img src="https://upload.wikimedia.org/wikipedia/commons/thumb/a/a7/Mandelbrot_set_image.png/220px-Mandelbrot_set_image.png" alt="Mandelbrot set (Wikipedia, useful sense." Rees's explanations are a bit clumsy in places - for instance, he confuses 'complicated' with 'complex' once or twice, which is a rooky mistake, and his example of the Mandelbrot Set as 'incomprehensible' is not convincing and rather misses the point about why emergent systems cannot be usefully explained by reductionism (it's about different kinds of causality, not about complicated patterns) - but he generally provides a good introduction to the issues.

These are well-trodden themes that most complexity theorists have addressed in far more depth and detail, and that usually appear in the first chapter of any introductory book in the field, but it is good to see someone who, from his job title, might seem to be an archetypal reductive scientist (he's an astrophysicist) challenging some of the basic tenets of his discipline.

Perhaps my favourite works on the subject are John Holland's Signals and Boundaries, which is a brilliant, if incomplete, attempt to develop a rigorous theory to explain and describe complex adaptive systems, and Stuart Kauffman's flawed but stunning Reinventing the Sacred, which (with very patchy success) attempts to bridge science and religious belief but that, in the process, brilliantly and repeatedly proves, from many different angles, the impossibility of reductive science explaining or predicting more than an infinitesimal fraction of what actually matters in the universe. Both books are very heavy reading, but very rewarding.