A 'Middle-Class' Elite
Perhaps the defining characteristic of the Bangladesh elite is that it has not, until recently, been particularly 'elite':
The word 'elite' begins to lose its meaning when it is applied to the middle-income
professional and business petty influentials who, under Sheikh Mujibur Rahman,
carried the message of the Awami League to the rural people of East Bengal.
(Nicholas 1973, 145)
Elites are groups among whom power or influence over decisions is concentrated; national elites tend to be identifiably wealthier, culturally superior, socially and sometimes racially or ethnically distinct from the masses.1 By contrast, the Bangladeshi elite has closely resembled the masses and been comparatively close to them, socially and spatially. The contemporary elite is mostly from Bengali-speaking Sunni Muslim stock, without important inherited or cultural status. Until independence from Pakistan, few of this group held great economic or political power. Compared to groups which have ruled the region in the past, and to contemporary elites elsewhere, the Bangladeshi national elite has been marked less by its distance and difference from the masses than by its affinity to them. 
This relative affinity has influenced the conception of the national project of development (Hossain 2005). Many Dhaka elites have recent peasant origins and associated ties. The nationalist political leadership and other elites are chiefly from recent peasant stock, families that entered the professions or had formal education in the past generation or so (see Siddiqui etal. 1990; Barua 1978; Sen 1986; also Sobhan and Sen 1988; Rahman 1991). The new nation-state of Bangladesh was an 'intermediate' regime, dominated by middle-class professionals, small businessmen, and wealthy peasants, rather than landed or large industrial interests (Sobhan and Ahmad 1980; Bertocci 1982).
Until recently, that group was close-knit, and multiply connected by ties of kinship and location. That connectedness comes with intertwined personal histories: 'batchmate' university friendships created some enduring ties that cut across sectors and ideologies within the present generation. A respected NGO leader once described to me a youthful friendship with a business leader whose own public profile was, shall we say, more ethically ambiguous. Past friendship trumped present differences, however, and the unlikely pair were seen side by side on pro-democracy platforms, demonstrating the importance of unity on issues of national importance. Such unlikely connections were once surprisingly common.
-  The classic account remains Mills (1956). Despite contextual differences, the sample framesused to study 'elites' are often similar (Verba etal. 1987). Most political science focuses on nationalelites, based on positions within institutional structures and networks.
-  The key text remains Ahmed's The Bengal Muslims (1996). It is not that there were noimportant social distinctions: Bengali Muslim society had its elite-mass distinction in the formof the ashraf ('foreign' aristocratic families claiming Central Asian extraction from way back inhistory) and ajlaf or atrap (local peasant and artisan converts) (Eaton 2001);it also had its ownversions of castes, with well-defined occupational groups and rules. But while hereditary in theory,these were never entirely rigid. Economic change and religious fundamentalism meant these weredeclining by the late nineteenth century (Ahmed 1996). They have little current resonance (alsoVan Schendel 1982).