SlideShare a Scribd company logo

Wrangling Complex Systems

•
•
FoCAS Initiative
FoCAS Initiative

This document discusses how lifelike behaviors can emerge in complex physical systems outside of biological organisms. It provides examples from fields like thermodynamics, astrophysics, chemistry, evolution, technology, and biogeophysics that exhibit self-replication, inference, and other behaviors we typically associate with living things. The author argues that we should treat these complex systems as if they were biological and investigate the mathematical similarities between living and non-living systems more rigorously, rather than assuming the differences. While acknowledging key differences, the author suggests different explanatory stances can be taken and that concepts from biology may help understand complex systems like economies and climate that seem to have behaviors of their own.

1 of 9
Download to read offline
Wrangling Complex Systems? Simon McGregor Centre For Cognitive Science, University Of Sussex
Overview • Some examples of the ubiquity and depth of life-like behaviour in the physical world. • A short argument that we should try treating complex systems as though they were biological organisms. • Some reflections on explanatory stances.
Examples In Natural Physical Systems • Thermodynamics: – Self-replication intrinsically fuels entropy production (England, 2013) – Any ergodic system with a Markov blanket provably conducts Bayesian inference on its environment (Friston, 2013) • Astrophysics: – Self-replicating vortex structures have been observed in simulated star formation (Marcus et al. 2013) • Chemistry: – Reaction-diffusion spots can exhibit precariousness, chemotaxis and heritable variation (Virgo et al. 2013) – Reaction networks can conduct approximate Bayesian inference (McGregor et al. 2012)
Examples In Biosphere Systems • Evolution: – The replicator equation in population genetics has the same mathematical form as Bayesian inference (Harper, 2009) – Evolution implements processes resembling neural networks (Watson et al., 2015) • Technology: – Human artefacts can be seen as actors interacting equally with humans (Latour, …) – Human artefacts can be seen as autopoietic (McGregor & Virgo, 2009) • Biogeophysics: – The entire planet Earth seems to self-regulate (Lovelock, …) • And more… – Social insect colonies are sometimes argued to be super-organisms – Human social institutions can appear to have their own agenda (so much so that companies are legally people)
Main Argument The prevailing mentality is that “lifelike” behaviour outside of biological organisms is a weak metaphor at best. • But in fact there seem to be mathematical isomorphisms. • We won’t know how strong the similarities really are until we stop presuming and investigate rigorously.
(Caveat: All Non-Human Systems Are Alien) This is not a licence to anthropomorphise. Our models of cognition shouldn’t be based on human peculiarities. • There are big differences even between human minds. • Organisms like insects are demonstrably not tiny people. • “Para-organisms” (like genetic populations) are more different still. ?
Dennett’s Three Stances Daniel Dennett points out that we can take different explanatory “stances” towards the same physical system: • Physical stance - understanding in terms of mechanisms; • Design stance - understanding in terms of function; • Intentional stance - understanding in terms of agency.
Physical & Intentional Systems Intentional Stance Unhelpful Intentional Stance Helpful Physical Stance Easy To Apply Physical Stance Hard To Apply Philosophers Orbits Arbitrary Fluid Dynamics Hurricanes? Bacteria? (But of course, some systems are probably just hard to understand.) We are making progress on understanding the mechanics of certain biological systems.Immune Systems? And understanding the “cognitive” abilities of others. Population Genetics? We tend to ignore the life-like properties of natural non- biological systems. Homing Missiles? We design certain artifacts to mimic biological phenomena. For very simple systems, the physical stance is great. For complex biological systems, the intentional stance is great. In all the natural world, is there really nothing in between biology and simple mechanics?
Key Questions Complex systems like economies and the climate are not like vehicles we can “steer”. • They seem to have a “life of their own” – are they more like animals we need to “wrangle”? We should continue using tools from the physical and mathematical sciences to understand them. • But, can we also use concepts from biology (including ethology) and husbandry? • We don’t know, because the question is taboo.

More Related Content

Wrangling Complex Systems

  • 1. Wrangling Complex Systems? Simon McGregor Centre For Cognitive Science, University Of Sussex
  • 2. Overview • Some examples of the ubiquity and depth of life-like behaviour in the physical world. • A short argument that we should try treating complex systems as though they were biological organisms. • Some reflections on explanatory stances.
  • 3. Examples In Natural Physical Systems • Thermodynamics: – Self-replication intrinsically fuels entropy production (England, 2013) – Any ergodic system with a Markov blanket provably conducts Bayesian inference on its environment (Friston, 2013) • Astrophysics: – Self-replicating vortex structures have been observed in simulated star formation (Marcus et al. 2013) • Chemistry: – Reaction-diffusion spots can exhibit precariousness, chemotaxis and heritable variation (Virgo et al. 2013) – Reaction networks can conduct approximate Bayesian inference (McGregor et al. 2012)
  • 4. Examples In Biosphere Systems • Evolution: – The replicator equation in population genetics has the same mathematical form as Bayesian inference (Harper, 2009) – Evolution implements processes resembling neural networks (Watson et al., 2015) • Technology: – Human artefacts can be seen as actors interacting equally with humans (Latour, …) – Human artefacts can be seen as autopoietic (McGregor & Virgo, 2009) • Biogeophysics: – The entire planet Earth seems to self-regulate (Lovelock, …) • And more… – Social insect colonies are sometimes argued to be super-organisms – Human social institutions can appear to have their own agenda (so much so that companies are legally people)
  • 5. Main Argument The prevailing mentality is that “lifelike” behaviour outside of biological organisms is a weak metaphor at best. • But in fact there seem to be mathematical isomorphisms. • We won’t know how strong the similarities really are until we stop presuming and investigate rigorously.
  • 6. (Caveat: All Non-Human Systems Are Alien) This is not a licence to anthropomorphise. Our models of cognition shouldn’t be based on human peculiarities. • There are big differences even between human minds. • Organisms like insects are demonstrably not tiny people. • “Para-organisms” (like genetic populations) are more different still. ?
  • 7. Dennett’s Three Stances Daniel Dennett points out that we can take different explanatory “stances” towards the same physical system: • Physical stance - understanding in terms of mechanisms; • Design stance - understanding in terms of function; • Intentional stance - understanding in terms of agency.
  • 8. Physical & Intentional Systems Intentional Stance Unhelpful Intentional Stance Helpful Physical Stance Easy To Apply Physical Stance Hard To Apply Philosophers Orbits Arbitrary Fluid Dynamics Hurricanes? Bacteria? (But of course, some systems are probably just hard to understand.) We are making progress on understanding the mechanics of certain biological systems.Immune Systems? And understanding the “cognitive” abilities of others. Population Genetics? We tend to ignore the life-like properties of natural non- biological systems. Homing Missiles? We design certain artifacts to mimic biological phenomena. For very simple systems, the physical stance is great. For complex biological systems, the intentional stance is great. In all the natural world, is there really nothing in between biology and simple mechanics?
  • 9. Key Questions Complex systems like economies and the climate are not like vehicles we can “steer”. • They seem to have a “life of their own” – are they more like animals we need to “wrangle”? We should continue using tools from the physical and mathematical sciences to understand them. • But, can we also use concepts from biology (including ethology) and husbandry? • We don’t know, because the question is taboo.