I know that I’ve been on blog hiatus for a while, and as my minions* have the troubling property of non-existence, I can’t opt for the Sully option of serving up a welter of outsourced stuff to hold down the space.
But I’m trying to dig out from under the delights of start-of-term and all that. To do so, I’ve decided to re-enter the fray with a kind of series, or at least a loosely-agglomerated guilty-by-association stack of posts to do with a real sorrow of mine, the decline of a once-great American marketplace of ideas, The Atlantic Monthly.
I’m going to indulge in my usual blog pleasure of burying the lede, so just to flag why you might want to wade through (or jump past) what follows, here’s my claim: no one at The Atlantic understands — or at the very least, cares to engage — science, whether as a mode of thinking or as a body of actual knowledge.
But, of course, The Atlantic as a brand is supposed to convey seriousness of thought and purpose, so they can’t actually ignore science altogether…and what has happened, as I hope to document, is that the place has become a playpen full of science-y stuff; appeals to the sound and appearance of rigor that masks business as usual…about which, details below.
To be fair-ish: the place still attracts good, some great writers: James Fallows is the genuine article; Ta-Nehisi Coates is putting together a world-beating career; Corby Kummer has been a favorite for a long time.and in my own area of particular interest, science writing, two or three articles from the mag show up every year in the “Best of” collections. I’m sure if you dig through the website long enough you’ll find plenty of other stuff that won’t make your eyeballs bleed.
But, but, but…much of the place has been turned into what aspires to be an upper-middle-brow advocate of right wing politics, and that beast uses the brand and something of the language of The Atlantic Monthly-as-was to obscure a fundamentally flawed understanding of what actually happens in the real world…because, as is well known, reality has a long-established liberal bias.
For a first example of this kind of rot, consider the Atlantic 50, a ranking of what that rag’s editors considers to be their “all stars,” (sic), “the fifty most influential commentators” in the country.
I got to this through DJ over at Balloon Juice, and he makes the necessary point that the list is bullshit — with a thumb on the scale to tilt the claim of influence towards The Atlantic’s pre-existing politics. As he writes,
Of the first 13, there are 9 conservatives, two liberals, and two other (Tom Friedman and David Broder).
It gets worse when you dive into the weeds of the list of course. Josh Marshall comes in at 29, which doesn’t seem hopelessly stupid…until you realize that Kathleen Parker is listed at 21, eight spots ahead…and there is simply no rational measure by which to justify ranking a middle-level member of a stable of writers at a declining venue above someone like Marshall, who is both redefining the form and the institutional structure of journalism, and has had demonstrable, potent impacts on daily political life.
Even more risible, Jonah Goldberg (thx, JRE), whose primary venue is a hack site overtly preaching to its choir, and is besides a writer whose influence even among his co-conspirators is muted by the fact that his writing is consistently awful, whether measured on prose style or clarity (or even detectability) of thought, weighs in (sorry) at number 34. Meanwhile, look for the ranking of, say, Markos Moulitsas….and you find it nowhere.
I mean, seriously guys. This isn’t even laughable as an editorial judgment.
Examining the rankings as a whole, it’s hard to avoid the sense of it as The Atlantic’s circle jerk — a celebration of those folks it likes whose influence is on each other.
Now, if The National Review, for example, were to publish a list of influential opinion leaders, one would expect it to have a bias in this direction — and no one would regard that as particularly untoward. (Worthy of sarcasm, perhaps, but not in itself risible behavior.)
But all of that is not what excites my scorn. The Atlantic is free to name whoever it wants as the people it wants to listen to.
What does push my snark button is the desperate plea to be taken seriously in its celebration of its own image in the mirror. Read this description of the rigor with which this list was compiled:
…our team spent months collecting and analyzing data, tracking a group of 400 names that eventually became our 50. Our in-house methodology relies on three streams of information:
- Influence: A survey of more than 250 Washington insiders – members of Congress, national media figures, and political insiders – in which respondents rank-ordered the commentators who most influence their own thinking
- Reach: Comprehensive data collection and analysis to measure the total audience of each commentator
- Web Engagement: In partnership with PostRank, a company specializing in filtering social media data, the Wire analyzed top commentators on 16 measures of webiness, including mentions on Twitter and performance on popular social media sites like Digg and Delicious
The final list is the result of an algorithm that brings together these three factors.
There are problems with each bullet point — and I’ll leave it as an exercise for the reader to tease them out.
But the point of this whole long screed, rests on very last line of the self-justification quoted above.
We are supposed to trust this list, to credit it with meaning, because all these essentially subjective factors of influence have been laundered through “an algorithm.”
I am so relieved.
I know this tune. It’s a rocker; it’s got a beat; I can dance to this fine, sexy algorhythm.
Give me a freaking break.
Bluntly: I believe the author(s) of that sentence may not know what an alogorithm is, or if they do, they’ve deliberately misused a technical term in a semi-technical context to obscure what they are actually attempting to describe/conceal.
The concept of an algorithm as used as a term of art in technical fields resists formal definition. But as a working notion, algorithms involve at a minimum, explicit instructions that can be carried out by a person or a machine which specify operations iterated through a sequence of steps, and produce an unambiguous correct answer (for a certain value of “correct”) within a finite time.
Algorithms in computing meet this cartoon definition. To put it even more simply:
“In computer systems, an algorithm is basically an instance of logic written in software by software developers to be effective for the intended “target” computer(s), in order for the software on the target machines to do something.”
That’s what The Atlantic’s people appear to have done: they wrote, or hired done, a program that took the numerical inputs – the measures of the three criteria above – calculated a single number for each candidate object (a pundit) based on a weighting scheme of The Atlantic’s editors’ devising, and then organized those numbers into a rank ordered list based on those numbers.
Smells like teen algorithm to me: iterated calculations over a body of data yielding a definitive answer – a “correct” one in the sense that the list corresponds to the appropriate numerical sequence from high to low calculated based on the inputs supplied.
Readers attuned to the rhetoric of science, and especially of popular communication of science will have long since gotten to where I’m going with this.
The use of the word “algorithm” on its own puts some English on the ball: it sounds authoritative, and is deployed in the same way and for the same purpose that Smartwater applies the epithet “vapor distilled/electrolytes” to justify charging a gazillion bucks for its carefully formulated dihydrogenmonoxide.
At a slightly deeper level, the fraud turns on the specific property of algorithms: they always give an answer, one that is true within the parameters of the calculation: if Paul Krugman’s calculation gives him a score of 1 (on a scale of 0-1) and Rush Limbaugh weighs in (sorry again) at .997, then Krugman is more influential than Limbaugh and the list will reflect that in its first and second place rankings…accurately, given the constraints, the data, selected as inputs in the first place. [Made up number alert.]
That correctness, that certainty, gives the term “algorithm” it’s fine glow. We don’t just get answers, ma’am, they’re the right answers – and we know that this is so because we’ve poured secret algorithm sauce into every bottle.
That’s the emotional affect of the word, certainly that sought by The Atlantic in its cheerfully sophomoric defense of the methods behind its list.
The reality, obvious I’m sure, is that which left so many traders wondering what happened when all those funny investments in the mortgage markets went to sleep with the fishes, after all their wonderful proprietary algorithms had declared them safe as houses.
Any computational algorithm contains a formalization of the assumptions of its authors, their perceptions and judgments about the nature of the reality of whatever is to be analyzed.
Any algorithm can produce valid results given such own assumptions, and, yet, as here, remain utterly unconnected with what those of us who live in it laughingly call “the real world.”
In other words: The Atlantic’s list is an exercise in conventional wisdom laundered through code. Its authors want to cloak themselves in the image of science without actually doing the work – or more accurately, without gambling that their presuppositions might not survive contact with experience.
And this is a trivial example of the way in which The Atlantic, and by extension a great deal of the right-web, has a problem with science — and hence, I would argue (will argue in some upcoming posts on problems more substantive than a gimmick-list of rank-ordered jaw-flappers) — prefers to treat it as a fashion accessory.
To take science seriously would require too much work, and would be inconveniently likely to confound certainties too good to check.
*One of my favorite words, with this definition from the sixteenth century: 1501, “a favorite; a darling; a low dependant; one who pleases rather than benefits” [Johnson], from M.Fr. mignon “a favorite, darling” (n.), also “dainty, pleasing, favorite” (adj.), from O.Fr. mignot, perhaps of Celt. origin (cf. O.Ir. min “tender, soft”), or from O.H.G. minnja, minna “love, memory.” Used without disparaging overtones 16c.-17c.
Images: Pieter Breughel the Elder, “The Tower of Babel” 1563
Scott Feldstein, Bottles, 2005