Reversals, Institutions and Change 105manoeuvre. Some steps can be taken; a start can almost always be made. Even if
the start is small and progress slow, it may be the seed of a self-sustaining move-
ment. In the spirit of the learning process approach to development, it is better to
start, to do something and to learn on the way, than to wait for better conditions
before acting.
In the spirit of pluralism, action can and should start in many places. But not
everything can be done at once. There are questions of how and where to start.
Two principles help here. The first is to start where it is easier, simpler and
quicker, while weighing the danger of biases against poorer farmers. It is better to
start and learn by doing and through mistakes than to wait for perfect conditions.
By starting, experience is gained and confidence built up.
The second principle is to change behaviour before attitudes. Preaching about
attitudes invites acquiescence without deep change. Action means experience
gained and that, more than exhortation, reorients attitudes and habits of
thought.
Taking these two principles together, analysis by and with farmers appears the
most promising point of entry, followed by search, choice and experiment. A basic
question to ask is what farmers would like in their basket of choices. From this
question follow demands which reverse the normal top-down flow. Whether a
department of agriculture, a university, an NGO or combinations of these can
handle such requests can then be put to the test. Activities and roles then have to
change. Procedures to accept and handle demands are required. Information sys-
tems for management from below have to be created and made to work. Subse-
quently, other elements of the paradigm become active, with testing and experiments
by farmers and consultative support by others.
It is one thing to start and establish a bridgehead. It is quite another to sustain
and spread it. The experiences reported by the International Service for National
Agricultural Research’s (ISNAR) On-Farm Client-Oriented Research project in
nine national institutes are sobering. They include difficulty maintaining an inter-
disciplinary focus, vulnerability to the withdrawal of special support, a tendency to
methodological stagnation, a loss of early enthusiasm and of farmer participation,
and a career ladder which leads away from collaboration on-farm and towards
specialization on-station (Merrill-Sands, 1988; von der Osten et al, 1988). With
farmer-first, similar problems can be expected but also differences. The approach
and methods described by the contributors to this book go further than most on-
farm research, exploiting as they do the comparative advantage of farmers’ knowl-
edge, continuity and capacity for innovation. When it is farmers, with their full
experience of their own farming systems, who analyse, experiment, monitor and
make judgments, it is less important to sustain an interdisciplinary focus; farmers’
enthusiasm and participation are more likely; and if outside support weakens,
farmers can carry on on their own, and make their own demands on the research
system, strengthened by their personal interest and participation. Compared with
on-farm research in the TOT mode, the farmer-first mode promises to be feasible
with a lighter touch and sustainable with less outside support.