I am a sociologist, working since 1983 in African countries on food security as well as on participatory rural development. So, a big thanks for this public debate and the really interesting views exposed. Let me bring some “critical” points.
Concerning inequalities within communities, for instance in Mozambique, “the largest part of the variation in per capita farm sizes and poverty levels in rural Mozambique is found within villages rather than between them” (Analysis of Adult Mortality within Rural Households in Mozambique and Implications for Policy; Research Paper No.58E, Ministry of Agriculture and Rural Development; Directorate of Economics; 2004).
Inequalities taking place within rural communities lead the most vulnerable to starvation through complex practices of access to labour, land and food stocks; generally assets and endowments. This is about social and power relations within the traditional system / community.
Social relations have much to do with labour, enter alia. Labour access is a major determinant of structural hunger in rural communities in Sub-Saharan Africa. Labour inequalities in rural communities lead the two lowest quintiles to trans-generational poverty and actual hunger. Labour control is much related to land control even when land apparently is “accessible” to all members of the community, even the foreigners. Much of the traditional institutions aim at regulating and controlling man-power / labour.
To give an example from Mozambique. In the Macua society, Northern Mozambique, a coping strategy through traditional social relations, called “o’lola”, takes place. This practice of “exchanging labour for food” enables considerable accumulation to those who benefit from the work of others. Suffice it to say that one day's work under "o'lola" can be paid with 3-4 kg. of cassava or sometimes with just a plate of beans, while the average flour production per workday corresponds to 7-9 Kg for cassava. Under these circumstances, a process of land concentration on the hands of few within the rural communities is taking place. For instance, in 1993, in Nampula Province, about 40-50% of the total land was held by only 25% of the subsistence producers that farmed between 4 and 5 times more land per household than the smallest 25%. The land accumulation has to be understood not in terms of property rights on land (in Mozambique the State is the only owner) but in terms of farming capacity, i.e. the capacity of a farmer to have access to labour during the peak season.
I disagree with any analysis that does not seriously rely on inter-disciplinarity and that takes a naïve approach to the traditional societies’ internal dynamics, picturing out a kind of Christmas-card image of an ideal word of perfectly functioning solidarity. This is risky when policies have to be designed and strategies implemented. Let’s take the case of the Social Safety Nets / SSN and schemes of social protection promoted in rural areas. Turn a blind-eye to social differentiation and internal inequalities, assuming, e.g. that traditional leaders are immune from any risk of power abuse, can be fatal to the most in need.