2.1. Institutional bricolage and path dependence

The definition of what constitutes an institution depends on the school of thought. However, in most theories, institutions are considered social arrangements that structure human behavior (North, 1990; Ostrom, 1990; de Koning, 2011). In line with critical institutionalism, we assume that institutions are complex, dynamic, and contingent. They are inherently place-bound, as they are shaped by local politics, culture, and history (Cleaver & de Koning, 2015; Hall et al., 2014; Jones, 2015; Whaley, 2018). Emphasizing spatio-temporal variability, we follow Frances Cleaver’s definition of institutions as “arrangements between people which are reproduced and regularized across time and space, and which are subject to constant processes of evolution and change” (Cleaver, 2012, p. 8). Cleaver’s empirical work focused mainly on understanding the introduction of new ‘external’ institutional arrangements and their effects on existing, often ‘local’ institutional arrangements (Wang et al., 2018). She shows how actors do not readily accept new institutional arrangements, but negotiate these in different ways. These negotiations and adaptations explain the spatial and temporal variability of institutions. To better understand this ‘negotiation’ process and its outcomes, Cleaver introduced the notion of institutional bricolage (Cleaver, 2001, 2002). This notion seeks to unveil how people, both consciously and un-consciously, modify institutional arrangements based on the social relations, resources, and materials available to them (Cleaver, 2001; Cleaver & De Koning, 2015). Cleaver (2002) discusses various institutional bricolage practices. De Koning (2011) categorizes them into three types: aggregation, alteration, and articulation. She argues that these processes occur when there is an encounter between newly introduced institutions and pre-existing socially embedded institutions which structure local practices (Cleaver, 2002; de Koning, 2011).

Aggregation is defined as the integration of existing institutions with newly introduced ones (de Koning, 2014). This leads to new hybrids that combine or add up unaltered elements from pre-existing ‘local’ institutions with those of the ‘newly’ introduced ones. (Boelens & Doornbos, 2001; Hoogesteger, 2015; García-Mollá et al., 2020).

Alteration is explained as the adaptation of institutions, which implies that the existing, socially embedded institutions and the newly introduced institutions exert a mutual pressure to change. This process entails a redesigning of rules, relations, and norms (de Koning, 2014).

Articulation happens when newly introduced institutions are rejected or resisted (Gebara, 2019). This process takes place especially when there is a disagreement between the existing, socially embedded institutions and the newly introduced ones (de Koning, 2011, 2014), leading to conflicts and confrontations (Boelens, 2015; Boelens et al., 2015).

Institutional bricolage elaborates and explains the concept of path dependence, which exists when choices made in the past affect choices in the present and the future. Institutional bricolage explains path dependence of institutions through the agency and interests of actors, who patch together institutions from the logics available to them, allowing for the simultaneous change and persistence of institutional elements in the process (Sehring, 2009). Power relations are vital in orienting the path dependence of institutions. Thelen (1999) argues that institutions often deepen power asymmetries by giving privilege to particular groups and excluding others (see Mirhanoğlu et al., 2022 in relation to irrigation). Sehring (2009) connects path dependence with these asymmetries, claiming that institutions are persistent because they both perform a certain function and serve certain interests. A particularly elaborate framework on the role of power asymmetries and path dependence has been developed from Jessop’s (2001) work on ‘strategic relationality’, which explains the dialectics between structurally oriented strategic calculation and structurally inscribed strategic selectivity. This body of work not only stresses the selectivity of institutions, which privilege some and hinder other actors; it also explains how actors’ capacities of imagining, creating, (re)producing, maintaining and changing institutions are structured by the very institutions they are embedded in (Jessop & Oosterlynck, 2008). Such a strategic-relational perspective on institutions can explain why institutional change takes the shape of bricolage. Communities of practice develop who share an interest in, and understanding of, certain institutions, defending their persistence even when pressured to change (Van den Broeck, 2010; see also Cleaver et al., 2021). Consequently, actors and institutions are connected through continuous and non-linear processes of institutional bricolage (Servillo & Van Den Broeck, 2012). In this process, Cleaver seems to reserve a role for a material dimension to institutional bricolage by asserting that “an understanding of the ways in which social arrangements also manifest physically (through labour, infrastructure and resource use practices) would significantly strengthen institutional analysis” (2012, p. 20). This opening to ‘the material’, which in irrigation mainly manifests as irrigation technology and infrastructure, is further explored in the next section.

2.2. Introducing ‘the material’ into institutional bricolage

Importantly, the work of Benouniche et al. (2014) shows that ‘bricolage’ in irrigation is not limited to institutions, but also happens in relation to material infrastructures. They discuss how farmers and welders tinker with infrastructures for drip irrigation to take control over the process of technological innovation. They emphasize how such material bricolage occurs in response to -and simultaneously shapes- social institutions, e.g. the adaptation of design standards and regulations. Implicitly, Benouniche et al. (2014) discuss examples of socio-material ‘aggregation’ (e.g., when imported drip irrigation equipment is inserted in local irrigation practices to ‘make it work’) or ‘alteration’ (when locally fitted irrigation equipment was adapted to subsidy standards, and subsidizing institutions adapted their regulations and control to allow for local ingenuity). This echoes earlier work of, for instance, Winner who argues that “individuals are actively involved in the daily creation and recreation, production and reproduction, of the world in which they live” (1986, pp. 14–15) through the use of technologies and infrastructures. The latter alters and (re)create work in social labor arrangements (Pfaffenberger, 1988). Therefore, technologies and infrastructure -and changes in them- bring about ‘significant alterations in patterns of human activity and human institutions’ (Winner, 1986, p. 6). Literatures on infrastructures also emphasize how infrastructure plays a key role in state-society relations (Harris et al., 2018; Kooy & Bakker, 2008; Larkin, 2013; Rodina & Harris, 2016). These insights have received much echo in the irrigation literature where several authors have studied different dimensions of socio-technological transformation (Boelens, 2008; Bolding et al., 1995; Obertreis et al., 2016). Bringing these insights back to the study of institutions, van der Kooij et al. (2015) emphasize “that technical modifications, innovations and tinkering form an intrinsic part of processes of institutional bricolage” (pp. 148–149). Van der Kooij et al. (2015)’s emphasis on the agency of infrastructure and its centrality in institutional bricolage has been taken up in several studies. Ortega-Reig et al. (2017) show how the introduction of drip irrigation triggers both organizational change and adaptations to operating rules (see also Garcia-Molla et al., 2020). Similarly, Kemerink-Seyoum et al. (2019) and Chitata et al. (2021) consider the lens of socio-technical tinkering a way to re-appreciate the materiality of water, its infrastructures, and technologies, and how it influences irrigation governance. From a somewhat different angle, Özerol & Bressers (2017) emphasize the mutual relationship between institutional and material changes triggered by the introduction of large-scale irrigation to a heterogeneous social structure. These studies highlight that technology and infrastructure exist through their replicability; albeit with permutations (Sese-Minguez, 2017; Van der Kooij et al., 2017). This reminds us that, “no technology can be said to exist unless the people who use it can use it over and over again” (Pfaffenberger, 1988, p. 241). The notion that technology exists only in relation to its use through the (altered) reproduction of associated technological behaviors and institutions (Hommes et al., 2022; Shah & Boelens, 2021) is important for a better understanding of the process of institutional bricolage; especially when it concerns socio-material bricolage (see below). Another important notion is that ‘new’ technology and infrastructures are interpreted, understood, adapted, and integrated by its users in very diverse ways that often do not correspond with the original ideas of those that implement them (Bolding et al., 1995; van Halsema, 2002; Venot et al., 2017). As Hommes et al., (2022) point out, this comes from ‘fissures and fractures’ in both how the infrastructure ‘behaves bio-physically’, and more importantly in how actors use, re-use and adapt the technology according to their own interests, knowledge, and institutional/power embeddings.

These and other studies (see, for instance, Dwiartama & Rosin, 2014) point at the realization that technological systems such as irrigation systems are neither social, material, or symbolic only, but ‘all at once’ (Boelens, 2014; Reyes Escate et al., 2022). However, through which mechanisms these different elements influence and relate to each other is often analytically not clear. In the next section, we elaborate on these mechanisms through the notion of socio-material bricolage.

2.3 Socio-material bricolage

In critical institutional studies, the technological or infrastructural dimension of institutional change, as described above, has not received sufficient attention. To bring analytical clarity and a better understanding of the processes through which institutions transform through infrastructural change (in this specific case irrigation modernization), we turn to the notion of ‘socio-material bricolage’.

Socio-material bricolage was coined by Johri, (2011) to describe how global software development depended on developers’ ‘making do’ with the ad hoc set of materials and colleagues at hand. Building on the attention to the agency of material artifacts in Actor-Network Theory (ANT) and Science and Technology Studies (Latour, 1987, 2005; Law, 1992; Winner, 1986) and the understanding of how institutional arrangements and actors’ practices co-shape each other in critical institutionalism (Cleaver, 2012; Cleaver & de Koning, 2015), we propose to translate this idea of socio-material bricolage to the study of irrigation systems. Our aim is not to investigate the ‘making do’ of irrigation actors, but to expand the concept of institutional bricolage and bring materiality to its heart, and to capture the interactions between actors, institutions, and infrastructures. The concept of socio-material bricolage helps us to understand the essence of materiality in institutional bricolage, which we reconceptualize as a strategic-relational dynamic of actors, institutions, and material infrastructures (Benouniche et al., 2014; Boelens & Vos, 2014; García-Mollá et al., 2020; Van der Kooij et al., 2015). It focuses on how both infrastructures and institutions constitute strategic selectivities in enabling or disabling the practices of human actors, and how actors relate to changes in either of them through socio-material bricolage, connecting the old and the new in various ways. In doing so, actors advance and safeguard their interests in, understanding of and control over the system.

Based on these notions, we define irrigation systems as assemblages of water flows, irrigation infrastructures, and institutions, bound together and interwoven with actors and their strategizing practices (see also Hoogesteger et al., 2023a). To analyze the dynamics of these relations, socio-material bricolage captures the material and immaterial interactions between existing and new actors, institutions, and infrastructures. Within these dialectical and differential interactions, power relations and power structures emerge from (1) the diversity of calculating but structured strategies of various actors, some hegemonic, some excluded, some counterhegemonic, (2) the agency of non-human material infrastructures, (3) the strategic selectivity imbued in institutions by actors and their interactions with material infrastructures, and (4) the strategic selectivity of irrigation infrastructures as materialized institutions. Changes in irrigation infrastructures introduced by certain, often new, actors through irrigation modernization projects, imply not only changes in water flows but also simultaneous and entangled alterations in actors’ access to resources, their social positions and power relations, and institutional arrangements (Figure 1). Expanding de Koning’s (2011) systematization of institutional bricolage to socio-material bricolage leads us to redefine aggregation, alteration and articulation as mechanisms not only occurring between new and existing institutions but also as mechanisms that can be identified where new and existing irrigation infrastructures, or new infrastructures and existing institutions (and vice versa) meet. Identification of these three mechanisms of socio-material bricolage brings, we argue, a detailed perspective to the study of socio-material transformations in irrigation systems, while also creating openings for broader socio-material explorations in critical institutional studies.

In this paper, we expand on and systematize the work of Benouniche et al. (2014) by discussing to what extent de Koning’s categories of institutional bricolage can add analytical clarity to the understanding of socio-material bricolage. To do so, we build on empirical data from a case study of a small-scale irrigation system in southwest Turkey, where drip irrigation is currently being introduced. We will expand upon the mechanisms of socio-material bricolage related to water access, water distribution and infrastructure maintenance and management in section 4.

3.1. Case study background

Ağlasun is one of the districts of Burdur province and located in the southwest of Turkey shown in Figure 2. The population of Ağlasun is 7,540 (The Turkish Statistical Institute, 2021), making it a somewhat larger town in the region. Situated in a rather narrow valley, farmland is distributed between valley bed farming and small plot irrigated terraced farming on the valley slopes. In combination with its relative distance to markets, this topography has inhibited Ağlasun’s smallholders from upscaling in the face of agricultural globalisation and neoliberalisation (Kocabıyık & Loopmans, 2019). Consequently, many younger families have either emigrated or found off-farm jobs in nearby cities, whereas tourism has increased the demand for construction land (Kocabıyık & Loopmans, 2021), leaving farming and irrigation largely in the hands of the older generation. Farmers in Ağlasun mostly use water to irrigate wheat, barley, corn, clover, beans, rose, walnut, apple, cherry, and vegetables, such as tomato, pepper, eggplant, and zucchini.

The Ağlasun municipal government is the only legal actor responsible for governing irrigation water in the studied irrigation system. Decisions about irrigation management are taken through the mayor and the municipal council, and there are water chiefs and water guards implementing the decisions in the field. The water guard system has been in force since the 1930s. The municipal government also determines the water fee per hour every year before the irrigation season starts. Water guards record the fee that farmers need to pay while distributing water. Each farmer is responsible for paying their fee to the municipal government by the end of the irrigation season, which is typically the end of September.

There are 24 water sources in the Ağlasun valley which have been used for irrigation over centuries. Since the 1970s, most of the irrigation canals are lined with concrete after important investments were made by the General Directorate of State Hydraulic Works (Devlet Su İşleri, DSİ), the national agency responsible for management of water resources in Turkey. Also, metal irrigation valves were installed along the concrete canal so that farmers could easily diverge water from the main canal to the secondary canals and their fields. Also in the 1970s, the municipal government constructed ponds at some sources to accumulate water and stabilize the flow rate. During this technological modernization, secondary canals, which are the canals that run through the fields to carry water to plots that are not connected to the main canal, remained untouched, and were kept as simple earthen trenches (see Figure 2). A lack of maintenance and difficulties to access canals for cleaning because of vegetation in and around the canals means that clogging and leakages are recurrent problems. Since 2009, the Ağlasun municipal government has invested in the pressurization of these irrigation systems with the aim to facilitate the introduction and use of drip irrigation. However, not all irrigation canals have been replaced by pressurised pipes and different technologies exist side by side in the studied irrigation system (see Figure 3).

The pressurised pipes have separate valves for each plot, which only the water guard is allowed to open and close. Farmers whose lands are irrigated by a pressurised pipe can switch to drip irrigation after installing the necessary equipment on their plots, whereas the farmers using water from the open canals need to dig ponds on their own land to install drip irrigation. Digging ponds, buying a water pump, and using energy to pump water cost farmers a lot of money, which few of them can afford. Hence, drip irrigation is mostly installed in the fields connected to the new piped system (see Figure 3). Farmers can receive state subsidies to switch to drip irrigation. However, one of the requirements to apply for those subsidies is to irrigate a plot of at least 1 hectare. In the mountainous, highly fragmented agricultural landscape around Ağlasun, plots of this size are rare, and few farmers qualify for these subsidies.

The relationships between actors, irrigation infrastructure and institutions were examined through a case study of one irrigation system shown in Figure 3. To conduct in-depth field work on how the materiality of irrigation infrastructure shapes and is shaped by institutional arrangements and actors after the introduction of drip irrigation in Ağlasun, we chose to focus on one of the 24 irrigation networks. The 24 Ağlasun networks are managed in the same way, are all more or less the same size and equipped with similar infrastructure. In the selected Ulupinar network, 128 farmers are involved. The municipal government installed pressurised pipes (110 mm diameter, Figure 4) in one canal, whereas three of them are still the original concrete open canals, allowing us to compare experiences of farmers affected and unaffected by pipe irrigation.

3.2. Data Collection and Analysis

During the fieldwork, data were gathered through semi-structured interviews, participatory observation, and from the archives of the Ağlasun municipal government. The irrigation networks shown in Figures 2 and 3 were traced with handheld GPS, measuring all main canal sections and their material characteristics, and, where allowed, also the earthen distributary canals in the fields. The types of the material infrastructures (earthen, concrete, pipe) of the studied irrigation network were mapped in detail (Figure 3).

The first author conducted the fieldwork during the summers (June–August) of 2018, 2019 and 2021, and the raw data of mapping of the irrigation network was gathered by two cartographers during the summers of 2019 and 2021. A total of 42 farmers, 3 water guards, 1 water chief, the current mayor and the former mayor were interviewed during the fieldworks.

Three interview guides were prepared before the fieldwork. First, an interview guide was prepared for farmers to understand what kind of problems they face during irrigation, specifically related to infrastructure and irrigation methods, how they solve these problems and their perception of the introduction of drip irrigation. Second, the focus of the interviews with the water guards and the water chiefs was mainly on how they manage the irrigation systems on a daily base, the maintenance of infrastructure and their strategies to solve the problems they received from farmers. Third, the interview guide for the mayors essentially included the questions about how they manage the irrigation systems, how they decide to introduce drip irrigation, and challenges they face during their mayorship.

The interview transcripts and field notes were analysed using NVIVO. During the first coding cycle, a general code tree was produced from the transcripts. Then, a map of relational concepts demonstrating the relationships between the codes was prepared and analysed together with the code tree during the second coding cycle.

4.1. Bricolage of collective infrastructures

During the mayoral elections of 2009, drip irrigation rose as an issue in the electoral campaign. The new mayor was elected partly based on the promise to facilitate the adoption of drip irrigation by farmers. He invested in the pressurization of the irrigation system by introducing a plastic pipe network to replace the main canals.

For lack of funds, the replacement became a matter of aggregation of collective infrastructures. The pipes did not replace the entire system but were installed in a part of the network high up the slope which had only earthen canals and was suffering from operational and maintenance problems and causing heavy water loss. Some nearby farmers had already arranged their land for drip irrigation and set up their boreholes and pressured the mayor to make the collective infrastructure fit to feed their drip irrigation. Additionally, most of the farmers owning lands on the newly replaced irrigation network were supporters of the new mayor. Therefore, the mayor consciously chose that part of the irrigation network, which can be considered as clientelism (Mirhanoğlu et al., 2022). What resulted from this aggregation of a new pipe network was a hybrid system where water is guided through pressurized drip irrigation pipes in some parts of the irrigation network and by concrete canals and earthen canals in others. Moreover, the plastic pipe network made good use of the existing concrete canals. Not only were these already connected to all the distributary earthen canals but also, they offered the necessary protection to the plastic pipes so that no new trenches had to be dug for the latter. Hence, the pressurized system of plastic pipes and tubes was installed following the trajectory of the existing main and secondary canals, and most tubes were placed inside the existing canals, an example of alteration of the concrete canals into protective instead of water-bearing infrastructure as shown in Figure 4.

Many farmers, even those not using drip irrigation, welcome the new infrastructure as it reduces their workload. They are no longer responsible for cleaning the canals (see below) and they also started to receive water without system conveyance losses, since water was carried through new and closed pipes.

Yet the collective drip irrigation infrastructure also meets with articulation or resistance, based on farmer’s worldviews, in particular their understandings about the material entanglements of the current irrigation network with the wider landscape. Farmers resisting the new infrastructure emphasize how the original canals do not only water their fields but also serve the rest of nature. They value the green areas in Ağlasun that sprout from them and emphasize how Ağlasun is well-known in the region for its lush landscape and fresh air. People living in neighboring cities visit Ağlasun over the weekend, and a significant group stays for the whole summer, renting apartments from the local population (Kocabıyık & Loopmans, 2021). Hence Ağlasun residents are proud of their green mountain slopes, which they do not only see as an additional source of income through this yayla (highland) tourism, but also consider to be part of their local identity. Some farmers in Ağlasun are strongly against a switch to drip irrigation since they believe that their “green Ağlasun” will be affected negatively. As explained by one of the farmers, whereas the original streams, or even the leaks in the concrete canals, water the plants in their surroundings, pipes do not leak:

“Our first target is to protect the green area in Ağlasun. The greenery is the heritage of our ancestors. You see how it is very green. If we switch to drip irrigation, most trees along the irrigation canals will not have access to enough water and we will lose these green areas” (Farmer Interview, 22 July 2018).

Some farmers argue that water is not only for the trees but also for the other living beings of Ağlasun. They consider drip irrigation a disruption of solidarity with animals and plants:

“They put pipes along the open canal to switch to drip irrigation and transferred water from the canal to the closed pipes. The fruit trees and other trees along the canal are getting dry. Also, the birds, the wolves, and many animals were getting a share of the pie from the open canal. Now, there is no water for them” (Farmer Interview, 2 August 2019).

4.2. Bricolage of irrigation infrastructure at plot level

Where the pressurized pipes were installed, farmers are stimulated to aggregate it with their private drip irrigation system. Farmers wishing to switch to drip irrigation must redesign the irrigation infrastructure in their own field and they must make investments to acquire and install drip lines and solve the problems and challenges that the new infrastructure brought along. First, they need to buy drip irrigation lines which are small, tiny pipes with small holes, connect them to the pressurized pipe system and lay them down on the field, as shown in Figure 5.

The aggregative character of the introduction of collective drip irrigation infrastructure forces some farmers to make alternative investments themselves. For farmers whose fields are not connected to the pressurized system, switching to drip irrigation is more difficult. They need to adapt the water flow from the concrete canal to create a pressurized system at plot level. This is done by constructing a water pond in between the main canal and their drip pipes. Buying drip irrigation pipes and building individual ponds require significant investments, and not all farmers can afford such costs. One of the farmers explains why he has not switched to drip irrigation:

“I do not have the water pond in my field. Constructing a water pond is very costly. Therefore, I am waiting for the municipality to replace the main concrete canal with the pressurized pipe system. Then, the costs will be relatively less. Otherwise, you need to construct a water pond in your field. Then, you need a pump to transfer water from the pond to your field and you also need electricity for your pump. These are all extra costs, and I cannot afford all of them” (Farmer Interview, 31 July 2019).

Farmers who adopt drip irrigation at the plot level are mainly those who are short in labour time. Farm labour has become a scarce commodity in Ağlasun, as the population is ageing and the younger generation seeks employment outside agriculture, and eventually also outside the region. Most farmers are now combining their farm work with other activities (Kocabıyık & Loopmans, 2021). Switching to drip irrigation saves time in irrigation, as drip irrigation requires less supervision, thus allowing for alternative livelihood strategies to aggregate with farming. Additionally, it allows farmers to plant water-sensitive, but less labour-intensive crops. Hence, changes in irrigation infrastructure also aggregate with material changes in land use. In the section of the irrigation system connected to the pipe network, many farmers turned their rose gardens into orchards with walnut trees. In the studied irrigation system, 92 out of 376 plots were recently planted with walnut trees. An older farmer, who recently switched to walnuts, explains how they save labour:

“When you get older, even if you do not shake the walnut, it comes down by itself. I said to my wife the other day: in the future, we get older, we will get our walking stick in our hands and weed walnuts with the walking stick. Gather them and come home. Put them in a bag.” (Farmer Interview, 06 August 2019).

Moreover, some farmers whose plots are adjacent to the pressurized pipe system do not necessarily switch to drip irrigation themselves, but resort to aggregation of the pressurized system with their traditional surface irrigation practice, as shown in Figure 6.

These farmers hesitated to switch to drip irrigation, not only for the additional costs but also because of their understanding of the interactions between water, the land and the crops growing on it. They fear their crops will not get enough water via drip irrigation. This points again at the importance of existing knowledge systems in the local ‘communities of practice’ formed around irrigation institutions and infrastructures. Ağlasun farmers have been doing surface irrigation for centuries, having learnt this method from their ancestors. Drip irrigation conflicts with their accumulated knowledge about how the wider socio-natural system of farming works. One of the farmers explains why he is not in favour of drip irrigation:“I do not use drip irrigation because it is like dripping water to a person who is dying of thirst. Therefore, I do not use it and I do not like it” (Farmer Interview, 08 July 2019).

Finally, some farmers claim that their soil type is not suitable for drip irrigation. Especially old farmers raised this argument during the interviews. One of the farmers explains as follows:

“I guess our soil is not suitable for drip irrigation because it is clay soil. The water does not go below the soil easily. So, look at my field, I irrigated it yesterday evening at 5 pm and you can walk on it now [it is dry]. I irrigated my field for 2 hours yesterday but if I had drip irrigation, I should have irrigated it for 4–5 hours” (Farmer Interview, 27 July 2019).

Resultantly, the new pressurized pipe system near their plots is not tapped into for drip irrigation. Instead, it is aggregated with the material infrastructure used for traditional surface irrigation, by opening the valves into the distributary soil trenches towards the field, or leading the main pipe directly into the field (Figure 6).

Depending on the farmer’s own know-how and interests, ‘modern’ and ‘old’ infrastructures, like institutions, are ‘patched together’ (Cleaver, 2012) rather than being designed and crafted as a whole.

4.3. Institutional bricolage related to collective infrastructure maintenance and management

The infrastructural bricolage discussed above is tightly related to instances of bricolage related to the institutions governing the collective infrastructure, particularly those related to cleaning and maintenance.

Cleaning of canals is essential to keep the irrigation system operational and avoid water losses. As a result, canal cleaning has been one of the most controversial issues in Ağlasun. Until 2009, canal cleaning was a collective responsibility of farmers. The farmers that use water from the same irrigation canal would gather at the head of the canal and walk down the hill, cleaning the main canal until its end. Each farmer was supposed to contribute to cleaning by providing manual labor until reaching the level of their own field. This arrangement often caused conflicts since head-ender farmers, whose field was close to the source, would work less compared to tail-ender farmers, who had to continue working until the end. One of the candidates running for mayor used the canal cleaning issue as a political campaign in the 2009 election and promised to make the municipal government fully responsible for canal cleaning. That campaign won him the elections, and farmers stopped cleaning the canals collectively by themselves. This institutional change triggered a cascade of other minor changes which facilitated the introduction of collective drip irrigation infrastructure. Indeed, whereas farmers could stop cleaning the main canal, they had to continue cleaning the distributary, mostly earthen canals (secondary canals) since the municipal government only took over the cleaning of the main canals. Simultaneously, farmers were dissatisfied with the cleaning by the municipal workers, who would not do it as carefully as they used to do themselves. Therefore, when the municipal government proposed to install the pressurized pipes for the drip irrigation system, the farmers considered this as a beneficial material alteration to the new situation, anticipating that they would no longer have to worry about the cleaning of the main and secondary canals and water losses would significantly decrease.

Subsequently, the introduction of pressurized pipes triggered institutional changes regarding the maintenance of the collective infrastructure. The municipal government is in principle responsible for the maintenance of the main canals, which is now composed of a network of pipes and the valves used to regulate water flows. But unclarity has emerged over the pipes connecting the main canals to the individual fields. The earthen canals were clearly the responsibility of farmers, although this also occasionally led to conflicts between neighboring farmers. However, this time, the conflict is about the institution regulating the responsibilities of municipality and farmers regarding maintenance, pressuring for alteration of old and new institutions:

“The T-pipe that channels water to my land was broken and nobody knows who broke it. I asked the water guard to change it, but they refused to change it since they claim that the T-pipe is not a part of the main pipe. I can’t change the direction of water and I think it is the responsibility of the municipality to channel water to my land” (Farmer Interview, 16 July 2019).

4.4. Institutional bricolage related to water access and distribution

Finally, drip irrigation brought new institutional arrangements for water distribution and use. These arrangements were adapted through redesigning and bargaining by both the municipal government and farmers. Water distribution has been disputed in Ağlasun, at least since the municipal government’s involvement in it since the 1930s. The municipal government appoints a water guard to each irrigation network who organizes the rotation system among farmers during the irrigation season, which typically starts at the beginning of June and lasts until the end of September. One of the main reasons for water sharing conflicts was the favoritism by the water guards (Mirhanoğlu et al., 2022). Typically, they were selected from among the supporters of the mayor and would prioritize the farmers who had voted for the elected mayor. Such unfairness was the cause of many conflicts, including severe physical violence, since the stakes were high: agriculture was the main economic activity in Ağlasun until well in the 1990s, and irrigation water was vital for the farmers. Proposals for a ‘fairer’ water distribution system became a recurrent element in mayoral election campaigns, although it took until the 1984 elections before a mayor succeeded in gaining broad support for a new system. From that time onwards, rules implied that the water guard starts rotation with the head-ender farmer, who calls the water guard after having finished irrigation. Then, the water guard informs the next farmer in line, and the water distribution continues until the last tail-ender farmer gets water. Farmers pay a fixed fee per hour that water is diverted to their fields. The clarity of the rule succeeded in seriously mitigating conflict. Therefore, although water guards and mayors still try to ‘bend the rules’ (Mirhanoğlu et al., 2022) to favorize specific farmers, this rule has now been firmly socially embedded in Ağlasun.

However, some farmers suspect that the materiality of drip irrigation infrastructure has given water guards again greater leeway for favoritism. Water guards decide about the network’s flow rate by controlling the opening of the valves of the two large ponds at the head of the irrigation networks. In the network’s concrete, open canals with a standard diameter, farmers can perform a visual check of the flow rate and confront the water guard if they suspect any unfair treatment vis-à-vis their fellow farmers. In the closed pipes, such visual control is impossible. Some farmers suspect that the introduction of drip irrigation has thus led to an alteration of the informal institutions governing irrigation water, reinvigorating clientelism and favoritism as water guards might abuse the system’s opacity to covertly increase the discharge rate for some while decreasing it for others. This issue is explained by one of the farmers:

“I follow water along the canal and want to see how much water is flowing to my land. When there is a closed pipe system, I will not be able to see the amount of water. If the water guard does not like me or if I am not the supporter of the mayor, the water guard may decrease the flow rate of water from the main valve and I will not be able to check since I will not see the amount of water” (Farmer Interview, 07 July 2021).

To be clear, the switch to drip irrigation is not only accompanied by informal institutional changes. To incentivize farmers to adopt the new infrastructure, the municipal government proposed two new formal rules in 2009. These stipulate that farmers using drip irrigation would be guaranteed access to water every 15 days and would only pay half of the hourly water fee. Under the earlier rotation system, farmers had to wait up to 70 days between two irrigation turns, meaning that the new rotation rule offers significant benefits to drip irrigation users, which allows them to plant more water-sensitive crops. However, the introduction of the rule caused resistance from other farmers. Very few farmers have switched to drip irrigation in the network under scrutiny. Only eight out of 128 farmers are using drip irrigation on 32 of the 376 plots in the system. Prioritizing these would mean that most farmers would have to wait even longer between turns, although drip irrigation farmers were expected to use less water. Since the drip irrigation farmers are located in different parts of the irrigation network (see Figure 3), the water guard needs to redirect water for each drip irrigation farmer every 15 days which causes more delays for surface irrigation farmers. Through their majority representation in the municipal council, surface irrigation farmers were able to resist the formalization of the rule:

“They decided to guarantee water access every 15 days for the drip irrigation users last year. However, they did not make the rule official because most of the council members are farmers, and they do surface irrigation. So, they did not want to give official priority to the drip irrigation users” (Farmer Interview, 15 July 2019).

However, most drip irrigation farmers are ‘social head-enders’, having close social and political relations with the authorities and controlling significant resources (Mirhanoğlu et al., 2022). Notwithstanding the resistance against the new rule, and the impossibility to have it officially adopted, water guards are told by the mayor and their water chief to implement it and prioritize drip irrigation users. However, articulation against the rule remained strong:

“I faced many problems because of the rule of accessing water every 15 days for the drip irrigation users. I asked the water chief and the mayor, and they said the new rule should be followed. But then the surface irrigation users went to the municipal government and made a complaint about me. The mayor told me that I should not have given water twice to drip irrigation users before giving one time to the surface irrigation users. They said the rule is valid, but they do not want to follow the rule they set up. Therefore, it caused a bit of a mess” (Water Guard Interview, 26 August 2019).

Forced towards alteration, the mayor has now agreed to a more informal, ad hoc regulation of water distribution. The new water distribution rule is implemented, until midway the irrigation season, when water becomes scarcer, protest from farmers using surface irrigation rises and forces the water guard to bend the rule, after which drip irrigation users may need to wait more than 15 days.

To avoid conflicts appearing on his desk, the new 2019 mayor decided to decentralize water distribution rules. He commanded all headmen⁴ to ask the farmers in their neighborhood whether they approve the new rule or whether they had different suggestions. This created an even more hybridized system: not only will the different parts of the network apply different rules (a case of aggregation), but also a new actor (the neighborhood’s headman) is included in the governance network. Some of the headmen resisted their incorporation and instrumentalization and did not consult (all the) farmers. This gave rise to additional conflicts between farmers and municipal government.