Here, we describe the costs and benefits of two interrelated cultural traits that might generate group benefits. The first trait, synchronized cropping, is a cooperative trait expressed at the individual-level. The second trait, a complex of institutions, rituals, and rites associated with subaks and regional water temples, is expressed at the group-level (Smaldino 2014).
Trait 1: synchronized cropping
The two main ecological constraints for rice production in Bali, water scarcity and pest damage, can be overcome by adopting either a staggered or a synchronized cropping pattern. When water is limited, it is in the best interest of upstream and downstream farmers to stagger their cropping (Lansing and Miller 2005) to spread peak water demand over a larger period of time. However, when pest damage is the greater threat to yields, it is in the best interest of both farmers to synchronize their cropping so that fields are flooded and fallowed at roughly the same time.
Both cropping strategies require collective action. Indeed, the individual costs and benefits of these strategies have been described as a two-player game-theoretic problem that is best solved when the upstream and downstream farmers coordinate their actions (Lansing and Miller 2005). Presumably because pest damage is a greater threat to yields than water stress in Bali (Scarborough et al. 1999), synchronization is the most common cropping pattern (Janssen 2007; Lansing and Kremer 1993). So, we focus on this variant of the trait for the remainder of the paper.
Trait 2: institutions, rites, and rituals
The second cultural trait is the set of institutions, rites, and rituals that govern water use and influence collective action. Formal institutions and complex religious rituals are important in Bali for minimizing free-riding (water theft), facilitating coordination and collective action, and providing a cultural foundation for maintaining group solidarity.
Formal institutions, i.e., subaks, are important for reducing water theft and maintaining infrastructure. The existence of upstream and downstream farmers within a group as well as upstream and downstream groups of farmers creates the incentive for individual farmers to free-ride (Lansing and Miller 2005). An individual upstream farmer can take more than his fair share of water because such cheating will have little impact on the downstream group (or on pest populations) provided that the remaining upstream farmers share. Lansing (2006: 67) notes that “…farmers in the upper subaks admit that they are often tempted to take a little more water. But such cheating is rare and usually occurs only in the tiny canals…” To minimize cheating, subaks use fines and other penalties to sanction infractions such as stealing water and missing subak meetings or work assignments. The existence of more severe penalties like having one’s water supply cut off or being expelled from the subak altogether (Lansing 2006) might reflect the historical existence of more substantial forms of cheating.
Perhaps more important for maintaining group discipline, however, is the intricate complex of rites and rituals that have developed around the water temples. These rites and rituals are examples of group-level traits “…not expressed by any single individual in the group, but (which) emerge from the structured organization of different individuals” (Smaldino 2014: 244). Importantly, rituals, institutions, and the maintenance of water temples can entail significant time and effort and impose a heavy financial burden on households (Lansing 2006). As Lansing (2006: 126) notes, “Assembling the many varieties of offerings needed for the annual round of agricultural rites, to take just one example, imposes a relentless series of obligations on households…”
While some rites and rituals play a direct role in coordinating cropping (Lansing 1987, 2006), others have a more symbolic meaning, facilitating group solidarity and cooperation (Norenzayan and Shariff 2008; Watson-Jones and Legare 2016). The internalization of the ritual aspects of a system can reduce the need for costly enforcement and sanctioning (Atran and Norenzayan 2004; Sosis 2009; Sosis and Bressler 2003). These public displays address the challenges that the pursuit of self-interest can pose for a system that is so dependent on cooperation and coordination. Lansing (2006: 195) suggests that rituals also play a cognitive role by helping individuals transcend personal limitations and by instilling, “… a state of mind that is…different from that of Homo economicus.” As Lansing (2006: 136) writes, the purpose of the ritual displays is to help contain individual desires by “…mobilizing the powers of the collective to gain control…” (p. 137). These rituals address the very tension between individual self-interest and the good of the group that lies at the heart of understanding the evolution of cooperation.
Organizational structure of the system
In this section, we examine how these two traits may have emerged and spread at multiple scales.
Three levels of social organization are important: (1) farmers/households, which are nested within (2) subaks, which are nested within (3) regional water networks (see Fig. 1). Whether farmers own, rent, or work farmland, they generally retain rights to what they produce (Lansing 2006) and bear the costs associated with synchronized cropping (i.e., sharing water, maintaining irrigation infrastructure, participating in and contributing resources to water temple rituals). Farmers are clustered into subaks, which can occupy 4 ha to over 800 ha (Scarborough et al. 1999) and include 50–400 farmers, frequently from different villages (Schoenfedler, 2001). Subaks coordinate cropping patterns and irrigation sequences and organize farmers to construct and maintain irrigation infrastructure (Lansing 1987). Subaks are an important first unit for regulating farmers’ behavior, but they are then clustered within regional water networks. One level of this network is an Ulun Swi temple, which serves several subaks whose canals are filled by a common weir or spring. Multiple Ulun Swi temples can be nested under Masceti temples, which unite a dozen or more subaks occupying a common stretch of river (Scarborough et al. 1999) (see Fig. 1). Finally, the main water temple, Pura Ulun Danu Batur sits at the edge of the volcanic lake that is the source of irrigation water for much of Bali (Scarborough et al. 1999). At this temple, priests arbitrate disputes and give approval and advice for the construction of new irrigation channels and tunnels. Subak members are summoned to this temple annually, where they make offerings to the water gods and receive guidance about when to start planting (Lansing 1987). While initial coordination occurs at the Pura Ulun Danu Batur temple, further coordination to synchronize planting times among member subaks occurs at Ulun Swi and Masceti temples at annual meetings of the elected subak leaders (Lansing 1987, 2006; Schoenfedler 2001). Thus, just as farmers cooperate and coordinate with other farmers within a subak, subaks also cooperate and coordinate with other subaks in the regional water network (Lansing 2006). The nested levels of decision making mirror the nested levels of physical irrigation works.
Lansing (2006: chapter 2) combines his own observations and ethnographic data with historic and archeological data (e.g., Scarborough et al. 1999) to convincingly argue that rice cultivation began in simple concave depressions, but increased in complexity to include large-scale coordination among the thousands of farmers in the region. In the first millennium a.d., Balinese farmers created channels and dug tunnels to irrigate terraced paddies on hillocks or “water mountains”. These engineered irrigation works required cooperation and coordination at increasingly larger scales as they connected villages downstream. As Lansing (2006: 42) writes:
“These systems of water control became more complex as new hillocks are added downstream. While a single water mountain is typically managed by farmers from one or two villages, several water mountains are often tethered to one or more irrigation systems, creating the need for water management at a larger scale.”
We suggest that the CMLS framework allows us to understand the emergence and growth of this system as the result of a shift in the dominant level of selection. When the dominant level of selection (e.g., group selection) is above the level of the social dilemma (e.g., individual farmers sharing water) selection can favor group-beneficial traits (Waring et al. 2015).