Complexity to Reflexivity:
Underlying Logics Used in
Science
Stuart A. Umpleby
The George Washington University
Washington, DC
www.gwu.edu/~umpleby
Four models currently used in
science
•
•
•
•
Linear causality
Circular causality
Complexity or self-organization
Reflexivity
1. Linear causality
• The way most dissertations are written
• Many statistical techniques, including
correlation and regression analysis
• Hypotheses can be falsified
• Propositions can be assigned a level of
statistical significance
• The objective is to create descriptions
which correspond to observations
2. Circular causality
• Essential to any regulatory process –
thermostat, automatic assembly line,
driving a car, managing an organization
• Can be modeled with causal influence
diagrams and system dynamics models
• Often a psychological variable is involved
– perception of, desire for
3. Complexity theory
• Primarily a method of computer simulation
– cellular automata, “the game of life”
• A very general concept – competition
among species or corporations,
conjectures and refutations in philosophy
• Differentiation and selection – creation of
new variety, selection of appropriate
variety
• Explains emergence
4. Reflexivity
• Requires operations on two levels –
observation and participation
• Involves self-reference, hence paradox,
hence inconsistency
• Violates three informal fallacies – circular
arguments, the ad hominem fallacy, the
fallacy of accent (two levels)
A further explanation of
complexity theory
Self-organization
• What is currently called complexity theory
can be seen as an extension of the work
on self-organizing systems around 1960
• There are two processes – differentiation
or the creation of new variety and
selection of appropriate variety
• The first is done within an organism or
organization; the second is done by the
environment
Earlier versions of the idea 1
• Adam Smith used the idea in The Wealth
of Nations when he described the process
of innovation and competition among firms
or nations (1776)
• Charles Darwin used the idea when
describing genetic mutation and selection
by the environment (1859)
• Karl Popper used the idea in philosophy -conjectures and refutations (1950s)
Self Organizing Systems
Early Conception
Self Organizing System
Environment
Ashby’s Conception
Organisms
Self Organizing System
Background on reflexivity
theory
Observation
Self-awareness
Reflexivity in a social system
Four reflexive theories
• Heinz von Foerster: Include the observer
in the domain of science (1974)
• Vladimir Lefebvre: Reflect on the ethical
system one is using (1982)
• Donald Schon: Management as reflective
practice (1983)
• George Soros: Individuals are actors as
well as observers of economic and political
systems (1987)
Von Foerster’s reflexive theory
• The observer should be included within
the domain of science
• A theory of biology should be able to
explain the existence of theories of biology
• “Reality” is a personal construct
• Individuals bear ethical responsibility not
only for their actions but also for the world
as they perceive it
Lefebvre’s reflexive theory
• There are two systems of ethical cognition
• People are “imprinted” with one or the other
ethical system at an early age
• One’s first response is always to act in accord
with the imprinted ethical system
• However, one can learn the other ethical system
and act in accord with it, when one realizes that
the imprinted system is not working
Soros’s reflexive theory
• Soros rejects Popper’s conception of “the
unity of method,” the idea that all
disciplines, including the social sciences,
should use the same methods of inquiry
as the natural sciences
• Soros says in social systems there are two
processes – observation and participation
• The natural sciences require only
observation
The informal fallacies
1. Fallacies of presumption which are
concerned with errors in thought – circular
reasoning, circular causality
2. Fallacies of relevance which raise emotional
considerations – the ad hominem fallacy,
including the observer
3. Fallacies of ambiguity which involve
problems with language – levels of analysis,
self-reference
Which models are acceptable?
1. Linear causality – the dominant
conception of science
2. Circular causality – used in first
order cybernetics, but involves circularity
3. Self-organization – Stephen
Wolfram’s “new kind of science,” complex
systems
4. Reflexivity – second order
cybernetics, violates 3 informal fallacies
Acceptable and unacceptable
models
• Models 1 and 3 – linear causality and selforganization – are acceptable. No
informal fallacies are involved
• Model 2 – circular causality – is suspect.
It involves circular reasoning. But it has
proven to be useful
• Model 4 – reflexivity – violates 3 informal
fallacies, so is highly suspect. Scientists
shun it, do not take it seriously
A decision is required
• Should traditions concerning the form of
arguments limit the scope of science?
• Or, should the subject matter of science
be guided by curiosity and the desire to
construct explanations of phenomena?
• Cyberneticians have chosen to study
certain phenomena, even if they need to
use unconventional ideas and methods
Some definitions
• A reflexive phenomenon – a boom and
bust cycle, an idea spreads through a
population
• Reflective practice – think about what one
is doing, observe consequences of
actions, and make adjustments
• Reflexive practice – think about one’s self
as a participant in a reflexive phenomenon
Describing what is occurring
• Reflexive phenomena can be described
using positive feedback loops
• Managers and policy analysts engage in
reflective practice but an epistemology of
reflective practice, separate from classical
science is not well-developed
• Metaphor of taking a photograph vs.
metaphor of riding a bicycle
• Classical science has constrained our
thinking
Two conceptions of
how to structure knowledge
• Most philosophers of
science
• Cause and effect
• If, then
• Analysis
• Reductionism
• Theory
• E.A. Singer, Jr.,
Churchman, Ackoff
• Producer - product
• Necessary conditions
• Synthesis
• Expansionism
• Method
Science one vs. science two
•
•
•
•
•
Observation
Description
Test knowledge
Extrapolate/ forecast
Reproduce
experiments
• Accuracy/ precision
•
•
•
•
•
Participation
Prescription
Solve problems
Create/ design
Achieve agreement or
acceptance
• Usefulness
Three changes are needed in
economics
1. Economists, and other social scientists, need to
accept the uncertainty that accompanies
violating the informal fallacies
2. Social scientists need to expand the philosophy
of science by including the observer in the
domain of science
3. Economists need a model of economic systems
which allows participants to be observers and
observers to be participants. This is a large step
beyond behavioral economics
Our conception of science is the
obstacle
• Practicing managers and social scientists will
readily agree that human beings are both
observers and participants in social systems
• Indeed, they say this idea is “not new”
• But this perspective is not permitted by the
current conception of science
• Our conception of science needs to be
expanded in order to encompass social systems
Contact information
Stuart A. Umpleby
Department of Management
The George Washington University
Washington, DC
www.gwu.edu/~umpleby
umpleby@gmail.com
Presented at the biennial conference of
the Washington Academy of Sciences
Washington, DC
March 27-28, 2010