The Food-as-a-Commons Discourse: Analyzing the Journey to Policy Impact

Increasingly, a discussion is emerging on new framings for food beyond food as a commodity. Several initiatives deem food a human right or a common good in the context of a variety of food issues, along the entire food system. This paper focuses on the development of discourses on food-as-a-commons and their success in influencing policy. We explore in detail four discourses: “Open source inputs in agriculture”, “Joint responsibility for food products”, “Reducing food waste” and “Safeguarding food culture and knowledge”. We examine and classify case studies in Germany based on a semi-systematic literature review including policy documents of 12 initiatives that apply and inform these four discourses. This allows us to present various levels of policy uptake, working at different speeds. Identifying characteristics based on commons theory helps us to describe the initiatives better, and especially explain the success of some discourses over others in influencing policy. Results show that discourses that invoke ideas of core human values and are aimed at changing relatively feasible goals (changing resource allocation, but not changing governance or institutions), may be the most likely new food discourses to have policy impact. A prime example of this is the discourse “Reducing food waste”.


INTRODUCTION
Policymakers around the world increasingly engage in food governance, aiming to address food system challenges such as obesity, food waste, or food insecurity (Moragues-Faus, Sonnino, & Marsden, 2017). From civil society too, many evolving initiatives on food are aimed at solving problems in the food system. Analyzing the different framings used in discourses (Tannen, 1993) to describe problems and their possible solutions, scholars (SAPEA-Consortium, 2020;Vivero Pol, 2013b) identify three main ones: food as a commodity, food as a human right and food as a common good. We focus on the latter.
The Group of Chief Scientific Advisors of the European Commission recommends in its newly released report "Towards a sustainable food system" (2020) that food should be considered a common good in order to achieve a sustainable food system. The term "common good" used in such recent policy papers in the EU is a term not used in scholarly literature and can be rather confusing. Authors of that document clearly intend to express with that term that the society should have a say and take part in decisionmaking processes on how the resource should be held and organized. This is closer to the idea of a "commons" than to the classical economic view of a "common-pool good." "Commons" is found for the first time in the literature in medieval times and referred to the use of land in Europe, a use shared by the community or by a group of individuals, necessarily accompanied by a set of rules developed by the user community (Anderies & Janssen, 2016, p. 3;De Moor, 2011. Over time, the term came to include not just shared land but many different and diverse resources, shared and held by a group of people (Anderies & Janssen, 2016, pp. 3-4;Hess, 2012).
In line with many authors (Boyd et al., 2018;De Moor, 2011;Marshall, 2012;Ostrom, 2012) we define commons as a broad set of resources, natural and cultural ones as long as they are held, managed and used collectively. We also believe that the governance arrangement used to provide the resource system and the coordination needed for adequate provision of these resources are key in identifying a commons.
When working towards conceptual clarity in the food-asa-commons debate, we observe different understandings of food-as-a-commons, mainly a resource-based stream of thought and a governance-based line (Vivero-Pol, Ferrando, De Schutter, & Mattei, 2018a). In agri-food policies, the debate navigates between two paradigms: the so-called post-exceptionalist approach (Daugbjerg & Feindt, 2017), in which food is still considered as a commodity but the understanding of the sector slowly moves from being state-protected to market-regulated, and that of the commoners who advocate for an holistic view of the agri-food sector, managed by relying on complex social arrangements among the community that administrates it, according to jointly designed rules. The commoners call for the de-commodification of the system (Vivero-Pol, Ferrando, De Schutter, & Mattei, 2018c).
"Commoning" according to Bollier and Helfrich (2012, pp. 8-19) is the institutionalized sharing of resources among members of a community. It is this institutionalized sharing that confers to any resource its commons consideration (Dardot & Laval, 2017). Commoning demands new institutions, goal setting and forms of interaction (Madison, Frischmann, & Strandburg, 2010;Vivero-Pol, Ferrando, De Schutter, & Mattei, 2018b), thereby presenting the potential to become a new paradigm of societal interaction and organization.
In this paper we would like to acknowledge the diversity in the food system governance debate, as intrinsic to the topic and its complexity, yet it calls for conceptualization.
Thus, food-as-a-commons can refer to natural or cultural resources and materialize at many levels: They can be a result of local action groups that establish their own rules to govern an urban garden or of the worldwide community designing rules to coordinate distribution of food. There must, however, always be an institutional arrangement in place to enable sharing of the good among users (Cheria & Edwin, 2011). In order to treat food and all inputs and outputs in the food value chain increasingly as commons, a transformation of the food system is required (Jackson et al., 2021). This would mean production, trade, marketing, distribution, processing, consumption and related social values, such as knowledge of nutrition, food culture and traditions, acquire joint discursive decision-making processes that go beyond supply and demand. In conceptualizing shared natural and cultural resources as commons, the food system has to be recognized at its various stages.
As food-as-a-commons discourses we define, drawing on Schmidt (2008), a set of ideas on food-as-a-commons informed by various initiatives applying them. There are various food-as-a-commons discourses, some being more successful in gaining recognition in the political agenda or even finding their claims reflected in laws and regulations. When several individual initiatives gain momentum and start being considered in policy debates, the overarching discourse is scaled-up and can transform into policy. Identifying this as success, we regard as one of the most successful food-as-a-commons discourses the one on "Reducing food waste". In contrast, discourses around "Open source inputs to agriculture" have been around for a longer time (Sievers-Glotzbach, Tschersich, Gmeiner, Kliem, & Ficiciyan, 2020) but have achieved less policy impact or influence.
The objective of this paper is to contribute to explaining why some discourses do enter the political agenda while others do not. To answer this question, we first need to define, describe and contrast a number of food-as-a commons discourses. Second, our research question is which characteristics of individual initiatives that underpin the discourses might explain the variances in success. Thereby we work towards a classification of initiatives that implement and inform food-as-a-commons discourses. With this we aim to contribute to conceptual clarity in the debate.
In Section 2 we present the theoretical concepts that help us analyze examples of successful setting of policy agendas, identifying criteria to use in assessing the success of food-as-a-commons discourses. In Section 3 we outline the study design and the methodology that is applied in this work. Section 4 explains our sample, i.e. the four discourses and corresponding initiatives underpinning the respective discourses. In Section 5, we present the results of our study. Finally, Section 6 discusses the results to explain the scaling-up potential of the discourses into policy.

THEORETICAL BACKGROUND
The level of transformation of selected discourses into policy can happen at various intensity levels, ranging from appearing in political campaigns, political statements or records to entering into law and regulation. The highest level of success for a discourse is production of such an intense transformation into policy that it creates a policy paradigm shift. We apply for this discussion Hall´s (1993) definition of a policy paradigm as the interpretive framework of ideas and standards that specifies not only the goals of policy and the kind of instruments that can be used to attain them, but also the very nature of the problems they are meant to be addressing. We further draw on discursive institutionalism, which focuses on the role of discourse and ideas to explain how institutions create and evolve (Schmidt, 2008). According to Schmidt (2008) the power of an idea is exerted through ideational elements such as discourse, practices, symbols, myths, narratives, collective memories, stories, frames, norms, grammars, models and identities. Indeed changes in discourse are always to be considered as indicators of some change taking place among the relevant actors in the struggle for ideas (Barbier, 2012). We thus look at the role and power of new ideas emerging in public discourses in the field of food and agriculture and how they can transform themselves into policy.
Some scholars, seek to combine the paradigm of the commons with an holistic approach of the food system and support the idea of the entire food system as a commons (Pettenati, Toldo, & Ferrando, 2018). They declare that in order to achieve change in the food system, every phase of the food chain must be redesigned under the commoning perspective (Pettenati et al., 2018). We follow a less holistic approach for analyzing the discourses around food-as-acommons. This allows us to identify initiatives that consider food, inputs required in the food chain, or even cultural values as outputs that can be shared and jointly provided as commons. Therefore, we can also define islands of commons in an otherwise commodified food system.
The Cambridge International Dictionary of English (Initiative, 2020) defines an initiative as "a new plan or process to achieve something or solve a problem." According to this entry, initiatives are aimed at producing changes or impeding them, at the individual, the social, the institutional or the political level. They are the instruments through which discourses are materialized and implemented. If we attempt to determine the success a certain food-as-a-commons discourses can bring about contrasted to another one, we need to find concepts that help us to define comparable levels of transformation into policy. We apply four theoretical concepts from political science to selected individual initiatives in order to determine their comparative level of success in policy influence: the heuristic stages model of Lasswell (1927), the multiple streams in the policy process by Kingdon (1984), the concept of policy paradigm shift by Hall (1993), and the dynamics of institutional change described by Lynggaard (2007). Lasswell´s policy cycle model divides the policy process into a series of stages and discusses some of the factors affecting the process within each stage. We follow these widely used stages which are a) problem definition, b) agenda setting, c) policy formulation, d) legitimation, and e) implementation and evaluation (Jann & Wegrich, 2007). Another concept we apply to the analysis is that of multiple streams in a policy process by Kingdon (1984). It separates the policy process into three streams of actors and processes: a) a problem stream in which problems are defined; b) a policy stream involving the proponents of solutions; and c) a politics stream consisting of elections and elected officials.
We also apply Hall´s theoretical approach that focuses on the concept of policy paradigm shift to explain institutional change (Hall, 1993). Once a new discourse has been articulated, a fundamental shift in the dominant policy paradigm may follow two quite distinctive trajectories: a sweeping, abrupt, crisis-driven change or a more deliberate, cumulative, negotiated pattern of change based on the concepts of policy feedback and policy networks (Coleman, Skogstad, & Atkinson, 1996;Hall, 1993). Hall (1993) proposes three different orders of policy change which we apply to describe the level of policy influence of our initiatives: a) the settings (1 st order of policy change) imply the normal policymaking of incremental changes until satisfying agreements have been reached, i.e., routine decision making relatively isolated from social pressure; b) the techniques and instruments (2 nd order of policy change) imply strategic action, i.e., relatively autonomous state action; and finally c) the 3 rd order of policy change, in which judgments are more political than scientific and depend on the arguments and positions of actors, due to a contest of authority over the issues at hand. If proponents of a new paradigm secure positions of power, it could lead to paradigm shift (Hall, 1993). Lynggaard (2007), in contrast, describes the dynamics of institutional change as strongly related to discourse creation: a new discourse usually develops alongside the old one, until the latter displaces the former one. For him, policy change occurs (1) as ideas are turned into discourse (articulation) and (2) as discourse is turned into institutions (institutionalization).

STUDY DESIGN AND METHODOLOGY
It is practically no longer possible to talk about food in an isolated way rather than in terms of food systems, especially if applying the commons theory (Ferrando, 2016;Jackson et al., 2021). For this reason and in order to overview the entire food system and all the potential initiatives and discourses on food-as-a-commons in it, we follow the social ecological systems (SES) framework as elaborated by Marshall (2015). We agree with the arguments that led Marshall to call for a modified version of the SES (socialecological system) framework to be applied to food systems -one in which transformation activities (e.g., food processing) could be accounted for as endogenous to an SES. The original framework (Ostrom, 2007(Ostrom, , 2009 assumed that only activities related to the provision and/ or appropriation of common-pool resources (CPRs) are endogenous to the focal SES. Six first-tier variables were included in the center of the original version. Here, (1) resource system characteristics, (e.g., fishery, lake, grazing area), (2) the resource units generated by that system (e.g., fish, water, fodder), (3) the characteristics of the users of that system, and (4) the governance system jointly affect and are indirectly affected by (5) interactions and (6) resulting outcomes achieved at a particular time and place (Ostrom, 2007). Marshall now adds, besides "products", another first-tier attribute named "transformation systems". Such systems are defined as predominantly human-driven. Transformation activities that add value to resource units appropriated from the resource system, like processing and distribution, are not explicitly covered in the original SES framework. Consistent with the food system definition by Ericksen (2008) which focuses on the full range of activities involved in such a system, Marshall (2015) describes four sets of food system activities: a) primary production of food, b) processing food, c) packaging and distributing food, and d) retailing and consuming food.
In line with that we used a literature review as a systematic way of collecting and synthesizing previous research and structuring it along various food system activities (Baumeister & Leary, 1997;Snyder, 2019;Tranfield, Denyer, & Smart, 2003). In particular, we applied a structured explorative review to find discourses on food that are likewise related to the concept of the commons in current policy documents at the German national level.
To be consistent with this strand of literature, we decided to analyze four food-as-a-commons discourses, which are categorized as follows: "Open source inputs in agriculture" as a primary production food system activity; "Joint responsibility for food products" as encompassing producing food, packaging and distributing food, and retailing and consuming food; "Reducing food waste" categorized as a retailing and consuming food system activity; and, finally, "Safeguarding food culture and knowledge" categorized as a food processing activity for it mainly refers to recipes and traditions on how to prepare food. These four discourses are described in detail in Section 4.
Thereafter, we identified three initiatives that are representative for each of the respective food-as-acommons discourse and can clearly be assigned to one of them. This was done with a cross-cutting analysis of the selected literature. We assigned to every discourse the initiatives in which the explicit wording of the discourse was found repeatedly in the documents upon which they were based. The selected 12 initiatives represent a convenience sample, a type of sample in which the first available primary data source is used for non probabilistic research (Henry, 1990;Saunders, Lewis, & Thornhill, 2016).
Second, we conducted a semi-systematic review, which is designed for topics that have been conceptualized differently and studied by various groups of interdisciplinary researchers (Snyder, 2019;Wong, Cummings, & Ducharme, 2013). A semi-systematic review differs from a systematic review in that it has softer requirements for search strategy and selecting articles for inclusion in the review, so that it could be a better strategy to map theoretical approaches coming from different schools of thought (Snyder, 2019). In this semi-systematic review, we focused on policy documents, i.e. a large set of literature informing political opinion related to the selected initiatives. Within the broad scope of policy documents we reviewed government websites including press releases; civil society organizations' declarations, statutes, and technical papers; website content including media releases, info-flyers, info-articles and position papers from the civil society organizations; general media reports such as documentaries; and studies and reports from research institutes. For each initiative we included 3-4 documents of different types, in total 40 across all the initiatives (see Appendix C). In order to focus the study we stopped analyzing additional material whenever saturation with new information concerning a criterion was achieved (Fusch & Ness, 2015). We explored both a) criteria that give us information on the degree of policy transformation caused and b) descriptive characteristics of the individual initiatives, including core commons characteristics, both shown in Table 1. As outlined in section 3.1 and 3.2, the criteria on policy implementation are derived from political science theory and the general descriptive characteristics are mainly derived from commons theory.
There are numerous methods that can be used to analyze and synthesize findings from a semi-systematic literature review (Snyder, 2019). In our study we decided to carry out a qualitative content analysis, which can be used to analyze and report patterns in the form of themes or categories within a text and identify theory-based analytical units for further research (Mayring, 2014). In order to do so, we used an interpretive and naturalistic approach, observing and counting frequencies, sequence, or locations of words and phrases (Constable et al., 2005).

CREATING POLICY IMPACT CRITERIA
To determine the extent of the policy impact of these initiatives, as shown in Appendix B, we started by applying Lynggaard´s (2007) concept to our sample of initiatives. In a first stage, Lynggaard (2007) states that before a discourse is created, there have to be some ideas on a certain topic emerging. Therefore, we first looked at ideas that were popping up in every initiative. If we could identify some ideas that were repeated in different texts and documents, we assigned them to the first stage of Lynggaard (2007), which is articulation. If these ideas were being articulated into more complex discursive elements and the initiative was explicitly working on the issue, we recognized in the initiative a second stage (institutionalization).
When applying Hall's (1993) model, we analyzed whether or not the initiatives have produced a political change and, if so, to what extent. We considered an initiative has reached a 1 st order of policy change if it had contributed to incremental changes, or different ways of allocating resources; the 2 nd order of policy change would have been reached if there is a strategic action that implies an independent move from the state due to the pressure of the initiative. Finally, an initiative could reach a 3 rd order of policy change, if it had led to a complete change in the policy paradigm. Lasswell (1927) could be operationalized in the same way as Lynggaard (2007): Some initiatives recognize certain problems or represent certain ideas, then these ideas are articulated into a discourse that is used explicitly by the organization in charge in their communication channels. These ideas might then be placed (or not) on the agenda. If these ideas first appear in the political agenda, we are talking about agenda setting. This might mean that the political actor is integrating this discourse into its communication channels, speeches, websites, etc. If laws or policies are written, then we are talking about formulation. But, agenda setting could also occur only in certain niches. That is, these ideas are already accepted in the mainstream of civil society organizations, but only in these particular circles and not in policy or political institutions.
Regarding Kingdon's (1984) model of multiple streams in a policy process, we assigned an initiative to the problem stream level when we found the process to be concerned with agenda setting; i.e., where some interest groups have identified and defined a problem. If these groups or others suggest different approaches to solve the problem, the policy stream applies. When these ideas have reached the political level in the form of regulation, laws or strategies, then we can locate them in the politics stream.

COMMONS THEORY RELATED CRITERIA DESCRIBING THE INITIATIVES
First, we provide criteria that show why the initiatives are initiatives that thus underpin and inform food-as-acommons discourses (see Table 1). To conceptualize, we identified for each of the initiatives the kind of resources they were mainly focusing on. We also labeled the initiatives as whether mainly concerned with the allocation of resource units (a package of seeds) or provision of the resource system (supporting seed banks). We next considered which property rights regime the initiatives propose to govern the respective resource (Feeny et al., 1990;Ostrom, 1986). A property rights regime expresses to which actor the majority of the various bundles of property rights are assigned to (Schlager & Ostrom, 1992). Resource governance can be distinguished by four kinds of property rights regimes -although in practice there are no clear cut categories but rather a mixture of property rights assignments: open access (OA) means that the good can be accessed by everybody and it is owned by none, private property regime (PP) means that the property rights are held by a private person or an organization, common property regime (CP) means that the rights are held by a group of people or any other organizational system, and state property regime (SP) means that the rights are held by the state. In that respect a common-property regime qualifies the initiatives as a commons. Then we drew on Vivero-Pol (2013a), in recognizing different conceptions of food: (a) food is a basic human need that should be available to all, (b) access to food is a fundamental human right that should all be guaranteed to every citizen, (c) consuming and producing food is a pillar and major feature of our culture, (d) food is a marketable product subject to fair trade and sustainable production, and (e) food is a global common good that shall be enjoyed by mankind. Another value we added was (f) food is a medicine that provides health to society (Tirado-von der Pahlen, 2018). We can then see which value judgments regarding food have been expressed in respect of the selected 12 initiatives.
We further differentiated our sample according to the part of the food system the initiatives sought to tackle. We applied Marshall's SESt framework (Marshall, 2015), differentiating the initiatives into categories: primary production of food (P), processing food (PR), packaging and distributing food (PD), and retailing and consuming food (RC).
Finally, we drew upon the four levels of social analysis by Williamson, which have been frequently applied by scholars to explain processes of policy and social change, including change in the agri-environmental sphere (Behera & Engel, 2006;Sultan & Larsen, 2011;Theesfeld & Jelinek, 2017). Williamson's (2000) four levels of social analysis are described as follows: level one (L1) is the top level and contains the norms and culture of society, level two (L2) is formal institutions, level three (L3) governance structures, and level four resource allocation (L4). Drawing on that we made a judgment as to which level in the Williamson framework the respective initiatives address a change and we focus on the predominant one, since some of them were aimed at changing several levels. When the initiative was mainly aimed at transforming -or perpetuating -underlying values of the society (customs, taboos or traditional norms), we assessed the initiative as addressing the level of embeddedness (L1). If the main target was to perform changes in the institutional environment meaning regulation or formal rules, we assessed the initiative as addressing level (L2) or institutional environment. Williamson's third level (L3) or governance would be assessed where an initiative was mainly aiming to effect changes in governance structures. His fourth level (L4) or resource allocation and employment would be assessed if the main objective of an initiative was to transform the way resources are allocated.
Lastly, we analyzed the ideas identifiable in the initiatives. The ideas can be found in ideational elements that, according to Schmidt (2008) change institutions. These elements can be discourse, practices, symbols, myths, narratives, collective memories, stories, frames, norms, grammars, models and identities.

FOOD-AS-A-COMMONS DISCOURSES AND THEIR CONTRIBUTING INITIATIVES
In the following we introduce four discourses that make up our sample including three initiatives for each of the discourses that are further described in the supplementary material (Appendix A). We chose initiatives that were reproducing the discourses in some way; i.e., they were applying the same ideas.

"OPEN SOURCE INPUTS IN AGRICULTURE" DISCOURSE AND ITS INITIATIVES
The term open source was firstly used for computer software, in order to provide free access to users (OpenSourceSeeds -AGRECOL, 2018). Intellectual property protection has been extended in the last three decades to a wide range of information, materials and products relevant to food and agriculture (Halewood, 2013). The extension of Intellectual Property Rights to agricultural inputs and products raises a number of socio-economic, environmental and ethical concerns, including: the increasing risk of a transfer of important knowledge from the public domain to the private domain; the likely negative impact of the agreements on the livelihood of poor farmers; the uncertain impact on sustainable access to affordable, safe, nutritious food for consumers with limited income; and the environmental impact, including the effect on biodiversity (FAO, 2007;Sievers-Glotzbach et al., 2020). In 2001, FAO established the International Treaty on Plant Genetic Resources for Food and Agriculture, creating a global network of seed banks which grants facilitated access for breeding purposes (Rabitz, 2017). The idea of treating inputs as open source grants farmers the freedom of growing (seeds) for replanting, sharing, trading or selling to others; studying and sharing or publishing information about those inputs and also of selecting and adapting the inputs at will as needed (OpenSourceSeeds -AGRECOL, 2018).
This discourse qualifies as food-as-a-commons discourse as genetic material and other inputs to the food system are treated as commons, thus they are made accessible and provided for all users in the community and are taken care of jointly by them. An open source governance regime e.g. for seeds, does however not mean not regulated, instead it follows clearly assigned rules defined by a community that holds the responsibility on how the resources are distributed.
A growing number of initiatives have been proliferating all over the world around the issue of "Open source inputs in agriculture". In the specific case of Germany, it was possible to identify the following ones: Opensourceseeds, Save our Seeds, and Community Seed Banks (for a detailed description see Appendix A).

"JOINT RESPONSIBILITY FOR FOOD PRODUCTS" DISCOURSE AND ITS INITIATIVES
Nowadays increasing numbers of people are taking stock of the food chain and are willing to accept their responsibility for how their food is produced. Consumers want to participate in the food system not only from the shop to the table, but from the field to the plate (Cone & Myhre, 2000;Evans, Welch, & Swaffield, 2017;Hayden & Buck, 2012;Thompson & Coskuner-Balli, 2007). This discourse deals with output resources, i.e. food units in a primary agricultural production process. Initiatives of this discourse focus on the participation of consumers (pro-sumers) in decision-making processes, shared production risk or in the setting up of distribution channels that link producers and consumers directly without intermediaries. The particular governance in place determines the possible level of collective decision on how food is produced and how the products are distributed among the society.
Some examples identified in Germany are: Community Supported Agriculture, Slow Food, Market Enthusiasts (for a detailed description see Appendix A).

"REDUCING FOOD WASTE" DISCOURSE AND ITS INITIATIVES
In 2011, the Food and Agriculture Organization (FAO) estimated that approximately one third of the world's food, meaning roughly 1.3 billion tons of food or enough to feed 600 million people (FAO, 2011;Stuart, 2009) was lost or wasted each year. According to the same source, the global perception of the issue had changed. Since then it has become a matter of even greater concern to public opinion. The 2030 Agenda for Sustainable Development's Target 12.3 of the Sustainable Development Goals calls for halving per capita food waste at retail and consumer levels by 2030, and reducing food losses along production and supply chains.
This discourse qualifies as a food-as-commons discourse since the governance of handling surplus food units can be created collectively and the responsibility on how to deal with food waste is held in common. One basic picture of this discourse is the joint responsibility of humankind to fight against food waste. Some of the initiatives proposed, rely on cooperative solutions to deal with the waste, to be redistributed and revalorized.
The topic of food waste has become increasingly prominent in initiatives all over the world, including in Germany. Almost every initiative, movement, or organization, working on food related issues includes the battle against food waste among its endeavors or at least mentions it. The examples of initiatives for this discourse in Germany that we chose in this study are: Plate instead of Bin, Empty Bin Campaign, Foodsharing (for a detailed description see Appendix A).

"SAFEGUARDING FOOD CULTURE AND KNOWLEDGE" DISCOURSE AND ITS INITIATIVES
There is growing recognition in society about the importance of preserving ancient knowledge and indigenous forms of food and food production. These traditional ways of producing and processing food are mostly examples of protection and sources of bio-cultural diversity, which presents the balance between biological and cultural diversity that makes the food system resilient and sustainable (Petrini & Waters, 2004). In this endeavor, rural women -who account for half of the agricultural labor force across much of the developing world -are vital to promoting diversity for food and agriculture; thus traditionally, they have been those in charge of preparing food and seeding and they have carried this knowledge over generations (FAO, 2019;Padmanabhan, 2011).
This discourse advocates for a preservation of traditions, craftworks and knowledge. Such knowledge and cultural habits are commons, too. They represent cultural resources that are shared, held collectively and are provided with a joint responsibility of the current generation towards the future ones. Initiatives to be found in Germany are Art of Taste, United Nations Educational, Scientific and Cultural Organization (UNESCO) Intangible Cultural Heritage, The Household License (for a detailed description see Appendix A).

RESULTS
In line with the criteria to depict policy influence, we first summarize the results of the empirical investigation on policy transformation achieved by the various initiatives of our sample (Appendix B). Thereafter, we compare the individual initiatives based on particular characteristics that illustrate their diversity, shown in Table 1 above. Likewise, the table provides first indications why some initiatives and the corresponding discourse might be more successful in transforming into policies than others.

DESCRIPTION OF POLICY TRANSFORMATION ACHIEVED
The analysis of our empirical material indicates that in line with Lasswell (1927) all initiatives concerned primarily with the discourse of "Reducing food waste" have attained the stage of agenda setting or even policy formulation. The initiatives aligned with this discourse have achieved a level of influence commensurate with what Lynggaard (2007) referred to as institutionalization. The evidence suggests also that the influence achieved by all these initiatives can be rated at what Hall (1993) referred to as the level of policy change; i.e., where there's an impact on normal daily policymaking. Turning to Kingdon's (1984) streams concept, the evidence from all the initiatives warranted their classification under the problem and policies streams. In the initiatives associated primarily with the discourse "Open source inputs in agriculture", we found that they had attained what Lasswell (1927) classified as the agendasetting stage, but still only in niches. This means that the aims of the initiatives are being articulated and set in the agenda of some interest groups and coalitions but still only in niches. Therefore, we do not perceive any policy change yet according to Hall (1993) but they are in an articulation stage, based on Lynggaard (2007).
In the empirical material that we studied, all initiatives that make up the discourse "Joint responsibility for food products" show no policy change according to Hall (1993). They are in an agenda setting stage according to Lasswell (1927) and the initiatives have been so far only articulated according to Lynggaard (2007).
The initiatives classified under the discourse of "Safeguarding food culture and knowledge" managed to reach an articulation stage (Lynggaard, 2007) around core values of the society and they are not demanding a change or transformation, but appeal to everyone's responsibility to preserve them. They do not present any policy change in accordance with Hall's (1993) concept.
In conclusion, the comparison of the four concepts helps to triangulate the findings. The initiatives that are classified in an agenda-setting or policy formulation stage according to Lasswell (1927), present themselves in an institutionalization stage under Lyndgaard's (2007), both pointing to an advanced level in policy change.
Further, the initiatives that belong to the same discourses present in general the same level of policy change, with slight deviations. Therefore, we can identify a first correlation between the discourses and the ideas the different initiatives present and the level of policy influence they exert.

CHARACTERIZING THE INITIATIVES
Looking at initiatives that belong to the same discourse we found a number of similarities (Table 1). For instance, the initiatives that are tackling the discourse of the "Open source inputs in agriculture" present options for seeds to be governed under a common-property regime. Likewise, the described "Reducing food waste", initiatives do urge a common property regime. Also, all these initiatives have been classified as mainly targeting level L4 (resource allocation) according to Williamson's (2000) levels of social change. The Leere Tonne campaign and the foodsharing initiative do to some extent aim at changing the institutional environment (Leere Tonne) and the governance structures (foodsharing), respectively. The revealed property rights regimes of the initiatives that work towards "Joint responsibility for food produces" are comparatively more diverse, depending what particular activity is focused on. Looking at Williamson (2000) all initiatives that make up the latter discourse seek to produce transformation at L3 and L4 levels. Continuing to regard the different perceptions of food presented by Vivero-Pol (2013a), we can observe that one initiative also incorporate the dimension of food as a commodity.
The initiatives relating to the discourse "Open source inputs in agriculture" present the perception of food in a very similar way. It is treated as a fundamental human right (b), a major feature of our culture (c) and as a global common good that shall be enjoyed by mankind (e). In relation to the Williamson framework (2000) we assigned a L3 level to them, meaning that their main target was specified as changing governance structures.
For the initiatives that build the discourse "Joint responsibility for food products", there is more heterogeneity concerning the perception of food. Whereas all present food in its dimension as a basic human need (a) and as a fundamental right (b) according to the classification of Vivero-Pol (2013a), some approach food as a commodity (d) and others as a common good (e). The Slow Food initiative does also introduce the cultural dimension of food (c) to this group.
There are different ideas presented by the initiatives, as shown in Table 1 (Schmidt, 2008). Here we find high similarities among characteristics of those individual initiatives that belong to the same food-as-a-commons discourse. For instance, we reveal similar ideas in all initiatives making up the "Reducing food waste" discourse: wasting food is a sin addressing responsibility, justice, solidarity, charity; wasting food is inefficient addressing shame and our moral imperative (Andriukaitis, 2017;Gjerris & Gaiani, 2014). In contrast, the ideas that appear repeatedly in initiatives under the discourse of "Open source inputs in agriculture" are: life should not be patented, nature and life as a commons, biodiversity as a commons, sovereignty, and knowledge as a commons. For the case of initiatives and movements under the discourse "Joint responsibility for food products" we find the following ideas: communality, sovereignty, decision-making, freedom, sustainability, proximity, distribution, community-based food production, and good food for everyone. Lastly, the initiatives building the discourse "Safeguarding food culture and knowledge" show ideas such as regionality, sustainability, tradition, self-empowerment, culture, and good food for everyone, knowledge as a commons, caretakers and knowledge keepers for the community.
In the same vein and in line with characteristics based on the categorization done by Vivero-Pol (2013a), those initiatives to which we have assigned an (a) and (a),(b), meaning the resource food we are dealing with is perceived as a basic human need or as a basic human right, thus makes it impossible to agree that it can be wasted. There seems to be a relation between a lack of counter-argumentation and the advantages that the initiatives around the "Reducing food waste" discourse have in entering the policy agenda.

DISCUSSION AND CONCLUSION
A core of ideas has been articulated around the concept of food-as-a-commons, that calls for conceptual clarity. We broke down the debate into several discourses, each harboring a variety of initiatives. We found that one such discourse has already been transformed into policy; other discourses are still early on their journey toward that goal.
We offer a classification of food-as-commons discourses and their initiatives, respectively, that allows for a systematic empirical study. We find that all the individual initiatives that have reached a higher level in either agenda-setting or policy change are found within the food-as-a-commons discourse "Reducing food waste." Thus, "Reducing food waste" is the most successful discourse -with success interpreted as having achieved political action and being on the political agenda. Their policy influence is such, that the Ministry of Food and Agriculture of the German government (BMEL) even initiated the Zu gut für die Tonne (Too good for the bin) program, which is the National Strategy to reduce food waste. The goal is to halve food waste per capita by 2030 and significantly reduce waste along the entire food supply chain (BMEL (Bundesministerium für Ernährung und Landwirtschaft), 2020). The initiatives that we analyzed for the other three discourses "Open source inputs in agriculture," "Joint responsibility for food products" and "Safeguarding food culture and knowledge" have achieved a lower level of policy change or no policy change at all.
There are two elements of the initiatives in the successful "Reducing food waste" discourse which may hint at what can make a new commons discourse successful in achieving policy impact.
The first element involves the ideas central to the discourse. The "Reducing food waste" discourse relies upon ideas that employ core values and beliefs in human nature and do not challenge them. In the more successful initiatives in this discourse we repeatedly find ideas of responsibility, justice, solidarity or charity. They invoke the common understanding that "Reducing food waste" is a moral imperative. By contrast, the ideas in the other three discourses touch less on moral imperatives, and in some cases suggest new maxims that are not (yet) tied to core human values. This seems likely to explain the success of the "Reducing food waste" discourse in being taken up faster on the policy agenda.
The second element that may have an influence on the success of this discourse is the fact its most successful initiatives aim solely at changing the allocation of resources level according to Williamson. Though resource allocation can be difficult to change, the level of difficulty is, Williamson suggests, not as great as can be encountered in attempting to change -in order of increasing difficulty -governance structures, institutional environment, or customs and traditional norms. The two less successful "Reducing food waste" initiatives attempt to change, in addition to resource allocation, governance structures or the institutional environment. All of the initiatives in the other three discourses similarly attempt to change governance structures, institutional environment, or even customs and traditional norms. That may explain why those discourses are journeying a slower road that may or may not ever take them to policy impact.
Our work shows how different initiatives engage to transform the food system and how some policymakers are starting to adopt a commons understanding as part of such a transition. It would be interesting for future research to scrutinize the governance models necessary to navigate towards a commons food system, how the private sector should be involved and how the public would act and react (Vivero Pol, 2017).
We believe that our work points to a research gap on foodas-a-commons discourses in agri-food policies in Germany, particularly when it comes to explaining their different paces of institutionalization. While we acknowledge and accept the inherent complexity of food-as-a-commons understandings, we provided a conceptual structure to classify the various initiatives. Further, case studies can contribute to theory development by understanding the mechanisms that help discourses to influence policy, and ultimately contribute to establish a new policy paradigm.

ADDITIONAL FILE
The additional file for this article can be found as follows: • Appendices. Appendix A, B and C. DOI: https://doi. org/10.5334/ijc.1100.s1