Plant breeding is a science of selection. This selection operates in two distinct dimensions: Plant breeders select plants, and society, by choosing to sanction and support specialists devoted to plant improvement, selects breeders. To understand plant improvement, therefore, we need to understand the way in which at any given historical moment these two kinds of choices are exercised.

Biological choice depends on researchers" understanding of the biological environment. A thumb­nail sketch of the biological history of plant improvement might include four distinct stages:

• the era of farmer­managed land races;
• the "hit­or­miss" phase of breeding (selection for yield and elimination of defects);
• the first designer phase in which breeders focused on modified plant architecture (crop ideotypes for monocropping in well­controlled farm environments);
• a second designer phase in which transgenic crops are "constructed" with no direct precedents in nature, but shaped to meet the needs of future economic environments (crop "econotypes").

The history of crop improvement presented this way is made to appear as a largely autonomous process (increased control over biological options). But this is misleading. Breeding is a social process; to take the most obvious sense in which this is true, there would be no professional breeders without societal support for the research stations, universities, and private businesses within which breeding takes place. What happens to the overall picture if we add the social history of plant breeding?
Social science makes a distinction between organizations and institutional cultures. Plant improvement takes place within organizations. These organizations are supported by institutional cultures. These institutional cultures can be both formal (e.g. courts and constitutions) or informal (e.g. a religious commitment to honesty and hard work). Organizations come and go, but change in institutional culture tends to take place more slowly.
The social theory of institutions is in its relative infancy, so many mysteries concerning institutional change persist. But there is good agreement that institutional cultures are the bedrock of human interaction, and that institutional cultures assume more and less efficient forms. More controversially, it has been argued that there is no natural tendency for better institutions to drive out the bad. This is because institutional culture is often a way of dealing with radical uncertainty (i.e. chronic lack of information about the true nature of the environment and what the future holds).

How have institutional cultures shaped the history of plant improvement? Two brief examples:
(1) In the 1780s the English merchant­reformer Henry Smeathman tried to introduce the sickle into rice harvesting on the Sherbro coast of Sierra Leone (West Africa). Farmers refused, claiming they would be accused of witchcraft. Farmers in Sherbro country persist to this day with panicle harvesting as a way of ensuring the separation of distinct classes of planting material, in order effectively to manage catenary soil sequences under conditions of labour shortage. In contrast, farmers in districts with the sickle now have to renew their planting material from central sources every few years, because types and off­types are reaped together.
Sherbro farmers in the 18th century had no such options. Witchcraft accusation was one of the few available sanctions against hasty and careless management of seeds. Perhaps needless to add, Smeathman saw only evidence of barbarity in the rejection of his innovation.
(2) During the 1980s the conservative government of Mrs Thatcher in Britain offered publicly­owned plant breeding research institutes for sale to the private sector. This was done in the belief that the market was the best (or only) institution to regulate the plant improvement agenda. In effect, a decision was taken that breeding choices should be shaped not by biological theory but by price discipline. In the event, the research facilities in question were acquired by agricultural business interests with a strong commitment to agricultural chemicals. No credence was given to the argument that this might discriminate against low­chemical options such as "durable resistance" breeding, nor were doubts entertained about the capacity of the market correctly to divine and discount long­term "sustainable" futures in British farming. This suggests that the decision to reshape British plant improvement research at a stroke owed little or nothing to biological rationality and almost everything to the political instincts of government ministers steeped in the institutional culture of finance capital.

Here are two examples that tell us about the way in which institutional cultures shape crop improvement options. But does anything change with biotechnology? It seems inconceivable that the influence of institutional culture on plant improvement will diminish. Biotechnology implies a further loosening of biological restraints on design and re­design of plants, so intensification of that influence is more likely.
Some international donor organizations support the establishment of biotechnology research organizations in developing countries, instead of using the transfer­of­technology approach. Although this is admirable in sentiment, it yet remains to be seen whether these organizations will become part of the local institutional landscape, and with what consequences for innovation. Will poor farmers, with an institutional culture of safety­first, take over and drive the research agenda, insisting (for example) on crop ideotypes designed not for monoculture but risk­spreading polyculture? Or will the biotechnology institutes serve, as Trojan Horses, to facilitate convergence on Western institutional norms? Perhaps the biotechnology institutes will be inconsistent with any local scheme of political or cultural values.
These are issues that remain to be addressed in a systematic way in the debate about biotechnology futures. It is time to recognise that the issue of institutional culture should be at the top of the biotechnology policy agenda.
Paul Richards

Paul Richards is Professor of Anthropology (University College London, UK) and Chair of the Group for Technology and Agrarian Development (Wageningen Agricultural University, The Netherlands)