During the second afternoon of the Science Forum, we plunge into a session that ought to have been scheduled on the first morning – a scene-setting plenary that delves into the key themes that are supposed to be framing the conference: forging partnerships and mobilising linkages. It’s great pity that the four useful presentations have come so late in the agenda, particularly since they take place after the parallel workshops are already over.

Nevertheless, the four speakers provide us with useful insights into the institutional and organisational frameworks that will be needed if the CGIAR is to transcend the discredited ‘linear model’ of agricultural research and development. In that framework, scientific innovations were understood to occur in laboratories and on field stations, subsequently being pushed out into farmers’ fields – but the new technologies often failed to make the transition from the lab into practice.

Effective innovation systems are now understood to be more diffuse and open. The speakers provide us with some useful insights into the kinds of organisational changes and new mindsets that will be needed, especially with regard to the power of modern information and communication technologies (ICTs) to enable new kinds of collaborative working, knowledge sharing and innovation.

It proves to be an interesting session, but unfortunately there is hardly any time for questions and discussion from the conference floor – a problem that has plagued this conference throughout. After a short break for tea, we begin the final session of the Forum – a long and gruelling one in which we endure detailed blow-by-blow accounts of the discussions that unfolded in each of the parallel workshops.

What comes out most clearly from the workshop reports is the extent to which all of the working groups have been dominated by technical discussions about agricultural and ecological problems and speculations about potential technological solutions, rather than the linkages and mechanisms that may equip the CGIAR to address those challenges.

The technological possibilities are indeed remarkable and exciting. For instance, as technical capacity increases and the related costs fall dramatically, the genomic secrets of crop plants can be deciphered more quickly and cheaply than ever before. Once uncovered, the information can be exchanged and shared worldwide, instantaneously, through the latest generation of collaborative ICTs and computer software. These technological breakthroughs hold huge potential.

But what will we use this spectacular potential for? The CGIAR needs to make choices about what it will do and how. What should be its targets and strategies? Where should be its vision for the strategic direction in which to go? How will its priorities be set? As the conference ends, I find that I am still looking for signs of the overarching strategy that should guide the future evolution of the CG system. From conversations with other participants, I learn that I am not the only one.

Amid all of the technocratic discussions, the insights of social science – the kinds of insights discussed by Bill Clark yesterday morning or the speakers of the first session this afternoon – have faded into the background. Organisational, managerial and institutional considerations are frequently invoked, but by and large social science seems to be regarded as a minor appendage of science – a vestigial organ rather like the human appendix: everybody agrees that it should be present, but nobody really knows what it is for. And, if it grumbles too much, it can always be cut out - the organism of normal scientific practice will carry on happily without it.

This is surely a disappointing outcome for an event that should be the CGIAR’s flagship, agenda-setting conference. However, the last speaker of the afternoon presents the audience with cause for hope. Mark Holderness, Executive Secretary of the Global Forum on Agricultural Research (GFAR) introduces us to the GCARD process – the steps leading up to the Global Conference on Agricultural Research for Development, scheduled to take place in Montpellier, France in March 2010. The conference aims at nothing less than the development of a completely new architecture for international agricultural research, focused squarely on development outcomes.

Holderness does not mince his words: the CGIAR must change; agricultural research must be for development; we must no longer accept the failure of technology transfer and agricultural extension; they must work, and we must make them work. Coming as the final note of the Science Forum 2009, one has the indelible impression that the GCARD 2010 will be a much more significant event for the future of global agriculture and international development.