In mainstream policy and corporate thinking, scientific knowledge and global markets are considered key for food security. This has resulted in the industrialisation and laboratory research-led intensification of agricultural systems, inputs and food-supply chains.
But intensified systems do not meet global food needs — they mostly suit export markets and corporate interests. The result is severe physical, but also economic, disconnection between production and consumption, or need, as well as private control of the crucial knowledge base that shapes agriculture.
This is effectively an industrial monoculture model of production — of both food and knowledge — that avoids its ecological and social costs, while suppressing more effective sustainable alternatives, and underexploits science's potential versatility.
To generate more sustainable pathways to equitable and healthy food production and access, agricultural diversification is needed, with food-supply systems decentralised and a move towards more localised networks.
This includes the strategic reorientation of agricultural research and development towards varied local conditions and needs, and towards farmers' knowledge — a global science for the public good.
Science as ideological tool
Agricultural systems are under severe, interrelated pressures, including chronic soil deterioration; water scarcity and overuse; chemical pollution; climate change; farmers' exodus to cities; population growth; and the global financial crisis.
In response to this, policymakers, funders and some scientists promote 'sustainable intensification' — which implies that science has a well-defined understanding of the technologies needed to overcome such challenges.
But while this approach builds on innovations from intensive laboratory-based research, it neglects the extensive diversity of local conditions and knowledge systems. And by presenting 'technification' as an undeniable scientific truth, it leads to one particular vision of so-called agricultural modernisation.
Proponents of this approach ideologically deploy 'science' against other types of knowledge and visions of innovation, dismissing alternatives as antiscience and antimodernisation.
In June, for example, UK environment secretary Owen Paterson, speaking on expert advice, exclusively equated genetic modification (GM) technology with 'science'. In doing so, he ignored the well-demonstrated scientific and practical credentials of non-GM, ecological approaches to agriculture, effectively villifying them as anti-science.
But recent research shows that this type of modern agriculture does not perform any better in terms of yields and reduced pesticide burdens than non-GM, especially for commodity crops such as soybean and maize. 
Unsustainable industrial agriculture
The Vision 2020 programme of the Andhra Pradesh state government in India is another example of promoting agroindustrial models. Supported by the Indian and UK governments, together with Monsanto, the world's largest GM seed and agrochemical company, the programme's stated aim was to reduce the state's rural population by 30 per cent as part of efforts to 'modernise' agriculture by minimising labour and eradicating so-called inefficient peasant agriculture.
The more than 25 million people designated for removal from the land were not included in any social impact assessment of this plan.
A similar scenario was articulated in a 2012 World Economic Forum report, which projected that small-scale farming's contribution to global food production would fall from 40 per cent in 2010 to zero by 2030 — being replaced by industrial large-scale (likely monocrop) farming. 
The report offered no comment on this — yet if the small-farm agriculture estimated as providing 70 per cent of current local food supply is intended to disappear, then who will produce food for most small-scale markets and poor people around the world?
These are symptoms of science understood and practised predominantly as technoscience — ignoring local conditions where food is produced, distributed and consumed, and devaluing less mechanistic and less reductionist forms of knowledge, production, exchange and access to food.
This technoscientific commercial vision also shapes technologies and production primarily for global market demands, which does not reflect diverse real needs.
Agroecology is effective
The alternative agroecological vision for agriculture and food supply is not just theory. Local expertise and knowledge integrated with leading scientific research have demonstrated a capacity to increase yields sustainably, decrease polluting inputs, promote food security and improve livelihoods.
One analysis compared 293 farms globally and concluded that agroecological farms can produce 96 per cent of a conventional yield while improving soil fertility.  A separate study reported that, under severe climate conditions, two different agroecological systems of maize and soybean yielded 137 per cent and 196 per cent more than conventional systems, respectively. 
In India, small-scale farmers are increasing productivity by up to 43 per cent using the System of Rice Intensification, which is based on low external input methods akin to agroecological ones. 
And in 2008, the UN reported cases where organic family farming doubled yields in Kenya, raised income by up to six times more than comparable conventional agriculture in Malawi and improved nutritional levels by 70 per cent in Ethiopia. 
These are only some indicative examples — there are many more.
Towards sustainable agriculture
How to strengthen these sustainable approaches? The key is a critical analysis of how science is understood, channelled and practised, and how it links with industrialised agri-systems and global markets. Socially and environmentally responsible food production and distribution requires global science policy to be reorientated towards locally adapted farmer knowledge, and essential local and regional networks.
As a first step, researchers, funders and policymakers need to identify and then re-evaluate assumptions about science, innovation, and food supply and access.
Second, they need to ask the right questions about sustainable global food security, which is more than a short-term production issue. Instead of asking almost exclusively how to increase short-term yields, we should ask why even existing global production is denied to many needy people, while generating mass obesity and enormous amounts of food waste elsewhere.
Third, they need to support small-scale and low external-input farming — both in terms of food production and research — providing tangible support for grassroots innovation. This will help integrate local practical knowledge systems with innovative scientific research and diverse food networks.
It is already known that scientific research can be integrated with grounded and farmer-orientated innovation systems. This approach could not only strengthen food security but also promote responsive food-supply networks more connected to real needs. And it could help initiate an authentic, sustainable science for society.
Brian Wynne is professor of science studies at the Centre for the Study of Environmental Change, Lancaster University, United Kingdom. Georgina Catacora-Vargas is a researcher at the SEED (Society, Ecology and Ethic Department) of GenØk — Centre for Biosafety, Norway. Wynne can be contacted at firstname.lastname@example.org
References International Journal of Agricultural Sustainability doi: 10.1080/14735903.2013.806408 (2013)
 World Economic Forum Putting the New Vision for Agriculture into Action: A Transformation Is Happening (World Economic Forum, 2012)
 Renewable Agriculture and Food Systems doi: 10.1017/S1742170507001640 (2007)
 American Journal of Alternative Agriculture doi: 10.1079/AJAA200345 (2003)
 World Bank Institute System of Rice Intensification (World Bank, video accessed 29 August 2013)
 UN Conference on Trade and Development and the UN Environment Programme Organic Agriculture and Food Security in Africa (UN, 2008)