Tools for Dialogue and Deliberation on Wicked Problems

Getting to the Root of Problem Solving.

It is difficult getting to the root of the problem when everyone's focus is on the fruit.

There are few holy grail solutions to complex global issues. The best that can be offered is mutual understanding and a desire to provide the most benefit in lieu of the least harm.

Rationally, I don't understand the benefit of a hierarchical argument structure where there is a root of a problem. A better visualization of a complex problem is a geodesic sphere, where each node represents a satisfactory solution for an individual or like thinking group. Connecting nodes represent alternative solutions which may or may not support each other. This geodesic model provides a better model of connectivity and analysis.

It is human nature to focus on self interest. Who wants to gnaw on a root when the fruit is within reach. A new matrix to problem solving needs to take into consideration self interest and mutual understanding. Our current linear model for political argument lacks the transparency to promote mutual understand. It allows the agenda of self interest to be hidden from view.

The debate model which our educational system promotes and our legal system fosters provides a linear method of argumentation which promotes limited polarized solutions. Complex problems are not guilty or innocent. They have thousands of ramifications and consequences. A geodesic argument model would better represent the complexities of the global issues which face us.

Next Post >

Add a Comment

You need to be a member of Global Sensemaking to add comments!

Join Global Sensemaking

Comment by Scott on August 24, 2009 at 12:20am: Human Input and Computer Automation

The heuristics for connecting arguments involves a mixture of human input and computer automation.
Automated Connectivity and Categorization

An Internet search engine indexes the words of web pages into a data store. This data store is instrumental to an automation for quickly retrieving a list of websites matching a keyword criteria.

Google provides an interface where humans supply a keyword criteria which is passed to Google's search engine. The search engine matches entries in the indexed data store with relevancy to the human supplied keyword criteria.

The Do Good Gauge will build upon the indexed data store. Whereas a human generates the search criteria for retrieving reference material from Google, the weighted words within an argument can be used to automatically connect related arguments.

Though this method automates the connectivity and categorization, it is not optimal for determining the relationship level or supportive direction.

Relationship and Supportive Direction

An author provides credit and reference to supporting concepts at the end of paper through a bibliography.

The Do Good Gauge introduces a concept similar to the bibliography. At the end of each article the reader is presented with a small list of compatible arguments. Along with each compatible argument a gauge is provided for measuring the relevancy. The gauge describes zero as having no relevancy, the positive extreme having a highly relative and supporting relationship, and the negative extreme having a highly relative and counter supporting relationship. The scores for the relevancy gauges is stored as an additional criteria for argument connectivity.

To reduce information overload, the number of these relationships is kept small or is limited by a parameter supplied to the reader. The choice of connected arguments is a combination of author suggestion, heuristics based on the automated categorization and connectivity, reader locale or topics of interest, prior argument visits by the reader, and the score of prior rel

Comment by Jack Park on August 22, 2009 at 6:17pm: Scott, I very much appreciate any contribution that is seeking fuzzy or other metrics of agreement; any move away from boolean truth values seems a towering contribution. Still, I'll argue that spheres are not an appropriate metaphor for visualization or representation of knowledge, even if spherical trigonometry brings valuable metrics into the dialogue.

You said it best, thus far: There are few holy grail solutions to complex global issues. The best that can be offered is mutual understanding and a desire to provide the most benefit in lieu of the least harm. I'll amplify that to, in my view, make it even better: There are no holy grail solutions to complex global issues. Herbert Simon said it this way: we satisfice. That there are contributions to ways to understand what satisficing means, to view the, um, degrees of agreement (or disagreement) between us, is, it seems to me, enormously valuable.

But, how do we go about measuring those, um, degrees? Do we not end up waltzing into a recursive and complex parse of our words, probing semantics ensconced in context? How do geo-positioning, great circles, and, indeed, spherical trigonometry help? Where does spherical trigonometry account for deceit (think: wormholes) in the public sphere? I'd love to think it does. Help me out here. How, indeed, did you arrive at 30 degrees?

Comment by Scott on August 22, 2009 at 5:41pm: Jack, I'm not convinced that we are looking at the same market. Jürgen Habermas' definition of Public Sphere as described on the Wikipedia web page does a much better job of portraying my vision better than I have been able to explain so far.

The geodesic concept is my metaphor to provide a mathematical formula for connecting arguments and measure supportive direction and level of relevancy.

Jack, I apologize if I come off as disagreeing with your premise. I would say that we are closer to 30 degrees in agreement than 180. Hopefully both our destinations are correct.

Comment by Jack Park on August 22, 2009 at 5:26pm: We appear to seek appropriate topologies that describe some "key architecture"; we look at the 2D properties of IBIS (whatever those might be), then argue for 3D information objects (whatever those might be). How does that help? Are we asserting that 2D is out and 3D is in? Why not 4D? Why not 'n'D, since there are arguments all over the place that knowledge is "high dimensional"? Where do the suppositions end?

Let's look at it this way, just for the sake of deep argument. Consider this bold assertion: a buckyball is a graphical (nodes and arcs) structure imposed on a surface (think 2D) that just happens to be stretched over a spherical object. As an example, take the right-left edges of a 2D sheet of paper and connect them and you just morphed the object into a cylinder. Now, and this takes a bit of imagination, take the top and bottom "edges" and connect them and you just morphed the object into a donut (torus). None of that morphing says anything about the "arcs" that lie between nodes elsewhere in the universe or even at the other "side" (whatever that means--think Karmakar solutions) of the 3D object. Algorithms for folding proteins are really hard precisely because the arcs (relations among nodes) cannot be shuffled onto a surface of any topology. A really bold assertion is this: neither can the subjects of our universe of discourse.

IBIS is cast in a 2D (tree) topology because it is easy, because it is necessary to engage people in structured dialogue in the most expeditious way, and because it deals with just one aspect of knowledge cycles: elicitation. Problem solving comes along in another cycle, where reflection, linking and annotation, and wiring into higher dimensional structures happen. Occasionally, that cycle occurs concurrently with elicitation; perhaps it shouldn't.

So what is a key architecture? What will it look like? I'd submit that it might actually turn out to appear (whatever that means) as nothing more or less complex than that of an ensemble of intersecting manifolds, the topologies of which are completely unknown. Ted Nelson's ZigZag framework might come the closest we have seen thus far to a way to model those intersections without even worrying about what is necessarily being intersected. He does, however, force those intersections to exist as lists of objects. It's not clear what that ultimately means, but it seems to be a start. After all, take away the lists and you can have trees, graphs of all sorts, all intersecting in interesting, possibly useful ways.

I will assert that Public Sphere is just a term, a confusing one at that, not a metaphor for knowledge representation. I'll likely burn in hell (after all, Scott is convinced that I'm stuck in the visualization aspect of this thread), too, but that's another story (lest we circle back on prior arguments)...

Comment by robert beckett on August 22, 2009 at 7:53am: thanks for the references...i think this is where the main field will be in 5-10 years time...3D information objects... IBIS and its 2D properties will be part of the code of design but certainly not the key architecture...in my opinion...

Comment by Scott on August 21, 2009 at 7:11am: I'm not ashamed to admit, I don't know what I don't know.

For whatever reason, I've never googled the term public sphere. I recently started reading David S. Allen's book called Democracy, Inc. This book describes the many issues relating to the press, laws, corporations, and the public sphere.

Comment by Scott on July 27, 2009 at 11:23pm: I continue to be impressed when so little can say so much. From the beginning of the DGG project I have collected quotes to help explain my motivation. On rare occasions I've been able compress a key idea into a sound bite. Here are a couple of sentences which clarify an advantage of a Buckyball matrix for argumentation.

If your desire is stagnation, finding an argument which is 180 degrees opposed to yours will help. If your desire is to move your argument in a specified direction, finding a similar argument within 15 degrees of separation is a better pool to recruit from.

Comment by Jack Park on July 12, 2009 at 6:36pm: There is a paper "Organizing a Fishnet Structure" that is available as a pdf by way of a web search. It speaks to the tensions that exist between our tendency to organize hierarchically in the context of heterarchical reality. From the abstract:
The fishnet organization is a new organizational form which, among others, emerged in scientific literature in the field of organization theory in the past 20 years. Questions to ask here include: How to develop and maintain such a dynamic and heterarchic structure, described with the metaphor of a fisher's net, in a real organization? How to find knowledge and abilities which are fundamental in constructing such a structure? It seems obvious that without an adequate information system this task would be almost impossible. Thus we present organizational and information tools which are needed for the development and maintenance of this dynamic organization.

Comment by Scott on July 9, 2009 at 9:56pm: Jack, I feel your getting lost in the visualization. Whether it is a sphere or a trojan horse, the 3 dimensional shape comes together from nodes and mesh. The nodes represents the argument domain and the mesh represents the networking or the conduit for communication with other arguments. What I'm attempting to do here is to pull the passionate members of this group away from their current project and to come to agreement on the definition an argument and the framework in which it lives. Until this is done we have a disjointed efforts at creating tools with no commonality. We need a standard API or SDK to build upon. Argument automation requires a Henry Ford assembly line framework for fitting the disjointed pieces into a working product.

Something that I feel has been lost in expressing the Do Good Gauge argument is the automation of connectivity. Connectivity will be a process of computer automation and human input. I've been pushing the concept of a relevancy gauge from the very beginning.

I realize the skepticism of the subjective nature of this relevancy gauge. I can only point out that subjectivity is human nature. Whether your argument is factual or not, it will be the subjective point of view of the populace, the champions, and the people who cut check to push the argument into a working model.

The Do Good Gauge relevancy instruments provide an interesting means of interaction for the reader to bring a human level of subjectivity to connecting arguments. Luis von Ahn, the inventor or ReCaptcha describes how he added a second term to the password phrase for the purpose of automating the OCR of deteriorating paper documents in the :

PBS Nova scienceNow episode 401

I would suggest the do good relevancy gauge is a similar form of "Crowd Sourcing"

Comment by Jack Park on July 7, 2009 at 4:50pm: Just mulling out loud (dangerous, I know), I am wrestling with a sphere being the preferred topology. Since you folks seem to be willing to eject trees as the end point in a topological analysis of argument space, then why stop at a sphere? In an earlier life, hilbert spaces came to mind; these days, category theory creeps in. After all, arguments can be viewed as members of categories, and arguments have social lives of their own; set theory alone perhaps doesn't provide sufficient horsepower to understand what set membership means in light of social lives of the members.
All that to suggest that DebateGraph, Compendium, and all the other "tree builders" provide a starting point. From a pragmatic perspective, they provide the "right tools for the right purposes at the right time" which is all about facilitating the capture of issues, positions, and arguments when people are ready for an elicitation phase. Arguments related to whether those captured trees are sufficient for a reflective/analytical phase are, I believe, an open topic of research. This thread, if situated in the reflective/analytical phase, makes sense; I am pretty sure that if audiences of combative colleagues were forced to wander some spherical or higher dimensional space during elicitation, then fewer benefits from the exercise would accrue (though I don't deny an interest in playing in such a sandbox--that I have reservations about moving beyond simple tree structures during elicitation should not mean it should not be tried).
Right now, I remain unconvinced from the arguments presented thus far, that spherical is better. I'm not sure I know what "better" means in this context.