Planning discourse: Integration of argumentation into systems models or systems modeling information into argumentative discourse.Posted: March 10, 2014
Various discussions about how complex societal problems and crises can be dealt with have revealed, among other things, a mutual shortcoming of two conceptual ‘models’ held to carry the best promises for overcoming the challenges: ‘Systems Thinking’ on the one hand, and the Argumentative Model of Planning on the other. Briefly, systems modeling tools are considered the best available tools for the understanding and analysis of complex systems behavior, while a carefully orchestrated argumentative discourse with wide participation appears to offer the best – because most familiar and accessible – vehicle for assembling the ‘distributed’ information and connecting that information forward to acceptable agreements and decisions.
The problem or shortcoming is the following: The detailed information embodied in complex systems models is not accommodated in the familiar patterns of argumentative discourse, and thus difficult to adequately bring to bear on the decisions reached at the end of such discourse. On the other hand, the disagreements (and thus conflicting, inconsistent information) that characterize argumentative discourse in the form of ‘pros’ and ‘cons’ are not accommodated in the typical systems models whose assumptions regarding variables, parameters, and their values and relationships have the appearance of being either valid on the basis of scientific verification, or ‘settled’ by other means (e.g. as goals ‘given’ by the clients of analysis projects, or opinion surveys).
The consequences of decision processes adopting either ‘model’ can be equally defective: decisions based on the output of model simulations, for example, run the risk of overriding critical disagreements and interests of parties whose information has not been included, or downplayed, in the model, and thus lead to future conflict. Decisions reached on the basis of argumentative discourse in which the complexity of the system in question has not been fully understood because it couldn’t be adequately represented in the tools of the discourse are equally likely to be flawed. This would be true even if the main shortcoming of the ‘parliamentary’ tradition were successfully resolved – that of the possibility of the final majority vote completely ignoring and overriding the concerns of the minority. (A possible solution for this problem has been suggested with the proposals for systematic and transparent argument assessment for planning arguments (Mann 2010), it will be assumed to be adopted in some form in the following.)
The mutual difficulties of these two models to appropriately accommodate each other’s content is considered to be a main obstacle for the successful development of a viable framework for planning / policy-making from the small scale, local level to the scale of global crises and conflicts. It has not, to my knowledge, received sufficient attention, analysis and discussion. The following two suggestions, exploring the possibility for each of the two models to be integrated into the other are intended as a starting point for this much needed discussion. The possibility for the emergence of a ‘third model’ that would resolve the difficulty is left open as a challenge for future thinking.
- A. How can Argumentation be integrated into ST tools?
Possibility: (using the example of simulation models for clarity, and
i) Starting with the model diagram:
ii) Each variable and parameter in the model diagram is shown in a ‘box’ with attached expansion symbols:
E for explanatory information about item x: What is x? Also: description?
F for factual information: What is (the value of) x currently; evidence, data?
O (instead of D) for deontic / ought information and arguments: should x be set (as part e.g. of an intervention package)?
H or I for instrumental (‘How to’) information: How can x be achieved?
Clicking on the symbol will open a discussion page where the question is stated and answers / arguments are listed.
Plan proposals are described as packages of variable and parameter values of the model that serve as the proposed ‘intervention’ settings whose performance will be simulated over time in the model.
Main menu symbols shown in the ‘legend’ box of the diagram show the links for the issues:
– What should be the plan proposal? (Described as initial intervention settings of model); clicking should link to follow-up questions:
– H- question: plan proposals (alternatives)?
– Evaluation work sheets for selected proposals. (To develop a pl-value for the proposal based on the assessment argument weight, argument plausibility and plausibility of argument premises).
– What is the critical performance variable that should be simulated with the model?
– What additional variable / parameter should be included in the model?
Subsequent additional links for the follow-up question:
– Should this item be included in the model?
– What are the values and relationships?
This information can be ‘automatically’ extracted from the discussion and shown in the model.
– Should the proposed variable be part of the intervention (plan) package?
– How can the initial / intervention variable setting be achieved (if not already in place…)
iii) These pages should have convenient ‘back’ links to the question from where they were accessed.
iv) The pages for these questions should be complemented or linked to issue maps showing the relationships between the various issues in the entire discussion, (with the ‘current issue from which the page was linked shown bold or highlighted).
v) These requirements imply that the different functions described: model diagram, issue discussions, mapping, evaluation etc. must be part of one single integrated software program.
- B. How can systems modeling information be integrated into argumentative discourse platforms and maps?
Assuming, as a starting point, that there is a discussion about whether a plan proposal X should be decided upon for implementation. The discussion support documentation (drawn from the ‘live‘ or conventional online discussion) is organized along the principles of adapted planning discourse IBIS (‘issue based information system’) resp. APIS (‘argumentative planning information system’).
Arguments pro or con the proposal will be raised and displayed in the ‘standard’ format:
“Proposal X ought /ought not be implemented because it is/is not a fact that X with help achieve goal Y, given conditions C, and conditions C are/are not (or will be) present.” Formally:
“+/-O(X) <— (+/-FI((X–>Y)| C) & +/-O(Y) & +/- F(C)”
Here, ‘conditions C‘ stand for the set of assumed variable and parameter values of a simulation model; and the proposal X will be described as the package of such model assumptions that are under the control of planners as the starting ‘intervention‘ into the situation and for which the performance over time is to be simulated with the model.
Successor questions about C will be answered by displays of the entire model, listing all variables and parameters with their assumptions and relationships so that they can be discussed, within the regular format provisions of the argumentative discourse platform.
The platform will be structured according to the main considerations described e.g. in Mann (2010) including the components of the verbatim file of contributions, the topic and issue lists, the discussion files of each issue in a condensed / formalized manner, argument maps, and evaluation worksheets and analysis tools.
This requires that the platform be structured so as to
i) allow discussion of each issue as a separate thread;
ii) permit visual displays of not only issue and argument maps but also systems model diagrams (and ideally, running provisions) within the same platform;
iii) allow convenient forward and backward linking between all its components.