This is a book of enormous ambition and should be read not just by those interested in the relationship between the mechanical philosophy and the life sciences, but also by those interested in the way modern science has acquired its distinctive image and character in modern civilization. Riskin covers her topic from the Middle Ages through to Erwin Schrödinger's What Is Life? (1944), and beyond to Pixar's Inside Out (2015). Along the way, there are chapters on the earliest mechanical moving figures in medieval churches, the development of elaborate clocks with moving figures, hydraulic garden machinery, the earliest ‘androids’ in the seventeenth century, the first robots in twentieth-century fiction (but, disappointingly, no mention of Robbie or Marvin) and the development of cybernetics and artificial intelligence. These chapters in themselves would have made a highly useful book on perennial attempts to imitate life, but Riskin's real concern is to do much more than that. Her main concern is with the phenomenon of what she calls ‘agency’ in the natural world, and what she sees as the overriding concern of scientists through the ages to exclude this ‘agency’ from nature.

What she means by ‘agency’ is pretty clear, although it is by no means sharply and unequivocally established (perhaps deliberately). Essentially, she is interested in the way scientists (even those specializing in the biological sciences) have come to insist that their subjects have no ability to direct or control their own behaviour. An early example of scientific thinking which seemed to run counter to this shibboleth was Kepler's second law of planetary motion. It seemed to imply that a planet must know how fast it is going and be capable of adjusting its speed to make sure it complied with the law. Alternative accounts were immediately introduced to avoid this embarrassment but none were as accurate as Kepler's seemingly anthropomorphizing law. Newton's universal principle of gravitation removed the difficulty, and everyone gave a sigh of relief as planets were shown to have no ‘agency’ in Riskin's sense of the term.

Much of the discussion is devoted to the difference between Lamarckism and Darwinism, from their respective originators to modern-day Weismannian neo-Darwinists, such as Richard Dawkins and Daniel Dennett. Perhaps deliberately (again), Riskin seems to prefer to collude with all the critics of Lamarck in accepting him as someone who allowed agency into biology. But, of course, he never upheld (contrary to what Cuvier, Darwin and almost everyone else has said) what Darwin dismissed as a ‘willing’ doctrine, a belief that creatures could wish themselves to have a handy long neck, or a useful hump for storing water. In spite of all the ‘bad press’, Lamarck's conformity to the scientific norm of excluding ‘agency’ comes through loud and clear to anyone who actually reads him without having an axe to grind (as Cuvier, Darwin, etc., etc., definitely did). But, presumably, it suited Riskin's argument to have someone she could use to represent an alternative view – someone who looked as though he had allowed ‘agency’ into his scientific theorizing.

Which brings us to the most remarkable and challenging part of Riskin's book. The tenth and final chapter of her book is called ‘History matters’, and she seems to want to do two things in this chapter. First, she wants to suggest that current science, in spite of the rigid strictures of its practitioners, might actually benefit from allowing that ‘agency’ might be a feature of natural things and might affect their behaviour. Second, she insists that a closer attention to the history of their subject will enable contemporary scientists to see the light, and will make them realize that their strictures against ‘agency’ are ill-founded. ‘This final chapter shows’, she writes, ‘that the history of these competing forms of science is profoundly relevant to current debates in artificial intelligence, cognitive science, and evolutionary biology’ (pp. 337–338 – I told you it was ambitious!). ‘By recognizing the historical roots of the almost unanimous conviction . . . that science must not attribute any kind of agency to natural phenomena . . . we can measure the limits of current scientific discussion, and possibly, think beyond them’ (pp. 338–339). What Riskin has in mind is what she sees as the undeniable historical fact that the prejudice against ‘agency’ in nature was first introduced as a way of proving the existence, or defending the idea, of God. Accordingly, ‘Those who continue to endorse this ban on natural agency do not seem to recall that its original justification was as much theological as scientific’ (p. 347; my emphasis on ‘recall’ – history is made to matter). She makes the point more forcefully a little later:

So the implication seems to be that if Dawkins and Dennett knew their history they would run screaming from their current position, recognizing it as one that has its roots in theology, and would champion Lamarck as a more thoroughbred, dyed-in-the-wool, scientist!

I feel bound to say that, for my own part, I wondered as I read this book why Riskin had chosen to present her history in what I took to be an unnecessarily convoluted, and sometimes confusing, way. It was only when I arrived at the final chapter that I realized that her book was, like Darwin's Origin, ‘one long argument’. I disagree with her argument (and I disagree with her version of history), but I really feel that all serious historians of science should read it for themselves and make up their own minds about it.