On Video Feedback - I Am a Strange Loop - Douglas R. Hofstadter

I Am a Strange Loop - Douglas R. Hofstadter (2007)

Chapter 5. On Video Feedback


Two Video Voyages, Three Decades Apart

THE loop of video feedback is rich, as I found out in my first explorations with our family’s new video camera in the mid-1970s. A few months later, my appreciation of the phenomenon deepened considerably when I decided to explore it in detail as a visual study for my book Gödel, Escher, Bach. I made an appointment at the Stanford University television studios, and upon arriving I found that the very friendly fellow there had already set up a TV and a camera on a tripod for me to play around with. It was a piece of cake to point the camera at the screen, zoom in and out, tilt the camera, change angles, regulate brightness and contrast, and so on. He told me I was free to use the system as long as I wanted, and so I spent several hours that afternoon navigating around in the ocean of “taboo” possibilities opened up by this video loop. Like any curious tourist, I snapped dozens of photos (just black-and-white stills) during my exotic trip, and later I selected twelve of my favorites to use in one of GEB’s dialogues.

Since that first adventure in video feedback, three decades have passed and technology has advanced a bit, so for my new book I decided to give it another shot. This time I was aided and abetted by Bill Frucht, who, because of (or in spite of) being my editor at Basic Books for a dozen years or so, has become a good friend, and who flew in from New York just for this purpose. Together in my kids’ old “playroom”, Bill and I spent many delightful hours sailing the same old seas but in a somewhat newer craft, and we wound up with several hundred color snapshots that archived our voyage superbly. Aside from the cover illustration, sixteen of my favorites, covering a wide range, can be found in the color insert.

Although both video voyages were vivid and variegated, I decided for this chapter to write up a “diary” of the earlier one, undertaken long ago at Stanford, since that’s when I first explored the phenomenon and learned about it step by step. So the story below involves a different television, a different TV camera, and in general an older technology than was used in making this book’s color insert. Nonetheless, as you will see, much of the old diary still pertains to the newer voyage, though there are a few small discrepancies that I’ll mention when I come to them.

Diary of a Video Trip

There happened to be a shiny metallic strip running down the right side of the TV set I was given, and the presence of this random object had the fortuitous effect of making the various layers of screens-within-screens easily distinguishable. The first thing I discovered, then, was that there was a critical angle that determined whether the regress of nested screens was finite or infinite. If I pointed the camera at the metal strip instead of the center of the screen, this gave me what looked like a snapshot of the right wall of a long corridor, showing a few evenly spaced “doorways” (which actually were images of that metal strip), moving away from where I was “standing”. But I was not able to peer all the way down to the end of this “corridor”. I’ll therefore call what was visible on the screen in such a case a truncated corridor.

If I slowly panned the camera leftwards, thus towards the center of the screen and perforce further down the apparent corridor, more and more doorways would come into view along the right wall, smaller and smaller and farther and farther away — and all of a sudden, at a critical moment, there was a wonderful, dizzying sense of infinity as I would find myself peering all the way down the corridor toward a gaping emptiness, stretching arbitrarily far away toward a single point of convergence (the “vanishing point”, as it is called in the theory of perspective). I’ll call this an endless corridor. (Note that essentially this same kind of corridor is also visible in the photo of the self-reflecting mirrors in Chapter 4.)

Of course my impression of seeing an infinite number of doorways was illusory, since the graininess of the TV screen and the speed of light set a limit as to how many nestings could occur. Nevertheless, peering down what looked like a magically endless corridor was much more enticing and provocative than merely peering down a mundanely truncated corridor.

My next set of experiments involved tilting the camera. When I did this, each screen obediently tilted at exactly the same angle with respect to its containing screen, which instantly gave rise to a receding helical corridor — a corridor that twisted like a corkscrew. Though quite attractive to the eye, this was not terribly surprising to the mind.

An unanticipated surprise, however, was that at certain angles of camera twist, instead of peering down a helical corridor punctuated by doorways, I seemed to be looking at a flat spiral resembling a galaxy as seen through a telescope. The edges of this spiral were smooth, continuous curves of light rather than jagged sets of straight lines (coming from the edges of the TV screen), and such smoothness mystified me; I saw no reason why a sudden jump from jagged corners to graceful curves should take place. I also noticed that at the very core of each “galaxy”, there was nearly always a beautiful circular “black hole”. (On our more recent video voyage, Bill and I were unable to reproduce this “black hole” phenomenon, to our puzzlement and chagrin, so you won’t see any black holes in the photos in the insert.)

Enigmatic, Emergent Reverberation

At some point during the session, I accidentally stuck my hand momentarily in front of the camera’s lens. Of course the screen went all dark, but when I removed my hand, the previous pattern did not just pop right back onto the screen, as I expected. Instead, I saw a different pattern on the screen, but this pattern, unlike anything I’d seen before, was not stationary. Instead, it was throbbing, like a heart! Its “pulse rate” was about one cycle per second, and over the course of each short “heartbeat”, the shapes before my eyes metamorphosed greatly. Where, then, had this mysterious periodic pulsation come from, given that there was nothing in the room that was moving?

Whoops — I’m sorry! What I just wrote is a patent falsity — there was something in the room that was moving. Do you know what it was, dear reader? Well, the image itself was moving. Now that may strike you as a fatuous, trivial, or smart-alecky answer, but since the image was of itself (albeit at a slight delay), it is in fact quite to the point. A faithful image of something changing will itself necessarily keep changing! In this case, motion begat motion endlessly because I was dealing with a cyclic setup — a loop. And the original motion that had set things going — the prime mover — had been my hand’s motion, of which this video reverberation now constituted a stable, self-sustaining visible memory trace!

This situation reminds me of another loopy phenomenon that I call “reverberant barking”, which one sometimes can hear in a neighborhood where many dogs live. If a jogger passes one house and triggers one dog’s bark, then neighbor dogs may pick up the barking and a chain reaction involving a dozen dogs may ensue. Soon the barking party has taken on a life of its own, and in the meantime its unwitting instigator has long since exited the neighborhood. If dogs were a bit more like robots and didn’t eventually grow tired of doing the same thing over and over again, their reverberant barking could become a stable, self-sustaining audible memory trace of the jogger’s fleeting passage through their street.

The dynamically pulsating patterns that I encountered in my video voyage were completely unlike the unwavering “steady-state universes” that I had observed up till then. Stable, periodic video reverberation was a strange and unanticipated phenomenon that I’d bumped into by accident while exploring the possibilities lurking in video feedback.

Even today, all these years later, the origins of such pulsation remain quite unclear, even mysterious, to me; for that reason, it is an emergent phenomenon, otherwise known as an epiphenomenon, as discussed in Chapter 3. In general, an emergent phenomenon somehow emerges quite naturally and automatically from rigid rules operating at a lower, more basic level, but exactly how that emergence happens is not at all clear to the observer.

I admit to feeling a little dense for not having fully fathomed what lies behind video reverberation, but at this point I am so accustomed to it that it “makes sense” to me. That is, I have a clear intuition for how to induce it on the screen, and I know that once it starts, it is a robust phenomenon that will continue unabated probably for hours, perhaps even forever, if I don’t interfere with it. Rather than trying to figure out how to account precisely for video reverberation in terms of phenomena at lower levels, I have come to just accept it as a fact, and I deal with it at as a phenomenon that exists at its own level. This should sound familiar to you, since it’s how we deal with almost everything in our physical and biological world.

Feeding “Content” to the Loop

As I mentioned at the outset, one lucky thing about the Stanford setup was the seemingly random metallic strip on one side of the television set I’d been given to use. That strip — a kind of interloper — added a key note of “spice” to the image that was being cycled round and round, and in that sense it was a crucial ingredient of Video Voyage I.

While Bill and I were conducting Video Voyage II, there were times, to our surprise, when the seas we were sailing seemed a bit too placid for our taste, and we longed for a bit more action, more visual excitement. This brought to my mind the crucial “spicy” role played by the interloping metal strip during Voyage I, so on a lark we decided to introduce something that would play an analogous role in our system. I picked up various objects around the room and dangled them in front of the camera without any idea of what would happen when the image was cycled round and round the video loop. Usually we got marvelous results that were (once again) unanticipatable. For instance, when I dangled a chain of beads in front of the screen, what emerged (the choice of verb is not accidental) was a random-looking swirl of pockmarked bluish-white globs that reminded me a bit of some kind of exotic cheese.

Of course each such interloping object opened up a whole new universe of possibilities, since we could vary its position as well as all the other standard variables (the amount of zoom, the angle of tilt, the direction of the camera, the brightness, the contrast, and others). I tried such things as a glass vase, a compact disk, and, eventually, my own hands. The results were quite fantastic, as you can see in the color insert, but alas, Bill and I didn’t have infinite amounts of time to explore the manifold universes we had uncovered and sampled. We played with the possibilities for perhaps a dozen hours and from that we got a 400-photo memory album, and that’s all. Like any excursion to a wondrous and exotic place, our trip had to end earlier than we would have preferred, but we were very glad to have taken it and to have savored it together.

A Mathematical Analogue

As might be expected, all the unexpected phenomena that I observed depended on the nesting of screens being (theoretically) infinite — that is, on the apparent corridor being endless, not truncated. This was the case because the most unpredictable of the visual phenomena always seemed to happen right in the vicinity of that central point where the infinite regress converges down to a magical dot.

My explorations did not teach me that any shape whatsoever can arise as a result of video feedback, but they did show me that I had entered a far richer universe of possibilities than I had expected. Today, this visual richness reminds me of the amazing visual universe discovered around 1980 by mathematician Benoit Mandelbrot when he studied the properties of the simple iteration defined by zz2 + c, where c is a fixed complex number and z is a variable complex number whose initial value is 0. This is a mathematical feedback loop where one value of z goes in and a new value comes out, ready to be fed back in again, just as in audio or video feedback. The key question is this: If you, playing the role of microphone and loudspeaker (or camera and TV), do this over and over again, will the z values you get grow unboundedly, sailing off into the wild blue (or wild yellow or wild red) yonder, or will they instead home in on a finite value?

The details need not concern us here; the basic point is that the answer to the question depends in a very subtle way on the value of the parameter c, and if you make a map by color-coding different values of caccording to the rate of z’s divergence, you get amazing pictures. (This is why I joked about the “wild yellow” and “wild red” yonders.) Both in video feedback and in this mathematical system, a very simple looping process gives rise to a family of truly unanticipated and incredibly intricate swirling patterns.

The Phenomenon of “Locking-in”

The mysterious and strangely robust phenomena that emerge out of looping processes such as video feedback will serve from here on out as one of the main metaphors in this book, as I broach the central questions of consciousness and self.

From my video voyages I have gained a sense of the immense richness of the phenomenon of video feedback. More specifically, I have learned that very often, wonderfully complex structures and patterns come to exist on the screen whose origins are, to human viewers, utterly opaque. I have been struck by the fact that it is the circularity — the loopiness — of the system that brings these patterns into existence and makes them persist. Once a pattern is on the screen, then all that is needed to justify its staying up there is George Mallory’s classic quip about why he felt compelled to scale Mount Everest: “Because it’s there!” When loops are involved, circular justifications are the name of the game.

To put it another way, feedback gives rise to a new kind of abstract phenomenon that can be called “locking-in”. From just the barest hint (the very first image sent to the TV screen in the first tiny fraction of a second) comes, almost instantly (after perhaps twenty or thirty iterations), the full realization of all the implications of this hint — and this new higher-level structure, this emergent pattern on the screen, this epiphenomenon, is then “locked in”, thanks to the loop. It will not go away because it is forever refreshing itself, feeding on itself, giving rebirth to itself. Otherwise put, the emergent output pattern is a self-stabilizing structure whose origins, despite the simplicity of the feedback loop itself, are nearly impenetrable because the loop is cycled through so many times.

Emergent New Realities of Video Feedback

Coming up with vivid and helpful nicknames for unexpected visual patterns had certainly not figured in my initial plans for my video voyage at Stanford, but this little game soon became necessary. At the outset, I had thought I was undertaking a project that would involve straightforward terms like “screen inside screen”, “silver strip”, “angle of tilt”, “zooming in”, and so forth — but soon I found myself forced, willy-nilly, to use completely unexpected descriptive terms for what I was observing. As you have seen, I started talking about “corridors” and “walls”, “doorways” and “galaxies”, “spirals” and “black holes”, “hubs” and “spokes”, “petals” and “pulsations”, and so forth. In the second video voyage with Bill, many of these same terms were once again needed, and some new ones were called for, such as “starfish”, “cheese”, “fire”, “foam”, and others.

Such words are hardly the kind of language I had thought I would be dealing with when I first broached the idea of video feedback. Although the system to which I was applying these terms was mechanical and deterministic, the patterns that emerged as a consequence of the loop were unpredictable, and therefore it turned out that words were needed that no one could have predicted in advance.

Simple but evocative metaphors like “corridor”, “galaxy”, and others turned out to be indispensable in describing the abstract shapes and events I witnessed on the screen. The initial terms I had tacitly assumed I would use wound up getting mostly ignored, because they yielded little insight. Of course, in principle, everything could be explained in terms of them, in a rigorous and incomprehensibly verbose fashion (like explaining a gas’s temperature and pressure by writing out Avogadro’s number of equations) — but such a boringly reductionistic, nearly pixel-by-pixel explanation would entirely leave out the wonderful higher-level visual phenomena to which a human eye and mind intuitively resonate.

In short, there are surprising new structures that looping gives rise to that constitute a new level of reality that could in principle be deduced from the basic loop and its detailed properties, but that in practicehave a different kind of “life of their own” and that demand — at least when it comes to extremely finite, simplicity-seeking, pattern-loving creatures like us — a new vocabulary and a new level of description that transcend the basic level out of which they emerge.