Sheri Markose was an invited speaker at The Economy as a Complex System at Institute of Mathematical Sciences (IMSc), Chennai, India

Novelty Production by ‘Smart’ Digital Agents: The Logical Foundations and Policy Design Implications of Complex Adaptive Systems

Abstract

Few will recognize that the logical foundations of complex adaptive systems (CAS) relate to novelty production by digital agents, which operate on encoded information and can exit from listable sets to produce new encoded objects that lie outside all machine listable sets.  Many complexity and dynamical systems theorists regard non-linear chaotic dynamics as being ‘complex’ and eschew the so called Type 4 Dynamics of the Wolfram-Chomsky Schema associated with novelty production by digital agents.  The Markose (2017, see below) thesis is that intelligence having originated from digital operations (viz operations on encoded information, think genetic code) associated with the immune system that deals with hostile agents, has the same computational principles governing social cognition with a parallel offline simulation system (aka mirror neurons ) and the capacity to recognize negation or, in short, hackers. She gives the formal conditions that permit ‘smart’ digital agents to trigger an arms race in innovations.  She shows how Gödel logic and the capacity of digital agents to encode the Gödel sentence are instrumental in genomic novelty and human  proteanism.  The encoding of  the Liar with the Contrarian agent implementing the negation of the predictable outcome of a code, is the trigger for a perpetual arms race of  innovation based undecidable structure changing dynamics wherein predictable formal rules will suffer demise. This has a long provenance in Kantian principles based political economy on the necessity of end-neutral coercive laws, and in the more recently and intuitively argued positions of the Lucas Critique, Goodhart’s Law on the failure of policy and Brian Arthur’s insights into the role of the Contrarian in generating heterogeniety and endogeneous boom-bust cycles. To prevent system failure from predictable rules, regulatory arbitrage from innovative rule breaking have to be eliminated.  If not, regulators have to abandon fixed rules and be prepared to co-evolve with rule breakers with horizon scanners in place.  Clearly, in the recent GFC,  authorities enamoured by pre-commitment to fixed rules, eliminated ‘surprises’ on their own part withdrew from the co-evolutonary arms race and gave the Liar qua rule breakers full reign to game the system. Such examples are given for the need of a CAS perspective in political economy and game theory.

Video link