The Alliance for a New Green Revolution in Africa and African government and CGIAR programmes oriented toward improving cassava production through intensification and the use of external inputs have the ultimate goals to improve food production, promote market integration, and increase incomes of small farm households. Essentially, AGRA’s arguments, which are either implicit or explicit in the policies and programmes of the Government of Cameroon and of several CGIAR institutes that the Government collaborates with, are that traditional farming systems and practices suffer from low productivity and are unsustainable. African soils are naturally poor, farmers use little or no fertiliser, and the fallow periods that, in the past, provided for nutrient recycling, are declining due to population pressure, leading farmers to mine the soil, which results in declining crop yields. Further, farmers’ local varieties are low yielding and are highly susceptible to pests and diseases compared to improved, high-yielding varieties (HYVs). Across Africa, per capita food production is declining, and families live in poverty and hunger. Population pressure is increasing, farmers are poor and thus in need of additional income and, if given the opportunity, they will seek to maximise their income from crops sales, which they in turn will reinvest in agriculture, given the right incentives. Farm households are food insecure and, by increasing their output and sales, they will become food secure. This dissertation challenges these underlying assumptions and questions the underlying parameters individually and as a whole by examining traditional and more commercial smallholder cassava agroecological systems and households in two study sites in rural Cameroon (where conditions are theoretically quite positive for the acceptance of such technologies) from agroecological, ethnobiological, economic, and cultural perspectives. The objective is to understand the implications of policies and programmes that promote Green Revolution-type technologies and market integration for the productivity and sustainability of such agroecological systems, for the conservation of crop genetic resources, and for the livelihoods, income, and food and nutritional security of smallholder farm households. The intention is to critically examine the assumptions and underlying parameters posited by AGRA, and to reformulate these on the basis of the findings to provide a more adequate framework for approaching and assessing agricultural innovations in the African context. The following questions orient the research: Are African farming systems, and farmers, characterised by attributes that AGRA ascribes to them? Are such farmers likely to accept the technologies that AGRA is promoting? Are AGRA technologies and strategies likely to lead to more sustainable, higher yielding farming systems? Are they likely to translate into greater market integration, higher incomes, greater food security, and renewed investment in agricultural intensification for small farm households? Are there trade-offs that farmers and their households and communities have to confront in adopting such technologies and, if so, how might these influence their strategies and responses to programmes that promote Green Revolution-type intensification of the ‘old’ or ‘new’ varieties? Findings presented in this dissertation show that Koudandeng and Malende farmers have barely accepted Green Revolution technologies and modern farming strategies and systems (including monoculture). The analysis of the findings proposes reasons for this, and attempts to explain farmers’ and households’ production systems and strategies from an emic (farmers’) perspective. It is argued that, if African farmers do not accept the Green Revolution-type technologies, or accept them only on their own terms and in accordance with the outcomes that they themselves desire that differ significantly from what governments and researchers and donors anticipate, then this may be attributable at least in part to the fact that the strategies and technologies that are promoted are based on erroneous assumptions, not least about the key parameters that define the performance of real African farming systems and real African farming households. These parameters are grouped under two main categories - agroecological and socio-economic – which, in AGRA’s discourse, are treated as if they were unrelated. There is thus an absence of attention to the relations between the agroecological (or what can be termed environmental, or ‘nature’), and the socioeconomic (or what can be termed ‘culture’), which in turn leads to an inattention to the diversity of cultures and agroecologies across Africa – its biocultural diversity – that permits blanket recommendations to be made on the basis of over-generalised and oversimplified assumptions. When emphasising the need to give greater consideration to the relations between culture and nature – that is, to the diversity of African cultures, agroecologies, and socioeconomic systems and relations, and to the relations between culture, agroecology, and socioeconomics - this dissertation proposes three different interacting sets of analytical parameters that must be considered if insights into real African agriculture and real African farm households are to emerge. Two of these sets of parameters emerge from a critique of AGRA’s parameters and a third arises out of a framework for assessing the acceptability of crop varieties that has its foundations in ethnobiology. This comparative research, which was carried out between 2002 and 2008, involved a total of 206 farmers in two different villages in two regions in the South of Cameroon. The methods for collecting and analysing data were both quantitative and qualitative, and were drawn from sociology, anthropology, and ethnobiology (cognitive anthropology). Qualitative data collection methods included a review of grey and published literature, as well as ethnographic interviewing and participant observation. Quantitative methods included four closed question surveys and cognitive ethnobiological elicitation (freelisting and triads testing). Qualitative interview data were coded and analysed narratively (description, explanation, interpretation, quotations) using Microsoft Word. The small household sample size that was used did not permit the use of sophisticated statistical analyses according to population sub-samples, which limited the analysis of survey data to that which would be done using descriptive statistics, such as proportions, percentages, and frequencies. Regression analysis was done sparingly. Cultural consensus analysis, proximities analysis, multidimensional scaling, quadratic assessment product, cluster analysis, and property fitting regression were used to analyse the ethnobiological data that was collected. The general conclusions of this dissertation assert that traditional African polyculture systems and their genetic diversity (crop species and varieties) are often environmentally sustainable, able to meet income and food needs of rural households and communities, and fulfil multiple cultural needs relating to identity, foodways, spirituality, and social reciprocity. The assumptions behind the promotion of AGRA-type technologies are reductionist; they do not take into consideration the complexities of African agriculture and livelihoods, or the interrelation between farmers’ social and cultural norms, resource access, and livelihood strategies, and how they carry out agriculture (e.g. spatial and temporal configurations, cropping patterns, crop and varietal choices, cultural practices). Across most of Africa, smallholders and their agroecosystems are firmly embedded in ethnic and tribal communities that adhere more or less strongly to cultural norms, beliefs, and kinship or lineage-based social relations. Their agricultural knowledge and practices are often based largely on local knowledge and resources. Such ‘traditional’ agricultural systems generally represent a long-term adaptation between culture and nature, where both have co-evolved over time. Farmers’ knowledge and practices are embedded in social relations where many modes of subsistence are characterised by forms of communalism that are relatively egalitarian, which tends to ensure that resources are distributed in such a way that people have sufficient means to meet socially defined, as well as biological needs. Unsustainable practices and inegalitarian social relations that may accompany the adoption of Green Revolution technologies and greater market integration are likely to be mal-adaptive over the long run. The assumptions underlying the ‘New Green Revolution for Africa’ drastically over-simplify traditional African farming systems and ignore their diversity and thus do not hold everywhere in Africa which, it is argued, may represent yet another threat to the integrity of traditional African cultures, agroecological systems, and biological diversity. Eight major critiques of this over-simplification and the resultant dangerous consequences for African farm households include: i) the inappropriateness (technical and practical limitations) of the recommendations for integrated soil fertility management practices and fertiliser use for most African contexts; ii) the lack of consideration for farm households’ social constraints: differential access to income, land, and labour, and investments in other livelihood activities that compete with investments in agricultural inputs, which consequently may have implications for soil fertility management; iii) the lack of attention to the pests and diseases of most significance to farmers; iv) the relative inattention to the need to develop varieties that conform with local foodways and food processing and storage conditions; v) the implications of mass production of the reduction of crop diversity and varietal diversity for food security and nutrition and the consequences for human health; vi) the lack of serious consideration of farmers’ knowledge and practices in crop breeding strategies and the lack of precise methodologies for effectively and systematically accessing and document farmers’ varietal knowledge, perceptions, and preferences and relating these to farmer behaviour when accepting crop varieties; vii) the improbability that prices for mass produced HYVs will increase income and investments in inputs; and viii) the consequences of conversion to monoculture for livelihood and food security that are entailed in widescale acceptance of AGRA-type recommendations. Based on these critiques, the major policy recommendation emphasised in this dissertation is to give greater consideration to real African farming systems and real African farmers and how and why they function as they do, which, it is argued, must serve as the point of departure for agricultural policies and programmes across the region if these are to succeed in supporting such farmers, their communities, and their nations. Farmers’ culture, social relations, knowledge, practices, and experiences that remain, in the ‘New’ Green Revolution, as in the ‘Old’, a black box, should be newly considered in policies and research and development as positive points of departure for increasing food security in Africa.