A reflective dialogue between the lesson study experts on their journey towards becoming koshi: learnings, experiences and the perceived roles Open Access

To contextualise my journey of becoming a knowledgeable other or a koshi in Japanese, I share the background of lesson study (LS) in South Africa (SA). LS was introduced to SA through the intervention of the Japan International Cooperation Agency (JICA) around the year 2010, focusing on one province (Ono and Ferreira, 2010). The purpose was to improve teachers’ skills in the teaching of mathematics and science in secondary schools. Due to systemic challenges, the effective implementation of LS was acutely affected. A few years later, LS was coordinated from the national office, i.e. Department of Basic Education, subsequently all the nine provinces in SA were exposed to this teacher development practice. Like many countries that adopted and modified the Japanese LS model to suit their needs, SA was not an exception. The current cyclical model of LS broadly used in SA comprises five stages: diagnostic assessment/analysis (Stage 1), collaborative lesson planning (Stage 2), lesson presentation/observation (Stage 3), post-lesson reflection (Stage 4), and lesson improvement (Stage 5) (Sekao and Engelbrecht, 2022). The unique feature of the LS cycle in SA is Stage 1, which I have given special focus to in Figure 1. Stage 1– Diagnostic assessment/analysis – entails explicit articulation of the process of identifying the goal of the lesson. The implementation of Stage 1 follows a two-pronged process involving teachers’ identification of the topic, which (1) they find difficult to teach or (2) poses a difficulty for learners to learn. The latter is a dominant approach to identifying the goal of the lesson. However, identifying a topic that poses difficulty for learners can be done either by designing and administering a diagnostic test and then analysing learners’ responses to gain insights into their difficulties or by diagnostically analysing learners’ responses stemming from summative assessment. In other words, although summative assessment is used for promotion purposes, i.e. from one grade to the next, the same learners’ responses can be analysed diagnostically to identify their (the learners’) misconceptions.

Generally, the role of koshi in SA is predominantly performed by the subject advisors because they work closely with the teachers within the education district. Subject advisors are education district officials who are responsible for teacher development in specific subjects. However, at the inception of LS within the education district where subject advisors are trained, the trainer is often perceived as the koshi by virtue of being knowledgeable about LS. The training referred to is the overall training on LS, which includes the roles of the koshi. In other words, for now, in SA, there is no exclusive/dedicated training system for becoming a koshi. Essentially, the primary knowledgeable other within the LS context in SA is the subject advisor; however, there are instances where a university lecturer and/or subject expert from the national office (i.e. Department of Basic Education) is invited to fulfil the role of the knowledgeable other.

Although LS has gained traction in some education districts, and the overall process thereof is understood by teachers, critical aspects that ensure deep understanding and effective implementation remain elusive. These critical aspects include quality of lesson planning, observations and reflections, as well as the explicit role of the knowledgeable other. The latter is the main focus of this study. In this qualitative self-study paper, I share my experiences on the journey towards becoming a knowledgeable other, henceforth a koshi, including personal learnings, challenges and perceived roles thereof. To this end, I explored the following research question through self-study: How did my journey towards becoming a koshi within the LS setting in SA enhance my view about the role of koshi? Since LS is a developing phenomenon in SA, the concept of koshi and the role thereof are also not deeply understood by many; therefore, self-study. Due to the insufficient proficient koshis in SA, and the numerous instances where I was invited to become a koshi, I deemed learning from experience through self-study (Kitchen and Butler, 2024) to be a viable option to answer the stated research question.

LS is a quintessential job-embedded and teacher-led professional development practice which originated in the 19th century in Japan (Toyoda, 2011; Ebaeguin and Stephens, 2014). The main attribute of LS is teacher collaboration – a team of teachers work together to plan, teach and reflect on their lessons to improve their teaching skills (Gutierez, 2015). A koshi is an important outside stakeholder in LS, without whom the improvement in teaching practice may be restricted. The fact that a koshi gives final comments after the post-lesson reflection session demonstrates a vital role he/she plays in ensuring teachers’ improvement in content, pedagogy, as well as LS as a whole (Groves and Doig, 2016; Takahashi and McDougal, 2016). Notwithstanding the importance of the koshi in the context of LS, it is necessary to share some insights into what koshi is, firstly, outside the LS setting and then secondly, within the LS setting. The concept of more knowledgeable other (MKO) is traced to the Vygotskyan era, used to unpack Vygotsky’s view of Zone of Proximal Development (ZPD) (Vygotsky, 1978). In fact, the term used by Vygotsky within the context of ZPD was “adult” to denote what the child can do with the guidance of an adult who was perceived to be more knowledgeable, which, within the education context, is a teacher. Of course, for Vygotsky, the emphasis on knowledge hierarchy between the child (less knowledgeable) and the adult (more knowledgeable) was necessary and justifiable to enable the child to navigate the ZPD (Jarrett, 2022; Abtahi et al., 2017; Shabani, 2016). Drawing from Vygotsky’s MKO, Jarrett (2022) dubbed MKO a catalyst to portray his/her (MKO’s) role to stimulate the less knowledgeable (child) to learn to a greater degree. However, within the confines of LS, the term knowledgeable other (KO) is broadly used (Watanabe, 2005) instead of MKO. My view is that using the modest term KO within LS portrays and affirms the entrenched collaboration as a “default setting” of LS where all teachers are considered to be knowledgeable and their collective contributions towards improving their practice are valued. The term knowledgeable other within the LS context, therefore, could be viewed as acknowledging that people participating in LS are knowledgeable; however, other knowledgeable persons were not part of the LS team. In addition, the term KO, instead of MKO, desists from exalting one person as being more knowledgeable, thereby defeating the purpose and undermining the entrenched collaborative attribute of LS.

Within the LS context, a koshi is highly revered (Seleznyov, 2018). However, a dearth of sufficient research on koshi, especially on the “nature and structure of the final comments” (Seino and Foster, 2021, p. 509), which often results in an incorrect understanding of the attributes and importance thereof within the LS context, could be attributed to the challenges occasioned by a borrowed cultural practice such as LS. In fact, Skott (2024) unequivocally and succinctly remarked that although LS was imported into other countries, one area that was not successfully imported is koshi. As Fujii (2014) puts it, the lack of detail could be attributed to the fact that LS is an ingrained Japanese classroom culture – an organic practice that every Japanese teacher is familiar with. Metaphorically speaking, for Japanese teachers, LS, including its instantiations such as koshi, “is like air, felt everywhere because it is implemented in everyday school activities, and so natural that it can be difficult to identify its critical and important features.” (Fujii, 2014, p. 66). Certainly, Japanese teachers may not need the roles of the koshi documented in detail because of, as Ebaeguin and Stephens (2014, p. 2) assert, Japan’s high-context culture where “meaning and information is drawn from the context or internalised by people instead of making them explicit.” For Japanese teachers, acquiring the skills of becoming a koshi, therefore, could be through participating actively in the LS activities (Takahashi, 2014) – an undertaking that they may not find difficult due to the entrenched culture of LS. In fact, Watanabe (2018, p. 10) makes an intriguing observation that in the Japanese context, “teaching is [perceived as] research and therefore teachers are researchers”. Possessing research skills is an important attribute of a koshi, therefore, for Japanese teachers, becoming a koshi may not be a difficult undertaking. However, Seino and Foster (2021) make a compelling argument that for teachers outside Japan, it may be a daunting task to understand, develop and actualise the role of a koshi. SA is experiencing a similar difficulty of a lack of proficient knowledgeable others as per the LS prescripts. Often, subject advisors share the final comments during post-lesson reflection; however, their role as subject advisors often infringes on their role as a knowledgeable other. For instance, as subject advisors, they are a leading voice or voice of authority in the subject they are responsible for – after all, they are the district officials; however, as knowledgeable others, they have to be modest and project the image of being a lifelong learner (Takahashi and McDougal, 2016). The tension between the subject advisor and koshi often muddies the deep understanding of koshi and koshi’s role within the LS context in SA; consequently, koshi is often viewed through the lens of a subject advisor. Essentially, understanding the role of koshi and discharging such roles/functions effectively remain a major challenge in the implementation of LS in SA.

Notwithstanding the insufficient research on koshi, the available studies provide valuable insights into the role of the koshi within the LS context. Drawing from the work of Takahashi (2014), Watanabe (2005), and Fernandez and Yoshida (2012), the following roles of a koshi can be distilled: (1) bringing in new knowledge from research; (2) connecting theory and practice; and (3) creating an opportunity for others to learn how to reflect on teaching and learning. Clearly, the first two roles emphasise the depth of knowledge and experience the koshi ought to possess in relation to content and pedagogy – not only about theory, but extended to practice, too. Supposedly, this justifies the view that koshi could be a university teacher, research expert, retired school teacher, district official or subject specialist (Ebaeguin and Stephens, 2014; Seleznyov, 2018; Seino and Foster, 2021). In addition to the knowledge domain that a koshi should have as articulated in roles (1) and (2), Takahashi and McDougal (2016) identified certain attributes that a koshi should possess, such as: reading and understanding the audience to pitch the comments appropriately; noticing critical issues emerging from the lesson that others may not have seen; and being modest and adopting a demeanour of or project oneself as being a (lifelong) learner thereby valuing teachers and their efforts. I contend that these attributes align with role (3) of a koshi, i.e. creating an opportunity for others to learn how to reflect on teaching and learning. Drawing from the conceptualisation of the roles of the koshi articulated in this section, as well as the proposed attributes that shape koshi’s demeanour, I perceive koshi as an integral role-player in LS. In fact, for Yoshida (2012, p. 148), “… if somebody who is knowledgeable in content and pedagogy can be involved in the lesson study, the quality of lesson study, as well as the quality of the kyozaikenkyu, increases.” Conversely, the absence of koshi within the LS process tends to significantly reduce its effectiveness (Takahashi, 2014). Skott (2024) justifiably takes a radical stance by arguing that the absence of koshi in LS disqualifies the practice from being branded as LS.

Considering the sparsely documented roles of koshi, it is evident that often this tends to dilute the attributes of LS. Again, due to LS being an imported cultural practice and the insufficient detailed guidelines on the role of koshi, alternative conceptions and omissions are likely to characterise the adaptation and implementation of LS in other countries, especially as it pertains to the role of koshi. I have intentionally used the term “alternative conceptions” rather than the conventional term “misconceptions” because I viewed koshi through the lens of the LS implementer. After all, as Fujii (2014, p. 453) contends, “[alternative conception] is a reasonable and viable conception based on [teachers’] experiences in different contexts or in their daily life activities”. In other words, the so-called teachers’ misconceptions about koshi are, in fact, for them (teachers) a reasonable understanding of what koshi is or is about. For instance, Gu and Gu (2016) discovered one such alternative conception of the role of koshi where, instead of engaging in dialogical comments, the koshi engaged in monological comments. In other words, the comments revolved around koshi’s knowledge about the topic taught, “rather than what students have learned in class and what teachers were concerned about in their teaching” (p. 451). According to Skott (2024), monological comments are often occasioned by socio-cultural and power-related practices where respect for authority is emphasised and telling teachers is a norm.

In addition, koshi’s comments are often gravitated towards the process of LS (Skott, 2024) rather than towards substantive issues such as learners’ thinking and teacher learning. Further, there are often inconsistencies about when koshi should be involved in the LS process: during post-lesson reflection to share final comments (Wake et al., 2013; Wake and Seleznyov, 2020) or during both planning and post-lesson reflection stages (Watanabe, 2005; Takahashi and McDougal, 2016). In fact, it would appear that, in most cases, a koshi is associated with post-lesson reflection to provide final comments.

The systematic literature review conducted by Seleznyov (2018) revealed that a significant number of studies reported that a koshi was not involved in the LS process. Takahashi and McDougal (2016) revealed numerous similar instances in the USA where the koshi was omitted from the LS processes. My view is that these alternative conceptions of a koshi and the role he/she ought to fulfil within the LS context are inevitable for as long as the journey towards becoming a koshi is based on experiential learning.

This paper emanates from my experiential learning on how I learnt to become a koshi as I got exposed to and gained interest in LS. My journey towards becoming a koshi can best be understood through the lens of Experiential Learning Theory (Kolb, 2015), which other scholars refer to as Kolb’s reflective cycle (Fergusson et al., 2019; Turesky and Wood, 2010). The notion of experiential learning is traced back to the Deweyan era through the seminal work titled Experience and Education (Dewey, 1938), wherein a high premium is placed on the value of experience to derive genuine learning. However, Kolb’s experiential learning model, which incorporates four modes of experience to foster learning (Egan et al., 2023; Kolb, 2015), provides an explicit account of my journey towards becoming a koshi in LS. The four modes are cyclical and are often represented on orthogonal axes. The first mode – concrete experience – which involves immersing oneself in a new learning activity, highlights a common parlance that experience is the best teacher. In relation to the current study, concrete experience included contending with and grasping the information about the role of koshi within the LS setting as the process of learning how to become a koshi unfolded, i.e. learning by doing. The second mode – reflective observation – involves processing, forming and transforming the learning experience. Immersing oneself in an activity such as becoming a koshi, is characterised by internal reflections during and after one’s participation in concrete experience. Reflective observation raises questions such as how my prior theoretical knowledge about the role of koshi is linked to my practical experience as koshi? What should one modify to enhance one’s koshi skills? The third mode – abstract conceptualisation – highlights the individual’s analysis and interpretation of and drawing conclusions about the learning experience, thereby making connections with the theoretical knowledge. Essentially, abstract conceptualization highlights “integrating ideas into a working theory.” (Egan et al., 2023, p. 26). The fourth mode – active experimentation – involves making decisions and using theories to improve skills (Long and Gummelt, 2020).

For Kolb (2015), “Learning is the process whereby knowledge is created through the transformation of experience.” (p. 49). From this statement, one can discern that learning is not a rigid event but a dynamic process where knowledge morphs and gets enriched as one is exposed to more experiences. Essentially, the modes of Concrete Experience and Abstract Conceptualization represent two ends of a continuum – connected dialectically to foster perceiving and grasping of experience, while the modes of Reflective Observation and Active Experimentation advance the transformation of experience (Carvajal et al., 2021; Long and Gummelt, 2020). To this effect, learning is characterised by grasping and transforming experience.

I am a teacher educator in a public university in SA, and I was introduced to LS over a decade ago, under the auspices of the Japan International Cooperation Agency. Subsequently, I introduced LS as part of, inter alia, the methodology of mathematics modules in the university where I am based. In addition, I train in-service teachers on LS as a community engagement programme. Yolanda (pseudonym) voluntarily agreed to participate in the reflective dialogue on our journey towards becoming koshi as a critical friend, which, within the context of self-study research, “is a valuable authority of data and analysis” (Alan et al., 2021, p. 321). Yolanda is a mathematics subject specialist based in the Department of Education in the public sector and has experience in using LS to support teachers. I introduced LS to the mathematics subject advisors (including Yolanda) in the year 2017 in one of the provinces in SA, at the time when Yolanda was a mathematics subject advisor. Yolanda was one of the few mathematics subject advisors who keenly embraced LS and worked with teachers to implement it. Through my support, Yolanda became the first subject advisor to host the first LS seminar at the district level, where teachers demonstrated their teaching skills, and I was invited to be a koshi. Both Yolanda and I learnt in practice to become koshis by virtue of having introduced LS to teachers and subject advisors, respectively, and not because we were trained to become koshis. In fact, contrary to on-the-job training, our experiences of becoming koshis were acquired through on-the-job learning. Through reflective dialogue between Yolanda – my critical friend (Kitchen and Butler, 2024; Bullock and Sator, 2018) – and me, I share the learnings, challenges and the perceived roles of koshi emanating from our journey towards becoming koshi in the context of LS.

The genesis and promulgation of action research, which self-study research is part of, is traced back to the work of Kurt Lewin and his colleagues in the 1940s (Adelman, 1993). As he explored and chronicled the history of action research, Jacobs (2018) noted three forms of action research and their related world views or paradigmatic traditions: technical action research, practical action research and emancipatory action research, which are situated within positivist, interpretivist and critical science worldviews, respectively. Within the practical action research with interpretive foundations, Jacobs (2018) locates self-study research, which I adopted in the current paper. The relevance of self-study for this paper is that I studied my own practice by engaging in reflective dialogue with a critical friend (Kitchen and Butler, 2024; Bullock and Sator, 2018; Ritter, 2017; Vanassche and Kelchtermans, 2015) regarding a phenomenon of koshi. Self-study, rooted in reflective dialogues, is credited for enabling practitioners, teachers in the context of the current paper, “to develop their own practices and analyze their experiences” (Alan et al., 2021, p. 319). A critical friend is a fundamental element of self-study research because he/she can assist in revealing the implicit information that the researcher alone may not be aware of. To this effect, by virtue of having learnt in practice to become koshi, Yolanda was instrumental in providing a different perspective during the generation and interpretation of the data, thereby enhancing the objectivity of the entire study.

As can be observed from the heading of the current section, I opted for the term data generation instead of data collection because data were produced from our reflective dialogue and not “collected” from the participants other than the researcher, as is the norm. We held five meetings, as presented in Table 1.

Before the actual reflective dialogue between Yolanda and me to generate data, I scheduled the first meeting during which I shared the overall idea about the proposed study and its potential contributions to the often-elusive concept of koshi. During the second meeting, we discussed how the reflective dialogue would unfold. To make the reflective dialogue productive and meaningful, we agreed to allocate two weeks towards individual recollection of our memories about the journey we traversed in becoming koshi. The individual recollection of memories included eventful encounters that translated into learnings and events that constitute challenges or pitfalls. Essentially, the recollection of memories was framed on: our pre-conceptions of koshi, formative experience of koshi, and abstracting the concept of koshi. The next session (third meeting) was the actual reflective dialogue, which spanned about two hours. The dialogue covered our individual initial understanding of koshi before we immersed ourselves in the LS process and accidently assumed the role of koshi (pre-conception of koshi). Subsequently, we reflected on our practical/concrete experiences and the observed challenges of becoming koshi (formative experiences). Lastly, based on our experiences, we shared some insights into the proposed attributes of koshi (abstracting the concept of koshi). I audio-recorded our reflective dialogue and later transcribed it into textual data for analysis. In the fourth meeting, we dedicated our time towards transcribing the audio-recorded data using Sonix software for data transcription, followed by a data cleaning process – i.e. identifying and correcting incorrect transcriptions.

Given the predetermined aspects framing our reflective dialogue, analysis of data was mainly deductive; however, this was coupled with inductive analysis because there were individual views emerging from the reflective dialogue, especially during the abstraction of the concept of koshi. Drawing from thematic analysis advocated by Terry et al. (2017), I engaged with the transcribed data iteratively and then identified key issues that characterised the three pre-determined themes of pre-conception of koshi, formative experiences and abstracting the concept of koshi. The findings were documented in a report, after which the fifth meeting was scheduled for the critical friend, Yolanda, to authenticate the content thereof.

Our data-generation meeting took the form of a free dialogue or conversation. Although Yolanda understood the purpose of our current meeting, at the beginning of our dialogue, I briefly reiterated the purpose thereof by emphasising that “this is not me interviewing you, but it is for both of us sharing our insights and our experiences in terms of becoming the knowledgeable others.” In addition, I reminded Yolanda that, for ethical reasons, I will conceal her name by calling her Yolanda; however, “I don’t mind if you refer to me by my real name, because I’m going to be the author of the article in any way”. I then reminded Yolanda about the three key areas that would frame our dialogue.

The initial stage of our dialogue was to share our experiences on how we became aware of the concept of a knowledgeable other. It should be noted that the term koshi was not a familiar term to us when we were introduced to LS; instead, the term that we first encountered was knowledgeable other. I have, however, used the two terms interchangeably in the findings. I was introduced to the concept of the knowledgeable other as it relates to LS in the year 2012 during training in Japan. However, the role of the knowledgeable other was not explained in detail; therefore, my limited knowledge confined it (the role) to the knowledge of the process of LS. I shared the following information during our dialogic interchange regarding how I was first introduced to the concept of koshi within the context of LS:

Essentially, although my participation in the LS training in Japan framed my initial knowledge and understanding of koshi, I gained deep insights into koshi and koshi’s roles from engaging with literature. Drawing from the work of researchers such as Takahashi (2014) and Fernandez and Yoshida (2012) became an inflection point in terms of my deep understanding of koshi and koshi’s roles as they pertain to LS. However, Yolanda’s case differs from mine in the sense that practical experience shaped her knowledge and understanding of koshi.

Yolanda’s initial encounter with the concept of koshi was in 2017 when the mathematics subject advisors were trained on LS. Yolanda’s province was the first of the nine provinces in SA to be trained on LS. It should be noted that for both of us, our first encounter with the concept of koshi was when it was merely flagged during the LS training. Therefore, it was not dealt with in detail in terms of the roles and processes. Considering that I received my training on LS in Japan – the home of LS – one would have expected to gain a deep and comprehensive perspective and understanding of the concept of koshi.

As indicated in how we encountered the concept of koshi, it was inevitable that we created a mental picture or preconceptions about koshi before undertaking the accidental journey – accidental because we were perceived as such by virtue of having introduced LS in our respective contexts. The fact of the matter is that no one prepared us on what it really entails to become a koshi.

One of the preconceptions we had about koshi was that he/she should only know LS very well. For instance, during the dialogue, I stated that “I thought being a [koshi] would mean a [koshi] in terms of only understanding the process of lesson study. Being an expert in the process of lesson study”. This perspective of being a koshi characterised my initial role and experience when requested to be a knowledgeable other. In addition, Yolanda had this to say:

In addition, our preconception of koshi was that he/she participates only in the lesson observation and in the post-lesson reflections. Clearly, we had narrow preconceptions about koshi before immersing ourselves into the practice and assuming the role of koshi by default.

As part of our dialogue, we reflected on how our participation in the LS processes, albeit unprepared, changed our worldview about koshi and his/her roles within the LS context. In other words, when we started our journey towards becoming koshis, we had certain preconceptions which were shared in the previous section; however, having gone through the experiential learning process, there were learnings acquired which resulted in our conceptual changes about koshi. Some of the impactful learnings pertaining to koshi are captured in the following dialogic exchanges between Yolanda and me.

My comment highlights two issues: the misconception about koshi’s focus of reflection and the later realisation of additional areas of koshi’s focus of reflection. For instance, the statement “Until I realised …” was based on the experience from participating as koshi, where the focus was mainly on the LS process. In addition, the other turning point in my journey towards becoming koshi is that koshi must possess a unique professional noticing skill – be able to meticulously notice the finer details of the goings-on during teaching, and then during the reflection session be able to interpret and respond to what was noticed (Jacobs et al., 2010). In fact, I realised that assuming the role of koshi (and being introduced as such to the LS team members) tends to justifiably raise expectations of the LS team members about the expert knowledge to be shared by koshi. This, therefore, was a moment of awakening to the realities of what it takes to be a koshi.

Yolanda had her inflection point too, as reflected in the following utterances:

Another intriguing dialogue ensued where Yolanda shared a new perspective about sharing the lesson with prospective koshi well in advance to enable him/her to plan thoroughly and adequately.

Essentially, in this dialogue, Yolanda acknowledges that sharing the lesson plan with the koshi within a short time before the lesson is taught, as she experienced it before, may compromise the quality of the remarks that the koshi shares during reflections. At the heart of this dialogue, where a paradigm shift is evident, is the utterance “which we never did at that time” when Yolanda was referring to the practice of sharing the lesson with koshi well in advance. In response to and in concurrence with Yolanda’s view based on experience, I added that once the lesson plan is shared with the koshi, it should not be changed just before it is taught because the koshi’s hard work of preparing for the final comments may be in vain or pointless.

The other issue that Yolanda mentioned, and that was thought-provoking and fascinating, was the importance of the knowledge and understanding of the curriculum policy by koshi. This was triggered by my comment that “koshi ought to be an expert in two areas – an expert in LS and a deep understanding of the process thereof; and a deep understanding of the subject [content and pedagogy]”. In fact, at the beginning of my journey to becoming a knowledgeable other, my focus was mainly gravitated towards the former – LS process. Yolanda asserted that:

Yolanda’s view about the need for koshi’s understanding of the curriculum policy was based on the observation that teachers often neglected the key curriculum prescripts during post-lesson reflection; therefore, it was her role as koshi to alert teachers to this anomaly. For instance, the mathematics curriculum in SA is organised according to conceptual progression across the grades, i.e. the same mathematics concept progresses across the grades according to the degree of complexity. For Yolanda, making final comments without considering the broad curriculum prescripts and narrowly focusing on the topic taught does not add value to teacher learning. The koshi, therefore, ought to be mindful of these similarities and variations and make them visible to teachers.

Towards the end of our dialogue and guided by our practical experiences as the knowledgeable others, we shared our views on koshi’s demeanour or traits. The key issues that characterise the possible traits that koshi ought to have included a passion for LS and a willingness to share knowledge. To this effect, I had the following to say:

This exchange points to the view that knowing the LS and the subject is not sufficient to become a koshi; instead, what appeals to the audience is the innate attributes. The issues of koshi’s innate attributes, such as demeanour, modesty and projecting an outlook of lifelong learner are advocated by Takahashi and McDougal (2016). Indeed, as we journeyed the path of practising as koshis, we were driven by passion and willingness to share knowledge.

We reflected on our overall impressions about our journey towards becoming a koshi. As I reflected on this journey, I compared it to “travelling without a GPS [global positioning system] navigator, literally walking in the dark and empowering myself in the process by reading a lot”. Yolanda concurred that, difficult as the journey was, she was mainly driven by passion towards LS and the mathematics subject. We concluded that relying exclusively on learning from experience to become koshi tends to be a cumbersome undertaking and takes too long to acquire the skill; rather, a blend of training courses and experience could yield a better outcome. Reverting to the metaphor of a GPS navigator, a training course could be a navigating tool – a lodestar that guides experiential learning towards becoming a koshi.

Undertaking and reflecting on the journey towards becoming a koshi was an insightful experience, often punctuated by intriguing moments and episodes. Immersing myself in the activity of becoming a koshi has enabled me to grasp, analyse and connect my practical learning experience with theoretical knowledge. This was made possible by dialectically related modes of concrete experience and abstract conceptualisation learning continuum (Carvajal et al., 2021; Long and Gummelt, 2020; Kolb, 2015). Through concrete experimentation, I grappled with the practical realities of what it takes to be a koshi, while abstract conceptualisation helped me connect my practical experience with theory. For instance, my initial participation as a koshi, as per the prescripts of concrete experience, was confined to the process of LS – i.e. not critically focusing on content and pedagogical issues pertaining to the subject, in this case, mathematics teaching and learning. However, when processing my experience and gaining more insights from theory pertaining to the minutiae of being a koshi as articulated by, inter alia, Takahashi (2014), my superficial understanding of koshi began to change for the better. My preconceptions were refuted, my mindset changed, and, most importantly, new ideas were generated pertaining to the koshi. For instance, a shift from koshi as only an LS expert to koshi as an expert of both LS and the subject taught. This was made possible by the related modes of reflective observation and active experimentation learning continuum to advance the transformation of my experience (Carvajal et al., 2021; Long and Gummelt, 2020).

LS and its instantiations, such as a koshi, are rooted within the Japanese educational culture (Fujii, 2014; Ebaeguin and Stephens, 2014); therefore, it is inevitable that implanting them into other cultures will be punctuated by challenges (Seino and Foster, 2021). For instance, it is unsurprising that for Japanese teachers, learning to become a koshi is achieved effortlessly by participating in the LS activities and observing experienced koshis, while it is a daunting undertaking for teachers from other countries. The major drawback of my journey towards becoming a koshi is that I developed the skills by learning from practice devoid of koshi, unlike with Japanese teachers who learn to become koshis by observing and being guided by other competent koshis in practice (Takahashi, 2014; Watanabe, 2018). Indeed, my journey was a daunting undertaking, as echoed by Seino and Foster (2021). Given the fact that LS is still developing in SA and the subsequent lack of sufficient competent koshis, observing and learning from other competent koshis is not a readily available option for now. In other words, on-the-job learning to become an effective koshi may not be an immediate viable option, not only in SA but in other countries too, where LS is advancing. The lack of opportunities to observe a competent koshi in action often deprives aspirant and emerging koshis of a full grasp of the role of koshi. In fact, although the importance of koshi is acknowledged and appreciated in the countries where LS is introduced, the underlying attributes of koshi’s role are often not fully understood and made visible (Seino and Foster, 2021). A case in point is our preconception that koshi’s requisite knowledge is confined to (1) the knowledge of the LS process, and (2) sharing final comments in relation to the LS process during the post-lesson reflection. Although the current study is situated within the South African context, this offers some insights into the possible default characteristic of koshi during the inception of LS in the absence of experienced koshi from whom the aspirant and emerging koshis can learn best practices. To mitigate this, especially in countries where LS is a novel phenomenon, on-the-job training in the form of a blended or integrated approach to develop and empower aspirant koshis should be considered and institutionalised, i.e. an LS training programme highlighting theory and then a practical enactment where the role of koshi is made visible. Essentially, structures such as the institutions of higher learning (e.g. universities) could play a leading role in collaborating with the local Ministry of Education (districts and schools) to conceptualise the training programmes that integrate theory and practice. In other words, in countries where LS is in its infancy, a dedicated and intentional training programme is essential, thereby avoiding the potential paradoxes of koshi. These potential paradoxes include the tension between koshi as an expert envisaged to bridge the gap between theory and practice, and the risk of the blind leading the blind phenomenon. Availability of formally trained koshis who have gone through a programme advocated above guarantees that subsequent aspirant koshis can learn correct practices from them through on-the-job learning, thereby mitigating the possible risk of confining the role of koshi to the knowledge of the LS process only.

While I embrace the roles of a koshi proposed by researchers such as Watanabe (2018), Takahashi (2014), and Fernandez and Yoshida (2012), I view the interpretation and deep understanding of the local curriculum policy proposed by Yolanda as a fundamental role of a koshi. In other words, koshi’s contributions based on fresh research-based ideas about pedagogy and content should be anchored on the prescripts and philosophical underpinnings of the local curriculum policy, especially as it pertains to grade specificity and conceptual progression. For instance, the mathematics curriculum in SA is based on teaching for understanding (Department of Basic Education, 2018); therefore, this differs from problem-based learning that undergirds the Japanese mathematics curriculum (Fujii, 2018). Of course, koshi’s final comments should, therefore, offer fresh research-based ideas about pedagogy and content (Takahashi, 2014); however, it should be done within the confines of the local curriculum prescripts.

While there is often overemphasis on koshi as an outside expert (Fernandez and Yoshida, 2012), the outside-ness needs to be qualified. Is it outside in terms of the school where LS is taking place or outside in terms of not being part of the LS team? Takahashi and McDougal (2016) align themselves with the latter. For instance, an experienced school principal in terms of all the knowledge, skills and attributes as per the roles of a koshi, of the school where school-based LS is hosted, can be a koshi by virtue of not having been part of the planning session. Of course, he is an outside expert relative to the LS team, but an internal person relative to the educational establishment – a school in this case. I suppose the lack of clarity of “outside” often creates confusion in understanding a koshi, which was evident in Yolanda’s case in our reflective dialogue – she could not perceive herself as a koshi because she was always supporting her teachers during the LS sessions, therefore not literally an outside expert relative to the educational establishment, the district in Yolanda’s context.

Researchers placed more premium on koshi as an expert who gives final comments during the post-lesson reflection (Takahashi, 2014; Fujii, 2016). This characterisation of a koshi conveys, intentionally or not, a view that a koshi is confined to a particular stage of LS. On the contrary, in the dialogue we had about our journey to become a koshi, and the view held by Takahashi and McDougal (2016), it emerged that a koshi has a role to play during lesson planning too. In fact, Takahashi and McDougal (2016) advocate for two koshis – internal koshi (part of the LS team) who contribute expert knowledge and skills during planning and the external koshi (outside the LS team) who provide expert knowledge and skills during post-lesson reflection. This will safeguard learners from being exposed to potential erroneous information and/or ineffective pedagogy that can only be corrected post facto. Essentially, having two koshis – internal koshi and external koshi – is appropriate in SA, where the implementation of LS is still developing. To this effect, Fujii (2014) cautions us not to disrespect learners as we implement LS by making them “pawns [or guinea pigs] in order to improve classroom teaching.” (p. 76). Although the context of this caution was on re-teaching the lesson, the same caution could apply to the absence of internal koshi.