Conference Agenda

This paper offers a philosophical response to the ongoing debate about how the cohesion of didactic research can be conceptualised in the face of its diversity. As established by previous research, the European field of didactics is characterised by long-standing fragmentation, which has intermittently impeded the exchange of scholarly thought. This diversity stems in part from cultural differences, in part from the plurality of theories and methods used, and in part from the fact that the research field encompasses both general didactics and a multitude of subject-specific counterparts (Gundem, 2011; Hudson & Meyer, 2011; Ligozat & Almqvist, 2018; Werler, 2025).

So far, the challenge posed by fragmentation has been met by efforts to articulate a common ground on which all forms of didactic research can rest. This is most evident in the field’s tendency to define didactics in terms of core elements or ideas. As a prevailing metaphor in contemporary research (cf. Sfard, 1998), the ‘core of didactics’ has featured in several key publications and centred on, for instance, the expanded didactic triangle, the didactical questions, or the contingent relationship between educational content and meaning (e.g. Hudson & Meyer, 2011; Hopmann, 2007; Klafki, 2000). While defining the field’s coherence is necessary, we argue that the use of ‘core’ as a prevailing metaphor implies an essentialist notion of didactics that risks limiting the originality of future research and the nature of the conversation among researchers from different scholarly traditions. As such, this paper aims to challenge the metaphor of a didactic core by providing an alternative way of conceptualising the cohesion of didactic research. In other words, we aim to suggest and exemplify a new metaphor by which the field’s commonality can be understood in a non-essentialist manner.

The suggested metaphor draws on Ludwig Wittgenstein’s (1953/2009) philosophy of language-games, which emphasises that they are interconnected through a series of family resemblances rather than a single common feature. Crucial to Wittgenstein's definition of language-games is the idea that words do not have inherent meaning but acquire it through the ways in which they are employed for a variety of purposes. The meaning attributed to words is therefore changeable, depending on the usage and human activity at hand. However, this usage is not completely arbitrary, as it is impossible for a person to say or do whatever they wish and still present themselves as intelligible within a given language-game. Instead, these games are guided by patterns, or rules of language, established through previous human interaction, much like the rules for chess, backgammon, and go guide what moves are possible on their respective boards (Wittgenstein, 1953/2009; see also Bowell, 2017). From this position, Wittgenstein (1953/2009) argues that meaning is established as a consequence of how words are functionally used within rule-governed language-games. In addition, he stresses that different language-games are not completely separated from each other, nor are they joined by a single common feature. Rather, Wittgenstein (1953/2009) suggests that their affinities are to be understood as a multitude of interconnected family resemblances, not unlike how eye colour, gait, temperament and so on criss-cross in various ways between members of an extended family. Thus, by using Wittgenstein’s concepts of language-games and family resemblances, this paper offers a conceptualisation of didactic research as consisting of many different discourses that are related to each other through a network of similarities.

Formative interviews with teachers are a critical lever for improving teaching quality, alongside observation frameworks and the identification of high-quality teaching features. However, in many observation-based professional development designs, these interviews rely on ancillary tools and/or narrow the discussion to pre-prioritized rubric indicators. This paper examines how the JAD-MTQ model can structure a professional development design aimed at strengthening content coherence across a teaching unit. Using the full set of JAD-MTQ rubrics across diagnosis, formative interviewing, and impact monitoring, we introduce the PACK–UNPACK–REPACK (P-U-R) protocol to structure a formative input. We argue that this protocol supports the formative process by providing a common structure of rubrics that can capture both content-specific aspects of teaching and how these aspects are organized hierarchically across different timescales within a teaching unit. Ultimately, we demonstrate that the JAD-MTQ rubrics (organized into dimensions and criteria) can fulfil various functions at different stages of the formative process.

In French-speaking research, the Joint Action framework in Didactics (JAD; Sensevy, 2014; Ligozat et al., 2018) typically investigates what knowledge contents are developed in learning situations and how. Studies carried out with this framework typically investigate how knowledge contents develop in the teacher and students’ classroom interactions. Over the years, the JAD has proved its capacity to describe, understand, and explain classroom practices in various subjects (mathematics, sciences, physical education, French language, etc.; e.g., Amade-Escot & Venturini, 2015; Ligozat et al., 2018; Sensevy, 2011; 2014).

The JAD-MTQ (Model for analyzing Teaching Quality based on the Joint Action frameworks in Didactics; Ligozat & Buyck, 2025) structures classroom observation through three dimensions—(1) selection of knowledge content and tasks, (2) structuring of learning situations, and (3) organization of teacher–student interactions—each further specified by criteria. These dimensions address content development across complementary levels of analysis within a teaching unit (Tiberghien & Sensevy, 2012) and are designed to be applicable across school subjects. When used for formative purposes, the JAD-MTQ model contributes to a formative input structured into a three steps protocol; 1) PACK (Seq.A), the teacher-trainer uses the JAD‑MTQ to diagnose strengths and weaknesses across the three dimensions. 2) UNPACK, a feedback meeting is meant to present the diagnosis dimension by dimension to the teacher. 3) REPACK (Seq.B), a planning meeting supports the teacher in redesigning a new unit on the same content by integrating key issues identified.

The P-U-R protocol itself contributes to the broader formative process experienced by the teacher (trainee). This formative process is generated by the overall TPD-design. This paper unfolds the conceptual and methodological aspects of the PUR protocol, in which epistemological aspects of the curriculum content and the specificity of the subject matter structure are core concerns during the whole process.

Subject-matter didactics is an academic field which is more or less compartmentalized by the subject area of the relevant discipline. Up to now, the discussion about research designs in subject‐matter didactics has mainly taken place within the disciplines of this academic field and the respective national contexts (Bayrhuber et al., 2012; Rothgangel et al., 2020). To stimulate inter-disciplinary and (inter-)national discussion, recently a generic model of research designs has been developed by two cross-disciplinary Delphi Studies. First, a Delphi Study has been administered within the German speaking discourse of subject-matter didactics, in which 305 colleagues from 21 subject-matter didactical disciplines have responded to an online survey with predominantly closed-ended questions and further 68 colleagues answered to a follow-up online survey with predominantly open-ended questions. The result of this study has been used for an international Delphi Study in 30 countries from four continents (Riegel & Rothgangel, 2022, 2024). Both studies resulted in a generic model that is based on three dimensions which are typical for scientific research: theories, subject areas, and methodologies. Its particular subject-didactical character is coined by the categories of these dimensions: These categories give the model the character that is typical of subject-didactic research. This generic model, however, still is deductively derived, while an inductive corroboration of this generic model is still pending.

As a first step to take up this desideratum, this project sets out to validate the generic model by comparing it with the latent themes within a sample of doctoral theses within the field of subject-matter didactics. This sample is a perfect test case because in such theses the early career researchers testify that they are competent to do proper research in this academic field. Additionally, the distribution of such a sample regarding theories, subject matter and methodologies may not be as influenced by publication bias as a sample consisting of research papers. The sample’s heterogeneity, volume and length of its single documents prevents traditional approaches for validating the model like qualitative content analysis. Instead, the application of computer-linguistic machine learning approaches is necessary to extract, clean and analyze the texts. The latter was realized with a latent Dirichlet topic modelling approach (LDA), a proven and widely used method to identify latent themes within literature collections (Blei et al., 2003). To ensure parallelism of the generic model and the LDA, instead of a single model, a total of three LDA models were conducted one for each dimension of theories, subject matter and methodologies each based on a set of fixed terms related to each method (i.e. correlation or interview linked to methods; religion, chemistry or language-learning to subject matter etc.).