The end of the world, a popular theme in philosophy, goes back beyond biblical times, has generated a massive literature of eschatology, is associated with recent cult tragedy, and figures in stage entertainment (Beyond the Fringe) and literature (On the Beach). No one knows how the world will end; in the Middle Ages comets were feared as harbingers of doom, as were other celestial portents, including solar eclipses, while records inaugurated by Chinese rulers, who kept watch on the skies, foretold no impending disasters, but proved useful for modern astronomers interested in comets, sunspots, aurorae, and supernovae. The notion that the heavens might threaten existence was revived in recent times by Jeans and Eddington, who, calling on the second law of thermodynamics, foresaw a heat death for the universe as it approached its isothermal fate.

John Leslie, a professor of philosophy, catalogues a variety of terminal scenarios, including the astronomical, and provides scholarly detail as he attempts to estimate the probabilities of occurrence. The list is rather sobering: Risks already well recognised 1. Nuclear war.

2. Biological warfare.

3. Chemical warfare.

4. Destruction of the ozone layer.

5. Greenhouse effect.

6. Poisoning by pollution.

7. Disease.

Risks often unrecognised Group 1: Natural disasters 1. Volcanic eruptions.

2. Hits by asteroids and comets.

3. Extreme ice age due to passage through an interstellar cloud.

51¹ú²úÊÓƵ

4. Nearby supernova.

5. Other massive astronomical explosions.

6. Essentially unpredictable breakdown of a complex system.

7. Something-we-know-not-what.

Group 2: Manmade disasters 1. Unwillingness to rear children.

51¹ú²úÊÓƵ

2. Disaster from genetic engineering.

3. Disaster from nanotechnology.

4. Disasters connected with computers.

5. Disaster from some other branch of technology, perhaps just agricultural, which had become crucial to human survival.

6. Production of a new big bang in the laboratory.

7. Possibile production of an all-destroying phase transition.

8. Annihilation by extraterrestrials.

9. Something-we-know-not-what.

Risks from philosophy These include: threats associated with religions; Schopenhauerian pessimism; negative utilitarianism; and the prisoner's dilemma.

Details are presented in two chapters that constitute a substantial resource. The treatment tends to uncover the darker aspects of what might seem to be lesser risks. Chemical pollution, for example, is certainly making parts of the world unlivable, is absorbing massive resources, and is poisoning oceans, important rivers and extensive agricultural regions. But, grim though this is, surely the vast, relatively uninhabited areas of the third world will allow humankind to avoid total extinction? Leslie reminds us that some effects of pollution are truly global and might indeed kill us all. Extensive detail, with literature references, will permit follow-up-though apocalyptic visions from the Bible, and God's destructive side ("I will destroy man," Genesis 6; "Behold, the Lord maketh the earth empty, and maketh it waste," Isaiah 24) are not mentioned, despite their having been influential on human behaviour.

In view of the gloomy emphasis and repeated assessments of high risk in case after case, it is a relief to read that the author "feels inclined to say that the probability of the human race avoiding extinction for the next five centuries is encouragingly high, perhaps as high as 70 per cent". After much severe discussion of risk assessment, estimation of probabilities of extinction, and mathematical talk about Bayesian analysis, Leslie has elected not to combine the separate probabilities mathematically on the grounds that the "estimated total risk of Doom Soon cannot possibly exceed 100 per cent". Yet, in the presence of several risks of death, each with its own probability distribution, one can certainly calculate the year-by-year combined probability of death; the mean survival time, or expectation of life, is a familiar and convenient way of characterising the composite probability distribution. When the author sidesteps standard actuarial methodology, telling us merely what he feels inclined to say, he fails to do justice to the preparatory emphasis on technical risk assessment.

Running through the solid material on the main risks are two scary theories, the anthropic principle and the doomsday argument. Brandon Carter's anthropic principle of 1974 was that "what we can expect to observe must be restricted by the conditions necessary for our presence as observers". The doomsday argument is less crisply enunciated. It starts with the supposition that "many thousand intelligent races, all of about the same size, had been more or less bound to evolve in our universe". It asks: "We couldn't at all expect to be in the very earliest, could we?" It asserts that: "We couldn't at all expect to be in the first 0.1 per cent, let alone the first 0.001 per cent, of all humans who would ever have observed their positions in time." The inference is Doom Soon.

51¹ú²úÊÓƵ

Here is an illustration of how Leslie associates the two theories. "Planets are few, and if intelligent life could arise only on planets, and if intelligent life could not arise inside stars or in interstellar space, then you and I would have to find either that we were on a planet or that our ancestors had been." This proposition has a familiar syllogistic structure but one notices some fuzzy elements. The conclusion would be the same if planets were not few but many, so the first premise is omissible; the third is too. The proposition reduces to this: "If intelligent life could arise only on planets, then intelligent creatures would have to find that either they were on a planet or that their ancestors had been."

This weak reasoning is offered in support of Carter's anthropic principle, which "can often help to persuade us that our position in space and in time is in fact unusual". Surely the fact is that we do not at present know whether our situation is "unusual"; is not that the motivation for the efforts now underway to find other intelligent life?

51¹ú²úÊÓƵ

Another quibble: if Carter's principle often helps to persuade, and sometimes does not, then it is on balance not a persuasive principle. Other lines of "reasoning" are offered as bolsters but a logician would surely find Leslie's support of the anthropic principle damned by faint logic.

The principles that are well known are rather terse. Pascal's principle: "In a fluid at rest in a closed container, a pressure change in one part is transmitted without loss to every portion of the fluid and to the walls of the container." Archimedes' principle: "Any body completely or partially submerged in a fluid at rest is acted upon by an upward force equal to the weight of the fluid displaced." The Columbus principle: "An explorer putting to sea in search of new civilisations discovers peoples less advanced in the technology of ship-building." The Peter principle: "An employee in an organisation tends to rise to his own level of incompetence." We acquiesce in these principles insofar as they conform with observation.

The anthropic theory does not depend on observation, indeed it cannot be invalidated by observation. Scientific hypotheses that are in principle falsifiable by subsequent observation are highly regarded because, whether they are accepted or rejected, they lead to progress. A theory for which a test cannot be suggested, for example that there are other universes beyond the observable universe, is not esteemed as a part of science. In the past such theories have found a home in philosophy and, sure enough, Leslie entertains this one.

Human mini-extinctions are occurring now, each event exacerbated by the population explosion. Whether death is due to starvation, epidemic, pollution, or gunshot is secondary; in any case many of the dead go uncategorised to mass graves. The world population as of mid-1996 was over 5.7 billion, rising at 1.4 per cent per annum, a rate that is restrained by both birth control and, increasingly, by starvation and civil war. Further mini-extinctions, which are inevitable, will not be distributed evenly over the globe; societies in some areas will collapse to conditions differing little from those of the Stone Age, except that there will be plenty of scrap metal.

It seems pointless to care whether Leslie's theoretical arguments are unconvincing. Whether the human mind can respond to the need for institutions to curb the population rationally, as local collapses multiply, is what matters. That is the unknown risk to focus on and proclaim.

Ronald N. Bracewell is professor of electrical engineering emeritus, Stanford 51¹ú²úÊÓƵ, California.

51¹ú²úÊÓƵ