August 28th, 2011
The Time of Pattern Recognition
When is it that the pattern is recognized? Was it at that first moment, the moment when the first element emerged from future possibility into present possibility? How might I recognize this element all on its own—without the links that make it part of the larger pattern?
Perhaps it’s the pattern that must first make its impression, such that the newly appearing element has a sensible place to stand. Having the pattern in mind, I wait for the last element to find its place and complete the pattern in its wholeness.
But having seen a pattern only once, I can’t yet say that it’s a pattern. It’s only having seen the pattern at least a second time, that I can look back in retrospect and say, this first instance was the earliest example of the pattern. That’s where it all started.
If we’re looking for the moment the pattern—as pattern, emerges, it’s never with its first appearance, but at a minimum with its second. A third appearance might supply some needed confirmation, a signal that it’s really a pattern and not merely a set of twins.
The time of pattern recognition seems to be backward looking, out toward the horizon of memory. These floating historical elements are gathered up and crystalized into a pattern, a new object for the present moment. And, of course, the pattern itself may become a part of another pattern, and so on.
Once we have the pattern in hand, can we project a future time of patterns? Could a single new event trigger the recognition of a pattern? To create certainty, the event would have to travel with an attached message that said, “save me, I’m always part of this pattern you’re interested in. I have a purpose (telos) that may not be apparent by just looking at me, but this message you’re reading vouches for my higher purpose. I am a part of a significant pattern. Recognize me.” What do we do to the thing when we pre-pattern its existence? In some ways, isn’t this the only way we can possibly recognize anything? A thing that wasn’t part of a pre-existing pattern might simply appear as noise to us.
Rather than demanding certainty, we might assign probabilities. A newly arrived element might have a calculated probability that it belongs to a certain pattern. We might provisionally treat it as though it does, until sufficient evidence accumulates. When the confirming evidence presents itself, we bring out the rubber stamp and certify that it’s a member of some particular pattern. Or perhaps we determine that it’s actually a member of a different pattern, or no recognizable pattern, and so we treat it accordingly.
As we think of the time of the pattern, we also might consider the time of the element. Is the element, once lodged firmly into a pattern, permanently defined by the pattern? Does the pattern exhaust all of the possibility of the element? Could the element change in such a way that it was no longer part of a particular pattern that had claimed it? Is a pattern a fixed constellation, or are the elements brimming with energy and possibility? Could they, at any moment, break off and find another pattern of which to be a part? Could the pattern itself suddenly change its requirements, excluding some heretofore members in good standing, and including others formerly considered outsiders?
We’ve been thinking of patterns as something a human recognizes in the stream of events surrounding it. What happens when the work of recognition is displaced to a machine built to recognize patterns and then take certain actions upon their identification? I might dream up a list of patterns and stuff them in the top of the machine, and then tell the machine very specifically what I’d like to have happen each time a pattern is recognized. The machine automatically churns through large quantities of material and digs up elements that fit into one of the specified patterns.
Imagine that we tell the machine to simply observe the flow of events around us and to detect emergent patterns. In this example, the machine isn’t working with patterns we consciously select, but instead with patterns we actually enact. Certainly this would provide us with a more real set of patterns, and it would save us the trouble of dreaming up patterns and feeding them into the machine. The patterns and their recognition would be entirely automated. This would allow anyone owning such a machine to simply turn it on and let the benefits of automatic pattern recognition accumulate over time.
One can image additional modules for the machine. There may be patterns I enact that I have no awareness of. Some of these patterns may be having a negative effect on my overall well being. A special sub-system that identified these patterns and integrated them back into my conscious awareness might be called psychiatric plugin. Or perhaps, I’m enacting a pattern that could be used to identify me as a target for certain kinds of advertising offers. The cost of the machine could be subsidized by auctioning these pattern matches to the highest bidder. There might be a module that pays me when I enact a certain set of patterns. Of course, the machine couldn’t reveal the substance of the patterns to me as this might encourage me to pretend to enact rather than really enact. We might call this a Skinner-box module.
If there’s an economics to information flow, it’s based on the production and consumption of patterns of bits. It might not even matter what the pattern consists of, if the cost of the transaction wrapper is sufficiently small, any pattern can serve as an economic vehicle. And once this has occurred, the value of the pattern is separated from its economy. All patterns, regardless of value, can have an economy in this model.
Philip Roth, writing some time ago about the state of literature behind the Iron Curtain, noted that when nothing is allowed, everything becomes important. And conversely, when everything is allowed, nothing is important. Having established that you can buy or sell anything, we find ourselves standing around without a measuring stick, asking whether it’s any good or not.