knowledge (Beal et al., 1986). The organization of knowledge processes in agricultural development has been subsumed in powerful mental models of how science, knowledge and technology "get agriculture moving" (Mosher, 1966; Borlaug and Dowswell, 1995). Each of the main models (Albrecht et al., 1989, 1990) has its own logic and fitness for purpose. They and their variants are discussed and compared; in each case for the sake of clarity they are first presented as commonly accepted abstractions followed by assessment of the dynamic ways in which the model has been applied within specific institutional arrangements in particular contexts. Institutional arrangements are important to the assessment because they provide different ways of distributing power and influence among sources of knowledge and hence are consequential for understanding the kinds of impact that can be expected and were in fact realized.

2.1.2.1 Transfer of Technology as a model for organizing knowledge and diffusion processes

One model in particular has dominated as a guide to the organization of knowledge processes in the public sector in developing countries, the Transfer of Technology (ToT) model. It was formally elaborated as a practical model for guiding action and investment in specific AKST arrangements on the basis of empirical studies of knowledge management and diffusion processes in the midwest of America (Lionberger, 1960; Havelock, 1969). Science is positioned in this model as a privileged problem-defining and knowledge generating activity carried out mainly by universities and research stations whose knowledge, embedded in technologies, messages, and practices is transferred by extension agents to farmers. The model assumes a linear flow of technological products and information. Each of the entities described in the model is treated more or less as a "black box." Although in practice much local level interaction takes place between extension agents, farmers and research specialists, the underlying assumption of the model is that farmers are relatively passive cognitive agents whose own knowledge is to be replaced and improved as a result of receiving messages and technologies designed by others and communicated to them by experts (Röling, 1988; Compton, 1989; Eastman and Grieshop, 1989; Lionberger and Gwin, 1991; Blackburn, 1994; Röling and Wagemakers, 1998).

     The model mirrored the prevailing AKST organizational arrangements of states gaining their independence in the 1950s and 60s. Many explicitly favored centrally-planned economic development and most relied heavily on state organizations as the catalyst of agricultural development and commodity marketing (Hunter, 1969, 1970; Dayal et al., 1976). Extension field staff were positioned on the lowest rung in a hierarchy of relationships under the direction of departments of agriculture and publicly funded research stations and universities (Maunder, 1972; Peterson et al., 1989). Social, educational and political biases reinforced the idea that lack of access to "modern knowledge" was a constraint to production (Mook, 1974; Morss, et al., 1976). District development plans and projects to develop cooperatives, farmer service societies and the like received considerable attention (Halse, 1966; Lele, 1975; Hunter et al., 1976).

     The ToT model assumes that wide impact is achieved on the basis of autonomous diffusion processes; this indeed can be so (Rogers, 1962). The classic study of diffusion of innovations was published in 1943 based on the rapid autonomous spread of hybrid maize among farmers in Iowa (Ryan and Gross, 1943). The diffusion of innovations became a popular subject for empirical social science research, generating well over 2000 studies and much was learned that was helpful concerning the conditions in which rapid and widespread diffusion can occur, what helps and hinders such processes and the limitations of diffusion for achieving impact. Diffusion research has continued even after the late Everett Rogers (well-known for his classic decadal overviews of research on the diffusion of innovations) (Rogers, 1962, 1983, 1995, 2003) himself spoke of the "passing of a dominant paradigm" (Rogers, 1976). The role of autonomous diffusion among farmers persists as one of the pillars of the common understanding of the pathways of science impact. The history of the rapid spread in Africa of exotic crops such as cassava, maize, beans and cocoa is added testimony to the power of diffusion processes to change the face of agriculture even without the kinds of scientific involvement of more recent years.

     The positive impact of the ToT model. The ToT model gained credibility from the rapid and widespread adoption of the first products of the Green Revolution (GR) emerging from basic and strategic research (Jones and Rolls, 1982; Evenson, 1986; Jones, 1986; Evenson and Gollin, 2003a). For example, in the poor, populous, irrigated areas of Asia the GR allowed Bangladesh to move in 25 years from a net importer of rice to self sufficiency while its population grew from 53 million to 115 million (Gill, 1995) and India, Indonesia, Vietnam, and Pakistan to avert major famine and keep pace with population growth (Repetto, 1994). In China, wheat imports dropped from 7.2 tonnes in 1994 to 1.9 tonnes in 1997 and by 1997 net rice exports had risen to 1.1 tonnes. The Green Revolution not only increased the supply of locally available staples but also the demand for farm labor, increasing wage rates and thus the work-based income of the "dollar-poor" (Lipton, 2005). National food security in food staples in the high population areas of developing countries throughout the world was achieved except in sub-Saharan Africa. The diet of many households changed as more milk, fish and meat became available (Fan et al., 1998). Investment in industrialized food processing and in agricultural engineering, often stimulated by heavy government subsidies, in turn began to transform subsistence farming into a business enterprise and created new employment opportunities in postharvest operations i.e., storage, milling, marketing and transportation (Sharma and Poleman, 1993). The ToT model clearly proved fit for the overall purposes of disseminating improved seed, training farmers in simple practices and input use and disseminating simple messages within the intensive, high external input production systems characterizing the relatively homogeneous irrigated wheat and rice environments of South and Southeast Asia. Positive impacts were recorded also in parts of sub-Saharan Africa (Moris, 1981, 1989; Carr, 1989).