10 of 10 people found the following review helpful
on June 14, 2006
In this small book of essays, Nobel laureate, Sir Peter Medawar defines what science is and what it is not. Using a familiar and engaging voice, he addresses some of the big questions about science, revealing its true nature to the learned scientist and the novice alike.
3 of 9 people found the following review helpful
on December 18, 2007
Although it begins "This is a serious book", this is not a serious book. Its first two essays do not even pretend to be serious but are instead concerned with such things as quipping about "the incredulous derision" with which a modern-day grant-giving body would have greeted "a research proposal 'to discover a means of making human flesh transparent'" in 1900 (p. 46), and quoting a 17th century Shadwell play on the tension between pure and applied science: "When the curtain goes up, Sir Nicholas Gimcrack is seen making froglike swimming movements on the table in his workroom. Does he intend to swim in the water? ... Never, sir; I hate the water. I content myself with the speculative part of swimming and care not for the practical." (p. 8). The third and "principal" (p. xii) essay at least has serious claims in it, but its arguments are so utterly underdeveloped that it can hardly be called an essay at all. The main claim is that "there is an intrinsic, built-in limitation upon the growth of scientific understanding" (p. 59), stemming from "the Law of Conservation of Information, which runs as follows: No process of logical reasoning---no mere act of mind or computer-programmable operation---can enlarge the information content of the axioms and premises or observation statements from which it proceeds." (p. 79). Medawar's only proof of this law is to debunk a possible counter-argument: if science could reach inductive conclusions of the type "all swans are white" then the law would be false, for the statement says more than the totality of observations that prompted it. But it is not credible to admit such induction as science (p. 80). For one thing, "Our sense of the fitness of things is offended by the idea that an induction such as 'All swans are white' can be corroborated by the discovery in a trash heap of an old black boot, yet so it is: for if all swans are white, if follows logically that all non-white objects are non-swans. If, then, any black object is discovered which anxious scrutiny shows not to be a swan, then we have confirmed a logical prediction from a hypothesis and given ourselves an extra incentive to believe in it." (p. 15). From Medawar's law it apparently follows, in some manner not further explicated, that science cannot answer "ultimate questions" such as "what is the point of living?" (p. 66), and that to answer such questions from within science is as impossible as "to deduce from the axioms and postulates of Euclid a theorem to do with how to cook an omelet or bake a cake" (p. 82). Medawar further maintains that there are no other limits on the growth of science, which he also purports to prove by equally lightweight arguments. For example, increased specialisation is not a problem because "There were always sciences and there were always arts, and no one man knew them all---no one man ever had the know-how to make glass, brew beer, dress leather, make paper and cast a bell" (p. 71). That Newton did not know how to brew beer is hardly very conclusive proof that pathological specialisation is not detrimental to scientific progress, but that is beside the point: this unserious book proves nothing but is delightful reading all the same.