**1. Introduction**

The primary focus of 21st century education is to support students to develop meaningful knowledge that can be applied to a range of evolving, real-world settings [1–3]. The world with all its complexity—including a rapid growth of information and knowledge, along with increased pressures on the educational system—creates a challenge to help students to develop the skills to navigate these complexities. Therefore, the key role of curricula at school and at university is to promote theoretical knowledge that underpins evolving practice, and to help students to navigate between theoretical and everyday knowledge and between di fferent kinds of theoretical knowledge [3]. Additionally, learners in higher education have to be prepared with appropriate, authentic contextual knowledge to ensure graduate employability [4].

In any discipline, novices tend to have loosely organized knowledge, where concepts and strategies are not well linked, while experts have a highly organized and well-structured knowledge base that allows them to use information meaningfully to solve problems [5,6]. Rather than adopting a trial-and-error approach that is typical for a novice, we need experts that can use a principles-based approach to solve problems [7]. With this in mind, several researchers have demonstrated the benefits of concept mapping in teaching, learning and assessing scientific subjects. The use of concept maps has been shown repeatedly to be an e ffective tool for improving conceptual understanding [8–12], developing higher-order thinking skills [13], revealing misconceptions [14,15] and eliciting achievements and grades [16]. Therefore, we ask: what do concept maps reveal if we explore di fferent types of knowledge (novice, theoretical, practical and professional) in students' concept maps?

The scoring of concept maps and the awarding of a single number to summarize map quality may give an indication of how much information a student has acquired during his/her study, but it does not provide any indication of the types of knowledge that have been acquired (e.g., conceptual or procedural) or the relationships that the student has identified between knowledge types. This recognition if different knowledges has been described as essential for developing the basic characteristic of the expert student [17] who needs to recognize the existence and complementary purposes of different knowledge structures. This has been overlooked in the research literature on concept mapping that has tended to foreground the development of conceptual knowledge to the exclusion of procedural knowledge. The focus on the development of a discrete single map structure has emphasized this bias in knowledge type, with procedural knowledge often being buried within a map of conceptual knowledge. Kinchin and Cabot have discussed how expertise requires the oscillation between linear structures of procedural knowledge and networked structures of underpinning conceptual knowledge, but they did not offer any framework to assess the relationship between the two or how this may evolve over time [18].

In this paper we explore how different types of knowledge are embedded within a concept map and interact to each other. Concept maps that represent learners' knowledge structures have been associated with meaningful learning theory [19] and the promotion of higher order thinking skills [13]. Here, we present a major shift in emphasis in concept map evaluation by considering the analysis of concept maps in relation to the semantics dimension of Legitimation Code Theory [20]. This not only provides a commentary on the student's progress, but also offers a critique of the curriculum experienced and the way in which it facilitates (or not) a student's development from novice to expert. Here, expertise is considered to be derived from the purposeful interaction of different knowledges (as described by [21]). We present examples of student maps that illustrate the way in which students may navigate the curriculum and argue that, in most cases, students do not reach the level of professional understanding.

The expert structure that represents professional knowledge is explicit in the integrated nature of theoretical knowledge and the way in which this underpins the procedural knowledge that constitutes the visible practice that defines a professional [22]. The derivation of chains of practice from theoretical knowledge is one of the hallmarks of expert knowledge [18]. However, we should not be surprised that this expertise is rarely exhibited by students, who grapple with their understanding of concepts before they are able to distinguish between conceptual and procedural knowledge, or that it is rarely depicted in concept maps that generally aim to combine procedural and conceptual knowledge within a single structure. The example of professional knowledge given in Figure 1 (of local anesthetics in dentistry) shows how knowledge that has a high semantic density and low semantic gravity, SD+SG- (such as physiology and pharmokinetics), determines the structure of the theoretical knowledge to the right, whilst the chain of practice to the left is composed of concepts such as instrument assembly and techniques, which exhibit lower semantic density and high semantic gravity (SD-SG+). In this paper, we explore the possibility of locating elements from the practical and the theoretical in students' emerging understanding of a discipline as an indicator of their current status on the journey through secondary and higher education towards professional knowledge.

**Figure 1.** The semantic plane in which each quadrant has been populated by the archetypal map morphology (spoke, chain and network) that is likely to be found there, with (inset below) an example of a well-defined expert knowledge structure in which practice and theory are clearly delineated as complementary chain of practice and network of understanding [17,20,23].

### **2. Theory of Concept Maps**

Concept maps have their roots in Ausubel's meaningful learning theory, and they emphasize the connections among concepts that represent individuals' knowledge structure [10,24,25]. There are three elements from Ausubel's theory that Novak and his research team found useful to develop in the concept mapping method:


(3) Meaningful learning takes place when relationships between concepts are explicit and are better integrated with other concepts and propositions [10].

Concept maps are composed of concepts that are written in boxes and connected with arrows that are labeled to indicate the relationship between concepts [26]. The labeled connections between concepts are called links, and each 'concept-link-concept' forms a proposition that can be read as a stand-alone meaningful expression. Cross-links, which might sometimes be formed, show the relationships between two di fferent areas of the map [27]. Concept mapping is a skill that encourages nonlinear thinking [28]. The construction process of concept mapping helps the learner to actively construct their knowledge and, as suggested by Hyerle [29], helps students to "think inside and outside the box". The important function of this graphic representation is to display the overall arrangemen<sup>t</sup> of concepts and the enhancement of metacognitive skills [7,12]. According to Salmon and Kelly, concept mappers with these skills are able to (1) define specific thinking process as recurring patterns; (2) support the transferring these patterns across disciplines; (3) guide the building of simple to complex mental models and (4) reflect how the frame of reference influences their meaning-making, thinking patterns and understanding [7].

Kinchin expresses the benefits of using concept maps by saying, "This is a tool that helps me not only to see how the students are putting ideas together (or not), but can also help the students to diagnose their own di fficulties" [17]. Much school learning is achieved through rote learning, while using strategies like note-taking, rewriting the textbook pages, summarizing as bullet points and completing 'fill-the-gap' test that are not as productive as concept maps to develop well-organized knowledge. Thus, learners who are used to learning through rote learning find the higher level thinking that is required to construct a concept map challenging [13]. Concept mapping has also been proposed as a useful tool to support the learning of complex topics, where learners have fragmented understanding and might face di fficulties integrating all components to form a meaningful overview [12]. The external sca ffolding that the concept mapping process involves can be very helpful to support deep thinking and complex learning [7].
