On the highly flaky origins of the humble mousetrap

In his earlier book, Nine Crazy Ideas in Science , Robert Ehrlich enjoyed mystifying and demystifying his audience by showing them how some apparently crazy theories were not so crazy after all. He rated them from zero to four on his "cuckoo" scale. For these "eight preposterous propositions", he has a new "flakiness" scale. An idea is flaky, Ehrlich writes, when it lacks "in empirical evidence, or internal consistency".

Quantum physics, for instance, is as crazy as a congress of mad hatters in a meadow of March hares (Ehrlich does not word it like that), but it is not flaky: there is empirical evidence, and it is internally consistent. After these preliminaries, Ehrlich equips the reader with a verbal sword of "ten questions to ask in judging whether A really causes B". So, off we go to cleave asunder the preposterous among eight hand-picked propositions.

Heading straight to proposition three - that people are getting smarter - Ehrlich gives a few pages of his attention to New Zealand philosopher James Flynn who claims that, well, we are getting smarter.

That's nice, but... what is this diagram with the IQ scores of Belgians climbing from 73 in 1940 to 100 in 1995, in a perfectly straight line? And Britons' IQs rising from 93 in 1955 to 100 in 1965, also in a perfectly straight line but with no figures outside that period? Alarm bells should ring. The caption for the diagram rings even louder alarm bells: "The scores are normed (sic), only the slopes are meaningful." In plain English, the figures have been cooked. Moreover, if that trend is not a short-term aberration, if it is real, then Isaac Newton had the IQ of a retarded chimpanzee and Archimedes that of a cockroach. This fact does not escape Ehrlich: "Were our ancestors really idiots by our standards?" The empirical evidence is dodgy, the internal consistency is that of a ripe, smelly rat.

Full marks on the flakiness scale, then. So why does Ehrlich give only one flake to the proposition that we are getting smarter, "because it seems a bit more plausible, given the Flynn effect"? This is not internally consistent.

The reader is then asked to consider the following proposition - that intelligent design is a scientific alternative to evolution. Much of this chapter is occupied by Michael Behe's case for intelligent design (read: God-directed evolution), and the case against. Behe's argument amounts to "the components of a mousetrap have to be well matched and fit together precisely for the trap to trap mice, so show me how a mousetrap could evolve without design". The answer should be plain: "Show me a mousetrap that eats mice and copulates with other mousetraps." But, because Behe is a professor of biochemistry his argument must be taken seriously, and Ehrlich spends much time on it, his reflection culminating with "How to ascend Mount Mousetrap" (a three-dimensional space of "mousetrap fitness").

But evolution does not flow from the survival of the fittest. It flows from the demise of the terminally inept, or just plain unlucky. If their island had not been overrun by predatory apes armed with firesticks and with a taste for dodo flesh, dodos would still be thriving, as they had been for thousands of years. It was just sheer bad luck.

Behe's mousetrap could have evolved from anything, cricket balls or Rubik's cubes... which could have evolved into anything but mousetraps, in which case Behe would be clamouring "Ha! show me how this garlic press could have evolved without design!" That is post hoc argumentation, and is as meaningless as claiming divine intervention in the hand of bridge you were just dealt because there was only one chance in 4 sextillion of receiving those 13 cards in that precise order. It makes one's blood boil to see Ehrlich walk into Behe's trap instead of showing how it has evolved from a red herring.

Never fear, proposition seven - that a sugar pill can cure you - reduces the blood pressure. It can, insofar as belief can heal. But belief can also kill: Japanese and Chinese Americans die of heart disease in significantly greater numbers (13 per cent) on the fourth of the month ("four" is unlucky, being homophonous with "death"). The figure rises to per cent of those of this ethnic community living in California, an enormous deviation. Remembering those "ten questions to ask" in the introduction, you expect Ehrlich to ask "Can that really be? What about other causes of death or just car accidents? Do they corroborate this? And if not...?" But he does not, and he eventually gives his blessing to the sugar pill that cures or kills: zero flakes.

Fine. But then, in proposition eight - that we should worry about our cholesterol - Ehrlich should warn us that we might die of believing that cholesterol is bad for us; and cholesterol doomsayers should be admonished.

Chinese Californians die from believing that "four" means death, don't they? Somehow, all this lacks internal consistency. I felt short-cheesed.

Jacques B. M. Guy is a computer scientist interested in natural language understanding. He holds a PhD in linguistics from the Australian National University.