From transfer to translation: Using systemic understandings of technology to understand drip irrigation uptake

Agricultural Systems 128 (2014) 13–24

Contents lists available at ScienceDirect

Agricultural Systems journal homepage: www.elsevier.com/locate/agsy

From transfer to translation: Using systemic understandings of technology to understand drip irrigation uptake Yaakov Garb ⇑, Lonia Friedlander 1 Blaustein Institutes for Desert Environmental Research, Ben Gurion University of the Negev, Sede-Boqer Campus, 84990 Midreshet Ben Gurion, Israel

a r t i c l e

i n f o

Article history: Received 2 October 2013 Received in revised form 15 April 2014 Accepted 23 April 2014 Available online 15 May 2014 Keywords: Drip irrigation Technology diffusion STS Innovation systems Development Socio-technical systems

a b s t r a c t Drip irrigation is a technology with great potential for improving the efficiency of water use, and for increasing crop production and food security by enabling agriculture on marginal land. Yet drip irrigation’s uptake is patchy, with conspicuous successes in some locations and failures in others. In this paper we compare the history and circumstances of the mostly failed uptake of drip technology in sub-Saharan Africa with those of its deep and robust uptake in the Israeli context in which many of the failed African systems originated. We do this not only to throw light on the contextual dependence of this particular technology, and highlight strategies that have been attempted to protect it from this dependence, but also, more broadly, to use the notion of ‘‘technology translation’’ to consolidate several streams of socio-analytic thinking that offer improved understandings of how technologies evolve and travel. Israel has long been a major player in the development and distribution of drip irrigation, with exceptionally extensive national level uptake. We suggest that this emerged from an integrated technology innovation system with a capacity for ongoing multi-leveled learning and dynamic evolution of the technology in light of context-specific potential and problems. Conversely, the failed uptake of drip irrigation in many sub-Saharan African countries can be viewed as a consequence of the transfer of static physical artifacts into new contexts lacking similar local systems into which these could be absorbed and evolve (re-innovated). We interpret two contrasting attempts to boost drip irrigation adoption as efforts to overcome this dependence: simplifying the hardware to become system-free, or creating a kind of remotely operated autonomous small-scale innovation system in which self-contained installations are bundled with resources and linkages to a directing hub. Drawing on several vibrant streams of literature in the sociology of technology and technical innovation, we suggest that the emerging metaphor of ‘‘technology translation’’ provides a better way of thinking about and improving what happens when technologies such as drip irrigation travel to new settings. Technology translation, rather than transfer, suggests a more dialogical approach emphasizing learning and using the local ‘‘languages’’ of the contexts into which artifacts will be translated, making artifacts supple enough to be readily modifiable within these, and finding ways to bolster the local innovation systems that will re-invent and re-link them into new relationships. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction The rising demand for food, and thus, irrigation water, presses against considerable constraint in many countries, which are already experiencing water stress and in which agriculture already dominates the allocation of freshwater resources (World Water ⇑ Corresponding author. Tel.: +972 547 560 667. E-mail addresses: [email protected] (Y. Garb), [email protected] (L. Friedlander). 1 Present address: Department of Geosciences, 255 Earth and Space Sciences (ESS) Building, Stony Brook University, Stony Brook, NY 11794-2100, United States. Tel.: +1 631 632 1196. http://dx.doi.org/10.1016/j.agsy.2014.04.003 0308-521X/Ó 2014 Elsevier Ltd. All rights reserved.

Assessment Programme (WWAP), 2009). As a result, increasing the extent of irrigated land to increase agricultural production cannot be the primary solution for responding to increasing demand for food as the corresponding rise in water demand cannot be sustained. Global climate change and increasing populations will only worsen the problem: the WWAP (2009) predicts that 47% of the world’s population will live in highly water stressed regions by 2030. Against this background, irrigation technologies and practices that increase the agricultural yield per unit of water are critical. Drip irrigation is one of the most promising options for increasing the efficiency of irrigation (e.g. Bucks et al., 1982; Goldberg et al., 1976; Ibragimov et al., 2007; Kang et al., 2004; Shoji,

14

Y. Garb, L. Friedlander / Agricultural Systems 128 (2014) 13–24

1977). The hardware delivers steady, low quantities of water directly to the root area of a crop in a precise and parsimonious manner that increases efficiency, allowing the irrigation of crops in areas where water costs or ecology might otherwise prohibit this, boosting and stabilizing crop production, and thereby increasing the amount and stability of the food supply (e.g. Burney et al., 2010; Polak and Sivanappan, 1998; Postel, 2001; Shah and Keller, 2002; Verma, 2004). Though advanced drip irrigation systems offer the greatest efficiency gains, systems have also been redesigned for simpler and smaller-scale irrigation with little reduction in observed benefits (Polak and Sivanappan, 1998; Polak and Yoder, 2006; Polak et al., 1997; Postel et al., 2001; Woltering et al., 2011). Despite academic and practitioner recognition of the advantages of drip irrigation and extensive promotion over the past two decades, global adoption remains below 4% of total irrigated land area (International Commission on Irrigation and Drainage (ICID), 2012). This paper explores the reasons for this patchy and marginal fulfillment of drip irrigation’s potential globally. We do this through a case study of the ‘‘failed transfer’’ of drip irrigation to two sub-Saharan African (SSA) countries (Ethiopia and Senegal) to offer not only pragmatic insights into how such failures might be reduced, but, also, to force a rethinking of the notion of technology transfer itself. We do this by drawing on a dynamic emerging cluster of new ways of thinking about technologies-in-context (i.e. technological systems), and what these imply for thinking about how technologies arise, travel, and evolve. A full generation has passed since major works such as Pacey (1983) and Blaikie (1985) demonstrated the extent to which agricultural technologies and practices are deeply mediated by and responsive to their cultural and political-economic contexts. Since then, various research traditions have offered increasingly nuanced and powerful understandings of socio-technical systems. We suggest that their collective interlocking insights render traditionally conceived notions of technology transfer and diffusion obsolete, and suggest that the notion of ‘‘technology translation’’ offers a more apt metaphor. Specifically, we review relevant aspects of Science and Technology Studies (STS) and allied vibrant literatures on technology and innovation, and agricultural innovation in particular. The goals and outline of this paper are the following. We begin with a review of how STS and other research traditions on sociotechnical systems have shifted our understandings of socio-technical systems and recast our understanding of the dynamics of what was traditionally referred to as ‘‘technology transfer.’’ We then discuss the fate of drip irrigation technology internationally, and our choice of two different cases for more detailed examination: the spectacularly successful establishment of drip irrigation technology as a mainstay of agriculture in Israel and the ways in which the very same hardware often turned out to be completely useless in the sub-Saharan African context. We then suggest that this discrepancy is due to the deep institutional embedding of the technology as it evolved in the context of its deepest and earliest emergence (Israel), and the vulnerability of sheer physical apparatus ‘‘transferred’’ to the African context in the absence of this broader systemic socio-technical envelope. We then describe two diverging strategies taken by drip irrigation practitioners, interpreting these as questionably successful attempts to protect ‘‘transferred’’ hardware from this kind of contextual dependence. We suggest that the literatures reviewed offer an alternative to this kind of brittle fortification: a series of organizational, communication, and policy efforts that would boost the ability for more fluid translation between the contexts in which technology travels and mutates. We believe this account may be useful not only for understanding the global uptake of drip irrigation, but for advancing the synthesis of theoretical efforts for rethinking technology transfer and their application in various domains.

1.1. Theoretical background: rethinking technology transfer The systemic nature of technology has emerged as a key theme in studies of technology, especially the academic sub-disciplines of STS (Science and Technology Studies), the History and Sociology of Technology, and agricultural systems research. Beginning with the early work of historians of technology in the 1960s, these literatures have offered an increasingly compelling and nuanced understanding of a technology as not simply the thing we usually point to (‘‘telephone,’’ ‘‘car,’’ ‘‘drip irrigation’’), but an eponymous artifact that emerges from, and, in a real sense, is constituted by an extended socio-technical network. These perspectives emerged as part of a broader challenge to earlier deterministic understandings of the trajectory of technical development. By underscoring the contexts in which technological innovation and adoption occur, these newer accounts challenge conceptions of technological development as unilinear (evidencing an inherent technological momentum from less to more advanced technologies), or deterministic (with overly simple accounts of technology shaping the nature of society or society alone determining the directions in which technologies develop). Drawing on these beginnings, the SCOT (social construction of technology) school within STS (Bijker et al., 1987) offered a more complex account of the socially-located development paths and meanings of technology—in fact, it advanced a notion of sociotechnical systems, in which the separation of technology and society is blurred. It argued that technology cannot be spoken of as simply hardware and its ‘‘functions,’’ but is co-constructed in a social context—each shapes and bears the imprint of the other. Thus, a technology’s functions and implications are not fixed but can be interpreted differently—indeed, are different—in different contexts and for different social groups. The question of the ‘‘best technology’’ is, therefore, to some extent an open one; for whom is this technology the best technology? When is this the best technology? How is this the best technology. . . and so on. This mediation by social contexts and processes of an artifact’s nature, function and effectiveness has obvious implications for our understanding of the dynamics of technology adoption, evolution, diffusion, and rejection. Several scholars, notably John Law (1999), Callon (1991), and Bruno Latour (1991), further elaborated these perspectives of technology as a heterogeneous socially-embedded system saturated with power relations. Over the course of the eighties and nineties, they added insights drawn from Foucault, semiotics, and ethnomethodology to forge a vibrant and evolving body of work that came to be known as Actor Network Theory (ANT). ANT proposes a processual, performative, and relational vision of socio-technical systems: ‘‘entities achieve their form as a consequence of the relations in which they are located. . . they are performed in, by, and through those relations’’ (Law, 1999). The links and structures of a system are not given but continually produced. But, at the same time, they gain a degree of solidity and regularity. A key notion in ANT’s description of socio-technical dynamics is that of stabilization. This occurs when a system or parts of it are robust enough so as to become routine and invisible: a black box that can be used and relied on, with no need to question or even examine its innards. Networks (human and nonhuman) struggle to achieve such stabilization by ‘‘enrolling’’ other actors into using and strengthening durable links that ‘‘serve’’ them. Stabilization (and the related notion of ‘‘closure’’) occurs when such an assembled web of alliances, linkages and understandings becomes too robust to challenge or unravel. The notions of technological closure/stabilization have been drawn on and developed in useful ways. Star (1999), for example, describes a kind of stable background system, which she characterizes as ‘‘infrastructure’’—an invisible support system that people rely on,

Y. Garb, L. Friedlander / Agricultural Systems 128 (2014) 13–24

deeply embedded in other structures, social arrangements and technologies. Infrastructure achieves this kind of fluent transparency by standardized interfaces to other tools and infrastructures, being built on an already installed base, and growing or being fixed in increments. It is not something created all at once, in a top down manner, but is pieced together, extending the degree to which it is increasingly and transparently relied on by members of a particular community of practice (Star, 1999). The increasingly systemic theories of socio-technical change emerging from the history and sociology of technology suggests a perspective that resonates, in some ways, with parallel developments in what has, until recently, been a largely separate set of literatures addressing innovations (emergence of new ideas, behaviors, and technologies) and their spatial and temporal diffusion. In particular, these literatures have evolved away from the classic Rogerian (Rogers, 1962) approach to the diffusion of innovations toward a more systemic nature of technology and technical change. The traditional diffusion paradigm was prominent in the 60s and 70s, implicitly or explicitly informing work by anthropologists and sociologists and other researchers who applied it to diffusion phenomena in a range of fields such as public health (the spread of practices, drugs, programs), rural sociology (the spread of farm technologies), marketing (knowledge and use of new products), and communication (the spread of news). The traditional diffusion/transfer model (which undergirds the ‘‘technology transfer’’ model) implicitly posited a technology that has been invented (once and for all) and is then transported or copied to other sites where it is rejected or adopted, in which case it will have impacts. The questions addressed under this paradigm revolve around the channels and rates of adoption, how costs and benefits shape adoption, the role of early adopters, their networks, and promotion, the nature of barriers, etc. Often implicit in this model (and the crux of the eventual opposition to it) is a conception of the technology emerging from a ‘‘core’’ (in developmental terms), as a stable object with fixed characteristics, whose benefits are obvious, while the recipients in this equation reside in a ‘‘periphery,’’ who are grateful adopters, unaware of the benefits, or hampered by barriers and market failures that foil what would otherwise be the rational choice of adopting a technology whose benefits exceed its costs. Early critiques of the diffusions approach were directed primarily at the dichotomies implicit in the diffusion metaphor, arguing that they mirrored an imperialist worldview (Blaut, 1987). More recently, the transfer/diffusion question has been tackled from perspectives that view the broader systems in which technology emerges. In the nineties, for example, the role of institutional contexts supporting innovation has been examined from an Innovation Systems approach, emerging from institutional theories, evolutionary economics, and the study of innovation clusters. This approach posits interactions of resources, people, and knowledge within spatial, cultural, and institutional crucibles as an important source of innovation. The interconnections and learning that occur within these systems accelerate the development and robust diffusion of new technologies. The approach is explicitly pitched against the technology transfer (or ‘‘technology supply push’’) approach, which emerged from the experience of striking growth in agricultural productivity in the USA, and out of which the Rogerian diffusion paradigm emerged (Hounkonnou et al., 2012), with its linear ‘‘extension’’ conception of agricultural research moving from village level workers to ‘‘end users.’’ Innovation systems perspectives represent a significant departure from traditional, linear perspectives on the creation, dissemination and utilization of knowledge (Soete, 2014). These shift focus away from central organizations (public sector research institutions) and the linear relationship between research, education and extension, emphasizing the bidirectional and cross-cutting

15

interaction between end-users and technology in a complex socio-economic network. There are obvious linkages between this and previous work within STS on technical development trajectories, and increased interplay between them (Geels, 2004). Innovation systems approaches have been extensively applied to industrial systems in developed-country contexts (e.g. Carlsson and Stankiewicz, 1991; Carlsson, 2006), but more recently have been brought to bear on the issues surrounding agricultural development, including in developing countries (e.g. Spielman, 2005), and SSA in particular (Spielman et al., 2009). This journal (Agricultural Systems) has emerged as a central venue in which this approach has been developed over the last decade (e.g. Hall et al., 2003; Horton and Mackay, 2003; Kilelu et al., 2013; Klerkx et al., 2010; Morriss et al., 2006; Ortiz et al., 2013; Spielman et al., 2008; Temel et al., 2003). This agricultural innovation literature has generated a series of concepts of value for the focus of this paper, namely a more nuanced and systemic alternative to the ‘‘technology transfer’’ model of how technologies migrate and mutate. For example, it offers technography (an ‘‘ethnography of technology’’) as a methodology for studying socio-technical and innovation systems that emphasizes how tools and techniques are performative and situated, distributed, and dependent on institutions, and offers approaches to closely study the shaping, use, and impact of technologies in specific social situations (Jansen and Vellema, 2011). The notions of innovation platforms (Kilelu et al., 2013; Spielman et al., 2009), boundary objects and actors (Jakku and Thorburn, 2010; Klerkx and Leeuwis, 2009; Klerkx and Proctor, 2013; Klerkx et al., 2012; 2010), and the concepts of socio-technical co-evolution (Kilelu et al., 2013; Leeuwis, 2013; 2004), have been elaborated and applied to the kinds of concepts we examine here. Of particular relevance is the reinvention of the conception of agricultural extension as centrally revolving around tasks of communication and innovation, network building, learning, co-design, and negotiation (Leeuwis, 2004). These tasks contrast with the traditionally conceived role of the extension agent as a link in the distribution network that moves hardware from (research) centers to peripheral end-users. Relatedly, the kind of multi-leveled multi-actor exploration and adaptation functions required in the ‘‘scaling out’’ strategies reviewed and illustrated by Millar and Connell (2009) represent precisely the kind of alternative to the linear model of efforts to simply increase adoption of an existing technology that can be used to embed a (transferred) technology in new contexts. Though the ANT conception of socio-technical systems and various innovation systems frameworks emerge from different research traditions, and retain some quite foundational–and perhaps irreconcilable–philosophical differences, both offer a fresh approach to the set of issues traditionally referred to as ‘‘technology transfer.’’ For a constructivist and, especially, ANT framework, technology is not a self-evident black box or stable object, but a network of various social and technical elements, which negotiate and align their respective interests as they work to enroll one another into chains of association that serve their individual and institutional ends. Some of these associations are stable enough (for the time being) to be treated and function as a durable whole, offering a ‘‘technology,’’ a black box that can be treated as a thing. The linkages and the efforts that go into the seemingly self-evident stability and function of such a black boxed technological entity may become invisible with regularization and stability, but they must be constantly (re)performed, otherwise these associations will dis-integrate, and the technology will ‘‘fail’’ or become obsolete or irrelevant. Understandably, a technology’s meanings and uses may change over time, or differ in different contexts, since it will be linked up in different ways and to different actors—it will be, in some real ways a different technology, despite a similar outward physical appearance. As Andersson and D’Souza (2013) have demonstrated with respect to

16

Y. Garb, L. Friedlander / Agricultural Systems 128 (2014) 13–24

Conservation Agriculture among smallholders in southern Africa, even the studies and figures on the degree of adoption of new practices and technologies are contextually shaped. Thus, in many ways, the process of technology ‘‘transfer’’ is not that different from the process of technological innovation. Both involve stabilizing a network through forging new social and technical connections between heterogeneous elements (ideas, actors— whether institutions or people—resources, and artifacts), which work to translate their own interests into those of others and vice versa. The difference between the work done in innovation and that done in relocation (as a kind of re-innovation) is that in the latter, an object is to a greater extent and more explicitly a given starting point for the enrollment of additional actors and the reassembly of a new system. Clearly the notions of ‘‘transfer’’ and ‘‘diffusion’’ are somewhat misleading metaphors for this process. The hardware that could be said to be diffusing or transferring is the tip of the network iceberg. In our case, the irrigation pipe and drip emitters might be the same in a new setting, but even were these to be spliced into new systems, their alignments and functioning might be quite different, so that the technology could very well carry new meanings, advantages, costs, and applications: at the extreme, one might argue that it is no longer the same technology. To reflect this perspective, it might be better to talk of technology translation, rather than transfer. (In fact, the enterprise of ANT has sometimes been referred to by Callon and others as the ‘‘sociology of translation’’; this notion, drawn from the philosopher Michel Serres, has several meanings, which we cannot unpack here (Brown, 2002)). A translated system is both continuous with and different from the ‘‘original’’ one—as Serres points out, translation allows a distortion, even betrayal, of original sources (Brown, 2002). We suggest, also, that the metaphor of re-invention, introduced by Rogers in his later writing (e.g. Rogers, 1978; Rice and Rogers, 1980), or of re-innovation might also be useful, and allow easier linkages with the ‘‘innovation systems’’ frameworks. We are not suggesting too dogmatic a cleavage between the ‘‘transfer’’ metaphor and a successor ‘‘translation’’ alternative, but a spectrum between two end points at which one or the other metaphor is most appropriate. In cases when a technology is relatively self-contained and/or amenable to linkage with analogous or adapted systems in new settings, its systemic complexity is a less dominant (or, at least, visible), factor, and it will be easier to conceptualize the process as one in which hardware is moving from one site to another. In other cases where the technology is deeply coupled to its surrounding material and social contexts, and/or not readily linkable to systems in a new context, its systemic character becomes critical for understanding what happens in new contexts. Here the ‘‘translation’’ and reinvention metaphor is far more suited to describe the process. We would argue that drip technology is one of these latter instances. 2. Israel and Africa as illustrative endpoints of global drip irrigation uptake We now turn to our two case studies as a more textured example of the relevance of this alternative perspective on technology uptake, and to throw light on the pressing and practical question of drip technology’s striking disuse in settings where it would seem to offer much promise. Our analysis is based on interviews and site visits in Israel and two African countries, and a reading of relevant primary and secondary documents on the development and transfer of drip irrigation between them. The Israeli case is not only an example of the deepest uptake of drip technology, but, also one that played a critical role in its development, offering a role model of drip irrigation’s feasibility and promise, at the forefront of conceptual and commercial efforts to propagate it to other contexts, especially in the developing world.

The second and in many ways opposite case is that of two SSA countries, where the lack of uptake, indeed, rejection, of the technology, is marked and well known. We suggest that the differing fate of the technology in these two settings can best be understood precisely through the lens of systemic understanding just described. Drip irrigation in Israel is rooted in and inseparable from an extensive infrastructural network so pervasive and successful as to be nigh invisible. Upon arrival in the SSA context, this same hardware is stripped of this contextual cocoon, and has not been able to reinvent itself in and into new contexts. Our emphasis on technology translation suggests that further efforts to incentivize and boost uptake will not be able to achieve the hoped for transfer of drip irrigation technology to African farmers unless these grapple with the need to translate—not merely transfer—drip irrigation technology into new systemic contexts. Large-scale adoption of drip irrigation is rare, and even in countries in which drip irrigation has been adopted by a significant proportion of farmers it remains a much smaller part of stable agricultural development than rain-fed agriculture and alternative irrigation technologies. In fact, Israel, which invested heavily in the early development of drip irrigation, is the only country in which it is used by a majority of the agricultural sector. According to the International Commission on Irrigation and Drainage (ICID), drip irrigation accounts for 73.6% of all irrigated land in Israel (sprinkler accounts for a remaining 26%). Thus, the Israeli drip irrigation experience represents a case of early and still unparalleled national scale adoption. As a comparison, India and the USA are two other countries that can be considered representative of successful drip irrigation adoption, though at substantially smaller scales relative to the overall size of the agricultural sector and available arable land. In 2009, the United States had 1.64 Mha under drip irrigation out of 24.7 Mha irrigated land (ICID, 2012), representing 6.64% of the total irrigated land in the United States, a respectable number well above the global average. In terms of total land area under drip irrigation, India has the largest drip irrigated area of any country in the world with 1.90 Mha under drip irrigation (followed by China and the United States, in that order). Relative to its total irrigated area, however, drip irrigation accounts for only 3.12% of the irrigated land in India. By comparison, China, has a similar area under drip irrigation (1.67 Mha), but this comprises only 2.82% of the country’s total irrigated land area (ICID, 2012). We move now from the case of Israel, with deep national-level adoption of drip irrigation, to consider adoption in developing countries. As we would anticipate, in the absence of favorable institutional environments, government incentives, and broad technological availability, these countries have very low or negligible levels of adoption, and this is the case for SSA, which constitutes the opposite end member case study in this paper. According to a 2012 ICID report (giving usage values for 2007– 2009), only four African countries reported any level of drip irrigation usage: Morocco, Egypt, South Africa and Malawi. Overall, drip irrigation accounts for less than 1% of African irrigated land area, which in turn accounts for under 5% of total arable land (Food and Agriculture Organization of the United Nations, 2012). The untapped potential for drip irrigation is particularly disturbing in SSA, where this technology could help mitigate seasonal or cyclical drought cycles and hunger, which affect nearly all smallholder households there (Burney and Naylor, 2012). In fact, most of the successful drip irrigation adopters in SSA are predominantly large commercial farmers, and despite concerted efforts to encourage adoption by smallholders, successful adoption is rare and particularly susceptible to being temporary, followed by abandonment (Belder et al., 2007; Kulecho and Weatherhead, 2005) . To summarize: at one end of the continuum we have Israel, with extensive and deeply institutionalized adoption; a few

Y. Garb, L. Friedlander / Agricultural Systems 128 (2014) 13–24

countries (in particular, the USA and India) in which proportional adoption is low, but gradually increasing; and then the rest of the world—with Africa as an end point—in which uptake is negligible and often in decline. In the following sections we examine these two endpoint cases in more detail, and, drawing on the perspectives reviewed in Section 1, underscore their differing degree of systemic coherence and embeddedness. 2.1. Drip Irrigation in and from Israel 2.1.1. Slow, sustained, and precise nourishment to the roots: the infrastructures that nourished drip technology in Israel The precise measured delivery of water drops from plastic pipe to a plant root in an Israeli field is undergirded by the existence of a complex, extended, and in many cases taken-for-granted network needed for the technology to ‘‘work.’’ This network or socio-technical system is physical, informational, institutional, economic, behavioral, and cultural. For example, the water is reliably of a certain quality and at a certain pressure. The spacing and width of the pipe and quantities of water are matched to soil, crop, and weather and, increasingly, adjusted in real time through automated delivery systems responding to sensors. Farmers can call on a range of institutions to advise on, sell them materials for and repair the system, or adjust the system to their changing needs. The crop is grown in a political-economic context with relatively stable costs of inputs and crop prices, as well as assured channels for delivery. The entire agricultural and water establishment is familiar with and endorsing of drip technology. Risk exists, but is buffered by insurance and assurances. The farmer is in dialog with salespersons, extension agents, other farmers and political representatives of his interests in order to increase the stability and profitability of his enterprise. A farmer’s skills, daily work patterns and cultural frames of understanding are all fairly well aligned with the demands and goals of the technology. This extensive drip irrigation ‘‘system’’ is not a static one, but integrated with and evolving through research and improvement efforts. The ongoing connection between researcher, field agent and farmer was (and is) critical to the success of Israeli drip irrigation and other agricultural technologies. In Israel, the Irrigation and Soil Field Service (I&SFS) was established explicitly to meet the needs of Israeli farmers for agricultural extension in irrigation. It is a highly responsive government agency that goes beyond merely providing agricultural extension services of static products and knowhow—it is a learning organization, actively conducting research and development. These activities are in addition to relaying farmer concerns and experiences with new technology back to the Israeli Agricultural Extension Service. Since its founding in the state’s first decade, the I&SFS has ‘‘. . . not view[ed] its role as merely a transmitter of know-how. When solutions were needed to crucial problems, it set up field trials and test plots. About one-third of its effort is devoted to research. . . Cooperation with researchers is essential both for keeping extension workers up to date and for conveying the problems that farmers encounter (Sne, 1989).’’ As summarized in Table 1, the development of this remarkably extensive and vertically integrated system supporting the evolution and use of drip irrigation was enabled by a unique and compelling confluence of social, economic and cultural circumstances operating at multiple scales (the individual farmer, the agricultural settlement, and the state). The demand for a technology that would enable agriculture on marginal land was over-determined by several factors: an arid environment; the Zionist ideological focus on working the land of Israel; and the geo-political importance of establishing viable settlement along disputed borders. Another enabling factor was the highly centralized nature of the water system—in terms of both control and pricing. This was a geographical necessity as Israel’s largest freshwater resources are located in the

17

north of the country, save the coastal and mountain aquifers. The early state’s geo-political situation made promotion of settlement in the much vaster and more sparsely populated Negev Desert in the south of Israel a strategic imperative, and to do this water resources had to be physically transported to the south and carefully allocated to be shared throughout the country. This was enabled by centralizing and regulating water allocation in the hands of a National Water Authority, which commoditized water and set in place a management infrastructure that continues till the present. At the heart of this effort was a pressurized national water supply system—the National Water Carrier—to convey water from the Lake of Galilee in the north to the southern parts of the country. Continuing till today, water allocations are rigidly enforced in Israel with farmers charged according to water used. Increasing water use efficiency is, thus, a shared national scale economic and environmental imperative for Israeli farmers. The government controls not only how much, but also the quality of water available to farmers in Israel, and water quality for agriculture is essentially guaranteed by the government. This makes it very easy to design the specifications of agricultural technology to match the standardized quality of water provided to farmers. In all these respects, the Israeli context was primed for the development of drip irrigation (Hillel, 1989; Melamed, 1989; Sne, 1989). The systemic coherence that supported the emergence and evolution of drip irrigation was also evidenced in (and deepened by) two episodes in the 1970s and 1980s. These propelled the technology, initially regarded as valuable only for use on marginal soils with marginal water (Pasternak, personal communication), into more mainstream uses. The first began in 1970 when the Israeli Water Commission—the government body mandated with assigning and enforcing water allocations and maintaining the National Water Carrier—initiated a World Bank funded project to improve national water use efficiency. This large-scale project allowed farmers to obtain grants and long-term loans for shifting to water-saving irrigation methods. Such grants helped farmers defray the inherent costs and risk involved in adopting a new technology. To receive their grants, farmers submitted irrigation plans to the Joint Regional and Central Judgment Committee, a body comprised of both Water Commission officials and agricultural extension workers. These bodies worked together to examine and make recommendations to farmers. The World Bank project brought about new, higher standards of irrigation system planning and implementation in the Israeli agricultural sector (Sne, 1989). In 1986, a second episode, this time an act of nature, further solidified the entrenchment of drip irrigation. A two-year drought led to the dwindling of aquifers and surface reservoirs, and the Water Commission declared emergency conditions and imposed a 15% cut in water allocations for irrigation in that crop year for the coastal strip of Israel. Here the coherence of the water system shone. The I&SFS took great care to try and minimize the potential damage of such an allocation reduction. Emergency irrigation recommendations were issued taking into account any factor that would reduce water consumption, including shifting to drip irrigation. Advanced water-use schedules were developed for individual farmers and updated monthly. Water-supply associations compared planned versus actual consumption in real time and made immediate adjustments. Inspection teams prevented violations of the allocation rules. The physical outcome was quite successful. Despite the substantial cut in water provisions, agricultural output declined only slightly (Sne, 1989). More importantly, the event seared drought-responsiveness deep into the Israeli national consciousness and led many farmers to voluntarily shift to drip irrigation for its greater water use efficiency. This unique conjuncture of circumstances in which drip irrigation was forged in Israel, and the complex infrastructures that evolved over many years to support and protect it, are, indeed,

18

Y. Garb, L. Friedlander / Agricultural Systems 128 (2014) 13–24

Table 1 Factors in overlapping spheres of influence, at various scales of the Israeli drip irrigation innovation system that supported the development and/or adoption of drip irrigation in Israel; this table provides a summary of our analysis of the Israeli case drip irrigation case study. Together, these factors make up the invisible ‘infrastructure’ of the Israeli drip irrigation innovation system that fails to be ‘translated’ with the drip irrigation hardware to other contexts. Scale

Environmental

Strategic alignment

Institutional support

Historic drivers

Farm

Settlements on marginal lands (i.e. Negev Desert)

Farms are communal and not individual. No individual ownership, lands leased from government, rotation of roles (including agricultural practitioner) among kibbutz members

Irrigation & Soil Field Service (I&SFS)

World Bank funds a national project within Israel to improve water use efficiency in agriculture, the money provides subsidies directly to farmers enabling them to switch to water efficient irrigation technologies (1970) A 15% cut in water allocations to the coastal strip after an extended drought in 1986 drove the Water Commission to develop individual water use plans for farmers in that region, which the I&SFS enforced in direct dialogue with farmers

Community

Self-contained communal rural settlements in a small country—isolated yet connected

Communal (socialist) agriculturebased settlements (Kibbutzim) Settlements dedicated to physical occupation of land, self-sufficient source of employment/resources

Agricultural Extension Service – provides new information and technology, responds to farmer concerns and experiences

Joint Regional and Central Judgment Committees are established at the community level using World Bank funds (1970) to evaluate farmers’ water use efficiency improvement plans, offer advice, and implementation support

National

Arid environment

Settlements establish national borders, protect, and employ immigrants Agriculture provides visible occupation of large tracts of marginal land Both techno-optimistic and rural ideologies are dominant

Department of Agriculture

Two-year drought (1984–6) forces the Water Commission to adopt an advanced, nationalscale water-use plan that results in an unavoidable 15% cut in water allocations for the Israeli coastal strip. The I&SFS and Agricultural Extension Service step into help farmers manage these cuts without severe loss of productivity (many are encouraged to adopt water-saving irrigation technologies)

Water resources concentrated in one region (North)

an ‘‘infrastructure’’ in the sense of Star (1999)—a taken-for-granted and partly invisible backdrop on which drip irrigation depends. In fact, the surprise of Israeli and other experts at the failure of such well functioning and simple hardware in other settings often corresponds to the degree to which this infrastructure is invisible to them. In some ways, the pipe and emitter, as ingenious as they might be, are just the ‘‘delivery point’’ of this broader and wellintegrated system extending into Israeli society. The technology’s apparent simplicity and viability are sustained by these massive and mostly submerged systems, absent in most other contexts. Or, to put it in a different theoretical terminology, Israel possessed—at least for agricultural and water-related innovations—a ‘‘national innovation system’’ able to exert the key functions needed to develop and diffuse this new technology. The presence—indeed, necessity—of this broader systemic envelope that enabled the emergence of Israeli drip irrigation technology was quite prominent in the awareness of the Israelis engaged in the early development and export of drip irrigation. For example, in an important conference on microirrigation convened by the World Bank in 1989, when Israeli experts were asked to comment on the Israeli experience of drip irrigation, the contextual shaping (and thus dependence) of drip irrigation in the Israeli context emerged as a leitmotif in their presentations. Drip irrigation, claimed one speaker. . . . did not appear suddenly as a full-fledged bright idea that immediately and automatically became a practical system. In fact, the idea itself was not new. What finally induced its evolution into a practical system was a combination of several synergistic factors. . . It was indeed a fortuitous confluence of conducive circumstances as well as of theoretical and technological developments that made drip irrigation an idea whose time had come (Ben-Meir, 1989). The authors described drip irrigation as a system, indeed a ‘‘set of systems (Ben-Meir, 1989),’’ emphasizing the fact that the Israeli drip irrigation system was developed in singular conditions

Water Commission

National Water Carrier National network of agricultural research institutions affiliated with the Agricultural Extension Service

(Table 1), so that the technology as it evolved in Israel was somewhat endemic to this setting. This led them to a certain caution regarding its transplantation to other settings—blind transfer would lead to failure, so that a ‘‘location specific and . . . holistic approach (Ben-Meir, 1989),’’ and sensitive adaptation would be critical to its successful ‘‘transfer’’ to other settings. [I]f Israeli experience and technologies are to be transferred to other countries, whether developed countries or areas where traditional gravity technologies still predominate, some knowledge of the Israeli background is needed. No agricultural or irrigation technology can be transferred without proper analysis of conditions in the source and target areas. This is especially so when adapting technology used in the Israeli context. A set of rather singular, almost unique, conditions in Israel favored the widespread and rapid adoption of sprinkler irrigation technology at the end of the 1940s and in the early 1950s and the continuous innovation, whereby new technologies were adopted almost as fast as they moved out of the experimental stage (Melamed, 1989). The following generation of irrigation professionals, however, who inherited these systems as taken-for-granted, was, perhaps, less sensitized to their rarity and precarious contingency in other contexts. But our analysis confirms the early insights of these speakers: the contextual sensitivity of the systems they spoke of can help explain the lack of ‘‘transfer’’ of drip irrigation to the African context, as described in the following section. 2.2. Drip Irrigation in the African Context 2.2.1. Pipes without a system If the Israeli case illustrates the deep embedding of drip technology in a site of early, deep and wide adoption, the fate of many drip irrigation systems in the African context offers a stark illustration of the opposite: what can happen when technology ‘‘transfer’’ is of a technical artifact stripped of this kind of broader systemic

Y. Garb, L. Friedlander / Agricultural Systems 128 (2014) 13–24

support, and with no local alternative into which it is grafted and reinvented. Drip technology may be particularly fragile in this respect, since its main advantage, water efficiency, is achieved through the precision and constancy of water delivery, which presupposes a broad socio-technical system to reliably maintain and regulate it. As one Israeli irrigation expert points out, ‘‘the fact that drip irrigation wets only a fraction of the soil volume can also be a problem. . . crops become extremely sensitive and vulnerable to even a slight disruption of the irrigation regime. There is thus very little latitude for error or malfunction. If the system does not operate perfectly and continuously, crop failure can result rather quickly, because the reservoir of soil moisture available to the plants is very small (Hillel, 1989).’’ Thus, this precision at the extreme tips of a delivery system is dependent on the robustness and substantial socio-technical resources of all that precedes it. Our field observations of both Israeli and American sponsored drip irrigation systems in two African countries, which are described in the remainder of this section, underscored these dynamics. Despite the aforementioned sensitivity of early Israeli experts to the technology’s contextual circumstances, and the aspirations and considerable efforts of commercial companies and NGOs, these systems were dogged by trouble and failure. We suggest that these should not be considered as a failed transfer of technology (in the sense of an artifact)—the pipes and filters were delivered—but as non-transfer of technology—that is, the overall socio-technical system, in which the concrete artifact figures as a necessary but dependent emblematic final link, is still absent. A leading drip technology supplier, the Israeli company, Netafim™ has three corporate offices in Africa (one in Egypt, two in South Africa) as well as distributors or regional offices in several countries, including Ethiopia and Zambia. In addition, multiple non-governmental organizations (NGO’s) actively promote various versions of drip irrigation technology in Africa. Yet they are fighting an uphill battle. Consider, for example, the T’esfa Hiwot Farm in the Oromiya region of Ethiopia, just south of the city of Nazret, visited in August 2009. The owners of the farm claim it to be one of the oldest commercial farms in Ethiopia. When they installed a buried drip irrigation system in 2005 they chose only the best technology, and hired a foreign farm manager to manage the irrigation scheme. We spoke with the current manager, another non-Ethiopian, who replaced the manager hired at the installation of the system. He showed us the progress being made at the site: in digging up this accurate and very sensitive, top-of-the line buried drip irrigation system. Without irrigation water of high quality (taken for granted by every farmer in Israel), the manager at T’sefa Hiwot farm made do with river water stored in an unlined, surface reservoir. The water could barely be pumped from the reservoir and was so thick with mud that the system’s fine-mesh, automated filtration components were quickly overwhelmed and stopped working. Thus, the farm’s drip irrigation system began failing (or, in our perspective, revealing its lack of connection to a system)—one section at a time—from the day of its installation. The 12 mm diameter drip lines, probably adopted for their increased irrigation efficiency, were now blamed for the blockages, a problem so severe as to render the system useless. The plan is now to gradually convert the farm back to furrow irrigation as each block of the buried drip irrigation fails. We could see that tubing that had already been unearthed was stiff from silt infiltration, and that many of the central joints had been severely damaged by warthogs. Such scenes, visible in many similar projects in Africa, are symptoms of the absence of a cohesive ‘‘system’’ around this newly introduced technology. The tight learning network linking farmers, agricultural extension workers and research institutions, which helped ensure the successful adoption of drip irrigation in Israel, was notably absent in the African environment. The links between farmers and the NGOs and commercial suppliers active in the irri-

19

gation sphere were haphazard, leading to all kinds of unwarranted assumptions and misunderstandings between various actors, as each of them conceived of the ‘‘technology’’ in a different way. The more material face of these networks, the taken-for-granted infrastructures on which the drip hardware is predicated, especially those relating to assuring the quality of water supply, were also absent. Without an integrated ‘‘socio-technical system’’ to animate drip ‘‘technology,’’ adapt it to the surroundings and make it cohere end-to-end, users were left with mere hardware—plastic tubes and drippers lying unused in the African sun. We found the views of actors on the ground regarding the reasons for such ‘‘failures’’ to be divergent and patchy, and similar findings are evident in the academic literature on drip irrigation uptake. There is broad agreement on the water supply problem, which leads to clogging, maintenance problems and eventual abandonment, but this breakdown appears differently to the stakeholders, who each see it from their own perspective. For example, in a study of smallholder irrigation by Kulecho and Weatherhead (2006a,b), farmers, representatives of the irrigation industry, Non-Governmental Organizations (NGO’s) and government officials each differently identified the problems with drip systems. Farmers most frequently reported problems with the maintenance of their systems, followed by water supply problems and problems marketing their produce. Industry representatives believed that the limited size of the market for small-scale irrigation systems prevented growth in the industry and uptake of smallholder drip irrigation technology. NGO representatives reported trouble with marketing and in spreading awareness to farmers about the existence and availability of the drip irrigation kits they offered. Government representatives claimed they were ignorant of any policies promoting smallholder irrigation. In our own fieldwork, we found a key discrepancy to be that between the perspective of farmers, who tend to consider the ‘‘technology’’ to be simply the hardware, and distributors, who had a broader view of the system surrounding this hardware. The farmer sees only the most visible manifestation of drip irrigation technology, the tubing and drippers; the linkages that could or should surround the system are invisible to him. In purchasing the very best ‘‘technology’’ for his farm, he focuses on this aspect, not thinking about connecting his farm to a local market to better sell his produce, or how to reliably get the best possible water to his farm. To manufacturers and suppliers of drip irrigation, on the other hand, it is obvious that an advanced drip irrigation system needs to be connected to an advanced water system. For them, the system includes pumps, filters and chemicals that feed the lines and drippers. In their eyes, the farmer who wants to purchase only tubing and drippers is buying half a system, a choice construed as misguided thrift or corner cutting. The seller knows before the deal is closed that the partial system he is about to sell will fail, but demand for drip irrigation is low and a sale is a sale. And in a few seasons, when the drippers are clogged and the drip irrigation system useless, the farmer blames the salesperson. While the business sector was frustrated with farmers who ‘‘refuse to pay for the technology they need,’’ NGO workers we spoke to saw the obstacle not as price, but, overwhelmingly, as an issue of farmer ‘‘capacity.’’ In their estimation, some farmers were simply more willing than others to invest the time in maintaining their systems. While it is certainly possible that some farmers have a greater capacity for managing a new technology or are willing to work harder than others or are more or less constrained in the extent of technology they can afford, we suggest that these gaps in perception and expectation are often the manifestations of a more fundamental issue: the farmers are simply not seeing the whole ‘‘technology,’’ and there are almost no links or mechanisms for the dialog and learning to create this aligned and supportive context for the technology, or even a sense of what it could or should consist of.

20

Y. Garb, L. Friedlander / Agricultural Systems 128 (2014) 13–24

The facilitation of these links and learning process is precisely the role of agricultural extension, at least as reconceived by Leeuwis (2004), as a process of ‘‘communication for rural innovation.’’ But even less progressively conceived forms of agricultural extension support were absent or insufficient. These kinds of difficulties with respect to drip irrigation we described by Kulecho and Weatherhead (2005), who found that maintenance trouble and dissatisfaction with the level of support were common reasons cited for discontinued use. Many of the farmers they surveyed reported that support agents could not help with repairs or did not know how to deal with blockages and other common problems. Another common set of problems they found, which can also be seen as a failure to establish the necessary broader support system required for the smooth operation of drip irrigation, with its requirement for sustained precise flows, was the lack of availability of spare parts and long wait times for replacements. When they pressed support agents on the reasons for these shortcomings, the researchers learned that support agents lacked adequate transportation and many offices lacked sufficient staff. We observed similar patterns, even in situations that might appear, initially, to be examples of flourishing drip irrigation system uptake. For example, while visiting the regional office of an NGO that promotes drip irrigation in Ziway, Ethiopia, service agents expressed pride in their sales record: after 4 years in the area they could list 5,136 farmers as ‘‘clients’’—a substantial step towards their ultimate goal of 8000 farmers. Closer examination, however, showed that only 464 of these farmers (less than 10%) had actually purchased their technology; the rest of these ‘‘clients’’ were simply farmers who had sought marketing assistance from the NGO staff. (NGOs and businesses have different business models and target functions, leading to these somewhat different definitions of ‘‘client.’’) And, even these purchased systems had a troubled fate. In the area of Ziway, a town of about 36,000 people, this NGO had established only 22 technology installation projects in the surrounding countryside, and our visits to some of these showed that they could not readily be considered as examples of successful uptake. Four farmers had lost their drip irrigation systems due to rat infestations as of August 2009, but field agents did not know how to advise farmers on preventing this. Several other farmers could not answer simple survey questions regarding ‘‘irrigation frequency’’ as they were using their drip irrigation system as ‘‘supplemental irrigation.’’ This is a relatively common practice, which occurs when a farmer, not trusting his drip irrigation system and not having been trained otherwise, relies primarily on traditional hand watering, with the drip irrigation system ‘‘keeping the soil wet’’ between watering. Thus, a system intended to replace hand watering and thus reduce labor inputs has become one more thing to care for on the farm, reducing its appeal and uptake. Finally, though this has not been addressed in great detail in the literature, we found the absence of research—a function central to the success of drip irrigation in the Israeli context—to be, perhaps, the most important weakness in the African drip irrigation environment. We observed that none of the agents in any of the visited NGO offices were actively conducting any sort of research—even in the simplest sense of principled evidence-based inquiry—to try and address the problems encountered by farmers in the region. Most viewed their role as maintaining a modicum of direct contact with farmers (‘‘customers’’) in order to encourage continued farmer ‘‘effort.’’ In addition, NGO’s tended to be organized hierarchically with a main office in the capital city and satellite offices in various regions within which the NGO operated. These satellite offices are run almost entirely separately from the central office. Technology, parts and information are disseminated outward from the main office in the capital, with little reverse or lateral flow. Thus, even if extension workers in one regional office were to solve a local problem through agricultural research there are no mechanisms for this innovation or

the allied technical modifications needed to be transmitted back to the main office nor out to other local offices and farmers. In short, our experience revealed hardware transplanted into an absent, fragmented or, at best, linear and deficient chain of communication, innovation and support, a situation quite different from the tight vertical alignment that allowed Israeli drip irrigation technology to become so deeply rooted in its country of origin. With none of the surrounding socio-technical system, farmers are without a voice in the process of technological adoption, purchasing brittle hardware stripped of the knowledge or means to develop or link this into a locally responsive technological system supple and ‘‘fluid’’ enough to flourish over time (de Laet and Mol, 2000). The following section describes two strategies through which development practitioners have attempted to make system-dependent technologies immune to contexts in which such systems are absent, while our concluding section suggests an approach that involves cultivating these necessary networks, rather than compensating for their absence. 3. When there is no innovation system: making drip irrigation installations self-contained The previous sections gave some theoretical context and examples from the case of drip irrigation for understanding certain common instances of failed ‘‘technology transfer’’ as occurring when hardware is relocated stripped of the broader socio-technical system needed for it to be useful. In this section, we describe two prominent drip irrigation research and development programs, which we interpret as representing two different strategies for overcoming the dependence of transplanted drip irrigation technology on this broader system. One approach suggested by Woltering et al. (2011) attempts to reduce stark dependencies on in-country systemic infrastructure, on the one hand, and bundle the physical system with a self-contained minimal support system on the other. The second approach promoted by Polak et al. (1997), and better known by the organization that promotes it, International Development Enterprises (IDE), tries to derive a version of the hardware simple enough to be viable without (almost) any support infrastructure, and, thus, more able to be relocated to settings where there is none. 3.1. The African Market Garden: bundling drip irrigation with a minimal self-support system Woltering et al. (2011) described four models of what they termed the African Market Garden (AMG), a ‘‘holistic horticultural production system for small producers.’’ The design of the system is similar to other low-pressure systems (Or and Rimon, 1991; Polak et al., 1997; Postel, 2001; Woltering et al., 2011). Such systems simplify conventional drip irrigation systems, which, according to Perry (1997) are complex and expensive mostly because they require water pressures of about 2 bars and are designed to meet the needs of field sizes of at least 2–3 ha. In contrast, smallholder farmers generally cultivate areas of less than 0.5 ha and cannot usually afford to invest in pressurized water sources. The technological features of greatest importance to these latter farmers include, low investment and operation cost, rapid return on investment, simple operation and low maintenance (Woltering et al., 2011). The authors’ goal in designing the AMG was to cater to these needs, balancing affordability and simplicity with quality and equipment longevity. An early attempt to disseminate 800 units of early AMG technology throughout Niger in 2002–3 failed massively. Two years later it was found that most systems had been abandoned as a result of poor equipment maintenance and farmer disregard for recommendations on fertilizer, water and crop management.

Y. Garb, L. Friedlander / Agricultural Systems 128 (2014) 13–24

According to Pasternak, an Israeli practitioner who was a key developer of the AMG system, and coauthor of the paper discussed above, the lesson to be learned from this initial failure was that success in projects involving low-income producers required the control of as many inputs as possible. He applied this understanding in a revised AMG implementation, and our observations are based on site visits to these AMG installations in Senegal (February 2010) and discussions with Pasternak himself. The AMG is made up of four components: the drip irrigation system itself, the water reservoir, an operation and management package, and high-quality vegetable seeds for adapted vegetable varieties. Thus, as signaled by its more inclusive name, the AMG approach is, essentially, a ‘‘batteries included’’ delivery of the product—an attempt to create a tailored self-contained bundle incorporating all the systemic factors deemed most important for viable drip irrigation. At the same time, installations are only done where the context is suitable for the survival of this bundle, and the bundle is tailored in advance to the particular site. Thus, dependence on the context and the need for real-time adaptation are reduced by two avenues: variability of the contexts of uptake is reduced, mostly by exclusion of unsuitable sites, and, to a much lesser extent by the co-construction of better settings; and for these suitable sites, as much preparation and pre-loading of necessary inputs is done in advance as possible, leaving as little as possible that is vulnerable to the vagaries of a particular setting. A key factor in the careful screening of potential AMG sites is finding locations where producers have access to credit, markets and technical support (Woltering et al., 2011). Where the water supply is irregular or poor quality, the AMG will be designed to include subsidiary investment in water infrastructure, particularly important in the African context where irregular or problematic water supply has already been linked to many cases of drip irrigation abandonment (Friedlander et al., 2013; Kulecho and Weatherhead, 2006a, 2006b; 2005). These investments may mean anything from a simple reservoir (when there is an adequate supply of nearby water) to a motor pump and all necessary connections (Woltering et al., 2011), including solar-powered pumps where connection to an electricity source is difficult or impossible (Burney et al., 2010). In Israel, farmers rely on government-supplied water, an invisible infrastructure enabling them to simply install drip irrigation with the expectation that it will work. In Africa the AMG approach partially translates drip irrigation technology to the African context by recognizing and accounting for the absence of this infrastructure, and ensuring that drip irrigation systems will only be installed in places where they can reliably be fed by alternative channels including motorized pumps, humanpowered pumps, rain-filled reservoirs and hand-filled reservoirs. In contradiction to the mismatched expectations regarding the cost and scope of systems, described above, the installation costs of the AMG system includes all of these ‘additional’ technologies necessary for it to function, and participants are explicitly expected to pay back these initial costs. The package also includes high quality vegetable seed varieties and drip hardware specifically designed to prevent maintenance difficulties, and the need for replacement parts. This higher reliability installation obviates the need to establish a network of locally provided maintenance workers or parts suppliers. The water reservoir is itself simply a tank sized individually to the average irrigation requirements of each AMG user. Thus, every detail of an AMG ‘‘kit’’ is extensively researched and designed before installation of the system; no two projects are alike and each responds to specific local conditions. This learning and adaptation occurs mostly at the hub, and by providers, rather than at the periphery by users. The latter are provided with 5 days of training in recommended management practices, including irrigation guidelines and fertility and pest management.

21

If many drip systems, including the early AMG attempts in Niger, failed because the hardware was abandoned in a landscape with no systemic support, Pasternak’s ingenious strategy could be regarded as sending the hardware out as part of an individually tailored and equipped self-contained package, which remains connected to a central hub through an umbilical cord of expertise and assistance. This linkage is through a single, highly effective circulating field technician brokering information about problems and advice back and forth between installations and the main office. Farmers receive this technical support every six months for at least two years from the technician, Aliyun Doufe, who travels extensively throughout the Western Sahelian region of Africa visiting AMG locations. Doufe works very closely with Pasternak who is also a researcher at the International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) and his team, constantly reporting back about his site visits. The traffic of information and guidance along this lifeline (reports of difficulties and success in one direction; site pre-selection and system tailoring, with ongoing advice in the other) enables a kind of a lean, targeted micro version of an ‘‘innovation system,’’ or, perhaps more accurately, an innovation or learning hub. That is, the system is not built around the capacities of a local innovation system, which might, over time, reshape and be reshaped by the technology, which could evolve divergently over time. Rather, after the system is ‘‘parachuted’’ in from the outside it relies on the self-contained adequacy of the resources it is already packaged with, and the activities and knowledge of a small close network of highly trained, external agricultural experts. This model echoes, at a much smaller and commercial scale, the national level government extension service that allowed Israeli drip irrigation technology to evolve and thrive, though more linear and with considerable less trust of the farmers at the capillary end of the system. Pasternak is explicit about the fact that while the drip irrigation hardware forms the core of the package (which he named TIPA—Techno-agricultural Innovations for Poverty Alleviation), the project is really constituted by the non-hardware components, the training, organization, and agricultural management support that the AMG package provides. The AMG approach seems highly successful in the sense that AMG adopters and TIPA participants express much greater satisfaction with their systems than drip irrigation adopters in other African contexts, that many TIPA participants have continuously used drip irrigation for 3 years or more, and that some participants in every TIPA project, usually around 4–7% or 4/60, with the means to adopt drip irrigation on their personal plots, eventually do (Friedlander, 2009; Friedlander et al., 2013). There are plans to expand the AMG network and establish regional offices to allow the Gardens to spread beyond current locations (Burney and Naylor, 2012). However, questions can be raised about the long-term sustainability of the TIPA approach. While current projects are ‘successful’ by several adoption metrics including length of use and continuous use (Friedlander et al., 2013), their success thus far has depended significantly on the input and decisions of one person (Dov Pasternak) and especially on the quality care that participants are afforded from their highly capable technical assistant, Aliyun Doufe—again only one person. There are plans to scale up TIPA through training additional technical support staff, but Doufe is a very singular character whose capacities will be difficult to recreate or outsource. In addition, as TIPA scales, it will be increasingly difficult to conduct the careful screening of potential recipient sites now so important for the relatively high adoption rates. In addition, the TIPA/AMG approach might reasonably be viewed as an extended attempt to secure the drip irrigation supply chain in Senegal and the Western Sahel in a company-specific manner. On the one hand, what allows the system its vigorous and systemic coherence is the fact that a single commercial company operates it. At the same

22

Y. Garb, L. Friedlander / Agricultural Systems 128 (2014) 13–24

time, there have been difficulties reported when the instincts of suppliers to ‘defend’ their technologies suppressed a more open and translational approach, which eventually proved more successful at encouraging adoption (Hall et al., 2007). 3.2. The IDE Strategy: independence through simplification Another response to the dilemma of the critical role of support systems in technology transfer can be found in the systems developed by Polak et al. (1997). The driver behind their low-cost drip irrigation system is the presumption that initial investment costs are the largest barriers to drip irrigation adoption in poor regions. Thus, the argument goes, if drip technology can be pared down to its barest essentials and sold cheaply, it will be successfully adopted. Considered from a more systemic perspective, this can be seen as a way of reducing the technology’s dependence on infrastructure through simplification. Polak’s design eliminates the technologically advanced components of the system and thus the infrastructure required to support these. If the AMG approach to placing hardware into new contexts is to screen the destinations and connect them to learning hubs, the IDE strategy is to reduce the technology to an idiot-proof core. Here we present the initial results from trials with drip irrigation technology pared down by Polak’s IDE approach. These early studies were conducted in Nepal and India, but our experience in Ziway, Ethiopia showed that IDE was attempting to transplant its original approach to sub-Saharan Africa nearly unchanged. Thus, the experiences described in the following section are still relevant to our Africa analysis although many refer to results from trials in India, Nepal and other countries on the Asian subcontinent. Initial results of trials with this kind of low-cost drip irrigation system were found to be very encouraging (Polak et al., 1997). The systems were distributed to ten farmers over two months in the hills west of Kathmandu, Nepal. Here, too, early recipients of the technology were selected for a predisposition to drip irrigation adoption: farmers farming small plots of land, with previous experience in vegetable farming, and access to a small source of water. Initial assessment showed that these farmers increased their irrigated area by as much as four times and reported a halving of irrigation workload, and all farmers were able to adapt to use of the technology. In a later more extended assessment, Polak and Yoder (2006) reported that over 100,000 IDE low-cost drip irrigation kits had been purchased in India, Nepal, Sri Lanka and Zambia. This, they claimed, was proof of the system’s success. However, a 2003 external assessment by Intermediate Technology Consultants (ITC) suggested that the outcomes were more ambiguous. ITC researchers found inconsistent adoption trends, which appeared highly susceptible to the local environment, both at the regional and at the farm scale. The differences between eastern and western India were instructive. In projects throughout eastern India (Jarkhand and West Bengal), the IDE goal of market-based distribution reached only middle and upper income farmers. Lower income farmers were reached only by cooperation with NGOs. According to the ITC report, there was no conclusive evidence of a self-sustaining low-cost drip irrigation market. Additionally, researchers reported that the adoption of drip irrigation seemed to depend on several preconditions: a feeling of water scarcity and limited water resources, well-established cultivation of vegetables and other horticultural crops suited to drip irrigation, initial and continued agronomic guidance and the support of local NGO’s. The findings on drip irrigation uptake in western India (Maharashtra State) stand in marked contrast to those from eastern India. In western India, independent, continued drip irrigation adoption is a reality. Poorer farmers work together to purchase more advanced drip irrigation systems requiring larger initial investments. The researchers immediately observed that several key

factors played an important part in priming Maharashtra for drip irrigation adoption. Firstly, the region is both drought-prone and poor in natural resources. Farmers perceive horticulture and the cultivation of intensive, high-labor crops as the only viable option for successful agriculture. Poor farmers in the region are aware of drip irrigation and other modern technologies in part as a result of the government micro-irrigation subsidies provided in the region, though these are generally aimed at high-income farmers. Farmers in the region perceive drip as the only viable irrigation option, given frequent water shortages. Given the choice between low-cost, small drip irrigation models and more expensive, more advanced drip systems, farmers in western India prefer to combine their funds and purchase the more advanced drip models rather than cheaper, smaller systems. This may be due to the higher profile that the drip irrigation industry enjoys in Maharashtra as result of the government subsidies program. This overall preexisting awareness of drip in the region provides a platform for the IDE approach, as farmers are already familiar with the technology and parts and assistance are readily available locally. Looking at these apparently disparate experiences through the innovation systems perspective, suggests, ironically, that where successful, adoption of the streamlined, low-cost IDE technology by lower income farmers was due not to a technology sufficiently pared down so as to be viable without local systemic support, but to the presence and nature of this local support! Where it succeeded, adoption by lower income farmers relied on the support of NGOs (as in eastern India), or else contact with the support network surrounding government-subsidized micro-irrigation aimed at higher income farmers or communally pooled resources for the purchase of sophisticated systems (as in western India). Success was not the result of simplification of the technology, but of the technology being able to find support with existing socio-systems outside of the boundaries of the project. Later studies support the finding that external factors have a large impact on drip irrigation adoption. For example, Namara et al. (2007) found that the five factors that most strongly affected the probability of drip irrigation adoption were years of schooling of the head of household, access to groundwater, cropping pattern, additional sources of income and social and poverty status. To put it differently, the confluence of all of these factors together can be considered to constitute a preexisting innovation system into which drip irrigation can readily fit. Our observations of IDE installations in regions of Africa in which these predisposing factors are lacking (Section 2.2) support this interpretation.

4. Conclusion: toward strategies for the translation and reinvention of drip irrigation technologies This paper synthesizes various literatures converging on a systemic, distributed, and relational perspective on technological innovation and change, and drawing on these, suggests that the metaphor of ‘‘technological translation’’ offers a useful shorthand for the kind of reinvention or ‘‘re-innovation’’ necessary for the process more traditionally referred to as ‘‘technology transfer.’’ We hope this perspective will be of value for policy makers and agricultural technology providers who would like to meet the needs of poor smallholders. We believe that such an approach avoids simply shipping hardware to contexts devoid of support, or efforts to reduce contextual dependence through umbilical life support systems or the crafting of hardware so simple that it is dependence-free. Instead, a technology translation approach would encourage shaping technologies that are fluid in the sense sketched in de Laet and Mol (2000)’s remarkable essay on the Zimbabwe Bush Pump. As these authors suggest, ‘‘. . . in traveling to intractable places, an object that isn’t too rigorously bounded, that

Y. Garb, L. Friedlander / Agricultural Systems 128 (2014) 13–24

doesn’t impose itself but tries to serve, that is adaptable, flexible and responsive - in short, a fluid object - may well prove to be stronger than one which is firm.’’ At least three strategies for facilitating translation emerge from this perspective, and these have been formulated in various terminologies in the literatures we have reviewed: 1. Setting technical and institutional fluidity as the goals of design efforts. 2. Encouragement of promoter fluency in the language of local needs and attentiveness to what these offer in the dialog of re-invention. 3. Support for local innovation systems along with lateral and vertical communication and integration to robustly weave existing technology into new systemic contexts. Agricultural development in SSA presents unique and difficult challenges for the promotion of new technologies; institutions are weak, infrastructure is irregularly distributed and often poor quality, farmers frequently lack access to markets and basic economic assurances (credit, insurance). Advanced irrigation technology cannot be promoted as a poverty alleviation and food security enhancement tool in this context by a ‘‘technology transfer’’ approach that relocates hardware, while ignoring the local circumstances that are so different from the ones in which the hardware evolved and operates in such an exemplary way. The attempts we have described to bypass this disjunction by simplifying or remoteoperating the hardware also seem to have their limitations. They create brittle artifacts, or ones that depend on remote hubs for their survival. They also forfeit the valuable and sometimes simpler innovations that can arise from those who know local circumstances intimately, people like the farmer we interviewed in the peri-urban surroundings of Ziway, Ethiopia, who claimed that he never experienced problems with dripline blockage; he repairs lines damaged by blockage by threading thin blades of grass through his blocked drippers. Unfortunately, this farmer has no way to communicate his innovation to the manufacturer or distributor of his system, and little opportunity to share his technique with other farmers. Promoting the lateral and vertical communication necessary to enable this farmer to share his technique would not only improve the experiences of other users in the Ziway, Ethiopia region, but also weave low-cost drip irrigation into a network of farmer experience and innovation-sharing. Technology promoters might be able to obtain more supple and long lasting support for the uptake of drip technology if they are prepared to have ‘‘their’’ hardware evolve (and to evolve with it), and to devote resources to the creation of ingenious and adaptable artifacts rather than idiot-proofed or fortified ones, and invest in the cultivation of and contextual dialog with the local innovation systems that would weave these into local lives and economies. This broader perspective adds additional complexity and effort to the promotion of improved irrigation technology, but in other ways can reduce the complexity and costs demanded by maintaining a static technology in a hostile setting. It may also demand less familiar kinds of expertise, innovation, and business models—more akin, perhaps, to the dispersed creativity of open source software than the develop and transfer models of proprietary software (Raymond, 1999). On a practical level, our conclusions suggest that agricultural development organizations, and drip irrigation promoters in particular, might begin addressing some of the shortfalls in their local innovation systems by investing in intermediary groups or organizations that can facilitate learning and change around hardware; establish networks to enable farmers using innovations (cleaning drippers with blades of grass as a concrete and also metaphorical

23

example) to connect and spread their innovations. Innovation platforms are an emerging organizational structure thought to encourage multi-stakeholder initiatives (Kilelu et al., 2013). The innovation platforms approach is in direct opposition to the ‘umbilical’ network of TIPA. It is closer to the IDE approach, but might encourage farmers to form collectives and purchase larger, more expensive systems. Thus, in the future, drip irrigation promoters should shift resources toward capacity building and away from providing more hardware. Unfortunately, such a shift would require dramatic changes in both policy and incentives for aid organizations. Instead of providing hardware and building more projects, aid organizations should be supported as learning organizations, and, instead of driving development, research should come in response to user experience. Such an institutional overhaul would be extremely difficult, but is likely to produce the kinds of technology that can have really lasting impacts on food security and poverty alleviation. To summarize, we suggest that the success of Israeli drip technologies in their originating settings and their failure in the SSA context are two sides of the same coin: well functioning artifacts are immersively dependent on broader socio-technical contexts and innovation systems in particular. Viewed in this perspective, promotion of the successful uptake of drip irrigation in new contexts would consist of building the policy, organizational, and communication contexts supportive of translation and re-invention, rather than improving a product to be copy-pasted or scaled up in new contexts. Fluid translation is a process that allows hardware to be continually remade to better suit the local contexts, inputs and needs of each user, especially the most vulnerable. Acknowledgements This paper grew out of ongoing joint analysis by the authors of Friedlander’s observations on the broader contexts of drip irrigation uptake made during two trips to sub-Saharan Africa to collect quantitative data on drip irrigation use for her M.Sc. thesis work. (This quantitative analysis appears in Friedlander et al. (2013)). The authors wish to gratefully acknowledge the financial support of Netafim™ Ltd. for this fieldwork in Africa, the Blaustein Institutes for Desert Research for the use of their facilities and additional financial support, and two anonymous reviewers for their constructive suggestions. References Andersson, J.A., D’Souza, S., 2014. From Adoption Claims to Understanding Farmers and Contexts: A Literature Review of Conservation Agriculture (CA) Adoption among Smallholder Farmers in Southern Africa. Agric. Ecosyst. Environ 187, 116–132. Belder, P., Rohrbach, D., Twomlow, S., Senzanje, A., 2007. Can drip irrigation improve the livelihoods of smallholders? Lessons learned from Zimbabwe (No. 33), International Crops Research Institute for the Semi-Arid Tropics, Global Theme on Agroecosystems Report. Bulawayo, Zimbabwe. Ben-Meir, M., 1989. Establishing research priorities. In: Le Moigne, G., Barghouti, S., Plusquellec, H. (Eds.), Technological and Institutional Innovation in Irrigation. The World Bank, Washington, DC, pp. 104–107. Bijker, W.T., Hughes, T., Pinch, T., 1987. The Social Construction of Technological Systems: New Directions in the Sociology and History of Technology. MIT Press, Cambridge, MA. Blaikie, P., 1985. The Political Economy of Soil Erosion in Developing Countries, 1st ed. Longman Group Ltd., London, New York. Blaut, J.M., 1987. Diffusionism: a uniformitarian critique. Ann. Assoc. Am. Geogr. 77, 30–47. Brown, S.D., 2002. Michel serres: science, translation and the logic of the parasite. Theory Cult. Soc. 19, 1–27. Bucks, D.A., Nakayama, F.S., Warrick, A.W., 1982. Principles, practices and potentialities of trickle (drip) irrigation. In: Hillel, D. (Ed.), Advances in Irrigation. Academic Press, New York, NY, pp. 219–298. Burney, J., Woltering, L., Burke, M., Naylor, R., Pasternak, D., 2010. Solar-powered drip irrigation enhances food security in the Sudano-Sahel. Proc. Natl. Acad. Sci. USA 107, 1848–1853.

24

Y. Garb, L. Friedlander / Agricultural Systems 128 (2014) 13–24

Burney, J.A., Naylor, R.L., 2012. Smallholder irrigation as a poverty alleviation tool in sub-Saharan Africa. World Dev. 40, 110–123. Callon, M., 1991. Techno-economic networks and irreversibility. In: Law, J. (Ed.), A Sociology of Monsters? Essays on Power, Technology and Domination, Sociological Review Monograph. Routledge, London, pp. 132–161. Carlsson, B., 2006. Internationalization of innovation systems: a survey of the literature. Res. Policy 35, 56–67. Carlsson, B., Stankiewicz, R., 1991. On the nature, function and composition of technological systems. J. Evol. Econ. 1, 93–118. De Laet, M., Mol, A., 2000. The Zimbabwe bush pump: mechanics of a fluid technology. Soc. Stud. Sci. 30, 225–263. Food and Agriculture Organization of the United Nations, n.d. FAO statistics . Friedlander, L., 2009. Field notes, unpublished. Friedlander, L., Tal, A., Lazarovitch, N., 2013. Technical considerations affecting adoption of drip irrigation in sub-Saharan Africa. Agric. Water Manag. 126, 125– 132. Geels, F.W., 2004. From sectoral systems of innovation to socio-technical systems: insights about dynamics and change from sociology and institutional theory. Res. Policy 33, 897–920. Goldberg, D., Gornat, B., Rimon, D., 1976. Drip Irrigation: Principles, Design and Agricultural Practices. Drip Irrigation Scientific Publications, Kfar Shmaryahu, Israel. Hall, A., Clark, N., Naik, G., 2007. Institutional change and innovation capacity: contrasting experiences of promoting small-scale irrigation technology in South Asia. Int. J. Technol. Manage. Sustain. Dev. 6, 77–101. Hall, A., Rasheed Sulaiman, V., Clark, N., Yoganand, B., 2003. From measuring impact to learning institutional lessons: an innovation systems perspective on improving the management of international agricultural research. Agric. Syst. 78, 213–241. Hillel, D., 1989. Adaptation of modern irrigation methods to research priorities of developing countries. In: Le Moigne, G., Barghouti, S., Plusquellec, H. (Eds.), Technological and Institutional Innovation in Irrigation. The World Bank, Washington, DC, pp. 88–93. Horton, D., Mackay, R., 2003. Using evaluation to enhance institutional learning and change: recent experiences with agricultural research and development. Agric. Syst. 78, 127–142. Hounkonnou, D., Kossou, D., Kuyper, T.W., Leeuwis, C., Nederlof, E.S., Röling, N., Sakyi-Dawson, O., Traoré, M., van Huis, A., 2012. An innovation systems approach to institutional change: smallholder development in West Africa. Agric. Syst. 108, 74–83. Ibragimov, N., Evett, S.R., Esanbekov, Y., Kamilov, B.S., Mirzaev, L., Lamers, J.P., 2007. Water use efficiency of irrigated cotton in Uzbekistan under drip and furrow irrigation. Agric. Water Mange. 90, 112–120. Intermediate Technology Consultants (project head: Heather Mackay), 2003. Low cost micro-irrigation technologies for the poor. Final Report. Intermediate Technology Consultants. Rugby, U.K. International Commission on Irrigation and Drainage, n.d. ICID . Jakku, E., Thorburn, P.J., 2010. A conceptual framework for guiding the participatory development of agricultural decision support systems. Agric. Syst. 103, 675– 682. Jansen, K., Vellema, S., 2011. What is technography? NJAS – Wageningen J. Life Sci. 57, 169–177. Kang, Y., Wang, F., Liu, H., Yuan, B., 2004. Potato evapotranspiration and yield under different drip irrigation regimes. Irrig. Sci. 23, 133–143. Kilelu, C.W., Klerkx, L., Leeuwis, C., 2013. Unravelling the role of innovation platforms in supporting co-evolution of innovation: contributions and tensions in a smallholder dairy development programme. Agric. Syst. 118, 65–77. Klerkx, L., Aarts, N., Leeuwis, C., 2010. Adaptive management in agricultural innovation systems: the interactions between innovation networks and their environment. Agric. Syst. 103, 390–400. Klerkx, L., Leeuwis, C., 2009. Establishment and embedding of innovation brokers at different innovation system levels: insights from the Dutch agricultural sector. Technol. Forecast. Soc. Change 76, 849–860. Klerkx, L., Proctor, A., 2013. Beyond fragmentation and disconnect: networks for knowledge exchange in the English land management advisory system. Land Policy 30, 13–24. Klerkx, L., van Bommel, S., Bos, B., Holster, H., Zwartkruis, J.V., Aarts, N., 2012. Design process outputs as boundary objects in agricultural innovation projects: functions and limitations. Agric. Syst. 113, 39–49. Kulecho, I., Weatherhead, E., 2006a. Adoption and experience of low-cost drip irrigation in Kenya. Irrig. Drain. 55, 435–444. Kulecho, I., Weatherhead, K., 2006b. Issues of irrigation of horticultural crops by smallholder farmers in Kenya. Irrig. Drain. Syst. 20, 259–266. Kulecho, I.K., Weatherhead, E.K., 2005. Reasons for smallholder farmers discontinuing with low-cost micro-irrigation: a case study from Kenya. Irrig. Drain. Syst. 19, 179–188. Latour, B., 1991. Technology is society made durable. In: Law, J. (Ed.), A Sociology of Monsters? Essays on Power, Technology and Domination, Sociological Review Monograph. Routledge, London, pp. 103–131. Law, J., 1999. After ANT: topology, naming and complexity. In: Law, J., Hassard, J. (Eds.), Actor Network Theory and After. Blackwell and the Sociological Review, Oxford and Keele, pp. 1–14.

Leeuwis, C., 2004. Communication for Rural Innovation: Rethinking Agricultural Extension, third ed. Blackwell Science Ltd., Oxford, UK. Leeuwis, C., 2013. Coupled performance and change in the making. Inaugural lecture upon taking up the post of Professor of Knowledge, Technology and Innovation at Wageningen University. Text published by Wageningen University, June 6, 2013. Melamed, D., 1989. Technological development: the Israeli experience. In: Le Moigne, G., Barghouti, S., Plusquellec, H. (Eds.), Technological and Institutional Innovation in Irrigation. The World Bank, Washington, DC, pp. 23–36. Millar, J., Connell, J., 2009. Strategies for scaling out impacts from agricultural systems change: the case of forages and livestock production in Laos. Agric. Hum. Values 27, 213–225. Morriss, S., Massey, C., Flett, R., Alpass, F., Sligo, F., 2006. Mediating technological learning in agricultural innovation systems. Agric. Syst. 89, 26–46. Namara, R., Nagar, R., Upadhyay, B., 2007. Economics, adoption determinants, and impacts of micro-irrigation technologies: empirical results from India. Irrigation Science 25, 283–297. Or, U., Rimon, D., 1991. Advanced technologies in traditional agriculture. J. Sust. Agric. 2, 1–10. Ortiz, O., Orrego, R., Pradel, W., Gildemacher, P., Castillo, R., Otiniano, R., Gabriel, J., Vallejo, J., Torres, O., Woldegiorgis, G., Damene, B., Kakuhenzire, R., Kasahija, I., Kahiu, I., 2013. Insights into potato innovation systems in Bolivia, Ethiopia, Peru and Uganda. Agric. Syst. 114, 73–83. Pacey, A., 1983. The Culture of Technology, first ed. MIT Press, Cambridge, MA. Perry, E., 1997. Low-cost irrigation technologies for food security in sub-Saharan Africa, in: Irrigation Technology Transfer in Support of Food Security, Proceedings of a Subregional Workshop. FAO Water Reports, Harare, Zimbabwe. Polak, P., Nanes, B., Adhikari, D., 1997. A low cost drip irrigation system for small farmers in developing countries. J. Am. Water Resour. Assoc. 33, 119–124. Polak, P., Sivanappan, R.K., 1998. The potential contribution of low cost drip irrigation to the improvement of irrigation productivity in India. In: Proceedings of the National Seminar on Micro Irrigation and Sprinkler Irrigation Systems. Delhi, India. Polak, P., Yoder, R., 2006. Creating wealth from groundwater for dollar-a-day farmers: where the silent revolution and the four revolutions to end rural poverty meet. Hydrogeol. J. 14, 424–432. Postel, S., 2001. Growing more food with less water. Sci. Am. 284, 46–49. Postel, S., Polak, P., Gonzales, F., Keller, J., 2001. Drip irrigation for small farmers: a new initiative to alleviate hunger and poverty. Water Int. 26, 3–13. Raymond, Eric, 1999. The Cathedral and the Bazaar. O’Reilly Media. Rice, R.E., Rogers, E.M., 1980. Re-invention in the innovation process. Knowledge 1, 499–514. Rogers, E., 1978. Re-invention during the innovation process. In: Radnor, M., Feller, I., Rogers, E.M. (Eds.), The Diffusion of Innovations: An Assessment. Northwestern University, Center for the Inter-Disciplinary Study of Science and Technology, Evanstion, IL. Rogers, E.M., 1962. Diffusion of Innovations, first ed. The Free Press, A division of Simon & Schuster Inc., New York, NY. Shah, T., Keller, J., 2002. Micro-irrigation and the poor: a marketing challenge in smallholder irrigation development. In: Sally, H., Abernathy, C.L. (Eds.), Private Irrigation in Sub-Saharan Africa: Regional Seminar on Private Sector Participation and Irrigation Expansion in Sub-Saharan Africa. International Water Management Institute, Accra, Ghana, pp. 165–183. Shoji, K., 1977. Drip Irrigation. Sci. Am. 237, 62–68. Sne, M., 1989. The role of extension in irrigation: the Israeli experience. In: Le Moigne, G., Barghouti, S., Plusquellec, H. (Eds.), Technological and Institutional Innovation in Irrigation. The World Bank, Washington, DC, pp. 128–136. Soete, L., 2014. Research and Innovation for Sustainable Developmentin. In: Bolay, J., Hostettler, S., Hazboun, E. (Eds.), Technologies for Sustainable Development, pp. 239–244, Springer. Spielman, D.J., 2005. Innovation systems perspectives on developing-country agriculture: a critical review. In: International Service for National Agricultural Research (ISNAR) Division. International Food Policy Research Institute (IFPRI), Washington, DC. Spielman, D.J., Ekboir, J., Davis, K., 2009. The art and science of innovation systems inquiry: applications to Sub-Saharan African agriculture. Technol. Soc. 31, 399– 405. Spielman, D.J., Ekboir, J., Davis, K., Ochieng, C.M.O., 2008. An innovation systems perspective on strengthening agricultural education and training in subSaharan Africa. Agric. Syst. 98, 1–9. Star, S.L., 1999. The ethnography of infrastructure. Am. Behav. Sci. 43, 377–391. Temel, T., Janssen, W., Karimov, F., 2003. Systems analysis by graph theoretical techniques: assessment of the agricultural innovation system of Azerbaijan. Agric. Syst. 77, 91–116. Verma, S., 2004. Promoting Micro Irrigation in India: A Review of Evidence and Recent Developments, IWMI Tata Annual Partner’s Meeting, IWMI-Tata Water Policy Program Annual Partners’ Meeting. Mumbai, India. Woltering, L., Pasternak, D., Ndjeunga, J., 2011. The African market garden: the development of a low-pressure drip irrigation system for smallholders in the Sudano Sahel. Irrig. Drain. 60, 613–621. World Water Assessment Programme, 2009. The United Nations World Water Development Report 3: Water in a Changing World. UNESCO, Earthscan, Paris and London.