That compulsion to draw circles and arrows

We resort to “circles-and-arrows” representations of important ideas precisely because:

  • Language and memory fail us … frequently and pervasively.
  • Even within a group of like-minded people, understandings of basic facts differ.
  • Deconstructing complex understandings into visual objects and relationships allows us to
    1. Point to the deconstructed elements that we want to address, whether they are assertions or relationships among those assertions.
    2. Substantially reduce ambiguity of meaning.
    3. Suppress details.
    4. When properly implemented, ensure that each discrete idea or fact exists in only one place.

In business environments, whiteboards lurk in every meeting room. Sometimes they are a convenience for listing meeting agendas, but they almost inevitably become brainstorming tools or surfaces for mapping important points, action items, and responsibilities. Key items are circled. Arrows are drawn to related items then re-drawn to reflect better connections. This activity seems so natural in modern business that no one thinks twice about why that is so.

Individuals have turned to technologies to support these acts of connecting words with other words via graphics for almost as long as there have been personal computers with graphics capabilities. Roy Grubb has recently taken over Vic Gee’s web resource, Software for mindmapping and information organization, which lists nearly 100 applications for graphic representation of concept maps, mind maps, brainstorming, and argumentation. There’s also “Idea Mapping” — a technique for creating “a colorful, single-page diagram that visually captures ideas in a non-linear format and engages both hemispheres of the brain.” (Source: The term Idea Mapping seems to be associated strongly with Mindjet software.

This is terrific. Even the brain behind the Web thinks this way. While writing an earlier piece on capture and organization of ideas, I came across an old email from Doug McCue of EPC Solutions, in which he writes of “comments from Tim Berners-Lee that (my paraphrasing) ‘everything in the world can be described with circles and arrows’.”

(Tim B-L uses the phrase “circles and arrows” in at least one of his papers: “The Semantic Web starts as a simple circles-and-arrows diagram relating things, which slowly expands and coalesces to become global and vast.” in “The Semantic Web lifts off” by Tim Berners-Lee and Eric Miller. ERCIM News, No. 51, October 2002. His original vision is for metadata for documents.)

Tim B-L is right that you can describe everything in the world with circles and arrows. Well, almost right. There are major constraints:

  1. You have to exclude subtleties and nuances. There are practical limits to representations of knowledge, and Tim, I am sure, would not deny that. Ontologists in general do not deny that. The most avid proponents of Linked Data do not deny that.
  2. Tim is thinking of the world of knowledge representation for computing, in which nesting might be any number of levels deep. Infinite nesting is not usable by people. That’s why we create broader abstractions … and use those instead.
  3. The challenges we face are not limited to representation of reality. We also have to deal directly with how people communicate ideas and evaluate those ideas. Circles and arrows are not sufficient for those purposes, even though they are helpful. And the tools I have seen for computer-supported argumentation and “brainstorming” are, conversely, inadequately grounded in why and how we use circles and arrows to represent reality — a domain well explored by ontologists.

This circles-and-arrows thing is both widespread and seductive. It’s very useful when you look at local connections in a “network of knowledge,” leading to rapid grasp of relationships that are hidden by words. (Also note that the IA [Information Architecture] folks use “Boxes and Arrows” to talk about their web site navigation frameworks. That’s a bit different.)

But the circles-and-arrows mindset is only part of the toolset of meaning that we need. We need a new, expressive, structured, and usable visual language for communicating meaning that builds on the work already done but is not constrained by (or dominated by) that mindset.

  • It should encompass what we have learned from “circles and arrows” and should be extensible and scalable.
  • It should give users a standard, easy-to-use method for treating sets of connected things as addressable objects. We should be able to make assertions about the connections among those objects, whether they are currently connected or not.
  • It should support Incremental formalization.
  • It should comfortably support the data for several work activites in which the circles-and-arrows model is currently applied, including activities that are very loosely described as argumentation support, concept mapping, mind mapping, and knowledge mapping. See Mapping madness.

Re: scalability, It should also be designed with patterns in mind. When an implemented knowledge model gets really big, you’ll want to examine patterns in that knowledgebase.

© Copyright 2017 Philip C. Murray


This entry was posted in Uncategorized. Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *