Norman Lee Johnson Chief Scientist Referentia System Inc. Honolulu Hawaii norman@SantaFe.edu

1. Complexity: A Diverse Description • Norman Lee Johnson • Chief Scientist • Referentia System Inc. • Honolulu Hawaii • norman@SantaFe.edu • http:// CollectiveScience.com ONR Workshop on Human Interactions in Irregular Warfare as a Complex System Apr 2011

2. My Background • Future of the internet • Self-organizing collectives • Diversity and collective intelligence • Finance applications • Effects of rapid change • Finance applications • Group identity dynamics • Coexistenceapplications • Leadership models • Bio-cyber analogies • Star Wars • Novel fusion device • Novel diesel engine • Hydrogen fuel program • P&G – Diapers • Large scale Epidemiology – Flu modeling • Biological threat reduction • Bio-Risk assessment • Cyber security

3. Counterinsurgency in Iraq:Theory and Practice, 2007 David Kilcullen Available online at: http://smallwarsjournal.com/documents/kilcullencoinbrief26sep07.ppt

4. Caveat: the logic of field observation in Iraq • Everyone sees Iraq differently, depending on when they served there, what they did, and where they worked. • The environment is highly complex, ambiguous and fluid • It is extremely hard to know what is happening – trying too hard to find out can get you killed…and so can not knowing • “Observer effect” and data corruption create uncertainty, and invite bias • Knowledge of Iraq is very time-specific and location-specific • Prediction in complex systems (like insurgencies) is mathematically impossible…but we can’t help ourselves, we do it anyway • Hence, observations from one time/place may or may not be applicable elsewhere, even in the same campaign in the same year: we must first understand the essentials of the environment, then determine whether analogous situations exist, before attempting to apply “lessons”.

5. Levels of Social complexity Identity, diverse, decentralized, collective survival and problem solving Collectively adaptable, self-organizing, emergent properties Collective: Memory, intelligence, deception, tools From a workshop on “The Evolution of Social Behavior” which covered a wide range of social organisms Example: All social organism when stressed are “programmed” to copy the behavior of others in the “organism” Individual: High intelligence, deception & emotions, tool making Individual Self-awareness & Consciousness

6. Complex Human Dynamics: Three Challenges • System Analysis • How different analysis perspectives can simplify the “complexity” of the system attributes or dynamics? • Behavioral-Social Models • How abstract models can help simplify the “complexity” of individuals components • OR provide relationships between components?

7. Gen William C. Westmoreland, COMUSMACV, MACV Directive 525-4, 17 September 1965 “Getting it” is not enough “[This] is a political as well as a military war…the ultimate goal is to regain the loyalty and cooperation of the people.” “It is abundantly clear that all political, military, economic and security (police) programs must be integrated in order to attain any kind of success” Understanding by leaders is not enough: everyone needs to understand, and we need a framework, doctrine, a system, processes and structures to enact this understanding.

8. Complex Human Dynamics: Three Challenges • System Analysis • How difference perspectives can simplify the “complexity” of the system attributes or dynamics? • Behavioral-Social Model • How abstract models can help simplify the “complexity” of individuals components OR provide relationships between components? • Tools for Decision Makers • How the above understandings lead to actionable knowledge for decision makers?

9. Complex Human Dynamics: Three Solutions • System Analysis • Connection between local  global • Behavioral-Social Model Features • Individual behavior models Tools for decision makers

10. Individual preference + Social drives + Options + Rationality = ? • Habitual repetition: • Classical conditioning theory (Pavlov), Operant conditioning theory (Skinner) • Individual optimization of decision: • Theory of reasoned action (Fishbein & Ajzen), Theory of planned behavior (Ajzen) • Socially aware: • Social comparison theory (Festinger), Group comparisons (Faucheux & Mascovici) • Social imitation: • Social learning theory (Bandura), Social impact theory (Latané), Theory of normative conduct (Cialdini, Kalgren & Reno) CONSUMAT model -- Marco Janssen & Wander Jager – Netherlands

11. What drives the changes? Historical comparison Satisfied Dissatisfied Certain Increased stress Uncertain

12. Individual Behavior + Network = Global Dynamics 1000 Consumers with the same behavioral tendency buying 10 products on a small-world network

13. Population of “Repeaters” - satisfied and certain Closest to “Rest state” Few products of equal distribution - highly stable

14. Population of “Imitators” - satisfied but uncertain Transitional individual => social state Few products of unequal distribution - highly stable

15. Population of “Deliberators” - dissatisfied but certain Closest to Homo Economicus High rationality, low social High volatility on all products

16. Population of “Comparers” - dissatisfied & uncertain Social and Rational Volatility over long times on few products But difficult to maintain - high energy state

17. Highly stable with sustained diversity High volatility “habitual” agent Homo Economicus Socially driven Social and Rational Highly stable - decreased diversity Longer time volatility - difficult to sustain

18. Complex Human Dynamics: Three Solutions • Behavioral-Social Model Features • Individual behavior models • Drivers & threshold transitions • Habitual behavior • Importance of habitual behavior & individual threshold transitions • System Analysis • Connection between local  global • simple models can lead to complex global behavior • Study of system thresholds • Tools for decision makers • Focus on system thresholds first • rather than quantifying the details between threshold states

19. Rat Studies of Maximum Carrying Capacity One “social” rule => Cooperative social structure Control group - no “rules” => Your worst nightmare NIMH psychologist John B. Calhoun, 1971 Both systems loaded to 2 1/2 times the optimal capacity. Social order system can carry 8 times the optimal capacity before going over the threshold.

20. Simple Ant Foraging Model Using NetLogo • Collective information • Evaporation • Diffusion • Agent internal state: • Current direction • Have food? • Three rules of action: • Carry food • Drop food • Search Nest n“Productive men” n“Salaried men” nInnovator nCollective structure Food supply • Key concepts: • Emergence, Productivity, Diversity, Structure

21. Quantified Environmental Change Infinite source moves at a fixed radius and fixed angular velocity

22. Slowly changing environment Productivity is only slightly less than an unchanging source Herd effect allows for quick utilization of new resource location Innovators are important at all times by sustaining optimal performance of the collective

23. Effect of Rate of Change on System Development Food Production Rate

24. Complex Human Dynamics: Three Solutions • Behavioral-Social Model Features • Individual behavior models • Drivers & threshold transitions • Habitual behavior • Emergent problem solving • System Analysis • Connection between local  global • Study of system thresholds • Developmental perspective • Emergent behavior via multi-level analysis • Tools for decision makers • Focus on system threshold changes first

25. Structural Efficiency - Boom and Bust Lower average production crash avoidance strategy Greater minimums and maximum when compared to extreme rates! Bust is proceeded by increased production

26. For the slowest rate of change 2500 2000 transient structure 1500 Total production(units of food) 1000 sustained structure 500 0 0 500 1000 1500 2000 2500 3000 Time(time units) Combination of Sustained Structure and Change How does the retention of structure change the collective response? Suggests that fixed evolutionary adaptations lead to inefficiencies in the presence of even small rates of change What would be the effect of a faster worker? What would be the effect of mass communication?

27. Original observations for chemical systems Equilibrium states are an attractor for non-equilibrium states. System near equilibrium cannot evolve spontaneously to generate spatial−temporal (dissipative) structures As the system is driven far from equilibrium, it may become unstable and generate spatial−temporal structure from nonlinear kinetic processes associated with flows of matter and energy. The possibility of new structures is determined by the system size. Bifurcation introduces “history” into the model (trajectories are replaced by processes). Every description of a system which has bifurcations will imply both deterministic and probabilistic elements. Generalized observations Perturbed systems will return to their normal state. The mechanisms for permanent change are not accessible to systems near equilibrium. Bang on a system hard enough, existing structures can be replaced by new structures. The evolution of new structures are limited by system size. Multiple outcomes change certainty into probabilities, and require a fundamentally different approach. Prigogine’s Laws of stasis, change and evolution:

28. Complex Human Dynamics: Three Solutions • Behavioral-Social Model Features • Individual behavior models • Drivers & threshold transitions • Habitual behavior • Emergent problem solving • Individual and collective structure • System Analysis • Connection between local  global • Study of system thresholds • Developmental perspective • Emergent behavior via multi-level analysis • Optimization vs. robustness • Tools for decision makers • Focus on system threshold changes first

29. Structure in a system increases over timefor decentralized, self-organizing collectives (nature, societies, technologies) • Structure declines because the number of new rules are limited by past rules. Structure (e.g., the rules required to “run” the system) Structure increases rapidly as components build structure together Structure increases first by components developing structure Time

30. The Structure of Structures Structures direct the evolution of the system by creating and limiting potential options Their definition depends on the time constant of exogenous/endogenous change.

31. Options around Structure also change Options are the free choices both created and limited by the structure (example: the rules of chess create an “environment” where many options are possible - while also limiting what choices are available) Options are reduced as more structure restricts options Options Options are greatest when structure connects the components Structure (the rules required to “run” the system) Little initial structure means few Options Time These ideas are captured by researchers studying “infodynamics”

32. Difference between Options and Diversity • Diversity is the the unique variety in the system • Options are when the diversity has multiple expressions of differences, often expressed as multiple connectivity in the network Diversity may be very high but Options are low Options Diversity and Options are high Structure Diversity and Options are low Time These ideas are captured by researchers studying “infodynamics”

33. Adaptability and Robustness of System • Robustness is a system-level ability to sustain performance in the presence of change • Adaptability is a component level ability to accommodate change Not robust or adaptable, but existing components can be rearranged for new features Options Both adaptable and robust Structure Robustness achieved by component replacement Diversity Time These ideas are captured by researchers studying “infodynamics”

34. Collective Response to Environmental Change Stages in Development Rate of Environmental Change

35. Why Care about Structure-Options? Studies of thresholds in structure: • Prigogine’s Laws of Stasis, Change and Evolution • Joseph Schumpeter’s Creative Destruction • Foster and Kaplan "Creative Destruction: Why Companies that are Built to Last Underperform the Market - And how to Successfully Transform Them”, 2001 • John Padgett life’s work on innovation in the Florentine (and world) finance system • Dynamic “structural” thresholds do the same

36. Complex Human Dynamics: Three Solutions • Behavioral-Social Model Features • Individual behavior models • Drivers & threshold transitions • Habitual behavior • Emergent problem solving • Individual and collective structure • System Analysis • Connection between local  global • Study of system thresholds • Developmental perspective • Multi-level analysis & emergence • Optimization vs. robustness • Interplay of structure and options • Tools for decision makers • Focus on system threshold changes first • Capture structure and options: • What has to be removed before change can occur • Diversity is not the same as options!

37. Remember article 15 “15. Do not try to do too much with your own hands. Better the Arabs do it tolerably than that you do it perfectly. It is their war, and you are to help them, not to win it for them. Actually, also, under the very odd conditions of Arabia,your practical work will not be as good as, perhaps, you think it is.” T.E. Lawrence, “Twenty-Seven Articles”, The Arab Bulletin, 20 August1917

38. Expert Performance in Finance Why can’t financial experts outperform the S&P 500 “collective” – good + bad – consistently? • Professional money managers fail to beat the S&P 500 at an average rate of 70% per year. • 90% trail the S&P over a 10-year period. • Over decades are only a few – Soros, Miller, …. • “These are the people who have more knowledge and more training than the vast majority of investors. And yet, neither the superior knowledge nor the superior experience helps them in the long run.” Bill Mann, TMFOtter

39. Ants Solving “HARD” problems Food Food Nest Nest • The ant colony (and individuals) finds the shortest path How does it work?

40. When individuals solve the maze again, they eliminate “extra” loops In “Learning” the maze, individuals create a diversity of experience.        But because a global perspective is missing, they cannot shorten their path. This is were diversity helps. End Start A Model for Solving Hard Problems • How can groups • > solve hard problems, • > without coordination, • > without cooperation, • > without selection? • The Maze has many solutions • > non-optimal and optimal. • Individuals • > Solve a maze • > Independently • > Same capability

41. Collective path Unlike in natural selection, no one individual is the fittest! How collectives find the Shortest path Paths of three ants

42. Using established information Ensemble (Averaged) Behavior 1.3 Using novice information, with two different collections 1.2 1.1 Normalized number of steps . Average Individual 1.0 0.9 0.8 0 5 10 15 20 Individuals in Collective Decision Performance correlates with high unique diversity

43. Expert Performance & Complexity Complexity Barrier Value of Collectives Where Experts Have Value Value of Experts Simple Complex Domain Complexity MichaelMauboussin - Legg Mason Capital Management

44. From a workshop on Complex Science for the Physician’s Alliance

45. Effect of Complexity in Stable Systems System goes to optimization via “collective” route “Complexity Barrier” requires Collective Solutions X Structure (the rules required to “run” the system) System goes to optimization via “expert” route time

46. Collective Error = Average Individual error minus Prediction Diversity “The Difference:
How the Power of Diversity Creates Better Groups, Firms, Schools, and Societies”

47. Collective Performance Value of Collectives Collective Error = Average Individual error minus Prediction Diversity Where Experts Have Value Value of Experts Simple Complex Domain MichaelMauboussin - Legg Mason Capital Management

48. Collectives in complex environments end • • • • • • • • begin         Options in infrastructure, societal structure, economies, etc. In complex domains: • People beginning points differ • Their final goals may differ • But local paths can overlay and find synergy

49. Why counterinsurgency is population-centric • This is not about being “nice” to the population, it is a hard-headed recognition of certain basic facts, to wit: • The enemy needs the people to act in certain ways (sympathy, acquiescence, silence, provocation) -- without this insurgents wither • The enemy is fluid; the population is fixed – therefore controlling the population is do-able, destroying the enemy is not • Being fluid, the enemy can control his loss rate and can never be eradicated by purely enemy-centric means (e.g. Vietnam VC losses) • In any given area, there are multiple threat groups but only one local population – the enemy may not be identifiable but the population is. Terrain-centric and enemy-centric actions are still vital and crucial to success. Enemy and Terrain still matter, but Population is the key.

50. Complex Human Dynamics: Three Solutions • Behavioral-Social Model Features • Individual behavior models • Drivers & threshold transitions • Habitual behavior • Emergent problem solving • Individual and collective structure • Conditions for synergy/conflict • Group identity models • System Analysis • Connection between local  global • Study of system thresholds • Developmental perspective • Multi-level analysis & emergence • Optimization vs. robustness • Interplay of structure and options • Tools for decision makers • Focus on system threshold changes first • Capture structure and options • New social consensus tools