Andrew Orlowski writes:
“A new study conducted at Cornell University suggests that we think in analog, not digital. It’s a bold claim which, if true, threatens to make thirty years of linguistics and neuroscience metaphors look very silly indeed.”
The fact that our cats can calculate the required muscle flex and velocity to leap onto a table with food, but don’t understand the meaning of ‘no’ and can’t do simple arithmetic has always puzzled me.
However it makes sense that any system based upon learned statistical reaction to sensory input would create sophisticated responses without understanding them. In this instance ability to extract logical rules would be based upon a enormous amount of analog input that produced binary certainty as an emergent phenomenon.
The Reverse Turing test or ‘captcha’, usually contains a noise filled image of a password with warped fonts is used to filter humans from computers, to stop spam. In this instance a very simple digital message is encoded in a very analog way.
As captchas have become more sophisticated, in an arms race against algorithms designed to crack them, humans often make errors reading them. The fact that these errors become pronounced may indicate that we use statistical responses to patterns rather than an internal algorithm for reading captcha’s – this would tend to indicate that our brains work in analog mode.
In other words, if we had an algorithm for reading captchas then it would be likely that we would either be entirely able or unable to decode any captcha created with the same algorithm, there would not be a statistical chance of failure.
The interesting thing about a captcha image is that it does not compress very well, while retaining fidelity – in other words the ratio of bits required for the analog message to the equivalent digital message is very high.
Current SETI searches assume that a signal would not have deliberate noise – that it would be separated from natural phenomena by its ‘artificial’ simplicity. One of the problems with this is that we have to look in areas that are naturally quiet, it is difficult to differentiate between a signal caused by a known or unknown natural phenomenon such as a quasar pulse or WOW signal and a deliberate message.
But what if someone were to try a Reverse Turing Test on us?