Eulogizing Theory

Posted in The Gnovis Blog

How does one produce truth? This summer I have put the finishing touches on a chapter about anonymity, identity and craigslist Missed Connections. I have now spent about a year studying how people use and abuse this potentially unintended social space (you can see the beginning of this journey here on gnovis), and (if I do say so myself) have some good ideas. In just a matter of days I will have completed my grand theory. But it’s just a theory, and according to Chris Anderson’s recent Wired article (new window), despite my supporting data, it’s just as passe.

In Anderson’s article entitled "The End of Theory" he argues that mathematics and massive amounts of data have eliminated the need to speculate on human behavior. With an unprecedented ability to capture data, theorizing makes no sense when algorithms can predict behavior with numbers. Sounds nice, right? Why theorize about the surfing behavior of netizens when you can just measure it? While this has certainly worked for Google, I am left me wondering if my academic aspirations are now somehow antiquated.

Theory can only get you so far. At the end of my Senior Thesis, having performed all of the correct analysis, I tentatively theorized that perhaps, under certain conditions, the coming out stories of gay men might homogenize over time. But what good is a perhaps-maybe-theory? Even after my endless hours of interviews, transcription, and analysis, not a single partial correlation coefficient could save me from the limitations section that would challenge the value of a year and a half of work. "But what about their zodiac sign?", I used to joke. "Can I get a regression measure on that?" In this social scientific world, these disclaimers were the unfortunate side effect of having to sample my population.

So what does our new theory-less world look like? Challenging George Box’s famous statement that "All models are wrong, but some are useful", Anderson’s article rebukes the academic primacy of models by asking "Who knows why people do what they do? The point is that they do."

But what are they doing? There is a problem in this new world that Anderson points towards, the "Petabyte age" as the Wired headlines announce. It presumes a world that is already defined. This might not seem immediately clear. Anderson, for example, shares a scenario in which an unknown species was discovered, one that ostensibly would never have been found were it not for this data-first, degrees-of-freedom-be-damned, approach to research. Yet these "discoveries" are always based on a predetermined knowledge architecture, in this case the limited perspective that the quantitative data communicates. Call it information architecture if you are a technologist, call it disciplinary categorization if you are a philosopher; these predetermined categories of knowledge allow us to mine the data, sure, but its foundation serves to only reify our ontological understanding of the world.

We have been down this road before. In the early 1990s, new media scholars were concerned with overly-defined approaches to the digital world. In her book Materializing New Media, Anna Munster recalls a computer graphics conference during which philosophers collided with engineers to discuss a unexpectedly controversial paper entitled "Grids, Guys, and Gals: Are you Oppressed by the Cartesian Coordinate System?":

The session that materialized on an oppressively humid Friday morning at the Los Angeles Convention Center was a testament to the disorder and contingencies of embodied life. A group of staunch neo-Cartesians filled the front rows of the audience and — bringing the seventeenth century into an odd alliance with late-twentieth-century computer geekdom — sported logos emblazoned across their mass-printed t-shirts declaring "I heart Descartes."

Munster described this t-shirt protest as a response to a session that, for them, represented an assault on the epistemological foundations of present-day computing.

On the other side of this protest was a far less organized, and certainly less starched, group of academics whose training taught them to question the validity of confining structures such as coordinate systems. X and Y axises let us define points, and these points allow us to measure relative distances. In the emerging world of computer graphics, objects were being created on Cartesian planes.

But think instead of the "space" created by a telephone phone conversation. Why is the location of, and the distance between, the two callers a more valid spacial concept than the virtual, non-Cartesian space shared by the two callers?

The outcome of this conflict is all around us. As we rush to reconstruct our physical lives in online spaces, digital worlds like Facebook and Second Life have largely forgone the potential freedoms of digital environments. When Anderson asserts that we should focus on what people are doing, he forgets Althusser’s Marxist response that "ideology is material" and as such, never offers a problem that is outside of material’s ability to respond. No wonder all of this data analysis is working. We are just measuring predefined user behavior in a digital world, a world we insisted on digitizing.

What then is the role of theory? I entered my current program with the understanding that theory should lead to truth. A year later, I am not so sure. I find much of my academic work more artistic in nature, and more often than not damning to any positivist vision of a universalizing whole. If, as George Box said, "All models are wrong, but some are useful", then maybe the theories and models we use, the work we create, should aspire to do more than represent the world we live in. It should transform it.