From Structural Representation to Meaning Reconstruction: A Framework for Kit-Build Concept Mapping

 From Structural Representation to Meaning Reconstruction: A Framework for Kit-Build Concept Mapping


Concept maps have long been used as a tool for externalizing learners’ knowledge structures and supporting meaningful learning (Novak & Gowin, 1984). In concept mapping, knowledge is represented as propositions composed of concepts and linking phrases, forming a graphical network that visualizes relationships among concepts. Because of this graphical representation, concept maps have frequently been used as an assessment tool for diagnosing learners’ conceptual understanding.


However, previous research has repeatedly pointed out a fundamental limitation of concept map assessment. The graph structure of a concept map does not necessarily correspond to the learner’s knowledge understanding (Ruiz-Primo & Shavelson, 1996; Nesbit & Adesope, 2006). Structural characteristics such as the number of links, hierarchical depth, and cross-links capture properties of the graphical representation, but they do not necessarily represent the semantic meanings intended by learners.


This discrepancy arises because concept maps are flexible representational systems. The same conceptual understanding may be expressed through different structures, and structurally similar maps may represent different meanings. In addition, concept map construction requires representational skills such as diagram organization and linking phrase selection, which may influence map structures independently of conceptual understanding. Consequently, differences between concept maps may reflect differences in representation rather than differences in knowledge.


One promising direction for addressing this issue is to constrain the representational space of concept maps. By limiting the set of concepts and relations available for constructing maps, the variability of representations can be reduced. Under such constraints, structural differences are more likely to correspond to differences in conceptual understanding.


The Kit-Build Concept Map (KB map) approach adopts this idea by decomposing a teacher’s reference concept map into a set of components that are provided to learners as a kit (Hirashima et al., 2015; Yamasaki et al., 2017). Learners construct their maps by reconstructing these components rather than freely generating concepts and relations. Because the representational elements are shared and constrained, the space of possible maps becomes significantly limited.


This constrained representation enables structural comparison between learner maps and the reference map to function as a reliable diagnostic mechanism. Missing links, incorrect connections, and misplaced relations can be detected automatically and interpreted as potential misunderstandings in learners’ knowledge structures.


More importantly, this structural diagnosis can trigger a process of meaning reconstruction. When learners compare their maps with the reference structure or receive feedback based on structural differences, they are prompted to reconsider the conceptual relationships underlying their maps. In this process, learners do not simply correct graphical structures; rather, they reconstruct the meanings represented by those structures.


Therefore, Kit-Build concept mapping can be understood as a learning framework consisting of three interconnected stages:


Knowledge Representation

Learners externalize their understanding by reconstructing concept maps from provided components.


Structural Diagnosis

Differences between the learner’s map and the reference map are automatically detected through structural comparison.


Meaning Reconstruction

Learners revise their conceptual understanding by reflecting on the detected differences and reconstructing appropriate relationships.


In this framework, structural comparison is not the final goal of assessment but a mechanism that mediates the reconstruction of meaning. By constraining the representational space and enabling precise structural diagnosis, the Kit-Build approach reduces the gap between graph structure and semantic meaning and transforms structural differences into opportunities for conceptual reconstruction.


Thus, Kit-Build concept mapping addresses the long-standing challenge in concept map research that graph structures do not necessarily represent knowledge understanding, by providing a learning environment in which structural differences actively support the reconstruction of meaning.

コメント

このブログの人気の投稿

答えから始まる学習:再構成学習・パズル性(適合的/生産的)・点検の設計

AIリテラシーとしての生成AI活用の二層化:プロンプトリテラシーから意味構成リテラシーへ

生成AIの成果物に「何があって,何がないのか」:四つの機序と意味点検としての自己再構成