Tuesday, January 17, 2012
Moneyballing knowledge: Or, Walmart epistemology
I want to make three disclosures at the outset. First, I buy a LOT of stuff from Walmart (call me a backslider, I know the issues, I just can't resist good stuff at those prices). Second, I love Wired magazine, and am a serious fan of Chris Anderson, even when I might disagree. Third, I like Brangelina and really enjoyed Moneyball – I knew I'd enjoy the movie, but did not expect to be as stimulated by it as I have been since catching it on a Toronto-Mexico City flight on the weekend.
OK, way back in 2000, during a collaboration on an Australian Research Council funded project with Chris Bigum and Leonie Rowan, on online and distance learning, we collectively coined a notion of “digital epistemologies”. Michele and I subsequently published a few pieces on this theme and, perhaps for this reason, it never caught on at all. Not even making it part of the title of our book series could breathe life into it. So it goes. We forgot it and let it rest in peace – unlamented.
However, funny things happen. Especially in Mexico (along, of course, with not so funny things). A couple of months back a collective charged with creating a Spanish language Dictionary of Philosophy of Education to be published in Mexico asked for an entry on “Digital Epistemologies”. I felt the earth move under my feet, and since then, from time to time, I have been giving the work some thought.
Way back in 2000 we used the term as a shorthand for a host of issues and questions about knowledge and belief and information and so on under the condition of ubiquitous digital media and mediation. More recently, however, we have been thinking of digital epistemologies in terms of four “ends of”: the end of truth, the end of ideology, the end of theory, and the end of individual(ist) knowledge. That's the story we'll largely focus on for the Dictionary entry.
A little while back we bought a couple of nice brushed steel cordless kettles for use down here in Mexico. They weren't cheap, even though they weren't top of the market. We bought one in the north, and one from a posh store here in Mexico City. After barely a year they both died, the same way, within a week or two of each other. It was a sad end to the sorts of kettles you wouldn't mind your friends and colleagues seeing in your kitchen. They had to be replaced.
Last weekend we arrived in Toronto in time to do a bit of Friday evening shopping ahead of Saturday's class and a quickfire return to Mexico City. I had done some online sleuthing: where is the nearest Walmart to the hotel? Check. What do they have in cordless kettles? Google, as always, was bounteous on both counts. I liked the look of a 1.7 litre white plastic kettle priced at $14.88 normally, but in this particular store for $12. We bought two and filed them away in the suitcase. Early indications are that they are robust, very effective, and should easily outlive their superiors. In dollar terms they'd only have to last 3 months each to do that, but I have much higher hopes.
Following my entranced viewing of Moneyball on the flight back I have been thinking about those Walmart kettles as somewhat analogous to using sabermetrics to get Major League Baseball wins from players who cost only a fraction of what cultural insider baseball lore would identify as “the best players”. Billy Beane wanted wins, just like I want a cordless kettle that works, and works for a decent period of time (a “win” in my terms). He got his wins, and an American League record, by buying the baseball equivalent of Walmart kettles. Wins being the objective, Walmart kettles are not inferior to higher end kettles; and where they win more and win longer they are superior. Screw any cultural snobbery that would insist on a better brand, a better look, a more “sophisticated” product, the kind of thing a chef to the elite would insist upon.
So, more than three years ago, Chris Anderson proposed the end of theory in a typical Chris Anderson gem published in Wired magazine. In the Petabyte Age, said Anderson, the scale of data is such that the Google model of agnostic statistics wins out. Google's algorithms can't tell us why one web page is better than another. They simply tell us which one is better from the perspective of people looking for epistemic “wins”. A kind of sabermetrics of information. When there is so much data and such strong computing power and algorithms as to throw up patterns that can underwrite successful decisions and courses of action, why would we miss “the theory”.
In Anderson's words:
Petabytes allow us to say: "Correlation is enough." We can stop looking for models. We can analyze the data without hypotheses about what it might show. We can throw the numbers into the biggest computing clusters the world has ever seen and let statistical algorithms find patterns where science cannot.
The new availability of huge amounts of data, along with the statistical tools to crunch these numbers, offers a whole new way of understanding the world. Correlation supersedes causation, and science can advance even without coherent models, unified theories, or really any mechanistic explanation at all.
Predictably, Anderson's approach brought on similar kinds of refusals and rebuttals from academics and (certain kinds of) scientists to those portrayed in Moneyball as coming from baseball purists faced with sabermetrics.
Two things are pretty obvious here. The first is that we are looking at a trifecta (assuming for the meantime that my Walmart kettles will go the distance). My consumer metrics are backing my new kettles to give me more wins at a cheaper price (less complication) than their brushed steel predecessors. The Boston Red Sox won the World Series in 2006 using sabermetric principles and, furthermore, Major League Baseball franchises have increasingly embraced sabermetrics, and teams that have done so have fared not too badly in recent years. And, as Anderson observes, there is a discernible drift within fairlysome mainstream areas of science toward finding statistical patterns and going from there. In other words, there is something of a trend that is going across wide stretches of everyday life: consumer economics (I am not the only person who shops Walmart when they have wider options); baseball; science. You will be able to think of others. But maybe there is a pattern here that goes below the surface of contemporary ways of doing and being.
The second is that the stakes are pretty obvious. It is clear what the gains are and where they lie – even though the games in question are by no means the same, they are nonetheless different versions of getting more with less. That is, “more” in some sense. The losses are in tradition, “culture” (in an elite sense, maybe), style, a certain kind of propriety, and so on.
The question for me, I guess, is whether there is a deeper loss -- maybe in something akin to the sorts of things John Stuart Mill was trying to redress when he sought an alternative to Jeremy Bentham's version of Utilitarianism. Or, again, maybe in a stronger sense still: a sense that speaks to caring and taking care in ways that we presume humans to be uniquely capable of at their best. We might get a good pragmatic decision from acting off a correlation, but is that decision as good – in a humanist, ethical, moral, or whatever sense – as one that is backed by the kind of caring that sometimes goes into hard core theoretical angst-ing?
Needless to say, I take at least a little solace from knowing that so far as cordless kettles are concerned there is in all likelihood no less exploitation of human beings involved in the creation of artifacts that turn on brushed steel chefs than in the creation of my plastic fantastic boiler. Nonetheless, the wider phenomenon leaves me uneasy. Can we have “purism” and “justice”? Can we all participate fairly in “noble traditions”? If not, what is the moral math around doing/getting more with/for less?
If I were you I'd give notions like digital epistemologies a big swerve … They can mess up your day.