Science and Morals 201 



our ethical notions are fundamentally based on a system 

 of individual responsibility for individual acts. The principle 

 of statistical correlation between two sets of events, al- 

 though accepted in scientific practice, is not usually felt to 

 be ethically completely vahd. If a man hits a baby on the 

 head with a hammer, we prosecute him for cruelty or 

 murder; but if he sells dirty milk and the infant sickness 

 or death rate goes up, we merely fine him for contravening 

 the health laws. And the ethical point is taken even less 

 seriously when the responsibility, as well as the results of 

 the crime, falls on a statistical assemblage. The whole com- 

 munity of England and Wales kills 8,000 babies a year 

 by failing to bring its infant mortaHty rate down to the 

 level reached by Oslo as early as 1931, which would be 

 perfectly feasible; but few people seem to think this a 

 crime." ^ 



Since he wrote these paragraphs, it has become more 

 and more obvious that we must begin to think of such 

 statistical responsibilities in much the same way as we 

 have thought in the past about individual crimes. The 

 people who die because of the pollution of the air or the 

 reckless driving of an automobile are just as dead as if 

 an individual had run a sword through them. Those re- 

 sponsible for polluting the air or driving the car improperly 

 are just as responsible for the deaths as if they had com- 

 mitted an old-fashioned crime. The only difference is that 

 they didn't know in advance the names of the particular 

 people who were going to die. 



Sooner or later, it will seem very odd that we bear down 

 hard on a man who lures a child into a car and later 

 strangles it in a park, while we shower high salaries and 

 other dignities on advertising managers and TV artists who 

 lure thousands of teenagers into the statistically fatal habit 

 of cigarette smoking. 



