Let me first define a "pundit" as "someone who writes about politics."
Do pundits read anything other than (1) other pundits, and (2) newspapers? I wonder this because lots of punditry is really bad. And the reference points for punditry tends to be other punditry; thus a blogger on National Review's "Corner" will reference some other conservative blogger, or lambaste some liberal blogger, or cite to the Washington Post or New York times and make some vague reference to "bias."
Am I wrong about this?
Does this serve readers well? Is this intellectually honest or intellectually stunting? It seems to me that there's nothing really wrong with scribbling solely about politics and party; but in order for that writing to be good -- that is, in order for it to be analytically or otherwise insightful -- it can't simply exist in a pundit vacuum.
It seems to me that one of the drawbacks of "the blogosphere" is that it gives pundits the illusion of reading widely, when in fact certain narrow insights, conventional wisdom, etc. are being repeated over and over again and so the cleverness arises either from the super-subtle insight, or from the same thing being said in more clever fashion.
I like Andrew Sullivan, almost despite myself. His writing can betray a certain amount of self-absorption, but I can't fault him for intellectual honesty. And while he somtimes succumbs to the easy dramatic writing flourish, that's not hard to forgive because he tends to write very clearly and his view is extremely well-informed. He's not just reading political bloggers; he seems to read everyone. I think that matters, so that he has not only Tom Friedman's latest column in mind, but also Oakshott or Foucault or Sen. And he seems as interested and curious about science, religion, and philosphy as he is about politics. That combination makes for great reading, so the flaws are forgivable.
Thus on the internet, as in the rest of written life, the better writers tend to be the better readers.