There was an article in a recent New Yorker (I think) about a system for analyzing new recorded music to predict whether it will sell. It's pretty accurate, apparently. That's a little different deal from this, of course, but it seems pretty likely that there are patterns to which we respond but of which we're not altogether aware.
I suppose the great limitation is that such an analysis is inherently conservative; that is, it can work only from what has already been judged, and really only from a small subset of that, the part that has been programed into it. The analysis couldn't say anything useful about Whitman, say, or Dickinson. Of course, neither could very many of the critics of their day.
Richard
|