Concept maps have been shown to have a positive impact on the quality of student learning in a variety of disciplinary contexts and educational levels from primary school to university by helping students to connect ideas and develop a productive knowledge structure to support future learning. However, the evaluation of concept maps has always been a contentious issue. Some authors focus on the quantitative assessment of maps, while others prefer a more descriptive determination of map quality. To our knowledge, no previous consideration of concept maps has evaluated the different types of knowledge (e.g., procedural and conceptual) embedded within a concept map, or the ways in which they may interact. In this paper we consider maps using the lens provided by the Legitimation Code Theory (LCT) to analyze concept maps in terms of semantic gravity and semantic density. Weaving between these qualitatively, different knowledges are considered necessary to achieve professional knowledge or expert understanding. Exemplar maps are used as illustrations of the way in which students may navigate their learning towards expertise and how this is manifested in their concept maps. Implications for curriculum design and teaching evaluation are included.