The contingencies of power: reformulating Foucault

Abstract

In a critique of Michel Foucault’s understanding of power relations, Andrew Sayer outlined an alternative that was both more conceptually rigorous and offered a clearer analytical framework with which to approach the analysis of power. Part of Sayer’s development was to distinguish not only between the holding and the operating of power(s), but also the three contingencies that always exist alongside power(s). The paper argues that Sayer’s critique of Foucault, while a welcome theoretical clarification, partially mistakes Foucault’s project (which was to analyse power ‘as is’, not power ‘in the abstract’) and also misses some of Foucault’s own clarifications to his understanding of power. The paper develops Sayer’s position by integrating important elements of Foucault’s approach – namely his concept of ‘dispositif’ – that illustrate Foucault’s more nuanced understanding of power. This is developed by arguing that there is, in fact, a fourth level of contingency that Sayer does not articulate. The paper uses De-facing Power, an early work of Clarissa Hayward, to illustrate how a combined ‘Foucaultian/Sayerian’ model of power and contingency can give extra explanatory and analytical strength. The outcome of the article’s argument is that contingency ceases to be a ‘problem’ in social scientific research; instead it becomes the raison d’être for social scientific explanation.

Keywords:

• Andrew Sayer
• contingency
• critical realism
• dispositif
• Michel Foucault
• power

Acknowledgements

The author wishes to thank Frank Pearce and Mira Bachvarova for many discussions on the topics of Foucault and power. All mistakes and errors, of course, remain the author’s alone. Early versions of sections of this paper were presented at the Society for Socialist Studies (Canada) annual conference, and at the Atlantic Provinces Political Studies Association (Canada) annual conference, both in 2014.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. For arguments expanding a similar position re causation, see Williams (2010) and Hardy (2014).

2. It may, at this point, be argued that Foucault does account for the non-absolute effects of power via his argument that power always creates resistance (Foucault 1980, p. 95, Sayer 2012, p. 187). However, it is worth noting that ‘resistance’ does not necessarily mean whole, partial or even minimal levels of effectivity.

3. The closest Foucault comes to articulating a ‘pure’ conception of power is to term it ‘an action upon an action, on possible [future] or actual … present actions’ (2003c, p. 137) or a ‘relation of forces’ (1980, p. 196, 1990, p. 92, 2003a p. 357).

4. And Sayer recognises this (2012, p. 180), stating that his account only focuses on certain of Foucault's key texts and was not meant to be an exhaustive engagement.

5. It is worth noting here that this is even before the advent of repetitive Fordist production techniques.

6. For no powerful group rules for eternity and power must, at some point, cease to operate.

7. It might be argued that since the power₂ exercise is so costly, then it does not even exist as a power₁ potential. I am not persuaded by this as the power₁ potential exists, even if at ruinous cost. To use a somewhat macabre example, I have the power₁ capacity for ending my life, even if the power₂ operation of this ability subsequently ends my existence and all the power₁ potentials that I have contained in me, including the ability to end my life.

8. For a detailed discussion of dispositifs, see Hardy (2015).

9. This does not have to be, and is not meant here as, a tautology. There will likely be external forces required to bring and arrange the component parts together, but once arranged and linked together the entity will likely become, at least partially, self-reinforcing. An example might be of the frame of a building, where the interrelationship of forces – gravity, weight, as well as any necessary joining materials, etc. – keeps the components together.

10. This raises an interesting (but non-refuting) point about ‘faked’ vs. ‘held’ power. For example, an agent may pose as a government official in order to deceive another agent to give them money, e.g. payment for a non-existent fine, etc. The fraudster does not officially have access to the governmental power relations themselves yet manages to obtain money from another party: so does that make the fraudster powerful? It does not, because the fraudster is only ever an approximation of another power relation. Power is not held by the individual fraudster, but the fraudster is a convincing mimic of power that is contained in another, already existing, government/agent relation.

11. Schmidt herself does not use the concept of dispositif. Her work is being used here as one possible means of demarcating the discursive ‘levels’ in a dispositif. Foucault outlined a difference between savoir (roughly ‘background principles and axioms’) and connaissance (roughly ‘experiential/applied knowledge’) (Foucault, 1972), but arguably this does not adequately give enough gradation to the types of discourses utilised in a dispositif. Ian Hacking makes a compelling case for the how the Foucaultian power/knowledge relationship ‘works’ in practice – he calls it a ‘looping effect’ (1995, pp. 369–370) whereby knowledge of, in his example, a ‘type’ of person gets applied to the person but in the process generates new knowledge that then feeds back (‘loops’) back to the type-category that person has been deemed to be in. The knowledge is then reapplied, with new knowledge generated, which loops back, etc. The lack of clarity that Foucault exhibits in his argument for the operation of connaissances, and then how this knowledge moves to becoming savoir, can be greatly aided by considering knowledge as ‘looping’ in this way. (My thanks go to one of the reviewers of this paper for bringing Hacking’s argument to my attention.)

12. For more on this topic, see Hardy (2011, 2015).