It’s unfortunate that the phrase irreducible complexity is most commonly known nowadays as the failed idea of creationists arguing against neo-Darwinian evolution. I’m greatly over-simplifying the story here because today’s blog post is not about this popular “controversy” between science and religion.
Instead I’m going to discuss some thoughts from reading Robert Rosen’s Essays on Life Itself, in particular focusing on the relationship between physics and biology. I’d like to think I have special insight as a chemist sitting between these two fields, but maybe I don’t, and maybe it’s an open question whether or not these fields are sequentially linked. The Venn diagram I have used with my students to discuss definitions of life may not be accurate either.
I discovered Rosen’s work after stumbling on Mikulecky’s definition of complexity, which implicitly contains the idea of the irreducible. This is distinguished from something that is “merely” complicated, which is reducible to its parts even if the disentanglement process is super-complicated. Rosen begins with Schrodinger’s What is Life? and argues that most contemporary physicists and biologists ignore a deeper fundamental question about the relationship between the two areas.
Schrodinger argued that we might need a “new physics” to characterize life. In the present reductive model, familiar to scientists and students of science, biology can be implicitly reduced into chemistry, which can then be reduced to physics. Living organisms are a subset of the larger “world” of non-organismic materials and forces; they are a special and perhaps even rare case – rare in the physical universe because of the special Goldilocks conditions of planet Earth. Rosen turns this idea around, making the enigmatic proposal that “organisms are more general than the non-organisms in the old physics, and that their apparent rarity is only an artifact of sampling.”
An argument made against modern-day proponents of irreducible complexity from scientific creationists affiliated with the IDEA institute and the Intelligent Design movement is that they are reviving the failed idea of vitalism, and attempting to smuggle in God as its sustainer – thus we can’t explain life without God, the author and creator of life. I find this view theologically impoverished; it seems like a God-of-the-gaps argument that relegates the supreme being to a tinkerer of parts within the “old physics”.
It’s hard for us to get out of the mindset that life seems rare and special. Perhaps that’s because we modern folk think in terms of material substances rather than functionality. To think of functionality smacks of teleology and the smuggling in of purpose. We the scientific-literates focus on the syntax but not on the meaning. Information is reduced to bits, bytes, and Shannon entropy. We’re all about counting backwards and forwards in precise discrete addition and subtraction, each piece independent from each other. We do acknowledge mysteries such as the nature of energy – we don’t know what it is but we can count it – but that’s only because we haven’t gotten through the process of disentangling its complications.
Rosen has many analogies as to why this sort of thinking is flawed. Here’s my favorite one that serves as an analogy to why the “old physics” might be a subset of unruly biology, rather than the other way around of subsuming biology within the larger realm of physics.
“This kind of argument rests on a confusion about, or equivocation on, the term rare, and identifying it with special. An analogous argument could have been made in a humble area like arithmetic, at a time when most numbers of ordinary experience were rational numbers, the ratios of integers. Suddenly a number such as π (pi) shows up, which is not rational. It is clearly rare, in the context of the rational numbers we think we know. But there is an enormous world of ‘new arithmetic’ locked up in π, arising from the fact that is much too general to be rational. This greater generality does not mean there is anything vitalistic about π, or even anything unarithmetic about it; the only vitalistic aspects show up in the mistaken belief that ‘number’ means ‘rational number’.”
There are many other thought-provoking ideas just in the first fifty pages of Essays on Life Itself, I could spend the rest of my lifetime exploring them in detail. As a reminder to myself, since my blog is a cyborgian extension of my otherwise poor memory, I will briefly note a few of these here in the hope that I will get back to them at some point.
Rosen grounds his definition of complexity by thinking about the non-commutability of relationships, and that analysis and synthesis are not exactly opposite processes. I’m reminded of when I discuss Hermitian operators and Heisenberg’s Uncertainty Principle in my quantum chemistry class, and it seems like there’s some quantum astrology going on. Rosen regularly refers to the difference between inertial and gravitational mass, with a hint to why the N-body problem exists. Most of us (me included) don’t think about the difference between these two types of mass because they are numerically equivalent. But they are not the same.
This brings me to Rosen’s thoughts on biomimesis and simulation. Can mind be reduced to machine? If an artificial intelligence passes the Turing test, does it mean it is alive? Does it matter if we humans can’t tell the difference? Rosen marvelously connects science and magic as he ponders the nature of objectivity: “[Mimesis] is animated by an idea that things that behave enough alike are alike, interchangeable. In biology, such ideas go back to the earliest historical times and, indeed, are intimately related to the archaic concept of sympathies as in ‘sympathetic magic’ … if we can produce a system manifesting enough properties of an organism, i.e., it behaves enough like an organism, than that system will be an organism… Indeed, mimesis treats such individual behaviors as a reductionist treats atoms, as syntactic units… [that] can be reduced and then recombined...”
I’m reminded of the parallel work in origin-of-life research that studies artificial life, some through carefully designed biophysical systems such following GARD, but more often through simulation – with the underlying assumption that software and hardware can be treated independently from each other. As a computational chemist, I also wonder if my particular approach to tackling origin-of-life questions is doomed to failure, given that one aspect of the complexity of biology, according to Rosen, is that it is non-computable. The irreducibility of life makes it so. I’m also reminded of one tricky aspect that most students don’t notice on the first day of my introductory chemistry class when we discuss the definitions of element, atom, molecule, compound. There’s a certain circularity to those definitions, and I feel a slight pang of helplessness every time we go through it, because the occasional student (one who has had little background in chemistry) will for good reason that they cannot easily articulate find the definitions confusing, and I will provide pat answers in class so we can move on.
Rosen employs notions of graph theory as he ponders how to think relationally about forces acting on materials which are themselves sources of forces. I’ve been thinking along these lines in my current research as I struggle to conceive the limitations of differential equations in studying the kinetics of autocatalytic systems and how they evolve. There’s a strange (by which I mean I don’t understand it) recursive relationship going on, that might lead to infinite regress on the one hand – you might need a larger system to explain what’s going on in your subsystem – and yet “infinity is not the same as large finite” as Rosen states, and it boggles my mind how to think about this. Even more mind-boggling, Rosen suggests a possible way out through replication as some way to stabilize open systems such that somehow the recursive loops close back into some sort of discrete entity (an organism!) with boundaries that remain fuzzy.
Murkily, rather than clearly, the approach to understanding complexity will not adequately proceed through pure reductionism. Reductionist models can give us partial understanding of the part, not the whole, and we need to remember the limitations of our model whereby we’ve conveniently hidden or ignored the irreducible parts. To grasp that whole, maybe a “leap of faith” is required. Not because of small gaps because the “old physics” is starting to see widening chasms as it explores complexity theory, but large gaps that Schrodinger began to ponder when he suggested a new physics is needed. Those lessons may come from life itself.
No comments:
Post a Comment