Mapping madness

There is an abundance of tools for creating visual “maps” of “knowledge.” The sheer number of tools available for this purpose indicates a strong need for such applications.

Categories of semantic mapping tools

The site “Software for mindmapping and information organization” at [Note the hyphen.] — which lists hundreds of applications — implies that nearly everyone with a computer, a whiteboard, or a large blank sheet of paper feels compelled to use circles and arrows to deconstruct, organize, explain, or remember the ideas most critical to him or her.

In January of 2018, I discovered a new resource: — which appears to be work of Roy Grubb, also the current owner of Grubb offers a very different categorization than you will find here.

  • Outliners evolved from a paper model for rhetorical organization of documents into a computer tool not long after desktop computers arrived. Some add linking from node-to-node across hierarchies.
    • Hierarchies form the core of the Insight-Centric Knowledge Model — just as they do for some other forms of knowledge representation, including computer ontologies.
    • TheBrain does have an outline viewing mode — which is helpful at times — but does not offer the ease of building and changing outlines that one finds in TreePad and some other products.
    • FreeMind is an excellent freeware/open source outliner with limited linking across subhierarchies, but (in my opinon) its outline-manipulation features are not very fluid and the cross-linking features are not very helpful. Like many “mind-mapping” applications (See below.), FreeMind organizes nodes radially from a top node into subhierarchies.
  • In most concept-mapping software, nodes (also referred to as “circles” or “bubbles”) and arrows (also “edges”and “arcs”) are the primary tools for representing the relationships among ideas and terms. Common explanations of concept mapping include:
  • Mind maps are, by definition, a graphical method of taking notes.( ) [Comment: “Notes” recorded in this way are usually Insights (in my terminology) — sometimes connected by explicit relationships.]
  • Semantic network applications can be traced back to the 1950s. The Free Dictionary defines semantic network as “A graph consisting of nodes that represent physical or conceptual objects and arcs that describe the relationship between the nodes, resulting in something like a data flow diagram.” John Sowa describes the technique as “a graphic notation for representing knowledge in patterns of interconnected nodes and arcs.” Semantic networks strongly influenced many of the products in the prior categories, but few vendors apply this label to their own products these days.
  • Knowledge mapping, more often than not, refers to an organization’s efforts to locate and connect the information most vital to its operations. But use of the term is not consistent.
  • [added 26-Jan-2018] Text zooming — of which iMapping may be the best (or only?) current example — visualizes subordination by displaying smaller blocks of text within the current block of text in a grid-like arrangement. You can zoom down into the smaller (subordinate) blocks and zoom out as desired. I just made this category up and I’m not happy with the name, but products like iMapping are very different from most “circles and arrows” approaches. See also, Plectica and TreeSheets.
  • Those more concerned with formal representations of knowledge turn invariably to ontology visualization tools .

The distinctions among these categories of “mapping” applications are often very fuzzy. For example, many tools labelled “concept-mapping” are really outliners with a different take on the interface. (Freemind is a good example.)

More about “concept mapping”

Of the types of mapping mentioned, concept mapping is among the more rigorously defined. Or so concept-mapping people believe. If you use the term Concept (See What is a “Concept”? ) in this way, most ontologists (Knowledge Representation [KR] specialists) might agree with the following explanation:

Concept mapping is a process of meaning-making. It implies taking a list of concepts – concepts being perceived regularities in events or objects, or records of events or objects, designated by a label (Novak & Gowin, 1984) – and organizing them in a graphical representation in which pairs of concepts and their linking phrases form propositions [emphasis by PCM] . Hence, key to the construction of a concept map is the set of concepts on which it is based. In educational settings, teachers often prompt students by providing an initial set of concepts that they should include in their map.

Source: Canas et al., “Mining the web to suggest concepts during concept map construction.”

KR people usually define concept mapping as the construction of relationships among concepts — in effect, “triples” that consist of a subject, predicate, and object. This is the underlying model of RDF (the W3C Resource Description Framework).

Inconsistent and poorly-modeled Concept Maps

But, in practice, the set of people who are interested in “concept mapping” is much broader than KR and the Semantic Web. For others, the goal of concept maps is “simplifying access to complex material” or building a personal “map” of relationships among ideas (very loosely defined). The representations they create are often far from consistent with the definition of Canas et al. Concepts, “propositions,” and the relationships among them are far less rigorously defined.

The opening page of the web site mentioned above — IHMC Cmap Tools — is a perfect example of inconsistent and unmodeled use of Concept Maps. Consider, for example, the following two set of links on that site:

  • (Software Toolkit) -> is –> (Documented)
  • (Software Toolkit) -> is composed of –> (CmapServers)

This could not be sillier or more arbitrary — a great irony in a resource on Concept Mapping itself produced by a distinguished group of people from the world of Knowledge Representation. I even recommend their list of resources below in the “Reading” section. But these two items are simply natural-language sentences fragmented arbitrarily into objects and relationships … badly. The are not about relationships among Concepts. Many other representations on this site are just as bad.

I recommend incremental formalization, but you have to start with some basic underlying model for what you are representing with these circles and arrows.

Lack of an underlying model for meaning limits the usefulness of mapping tools

“Mapping” tools of many kinds can be very useful. However, tools for individuals often fall into disuse rather quickly — sometimes because they are limited, clumsy, and/or don’t provide a return on investment of time spent. But even with sophisticated, well-designed mapping applications, extremely narrow definitions of target activities limit the ways in which the products of these tools can be applied.

The biggest limiting factor in using the various mapping tools is that there is no overall framework and model for representing the meaning of communications among people . What, exactly, are you representing when you use such tools?


© 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 *