Mechanistic models [in physics]
Until quite recently, it was generally assumed that if one were able to get at the microscopic constituents of the universe they would look essentially like small-scale analogs of objects familiar from everyday life. And so, for example, the various models of atoms from the end of the 1800s and beginning of the 1900s were all based on familiar mechanical systems. But with the rise of quantum mechanics it came to be believed throughout mainstream physics that any true fundamental model must be abstract and mathematical—and never ultimately amenable to any kind of direct mechanistic description. Occasionally there have been mechanistic descriptions used—as in the parton and bag models, and various continuum models of high-energy collisions—but they have typically been viewed only as convenient rough approximations. (Feynman diagrams may also seem superficially mechanistic, but are really just representations of quite abstract mathematical formulas.) And indeed since at least the 1960s mechanistic models have tended to carry the stigma of uninformed amateur science.
With the rise of computers there began to be occasional discussion—though largely outside of mainstream science—that the universe might have a mechanism related to computers. Since the 1950s science fiction has sometimes featured the idea that the universe or some part of it—such as the Earth—could be an intentionally created computer, or that our perception of the universe could be based on a computer simulation. Starting in the 1950s a few computer scientists considered the idea that the universe might have components like a computer. Konrad Zuse suggested that it could be a continuous cellular automaton; Edward Fredkin an ordinary cellular automaton (compare page 1027). And over the past few decades—normally in the context of amateur science—there have been a steady stream of systems like cellular automata constructed to have elements reminiscent of observed particles or forces. From the point of view of mainstream physics, such models have usually seemed quite naive. And from what I say in the main text, no such literal mechanistic model can ever in the end realistically be expected to work. For if an ultimate model is going to be simple, then in a sense it cannot have room for all sorts of elements that are immediately recognizable in terms of everyday known physics. And instead I believe that what must happen relies on the phenomena discovered in this book—and involves the emergence of complex properties without any obvious underlying mechanistic set up. (Compare page 860.)