This page contains notes and general information about concept maps.
Concept Map: a two-dimensional graph of a body of knowledge meant to represent and organize knowledge in a fundamentally hierarchical fashion but also supporting cross-linking between branches of the hierarchy.
Concept: a perceived regularity in events or objects, or records of events or objects, designated by a label - usually drawn in a box, circle, or other shape (sometimes called a node).
Relation: a word or clause denoting a relation between two concepts - usually drawn as a directed arc joining two nodes.
Proposition: a statement about some object or event in the universe, either naturally occurring or constructed, and that contains two or more concepts connected by relations to form a meaningful statement. Propositions are sometimes called semantic units or units of meaning.
Cross-link: A link that connects a node in one branch of a hierarchical concept map to a node on another branch - essentially creating a loop in the graph.
CMs assume that David Ausubel was right: “…learning takes place by the assimilation of new concepts and propositions into an existing concept propositional framework held by the learner…. This acquisition is mediated in a very important way when concrete experiences or props are available; hence the importance of hands-on activity for science learning with young children, but this is also true with learners of any age and in any subject….” [Nov00]
“…the most critical memory system for incorporating knowledge into long-term memory is short-term, or working memory…. working memory can process only a relatively small number of psychological units at any one moment. This means that relationships among two or three concepts are about the limit of working memory processing capacity. Therefore, to structure large bodies of knowledge requires an orderly sequence of iterations between working memory and long-term memory as new knowledge is being received.” [Nov00]
“While it is true that some students have more difficulty building concept maps and using these, at least early in their experience, this appears to result primarily from years of rote-mode learning practic in school settings rather than a result of brain structure differences per se. So-called learning style differences are, to a large extent, differences in the patterns of learning that students have employed varying from high commitment to continuous rote-mode learning to almost exclusive commitment to meaningful mode learning. It is not easy to help students in the former condition move to patterns of learning of the latter type. While concept maps can help, students also need to be taught something about brain mechanisms and knowledge organization, and this instruction should accompany the use of concept maps.” [Nov00]
As a tool for design, concept maps exhibit several problems.
Uniqueness: Concept maps are not unique representation forms. (They do seem to be, however, non-ambiguous.) This is both a weakness and a strength. It is a strength because it encourages users to create maps that are consistent with whatever cognitive structures they hold - the map doesn't force the user to think in a prescribed way. However, the extent to which this is taken by the method encourages a certain sloppiness. For example, there may be on one map both (a) a 1-to-many link to indicate multiple instances of a relation, and (b) multiple 1-to-1 links to represent the same thing. This is confusing because it suggests there is some difference in meaning between single 1-to-many links and multiple 1-to-1 links. Multiple 1-to-1 links also waste space and add to map clutter.
No mechanism for multiple contexts: Concept maps work best for single contexts. However, there are many different contexts in a design project. How can a single concept map accomodate multiple contexts? What exactly are the kinds and nature of the contexts that occur in design projects?
Emphasis on hierarchical arrangements: Design information is intensively cross-linked. Concept maps treat cross-links as important but secondary to the hierarchical structure. More support for cross-linking should be provided.
Connection to CBR: See [LW01]. Concept maps can be used to structure cases suitably for searching via case-based reasoning.
Using CM for PDS: I'm not sure I need to represent the universal PCs on a CM. Maybe they should just be used as a checklist to make sure the specific PCs are addressed. …no. In fact, I do need the top level PCs because the connect directly to issues addressed by phases, or segments of an engineering enterprise or process.
Developing a PDS CM: The actual process for developing a PDS CM may best be done by starting with one characteristic and exploring its implications (FRs, C, PMs) rather than going breadth first. I should actually develop some PDS CMs of simple products to explore this.