In lieu of an abstract, here is a brief excerpt of the content:

7 THE UNEXAMINED RHETORIC OF ECONOMIC QUANTIFICATION Even in the most narrowly technical matters economists have shared convictions about what makes an argument strong, convictions which they have not examined, which they can communicate to graduate students only tacitly, and which contain embarrassments to the official rhetoric. Rhetorical Standards, for Example, Are Necessary to Measure the Integration of Markets Do numbers tell? According to the official rhetoric, yes: only numbers. Most economists believe that once you have reduced a question to numbers you have taken it out of human hands. That's where the rhetoric of quantification goes crazily wrong. The best quantitative economists know it. The rhetorical point is idiotically simple. It is that in a human conversation a number is high or low relative only to some standard, and the only relevant standard is provided by the humans involved. Ten degrees below zero is paralyzing cold by the standard of Virginia, a normal day by the standard of Saskatoon in January, and a heat wave by the standard of most interstellar gas. Everyone knows this. A New Yorker cartoon shows faucets labeled "Hot: A Relative Concept" and "Cold: A Relative Concept." A thing is not large in itself. It is large (or yellow, rich, cold, stable, well-integrated, selfish, free, rising, monopolistic) relative to something else, and this something has to be specified. The question "But how large is large?" applies to any quantitative argument. It gets some of its excellence from its father in thought, the terrifying, mind-stunning "So what?" and from its Jewish mother, "So what else is new?" Few better questions can be asked, because most inadequate scholarship errs more in relevance than in execution. What is remarkable about this obvious question is how often it is not 100 101 The Unexamined Rhetoric ofEconomic Quantification asked. On the issue of how much better black children do in nonsegregated schools, for example, Robert Crain of the Johns Hopkins Center for Social Organization of Schools remarks that "there is a great deal of debate about when improvement is a big deal and when it isn't a big deal." Complaining that social scientists have been trained to think in terms of merely statistical significance, he notes that they "have never arrived at a consensus on how big a number is big" (1984, p. 12). The same point can be made about the collateral controversy over race and I.Q. revived again in the debate over The Bell Curve. The technical issue is whether or not the averages of white and and black I.Q.s are different statistically. I.Q. is a questionable notion to begin with and hard to measure free of cultural bias. The point I am making here, however, observes that the distributions of black and white I.Q.s largely overlap. An alleged difference in averages, however certified by the standard of merely statistical significance, might not therefore be a Big Deal. It has no practical use. On the basis of a statistically significant difference between the races in average I.Q., for instance, you would hardly propose to use race as a criterion for excluding certain children from certain schools. Under such a policy, even accepting its repulsive moral base, most of the students would be placed in the wrong school. Statistically significant or not, the difference is too small to matter. The point comes up repeatedly in statistical thinking. The rhetorically savvy scientist asks every time, "So what?" "How large is large?" "What does it matter for the intellectual or political or moral issue at hand?" Much of economics turns on quarrels of characterization: Is America monopolistic? Were medieval peasants selfish? Is the market for goods worldwide? Is capitalism stable? These are quantitative questions, all depending on answers to the question "How large is large?" That the quarrels of characterization go on and on, passing from one century to the next unanswered, suggests that the rhetoric has failed. No one answers the question "How large is large?" Everyone knows it has to be asked, but no one answers it. The last step of most calculations in economics or history therefore is sleight of hand, the more convincing because the magician performs it so absent-mindedly: "The coefficient in a regression of domestic prices on foreign prices is statistically insignificantly different from 1.00, and therefore purchasing power parity is true." "The number of formal whippings of slaves was less than 0.7 [or perhaps 1.2] a year, and therefore...


Additional Information

Related ISBN
MARC Record
Launched on MUSE
Open Access
Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.