my first dead vole: synchrony, structure, snow.

by Jacob Lee

I want to first thank Carl, John, and Asher for giving me the opportunity to join Dead Voles as its fourth member! I am looking forward to a future of exciting exchanges and world domination! I have been a little shy about starting- a funny feeling in someone with a relatively visible amateur online presence. What should I write about? What if everyone thinks its boring? What if its a just a dead vole? Oh wait

Lately I have been working on a little problem. I wanted some way to say that two configurations in some space are structurally identical but in a way which abstracts from any particular way that space happens to be structured. I do not say independent of how that space is structured, since the underlying structure of a space is what gives any configuration of locations its form.*

Anyway, while I was pondering these things, and doing a little background googling, I came across something completely different, the fascinating work by Sang-Hun Lee  and Randolph Blake on how the perception of spatial structure can be induced by temporal synchrony.

Picture an old black and white analog television filled with static, what is often described as ‘snow’. Each pixel on the screen changes its luminescence by some random amount at random intervals. It is unstructured chaos. Such a sequence cannot be compressed without loss, because it does not have any regular structures.

Temporal Synchrony and Random Luminescence

Temporal Synchrony and Random Luminescence Source:

Imagine  synchronizing a block of pixels’ random changes in luminescence, and voila! you see the block of pixels as a distinct visual object! It must be emphasized that the synchronized changes happen at random intervals, and each pixel’s change in luminescence in the synchronized block was individually random. You can see an animation here.

Lee and Blake do a bit more with this than I have described. actually did was a little more complicated and involved. Check it out (PDF).

It reminds me of the fact that staring at that old television set filled with static, if you try you can see any shape or scene that you desire. Perhaps the brain is selectively picking up random synchronizations in the visual field. Lookie lookie! Its a couple dancing! Lookie lookie! It’s a coyote chasing a rabbit! Lookie lookie, its Jacob’s first dead vole!

*We might specify that two spatial configurations are equivalent if there is some kind of structure preserving map between them: in particular, you define a set of invariant transformations of the space. Which of these aspects need to be invariant depends on the space at hand. For example, a linear transformation of a geometric figure, like a triangle, in a Cartesian plane preserves shape, size, orientation. Jerry Seligman uses a similar approach to derive the equivalence class of situations in his paper Physical Situations and Information Flow.


Seligman, Jerry. 1991. Physical situations and information flow. In Situation theory and its applications, ed. Jon Barwise, Jean Mark Gawron, Gordon Plotkin, and Syun Tutiya, 2:257-292. CSLI Lecture Notes 26. Stanford, CA, USA: Center for the Study of Language and Information (CSLI).


8 Comments to “my first dead vole: synchrony, structure, snow.”

  1. Intriguing. It will take a few days before I can say much more, since I am still stuck in this madcap adventure in Italy where the men’s chorus I belong to has sung at Milan Cathedral, done a reduced version of Saegusa Shigeaki’s requiem broken into parts interspersed in a high mass performed at the Vatican, had my pocket picked in the Roman subways, and been overwhelmed by the Vatican Museum of which the Sistine Chapel is only a fragment. Today it is off to Amalfi, Capri and Naples in the company of a Japanese _kuishinbo_ (glutton/gourmand), a former banker/rugby player whose approach to travel is reminiscent of Calvin Trillin, I.e. The constant question is what and where to eat next. I expect recovery to take a while.

  2. Cool. OK, so how then do we tell the difference between the perceptions that are basically imaginative projections on chaos (Jacob’s first dead vole on the staticky tv) and the ones that are actually ‘out there’, albeit conditioned by the same sorts of physical structures (Jacob’s first dead vole right now on my laptop screen)?

  3. John, in my view Trillin and your kuishinbo are correct, but it may be necessary to visit a couple of non-culinary sites between meals to speed digestion and quiet the ghosts of Puritans past. I’m also a big fan of sitting and people-watching, preferably with a capuccino or carafe of local wine.

  4. Interesting question. I can’t talk about the visual system itself, but suppose that I am looking for the pattern 1101 in the string : 10011010001011001111100101010010110. The pattern can be correctly matched just once. But I can be a little more truthy about it. I can ignore some bits as mere inconveniences and cherry pick the bits that construct the pattern I’m looking for. For example, I can construct 1101 by pulling out the 1st, 3rd, 5th, and 6th bits (from the left). But in so doing, I am ignori…, uh deconstructing the underlying logic of the string-space. The antidote to this sort of thing begins with actively seeking out disconfirming evidence, and maintaining a skeptical attitude about the veracity of one’s beliefs and perceptions. A favorite bumpersticker of mine is: “Don’t believe everything you think.” It is a hard habit to acquire.

  5. The paper to which Jacob points us begins with the following paragraph:

    For objects to be seen, they must stand out from their backgrounds. The role of spatial structure in scene segmentation has been exhaustively studied this century, beginning with the landmark work of the Gestalt psy- chologists. From that work, we know a great deal about how figure/ground segmentation is influenced by spatial factors including prox- imity, continuity, and symmetry (1). At the same time, we know that vision is dynamic: Objects can move about the visual environment unpredictably, and as observers we are chronically moving our eyes and heads to sample that environment. Consequently, the optical input to vision also contains rich, complex temporal structure. Can the visual system exploit that temporal structure to segregate objects from their backgrounds?

    From an evolutionary perspective, the answer to the question is yes, of course, how could it be otherwise? Vision is adaptive precisely because it enables mobile animals to detect changes in the visual field and respond to possible dangers or prey. How is, of course, the serious question and the one on which this fascinating article is, I believe, properly focused. Simultaneous, parallel movements suggest the presence of structure lurking behind apparent chaos; the shadows move, we see stripes, we infer the presence of the tiger.

    But is the tiger really there? Evolutionary biology suggests that whether the tiger present or not, it is better to assume the tiger’s presence and freeze, flee or prepare to fight. Stopping to ponder will, if the tiger is present, reduce the chances of survival. Thus it is that training is required to improve decisions and, in the best of all possible worlds, scholars lead protected lives that allow them to pause and ponder without immediate exposure to danger. Thus it is that overcoming prejudice is hard.

  6. “scholars lead protected lives that allow them to pause and ponder without immediate exposure to danger”

    And because the systems are programmed for threat detection and avoidance, scholars then must invent dangers of their own to maintain homeostasis, haha… loved this point and my new post riffs a bit on another dimension of this.

  7. “The antidote to this sort of thing begins with actively seeking out disconfirming evidence, and maintaining a skeptical attitude about the veracity of one’s beliefs and perceptions.”

    I completely agree. The problem is we do our seeking, our perceiving, our believing and our doubting all with the same brain, so getting twisted into the pretzel that productively doubts itself without slipping into paralysis and/or psychosis is in principle not a trivial exercise. In practice, we have each other to check with (although delusions are communicable) and the helpful influence of the brute facts our evolutionary history has prepared us to cope with realistically on pain of extinction. But even there, we’ve civilized ourselves well past the point of immediate remediating consequence for a lot of pretty important stuff, e.g. the climate.

  8. OK, at last engaging with the material itself rather than bs-ing by going meta,

    “Can the visual system exploit that temporal structure to segregate objects from their backgrounds?”

    Some thirty years ago I went to an optometrist complaining that my asymmetric near-sightedness was messing with my depth perception for hitting a tennis ball. He told me in effect to stop whining because depth perception is not so much a binocular function as a dynamic calculation of movement against background. So I guess his answer to this question would also have been yes, of course; and how is indeed the further critical question.

Leave a Reply!

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: