BY ADRIAN BERRY
Posted: May 14, 2008
In 1961, the meteorologist Edward Lorenz totally changed our thinking about the future – and indeed about the nature of existence – by the simple act of getting up from his desk to fetch a cup of coffee.
In his office at MIT he was using a computer to make a long-term weather forecast. Using twelve equations based on the relationships between temperature, pressure and wind speed, he relied on physical determinism to produce an absolutely accurate forecast.
After breaking off from his work for refreshment, he re-entered his data from an earlier printout. When he looked at the result he could barely believe his eyes. The new weather predictions were totally different from the old.
At first he thought his machine had malfunctioned. Then he realised the truth. In the original program the numbers had been stored to six decimal places. To save time he had rounded them off to three, assuming that this difference of one in a thousand would be inconsequential.
He had unwittingly discovered the spooky theory of Chaos, the great truth that when making a long-term prediction, no matter what the subject, there will always be tiny errors in the initial observation of conditions, and as time goes forward these will utterly overwhelm the accuracy of the forecast.
More accurate and faster machines have made no difference. Nor will any machine, however powerful and advanced, have any chance of doing so.
Today the biggest weather computer in the world, the European centre for Medium-range Weather Forecasting in Reading, does as many as 400 million calculations every second. It is fed 100 million separate weather measurements from around the world every day, and by processing data in three hours of continuous running it produces a ten day forecast. Yet beyond two or three days the forecasts are speculative, and beyond six or seven they are worthless.
Chaos completely repudiates the optimistic confidence of Newton who believed the Solar System ran like clockwork. And of his French counterpart the Marquis of Laplace who believed that there could exist a God-like intelligence who would know everything and could predict everything if only it could have access to an infinite amount of data. He imagined a super-intelligence . . . .
". . . which knew at a given instant all of the forces by which nature is animated, and the relative positions of all the objects, and which, if it were sufficiently powerful to analyse all this information, would include in one formula the movements of the most massive objects in the Universe and those of the lightest atoms. Nothing would be uncertain to it. The future, as the past, would be present to its eyes.''
However verbosely he might have expressed his opinion, it seems at first sight so obvious and so self-evidently correct that it is hard to imagine it being wrong. Surely, one reflects, the only reason that we do not understand everything is that have not yet found out everything. We lack only the tools with which to obtain complete knowledge.
But in fact the full truth about the Universe is not only unknown but unknowable. We will always be like the machine in Reading. However deeply we we try to predict the future we will always be misled by tiny errors whose effects grow until they produce infinitely ignorant answers.
Previous Adrian Berry
Apr Time Without End read more
Mar Ugly Light read more
Visit Adrian Berry's website at www.adrianberry.net He is Consulting Editor (Science) of the Daily Telegraph.