Why Do We Believe What Science Says? 25 



in the past so that we can modify them in such a way that 

 it is unlikely to occur in the future. Further studies have 

 shown, for example, that the mosquitoes transmit a virus 

 from one person to another, that the virus circulates in 

 the blood of the sick person for a certain number of days, 

 and that the mosquito must bite during that time in order 

 to pick up the virus. Furthermore, only certain types of 

 mosquitoes can carry the virus. The most important one 

 happens to hve in and about houses and to breed in rain- 

 water which collects in household vessels of various sorts. 

 The existence of yellow fever, therefore, depends on an 

 intricate pattern of rain, tin cans, the presence of large 

 numbers of nonimmune individuals hving close enough 

 together so the mosquitoes can fly easily from one to an- 

 other, the presence of the virus, and so on. It is the knowl- 

 edge of this entire chain of events that we really need in 

 order to control the disease, and it is naive to talk about one 

 or another of the steps in the chain as The Cause. 



It is important to note also that we really don't have 

 to worry very much about the fact that there is some degree 

 of uncertainty about each step. It must be admitted that 

 most people do suffer from the urge to be absolutely sure 

 about things or at least about some things. Skeptical people 

 like Hume, who have repeatedly pointed out that it is 

 impossible to be absolutely sure about anything, have 

 never been very popular. Hume, of course, allowed that 

 the fact that the sun has come up over the horizon every 

 morning for several million years makes it very likely that 

 it will come up tomorrow morning. Since his time science 

 has made great progress in analyzing and refining the 

 meaning of the phrase "very likely." These studies have 

 been gathered together in a body of knowledge known as 

 the theory of probability, which enables us to state the 

 likelihood or probability of future events in terms of num- 



