I am Alp Baysal, a researcher and lecturer at The School of
Engineering Management (SEM) of University of Cape Town (Republic of
South Africa). At SEM, most of our activities are based on systems
principles, and they are structured according to systemic frameworks.
In addition to System Dynamics, we are also using extensively other
systems approaches like Viable System Methodology and Soft Systems
Methodology, in teaching, facilitation and consultancy works.
To start some discussion and also to give some background about my
thinking, I have attached an edited version of a message that I have
previously sent to k12all list.
-----edited message---------------------------------------------------
On the topic of Causal Loop Diagrams and Flow Diagrams, I would like
to share some of my experiences and opinions with you
The use Causal Loop Diagrams or Flow Diagrams is a context dependent
topic. Some of the possible contexts that immediately comes to my
mind are (I am sure that one can think of many others) :
- introducing systems ideas to people who are not familiar with them,
- facilitating a problem solving exercise in a complex system
(teaching or consultancy) where people are fairly conversant with
systems ideas,
- teaching technicalities of computer simulation (Dynamo, Stella, Ithink)
- ...
I think that the first one (i.e. introducing system ideas to people
who are not familiar with them) is more directly relevant to the
original question addressed to the list (whether to introduce causal
loop diagrams or flow diagrams first when teaching system
dynamics). Before elaborating further on this point I would like to
highlight Kauffman's book "Systems 1, An Introduction To Systems
Thinking" as a very useful, conscise introductory document that can
be used in this context.
I believe that, among all system methodologies, System Dynamics is a
more suitable one for this purpose. Core systems concepts like,
interdependency, interaction, circular causality, complexity,
systemic behaviour (emergence), patterns of behaviour, etc... can be
illustrated in a tangible way by Causal Loop and Flow Diagrams. I
would argue that both types of diagrams are complementary and
unseparable by their nature.
>From sequencing point of view I tend to introduce Causal Loop
Diagrams first, and I have two major reasons for this.
An important benefit of System Dynamics is to allow us to move away
from linear causality relationships (that can sometimes form an
infinite string of causes and effects, that can only give a limited
insight about a situation) towards circular causality structures that
are more powerful in explaining systemic behaviours. In Flow Diagrams
the main focus is on the linear structures of Levels and Rates. Feed
back links (that I believe to be crucial in systems structures,
because the regulation and thus the viability of the system depends
on them) are less emphasised than they are in Causal Loop Diagrams.
Therefore to start with Flows and Rates and to move to computer
simulation, may mean to offer the tools of a different approach
before introducing one of its important principles.
My second reason is the following one.
Synthetic Thinking (which is one of the two major streams of thinking
underlying Systems Thinking -the other is Analytical Thinking-)
requires to put the system that is under study in its overall
context. In other words, it suggests to see a system as an element of
a containing whole, then to develop an understanding of the
containing whole, and from this holistic understanding, to get some
insights about the system that is under study. It seems to me that
Causal Loop Diagrams are more appropriate for achieving this set of
objectives. They allow to build a bigger picture easier than Flow
Diagrams. They can also tap on the collective intelligence of people
who are involved in the modelling exercise, and thus generate a
shared vision about given a situation.
Once an overall understanding has been developed through Causal Loop
Diagrams, Flow Diagrams can be introduced to highlight the dynamic
behaviour of the system, and to model the systemic behaviour on the
computer.
I cannot see any modelling effort in isolation (i.e. modelling for
the sake of modelling). At the end of the day, the purpose is to
develop some systemic understanding of a problematic behaviour, in
order suggest meaningful intervention(s) that are likely to improve
it. From the learning and understanding point of view, the process of
building a model is as valuable as (if not more) the final model
itself. In order to get the full benefit of this process, I believe
that one should start with causal loop diagrams, and then move to
flow diagrams. Otherwise, it may be difficult to develop an
appreciation of the overall context, and without such an
appreciation, modelling effort can become an objective rather than
being a mean.
Alp Baysal
School of Engineering Management
University of Cape Town