Book Review: A New Kind of Science
By Rutger-Jan van Spaandonk
This month marks the first anniversary of the release of “A New Kind of Science”, a book written over a ten-year period by Stephen Wolfram, and which has attracted a lot of attention in (semi-) scientific circles. You seem to either love it or hate it.
It is not often that you come across a book that leaves you utterly confused, but this one certainly does. Not only because most users will get lost in the intricate scientific details, but also because one is tempted to think that if (and that is a big if) Stephen is right, we will soon solve all the mysteries left in the universe, including that of the universe itself.
The book is a pleasure to read. First and foremost, because the author put it together with attention to detail and aesthetics: he takes care to mention that “the book was printed on 5-pound Finch VHF paper on a sheet-fed press. It was imaged directly to plates at 2400dpi, with halftones rendered using a 175-line screen with round dots angled at 45o.”
It is not easy to read, though, and also very long: 1280 pages and 583,313 words. Main text: 227,580, notes 283,751 – yes, that’s why I mention the stats; there are more words of notes than of main text, and at one stage the author wanted to add notes to the notes.
And all this verbiage to show that simple rules can drive systems that exhibit very complex behaviour. That, in essence, is the main tenet of his book. However complicated the environment might seem to us, Stephen tries to convince us that even the most intricate structures and processes can be the result of an infinite loop of very simple rules. And from this follows, in his opinion, that even the most complex structure of all, our universe, came about because of a couple of lines of programming.
He explains his theories with the help of cellular automata. A cellular automaton in its simplest format consist of a matrix in which the first line of each cell is coloured either black or white (or different shades of grey or even colours in more complicated automatons), and in which a definite rule determines the colour of cells in subsequent lines.
By changing the instruction sets only incrementally every time, Stephen found that, while in the majority of cases the rules resulted in symmetric, or at least neat patterns, in some instances they resulted in patterns without any recurring structure, and that never settled down in a steady state. Traditional mathematics and statistics are unable to capture such chaos in a formula, and thus should lead us to believe that serious complexity is at play. However, the underlying simple instruction set shows that this is not the case.
This one observation sets off a whole gamut of other experiments with more complicated 2D and 3D cellular automata, as well as analyses of biological phenomena, and his conclusions are always the same: simplicity can breed complexity.
Many colleagues and observers have belittled his work as an attempt to reduce God’s creation to a couple of lines of code, and that it is merely the result of his infatuation with computers. As one reviewer commented, “So might a carpenter, looking at the moon, suppose that it is made of wood”, to which Stephen retorted in a recent e-mail to devotees that “introducing a major new intellectual direction is never easy”. He is convinced that his discoveries place him in the same league as Newton, Darwin and Einstein. Interestingly enough, some of his colleagues seem to concur.
Stephen has the CV to match his oversized claims. He is a renowned scientist, published his first physics paper at the age of 15, went on to earn a PhD in physics from the California Institute of Technology at the age of 20 and then, to top it off, won the prestigious MacArthur Foundation grant a year later in 1981. After an unrewarding stint in academics (he got for example into a fight over the intellectual property rights of his many findings) he started Wolfram Science, the software company that developed and markets the leading mathematical software-programming environment, Mathematica.
But despite these credentials, some of his claims are hard to swallow. For example, he denounces evolution theory – in one interview famously asserting that the vast diversity of zebra-stripe patterns is not the result of natural selection that confers a survival advantage, but rather results from a simple set of colouring instructions. Also (and this one is for the die-hards), he attempts to prove the Second Law of Thermodynamics wrong by showing that not all complex systems are irreversible – something which in certain circles is considered outright heresy.
However, as the Economist rightly noted in its review, “A book with an interesting new theory does not have to be right for it to be worth reading”.
More information “A New Kind of Science” and its author can be found on www.wolframscience.com.
Rutger-Jan van Spaandonk is the founding director of FutureForesight Group, a boutique strategy consultancy. He can be reached via firstname.lastname@example.org.