**6. Conclusions**

This new approach to concept map analysis raises a number of new opportunities and challenges for the research community:

By considering concept maps to be composed of di fferent types of knowledge, it o ffers the possibility of asking a new set of research questions that might be addressed through concept mapping. Where powerful knowledge [47] is seen as the goal of professional education, then the semantic weaving between theory and practice is required to achieve expertise [21]. The assessment of this plurality of knowledges requires the mapping of semantic density and semantic gravity.

Beyond just assessing the 'correctness' of propositions within a map, the application of Legitimation Code Theory to concept mapping allows for the assessment of the ways in which the mapper is able to link theoretical knowledge with practical knowledge. This lifts the map above the assessment of factual recall and considers the higher order thinking skills that are required for students to achieve mastery of their discipline. This mastery has been shown to be dependent upon the learner's ability to oscillate between complementary knowledge structures consisting of chains of practice (exhibiting low semantic density and high semantic gravity), and underpinning networks of understanding (exhibiting high semantic density and low semantic gravity) [18,22]. The method of applying Legitimation Code Theory to concept mapping described in this paper provides a way to make the knowledges that underpin that expert practice explicit, so that they may be modeled for students. Further, this paper suggests that when assessing students' knowledge using concept maps, the use of a single map may be insufficient in order to obtain an authentic representation. As procedural and conceptual knowledge may be constructed differently and activated in different contexts, it may be better to encourage students to separate them structurally, whilst also recognizing the ways in which they interact in expert practice (as in Figure 1). This represents a significant methodological shift from many of the research papers that have previously explored learning using concept maps and that had assumed that complex knowledge may be captured in a single map structure.

**Author Contributions:** Conceptualization, I.M.K.; methodology, I.M.K. and A.M.; formal analysis, I.M.K. and A.M.; data curation, A.M.; writing—original draft preparation, I.M.K. and A.M.; writing—review and editing, I.M.K. and P.R.; supervision, P.R.

**Funding:** This research received no external funding.

**Conflicts of Interest:** Authors declare no conflict of interest.
