3. ComplexityScienceandtheSocialWorld- WillMedd
EncyclopediaofSocialMeasurement(2005)
• ‘‘understanding the emergent patterns of system behavior and in
developing models that can be applied to a whole host of phenomena”
• ‘‘complexity social scientists’’ “want to show how social science can
contribute to complexity science and how complexity science, applied and
developed for the social world, can illuminate real processes of social
dynamics”.
3
4. Complex systems key features
• 1. Self-organized emergence
• 2. Nonlinearity
• 3. Self-organized criticality
• 4. Evolutionary dynamics
• 5. Attractor states - trajectories that tend to be restricted to (attracted
to) particular parts of phase space (abstract multidimensional space with
coordinates for all the variables of the system)
• 6. Interdependent systems - patterns of system dynamics which are
repeated through different scales (systems within systems)
4
5. Social measurement
Quali
• ethnographic
• participant- or non-participant observation
• content analysis
Quanti
• Mathematics
• Simulation
5
bifurcation diagrams
non-linear modeling
computational models
genetic programming
cellular automata programming
distributed artificial intelligence
evolutionary game theory
parallel distributed cognition
artificial life
network analysis
sociocybernetics
6. Mathematical tools
1° General equilibrium theories – Equations to define a dynamical system -
Systems of differential equations describing the behavior of simple
homogenous social actors. Change occurs as a result of perturbations and
leads from one equilibrium state to another.
• Weakness: the assumption of equilibrium is forced by the mathematics, not
by the observation of social behavior
• 2° Statistical theory - indirect method for discovering causal relationships
and observe regularities in social phenomena
6
7. Computer technology - solve nonlinear equations
numerically (by trial and error) rather than analytically
(finding a formula to a solution).
Objective:
• to examine graphic representations of time series data in order to expose
‘‘attractors,’’ or forms of underlying order, pattern, and structure.
Phase space
• point attractors - stable equilibrium
• cyclical attractors - stable periodic oscillation between the same points
• strange attractors - figure-eight pattern.
7
Nonlinear Mathematics
8. Three particular difficulties
• 1° - such analysis requires large amounts of time series data that are
consistent and reliable, which is difficult to achieve in the social
sciences
• 2º - some of the emergent patterns are as much a function of the
nonlinear equation as they are of the particular data being explored
• 3° - equations assumed deterministic relationships in their structure,
meaning that the structure of the equation does not evolve; this is
problematic for social analysis because underlying social structures
may change in ways that the equations supposedly representing
them cannot
• Models could not provide objective viewpoints of the social world,
and would therefore become by necessity tools in the social world
8
9. Simulation
• Artificial societies - specialized type of simulation model that typically
employs an object-oriented, agent-based system architecture.
• Interests - self-organized patterns that emerge in running the simulation,
including the emergence of individual behaviors of the agents.
9
10. Critics
• Simulations are always too simplistic —simulation requires a
simplification of the agents to rule-based behaviors, of the agents
interacting in particular ways, and of what is seen as important in the
environment (what is important?)
• Simulations - may fail to capture a key aspect of the social world (human
beings may change their very behavior in unpredictable ways as a
consequence of the emergent system)
Simulation does not claim to be the real system, but instead provides a
method for exploring possible dynamics that might be looked for in the real
system
10
11. • Nonlinear mathematical models to the social world - establishing
determinate sets of relations to represent complex social interactions
• Developing simulation models -If it is the case that those boundaries
are dynamic - continued adjustment / continually - adaptive to
changing internal and external environments.
Challenges
11