose a microphone and a loudspeaker have been connected together so that even a very soft noise will cycle around rapidly, growing louder and louder each pass through the loop, until it becomes a huge ear-piercing shriek. But suppose the room is dead silent at the start. In that case, what happens? What happens is that it remains dead silent. The loop is working just fine, but it is receiving zero noise and outputting zero noise, because zero times anything is still zero. When no signal enters a feedback loop, the loop has no perceptible effect; it might as well not even exist. An audio loop on its own does not a screech make. It takes some non-null input to get things off the ground.

Let’s now translate this scenario to the world of video feedback. If one points a TV camera at the middle of a blank screen, and if the camera sees only the screen and none of its frame, then despite its loopiness, all that this setup will produce, whether the camera stands still, tilts, turns, or zooms in and out (always without reaching the screen’s edge), is a fixed white image. As before, the fact that the image results from a closed feedback loop makes no difference, because nothing external is serving as the contents of that loop. I’ll refer to such a content-free feedback loop as a “vanilla” loop, and it’s obvious that two vanilla video loops will be indistinguishable — they are just empty shells with no recognizable traits and no “personal identity”.

If, however, the camera turns far enough left or right, or zooms out far enough to take in something external to the blank screen (even just the tiniest patch of color), a bit of the screen will turn non-blank, and then, instantly, that non-blank patch will get sucked into the video loop and cycled around and around, like a tree limb picked up by a tornado. Soon the screen will be populated with many bits of color forming a complex and self-stabilizing pattern. What gives this non-vanilla loop its recognizable identity is not merely the fact that the image contains itself, but just as crucially, the fact that external items in a particular arrangement are part of the image.

If we bring this metaphor back to the context of human identity, we could say that a “bare” strange loop of selfhood does not give rise to a distinct self — it is just a generic, vanilla shell that requires contact with something else in the world in order to start acquiring a distinctive identity, a distinctive “I”. (For those who enjoy the taboo thrills of non-wellfounded sets — sets that, contra Russell, may contain themselves as members — I might raise the puzzle of two singleton sets, x and y, each of which contains itself, and only itself, as a member. Are x and y identical entities or different entities? Trying to answer the riddle by defining two sets to be identical if and only if they have the same members leads one instantly into an infinite regress, and thus no answer is yielded. I prefer to brazenly cut the Gordian knot by declaring the two sets indistinguishable and hence identical.)

Baby Feedback Loops and Baby “I” ’s

Although I just conjured up the notion of a “vanilla” strange loop in a human brain, I certainly did not mean to suggest that a human baby is already at birth endowed with such a “bare” strange loop of selfhood — that is, a fully-realized, though vanilla, shell of pure, distilled “I”-ness — thanks to the mere fact of having human genes. And far less did I mean to suggest that an unborn human embryo acquires a bare loop of selfhood while still in the womb (let alone at the moment of fertilization!). The realization of human selfhood is not nearly so automatic and genetically predetermined as that would suggest.

The closing of the strange loop of human selfhood is deeply dependent upon the level-changing leap that is perception, which means categorization, and therefore, the richer and more powerful an organism’s categorization equipment is, the more realized and rich will be its self. Conversely, the poorer an organism’s repertoire of categories, the more impoverished will be the self, until in the limit there simply is no self at all.

As I’ve stressed many times, mosquitoes have essentially no symbols, hence essentially no selves. There is no strange loop inside a mosquito’s head. What goes for mosquitoes goes also for human babies, and all the more so for human embryos. It’s just that babies and embryos have a fantastic potential, thanks to their human genes, to become homes for huge symbol-repertoires that will grow and grow for many decades, while mosquitoes have no such potential. Mosquitoes, because of the initial impoverishment and the fixed non-extensibility of their symbol systems, are doomed to soullessness (oh, all right — maybe 0.00000001 hunekers’ worth of consciousness — just a hair above the level of a thermostat).

For better or for worse, we humans are born with only the tiniest hints of what our perceptual systems will metamorphose into as we interact with the world over the course of decades. At birth, our repertoire of categories is so minimal that I would call it nil for all practical purposes. Deprived of symbols to trigger, a baby cannot make sense of what William James evocatively called the “big, blooming, buzzing confusion” of its sensory input. The building-up of a self-symbol is still far in the future for a baby, and so in babies there exists no strange loop of selfhood, or nearly none.