Nancy Cartwright on Hunting Causes

摘要

Nancy Cartwright's recent book Hunting Causes and Using Them comes to causality in an interesting stage of its stormy, century-old courtship with economics. A survey article by Hoover (2004, 'Lost causes') counts the frequency of causal terminology in econometrics articles and asks, 'Where have all the causes gone?' Hoover notes a steady decline from the 1930s to the early 1980s, a period of over half a century, followed by two decades of sluggish recovery, to be followed by a marked upsurge of usage in the year 2000. This pattern is both enigmatic and pathological considering that the central aim of econometrics is to provide a methodology for understanding and controlling economical phenomena, and that the founding fathers of econometrics were the ones who developed the basic mathematical tools for causal analysis, in the form of structural equation models (Haavelmo 1943; Marschak 1950; Strotz and Wold 1960). The decline of causal understanding in econometrics has in fact been so severe, that in 1995, while I was working on my book Causality (Pearl, 2000a), Jim Heckman had difficulty naming a single econometric textbook that adequately treats causality.