This article is a conceptual review of a decade of work on Farmer Participation in Research and Extension (FPR/E). Though FPR/E does have an important role, it has generated over­optimistic expectations. This is caused by lack of clarity in the objectives of different kinds of participation; the way participatory approaches relate to other modes of client orientation; and the various roles of different organizations in promoting participation. A major unresolved issue is the need to complement depth of participation with breadth of coverage. Inter­agency collaboration may hold some, but not all of the solutions to this dilemma.

Stronger participation by farmers in agricultural research and extension is fuelled by the realisation that the socio­economic and agro­ecological conditions of (especially low­income) farmers are complex, diverse and risk­prone. Conventional approaches, based on research station trials followed by unidirectional technology transfer, are unlikely to be fruitful. Close engagement with farmers is needed throughout the cycle of diagnosis, experimentation and technology dissemination. This increases the understanding of the opportunities and constraints farmers face, and of their own technical knowledge. This in turn enhances the prospects that externally promoted technologies will be adoptable, and environmentally and institutionally sustainable. The approach may thus, enhance the efficiency of the technology development processes.

What is participation?
'Participation' is becoming a devalued term. For instance, senior administrators, partly in response to donor exhortation, deploy much of the rhetoric, and occasionally the form, of participation without the substance. For this paper, participation conveys that the intended clients of agricultural research and extension (R&E) influence the focus and content of R&E. Public sector, private commercial and private non­profit organizations involved in R&E serve a wide range of clients. Their clients are not only farmers. They also include processing industries; scientists; and government departments concerned, for example, with land rehabilitation. With all types of client, the crucial link  and the most difficult to make is between what 'science' has to offer and what clients require.
Clients participate with different technology suppliers in numerous ways. There is a subset of types of research (applied and adaptive), of 'suppliers' (mainly public sector and Non­Governmental Organizations (NGOs), and of clients (farmers). Our focus here is an adaptive/applied research for farmer clients. Within these sub­sets, several types of participation exist. For instance, farmers in the middle and higher income ranges, may participate in R&E through the market by contracting advisory services, or by buying inputs which incorporate new technologies. Additionally, they can exert pressure through lobby groups. They can also respond vocally to the technologies offered during, for example, research station visits. These farmers tend to specialize in a small number of market­oriented commodities, and tend to operate on an individualistic, not group, basis. For these situations, the functions of public sector research organizations are easily defined. If these public sector research organizations can identify the needs of client farmers, and if these organizations efficiently manage the research project cycle, it should not be difficult to deliver relevant and adoptable technologies.
However, the situation of low­income farmers is more complex. In biological and physical terms, it is characterized by:

• poor infrastructure;
• diverse and risk­prone agro­ecological conditions;
• strong interaction between crop, livestock, tree and fodder components of the farming system, and between on­ and off­farm resource management.

• a degree of political and economic marginalisation, implying limited access to markets;
• diverse socio­economic conditions: some households being fully committed to farming; off­farm employment being important for others;
• the importance of group action in some traditional practices (e.g. exchange labour) and for soil and water conservation through the management of common resources;
• a high proportion of female headed households and of female farm labour since men are often employed in off­farm activities;
• strong local knowledge underpinning traditional farming practices.

Different interpretations and implementations
The objectives of the public sector in pursuing participatory FPR/E are primarily functional. It aims to enhance the efficiency of research services in delivering adoptable technologies that are environmentally and institutionally sustainable. Within this functional context, group approaches have occasionally been used, such as in integrated pest management. However, the public sector largely work with farmers in an individualistic way.
By contrast, the objective of participation for most NGOs is the social, economic and political empowerment of the disadvantaged and marginalized. Almost universally, NGOs have used a range of group building techniques. This includes awareness creation, conflict resolution and the development of leadership skills. NGOs have pioneered the use of Participatory Rural Appraisal (PRA). In much of South Asia, NGOs have taken the lead in promoting group management of common resources. This includes trees and grazing land, with a focus in undulating areas on raising water tables to permit agricultural intensification. NGOs' strengths lie in group formation and diagnosis. However, their capacity for experimentation, wide scale replication of approaches, or adoption of technology remains limited. They particularly need support in sourcing new candidate technologies and management practices.
A further weakness is that many NGOs ignore the fact that farming makes only a marginal contribution to the livelihoods of many poor households. This misperception generates mistaken confidence among many NGOs. They think that sustainable technical change through participatory approaches, can make all farmers succeed.
NGOs' pioneering of PRA has led some to equate PRA with FPR/E. There remain important distinctions, however. PRA has been used almost exclusively at the diagnostic stage of the research cycle. PRA has powerfully demonstrated the ability of village households to contribute to rural development planning. It has generated a sense of community ownership of development projects and processes. It has further created a recognition among administrators that participation enhances the prospects of success. However, PRA is increasingly seen as a 'new orthodoxy'. Like all orthodoxy, it attracts diverse challenges. One has to do with intellectual property. Namely, the argument that a number of its methods predate the term 'PRA'. Another is that some PRA enthusiasts create bias by carelessness in phrasing questions and facilitating discussions. A third is that enthusiasm for methods has led many to ignore differences in objectives and in the comparative advantage of different kinds of organizations. Generally, NGOs can mandate themselves to spend considerable resources in a few villages. Hence, they can pursue costly, empowering face­to­face types of participation. Many see the public sector's efforts as deficient since these are not as fully empowering. This ignores the much wider mandate of government departments that requires them to spread resources more thinly. The concern of government is necessarily with functional rather than with empowering types of participation.
A fourth is the realization that diagnostic methods (e.g. PRA) are not implemented within social and political voids. The outcomes of PRA meetings are determined by the 'mix' of community groups. For instance, in some cultural settings, it is difficult to assess women's needs. Overall, there is a need for better understanding of the processes of institutional, political and economic change at local levels. Additionally, a more judicious selection and application of participatory methods is needed.

Policy implications
The public sector can validly limit itself to functional types of participation, as distinct from the empowering types undertaken by many NGOs. Several conditions have to be met before public sector researchers can implement participatory approaches effectively.
First, the institutes should be committed to produce results which are of use to the identified clients.
Second, the performance criteria, reward, and incentives must be provided in delivering technologies that meet clients' needs.
Third, scientists will need training in participatory methods. Providing that the potential shortcomings of PRA are made clear, training in PRA methods is a good first step. However, scientists must have the resources to pursue the research issues identified by PRA. They must be able to maintain participatory approaches throughout the research project, and so go beyond mere diagnosis.
Participatory approaches can only become fast and cost­effective, if they are part of the daily practice of researchers. It will not be adequate for researchers to rely on specialist participatory units. Some farming systems research fell into that trap. Lessons should be learned from this.
Different approaches to participation are likely to be needed according to the biophysical setting. Where the intention is to increase crop yields, approaches can be largely individualistic. However, in many undulating semi­arid areas, the scope for agricultural improvement will be limited unless water tables can first be raised. To increase percolation, soil and water conservation measures are needed. Almost invariably, such measures require joint action. However, the public sector rarely has group formation skills. Of course, training can be given. Nevertheless, a potentially better alternative is to collaborate with organization(s) (such as NGOs) which already have many of the requisite skills.
The major dilemma in FPR/E is that of combining breadth with depth. If it concerns technological change but not with empowerment, this is not difficult to address. If participatory approaches have succeeded in identifying technologies acceptable to farmers of known socio­economic and agro­ecological characteristics, then the same technology could be offered to other similar groups. In more empowering approaches it is problematic. As the experience of several NGOs suggests (see e.g. Fernandez, 1993), more than a year of intensive face­to­face interaction is needed with small groups of low­income farmers to diagnose and identify pathways for addressing their needs. Does this suggest that the only expansion path is to repeat the exact process elsewhere? If so, the spread of participatory approaches is likely to be slow and resource­demanding.
Can participatory approaches spread in a less resource intensive fashion? Experimentation with different approaches is essential. For example, 'lateral spread' may be achieved by cross visits between villages. The rural spread of media such as radio, television and video can reinforce group­based approaches. For instance, groups could record how they have become organised and how they have introduced technical change. Intensive, face­to­face participatory methods have become part of the raison­d'être of NGOs. Not surprisingly, they have tended to dismiss mass media approaches as 'top down'. In reality, these may usefully supplement face­to­face approaches. This is not to disregard other settings, wherein only face­to­face methods can create the necessary confidence and negotiating skills to redress the biases against low­income groups.
Different types of organization have different strengths and weaknesses.  NGOs' strengths in diagnosis and group formation could be complemented by the technical skills of public sector R&E services. Exploitation of these complementarities would make farmer participatory research more effective and wider spread. Equally, complementarities among government departments could be exploited. These deal with agriculture, horticulture, livestock, water resources and trees. A number of pilot efforts towards multi­agency and participatory approaches are being tried with some success. The process requires careful building of trust and monitoring of progress against expectations. Likewise, it is vulnerable to changes in personnel on each side. Despite the difficulties facing multi­agency approaches, this merits extensive donor and government support.

Conclusions
Much of the confusion, and in some cases, excessively high expectations, surrounding FPR/E can be resolved by examining it within its agro­ecological and socio­economic settings. It is important to examine the objectives of the organizations implementing FPR/E. The functional approaches of public sector researchers can be complemented through collaboration with those organizations (such as NGOs) taking a more empowering approach. However, collaboration among different types of organization, no matter how appealing, requires much patience in practice. To engage effectively in participatory approaches, implies several preconditions for the public sector. Scientists must have rewards and incentives to work towards close client­orientation. Authorities must be sufficiently decentralised for rapid and flexible decision making.
John Farrington

Research Fellow, Overseas Development Institute and Visiting Professor, Agricultural Extension and Rural Development Department, University of Reading ODI, Portland House, Stag Place, London SW1E 5DP. Fax (+44) 171 393 1699 E­mail j.farrington@odi.org.uk

Sources
R.G. Alsop, R. Khandelwal, E.H. Gilbert, and J. Farrington, (1996), "The Human Capital Dimension of Collaboration among Government, NGOs and Farm Families: Comparative advantage, complications, and observations from an Indian case". Agriculture and Human Values, 13 (2) pp. 3­12.

J. Farrington, and A. Bebbington, with D. Lewis (1993), Reluctant Partners? Non­Governmental Organizations, the State and Sustainable Agricultural Development. London: Overseas Development Institute/Routledge.

A.P. Fernandez, (1993), The Myrada Experience. The interventions of a voluntary agency in the emergence and growth of peoples' institutions for sustained and equitable management of micro watersheds. Bangalore: Myrada.