从我的日中外交经历谈起

我从事日中外交方面的工作大约有四年,先后担任过外务省亚洲局中国课课长和日本驻华大使。这段经历虽然在时间上只占我40年外交生涯的十分之一,但是若论工作内容,似乎是这个数字的三倍甚至五倍。一任亚洲局中国课课长时期我被任命为外务省亚洲局中国课课长是在1973年初。那时正     

Game Theoretical Formulations of The Olson Problem*

Abstract This paper takes up one of the basic themes of Mancur Olson's Logic of Collective Action (Harvard University Press 1965). that is a group size as a cause of suboptimal provision of collective or public goods. A general framework is developed for classifying collective action situations involving public goods provisions. This framework focuses on the two characteristics: relations between contribution and provision, and rivalness or jointness in consumption of the collective goods. This framework distinguishes six types of collective actions, for each of which a game theoretical formulation is developed to obtain. models concerning social movements against (or for) new legislations, a petition for the recall of an official, a strike, lobbying, building a lighthouse, creation of a database, etc. These models, formulated either as an N-person chicken game or as an N-person prisoner's dilemma game. are examined with respect to how a group size affects non-cooperative equilibria and their Paretooptimality. There is no group size effect in the collective action situations formulated as an N-person chicken game, while large groups may suffer from suboptimal provision of the public goods in the collective action situations formulated as an N-person prisoner's dilemma game. Two types of the group size effect in N-person prisoner's dilemmas must be distinguished. In some cases. “no contribution” is the equilibrium regardless of the group size. but increase in the group size makes the equilibrium Pareto-deficient. In other cases, increase in the group size changes the equilibrium from the Pareto-efficient one with N contributors to the deficient one with no contributors.     

Dose Response for Infection by Escherichia coli O157:H7 from Outbreak Data

In 1996, an outbreak of E. coli O157:H7-associated illness occurred in an elementary school in Japan. This outbreak has been studied in unusual detail, making this an important case for quantitative risk assessment. The availability of stored samples of the contaminated food allowed reliable estimation of exposure to the pathogens. Collection of fecal samples allowed assessment of the numbers infected, including asymptomatic cases. Comparison to other published dose-response studies for E. coli O157:H7 show that the strain that caused the outbreak studied here must have been considerably more infectious. We use this well-documented incident as an example to demonstrate how such information on the response to a single dose can be used for dose-response assessment. In particular, we demonstrate how the high infectivity limits the uncertainty in the low-dose region.     

PARTIAL CORRELATION AND CONDITIONAL CORRELATION AS MEASURES OF CONDITIONAL INDEPENDENCE

This paper investigates the roles of partial correlation and conditional correlation as measures of the conditional independence of two random variables. It first establishes a sufficient condition for the coincidence of the partial correlation with the conditional correlation. The condition is satisfied not only for multivariate normal but also for elliptical, multivariate hypergeometric, multivariate negative hypergeometric, multinomial and Dirichlet distributions. Such families of distributions are characterized by a semigroup property as a parametric family of distributions. A necessary and sufficient condition for the coincidence of the partial covariance with the conditional covariance is also derived. However, a known family of multivariate distributions which satisfies this condition cannot be found, except for the multivariate normal. The paper also shows that conditional independence has no close ties with zero partial correlation except in the case of the multivariate normal distribution; it has rather close ties to the zero conditional correlation. It shows that the equivalence between zero conditional covariance and conditional independence for normal variables is retained by any monotone transformation of each variable. The results suggest that care must be taken when using such correlations as measures of conditional independence unless the joint distribution is known to be normal. Otherwise a new concept of conditional independence may need to be introduced in place of conditional independence through zero conditional correlation or other statistics.