[PREV - DEGENERATE_TRIANGLE]    [TOP]

SHADE_CAST

                                               February 5, 2010

If you skip Judge Posner's rather
arbitrary judgements, The unique
material that remains in Posner's
"Public Intellectuals: A Study of
Decline" is:

   (A) a relatively large list of "intellectuals"
   (B) a ranking of them by popularity,
       (essentially via citation stats).

This is an attempt at putting a normally entirely
subjective critical process on a more objective basis.

It's a very weak attempt in
many respects.  The initial
list can be criticized on              Including too many people:
many grounds...                        Power brokers like
                                       Henry Kissinger compete
                                       with arguably more
                                       authentic intellectuals
                                       such as Noam Chomsky.

                                       Excluding too many people:
                                       people with unusual backgrounds
                                       are skipped: John Lennon, Bono,
                                       Rush Limbaugh, Werner Earhardt...


Worse: surveying major media sources
for names dropped is fundamentally a
popularity contest, and it seems           On the other hand if you
profoundly uninteresting if what           aspire to riches and
you're interested in is truth rather       fame, you might look
than influence.                            through the list and
                                           launch attacks on nearly
   So: how might one improve               all of them, in hopes of
   on this proceedure?                     provoking lots of
                                           responses: instant notoriety.
   A few general
   approaches occur                            TRAITORS_OF_INTELLECT
   to me:

     Use the network of
     references to evaluate       A difficulty: you must be
     the significance of one      able to distinguish between
     intellectual in terms        approval and approbation.
     of another.                  Counting name drops alone
                                  is not adequate.
        Citation counts seem
        rather crude in the
        post-PageRank era.

        Can this be done dynamically?
        Given different initial           "If you like Paul Krugman,
        assumptions?                      Noam Chomsky, Jane Jacobs,
                                          and Bruce Sterling, you may
                                          also like...."


   Use third parties organizations
   (such as FAIR, AIM, factcheck.org, etc.)
   to evaluate quality of output:

      Factual accuracy
      Accuracy of predictions                      Posner has a section
                                                   where he talks about
      Ideological classification?                  that, but proposes
                                                   no way of formalizing
                                                   the "keeping score"
                                                   on predictions.
          It may be possible to combine the
          two: Use one agency's approvals
          to seed the process, analyze the
          graph to rank others they don't
          mention.



   What will not work:

   A wide-open social network, wisdom-of-the-
   crowds approach without anything like            THE_TOY_WEB
   verification of the identity of sources.



--------
[NEXT - LONG_NOSED_BEAGLES]