Reality Bites Or Fewer Than Five Hundred Generations If Sal in the Past Few Decades)

SPRING BOOKS on inequality in the United States tions with our present size of population, ( although this neglects the startling rever Reality bites or fewer than five hundred generations if sal in the past few decades). Third, it the population grows larger. Leslie elab Freeman J. Dyson rightly takes to task the self-appointed orates this conclusion with numerous doomsters who perhaps do not deliberate variations. He examines many different ly exaggerate, but who exaggerate because The End of the World: The Science and hypotheses and deduces the correspond of poor research and inadequate ques Ethics of Human Extinction. By John ing probabilities of survival. But the tioning. The general laissez-faire tone of Leslie. Routledge: 1996. Pp. 305. essential core of his reasoning is as I have the volume will offend some, but, if so, the £16.99, $23. To be published in the stated it. The reasoning stands or falls offended should be prepared to answer United States in May. according to whether the application of the case made in the book. the Bayes rule is valid or invalid. Holdgate accepts there is an array of THE subject of this book is a philosophi After careful consideration, I state environmental problems and spends more cal speculation that the author calls the unequivocally that the application of the time telling us what has to be done. The "doomsday argument". The same specu Bayes rule is here invalid. The reason the recommendations are familiar but no less lation was independently put forward in rule fails is simple. The two quantities important for that: empowering commu Nature (363, 315; 1993) by Richard Gott, (probability of B) and (probability of B nities, moving from outright conservation who gave it the more appropriate name given A) are not based on the same div to sustainable use, education and new of the Copernican principle. The idea is ision of the world into known and alliances between interest groups. He is to estimate the probability of the survival unknown. When we estimate (probability sincere in his belief that we need a 'new of the human species by using the Bayes of B) to be close to unity, we assume that ethic' of care, but the menu of action rule. The Bayes rule says that (probabil we know enough about palaeontology to looks strangely naive: building a new con ity of A) multiplied by (probability of B be sure that we are living among the first science and spirituality, reforming the given A) is equal to (probability of B) hundred billion humans. When we esti international economic order, and "new multiplied by (probability of A given B). mate (probability of B given A) to be economic thinking" (when we have hardly The rule is supposed to apply to any two small, we assume that we know nothing of made use of the power of the 'old' eco hypotheses A and B, because both sides our place in the history of our species. The nomic thinking) to improve our lot and of the equation are equal to the probabil two estimates of probability are based on the lot of other species. ity of the hypothesis (A and B). Formally, incompatible assumptions and are there Holdgate's volume is a worthwhile read the rule is almost a tautology. But when fore not comparable. The application of but it has an inner risk. Spiritual shopping it is applied to hypotheses in the real the Bayes rule is invalid because the rules lists calling for 'changes in values' invite world, it can easily lead to absurd conclu of the probability game have been others to philosophize and debate while sions, because of ambiguities in the changed between the left and right sides the environment burns. Holdgate is wise meaning of probability. Every estimate of of the equation. enough to see some of this risk and he probability must be based on a certain Half of the book is devoted to theoreti calls for other action too. But one won division of the world into known and cal discussion of the Bayes rule and its ders if the dreamers really understand unknown. The Bayes rule is valid only if application to the question of human sur that it is all about competition between the four probabilities in the equation are vival. This discussion is worthless, being humans and other species for space, for a all based on the same division of the based on an invalid use of the rule. The slice of the Earth. Until environmental world. The phrase (probability of A) has other half of the book is a recital of the assets secure economic value greater than no absolute meaning independent of many real dangers to which our species is that embodied in the competing uses of previous knowledge. exposed. Here Leslie has nothing new to the land on which they rely, the state of Leslie applies the Bayes rule to a vari add. He refers to a large number of humanity really is at risk, as much, it ety of situations of which the following is sources from the literature of science and seems, from the concerned as from the typical. Hypothesis A says that the risk assessment, but does not examine benign optimists. D human species will ultimately give birth them critically. to less than X people. Hypothesis B says Almost two hundred years ago, in David Pearce is at the Centre for Social and that Leslie and his readers live at a time 1798, Robert Malthus published his Economic Research on the Global Environ• when the human species has given birth famous Essay on the Principle of Popula ment, University College London, Gower to fewer than Y people. Evidence from tion as It Affects the Future Improvement Street, London WC1E 6BT, UK. history and palaeontology indicates that of Society, deducing dire predictions of (probability of B) is close to unity when Y future misery from two hypotheses New in paperback is of the order of a hundred billion. Also, that he called the geometrical increase The Strange Case of the Spotted Mice (probability of A) cannot exceed unity. of population and the arithmetical and Other Classic Essays on Science by Thus the rule implies that (probability of increase of subsistence. His predictions Peter Medawar. OUP, £7.99. A given B) does not exceed (probability of inevitable poverty failed because his of B given A). Now, the probability that hypotheses were faulty. The increase of From Here to Infinity: A Guide to Today's Leslie and his generation should happen Mathematics by Ian Stewart. OUP, subsistence turned out to be much faster £7.99. A revised and retitled edition of to be among the first Y people in order of than arithmetical. Nevertheless, uncriti The Problems of Mathematics. birth, out of a total population greater cal belief in Malthus's predictions helped than X, is less than Y/X. That is to say, to hold back political and social progress Natural Acts: A Sidelong View of (probability of B given A) is less than in Britain for a century. Because of this Science and Nature by David Y/X. Therefore the Bayes rule leads to unhappy precedent, I consider it impor Quammen. Avon, $11. "Quammen's book is full of strange information about the conclusion that (probability of A tant to call attention to the fallacy in animals and their habitat, about Tycho given B) is also less than Y/X. That is to Leslie's argument. A mistaken philo Brahe's artificial nose, and other such say, there is less than a ten- per-cent sophical argument may have serious con oddities. His exuberance carries one, chance that the human species will sur sequences in the real world. D sometimes unwillingly, along", wrote vive long enough to give birth to a trillion Walter Gratzer in a review in Nature 318, people. The conclusion is that our Freeman J. Dyson is at the Institute for 116 (1985). species has only a ten-per-cent chance of Advanced Study, Princeton, New Jersey surviving as long as five hundred genera- 08540, USA. 296 NATURE · VOL 380 · 28 MARCH 1996