1 Introduction

The global demands on modern agriculture, particularly those in developed countries, have become extensive. After the agricultural productivism of the 1940s and the 1950s and the food policies for the market and economic development of the 1970s, food policies have recently been challenged through an examination of the interrelationships among climate change, ecological degradation, urbanization, population growth, and ecological public health (Lang et al., 2009). This is consistent with the fact that over the past 50 years, technological development and economic support, including large agricultural subsidies in the USA, the European Union, and Japan, have reduced agricultural production costs and increased overall food availability. However, society has identified negative environmental impacts that result from agriculture (Tilman et al., 2002).

According to the typology by Silvestre and Ţîrcă (2019), recent agriculture requires sustainable innovation that considers both the environmental and social dimensions. Rather than maximizing one particular dimension, an integrated and holistic perspective that gives equal weight to environmental, social, and economic goals is required (Jabareen, 2008). However, to achieve this goal, compromises may be required in each of these dimensions (Silvestre and Ţîrcă, 2019). Therefore, multisectoral and cross-boundary decision-making is necessary. Predicting the outcome of interventions and uncertainty is particularly difficult under highly complex ecosystems and localities.

Thus, collaborative processes that involve diverse stakeholders and promote equality are necessary (DeFries & Nagendra, 2017). Moreover, collective action and partnerships are key drivers of innovation processes (Kutoma et al., 2021). The engagement of diverse stakeholders in a local community is crucial for the development of innovations to address societal challenges (Cipolla & Heloisa, 2011; Kutoma et al., 2021). Therefore, it is important to involve multiple stakeholders, including farmers, to find a balance between productivity and sustainability and achieve sustainable innovation in agricultural technology development.

1.1 Farmer participatory research and multistakeholder platforms

Farmer participatory research has received increasing attention and recognition since the concepts of “Farmer First” (Chambers et al., 1989; Van De Fliert and Braun, 2002) and participatory technology development (Jiggins & De Zeeuw, 1992; Van De Fliert and Braun, 2002) concepts were first introduced in the late 1980s (Van De Fliert and Braun, 2002). In the 1990s, cutting-edge scientific research was combined with various forms of more conciliatory stakeholder participation as participatory approaches advanced (Neef & Neubert, 2011). Platform-based research and development is one such approach. In the context of collaborative development, a platform is defined as a tool or mechanism that serves as a communicative infrastructure and facilitates the collaboration of multiple actors (Kokuryo and Platform Design Lab, 2011). A platform is a systematic technology designed for the development and provision of efficient and individualized solutions through the recombination of different modules and media (Cantù et al., 2021). Multistakeholder platforms (MSPs) have been developed for the context of the collective management of natural resources. Mutual understanding, trust building, social learning, and collaboration in MSPs can lead to actions that cannot be achieved in isolation (Martey et al., 2014). These factors enable farmers to learn together about technologies and options for addressing complex biological, technological, sociocultural, economic, and political problems in agriculture and developing problem-solving and decision-making capacities (Röling & Jiggins, 1998; Braun and Hocde, 2000; Van De Fliert and Braun, 2002).

1.2 Concurrent self-evaluations of multistakeholder platforms for solidarity

Unfortunately, MSPs do not always work. Because different people have different ideas, concerns, and interpretations, the higher the number of involved stakeholders is, the greater the likelihood of conflict and project failure. While there are many different definitions of "failure," in this paper, failure is defined as the lack of solidarity and stakeholder engagement in participatory technology development. Mutual noncooperation is a common failure factor in participatory technology development, although the definition of failure varies among stakeholders. Therefore, maintaining mutual cooperation among stakeholders and avoiding disengagement are necessary to communication workers, as they serve as mediators and facilitators connecting stakeholders in participatory technology development.

Previous research has identified various failure factors in participatory development, including inappropriate goal setting, differences in standards between scientists and farmers, participant categories, insufficient negotiations prior to project initiation (Nederlof and Dangbégnon, 2007), insufficient definitions of appropriate problems and opportunities, a lack of prioritization, and the insufficient establishment of evaluation criteria (Van De Fliert and Braun, 2002). Clearly, the settings are crucial. However, in reality, projects are often conducted without adequate settings. Although the failure of a setting is noticeable during the project, reconfiguring the settings in real time is difficult. The first information received regarding a failure takes the form of complaints from other stakeholders. However, by that time, those stakeholders have often already adopted a certain attitude. In addition, depending on their cultural background, they may not complain but rather act reluctantly. Ngai et al. (2007) found that Asian guests were less likely to directly complain due to the fear of losing face when examining the differences in hotel guest attitudes toward complaint behavior between Asian and non-Asian cultures.

Moreover, reviewing of the cases is important for identifying the causative factors when a project fails. Such factors can be used as lessons for researchers and the government regarding the next project; however, farmers who have participated in “failed” projects are often not given another opportunity to participate. This means that farmers lose the opportunity to learn from the lessons gained through postresearch cooperation. Therefore, it is important for each stakeholder to understand the situation in real time. In the case of an irreversible situation, this provides the opportunity to fix the discrepancies in real time if necessary. The development of solidarity among stakeholders in participatory technology development can help to avoid mutual noncooperation. Thus, when stakeholders do not provide direct feedback, real-time concurrent self-evaluations are important for capturing stakeholder intensions.

A participatory monitoring and evaluation methodology where project objectives and activities are aligned with evaluation indicators based on both planned and implemented indicators and defined jointly with farmers has been advocated by Van De Fliert and Braun (2002) as a means of ensuring the capacity to make adjustments. Furthermore, Re et al. (2023) showed that disasters can be easily overcome through joint coordinated action by creating a common vision for the multistakeholder monitoring and evaluating of water resource management. However, this also requires a setting process. In the process, farmers or citizens are gathered by the planners, such as researchers, communication workers, or local government officers, prior to the start of the project. Thus, collaboration is often led by business elites, as noted by Nop et al. (2023) notes. Therefore, the goal of participatory development should be solidarity rather than collaboration among stakeholders.

A common definition of solidarity is the level of cohesion among group members in regard to specific normative objectives (Bayertz, 1998; Garbe, 2022) and the intermediation between communities and individuals, which results in a form of unity and involves positive morality with positive moral obligations (Garbe, 2022; Scholz, 2008). The concept of social solidarity is normative and descriptive, and it is used to measure the interdependence between groups and individuals (Garbe, 2022; Scholz, 2008). It is also moral (Garbe, 2022), insofar as it (re)produces long-term, mutual, horizontal, peaceful, and interdependent social relations and demands respect for the most disadvantaged members in society (Kolers, 2014; Navin et al., 2018). Social solidarity requires epistemic humility (Navin et al., 2018; Scholz, 2008) and judgment regarding both the means and ends of solidarity work.

Recently, solidarity has been proposed as an alternative to donor-recipient relations (Kajese, 1987; Van Wessel & Kontinen, 2023) and managerialism (Dar and Cooke, 2008; Van Wessel & Kontinen, 2023) in development contexts. It has been interpreted as a group confrontation of oppression, marginalization, and vulnerability. Solidarity has drawn attention in its broader sense in the context of working together to solve common challenges rather than simply helping. However, with regard to development in MSPs, those who decide to establish MSPs are increasingly developing methods for and with vulnerable farmers to promote their participation. This is hardly social solidarity. It is simply an acknowledgment of how solidarity is rooted in inequality and how it can reproduce inequality (Garbe, 2022; Van Wessel & Kontinen, 2023; Wilson, 2017). Concurrent self-evaluations should not be established from a planner’s perspective for the involvement of farmers and the reporting to funders but rather to enable stakeholders to horizontally recognize their respective interpretations and discrepancies and determine whether revisions are needed in real time.

As an indicator of the concurrent self-evaluation method, the focus of this study is on discrepancies in technology. These may indicate a problem with the technology itself. However, in the context of technological development, participatory development is the phase of flexibility in the expectations and interpretations of the technology in each linkage group (Bijker et al., 1989). Stakeholders themselves remain unaware of the discrepancies in each linkage group's interpretation of the technology, which can cause solidarity to become discrete. This leads to a better understanding of the impact exerted by the social uncertainty phase on sustainable innovation (Silvestre and Ţîrcă, 2019). Therefore, the specific drivers that explain the coexistence and conflict of narratives that affect the decision-making process, consensus building, and mutual understanding and affinity among stakeholders must be considered (Ricart and Rico-Amorós 2022).

To summarize, the technological development of MSPs, proactive research taking a narrative approach, and feedback to stakeholders are particularly important in countries with cultural contexts in which discrepancies and conflicts are not easily manifested and where people do not directly express their dissatisfaction but rather not to engage in similar projects in future. However, there is a lack of research on the capacity of concurrent self-evaluations, even if they are poorly set up, to proactively detect potential discrepancies and enable stakeholders to define the situation and make adjustments by themselves in real time. This paper is aimed at extracting the discrepancies from a “failed” case regarding agricultural technology among the stakeholders. Thus, concurrent self-evaluations that can be recognized in a fair and objective way by each stakeholder in ongoing participatory development projects are proposed.

2 Case and methods

2.1 Case overview: consortium for biological soil diagnostic technologies in Hamamatsu, Japan

The focus of this paper is on a consortium, namely the Consortium for Biological Soil Diagnostic Technologies, which was constructed in Japan from 2017 to 2020. This consortium was financially supported by the Ministry of Agriculture, Forestry and Fisheries of Japan (MAFF) in Japan under the “Development and Demonstration of Soil Analysis and Diagnostic Technologies” program (Fig. 1). The consortium verified and implemented the soil fertility index (SOFIX) developed by Ritsumeikan University as a means of diagnosing the biological properties of soil by calculating the total number of microorganisms using environmental DNA and modeling the nitrogen and phosphorus cycles in the soil. In addition, soil diagnoses such as chemical and physical analysis are performed. The biological properties of local organic resources, such as organic manure, can also be evaluated, and these resources can be matched with the appropriate soil (Kubo, 2020).

Fig. 1
figure 1

Overview of the Consortium for Biological Soil Diagnostic Technologies

Full size image

The following explains three characteristics of participatory technology development in Japan. First, Japan has a complex and highly diverse environment. In Japan, the soils are highly diverse and include black box clay, brown forest soil, lowland soil, red‒yellow soil, immature soil, podzolic soil, standing water in soils, and dark red soil (National Agriculture & Food Research Organization, 2021). Motooka (1956) showed that the ratio of reclaimed area to cultivated area is low. Thus, farming systems have been constructed to suit each area’s environment. Consequently, farmers adopt practices that consider the interactions between soil characteristics and inputs.

Second, the development of biological soil diagnosis technologies has progressed. There has been a long history of research and discussion on the chemical and physical properties of soils and technological intervention systems. However, most of the biological properties of these plants have not been fully elucidated. Microorganisms convert organic forms into inorganic forms. They store nutrients in their bodies as biomass and release them when they die in a way that supplies plants. For example, when the oxygen in the soil in rice fields is depleted, microorganisms carry out fermentation and anaerobic respiration using nitrate ions, manganese and iron oxides, sulfate ions, and carbon dioxide to substitute for oxygen (Asakawa, 2021). Thus, in relation to geographical and environmental conditions, soil microorganisms are adaptive and reactive. These adaptive properties have made the development of technological interventions challenging. Recently, however, biological soil diagnostic technologies have been developed and implemented in Japanese farmlands.

Finally, Japan and other Asian countries have cultural backgrounds in which people do not always express direct complaints to other stakeholders. Pelham et al. (2022) reported that there are no differences in the collectivism scores between the USA and Japan among a sample of young adults. However, they also acknowledge that the relationship between agriculture and collectivism is complex and that agricultural engagement and urbanization are drivers of collectivism in both directions. In Japan, even among “certified farmers" who have submitted a business management plan and conduct agricultural activities as a business, 57% of certified farmers are 60 years old or older, and more than 80% are not incorporated (Online document Ministry of Agriculture, Forestry and Fisheries of Japan, 2023). This finding does not correspond to collectivism as a national unit from the dimensions of young adults, urbanization, and collectivism in the rural agricultural communities of Japan.

This cultural context makes participatory development with farmers difficult in Japan and throughout Asia. If concurrent self-evaluations work in the agricultural community of Japan, they could be applied in various Asian countries in which people are less likely to directly voice complaints. The selection of areas in Japan was based on the following considerations. Areas with soil diversity and associated crop diversity were selected. This study aimed at examining how stakeholders differ by level of engagement in MSPs of sustainable technologies. Thus, comparisons of the engagement of farmers growing multiple crops in a single area are appropriate, as comparisons of the same crop in multiple areas involve other variables, such as local government response.

Based on the above criteria, Hamamatsu city in Shizuoka Prefecture was selected as the research area for this paper. Hamamatsu city is located between Tokyo and Osaka, almost precisely in the middle of Japan. The city annexed of 12 municipalities in 2005 m which resulted in a population of approximately of about 800,000 and an area of 1558 km2. Consequently, the area has diverse topography, geology, and natural environments (Hamamatsu City, 2011). As a result, there are many different types of soils in the area and a wide variety of crops suitable for these soils are grown. Furthermore, the SOFIX project in this area represented the largest-scale verification test, and it involved 44 farmers with different crops and farming practices (Fig. 1). Thus, identifying and analyzing the variation in discrepancies in this area is important.

The consortium in Hamamatsu was initiated in April 2017 as a local project called “A Challenging Project of New Soil Cultivation”. The main members of this consortium in Hamamatsu were (1) full-time farmers who make their living from farming and belong to the Hamamatsu Council of Certified Farmers, (2) researchers and a University Research Administrator (URA) at Ritsumeikan University and the General Incorporated Association SOFIX Promotion Organization, and (3) the Agricultural Promotion Division of the Hamamatsu City Government, which took over an office at the Hamamatsu Council of Certified Farmers.

However, three farmers, including the farmer who initiated the project, left during the consortium. The engagement of other farmers decreased in turn. Participation in debriefing meetings decreased each year, not only among farmers but also among people of research institutes and local governments (Table 1). The evaluation of the project by MAFF, which funded the project, showed that the goals were not sufficiently achieved, that further efforts needed to be made in 2018 and 2019, and that the goals of the verification test and research plan had also failed to be achieved. In the final year of 2020, further efforts were still needed to develop a local strategy (Bio-oriented Technology Research Advancement Institution (BRAIN), 2018, 2019, 2020). Stakeholders in the consortium were thus not able to build solidarity. When the project ended in March 2020, the relationships were disconnected and the technologies were not implemented in the area.

Table 1 Participants in the annual debriefing meetings, based on attendance records per stakeholder as provided by Hamamatsu City Hall
Full size table

2.2 Methods

2.2.1 Stakeholders interviewed

The primary data were collected through a qualitative approach from the following stakeholders who participated in the consortium. To the extent possible, all of the stakeholders who were directly related to the area (44 farmers, two local government officials, two University Research Administrators (URAs) as the secretariat of the representative research organization, six researchers, and two organization officials) were contacted by phone call, e-mail, or a freeware application enabling instant communication and other social networking messaging features. With their consent, data were obtained from 18 households (20 farmers), two local government officials, one URA, four researchers, and one organizational officer. From March 2022 to September 2022, individual semistructured interviews were conducted with 18 farming households (20 farmers). The structured questions are as follows: (1) What did you aim for and expect from implementing or participating in the project? (2) What were your perceptions in the first year, the second year, and the third year of the project? (3) What did you explain or understand about the biological soil diagnostic analysis and how to design fertilizer based on it? (4) How did you perceive soil, “good” soil, and “good” crops, and to what extent did you think soil and biological properties contributed to growing “good” crops? (5) Did anything change as a result of the demonstration experiment? Only one interview was conducted online (via Zoom). The other interviews were conducted on-site and ranged from 21 to 120 min in length. In terms of farming practices, 15 households reported practicing conventional farming (11 practice hybrid farming using chemical fertilizers and local organic manures or organic fertilizers, and four use only chemical fertilizers). Three practice organic farming (two are certified according to the Japan Organic Agricultural Standard and one simply practices without chemicals). Regarding the farming environment, eight households used traditional greenhouses with soil, eight farmed in an open field, one farmed in a rice paddy, and one farmed in a tree. In terms of crops, there are 11 vegetable, one fruit, four flowers, one rice, and one tea grown.

Individual semistructured interviews were conducted with three researchers (one from the life sciences, one from the science and engineering field, and one from the social sciences) from Ritsumeikan University. These included the technology developers, including both a researcher from the university and one from the promotion organization. All interviews were conducted online (via Zoom) and ranged from 27 to 72 min.

Individual semistructured interviews were also conducted with two people in charge of the Hamamatsu Council of Certified Farmers. The council is part of the City Agricultural Promotion Division. One of these individuals was in charge prior to the start of the project in March 2018, and the other was in charge from April 2018 until the end of the project. The interviews were conducted online (via LINE, a messaging application) and face to face, and they ranged from 18 to 117 min in length. The interviews were recorded and transcribed. All the interviewees provided consent for the recording.

The data sources used for analysis were agricultural censuses, Hamamatsu city issues, and four paper-based materials. They were prepared by the local government and university researchers, and a preliminary explanatory meeting was conducted with the farmers on 23 January 2017. Three materials published by the MAFF were also used, which related to the development and demonstration of soil analysis and the diagnostic techniques that can be used to evaluate biological properties.

2.2.2 Narrative approach and co-occurrence network

A narrative approach has been applied as a conceptual framework to this research. This framework is not only focused on the subjective world of the narrator but also on the reality constructed through linguistic activity, as multiple meanings exist simultaneously in reality. In other words, the framework provides theoretical elements from what appear to be contingent events up to a macroscopic perspective. This approach does not guarantee generalizability. However, it can provide insights for understanding and explaining individual cases (Nadamitsu et al., 2014). Bamberg (2012a and 2012b) and Hosaka (2014) classified the narrative approach into three categories: (1) narrative that focuses on the subjective semantic world of individuals; (2) narrative conducted in the context of the here and now; and (3) narrative as a linguistic or cognitive structure, background, or context. The third cognitive approach is from the formalist or structuralist stream, such as Labov, who focused on the structure of narratives (Yazaki, 2016), and this is the approach that has been applied to this research. Regarding the conceptual structure, narrators follow existing cognitive conventions to adequately verbalize the content. Therefore, the use of narratives can serve as an exploration of what is considered to be normal in certain social contexts (Hosaka, 2014).

The narratives regarding the biological soil analysis, biological soil diagnosis, the SOFIX, and technologies were extracted from the collected data, the materials published by MAFF related to the project, and the printed materials of the research consortium prebriefing that occurred on January 23, 2017.

As the analysis methodology, a co-occurrence network with KH Coder 3.Beta.03i (Higuchi, 2016) was used. The extracted narratives were analyzed and connected to each other in a line. The co-occurrence network can be useful for visualizing the narratives of each group of stakeholders. In this research, the stakeholders were categorized into (1) high engaged farmers who attended meetings other than the mandatory meetings for three years and conducted outreach, (2) less engaged farmers who attended meetings or study programs other than the mandatory meetings in the first year but not in the second or third years and who did not conduct outreach, and (3) SOFIX members affiliated with the research institute. The engagement of stakeholders best reflected by attending meeting in which they participate of their own voluntary will and without obligation. Outreach refers to the continued use of technologies and/or social learning, such as changes in farming practices. This approach is applied as an attitudinal element of behavioral intention and to the retention behavior carried out by stakeholders, as defined by Blanco-Gonzalez et al., (2020; Cachon-Rodriguez et al., 2019). The analysis of less engaged farmers is important. This analysis leads to a disruption of the consortium, which is the aim of this research. As an important comparison, highly engaged farmers and SOFIX stakeholders were analyzed.

3 Results

In this case study, rather than leaving the project, some farmers simply engaged in minimal participation, and their engagement decreased without any claims. Statements made by farmers and one person from the Research Institute supported these findings.

There were a number of study programs. I was interested in them at first. However, I had no confidence in the person in charge of the research institute. Therefore, I gradually left (Farmer I).

The person in charge of the research institute told me to come for sure. Therefore, I went to a meeting (Farmer O).

To this end, we had no additional cooperation with each other. Working and exchanging ideas with farmers did not work well. There are two types of people involved: Some were willing to talk openly, and others were willing to let it go if it did not go well (Researcher M).

This type of engagement, which we refer to as less engaged, is associated with perceptions of technology. As Table 2 shows, there were discrepancies among the linkage groups, namely, highly engaged farmers, less engaged farmers, research experts, and government officers, in terms of the perception of the technology, the uncertainty of the technology, and the degree of technology completion. The same interpretation of the technology is not mutually shared among stakeholders. This has led to network disconnection. Distinctive keywords were also extracted for each linkage group. Keywords such as “balanced,” “not simple,” and “no dramatic change” were preferentially used by highly engaged farmers. The keywords for the less engaged farmers were “immediate impact” and “result by number.” The words “completed” and “established” were distinctive keywords provided by the government and researchers. The results of the holistic analysis in support of the above are presented below.

Table 2 Interpretation of the technology by each of the stakeholder groups
Full size table

3.1 Highly engaged farmers

As shown in Fig. 2, farmers who maintain high engagement with the MSPs showed that the SOFIX is an approach to put in composts and quantity of fertilizer and then go in the quality of organic matter, nitrogen, and carbon, and increase in nitrogen. However, there are many gray areas in this approach (Fig. 2: Subgraph 01). The balance between fungi and bacteria is different due to their organic origins, even if the cultivated field is the same (Fig. 2: Subgraph 03). In contrast to the chemical soil analysis (of the soil conducted) by the Japan Agricultural Cooperatives (JA) (Fig. 2: Subgraph 05), the results are changed by the person who uses it. Humus was put to the cultivated land, but decomposition was necessary and the results were unfortunately bad (Fig. 2: Subgraph 02). The theory of the right practice of biological soil diagnosis has been taught, but I made a place for myself. Therefore, if it is bad, I am satisfied (Fig. 2: Subgraph 08). I studied it in the project, but I did not see any drastic change (Fig. 2: Subgraph 04).

Fig. 2
figure 2

Co-occurrence network of highly engaged farmers (n = 4)

Full size image

From the co-occurrence network of highly engaged farmers resulting from the overall analysis, it is clear that such farmers are receptive to technological uncertainty and share the perception that a single technology cannot result in the desired change. Therefore, as a back test, an analysis was also conducted on a few individual highly engaged farmers.

According to the co-occurrence network of Farmer G (Fig. 3), the result of the technology is perceived as being dependent on the person using it as well as the balance between the situation of the farmland and the inputs.

There are too many gray areas in the technologies, so if farmers in Hamamatsu want black or white, many people probably do not understand the technologies. The results will depend on the farmers who are using the technology. Everyone would make something different even if they cooked the same meal with the same pan and tools. (Farmer G)

Fig. 3
figure 3

Co-occurrence network of highly engaged Farmer G

Full size image

The co-occurrence network of Farmer K (Fig. 4) also shows that technology cannot cause drastic changes. However, Farmer K has learned from the study of microbes and got outreach in regard to these the projects.

I am sorry that I could not provide dramatically good results to researchers, even though I have studied microbes extensively and contributed to the project. However, the soil preparation and making has led to very slow results. Therefore, it is difficult to obtain the desired results immediately. (Omitted) Studying microorganisms made me realize the importance of the early detection of diseased onions. (Farmer K)

Fig. 4
figure 4

Co-occurrence network of the highly engaged Farmer K

Full size image

A similar statement in regard to learning was offered by Farmer E.

I have started to think very strongly that we have to look at other aspects of the soil in addition to its chemical properties. (Farmer E)

The factors identified in the individual analysis were in line with the results of the overall analysis.

3.2 Less engaged farmers

As shown in Fig. 5 for farmers who decreased engagement with the MSPs, SOFIX puts compost to the soil and, microbes increase in a few analyses handed (Fig. 5: Subgraph 01). The technology is visible to the farmers under actual cultivation conditions (Fig. 5: Subgraph 04). Soil analysis and fertilizer application were designed for the environments to produce crops (Fig. 5: Subgraph 05). If the number of soil mediums is good, the crops will grow healthily, but there is a problem that carbon is needed by crops, and a different fertilizer is ignored (Fig. 5: Subgraph 03). “ I use a lot of organic matter and do not use decontamination pesticides on the cultivated land”, as the professor said. However, I did not get any results from improving the crop or the soil in the cultivated land (Fig. 5: Subgraph 02).

Fig. 5
figure 5

Co-occurrence network of farmers with less engaged farmers (n = 9)

Full size image

From the co-occurrence network of the overall analysis, it can be seen that farmers expected to obtain numerical results, such as an increase in the number of microorganisms and healthier crops. They noted a lack of realistic and specific strategies for achieving such goals. Therefore, as a back test, an analysis was also conducted on small number of less engaged farmers.

Farmer N needed to demonstrate a means of increasing the number of microbes as a concrete strategy, as shown in the individual co-occurrence network (Fig. 6).

Since researchers count the number of microbes, I wanted them to be more specific about how to increase the number of microbes: What should I do and what should I not to increase microbes? (Farmer N)

Fig. 6
figure 6

Co-occurrences network of less engaged Farmer N

Full size image

Farmer A and Farmer I expressed a similar opinions to that of Farmer N.

The numerical value of how much compost was added was not given. This was not similar to the fertilizer design of the Japan Agricultural Cooperatives (JA). I did not have any idea how much compost would increase the number of microbes. (Farmer A)

I thought that the microbe number would increase if I added a large amount of organic fertilizer. (Farmer I)

Farmer H noted that the technology should be implemented after it is established, as both the results obtained in the field and the project itself remain unfinished (Fig. 7). Individual analysis and statements led to these findings.

I think it is a very interesting approach. However, I did not get any results in the field. (omission) In general, the technology is handed over to farmers for implementation after the results of the prefectural trails are obtained, isn’t it? In terms of implementation, this process is meaningless if it is incomplete. However, when I look at the details of the current subsidy program, it seems that farmers are being used without going through this process; the platform is set up and the farmers are told, "We will give farmers money to develop and establish the technology”, so the farmers have to go on by themselves. (Farmer H)

Fig. 7
figure 7

Co-occurrence network of less engaged Farmer H

Full size image

The factors identified from the individual analysis were consistent with the results of the overall analysis. Farmer B, who left the project, and Farmer L both noted that the technology was not yet established and that it was not in line with the existing system.

The seasonal variability of the microbes was that emerged during the project. However, nothing was taken into account. (Farmer B)

The analysis results were hard to understand. These results are completely different from the analysis results we obtained from the Japan Agricultural Cooperatives (JA). It's like, "What is this? (Farmer L)

Additionally, interpretation technologies were associated with the engagement, but crops were not.

3.3 Research institute

As shown in Fig. 8, stakeholders in research institutes showed that when biomass goes in, microbes enter the soil (Fig. 8: Subgraph 02). When the biological, physical, and chemical properties are adjusted, elements do not accumulate and biological substances are abundant (Fig. 8: Subgraph 07). The results differs by local area. Hamamatsu city has a lot of sandy soil (Fig. 8: Subgraph 03). The data were looked at and a good soil medium and design of fertilizer application for agricultural production was proposed to the farmer depending on the purpose. However, there are conditions and climatic problems throughout the year. Farmers have experience with this (Fig. 8: Subgraph 01). The basic research technologies were completed. The project was at the stage of empirical research (Fig. 8: Subgraph 05). It was difficult to discuss consumer appraisals of the correlation between soil and crops (Fig. 8: Subgraph 04).

Fig. 8
figure 8

Co-occurrence network of researchers and research officers (n = 5)

Full size image

It was recognized that the technological base was established when the project was initiated. One researcher, who is one of the developers of the technology, made the following statement.

If the writing of a paper is basic research, obtaining a patent filed and granted is demonstrated the use of the technology. We were already at the stage of patenting the technology when we started the project. So I think we can do it 100% as a business. In brief, I think we are at the stage of finishing the basic research. (Researcher K)

However, Researcher M stated,

In the project, we performed approximately 80% of the empirical research and approximately 20% of the social implementation. At the time, the technology was somewhat established. However, it was not 100% complete. (Researcher M)

3.4 Documents of the local governments and research institutes

As shown in Fig. 9, the documents were prepared by the local governments and research institutes for the preliminary meetings. The documents showed that the SOFIX can be used to analyze the circulation of the soil biomass and evaluate its possible biological biomass and microbes, phosphorus, and nitrogen in the environment (Fig. 9: Subgraph 03 and 06). SOFIX follows fertilizer design using organic matter instead of chemical case (Fig. 9: Subgraph 05). This technology was developed in vegetable cultivation to (determine) fertility indicators toward organic agriculture, which is an approach established through the analysis of fertilizers and resources in the local area and through the decomposition of the fertilizers types. A reduction in the cost of fertilizer for agricultural products was also established (Fig. 9: Subgraph 01). The fertilizer design is based on the measured diagnosis and help the plants to become more active and healthier through improved plant nutrition (Fig. 9: Subgraph 02 and 04).

Fig. 9
figure 9

Co-occurrence network of the documents provided by city halls and the research institutes

Full size image

This project is aimed at developing technology for fertility indicators. However, this approach can also contribute to crop production by reducing input costs and improving plants and nutrients.

The advantages of using SOFIX include an accurate soil analysis of 19 items, a 20% increase in yield, a 30% decrease in fertilizer costs, a reduction in nitrate levels in vegetables, and an increase in the value of agricultural products. (Summary Explanatory Materials of New Soil Cultivation Challenge Projects Explained by the Local Government to Farmers, 23 January 2017)

In response to these quantitative figures, the administrative successor in Hamamatsu City Hall said that the figures and the completion of technologies were interpreted in accordance with stakeholders, as follows:

The technology was explained as established, and the results were verified for various crops. However, the technology has not yet been fully established. (Omitted) The numbers are not guaranteed. This figure was the target of the project. However, farmers interpreted the technology differently. Some farmers perceived that the technology was capable of achieving these numbers. Maybe my predecessor communicated in such a way that the farmers arrived at different interpretations. (Local government N)

3.5 Document published by MAFF

As shown in Fig. 10, documents published by MAFF indicate that the development of new analyses though soil diagnostic technology is needed to decrease the cost of pesticides and fertilizers and increase yields (Fig. 10: Subgraph 03 and 04). These results are verifiable in terms of the number of crops. Verification is being conducted for easy application in agriculture nationwide and to motivate farmers (Fig. 10: Subgraph 01). The approach leads to improved biological appraisal in addition to chemical and physical improvements based on microbial indicators (Fig. 10: Subgraph 02). Furthermore, the objectives are accomplished with evidence (Fig. 10: Subgraph 05). Relevant manuals, support for transition approaches, and formulas for producing and growing are obtained (Fig. 10: Subgraph 06, 07 and 08).

Fig. 10
figure 10

Co-occurrence network of documents published by the MAFF

Full size image

Quantitative results, such as reduced inputs and increased yields, and the rationale underlying such results are needed. Social implementation throughout the country is the goal of the project.

The development goal is to demonstrate soil preparation methods that produce highly sustainable soil based on scientific evidence through advanced soil diagnosis. This will increase crop yield by 20% or reduce fertilizer and/or pesticide costs by 30%. (Details and Objectives of the Research Proposals in 2016 Supplementary Budget, November 2016)

4 Discussion: a concurrent evaluation of MSPs using a narrative approach and a KH coder

Our results are relevant to the discussion of Brown and Mike (2003), who showed that researchers and engineers who are closely involved in technologies report high uncertainty estimates, whereas potential users such as consumers report low uncertainty estimates. They also found that researchers have far-reaching goals and communicate with policy-makers and the public (Brown & Michael, 2003). This is consistent with the low estimation of uncertainty in the preliminary explanatory meeting held by the research institutes and the local government as well as in alignment with the quantitative targets required by the MAFF. The potential users perceived the uncertainty of the technology as low. Their engagement shifted to a low level when uncertainty was detected. Therefore, our result is aligned with the general conception of projects involving multiple stakeholders, including farmers.

Concurrent self-evaluations with a capacity for enabling internal adaptation in real time could be effective at preventing disconnection and creating solidarity. The workshop or discussion forum could serve as a common method of concurrent self-evaluation of MSPs, but as noted by Roncoli et al. (2011), a participatory approach may not be beneficial for all farmers and may be of little benefit to marginalized sectors of society in particular. In some cultures, participation involves cultural social interaction norms that emphasize politeness, modesty, and respect and employs differences in the number of voices based on gender, status, and other factors. In contrast, expression and affirmation are based on western-style values of fairness, equity, and legitimacy. This leads to the concern that in a workshop, the opinions of those with the loudest voices are heard, while the opinions of those with more moderate voices are not. However, a concurrent self-evaluation involving qualitative research from each stakeholder regarding their interpretation of the technology and analysis of narrative co-occurrence might be expected to achieve fairness and objectivity. This approach enables the comparison of various stakeholders’ definitions of the situation and involves the construction of reality or the organization of experiments (Goffman, 1974).

Hoffmann et al. (2007) found that the lack of understanding of farmers' tacit knowledge on the part of researchers was a factor that hindered relationship building (Hoffmann et al., 2007). However, the findings of this study, which has empirical implications for linking stakeholder groups rather than considering researcher–farmer relationships in a dichotomous manner, support the importance of stakeholder interpretation. The presentation of interpretational discrepancies to all stakeholders, regardless of whether a stakeholder adjustment of frames has occurred, can clarify the mutual perceptions of the discrepancy. This approach can lead to an acceptance of the discrepancies among the stakeholders themselves. We can see from the argument that objective concurrent self-evaluation and learning on the part of each stakeholder are crucial for the understanding of each stakeholder of his or her situation.

The research institutions and local governments both recognized that the technology developed represented completed basic research, while it was the field demonstration that was actually insufficient. The researchers and local governments explained the quantitative advantages of the technology, such as the reduced cost of input reduction, in the preliminary explanatory meetings in January 2017. This was based on the funding requirements of MAFF. In this way, misleading statements, such as statements regarding the level of completeness of the research, have the capacity to result in network disconnection.

Some farmers who participated in the consortium on the basis of this information perceived the technology as being complete. They expected immediate results, such as quantitative figures, and believed that a particular technology would lead to drastic changes. These individuals were potentially interested in new, sustainable and innovative technologies. However, when their expectations were too high and when they realized that there was a gap between the explanation and the field implementation of the technology, they could not tolerate the discrepancy, which resulted in reduced engagement. This variance arises from the gaps in conventional agricultural technology dissemination processes. The conventional linear model (Kline and Rosenberg 1986), whose process was established through verification tests conducted by prefectural institutions, was then sent to farmers. However, there was no common understanding of the differences between the conventional model and the current platform model, which involves farmers from various developmental stages, among stakeholders. On the other hand, farmers who accepted the uncertainty of the technology continued to exhibit a high engagement. Their expectations for a single technology were not high. Even though they did not continue to use the technology, they learned from the MSPs.

5 Conclusion and future works

The analyses presented above clearly demonstrates that discrepancies in technology interpretation can affect the engagement with MSPs in verifying or socially implementing technologies related to sustainable innovation for highly uncertain environments, such as soil. For researchers, collecting data and developing technologies via MSPs are high priorities. Therefore, researchers and local governments tend to use words such as “established” and “completed” to engage many farmers and appeal to financial supporters as a sign of success. However, soil-related technologies cannot be equated with other technologies. Soil is the foundation of farming for farmers, and technological intervention in soil has the potential to cause critical effects. A multiplatform approach to soil-related technology development, in particular, should be carefully conducted because of the inherent tensions among stakeholders. Therefore, concurrent self-evaluations during a project are important for identifying discrepancies in stakeholder interpretations. In this way, a space for learning and change can be created.

A prototype of the concurrent MSP self-evaluation methods was developed. Although this was ex-post research, it was found that this methodology could be applied to qualitative research through the use of a narrative approach. An analysis was performed for each linkage group. However, the same distribution could be observed for individuals as this represents a correspondence with the results with similar signs of the individual analysis that had been conducted as a backtest. Therefore, the methodology should also have a certain degree of reproducibility in predicting the risk of disengagement. This methodology can be widely applied around the world because it enables an objective interpretation of technology, although some adaptation may be required depending on cultural context.

In addition, the method can be used to identify stakeholders who are likely to leave MSPs or reduce their degree of engagement through the use of distinctive keywords. If detected, communication workers should conduct follow-ups that are designed to enable stakeholders to notice the discrepancies. Additionally, stakeholders themselves may be aware of the discrepancies, which could create solidarity. Along with concurrent self-evaluation methods, the design of a follow-up system should be prioritized. In future research, ongoing MSPs should be investigated to determine whether this method of concurrent self-evaluation is broadly applicable to participatory research in Asian countries. This methodology is expected to be used by communication workers in the development of MSP technologies for fostering solidarity among stakeholders.