Power: Three Faces and Beyond

Political sociology begins with the concept of power, and the concept of power turns out to be considerably more contested than its everyday use suggests. The classical sociological definition comes from Max Weber: power (Macht) is the probability that an actor in a social relationship will be able to carry out their will despite resistance, regardless of the basis on which that probability rests (Weber 1922/1978, pp. 53-54). The definition is deliberately broad. It covers coercion, persuasion, authority, and economic dependence. Weber distinguishes power from domination (Herrschaft), which is power exercised through commands that are actually obeyed, and which is usually legitimated through one of three ideal types: traditional, charismatic, or rational-legal authority (Weber 1922/1978, pp. 212-216).

Weber's definition is the starting point but not the whole story. Twentieth-century political sociology developed the concept through a famous sequence of elaborations sometimes called the 'three faces of power.'

First face: power in decisions. Robert Dahl's Who Governs? (Dahl 1961) studied political power in New Haven, Connecticut, by examining who actually prevailed in observable decisions across three policy domains — urban renewal, public education, and political nominations. Dahl found that different coalitions dominated different issue-areas and concluded that New Haven was a pluralist polity in which power was dispersed rather than concentrated in a single elite. The first face of power, on this account, is what can be observed in the resolution of overt conflicts: who gets what they want when they disagree with others who want something else.

Second face: agenda-setting. Peter Bachrach and Morton Baratz (Bachrach and Baratz 1962, pp. 948-949) argued that Dahl's approach missed a crucial dimension. Power also operates by keeping certain issues off the agenda — by non-decision. If an issue is never raised in the formal political process because actors anticipate that raising it would fail, or because mobilization around it is suppressed, the observable conflict that Dahl relied on never occurs. The second face of power is the power to define what is and is not a political question, to structure what Schattschneider earlier called 'the mobilization of bias' (Schattschneider 1960).

Third face: shaping preferences. Steven Lukes (Lukes 1974/2005, pp. 25-29) extended the analysis further. Power, he argued, also operates by shaping what people want in the first place. If people do not perceive their interests accurately — because ideology, socialization, or the control of information shapes their preferences — then neither overt conflict nor suppressed conflict will reveal the operation of power. The third face raises a difficult methodological problem: how does an analyst determine what people would want absent the influence of power, and how does one observe a force whose effect is precisely to prevent observable opposition?

Beyond the three faces: Foucault and power as productive. Michel Foucault pushed the analysis in a different direction. Power, on his account, is not primarily a property held by some actors and wielded over others. It is a diffuse set of relations running through all social interactions, institutions, and forms of knowledge (Foucault 1977, pp. 26-28). Power does not merely repress; it produces — it produces subjects, disciplines, categories of identity, and the terms in which opposition itself is articulated. The Foucauldian analysis is not incompatible with the Weberian one but it reorients attention from sovereign actors (the state, the ruling class) to the capillary operations of power in schools, clinics, prisons, and statistical bureaus.

Why does sociology need all of these rather than selecting one? Because different political phenomena become visible at different registers. The passage of a specific bill is a first-face question. The long absence of universal health care from the US agenda is a second-face question. The persistence of mass consent to extreme inequality is a third-face question. The construction of 'the citizen,' 'the voter,' or 'the criminal' as social categories through which politics is conducted is a Foucauldian question. A political sociology that restricts itself to one register will miss most of what it is supposed to explain.