Introduction

Debates about the place of scientific knowledge in teacher training and professional development have historically been driven, on the one hand, by applicationist approaches, with scholarly knowledge having to be implemented in action in one way or another, and, on the other hand, by those in which the expert know-how of experienced teachers is to be promoted. Today, the professionalisation of the teaching profession implies the permanent learning of a relevant and efficient way of acting in an ever-changing situation (Etienne, 2008), with inventiveness thus being put forward. In this context, training becomes support. Despite some discordant voices, a consensus around the paradigm of the reflective practitioner (Schön, 2017) seems to be relatively established. But what place can scientific knowledge have in this paradigm? What kind of science are we talking about, and how can it be linked to training? What kind of research results are employable by the practitioner, and in what ways?

Huberman (1982) points out that the results of educational research are difficult for teachers to use as they are. Moreover, teachers are not very familiar with this type of research, and do not see how it could be useful to them (Donnay et al., 2002). The influence of educational research on teachers’ practice is thus not very clear for some authors (Shkedi, 1998). Teachers would especially like to read research that relates examples of pedagogical actions or results in knowledge that is directly transferable to their practice (Zeuli, 1994). Some practitioners even question the validity of educational research because the results do not correspond to what they experience in the field (Hemsley-Brown & Sharp, 2003).

According to Rayou (2008), the ‘real’ training of trainee teachers is in the field, not in training institutions. Very little transversal knowledge relating to the humanities is considered by these trainees. Knowledge derived from research is rejected as not being directly applicable in the classroom. Thus, many young teachers find their training too theoretical, because it does not really enable them to teach. They would like to learn the tricks of the trade. However, some trainee teachers found the limits of ‘cobbled-together knowledge’ and wanted elements that would allow them to generalize. They would like to be able to obtain knowledge that is sufficiently general to solve problems of the same type.

For Mayen (2008), the knowledge proposed from outside is subject to criteria of practical efficiency. On the other hand, among professionals, some knowledge of action is more or less conscious, some not at all and others strongly so. Thus, scientific knowledge can be forgotten while it is integrated by a pragmatic transformation into the organizers of the action. For Durand (2008), teacher training is often based on intermediate knowledge, between practice and research.

Some researchers share their thoughts on the goals of educational research and how it should be implemented. Heron (1996) believes that it is necessary to improve practice first, and only then to identify the knowledge produced. According to Carr & Kemmis (1986), education should be a practical discipline, which will be evaluated according to its relevance to practice. Van der Maren goes one step further and calls for a discipline that focuses on education and that can distance itself from other reference disciplines, such as psychology and sociology. The creation of a ‘pedagogy’ (Van der Maren, 1995) is not unrelated to the professionalisation of the teaching profession, since practitioners must base themselves on scientifically founded knowledge specific to their field in order to form a recognized group. The author states that the teaching of research methods in faculties and departments of education should primarily serve to produce data that can be used as a basis for pedagogical theories and action in the field (Van der Maren, 1995, p.12).

However, other researchers believe that the work of the researcher, the trainer and the practitioner must be clearly differentiated, even if they are closely linked. The researcher generates scientific knowledge that the trainer popularizes and ‘tinkers with ‘ to make it directly useful in practice (Talbot, 2004). For Barbier (2008), the field of research is that of knowledge, i.e. formalized, cumulative statements, distinct from the subjects who state them, which structure the universe of research, whereas the constructions indissociable from the subjects, which structure the universe of training, are personal knowledge.

What knowledge is then necessary for teaching practice? For Tardif et al. (1991), curricular knowledge, professional training knowledge and knowledge of experience constitute teacher knowledge. Altet (2000, 2008) distinguishes between knowledge to be taught, knowledge for teaching, knowledge about teaching and knowledge of practice. Knowledge to be taught is scientific knowledge, which corresponds to the content to be transmitted and which has undergone didactic transposition in order to be taught to pupils. Knowledge for teaching is knowledge derived from research or formalized practices: it promotes the intelligibility of the teaching–learning process. It can then serve as a guide for action. Knowledge about teaching is the result of the formalisation of practices and is intended to be communicable and usable by peers. It is pragmatic knowledge built from daily experience. The knowledge of practice is the contextualized knowledge of teachers’ experience, which is sometimes partly implicit. For Tardif & Lessard (1999), professional knowledge is mainly derived from daily work experience. Scientific knowledge is then reconfigured to fit into practice while mixing it with other types of knowledge. According to Altet (2008), the appropriation of scientific knowledge or even other knowledge in training differs according to the profile of the trainee, some expecting recipes, a homogeneous and non-contradictory discourse, while others take advantage of possible contradictions between knowledge from experience and scientific knowledge. It should be noted that, for this author, the training approach has a strong influence on the way trainees appropriate and construct knowledge. For example, two conceptions of the accompaniment of trainee teachers by trainers co-exist. In the first case, teaching practice is considered to be the implementation of the preparations developed, whereas in the second option, teaching activity is not reduced to the application of the preparation previously developed. Resulting from trial and error, the practice is then based on problem situations, a reflexive relationship to the practice being thus privileged.

What approach then should be favoured for educational research that can be used by practitioners, while maintaining a form of scientificity that allows the term research to be retained? How can we ensure that educational research can be useful to the practitioner’s situation, which is in part singular?

We will first examine Evidence-Based Practice, which aims to find educational programmes that work and are intended to be generalized in all classes. After mentioning the difficulties of this type of approach, we will look at the approach to singular educational situations, asking how they can be scientifically apprehended in order to serve teachers.

Strengths and Limitations of Evidence-Based Practice

Evidence-Based Practice in Education

Evidence-based practice (EBP) in education derives from evidence-based healthcare and aims, according to its promoters, to overcome dissatisfactions with educational research, in particular the fact that it is non-cumulative and not useful enough for practice (Davies, 1999; Gersten et al., 2009; Hargreaves, 1996, 1997; Pring & Thomas, 2004). According to these authors, the research evidence to be considered depends on the question being asked: if the question is which educational method brings the best results, randomized controlled trials are the best approach (Slavin, 2020). A number of researchers advocates that educational practices and policies should be evidence-based (Saussez & Lessard, 2009). To do this, the randomized controlled trial should be systematically used to build the evidence, and then systematic reviews on a question, specifying which criteria are made explicit for the selected studies, should be developed to answer the question. Subsequently, public policy in education should advocate those educational methods that have been proven to be effective through this type of approach. These practices should be used by professionals because of their proven effectiveness. Slavin (2002) enthusiastically presents the strengths of EBP in education, making a historical diversions, criticizing the habits of educational research that ignore rigorous research. Also, a great deal of EBP research would enable optimal policy-making in education, in all its aspects.

EBP has been implemented in both the US and the UK. In the UK, Tony Blair, then in government, wanted public policy to be based on research evidence. One researcher (Hargreaves, 1996) spearheaded this movement and strongly criticized the educational research of the time as being far too ideological and failing to produce scientific results or help professionals. The idea was to conduct literature reviews of reliable work and to conduct research with randomized controlled trials to identify the most effective teaching practices. In 2011 the Education Endowment Foundation (EEF) was founded, supported by the Department for Education, which aims to improve teaching based on evidence from scientific research. A website, which is a toolkit for professionals, shows which teaching strategies have the strongest evidence of effectiveness.

In the United States, the EBP approach has been developed since the late 1970s, and especially since 1990, with funding from the Federal Department of Education for research with randomized controlled trials, such as Sucess for all (Slavin, 2002). The No Child Left Behind Act in 2001 affirmed the intention to fund predominantly educational research that met the standard of randomized controlled trials. The Department of Education then funded the What Works Clearinghouse to evaluate different educational programmes and provide guidance to policy makers and teachers.

A hierarchy of methods has been developed by proponents of EBP (Bissonnette et al., 2005; Ellis & Fouts, 1993). Level 1 research is descriptive, and is therefore useful for describing a phenomenon. It can introduce educational hypotheses but not validate them. Level 2 research makes it possible to establish cause and effect relationships between several variables. According to these authors, it has more scientific validity than level 1 research. Level 3 research corresponds to large-scale studies and is considered by far the most reliable. Qualitative research is not disregarded but is considered exploratory research to propose hypotheses, which in turn should be tested by randomized controlled trial methods (Slavin, 2002).

EBP Controversies

Researchers have invested in the EBP approach through research projects or synthesis books. Visible Learning (Hattie, 2008) is a synthesis of more than 50,000 studies of over 80 million students. This book is widely cited as the gold standard for EBP. It highlights the factors that promote or hinder students’ progress. The power of the effect is calculated for each factor. Nevertheless, the book has been severely criticized. Simpson (2018, 2019) has criticized Hattie’s work on effect size. Bergeron (2017), an educational statistician, refers to Hattie’s work as pseudoscience. There would be miscalculations in the meta-analyses and inappropriate comparisons. Bergeron’s conclusion is very harsh, indicating that Hattie lacks the competence to conduct valid statistical analyses.

The FollowThrough Project was a programme to evaluate teaching methods that took place in the United States between 1968 and 1977 (Bissonnette et al., 2005; Watkins, 1997) and which highlighted teaching methods that worked better than others. This project has also been subject to several criticisms. The effectiveness of pedagogical approaches varied from one district to another (House, 1978). There were also problems with measurement, including testing for more sophisticated skills than basic ones. According to House, the assessments focused too much on basic skills and not on more advanced ones (reading comprehension, problem solving, etc.). Another article states that the variance in student outcomes was found to be greater within programmes than between Follow Through programmes. No programme was able to produce consistent effects across sites (Berliner, 2002).

As a result, there has been much debate among researchers about this project. According to Paquay, ‘Thomas Cook, known for the methodological rigour of his work with Campbell on quasi-experimental research designs, was clear that, given its methodological shortcomings, the findings of ‘Follow Through’ have no scientific value (Cook, lecture at the Open Colloquium, Paris, Oct. 2006)’ (2008, p.159). If these criticisms propose a controversy about emblematic works of Evidence Based Practices, they do not call into question the EBP project itself, as methodological criticisms are useful to advance science.

Other EBP-type studies, using meta-analyses, show that in reading, writing and mathematics, direct instruction is the most effective teaching approach (Bissonnette et al., 2005). However, this finding is not shared by all researchers. Some claim that the synthesis of effective practices is contradictory, and that there is even some evidence that socioconstructivist approaches are sometimes more effective (Talbot, 2012). Direct instruction approaches would be more effective for highly structured tasks, but not for other types of tasks. Furthermore, intra-teacher variability should be taken into account, as a teachers’ practice is not reduced to a single pedagogical approach, as approaches vary according to the time of the class.

According to Kahn (2012), elementary skills, subtasks and algorithms are measured in research on effective practices (explicit teaching is cited), and not skills, i.e. the ability to choose from among the operations that have been practiced. Research was conducted to compare teachers’ effectiveness in teaching complex, novel tasks, not direct application (Carette, 2008). The aim was to be able to solve new tasks, while mobilizing one’s knowledge. In this study, the pupils were assessed on tasks in which they had to take initiatives and organize themselves. Each class involved in the research had a score on the assessments and the best performing classes were determined. These high-performing classes were defined on the basis of criteria for success in tests involving complex skills. At the same time, teachers were asked to complete a questionnaire on their pedagogical concepts and were interviewed to determine their pedagogical profile. This research showed that the teachers who led the pupils to the best results defended the principle of educability, with the school contributing to the progression of the intelligence of all pupils. They used a problem-based and project-based pedagogy, and encouraged students to compare their views. The teachers whose students performed the worst were those whose main focus was on students’ academic performance. The author concluded that teachers who are effective in teaching complex skills tend to be ‘constructivists’, which contradicts most meta-analyses indicating that the most effective teachers are those using direct instruction methods.

This latest research shows us that when it comes to decisions to adapt evidence-based practices, it is appropriate to be subtle. Thus, rather than asserting that a particular teaching method is good in itself, it would be better to indicate precisely for which types of skills this method has been shown to be beneficial, and at the same time to indicate that for other types of skills, no evidence exists regarding the superiority of this approach over others.

From Research to Practice

Beyond the controversies over the data, some authors criticize the EPB approach on its merits. According to Biesta (2007, 2010), EPB limits the teacher’s professional judgement of what is desirable in education, including the right not to act according to EPB if the teacher believes that this practice is not educationally desirable. In education, ends and means are inseparable in the teacher’s deliberation. Biesta regrets that EBP is based only on a certain type of research, quantitative research and mostly with randomized controlled trials. Furthermore, he criticizes the representation epistemology of EBP (experimentation shows the world as it is) and opposes it to a transaction epistemology (the perceived world is related to our interventions on it). This would imply that EBP gives evidence about what happened in the past, but could not predict what would happen in the future. His second argument is ‘ontological’. According to him, the social, and thus educational, domain is open, undetermined, recursive and semiotic, i.e. not governed by physical causes, but by meanings. However, Biesta finds that the advocates of EBP in education consider the field of education to be closed and deterministic. Finally, he contrasts EBP with value-based education, pointing out that values come first and means second. Other authors make similar criticisms (e.g. Hammersley, 2005).

Some supporters of the Evidence Bases Practice ignore Biesta's arguments when it comes to implementing practices in a school. However, teachers act according to values, implement school programs on a daily basis according to their own values. Also, another problem with the EBP approach concerns the implementation of educational programmes, and the problem of implementing a programme that has worked based on experimental studies.

‘The Research—Development—Dissemination’ reference model is simple: on the basis of new scientific knowledge, a pilot project is developed; when it is perfected, it is implemented in the entire education system. However, research shows that an experimental programme, a new method or a strategy tested and found to be particularly effective in pilot conditions often does not have the desired effects when ‘disseminated’ throughout the education system (Ross et al., 2004). This can be explained by the fact that the conditions for a pilot programme are not necessarily met in other contexts. Also, for decades, experts in programme development have regularly made this observation (Paquay, 2008).

Researchers will then try to determine how to successfully implement a pedagogical method that is considered to have proven itself in the sense of the EBP. The difficulty of implementing a pedagogical strategy on a large scale is a real problem, and in particular the fact that students’ progress is made over a long period of time, after the year of implementation by the researchers (Klinger et al., 2013; Vaughn et al., 2000). According to some authors, a key criterion is implementation fidelity (Benner et al., 2011). In addition, a successful implementation requires approval by the team implementing it. It is therefore necessary to take into account the complexities of practices in an institution. In this context, the following questions need to be answered: Under what conditions and with whom does the programme operate? What is needed to support teachers’ implementation of the programme? What is needed to build the capacity of districts to support teachers in implementing the programme under different ecological and demographic conditions? What is needed to support sustainable implementation of the programme? Then it is worthwhile for researchers and practitioners to reflect together to identify the support structures that would enable them to move students forward and adopt the programme over the long term; this is an adaptation to each school context. Mixed, quantitative and qualitative methods are then advocated to do this. Rather than applying one method to the letter, it is better to be flexible about fidelity, to adapt to the context.The difficulties of trying to implement a practice in schools show that the context is not a minor factor.

The Unsurpassable Context Variable?

The claim that randomized controlled trial methods (and meta-analyses, reviews of issues based on these methods) are the only research methods that can lead to results that can provide guidance for practice and that there is a hierarchy of methods, ranging from meta-analyses of studies with randomized controlled trials as the most valid, rigorous, to qualitative case studies as the least valid, is challenged (Cartwright, 2019). Cartwright criticizes the over-arrogance of EBP proponents, who indicate that only randomized controlled trials are rigorous and can be used as a basis for practice. The author repeats the argument that studies, or meta-analyses, only indicate what has worked in some contexts, but not what will work in a new context: ‘making local predictions requires local thinking’ (2019, p.15). It is an illusion to think that general results, averages, from EBP work can predict what will happen in a particular context. It would be more relevant for some authors (e.g. Cowen et al., 2017) to identify a successful experiment at the local level and then conduct research, not just experimental, to understand what worked, how, why it worked in that context. For Cowen et al. (2017), the general, abstract findings from EBP approaches need to be ‘de-abstracized’ to be usable in practice. The authors interviewed UK teachers about their views on using Education Endowment Foundation (EEF) EBPs, some of whom indicated that the propositions were general, vague, and made little contribution to what they usually did. For example, one said: who is against the use of feedback? As Bru (2019) points out, teacher practice does not point by point obey one and only one category of method (this practice not always being methodical and conforming to the rational choice model) and reveals, on the other hand, that the so-called independent explanatory variable is not so independent.

Despite the criticisms, EBP-type research has its value in providing general trends, provided that care is taken not to extrapolate the findings too far. However, it has the limitation of not responding sufficiently to the highly contextualized situations of the classroom. It is therefore in this consideration of the singular that I will orient my reflection.

The Singular in Educational Research

Taking Context into Account

Thomas (2016) and Wrigley & McCusker (2019) advocate not rejecting EBP, but using several complementary research methods to gain insight into a given topic. Higgins & Katsipataki (2016) argue that syntheses and meta-analyses do not provide a guarantee that an educational approach will work well in a particular context, but rather which approaches are most likely to work well in general, and in relation to a well-specified question. They point to the value of meta-analysis, particularly as a way of allowing small studies that would not, because of their size, have shown statistically significant effects, to show an effect in aggregate. One limitation is that the data from these studies are not equivalent for a single meta-analysis of several studies. Attention must also be paid to the drawer effect, i.e. experiments that were not published because they did not show an effect. Only published studies are taken into account in meta-analyses, which may show a stronger effect than is warranted, because of these unpublished studies.

Some authors call for an improvement science with evidence based on practice in networks of educational communities (Bryk, 2015). Bryk, referring to EBP-type research, states that these studies do not tell what it would take to make the intervention that has shown positive effects in general, work for different subgroups of students and teachers or in different contexts. Also, according to Bryk, educational community networks involving both teachers, researchers, and other professionals could uncover the knowhow needed for school improvement. It is then a question of reflecting on how teachers appropriate good practice. What alterations and modifications do they make to personalize an often ‘ready-made’ system? (Toullec-Théry, 2017). The author recommends starting from the teachers’ concerns and conducting a detailed description of their practice. But how can we go beyond the singularity of each practice, so that it can be useful to other professionals?

Teachers are reflective practitioners. They use what Aristotle called phronesis which involves tacit knowledge and rules of thumb, which vary with context (Fish, 1989). Thomas (2011) speaks of exemplary knowledge that is interpreted in the context of one’s own experience. Some authors distinguish statistical evidence from anthropological evidence (Sensevy et al., 2018). They also distinguish the treatment paradigm from the achievement paradigm. If one respects the dosage of the treatment, it works, regardless of the skill of the person taking it. The achievement paradigm is different. Indeed, in the context of education, the success of the achievement is uncertain, very much linked to the skills of the person implementing the pedagogical approach. Therefore, it is not surprising that pedagogical methods that have been proven in the laboratory sometimes give different results in the field. The success of the approach depends on the implementer. The precise description of the intentions and teaching strategies is therefore fundamental, and can constitute an example for other teachers who wish to appropriate and implement such an approach.

The Logic of the Singular

According to Passeron & Revel (2005), the ‘case’ is problematic and calls for a new framework of reasoning. Case thinking chooses to explore the properties of a singularity accessible to observation and understanding, which constitutes the case. In the twentieth century, this type of approach was often denigrated because homogenisation, the universalist option of science, seemed to be de rigueur. However, if the case must be described, it resists this description, which can never circumscribe it in its entirety and cannot be dissolved into an existing general discourse. How then can we claim to be able to extend what is said about the case under consideration to other cases?

The ‘case’ is an original configuration of an arrangement of facts that breaks into a descriptive, argumentative or prescriptive discourse. It is most often the result of a conflict between general rules and the applications that should be deduced from them, and gives rise to a certain undecidability. Case-based reasoning can lead to the invention of a generalisation path of its own. We can nevertheless draw some common points from these routes. The first is that a distinction must be made between cases and examples, the latter being a simple illustration of a general proposition. The second concerns the historical character that must be attributed to the case: the latter can only be understood by tracing its genesis. Case thinking aims at the intelligibility of a situation, the lived experience around the case being reconfigured by the narrative.

An Articulation Between Methods

From my point of view, there is an interesting dialectic between the general and the singular, with all the intermediaries between the two. The general (and therefore research approaches using randomized controlled trials and meta-analyses) gives fairly reliable global trends. But by wanting to be so general, by taking only what is common to each singular experience, we no longer know precisely what we are talking about. And ‘it works’ globally, but not necessarily in the singular situation in which we find ourselves. The singular, if it is well documented, can make us see, understand in a precise way an intervention, what it brings to the pupils concerned. But if we stick to this singular, nothing proves that what has been found will be valid for other singulars.

I think it is interesting to conduct research at all levels of singularity/generality, from case studies to meta-analysis. It seems to me that proponents of EBP emphasize meta-analyses of randomized controlled trials as the ultimate proof, some not denying the value of more qualitative approaches but rather as exploratory research, as the beginning of a series of research that would be enshrined by meta-analyses that would tell ‘the truth’. In my opinion, this vision makes too great an analogy with the hard sciences because, indeed, if I know the laws of physics (the general) I can predict what will happen to such and such an object at such and such a time (the singular). This is not valid in the human sciences, and particularly in education. We have never seen a law of education that would allow us to predict, with certainty, what we will see in the field in a singular way, for example, a student’s progress in spelling.

A sketch of what could be the complementarity between the different types of research can then be made. A first complementarity is that advocated by researchers in favour of the EBP approach: contextualized, qualitative, ‘level 1’ studies make it possible to explore and form hypotheses, for example that a new intervention might prove interesting. This hypothesis is then tested with quantitative studies, and literature reviews and meta-analyses are carried out on the basis of these quantitative studies to see what the hypothesis developed is like overall.

When we have very general interventions advocated (such as setting up feedback), it would then be a matter of giving qualitative illustrations to show concretely how, in practice, we can implement this general practice. For each teacher, it is then not a question of executing a method to the letter, but of having research results in mind to carry out their practice.

Evidence-Informed Practice

Evidence-informed practice is about the professional’s judgement and imagination. It draws on a variety of sources of information, empirical findings, case studies, clinical narratives and experiences. Professionals are encouraged to learn about and use the results of different types of studies in their work, also taking into account experience and judgment. The Research Learning Communities (RLC) model is one approach to achieving Evidence Informed Practice (Brown, 2015). There are many elements outside of research evidence that are involved in professional practice. EBPs based on randomized controlled trials tend, according to some authors, to focus on ‘what works’ rather than ‘what matters’. According to Brown et al. (2017), Evidence Informed Practice is about examining how teachers can use research alongside other forms of evidence, such as their tacit expertise, to make effective educational decisions in a specific context.

Transduction

To define transduction, we will compare it to deductive reasoning, inductive reasoning, which are the reasoning most used in science, and abduction. Deductive reasoning is reasoning in which the conclusions are more specific, or not more general, than the premises. It is reasoning oriented toward the application of existing knowledge to particular content, toward the production of knowledge from other knowledge. In research, the hypothetical-deductive approach involves hypothesizing, collecting data, and then testing the results obtained to disprove or support the hypotheses. Inductive reasoning is reasoning that produces generalizations: the conclusions are more general than the premises. In research, the inductive approach is oriented towards the construction of knowledge. It starts from facts, observable data, to go towards the explanation of them. Abduction is a type of reasoning that consists in inferring probable causes to an observed fact. We start with one or more observations, and then we look for potential causes, the one that seems the most plausible, even though it is not certain.

Several authors have used the term transduction with rather diverse meanings, but some of them go in the direction of passing from a singular to a singular by analogy. The notion of transduction was created by Stern (1914) and taken up by Piaget (1924): it allows, in the Piagetian sense, the child to pass, by analogy, from one singular to another singular, without passing through a level of generality. For Simondon (2005), transduction is a process of the discovering mind; more generally, it is a physical, biological, mental, or social operation by which an activity propagates from one place to another, basing this propagation on a structuring of the domain from place to place: each constituted region serves the following region as a principle, a model, and allows it to be constituted. Taking up Simondon’s definition, Lourau (1997) states that the transductive approach attempts to overcome the contradiction induced by the inductive/deductive logic that distances, by taking into account all the elements and events that propagate from near to near, in the singularity of a situation. He characterises transduction as a movement, a sequence of potentialisations and actualisations. Also for Lefebvre (1962), transduction is a reasoning that is irreducible to deduction and induction. It employs imaginary variations around themes and requirements defined by reality. It goes beyond empirical observation and makes room for the imaginary while maintaining a critical concern for the real.

Synthesizing the words of the authors quoted, we can say that transduction is a type of logic, reasoning, and approach of the mind, which is not reduced to deduction and induction, and which allows the passage from one singular situation to another, from close to close, while leaving a good place for the imaginary. How, in practice in educational research, can we implement transductive research?

Ricoeur (1983) uses the term mimesis to mean both the imitation of action and the creation of plot. In this sense, it is not only a reproduction but a creation. The interaction between action and narrative is realized in a three-step mimesis. The first time is the one of prefiguration of the practical field, the second the one of configuration of the narrative and the third is the one of refiguration of the practice by the reading. An action, a suffering, will be transformed by the narrative's intrigue and then given back to the reader who will be able to transform his action. Thus, the researcher-trainer can use the narrative to configure his or her path, a narrative that can be meaningful for a reader of the research, for example, another researcher-trainer. However, Christine Delory-Momberger and Hess (2001) warns against the idea that the person who tells his or her story is transparent to the person who receives the story. Understanding another person's story is an issue between the other person and me, and between me and myself. One can only understand the life world of the other person by calling upon one's own biographical memories, from what Christine Delory-Momberger calls the "biothèque" (2003), that is to say the whole of one's biographical experiences. What we feel when we listen to the other is specific to us, is linked to our system of interpretation. The example is here a source of being and potential action: by reading, the reader wants to become close to the person for whom he reads the course. The virtuous actions in the other person incite to want to appropriate them.

Through this paradigm of exemplarity, we see the possibility that the professional path of one teacher, when written down, can be an inspiration for another teacher. Without explicitly using the term transduction, a study by Desgagné (2005) puts the paradigm of exemplarity into practice. In the framework of a research on exemplary accounts of practice, Desgagné, in a first phase, helps experienced teachers to choose and to plot a teaching situation judged as exemplary, particularly representative of a type of problematic situation where the practitioner had to deliberate to overcome the difficulties encountered. These stories can then be inspiring for other teachers, not through generality, but directly through transduction, with the teacher readers having to recontextualize what they have read in relation to their own context.

In order to distinguish, in practice, between hypothetical-deductive research, inductive research, abductive research and transductive research, let us take an example found in Desgagné's research. One of the themes addressed in the research is the ability of a teacher to deal with a student with behavioural difficulties. In a hypothetical-deductive research, a hypothesis would be developed, for example, the implementation of a certain pedagogical approach would make it possible to manage behavioural difficulties, and this hypothesis would be tested to see if it is verified or not. Inductive research might involve interviewing teachers about how they deal with behavioral problems and creating categories that correspond to the issues these teachers address. In abductive research, for example, it would be to find the most likely causes of problematic behaviour in the classroom.

In Desgagné's perspective, which we describe as transductive, it is a matter of reconfiguring a teacher's account of a student's behavioural difficulties so that it is both a research product, since it is the result of a typological analysis of exemplary accounts of practice, and a practical case in the spirit of the case method in teaching. Thus, Desgagné collected professional narratives from several teachers and then produced a typology of these narratives. He then selected two stories for each type and reconfigured them into a version he called prototypes, which served to illustrate the type. The researcher, who is also a trainer, has several concerns in his reconfiguration of the narrative: to remain faithful to the original narrative, but also to ensure that the reconfiguration becomes a source of training for teachers, especially the more novice ones. With regard to the story of the teacher who is confronted with a student's behavioral problems, Desgagné presents the problematic situation that is not initially resolved by the teacher, who becomes exhausted. The researcher then presents the teacher's new innovative attempts, such as giving more responsibility to the student and the resulting beneficial effects. This reconfigured narrative allows any teacher reading it to draw on it if faced with a similar problematic situation. This teacher can thus use a contextualized situation to try to resolve his or her own situation. It is thus a question of transduction, of moving from one singular to another singular.

Conclusion

Despite the controversies, EBP-type approaches provide some trends in what can work in the school setting. It is worthwhile to make teachers aware of this work and to advise them to use EBP interventions while making them understand the limitations of these studies; they are interesting avenues that have worked in some contexts and perhaps less so in others, and it is worthwhile for teachers to be aware of this work in order to implement an evidence-informed practice.

According to Van der Maren (2003), the best theories only apply correctly to situations close to the conditions of experience in which they have been tested. Although knowledge is not completely generalisable, it can nevertheless be transferable. A distinction is made between practical, singular, contextualized knowledge and general knowledge. Practical knowledge is transmitted by observation, imitation, by reading experiences in the field and the views of professionals. The logic of transduction makes it possible to pass from a singular case to another singular case, with a work of recontextualisation. Any reader of this type of research highlighting the singularity of a pedagogical experience carried out by teachers, seeking to understand what their projects, difficulties and thought processes were, will be able to take hold of the research and draw inspiration from it to carry out their own practice.