Cultures of Knowledge and Science

Three cultures of knowledge thrive in the modern university: natural sciences, social sciences, and humanities (Kagan Reference Kagan2009). There are commonalities and differences between these cultures. They are all supposed to be driven by truth in their attempt to describe and understand the world and motivate improvement of human life. They all involve a long history of scholarship which generates complex and demanding systems of knowledge and related professional activity. Mastering a discipline in each of them requires learning a body of knowledge which is often highly abstract and very distant from observable reality. One might mention here relativity in physics, DNA in biology, general intelligence in psychology, deep structure in linguistics, Keynesian theory in economics, literary theory in literature, etc. It also requires learning methods and codes of operation ruling how theories may be evaluated, combined, improved or rejected. In their professional applications, they involve knowledge and rules underlying problem-solving in related professions, such as the relations between the natural sciences and engineering, biology and clinical practice in medicine, theories of learning and education, etc.

These three cultures differ drastically, despite their similarities. The natural and social sciences are taxonomic, causal, explanatory, and predictive. Their primary aim is to model causal relations between variables and allow predictions for some of them based on others and their control, if needed. In the natural and social sciences, truth is approached by systematic observation, controlled experiment, and evaluation of findings by formal methods (mathematical or statistical modelling) allowing one to test if predictions and evidence fit. Theories are chosen based on their predictive accuracy, cohesion, and scope. However, modelling in the social sciences is more complex and less solid than in the natural sciences because the source of all units of interest, the individual, often unpredictably and uncontrollably, introduces variables not accounted for by the model being evaluated. Thus, the social sciences involve an extra layer of uncertainty that is not involved in natural science.

The humanities also try to understand and account for individual and social behaviour and its products such as literature, historical events, and ideas. However, this does not come through the scientific methods of truth control of the natural and the social sciences. In addition to personal observation, the primary source of information here lies in written texts or historical evidence about the phenomenon or the condition of interest. Their primary aim is rather to relate and evaluate approaches, interpretations, or sources of evidence rather than the behaviours or actions themselves. The very products of human activity, such as literary products, traditions, and ways of life are the primary object of analysis. Thus, interpretations are often perspective-based rather than based on ‘objective’ methods of testing truth. In the natural and social sciences, the concern about bias is central and, thus, methods for bias elimination are important in the process of discovery and modelling. In the humanities, bias may be part of the inquiry as such. Hence, the current discussions and theories about ‘post-truth’ or ‘meta-truth’, assuming that no objective truth is possible (Higgins Reference Higgins2016).

Obviously, there may be large variations within each of the three cultures. For instance, mathematics is very special because exploring and specifying formal relations between constructs dominate over empirically based relations. Legal science, placed between the social sciences and the humanities, prioritizes institutionalization of moral or social principles governing human behaviour, based on political decisions as prescribed by political institutions and processes. As a result, the study of law is a formal rational system guided by cohesion of principles, laws, and rules and their implications for the individual, social groups, and institutions.

Any aspect of reality may be understood at different levels. For instance, the pre-schooler understands that rain is water falling from the sky (realistic representational thought); the primary school child understands that it is recycled evaporated water going from sea to clouds and back (rule-based relational thought); the secondary school student may understand that it is a dimension specifically related to other dimensions, such as atmosphere temperature and pressure; the meteorologist (and the university student) understands that it is a parameter in a model allowing predictions of possible precipitation based on the values carried by other related parameters in the model. Also, any aspect of reality may be the object of any science: rain may be the object of meteorology, as above; the social sciences, when we explore how weather interacts with social behaviour and fashion; the humanities, when we explore tales, myths and traditions about rain across the globe and possibly related literature, etc.

Therefore, understanding across the three cultures of knowledge, the natural sciences, the social sciences, and the humanities, may be equally demanding. Concepts, constructs, models, and theories are extensively remote from first-hand observable reality. Regardless of the specific paradigm, elements of reality are reduced in generalized or idealized constructs multiply related in complex hierarchies, relations are often variably quantified, expressed in domain-specific language or symbol systems which must be mastered ad hoc. Further, models must be understood as such in each paradigm and approached as temporary: they are tools of interpretation for as long as no better model is available; however, they are always subject to improvement in case of a better interpretation. Therefore, a critical stance is part of understanding and improving knowledge; reasoning of different kinds is a tool in the service of both. Admittedly, the ultimate level of abstraction and complexity may differ across paradigms, depending upon tradition, level, and language of description, and methods of truth and bias control. For instance, models in high energy physics or astronomy are far more remote from reality and they are stated in a very idiosyncratic language as compared with models about literature, which express human experience and are stated in natural language. These differences may relate to differences between disciplines in their mental demands for understanding and operation, which in turn may differentiate between students in their possibilities of success in different programmes of study. We will discuss these questions below.

Educational Gaps in the History of Education

Any time there was an increase in the demand for a new level of education, many students entering this level were not spontaneously ready to cope with the cognitive and social demands involved. With the passing of time, education gradually acquired the experience to minimize the mismatch between educational and learning demands of the school and the cognitive and social ability of students, enabling the majority of them to cope. I outline the major milestones in the history of education in the last century to highlight my argument. The solution of the problems caused by these milestones may point to the solution of the problem discussed here.

Primary education has expanded drastically since the late nineteenth century and became compulsory in most European countries in the early twentieth century (Benavot and Riddle Reference Benavot and Riddle1988). The second industrial revolution in the late nineteenth century resulted in drastic changes in the knowledge and skills required for work in the industry of the time. As a result, basic literacy and numeracy skills and knowledge offered by primary education were necessary for everyone to be integrated in the new work and social environment. By the beginning of the twentieth century, primary education was compulsory in most European countries and North America (Murtin and Viarengo Reference Murtin and Viarengo2011). Compulsory primary education caused the first educational gap at that time, because many children entering primary school faced serious difficulties in learning to read, write, and grasp the basics of arithmetic and other concepts beyond everyday experience. Intelligence testing was initiated by Binet and Simon in France to help identify students in need of special support to learn at school. Testing spread quickly in other countries, such as the United Kingdom and the USA, because the need for it was high (Anastasi and Urbina Reference Anastasi and Urbina1997).

This gap does not exist anymore. Two types of solution were implemented. First, research on the learning processes related to the basic skills of numeracy and literacy and related teaching methods enabled education to develop more efficient methods for the teaching of these skills (e.g. Dehaene Reference Dehaene2010, Reference Dehaene2011; Kuhn Reference Kuhn1979). Second, institutionalization of pre-school education prepares young children to adapt socially and cognitively to the social and learning demands of primary school (Saracho Reference Saracho2015). Thus, transition to primary school is generally smooth for most children because education developed specific methods geared to the capabilities and weaknesses of children entering primary school, such as modern methods for teaching reading or arithmetic at first primary school grade. There may be difficulties for some children in several school subjects, such as difficulties in learning to read or do arithmetic, but these are treated as special learning difficulties that may be dealt with by special education (Demetriou submitted).

The next gap came with the third industrial revolution in the mid-twentieth century and the fast social, economic, and technological changes following the Second World War. These changes raised the standards of knowledge and skills needed relative to those provided by primary school. They include the ability to read and understand complex texts in different areas, think abstractly to deal with changing and varying demands at work, and function across different social conditions of the modern city. Initially, these changes resulted in a drastic expansion of the demand for secondary education in the 1950s and the 1960s. At that time, it was soon realized that many students entering secondary education could not cope with the new demands. The gap here was related to the ability for abstraction of underlying relations or themes beyond the apparent to give meaning to unpredictable changes at work or in the social environment, grasp the role of symbols standing for abstract ideas and operate on them as in algebra, and grasp non-observable ideas, such as electricity or radio frequencies from physics, which became part of everyday life. These needs caused serious difficulties to many students moving from primary to secondary education.

The need to bridge this gap coincided in time with the cognitive revolution of the 1960s and the emergence of several theories related to education, such as Piaget’s (Reference Piaget and Mussen1970), Vygotsky’s (Reference Vygotsky1986), and Bruner’s (Reference Bruner1973) theories. From a twenty-first century perspective, these theories may have proven inadequate to guide the development of specific learning environments tailored to the needs of different school subjects at different school grades. However, at that time, these theories were very instrumental in raising awareness about the cognitive and learning demands of several important goals of secondary education: dealing with possibilities rather than observed facts; grasping underlying relations rather than noting obvious properties and characteristics; using arbitrary symbol systems standing for abstract entities rather than realistic representations standing for personal experiences. These theories, at that time, motivated extensive research in cognitive development and learning of school-related concepts which, in turn, resulted in extensive improvements in the organization of primary and secondary school curricula (Kuhn Reference Kuhn1979; Chapman Reference Chapman2001). Additionally, junior secondary education became compulsory in most countries. Nowadays, compulsory education is nine years in most countries of the world (six years of primary education and three years of secondary education) (Murtin and Viarengo Reference Murtin and Viarengo2011).

Technically speaking, the primary–secondary education gap does not exist anymore. Bridging this and the earlier gap is obviously related to the fact that general intelligence increased in the general population by about 25 IQ points over the twentieth century, as discovered by Flynn (Reference Flynn2009). According to the results of the Program of International Student Assessment (PISA 2012), only a minority of the students at the end of their nine-year compulsory education operate at the two top levels of reasoning and problem solving addressed by the PISA tests in reading, mathematics and science. Specifically, a considerable number of ninth-grade students graduating from compulsory education in the OECD countries (about 20%) operate at or below level 2 in all three domains: in reading, students can identify the main idea of a text; in mathematics, they can connect similar pieces of information, extrapolate by inference, and apply basic algorithms to solve simple problems; in science, they can interpret simple observations or simple manipulations, but they cannot design controlled experiments manipulating extraneous variables. The majority of students, i.e. about 60–70% operate at levels 3 or 4: they may abstract a relation or idea running through several seemingly different contexts; they can grasp the meaning of abstract symbols and connect them to different realities or contexts as in algebra, and they can understand that things may be more complex than they look. These figures suggest that the transition from primary to secondary education works rather well.

The third gap, which appeared in the last 20 years, lies between secondary and tertiary education and is related to the so-called fourth industrial revolution. This revolution resulted in the drastic expansion and use of scientific knowledge in practically all aspects of everyday life (e.g. programming needs of home appliances and cars, acquiring or selling products through the internet, financial possibilities related to banking and the stock exchange, which require complex mathematics); unprecedented connectivity (e.g. internet, mobile phones, social media, and travelling); easy access to knowledge (e.g. Google). Access to problem solving through machine learning may be imminent. These changes created new jobs requiring long and complicated forms of education that can only be provided at the tertiary level (e.g. jobs related to computer engineering and information technologies) and necessitated upgrading studies previously provided by vocational or professional schools into university studies (e.g. car mechanic, nurse, accountant, or even teacher training). Also, the globalization of production and the movement of people transform societies and economies worldwide. Moving production from the West to the East and Africa resulted into the expansion of secondary education in these parts of the world. At the same time, it transformed western economies into service economies where science-based jobs (e.g. lawyer, engineer, biologist, and researcher in all domains) are in much higher demand than in the past. As a result, the increasing demand for university studies resulted in the university boom of the late twentieth and early twenty-first centuries, sending about 50% of secondary school graduates in many western countries to university. The EU target for 2020 is 40%; in some countries, such as Lithuania and Cyprus, it is already close to 60% and nowhere in the EU is it lower than about 25%.

In response to this demand, university systems expanded drastically. New universities were created in most countries, new programmes of studies were added to the traditional ones, and many professional schools in many countries, such as the UK, Greece, and Cyprus, were upgraded into universities. However, only a minority of students operate at the two higher levels of the PISA studies (10–15% in all three subjects), which require principle-based thought or formal thought, in Piagetian terms. Only at these two levels can students abstract ideas from texts in different domains and build a differentiated representation including central themes and complementary and conflicting modules. In mathematics, only students at these levels can model complex problem situations using formal mathematical language, specify similarities and differences between different problem situations, and reflect on, evaluate, and communicate their work to others. In science, students must operate at these levels to grasp proper scientific ideas such as evolution, test hypotheses by controlled experimentation, and integrate findings and hypotheses into models (NCES 2014; van Damme Reference van Damme2019). In other words, only a few of the students (about 15%) operate at the level of abstraction and mental flexibility required to follow university education as defined by the university. Yet, about 50% of each cohort continue to university. Simple arithmetic suggests that many students would face difficulties in following university studies. Below, we will first exemplify the cognitive basis of the gap and then advance several proposals that might help bridge it.

The Cognitive Basis of Educational Gaps

The human mind is a complex hierarchical system involving general and specialized mental processes carrying out different tasks for understanding, learning, and problem solving. Specifically, there is a powerful general cognitive ability guiding action, decision making, and problem solving. It allows the following: focus on information and process according to goals; select goal-relevant information and inhibit irrelevant information or actions; rehearse and relate information by inference; self-monitor processing and reflect on concepts and experience to reconstruct them so that they better represent the environment and guide action; when concepts resist understanding and problems resist solution, repeat recursively the processes above, varying concepts and ideas until understanding is recognized as better; encode new concepts and reconstructions into symbols, making their future use more efficient (see Carroll Reference Carroll1993; Demetriou and Spanoudis Reference Demetriou and Spanoudis2018; Jensen Reference Jensen1998).

Second, the processes are constrained by the representational and procedural specificities of different domains. For instance, social, quantitative and spatial information involve different mental objects and relations and they are prone to different types of representations, such as language in the social domain, mathematical symbolism in the quantitative domain, and mental images in the spatial domain. To function efficiently, general mental processes need to have facility with the representational and relational units defining each domain, such as words, grammar, and syntax in the verbal domain or numbers, arithmetic operations and algebraic relations in the quantitative domain, and visual images and mental rotation in the spatial domain (see Demetriou and Spanoudis Reference Demetriou and Spanoudis2018; Demetriou et al. Reference Demetriou, Makris, Kazi, Spanoudis and Shayer2018a, Reference Demetriou, Makris, Kazi, Spanoudis, Shayer and Kazali2018b). Variations in the facility to use different representational systems may end up in large variations in performance in learning concepts in different domains. Individual differences in learning and school success in different school subjects such as language-based subjects in the humanities and symbol-based systems such as mathematics relate to differences in this facility. Academic tilt at the university may derive from these differences (Coyle Reference Coyle2019; Demetriou submitted).

General cognitive ability changes with development, drawing on a different mix of general processes. Overall, it gradually shifts from executive processes underlying attention and working memory control to inferential processes underlying inductive and deductive reasoning and self-awareness processes underlying self-monitoring, self-evaluation, and self-regulation. Specifically, at preschool, thought is marked by attentional control and awareness of the perceptual origins of knowledge. Children understand that they know what they perceive. In primary school, thought is marked by inductive reasoning and awareness of inferential processes. Children start to understand that they may fill in lags in information by inference. In secondary school, thought is marked by deductive reasoning, mathematical reasoning, awareness of logical constraints of reasoning, and accurate self-evaluation. For instance, they understand that starting assumptions constrain conclusions and they know their strong and weak points. The cognitive profile of different age phases constrains what can be learned and what problems can be solved (Demetriou and Spanoudis Reference Demetriou and Spanoudis2018; Demetriou et al. Reference Demetriou, Makris, Kazi, Spanoudis, Shayer and Kazali2018b; Markis et al. Reference Makris, Tahmatzidis, Demetriou and Spanoudis2017). Notably, the processes dominating in different developmental cycles corresponding to different school levels, such as preschool, primary, and secondary school, are the best predictors of school achievement at this school level. Therefore, training programmes directed to cognitive enhancement at different school levels must be adapted to these priorities (Demetriou and Spanoudis Reference Demetriou and Spanoudis2018).

Changes in developmental priorities are associated with different representational needs and possibilities. For instance, speaking at the age of two years; reading and writing at 6–7; grasping underlying relations and using mental rules at 8–10; mental production of possibilities and grasp of principles allowing their use and evaluation at 11–17; epistemic understanding, that is grasp of the origins of knowledge from different sources, such as theory, experiment or reasoning, and understanding of constraints and possibilities of different conceptual or knowledge systems based on their origins, at 18–25. Mastering a new representational medium opens new cognitive possibilities and it may also be the source of new challenges in approaching the world and learning. The major argument of this paper should be clear by now. Successive developmental profiles are associated with major historical gaps in the development and expansion of education over the twentieth century. Bridging the first two gaps at the collective level resulted into major economic and social changes observed in the twentieth century, such as the massive improvement in prosperity, quality of life and health in those parts of the world where this bridging was successful. Nowadays, these gaps may still be present at the individual level as experienced by children facing difficulties in reading and writing and in mathematics at the beginning of primary school or learning the abstract concepts and symbol systems required at secondary school. In a sense, these difficulties are the residuals of the gaps described above that still need to be dealt with by modern cognitive developmental and educational science. However, the third gap, associated with the grasp of principle-based and epistemic thought by the general population, is still present, hindering the grasp of learning scientific theories and methods at university.

It took humanity a few hundred thousand years to move from oral (when Homo sapiens appeared about 200,000 years ago) to written language (about 5400 BCE). Likewise, it took humanity a few thousand years to move from written language to the invention of abstract symbolic systems, such as those used in modern mathematics and science (at about the sixteenth century BCE or much later). Also, it took a few thousand years to move from the natural philosophy of the Greeks to modern experimental science. However, each of these changes is condensed in about four to five years in individual development as coached by school (e.g. about four to five years to move from speaking at 2 years to reading or writing at 6–7 years or four to five years to move from reading to using abstract symbolic languages as an object of thought in secondary school and another five to six years to having to assimilate scientific theories and methods at university. Obviously, evolution did not endow humans with specific brain networks dedicated to the learning of these feats. They are cultural products requiring enormous brain-rewiring by individuals. It is not an exaggeration to say that modern education in a very short period of time wires cultural products formed over millennia into individual brains that did not evolve for them.

Here we focus only on the development of inferential and knowing processes and their vicissitudes after the age of 7–8 years. In cognitive developmental terms, the lower levels of the PISA studies address rule-based thought and the higher levels address principle-based thought. It is important to specify these types of thought because they are intimately related to the theme of this paper. Rule-based thought enables one to access individual representations (e.g. a series of numbers in an arithmetic task, multiple observations in a causal task aiming to interpret a phenomenon, threads of a story), align them and identify relations between them (e.g. numbers double, a particular button causes the effect to be interpreted, all heroes of a story work for the same aim). Inductive and analogical inference are the major tools of rule-based thought (Gentner and Hoyos Reference Gentner and Hoyos2017). Rule-based thinkers can think deductively (e.g. if they know that ‘if A then B’ they infer that B will occur if A occurred and that if B did not occur A did not occur either). However, they may fall prey to logical fallacies, assuming, for instance that if B occurred, A will also occur, ignoring that B may be caused by other factors as well (Stanovich Reference Stanovich2011). Rule-based thinkers are aware of inference and its role in generating new knowledge (Demetriou et al. Reference Demetriou, Makris, Kazi, Spanoudis and Shayer2018a). However, they are inaccurate in self-evaluations and self-representations, often resulting into optimistic characterizations of their own performance. Thus, special support is needed to help them regulate their learning behaviour so that their learning efforts are successful.

Principle-based thought allows a grasp of higher-order relations between rules, generation of alternative possibilities, and evaluation of the cohesion of concepts based on logic. Principle-based reasoning is a logical metaprocess that defines acceptable and non-acceptable inferences. Thus, deductive reasoning at this level allows resisting logical fallacies because each inference is considered from the point of view of alternative complementary inferences. For instance, they understand that ‘if A then B’ does not imply that ‘when A does not occur B does not occur either’ because B may occur for reasons other than A. Thus, truth control is integrated into a system of rational analysis and related actual empirical control that may rule out confounding variables. Principle-based thinkers are aware of the logical constraints underlying the relations between concepts and they systematically search for them to examine if they are present in an argument. Moreover, they are relatively accurate in their self-representations and self-evaluations. They have a differentiated self-concept where strengths and weaknesses are explicitly specified. Thus, they are able to monitor and evaluate their problem-solving endeavours and self-correct, if necessary. This allows them to make long-term choices, such as a course of studies or a profession.

Epistemic thought uses principle-based reasoning for the sake of grasping systematic fields of knowledge, specify similarities and differences between fields, and to locate gaps between concepts and observations, or inconsistencies between concepts, and tries to remove the latter by inventing new concepts accommodating and explaining gaps and inconsistencies. Epistemic thinking also enables one to choose between the form of solution appropriate for different types of problems (Kitchener Reference Kitchener1983; Seppälä et al. Reference Seppälä, Lindblom, Kallio and Kallio2020). Epistemic thinking and underlying principle-based reasoning allows systematic long-term work on competing or complementary systems of knowledge for the sake of integrating them into new knowledge systems or paradigms. One may invoke examples from the history of science here, such as Maxwell’s equations for uniting electricity and magnetism, Newton’s theory of gravity integrating force and movement, Einstein’s theory of relativity for gravity, movement, space and time, Darwin’s theory of evolution integrating palaeontology, zoology and heredity, or Freud’s theory of behaviour and mental illness (see Kitchener Reference Kitchener1983; Richards and Commons Reference Richards, Commons, Commons, Richards and Armon1984; Seppälä et al. Reference Seppälä, Lindblom, Kallio and Kallio2020; Kallio, Reference Kallio2020).

Prevalence of Cognitive Developmental Levels and Cognitive Ability

How are these patterns related to cognitive ability in the general population? Answering this question requires an integration of psychometrics with the developmental expression of intellectual attainment. This would enable transferring knowledge from developmental research to learning at school. In view of this aim, we transformed attainment on tests of reasoning development into an IQ-like score. According to our transformation, an average IQ of circa 100 points corresponds to rules-based thought attained at the age of 9–10 years. Provided that intelligence is normally distributed in the population, this implies that the majority of people function with rules-based thought. Intelligence higher than 120 IQ points corresponds to principle-based thought. According to many studies (Demetriou and Spanoudis Reference Demetriou and Spanoudis2018; Demetriou et al. Reference Demetriou, Makris, Kazi, Spanoudis, Shayer and Kazali2018b; Shayer and Adey Reference Shayer and Adey2002), only the top 15% of 18-year-old students reach the top levels of principle-based thought in different domains, such as mathematics and scientific reasoning, even when systematically trained.

Understandably, the concentration of principle-based and epistemic thinkers is higher in universities than in the general population. In selective universities where students are admitted on the basis of nationwide competitive examinations, such as the SAT (USA), the GCSE (UK), or the Baccalauréat and Grandes Écoles (French examinations), about half of undergraduate students (48%) operate at principle-based thought level; about 10% in the humanities but 40% in the sciences operate at various levels of epistemic thought (see Demetriou Reference Demetriou, Commons, Richards, Armon and Sinnot1990). These levels correspond to an IQ that is higher than 130.

These findings suggest that there may be multiple gaps between secondary and tertiary education. The first gap is related to the attainment of principle-based thinking. Principle-based thinking is the minimal requirement for grasping science concepts as required by the university. According to the evidence above, about 15% of each cohort functions at this level when they graduate from secondary school. Yet, depending upon which Western country we are looking at, about 50–60% of secondary school students go to university. Overall, then, about 35–45% of students admitted to a university would face difficulties coping with the demands of university education. Obviously, there are differences between universities, because highly selective universities do have a high concentration of students coming from this upper 15% of cognitive ability. This may be as high as 80% in Greek (Demetriou Reference Demetriou, Commons, Richards, Armon and Sinnot1990), Finnish (Seppälä Reference Seppälä2013), or American (Wai Reference Wai2013) selective universities. The second gap is related to the attainment of epistemic thought, which is much rarer than principle-based thought. The majority of students operating with principle-based thought (about 60%) do not have the epistemic understanding required to grasp science as a model-theory building enterprise that is under continuous revision. These students may grasp abstract science concepts as such, but they do not embed them into the epistemological constraints and nuances of a discipline, let alone science at large. Seppälä (Reference Seppälä2013) found that there is a difference between classic universities (60–80%) and universities of applied science (50–60%). Supposedly, going through the years of undergraduate studies aims to build the epistemic mind required to grasp these aspects of science and operate with them. However, according to Seppälä (Reference Seppälä2013) this does not occur; she found that significant changes in the conceptions of scientific thinking do not necessarily take place during the years of university studies.

Expanding demand for university education leads many people to university although they are not ready to follow university studies as required. The challenge of our time is to bridge this gap by preparing these people to study as required. This may be the revolution of twenty-first century education in the way that bridging the primary and secondary education gaps was the basis of the revolution of twentieth-century education. Arguably, there is a Darwinian-like market force activated by this increased demand for university studies. The university system expanded by creating thousands of new universities catering for students who fail the standards of selective universities. The implications are grave for individuals and society: graduates miss the opportunity to expand their capabilities and knowledge and professional skills at the level expected from university studies; society suffers from problem-solving inferior to the needs and expectations of the complex world of our time (Arcidiacono et al. Reference Arcidiacono, Aucejo and Spenner2012; Wai Reference Wai2013).

What Must be Done

Bridging the secondary–tertiary education gap must draw on how the two previous gaps were bridged. I recall that changes occurred on both sides of these gaps. To bridge the first gap, preschool education was introduced, and many adjustments were made in the primary school curriculum and teaching methods. To bridge the second gap, junior high school became compulsory and many changes occurred in curriculum and teaching methods on both sides of the gap. To bridge the secondary–tertiary education gap, deeper changes are required, probably starting at the end of primary school. These changes are outlined below.

We noted that many secondary school students who do not attain principle-based thought end up at a university. Therefore, education must be redesigned to ensure that all students going to university grasp principle-based thought at the minimum level required for university studies. This requires specific programmes aiming to support the transition from rule-based to principle-based thinking upon transition from primary to secondary school. At junior secondary school principle-based thought must be consolidated well as an approach to problem solving. Examples of relevant programmes are given below (see also Herrnstein et al. Reference Herrnstein, Nickerson, de Sanchez and Swets1986; Klauer and Phye Reference Klauer and Phye2008). In senior high school, which is the doorstep to the university, principle-based thought must be systematically practised in all subjects, i.e. language and humanities, mathematics, and natural and social sciences. Probably, an extra pre-university preparatory year may be compulsory for students planning to follow university studies involving programmes designed to train epistemic thought. Some examples follow below.

Where the Gap Must be Bridged

Understandably, one might object that the intervention programmes addressed to bridging the gap between secondary and tertiary education may be demanding for secondary school so that it might be preferable for them rather to be implemented in the early years of university studies. Obviously, the terrain of these problems is not well charted, and more research is needed to guide decisions about how to prioritize actions that are more suitable for secondary education and actions that are more suitable for university. Perhaps, a combination might be better. On the one hand, a long-term programme for training principle-based thought starting from primary school and epistemic thought starting in late secondary school. This might culminate in a preparatory year at the transition from secondary to tertiary education for the students planning to go to university.

On the other hand, this would require that the university recognize its responsibility for preparing admitted students to grasp science at the level required. This would require a new stance from universities and related services. For instance, using diagnostic tools to identify students in need of the kind of support discussed here comparable to the way such tools are used, for instance, for pupils needing special support to learn reading and writing at the beginning of primary school. Based on the needs and individuals identified, the university would have to provide teaching programmes implementing the goals outlined above and staff trained to implement these programmes. I am aware that for many professors this approach may sound farfetched and alien to the university. It must be recognized, however, that these needs are here to stay and will still increase, and we better receive the message rather than kill the messenger. Obviously, more research and thought are required before these issues are resolved at the level of policy making.