In the last two decades much work has been published on what history should be taught, the educational value of historical knowledge, and the importance of learning about history. We support a learning process based on abilities that allow students to interpret the past, and which go beyond factual or conceptual knowledge based on the memorization of dates, characters or facts. It is about an approach focused on history comprehension as a method. It consists of learning history as a form of research and learning to think and reflect on the past (Barca, 2000, 2011; Carretero & López, 2009; Carretero & Voss, 2004; Clark, 2011; Gómez, 2014; Gómez & Miralles, 2015; Lévesque, 2008, 2011; Peck & Seixas, 2008; Prats & Santacana, 2011; Saiz, 2015; Sáiz & López Facal, 2015; Schmidt & Barca, 2009; Stearns, Seixas & Wineburg, 2000; Van Sledright, 2011; Wineburg, 2001). These studies have distinguished between two types of historical content.

On the one hand, substantive or first-order content is related to concepts or principles such as dates and specific historical facts. On the other hand, studies distinguish the strategic content of second-order or historical meta-concepts. These are defined by the knowledge and use of different capacities or competences to give an answer to historical matters and gain a more complex understanding of the past: this is knowledge related to historical skills, such as searching for, selecting and processing historical sources, empathy or historical perspective (Van Sledright, 2014; Wineburg, 2001).

According to Seixas and Morton (2013), historical thinking can be defined as the creative process that historians go through to interpret sources from the past and generate historical narratives. To develop this process, six key concepts must be considered: historical significance; historical evidence; change and continuity; causes and consequences; historical perspective; and the ethical dimension of history. This proposal fits into the context of an ever greater interest, in the usa and Canada, in the definition of knowledge about history (Barton & Levstik, 2004; Stearns, Seixas & Wineburg, 2000; Van Sledright, 2011; Wineburg, 2001). In addition to that, from an empirical perspective, we wanted to place emphasis on the abilities of thinking, reading, writing and giving historical support, which are considered historical literacy (Monte-Sano, 2010; Wineburg, Martin & Monte-Sano, 2013).

In Britain a great effort has also been made to analyze second-order concepts in student reasoning. In that regard it is important to mention the results of the project Concepts of History and Teaching Approaches (Lee & Ashby, 2000; Lee, Ashby & Dickinson, 2004; Lee 2005). The early stages of this project were based on an analysis of students’ historical skills through the use of sources and historical empathy exercises, and more recently through different ways of setting out the fortuity of historical facts. On the other hand, the work by Chapman (2011) and the project Constructing History 11-19 (Cooper & Chapman, 2009) have furthered this line of research through reasoning on the use of sources. A thematic line has also been applied in countries such as the Netherlands (Van Drie & Van Boxtel, 2008).

These studies have agreed on the need to have teachers that question the cognitive learning model of history (VanSledright, 2014). Contrary to common belief, knowing history does not mean memorizing facts, concepts and dates. Understanding history implies other aspects related to complex thinking processes. This type of historical thinking is acquired through gradual participation in the practice of the history (Gómez, Ortuño & Molina, 2014). History is a discipline that takes as its object the study of the past and is formed in a community of researchers that set out questions and approach the footsteps of the past through critical methods.

The perspective through which the past is treated depends on sociocultural and temporal factors that vary with the identity, beliefs and values of the researcher himself. History is a construction and it must be taught as such in the classroom. History teachers need to have a solid theoretical grounding in the formation of historical thinking and comprehension in students, the learning process of the discipline, and in identifying indicators of cognitive progression.

However, in recent years several authors have warned of the low value of a "skills versus conceptual content" debate (Cain and Chapman, 2014). As Lee (2011) indicates, historical knowledge implies reflection and critical judgment, and cannot be acquired simply through repetitive practice of unconnected skills lacking proper reasoning and knowledge of key historical processes and concepts. Students must have a clear understanding of the historical discipline and the key ideas that make knowledge of the past possible, including how historians explain the past and construct their historical narratives (Seixas and Morton, 2013). Students should be inclined to make the best possible arguments based on evidence, and notably historical evidence. But they also need to have coherent and substantive historical knowledge that allows them to use this evidence properly, at different scales, and with correct temporal understanding (Lee, 2011). However, teachers often find themselves up against a major sticking point: students’ misconceptions about historical reasoning and interpretation. Lee (2001) and Lee and Shelmit (2004) have studied in depth the ideas and preconceptions of students in this respect, and the epistemological problems that result from them.

Students' historical reasoning and their narrative representation of the past are key aspects of recent studies on history teaching. Narration has begun to be valued as a teaching approach (Henríquez, 2004) and as an important subject of research (McEvan & Egan, 1998). As stated by Carretero & Van Alphen (2014), knowledge of history and the construction of narratives are closely related. The exclusion of narratives from history teaching impoverishes the focus of classes and deprives social events of their singularity. In addition, it should not be forgotten that the development of narratives can be a fitting way of explaining, and introducing children to, historical processes and concepts so they may begin to understand them.

Narrative thinking is a basic mental operation that provides meaning and enables organization of the past (Carretero & Atorresi, 2008; Ricoeur, 1995; Rüsen, 2005 and 2010). If we ask students for a written synthesis, we discover not only what processes the students know, but also how they represent and organize them. As underlined by Barton and Levstik (2004), implicit in the narrative discourse are the use of the time dimension, the establishment of causal nexuses, and some means of granting historical importance and of connecting present and past from a moral perspective. Thus, using narratives with students allows one to reflect on history as a construction, because the narration is not simply an account of what happened but implies much more about the psychological perspective adopted by the narrator (Bage, 1999; Monte-Sano, 2010; Topolsky, 2004).

It is precisely the construction of narrations that enables us to value the internal processes of knowledge building. According to Bruner (1986; 1990), it enables us to discern both the level of substantive knowledge (enumeration of sure content about an event) and the level of awareness of second-order content (what we know about beliefs, motivations, emotions and willingness in past actions). At the same time, accounts provide information about progress in students' learning and their capacity to organize and understand the past. As McAdams (2006) states, the structure of the narration, its justification and causality, and the paradoxes reflected in narrative composition are a clear indicator of how mature a student's thinking is. As Bruner (1997, p. 168) puts it “We live most of our lives in a world constructed according to the rules and devices of narrative”. Given these circumstances the analysis of narration in history teaching is gaining ever greater importance, not only in the research into the accounts constructed by students, but also in the analysis of how the main national narratives are chosen and their interconnection through formal and non-formal teaching (Carretero & Bermúdez, 2012).

Skills related to narration, such as "narrative competence", are a way of representing or bestowing meaning on the past through three specific competencies that narration encompasses (content, form-meaning, and function): “temporal competency” (content), as knowledge of past phenomena or content; “interpreting competency” (form-meaning), as the capacity to find meaning in this content; and "orientation competency” (function), as the disposition to put the past to use in guiding the present and in moral values (Rüsen, 2010). Students' narrative examinations show how they express different historical meta-concepts. The influence of cultural studies and the linguistic turn have meant that increasing attention is being paid to the discursive analysis of students as narrators, and to the ways in which they express their ideas about the past (Barton, 2010). In this sense, great importance is attached to the “schematic narrative template” concept used by Wertsch (2002) as a cultural mediation tool generated and spread among a social group, and which is perfectly applicable to historical and national narratives (Carretero & López, 2010; Carretero, López, Fernanda & Rodríguez-Moneo, 2012). There is a powerful narrative mediation in knowledge and school historical memories on account of the very weight of national narratives in the curriculum and textbooks, and this conditions the content of the historical discourse that students can provide. As a result, no analysis of students' narrative productions can in any way be divorced from the "mediating" elements of the historical discourse.

The results confirm the low level of narrative skills and historical knowledge of students in the degree in elementary school teaching at the Universities of Murcia and Valencia. First, we should note the low level of substantive content, what students knew about the content requested, even though it came from their basic training and is informal knowledge. Second, there was an absence of skills in organizing an account, in how they narrated or presented the content, from the standpoint of SOLO taxonomy and in terms of the second-order or historical meta-concepts present.

Table III

Historical content and historical metaconcepts (1=null; 2=basic; 3=medium; 4=high). SOLO Taxonomy (1=Pre-structural; 2= Monostructural; 3=Multi-structural; 4=Relational).

Half of the students offered no noteworthy information and a similar percentage showed minimum communication skills; 46% of students were unable to articulate their account acceptably when the SOLO taxonomy was applied (table III). Apart from spelling and grammatical errors, the most outstanding feature of the accounts was their disjointedness. A large number did not meet the minimum requirements expected of higher education training for students to become elementary school teachers. Similarly, 50% of the narratives were seriously deficient in substantive historical content. Some students claimed the Muslims entered the Iberian Peninsula from the north (through France) and even spoke of the Christian kingdoms as outside entities. Several set the events in some pre-Christ period, while others had them occurring much later than they did. Also indicative was the fact that many of the accounts claimed that the Muslims had ousted the Romans from the Iberian Peninsula, when they actually arrived almost 250 years after the fall of the western Roman Empire.

Approximately 40% of narratives did meet the minimum levels of communication expected of higher education students, although these were mostly linear, descriptive, and one-dimensional accounts. They dated the Muslim invasion (711) and their final expulsion (1492) correctly, and the accounts included names of considerable importance. Yet they lacked control of second-order concepts; the historical phenomena included were limited to basic political-territorial facts, which were at times very fragmented with wayward time references. The events and figures most referred to were “Pelayo”, the “Battle of Covadonga”, “El Cid”, the “Catholic Monarchs” and the fall of Granada in 1492. In Valencia there were also allusions to James I (Jaime I).

When analyzing historical thinking capacities, we found that just 9% of students showed any consistent mastery of one or more second-order concepts, i.e. an average to high level. Indeed, 70% of accounts did not include markers showing the importance of the historical process at even a low level. Most students narrated the process disjointedly in terms of its social and economic reality or with respect to today's territories. The most frequent interconnection is the Arab cultural and artistic legacy (monuments, mosques, etc.), although it is anecdotic. Furthermore, 85% made no reference to causes and consequences, to the motivations that led to the advance of the Christian kingdoms in the Muslim territory, or to the different paces of the process. Many of the reasons outlined are lacking in historical thinking skills. For example, some students wrote that the Christians “decided one day that they had to reconquer the Iberian Peninsula, and they did so quickly", a crass simplification of an 800-year process that left a footprint that is still very present in the Spain of today. Likewise, 86% of narratives did not develop arguments about the changes and continuities associated with the process, while 96% included no marker associated with the concept of historical awareness or the ethical dimension of history; these students’ work showed no value judgments connecting past events to current social, economic, political or cultural configurations. Furthermore, students that did use these historical meta-concepts did so, for the most part, at a very low level.

Contingency tables were used in spss to test variable interaction and reveal significant results. First, the complexity of the accounts and the presence of historical thinking markers increase proportionally to an increase in substantive content. This would point to a close association between the two variables and to a need for a teaching that combines historical content with historical thinking competencies, but without ever losing sight of the interrelations between them (Lee, 2005). To continue to put forward dichotomies between content and competencies when teaching history is reductionist and counterproductive (Cain & Chapman, 2014).

Second, students who displayed the most historical information (in first and second-order concepts) showed greater organizational capacity according to the SOLO taxonomy. This is seen in tables IV, V and VI, which show a contingency table, the Pearson chi-square and the Kendall Tau-b in order to verify the hypothesis of independence between the variables (level of historical thinking in the narratives and the level of complexity in the accounts) and the degree of dependence between them. The chi-square value (487.475, with a bilateral asymptotic significance of .000) implies that the variables “historical thinking” and “SOLO typology” are dependent. The presence of second-order historical content is conditioned by the complexity of the narrative. Dependence between variables, according to the Kendall Tau-b, reached .906, confirming a strong dependence.

Table IV

Table V

Table VI

a. Assuming the alternative hypothesis.

The scarce skills shown by students in organizing historical discourse is due to the design of the syllabus for the degree in elementary school teaching in Spain, and an irregular and insufficient history education at lower levels. There is still a predominance of history teaching based on memorizing and reproducing factual and conceptual content, as is borne out in elementary and secondary school examinations (Gómez & Miralles, 2013; Sáiz & Fuster, 2014). Furthermore, incorporation of basic educational competencies is not remedying this deficiency, which is the fruit of school routine arising from a treatment of history in curricula that has been repeated and maintained in textbooks and teaching practices (Gómez, 2014; Sáiz, 2011 and 2013).

Both Wineburg (2001) and VanSledright (2011) have pointed out the dangers besetting a linear teaching of history, based on the building of a nation, that seeks to compete with a mass consumption culture. An approach to historical content that seeks to consolidate a collective memory of a nation inculcates a passive role in students. Besides, as VanSledright (2011) points out, this converts the teacher into a narrator of the nation's deeds, with an emphasis on its most emotional aspects, and aims, in the long run, to capture the attention of a student accustomed to audiovisual consumption. This has several dangers: the low cognitive level demanded by this type of teaching, and students’ reluctance to understanding, from multicultural contexts, history that has been already written from a specific perspective.

The results from the analysis of the narratives are a direct consequence of the hegemony of this descriptive historical discourse, which treats historical time one-dimensionally and is based on the building of a nation and the reproduction of the master narratives. This causes an epistemological conception of history as closed knowledge. There is a dominance, in the teaching of this subject, of an approach based on a linear account of past facts that have been selected by the official curriculum, the textbook, the center's programme, and finally, the teacher. Here we should remember the concept of narrative mediation and the narrative schematic template (Wertsch, 2002). The general structures of these master narratives, as is the case with the historical synthesis requested herein (the Christian expansion over Islam in the medieval Iberian Peninsula), changes very little over the years. Moreover, these national narratives often weigh heavily in both formal and informal education (Sáiz, 2015), hence their huge influence on how students interpret historical phenomena (VanSledright, 2008). Indeed, some of the results of the study bear this out. In recent decades the most significant change in the structure of the national narrative is that textbooks have incorporated more information on the Muslim cultural and artistic legacy (the Mosque of Cordoba, the Alhambra in Granada…) and an idyllic vision of coexistence between cultures (Sáiz, 2012). This has been confirmed in students’ accounts, with almost 30% making reference to this issue.

Neither is it surprising that almost 65% of the students referred to the process as the “Reconquista”, or reconquest (table VII). Those using this term also produced accounts of greater narrative complexity and with more substantive content. In other words, the lack of references to the process as the “Reconquista” was not due, in most of the accounts, to doubt over its historiographical validity, but because students probably did not situate the historical process correctly. It is true that the term “Reconquista” continues to provoke academic controversies historically and in the teaching of history. Its detractors claim that it is a historiographical construction that seeks legitimacy as a national narrative (López et al., 2015; Ríos, 2011; Sáiz, 2012 and 2015). Others, however, accept the inexactness of the term but hold the opinion that it remains an operative one for defining a key process in the Iberian Peninsula in medieval times (García-Fitz, 2010). In any case, the weight of this denomination when narrating the historical process in both textbooks and in student assessment shows the pervasiveness of a traditional and very classical vision of the historical phenomenon. Much research confirms the use of the term “Reconquista” and events and characters involved in the process. The denomination enjoys a hegemony in secondary education history manuals (Sáiz, 2012), and examinations and manuals in elementary education continue to feature mythical names such as “Pelayo”, who is presented to students as a real historical person, despite his scant presence in primary sources. The knowledge imparted, on the other hand, is practically reduced to accounts of the beginnings of the Christian conquest and the similarly disputed Battle of Covadonga (Gómez, Rodríguez & Miralles, 2015). The Catholic Monarchs appear as the great unifiers of what will become the Spanish nation, with the climax being the fall of Granada and the end of Arab rule in the Iberian Peninsula in 1492. Again, these narrative mediators - the textbooks and student assessment - have an impact on the direction taken by the students' accounts.

Table VII

With respect to the aims of this study, the results show that future elementary school teachers have not received an appropriate history education in substantive content or historical thinking competencies. Moreover, their accounts reproduce the identity topics set out in the textbooks and which are required of them in the examinations. It is difficult, therefore, for them to meet the minimum subject and teaching levels required, at least with regard to the standards that today's research into the didactics of social sciences and history education deems acceptable (Wineburg, 2001; Barton & Levstik, 2004; VanSledright, 2014). In these circumstances there is a need for a transversal intervention to increase the presence of history education (in both substantive content and historical competencies, which are closely related) in basic levels of education (elementary and secondary), as a complement to future teachers’ training.

The students' accounts reflect the dominance of a national narrative that presents itself as a closed, unquestioned issue. There is an insistence on the achievements of a nation of classical origins (with a particular influence from Romanization), which was enriched through mixing with other cultures, and then reaffirmed itself by driving out the Muslims before gaining fame as the discoverer of new lands and finally becoming an exemplary democracy. The historical downturns and discontinuities are glossed over or played down. One of the main reasons is to be found in the curriculum, which is encyclopedic and culturalist. Textbooks and examinations are the fruit of a legislation that, in Spain, is under continuous change, but which in the sphere of school narrative history shows more continuities than changes. Textbooks and examinations in basic education, as narrative mediators, leave their traces in the accounts analyzed in this study. The persistence of these identity topics stretches beyond compulsory education. The fact that it still persists in future teachers is a serious problem, as the likelihood is that these will pass it on to their future pupils.