Abstract:

The first part of the talk describes the paradox of robustness. When a system robustly corrects errors in its components, the direct selective pressure on those components declines. Weakened selection on components causes them to become less reliable, maintain more genetic variability, or change in design. Loosened constraint on component-level design allows complexity to creep in arbitrarily, without natural selection to remove unnecessary elaboration. New forms of organismal complexity appear, seemingly without any clear logic. In the second part, I link the paradox of robustness to similar ideas developed by others, such as the theory of constructive neutral evolution. In the third part, I unify previous ideas into a more general perspective. I then link the paradox of robustness to the observation that, in evolution, intermediate stages in development tend to be more strongly constrained than earlier or later stages. This leads us to John Doyle’s discussion of the universal architecture of complex robustly designed systems in engineering and in biology, in which there is an important distinction between hardware and software layers. I link the observations about evolutionary constraints in development to Doyle’s hardware vs software distinction. Finally, I argue that constructive neutral evolution is mainly about genomic hardware, in which neutrality likely dominates, whereas the broader paradox of robustness also includes functional components in the software layer, in which nonneutral economic costs and benefits likely dominate.

This talk will be given via Zoom. Please contact Thomas Pradeu if you’d like to join us for this seminar.