Requisite VARIETY: Attributed to Ashby
A system is described by a large number of variables and states. In the case of complex systems these are also intensively interdependent. To be able to deal properly with the diversity of problems the world throws at us, we need to have a repertoire of descriptors or responses, which is (at least) as nuanced as the problems we face. The Law of Requisite Variety, as formulated by Ashby[1], states “the number of states of a system’s control mechanism must be greater than or equal to the number of states in the system being controlled.” The application of this requirement in the context of DDS demands that the organizers ensure that those invited to participate and contribute their observations (which correspond to the variables) represent as much as possible the entire spectrum of perspectives, conflicting points of view, ideas, and more important interests. Thus, a key requirement for a successful outcome is that rigorous stakeholder analysis and consideration of appropriate selection criteria should precede the organization of an SDDP.
Scaling-up: Perspectives and interests represented through those who participate might not reflect those of the whole population of stakeholders for the particular subject of inquiry. In a completely open system, participants are self-selected.
Challenge: A future scaled-up platform will provide: (a) Assessment and visualization capabilities to support users “seeing” how well the various points of view and interests are represented. To this end, indices will be developed to assess variety in perspectives, interests and other characteristics of the population. (b) Mechanisms to identify imbalances, e.g., missing perspectives or interests, gender/age imbalance, etc., and (c) Easy-to-use options for participating users to suggest (to a particular other user, or to the general population of participants) to identify and invite additional [missing] participants whom they might happen to know.
Requisite PARSIMONY: Attributed to Miller and Warfield
Neuroscientists have known for decades that an important limitation of the brain is its short-term memory capability. This explains for example why we cannot recall more than 5-7 items after a short time[2],[3],[4], or that our cognitive systems cannot process too much information at the same time. For this reason, in the implementation of a face-to-face SDDP, participants offer their contributions in a robin-round manner giving others the opportunity to listen actively. For the same reason, during the Clustering and Structuring phases, participants are expected to make simple binary comparisons, so they focus on only one relationship at a time.
Scaling-up: Participants interact with information individually, which means there are debates or collective considerations. They engage with the system for short sessions distributed over a longer time span, which leads to forgetfulness. Web-based platforms tend to be cognitively overloading for users.
Challenges: A future scaled-up platform must comply with latest User eXperience (UX) design; flow of information rate should be carefully controlled using a strict seQuenced dEsign process (QE).
Requisite SALIENCY: Attributed to Boulding
Boulding suggested that when designing a new system, designers should be alerted that not all salient (i.e., important, primary, prominent) features are of equal importance[5],[6]. Therefore, one should examine their relative salience[7]. In order to reveal and discover the relative salience of factors contributed in the context of a DDS application, the participants are asked (e.g., in the Clustering phase) to discuss and debate whether two factors have SIGNIFICANT features in common to justify putting them in the same cluster. This process helps them to focus on features that distinguish two factors, thus forcing them to discuss their relative saliency.
Scaling-up: Because participants will be interacting asynchronously, we will be missing the lively debates that take place in a face-to-face implementation. Without open debates and the option to listen to other points of view, participants have to rely primarily on their own judgment.
Challenges: The future platform will test a number of different clustering algorithms for their capability to highlight distinctive attributes. For example, the system can offer factors A, B, and C, and ask the user to decide whether B is closer to A or C, thus combining judgment for similarity with judgment for non-similarity. A different algorithm may ask user to make a single-sentence statement why s/he thinks that A does not belong in the same cluster as B. The consideration of similarity matrices[8] produced by individual participants will allow us to identify cases where meanings and interpretations among participants deviate more and thus select those cases for further inquiry and investigation in smaller groups of users.
Requisite AUTHENTICITY & AUTONOMY: Attributed to Tsivacou
Most dialogues that involve a group of people usually suffer from phenomena known as Groupthink[9],[10] and Clanthink[11]. Groupthink is when a portion of participants follow and support certain ideas or arguments or simply remain silent, out of fear of rejection or persecution by someone with relatively more power. Clanthink, is the extreme case of Groupthink, when practically all members of a group support an idea even though they know that it is outright wrong, because otherwise they might be ostracized. The law of Requisite authenticity and Autonomy, attributed to Tsivacou[12], states that during the Clarification and Clustering phases, others should be allowed to ask questions only about meaning; no judgments allowed. This facilitation technique is intended to protect the autonomy and authenticity of participants so that no participant is discouraged, and no idea is prematurely evaluated and/or rejected.
Scaling-up: Participants will be interacting asynchronously, which means all non-verbal communication is lost. This can be positive or negative depending on the specific situation; in this case easy to comply. Without personal contact, however, the possibilities for misunderstandings (and life cycle of a possible conflict) are higher. At the same time, if the options of interactions are restricted and structured, these possibilities might be far less.
Challenges: A future scaled-up platform should restrict and structure the types of interactions allowed between participants. For example, actions allowed in response to “reading” another participant’s contribution maybe structured and restricted to questions like: How do you …?, Why do you think.. ?, What do you mean ..?
Requisite MEANING & WISDOM: Attributed to Peirce
This law states that meaning and wisdom are produced during a dialogue only when observers search for relationships of similarity, priority, influence, etc. within the set of their observations, as depicted in Charles Saunders Peirce’s Lectures at Harvard in 1903[13]. One way in which individual wisdom is externalized is the fact when asked to select the most important ideas out of the total pool, participants do not choose their own. Another way, takes place when following debates on a certain relation they change their mind after having heard multiple arguments.
Scaling-up: Because participants will be interacting virtually and asynchronously, the search for relationships of similarity, priority, and influence cannot happen in real-time face-to-face manners, therefore debates are weakened or lost. This inhibits emergence of shared meaning and learning. Furthermore, the group of individuals engaged in the exploration of relationships and also their state of mind is constantly changing.
Challenges: In order to ensure that the gradual search for relationships progresses in a manner approximating face-to-face interactions, a future scaled-up platform must (1) control that the number, composition, and variety of participants, who engage in exploring a specific binary influence is at least relevant and sufficient (thus enabling meaningful interactions); (2) attempt to construct shared meaning through an iterative learning process by which participants review and extend the decisions of the small group in a piecemeal manner.
Requisite EVOLUTION OF LEARNING: Attributed to Dye
While analyzing SDDP data across many applications, Dye and his colleagues[14],[15],[16] observed that ideas that receive high votes immediately after the Clustering phase (i.e., importance voting), do not usually end up at the root of the Influence Map. This means that if the decision for taking action is taken on the basis of aggregating individual “importance voting”, the action will be ineffective. This is called the “Erroneous Priorities Effect”. The relevant law postulates that participants learn through an evolutionary process that renders them capable of discovering the most influential factors only when they go through this process of exploring binary influences of one factor on another (Structuring phase)
Scaling-up: Because participants will be interacting asynchronously within a virtual medium, the exploration of influence relationships cannot happen as a group process. This compromises evolutionary and group learning. Also, the group of individuals engaged in deciding influence relationships and their state of mind is constantly changing.
Challenges: This represents probably the biggest challenge of a scaling-up effort. The future platform should ensure that the logic of the ISM algorithm is implemented and that it works within a virtual space in a manner that approximates face-to-face interactions. Section 1.3.1.4 explains in detail the proposed model of organizing and managing multiple, distributed ISM sessions, and developing the necessary mathematical logic to merge distributed influence maps using an iterative process, which gradually produces a larger map.
Requisite ACTION: Attributed to Laouris
Participants of face-to-face SDDPs are almost always willing to assume some kind of responsibility and take action. Organizers have no power to require participants of a co-laboratory to take action, but adherence to the laws of DDS seems to be setting up the stage. Parsimony, autonomy, and evolutionary learning assists them to achieve meaning and wisdom and out of these, largely cognitive processes, action emerges as a natural consequence. The relevant law is related to the Engagement Axiom (credited to Özbekhan[17]) that states that “it is unethical to try to change any socio-technical system without the explicit permission and participation of those whose lives will be influenced by any changes”, and predicts that “any attempts to do so have a higher risk of failure.” (i.e., Requisite Law of Action; credited to Laouris[18]).
Scaling-up: It is difficult for people who have never met, to agree on an action and assume joint responsibility to execute it. Examples, such as the Arab Spring and analogous, show that this is not possible when the virtual medium is used only for coordination rather than for deliberation.
Challenges: The premise of the new platform is that online DDS will engage participants in meaningful interactions
[1] Ashby, W.R. 1956, An Introduction to Cybernetics, Chapman & Hall, 1956, ISBN 0-416-68300-2 p207. Available online: http://pespmc1.vub.ac.be/ASHBBOOK.html Retrieved 15 Mar 2019.
[2] George A. Miller, "The Magical Number Seven, Plus or Minus Two", Psych. Rev. 63(2), March, 1956, 81-97.
[3] Herbert A. Simon, "How Big is a Chunk?", Science, 183, February 8, 1974, 482-488.
[4] John N. Warfield, "The Magical Number Three plus Minus Zero", ISGSR meeting, Budapest, June 1987.
[5] Hester, P. T., & Adams, K. M. (2014). Systemic thinking: Fundamentals for understanding problems and messes (Vol. 26). Springer. p.70
[6] Boulding, K. E. (1966). The impact of the social sciences (No. 316). Vakils, Feffer and Simons Private.
[7] Warfield, J. N. (1999). Twenty laws of complexity: Science applicable in organizations. Systems Research and Behavioral Science: The Official Journal of the International Federation for Systems Research, 16(1), 3-40. pg34
[8] https://www.sciencedirect.com/topics/computer-science/similarity-matrix
[9] Whyte, W. H. J. (1952). Group think. Fortune, 45, 145–146.
[10] Janis, I. (1983). Groupthink: Psychological studies of policy decisions and fiascoes. Boston: Houghton Mifflin.
[11] Warfield, J. N., & Teigen, C. (1993). Groupthink, clanthink, spreadthink, and linkthink: Decision-making on complex issues in organizations. Fairfax: Institute for Advanced Study of the Integrative Sciences, George Mason University.
[12] Tsivacou, I. (1997) ‘The rationality of distinctions and the emergence of power: a critical systems perspective of power in organizations’, Systems Research and Behavioral Science, Vol. 14, pp.21–34.
[13] Turrisi, P.A., Ed.,1997. Pragmatism as a Principle and Method of Right Thinking, State University of New York Press, New York.
[14] Dye, K. (1999). Dye’s law of requisite evolution of observations. In A. N. Christakis, & K. Bausch (Eds.), How people harness their collective wisdom and power(pp. 166–169). Greenwich: Information Age Publishing.
[15] Dye, K. M., & Conaway, D. S. (1999). Lessons learned from five years of application of the CogniScope. Approach to the food and drug administration. Pennsylvania: CWA Ltd.
[16] Laouris, Y., & Dye, K. (2017). “Democratic” voting without prior exploration of relationships between alternatives favors ineffective actions. Systems Research (submitted)
[17] Özbekhan, Hasan, 2019. The Engagement Axiom. https://en.wikipedia.org/wiki/Hasan_Özbekhan#Work
[18] Laouris, Y., Laouri, R., & Christakis, A. N. (2008). Communication praxis for ethical accountability: The ethics of the tree of action: Dialogue and breaking down the wall in Cyprus. Systems Research and Behavioral Science, 25, 331–348.
