I do not believe that my pointing out that Herbert Simon's use of an invalid mathematical argument "completely misses the point" about self-consistent prediction, as claimed by Steven Durlauf in his review of my book Impossibility (THES, September 4).
The Brouwer fixed point theorem simply cannot be applied to discrete systems like choices. My concern was to clear up the apparent discrepancy between the general theorems of Popper and Mackay forbidding predictions of this self-embodied sort and the claimed "theorem" of Simon that appears in some texts on mathematical aspects of politics and economics.
John D. Barrow University of Sussex