Risk Markets And Politics

Wednesday, August 09, 2006

On "Digital Maoism"

It's difficult to boil-down Jaron Lanier's essay on the limits of online "collectives", but these are the most salient points:

  • Collectives are stagnant and threaten to smother the voice of genius by enshrining consensus. Genius and creativity will naturally not correspond to average opinions.
  • Collectives are too volatile and chaotic, and need to be filtered by individuals
  • It is misleading for collectives to masquerade as objective when they might well contain some personal bias.
  • The collective fails in matters of subjective taste.
Notice the tension among these assertions, especially between the first two. This essay has received plenty of criticism already. Here are some paraphrased examples:

  • Lanier is attacking strawmen and taking phrases like the "Wisdom of Crowds" too seriously without considering how they are really used.
  • Voluntary collective efforts should not be identified with coercive forms of "collectivism". The word "collectivism" in the essay is generally spread too thin over too many distinct endeavors and technologies.
  • Projects like Wikipedia are not actually anonymous collectives, but are largely the undertaking of known, independent individuals operating within some degree of hierarchy and editorial control.
  • All minds are ultimately "hive minds" insofar as they are connections of neurons.
There's no need to rehash all of the criticism here, so let's focus on what specific relevance the essay might have for prediction markets. It's not clear that any of Lanier's concerns particularly confront these examples of crowd "wisdom". Prediction markets always resolve on something objective and have a strong meritocratic element. Even if subjectivity underlies a market, as with predicting box-office receipts or album sales, those traders with "good" taste who recognize quality work by underappreciated and underpriced artists will be rewarded far more in the long run than those who merely bet with the consensus. No-one is advocating using any sort of collective effort to legislate the aesthetic and reverse the long-tail trend in entertainment and culture. It's also unclear how "collectives" would hinder the emergence of a Dylan or Hendrix or Ellington as Lanier suggests. For all the talk of Googlization, older forms of media where those stars rose were much more centralized.

The title of "Digital Maoism" was a very strange choice and one guesses that it was chosen to emphasize the suspect point that collaborative, anonymous endeavors might present themselves as objective while containing significant biases and personal imprints. Lanier of course prefers that collective efforts be guided by individuals, so long as this takes place in the open. Apart from concerns about quality and innovation, he also fears instability:
One service performed by representative democracy is low-pass filtering. Imagine the jittery shifts that would take place if a wiki were put in charge of writing laws. It's a terrifying thing to consider. [...] The calming effect of orderly democracy achieves more than just the smoothing out of peripatetic struggles for consensus. It also reduces the potential for the collective to suddenly jump into an over-excited state when too many rapid changes to answers coincide in such a way that they don't cancel each other out.

First, this passage gives an idea of the measured world-view of the piece, which sneers at both socialism and anarcho-capitalism. (The latter would develop its own mechanisms of regulation and stability much like wikis have.) Second, the phrase "answers coincide in such a way that they don't cancel each other out" should pique the interest of prediction market advocates, as this is a textbook problem-case for such markets. Fortunately, these scenarios should be relatively easy to identify. Logically, one can ask questions like "Are all relevant interests/departments represented?" or "Is there a predominance of hedgers?" Empirically, if one has market data over a long enough period covering a set of sufficiently similar questions, one might be able to weight opinions based on similarities in past biases.

There is one further idea to parry: that there is really nothing emergent in an aggregation of opinions. This is best summed-up in one of the comments on Edge, by Yale Law Professor Yochai Benkler:
This leaves the much narrower set of moves that are potentially the legitimate object of Lanier's critique: efforts that try to depersonalize the "wisdom of crowds," unmooring it from the individuals who participate; try to create ever-higher-level aggregation and centralization in order to "capture" that "wisdom;" or imagine it as emergent in the Net, abstracted from human minds. I'm not actually sure there is anyone who genuinely holds this hyperbolic a version of this view. I will, in any event, let others defend it if they do hold such a view.

James Surowiecki, earlier:
I would question the implicit assumption in Tyler's post that the "knowledge" these markets are trying to tap is something that one employee (or a very few employees) have and that most employees are completely without. No one person knows what a company's sales are going to be in forthcoming quarters, or whether a given product is likely to be a hit or not, or whether a software project will be done in time. The information that's relevant to answering any of these questions well is not concentrated in the hands of one person or a few people. Instead, it's distributed among lots of people, each of whom likely has a small bit of knowledge that's germane.

This seems to just be another terminological straw-man; Surowiecki is talking about specific objective knowledge here, not "wisdom" in the broad sense. Prediction markets might be good at providing objective answers in many situations, but as Jaron Lanier aptly points-out, asking the right questions can be more important.


Post a Comment

<< Home