Observation and Inference (Where the Bits Go)

See these amazing images of individual bits recorded on a hard disk, obtained using a technique known as Magnetic Force Microscopy (MFM). As I understand it, MFM is not a photographic process but rather a form of visualization: magnetic sensors at the tip of the measuring instrument are able to detect the tiny fluctuations in the magnetic field on the surface of the disk, and thereby generate the images you see here.

But what are we really looking at? Douglas R. Hofstadter (best known for Gödel, Escher, Bach), has the following to say in his Fluid Concepts and Creative Analogies: Computer Models of the Fundamental Mechanisms of Thought:

Today, for instance, ultrasound allows us to see a fetus moving about inside a mother’s womb in real time. Note that we feel no need to put quotes around the word “see” — no more than around the word “talk” in the sentence “My wife and I talk everyday on the phone.” When we make such casual statements we don’t for a moment consider the weirdness of the fact that our voices are speeding in perfect silence through metallic wires; the reconstruction of sounds is so flawless and faithful that we are able to entirely forget the fact that complex coding and decoding processes are taking place in between the speaking mouth and the listening ear. [. . .] If, fifty years ago, high-frequency sounds had been scattered off a fetus, there would have been no technology to convert the scattered waves into a vivid television image, and any conclusions derived from measurements on the scattered waves would have been considered abstruse mathematical inferences; today, however, simply because computer hardware can reconstruct the scatterer from the scattered waves in real time, we feel we are directly observing the fetus. Examples like this — and they are legion in our technological era — show why any boundary between “direct observation” and “inference” is a subjective matter. (488)

Hofstadter goes on to note that, “much of science consists in blurring this seemingly sharp distinction” (488). As much or more than any of the scores of better-known prophets and pundits of the Information Age, Hofstadter has succeeded in putting his finger on one of the central dynamics of our times: that “information,” which was once explicitly defined by computer scientists as a quality independent or indeed exempt from meaning has now, as a direct consequence of advances in computing technology, become meaningful in and of itself. By this I mean not that data is useful or intelligible without context and structure, but rather that the continuum involved in the creation of meaning through the process of interpreting data now encompasses degrees of abstraction and representational artifice which would have heretofore been considered “meaningful” only after the prior imposition of some second-order procedure or analysis. Or to put it another way, advanced computing technologies, visual and graphical for the most part, now routinely allow information’s artificial or referential dimension to function as a kind of permeable membrane through which a dynamic of inference and direct observation can operate. Moreover, as Hofstadter rightly notes, examples of the oscillating dynamic between inference and direct observation are “legion” at the present moment—indeed, they define the most pedestrian behaviors of the wired lifestyle. When I first wrote this, for example, I was tracking rain cells over central Virginia using online Doppler radar available from the Weather Channel’s site (I had a chronically leaky window seam and was wondering whether I’d need to get up to arrange the pots and pans I use to catch the run-off).

Precipitation displays on Doppler radar as colored blobs, the color varying from a light green to deep reds and pinks depending on intensity. Note that unlike Hofstadter’s example of the ultrasound, what I see on the Doppler map does not, in any mimetic sense, “look like” rain. Yet not only do I accept the Doppler images as an accurate depiction of the weather conditions in my area (and thus a reliable indication of whether or not I’ll need to mop up my windowsill), I also tend to regard what I am seeing as simply “rain,” and not as “a computer-generated visualization of the prevailing atmospheric conditions.” Though I understand of course that the Doppler display is precisely that, the truth is that as a practical matter my phenomenological apprehension of rain has expanded to include images produced by radar waves reflected from bands of precipitation within what is essentially the same referential horizon as the puddles that I know will eventually appear on my windowsill.