Lund, R., Westfall, P. H., Hilbe, J. M., Brown, A., and Taleb, N. N. (2007), Special Section: Reviews of the The Black Swan, The American Statistician, 61,189-200: Comment by Finney.

摘要

On a day in 1919 or 1920, as I stood on the red-tiled floor of my home kitchen, a thought occurred. I murmured silently to myself: "David, you are becoming a clever boy. For a long time, you have been able to stand on two legs and now you have learnt to stand on one leg. What next?" I still think that, for a 2 or 3 year old, I showed creditable numerical ability in saying to myself "2 - 1 - 0." I was at the moment stably on one leg. Somehow I willed myself to make a rapid muscular contraction at the knee in an effort to lift the foot. I rapidly became aware of a sore bottom—probably I yelled. I never tried again but memory of the pain has given me a prejudice against any form of extrapolation from particular to general. Throughout a 40-year career in biometric practice, I have always taught that uncritical extrapolation must be avoided and that extreme deviations-from expectation or preconception need investigation as possible clues to new important phenomena. I suggest that Taleb should class me among the angels until he discovers that I believe the Normal distribution often to be an aid to processes of quantitative generalization from sets of data that can support such inferences.