Clarifying Power, Domination, and Exploitation: Between “Classical” and “Foucauldian” Concepts of Power

The paper examines two ubiquitous concepts of power: the “classical sociological” concept which draws on Max Weber’s definition of power, and the “Foucauldian” concept which stems from Michel Foucault’s genealogical works. Three main theses are argued for. First, the two concepts are not, in most respects, as radically different as it is usually claimed. It is demonstrated that both can make room for different sources of power, for understanding power in a non-reified way, for the fact that power is rarely completely centralised, etc. Second, in those respects in which the two concepts actually differ, the classical view of power is more convincing and useful than the Foucauldian one. It is demonstrated that the Foucauldian view is implicitly positivist in the normative domain and thus unable to differentiate between power and domination, and that it succumbs to errors of methodological holism (i.e. undertheorising agency). Third, it is argued that the classical sociological view allows to analytically distinguish between power, domination and exploitation. These three categories are shown not to be synonymous and to carry with them importantly different sociological implications. It is demonstrated that exploitation cannot merely refer to any process of unpaid appropriation of surplus as obvious false positives are generated from this definition. Nonetheless, such appropriation is the fundamental characteristic which differentiates exploitation from domination (but not power itself), and this reveals an important sociological implication for the dynamics of struggle of the exploited against exploitation in contrast to the struggle of the dominated against the dominators.


Introduction
The term "power", usually adjoined by the adjectives "social", "economic" or "political", has for a long time been among the most widely used social-scientific concepts.This should not come as a surprise, especially to sociologists, since the "history and theory of power relations [are] likely to be virtually synonymous with a history and theory of human society itself" (Mann, 1986: 1).Indeed, despite their many disagreements, even fundamental ones, sociologists largely agree and have always agreed that, as Haugaard and Clegg (2009: 1) recently put it in the Sage Handbook of Power, "[t]he concept of power is absolutely central to any understanding of society", since power is, indeed, "chronically and inevitably involved in all social processes" (Giddens, 1995: 268).
Even so, the concept of power is a hard nut to crack and many important issues regarding power remain unresolved.One reason for this is that discussions about that fundamental social phenomenon tend to be plagued by vague, unclear, unsystematic or even internally inconsistent definitions of power.To give just a few prominent examples, does power always imply domination or should the two be distinguished?Should social scientists concern themselves with the normative dimensions of power or should they heed Weber's positivist injunction of value-neutrality?Is power an abilityan ability which can remain dormant, unexercised -or can it only exist as exercised, actualised?What about the related notion of (economic) exploitation?How is exploitation to be related to power?Is it a subset of power?For instance, is exploitation just that exercise of power which involves "the separate appropriation of surplus labour" (Therborn, 1980: 9)?This classical Marxist definition which has become commonplace in sociology is, in fact, seriously lacking as it stands.For one, it implies that all those marginalised social groups which receive welfare, i.e. the handicapped, the elderly, the poor, are exploiters -after all, they appropriate, without remuneration, the surplus labour of the taxpayers.In order not to belabour the point any further let me, lastly, observe that even influential and important discussions of power tend to define it so abstractly as to make any concrete application of the concept specious.For example, Foucault (1978: 93) somewhat mysteriously announced that power "is the name that one attributes to a complex strategical situation in a particular society".Such definitions of power may sound tantalising but, by themselves, they merely rephrase the problem.The question of what power is -what kind of complex strategical situation it represents -remains unanswered.
The main purpose of this paper is, therefore, to clarify as much as possible what power is, how it is most usefully conceptualised, and what the distinctions are between various forms of power.In the following two sections of the article two influential definitions of power are presented and examined.For the sake of convenience these two were named, respectively, "classical sociological" and "Foucauldian".There, the differences between power and domination are teased out, and an argument is put forth in defence of making normative statements.At the end of the article the related concept of exploitation is examined and an important sociological implication of distinguishing between nonexploitative domination and exploitation is laid bare.

Power and domination: Refining the classical notion
To present what I call the classical sociological concept of power it is useful to begin with Max Weber's definition. 1 I should point out that I will not be reconstructing Weber's own account of power, nor the accounts of other classical sociologists, in any detail in this article; I start with Weber's definition simply as a productive way to launch my own discussion.For him (Weber, 1978(Weber, [1922]]: 53) power is, as is well known, "the probability that one actor within a social relationship will be in a position to carry out his own will despite resistance, regardless of the basis on which this probability rests".This can be, and usually is, further condensed.Social power is simply the ability of agent A to influence agent B in such a way (with the help of either personal or impersonal means) that agent B does something he/she otherwise would not have done, or does not do something he/ she otherwise would have done (cf.Dahl, 1961).An everyday example of power thusly defined is the power wielded by parents over their children.There are at least four important and more concrete points about power that should explicitly accompany the abstract definition provided.
First, as should be obvious from both the general definition and the concrete example, power is not limited to its exercise.Parents do not always exercise the power they have over their children, but they do not thereby stop possessing it.Within the classical sociological definition, power should 1 There are many layers to contemporary debates about power which cannot be discussed here in much detail due to limitations of space.One such layer concerns the nature and legitimacy of the distinction between two kinds of power (Dowding, 1991).On the one hand, power is the ability of an agent to pursue his/her own goals regardless of the actions of other agents; this is power to.On the other hand, there is power as the ability of an agent to steer the actions of other agents; this is power over.For the purposes of this article I will not explicitly distinguish between the two, especially since, as Pansardi (2012: 82) convincingly argued, "although the two concepts do not stand in a relation of perfect logical equivalence, they have an extremely high degree of correspondence".be seen as a capacity which can be either activated or remain latent, just as a sports car possesses the capacity to travel at 250 km/h even in the case when it is actually never driven at that speed.
Second, the means or sources of power are manifold.With the category of impersonal means of power I refer particularly to the vulnerable structural position of an agent over whom power can come to be exercised.Such structural vulnerability exists, for example, when access to the means of production and/or subsistence is unequally distributed so that there are those who are dispossessed.These property-less agents are in a vulnerable structural position in relation to those who have a smaller or larger monopoly over property because the latter can, owing to their privileged structural position, make the former an offer they cannot refuse.In other words, the property-less have to curry favour with the property-holders if they are to make a living which gives the latter the upper hand.With the category of personal sources of power I have in mind any threat of violence or the actual use of violence, whether this violence is physical in nature or not.Personal means of power also encompass juridical and economic sanctions (financial penalties, jail time or prison, loss of employment, salary reductions, etc.), or threats thereof, inequalities of status, manipulation, agenda-setting, more general threats of revoking privileges, promises of benefits and so on.
Third, the exercise of power need not be in conflict with the interests of the affected agent.Lukes (2005Lukes ( [1974]]: 30) defined power too narrowly when he claimed in 1974 that "A exercises power over B when A affects B in a manner contrary to B's interests".In fact, power is not necessarily negative -it can be, but it need not be.What Lukes' original definition actually points to is domination, not power, which he recognised in the introduction to the second edition of his 1974 book (Lukes, 2005(Lukes, [1974]]: 12).Domination is, indeed, that form or subset of power which contradicts -I should say nontrivially contradicts -the objective interests of the affected agent (on this see also Arnold and Harris, 2017).In order to illustrate the point, we can continue with the example of the power relationship between parents and children.When a parent exercises his or her power over the child by denying the child primary education, the parent hurts the objective interests of the child.The child is being dominated.The parent's power becomes domination.However, if a parent uses his or her power over the child to make the child who has badly broken an arm visit a physician, even though the child refuses to do so, the parent is promoting the child's objective interests.In this case, he or she is still exercising power over the child but is not, pace Lukes' older definition, dominating the child.
Fourth, one of the implications of the second observation above is to reveal how variegated power can be.There are many "faces" of power, as they have come to be called.There is power as simply (a) the overt assertion of will of some agents over others as happens in, for example, cases of lobbying for the government (Dahl, 1961).Then there is power (b) in a more covert, negative sense as excluding issues or participants from the decision-making process as happens when, for example, "barriers to the public airing of policy conflicts" (Bachrach and Baratz, 1970: 8) are erected.Third, power can also be exercised not simply by overtly or covertly affecting agents' action but also by (c) shaping the very subjectivity and dispositions of agents.So, power can be exercised by (a) successfully imposing someone's (or some group's) preferences over the preferences of others.Here many preferences are contending to be realised but only some get through in the end.Power can also be exercised by (b) making the preferences of someone (or some group) invisible.Here the range of preferences that are in play is restricted so that their potential assertion is neutered from the start.Lastly, power can be exercised by (c) changing the preferences of agents (or groups) themselves.
With regards to the third face of power the work of Pierre Bourdieu has been very influential, as has Foucault's, the nature of which will be examined in more depth in the next section of the article.According to Bourdieu (2001: 37) "symbolic violence" can shape how people perceive the world and act in it by triggering their "habitus".Symbolic violence is defined, somewhat mystically, as "a form of power that is exerted on bodies, directly and as if by magic, without any physical constraint" (Bourdieu, 2001: 37), while habitus is taken to be a system of "durable, transposable dispositions, structured structures predisposed to function as structuring structures, that is, as principles which generate and organise practices and representations that can be objectively adapted to their outcomes without presupposing a conscious aiming at ends [...]" (Bourdieu, 1990: 53).
What is crucial is that this process of symbolic violence shaping the agent's subjectivity through his/her habitus, as Bourdieu (2001: 37) noted, does not require conscious manipulation neither is it consciously experienced by the subject (see also Bourdieu, 1990: 53).Unfortunately, there are many problems with Bourdieu's theorisation of habitus so that it is more of a necessary starting point for new theories of how power shapes the very preferences of subjects, instead of being a fully worked-out theory ready for use.For one, Bourdieu never provided the exact micromechanisms which are ostensibly responsible either for the initial formation of an agent's habitus by what he calls a social "field" or its potential activation by symbolic violence.Without an at least moderately more finegrained explanation of why and how social forces such as fields are able to deposit durable dispositions in an agent's body, or how and why forms of violence are able to activate and guide these dispositions, all we have is an allusive starting point (Hedström, 2005: 4).Furthermore, and less important for the present case, Bourdieu's concept of habitus is liable to collapse in a kind of soft socially determinist account, which he otherwise sought to avoid, whereby an agent's reflexivity and intentionality are neglected, if not wholly effaced.In other words, people's actions and perceptions tend to be seen as completely guided by unconscious habitus, itself determined by the social field, in almost all situations.Fortunately, there have been attempts, successful in my view, to synthesise Bourdieu's theory of habitus with more explicitly "actionist" theories of subjectivity -such as Margaret Archer's (1995Archer's ( , 2000) ) social realism -so as to avoid either social determinism or individualist voluntarism (see, for example, Elder-Vass, 2010b: 87-114).
Before moving on to the Foucauldian view of power, a few potential analytical concerns should be addressed relating to the definition of domination I proposed above, following the later Lukes.It could be asked what exactly the qualifier "nontrivially" refers to when I say, refining Lukes's definition, that domination is that form or subset of power which nontrivially contradicts the objective interests of the affected agent, or how I can insist on the category of objective interests after decades of anti-humanist critiques of precisely such humanist concepts.Let me take each in turn.
The qualifier "nontrivially" is put in the definition of domination because almost all forms of power contradict at least one objective interest of the affected agent to some extent.Whenever one agent influences the actions of another, even in the case of a concerned parent who is forcing his/her child to go to the hospital due to a serious injury despite the child's protestations, the short-term autonomy of the affected agent is reduced.However, in this case -and many others -it should be clear that the fact that the parent is "violating" the child's need for autonomy does not amount to domination.The child's need for autonomy is only interfered with in a trivial way: it is only denied for a short period of time that is required to get to the hospital and take care of the injury, and it is only denied because otherwise the child's bodily autonomy and welfare would be fundamentally reduced down the line -not only for a few hours, but for a lifetime.Admittedly, the category of "nontriviality" is fuzzy around the edges, so it should be settled in each empirical case whether the exercise of power only trivially conflicts with the interests of the affected agent or not.But the category itself is indispensable for the general definition of domination.
As far as the category of objective interests is concerned, it is not necessarily empirically ungrounded or analytically suspect, common objections to the contrary notwithstanding (for quality in-depth discussions see Nussbaum, 1992, Archer, 2000, Callinicos, 2004: 139-151, Sayer, 2011, Durkin, 2014and Geras, 2016[1983]).If it is accepted that there exists a common human nature, at least in the minimalist sense of humans not being infinitely plastic and malleable, i.e. of having certain unshakable needs simply in virtue of the kinds of biological bodies and brains they possess, we can infer from these at least a few objective human interests.It can scarcely be denied that people need, for example, sustenance, housing and basic health care if they are to survive, let alone live.Furthermore, it seems particularly insidious -in the midst of serious, multiple humanitarian crises across the globe -to claim people are wholly indifferent to things such as lack of housing, experiencing chronic pain, lack of personal autonomy, the denial of human dignity and so on.But if these fundamental human needs are admitted to exist, it follows logically that people have an interest, an objective interest, in their not being violated by various practices such as exploitation, domination, torture, mutilation, rape and violence in general.Assigning basic objective interests to persons need not be an exercise in philosophical naiveté, especially if two things are kept in mind.First, that what is universal and objective can have a whole multitude of concrete, culturally specific expressions which are not to be prejudged; second, that there exists a plethora of much less universally given human characteristics besides the universal ones.It only means rejecting the ungrounded antihumanist claim that people are nothing but a contingent cultural construct.As Geras (1995: 153) put it more generally: "It is not on account of any special forms of acculturation, historically particular social structures or types of learned behaviour, that people generally do not want to die of starvation or disease, or to lose their loved ones so, or to be cruelly humiliated, or to die or be permanently damaged physically or emotionally at the hands of a torturer, or to be persecuted for what they are or what they believe, be forcibly confined for it, be violently destroyed.These are just afflictions for members of our species in virtue of characteristics which we cross-culturally and more or less universally share".

Enter Foucault: Radicalising power
What is usually counterpoised to the classical sociological concept of power which I have been discussing up to now, is the Foucauldian view.It derives from Foucault's "genealogical" works of the early to mid-1970s.The Foucauldian concept of power is, so it is claimed, distinguished from the classical one by many features, the most important of which are listed below (see Foucault, 1980: 88-89, 96-98 andFoucault, 1978: 92-96): (1) Power is not only repressive but, and especially, productive; agents are not subjugated by power, they are created by it; (2) Power is not a thing, a commodity, a rare resource, an institution or a structure, but a social relation; (3) Power is not something which could be possessed by individual agents and exercised by them over others; (4) Power does not derive from one concentrated point in a society (for instance, the state or the economy); instead, it is uniquely dispersed; (5) Power does not only encompass politics, it permeates everyday life; (6) It cannot be judged whether an instance of power is unethical or not, although in some sense resistance to power is preferred.With regards to this I will argue for two claims.First, that seemingly radical differences between the classical and the Foucauldian conceptions of power are, at least in certain respects, less radical and less real than it is thought.Second, that on those counts where the two concepts do, in fact, drastically differ; it is actually the classical concept that is more sociologically convincing and useful despite the longstanding popularity of the Foucauldian view.
Let us begin with the first claim.What has to be noted from the outset is that point (1) from the list above, if interpreted in a suitably moderate way, along with points (2), ( 4) and ( 5), is wholly consistent with the classical view of power.Power, based on the classical view, is definitely not solely negative and repressive; it positively contributes to the makeup and shaping of subjects via many social processes, the most notable being primary and secondary socialisation.What is more, power is also productive in the sense that it enables those who possess it to act much more freely.So, power constrains and enables, it represses and produces.Relatedly, power not only forbids but also commands.Feudal lords did not only forbid their dependent tenants from fleeing to the cities.They also commanded them to use their (i.e. the lords') mill for milling wheat or oven for baking bread.Furthermore, it is very true, as the Foucauldian view suggests, that power is not a synonym for commodities, institutions or tangible things in general.Instead, power usually denotes a social relation among persons or groups, and some of the consequent capacities of persons and groups. 2Power rarely exists without being based on and emerging out of institutions such as the state, or commodities and things such as weapons, money, precious stones, oil, etc., but it is not identical with these institutions and commodities.It is also true to say that power does not flow from one single place in a society, as it is correct to view power as occurring in spheres other than politics or the economy.Power, at least in the form of non-dominating power -but, regrettably and to a lesser extent, also as domination -is present in almost every aspect of human lives, from families, relations of friendship and love, not to mention schools, workplaces, the streets, demonstrations and, of course, wars.
Much is made of this last point, i.e. the infamous Foucauldian emphasis on the ubiquity of power (Foucault, 1978: 93), as if it was a grand insight that is (unfairly) missed by most analyses of power.In fact, there is nothing unique or special about it.Although it is true that there are many different ways of interpreting the thesis that power is everywhere, some of which are definitely not compatible with the classical notion, the most sensible one (whatever its exegetical status) is almost trivial.Power is ubiquitous simply because power is, as per the classical definition, essentially a capability of individuals to affect each other's actions and is, therefore, either latently or actually present wherever there are relations among people.In this rather unspectacular sense, at least, power really is inescapable.
All these Foucauldian observations do not in the slightest contradict the classical conception of power and are, instead, central to it.Only points (3) and ( 6) are at odds with it.But these two points are arguably the weakest part of the Foucauldian concept of power and should therefore be rejected.Take point (3) first, the claim that power is not something that can be possessed and wielded by individual agents over other agents.As Akram, Emerson and Marsh (2015: 355), drawing on Foucault and Hayward, recently put it: "Power is not an instrument that agents use to prevent the powerless from acting freely".However, what else but the possession and wielding of power are going on when power is present and exercised, as in the paradigm cases of an employer exploiting a worker, a rapist subjugating a victim, a police officer beating a peaceful protester, a state official cancelling poor-relief, a gang intimidating a group of refugees?It is absurd to deny that one of the fundamental aspects of power is the fact that it can be possessed and wielded by individual agents.
Some might nonetheless object to my notion of power being possessed and wielded by people, as it seems to reify power, transform it into a thing or thing-like entity.This appearance is deceiving.This is so because I agree that power is a relation, not an entity, and relations cannot be possessed or wielded.However, I claim it is still useful to conceptualise power in these terms -even though we have to recognise that the terms of possession and wielding are just metaphors -because we lack better ones.So, to make my point here more precise and less metaphorical let me add two notes.First, a power relation is not a subject and therefore cannot itself act.Only people can act and, more importantly for my case here, they can act on the basis of a power relation in which they are embedded.In other words, power as a relation affords human subjects to act in a certain (i.e.powerful or powerless) way and in this sense people can possess power or lack it.An employer possesses the power to harass or fire his/her workers, while a worker who is harassed by his/her employer (usually) lacks the power to stop the employer's action, especially if he/she cannot afford to lose the job.Second, and as already pointed out above, power relations are based on, and emerge out of, the distributions of various resources which can definitely be possessed and wielded in more than just a metaphorical sense.
As far as point ( 6) is concerned it is true that not all power is necessarily illegitimate.However, can we really insist on that when it comes to domination?On this point Foucauldians surprisingly strike a position that is remarkably close to positivist social science, at least in one respect.As Coole and Frost (2010: 36) have recently pointed out, even Foucault himself insisted "on his own nonnormative positivism".Sayer (2012: 190) explained: "For positivists, facts and values are different and incommensurable.
[...] Given this, they argue that normative statements (oughts), about what is good or bad should therefore be avoided: 'no ought from is' is the slogan.[...] Foucault was a post-positivist but he also attempted to avoid normative evaluation.Some critics have claimed that his work is 'crypto-normative', to use Habermas' term; that is, while it provides description of power and its ubiquity in the social world that is bound to seem ominous and vaguely dystopian to the reader, he studiously refrains from passing judgement on whether the power is good or bad".
(See also Philip, 1985: 79;Federici, 2004: 15.) Now, this does not mean that Foucault and the Foucaldians are positivists, but it is true that they accept the incommensurability of facts and values usually associated with positivism. 3This conundrum can be resolved by recognising that although there is a gap between facts and values preventing us from logically deducing values from facts, this gap can be, pace Hume, Weber, the positivists and Foucault, bridged.In our case it can be bridged by asking under what circumstances do human beings suffer and live impoverished lives?The answer clearly has to do, in the first instance, with how their fundamental human needs are being denied, and this is a matter of objective factual analysis, not speculation.It is true that there can be no factual justification for the starting ethical premise -that is, that human suffering should be avoided while flourishing and the good life are to be promoted -but if we decide to take on board this (reasonable) premise we can bring the facts in to the very next step.There are right and wrong answers, that is to say, objective, factual, scientific answers, to the question of what constitutes human flourishing as opposed to suffering.This ethical procedure in which values and facts become entangled is usually termed either "qualified" (Sayer, 2005: 212-224) or "weak" (Elder-Vass, 2010a) ethical naturalism and there is nothing naïve about it.
3 I do not invoke positivism here as an ill-defined swearword, so let me be more concrete about it.I identify positivist social science with any body of work which, explicitly or implicitly following the empiricist or logical positivist tradition, (i) denies the connection between facts and values, (ii) characterises causality in a non-realist, Humean way, i.e. as simply the "constant conjunction" of events instead of referring to the often unobservable causal mechanisms which sometimes produce regular events, but most often do not, and (iii) accordingly seeks to explain social phenomena along the lines of Carl Hempel's covering law or deductive-nomological model (Hempel, 1965).
It might be countered that Weber was nonetheless right to insist on scientists needing to be value-neutral with respect to their research in the sense that evidence should not be ignored (or inflated) so as to justify one's preconceived values or political stances.This is most definitely the case.But with this claim Weber implicitly undermined his original claim that there is an unbridgeable gap between facts and values.If scientists can recognise which facts (evidence) conflict with their values and so, in cases where they succumb to dogmatism, reject or inflate them, this is possible precisely because there is a connection between facts and values.If facts and values were truly and completely separate matters no scientist could ever be led astray by value-considerations in his/her process of scientific research.But, in reality values depend on facts for their validity and are also contradicted by them which means that impetuous scientists can be misled by their value-commitments.It is exactly this that shows the gap separating facts and values not to be as wide and unbridgeable as Weber and the positivists are wont to say.Now, if points ( 3) and ( 6) should be rejected, so should a radical interpretation of point (1).According to this version of point (1), which Lukes (2005Lukes ( [1974]]: 12) calls "ultra-radical", power not only contributes to the overall makeup of subjects but determines subjects all the way down so that subjects are nothing but "relays" of power, the "conduits" through which power exerts itself.This is the claim which is articulated most clearly in Discipline and Punish (Foucault, 1995(Foucault, [1975]]), in the first volume of The History of Sexuality (Foucault, 1978) and in Foucault's famous "Two Lectures" (Foucault, 1980), and for which the Foucauldian concept is most celebrated.There are two main and related parts to Foucault's radical notion of power and subjectivity which should be noted.
First and most importantly, any kind of independence or autonomy is emphatically denied to subjectivity. 4The subject "is not the vis-à-vis of power" and is, instead, "one of its prime effects" so that the individual 4 It is usually claimed in defence of the Foucauldian view on this point that it "does not eliminate the subject, or a concern with agency.[...] Foucault's focus is on the very construction of the subject" (Akram, Emerson and Marsh, 2015: 355).It is true that Foucault is concerned with the construction of subjectivity and that in this sense he does not eliminate it.But this is not what is at stake.My charge is that for Foucault subjectivity is nothing but what power makes of it; subjectivity is not relatively autonomous, agential.As Akram, Emerson and Marsh put it: power is "the rules which govern the various social practices that produce subjects and their preferences" (Akram, Emerson and Marsh, 2015: 355).The subject is a ruse of power, completely determined by it.
"which power has constituted" is merely "the element of its articulation [...], its vehicle" (Foucault, 1980: 98).As Weberman (2000: 260), a scholar very sympathetic to Foucault, put it: "According to the relay model, the conscious subject is a 'relay point' or 'conduit' (my terms) or 'vehicle' or 'empty synthesis' (Foucault's terms) where anterior forces converge and combine to yield some behavioural output".Society -i.e.power -completely determines the thoughts and actions of individual subjects (see also Sharpe and Boucher, 2010).
Secondly, for Foucault there is no such thing as human nature, no universal and relatively unchanging needs or capacities possessed by the subject.Not only are human beings not partly autonomous agents with their own intentions, reasons and conscious choices, but even human bodies are mere blank slates awaiting the inscription of history.The task of genealogical analysis is, as he famously said, "to expose the body totally imprinted by history and the processes of history's destruction of the body" (Foucault, 1984: 83;see, on this, Rehmann, 2013: 218-219).Even Stuart Hall (1996: 11), who was wary of Foucault's radical "theorisation of the subject" and power -going so far as to call it, pejoratively, "a very Durkheimean [sic] concept of power [...] [which] leads to an equally generalised Durkheimean concept of social control" (Hall, 1988: 52) -happily agreed that according the Foucauldian view "the body becomes infinitely malleable and contingent" (Hall, 1996: 11;emphasis added).
The main problem with all this is that Foucault's account of the subject and its body is, in a sense, Durkheimian, as Hall notes.More specifically, it is methodologically holist, i.e. socially determinist, in the double sense that what people do (subjectivity) and what characteristics people possess (human nature) is wholly determined from the outside.In fact, one struggles at times to see what the big (methodological) difference is between the long-defunct Parsonian structural-functionalism and Foucauldian genealogy.As Mouzelis (2008: 241) commented: "If instead of 'subjectless practices' one posits subjectless social processes, instead of 'objectives', system requirements, and instead of the 'construction of subjectivities', socialisation, the methodological similarities between Foucault and Parsons become quite striking.Both underemphasise agency, and as a result both have to resort to teleologically oriented functionalist explanations".At this point it might be objected that any critique of the genealogical Foucauldian view of power as holist misses the fundamental fact that Foucault himself was a nominalist.Indeed, Thomas Flynn (1989) countered that not only was Foucault not a methodological holist, but that his stance is actually, owing to his nominalist commitments, the very opposite -an individualist one.He explained: "What Foucault calls his 'nominalism' is per force of a kind of methodological individualism.It treats collectivities such as the State or abstractions like 'man' or 'power' as reducible, for purposes of explanation, to the individuals that comprise them.[...] These are standard claims of nominalism as commonly understood" (Flynn 1989: 134-135).Foucault (1978: 93) was himself quite explicit about his nominalism: "One needs to be nominalistic, no doubt: power is not an institution, and not a structure, neither is it a certain strength we are endowed with; it is the name one attributes to a complex strategical situation in a particular society".So how can I maintain that the Foucauldian view of power and subjectivity is holist when, in fact, for Foucault there is no monolithic power, institution or structure which would determine subjectivity in a Durkheimian, top-down fashion?There are at least two points that can and need to be raised in response.
First, it is true that, being a nominalist, Foucault does imply that power is simply a name, not a real entity (or, as I would have it, a relation among people).If that is all there is to it, Flynn is right and Foucault cannot be a holist since for him power was not a supra-individual force capable of influencing individuals.In fact, power on this view seems not to exist at all, either as a supra-individual entity or anything else for that matter.Power is simply a name.However, this cannot be.The strict nominalist position is unsustainable even from Foucault's own perspective because why study power if it does not actually exist (at least not as something more than a name)?Furthermore, the term "power" might simply be a name, but it is a name that refers to, according to Foucault (1978: 93) himself, "a complex strategical situation in a particular society".So, it turns out that power -or, rather, that to which the term power points -does exist.But how does it operate?What is its relation to individual subjects?
Here I come to the second point.Foucault can deny power being an institution or a structure and thereby circumvent, on the face of it, my claim of him being a methodological holist.But my charge is more general than this and is thus compatible with Foucault's denial of power being a totalistic structure that defines everything.What I find holist in the Foucauldian view is the thesis that what individuals do is wholly determined by extraindividual forces.What might these forces be if not institutions and structures?Flynn (1989: 135) called them practices, "socially sanctioned body of rules that govern one's manner of perceiving, judging, imagining and acting".So, it's not monolithic institutions, Durkheimian "social facts" and structures that determine the subjectivity of an agent.It is social "practices" that do so.Well, this is exactly what I find objectionable and what I call methodological holism, although I have no particular attachment to this exact term.It therefore seems to me that far from nominalism and holism contradicting each other, they are actually quite compatible.Flynn himself revealed as much when he went on to say that the characteristic move "of a historical nominalist" is the: "[e]vacuation of the creative subject from significant historical agency.
[...] The most profoundly nominalistic aspect of his project is perhaps this distillation of the subject into the point of intersection of various practices.[...] Foucault has reduced the linguistic subject to a placeholder in a shifting series of practices" (Flynn, 1989: 140-141).Whatever we call this thesis -holism, anti-humanism, etc., the Foucauldian view that individual subjects are not relatively autonomous agents who carry out practices, but are instead themselves carried out by practices, defined as a "socially sanctioned body of rules", is deeply problematic.It is so because the single biggest flaw of methodological holism, most evidently on display in Parsonian sociology but also in Althusserian Marxism, is that it cannot explain the omnipresent empirical fact of human resistance to power, subjugation, socialisation, etc. 5 This is all the more problematic for Foucault who constantly referred to resistance and stated that wherever there is power there is resistance to it.How can this be if subjects are mere "relays" or "vehicles" of power?As Harpham (2006: 3) commented: "This emphasis on the ways in which the self is shaped, formed, corrected, normalised, and documented by agencies and structures beyond its knowledge or control made it difficult to imagine how such a thing as a self-aware or self-determining individual might arise within the ʻregime' of modernity, with its proliferating sites of control".
5 One of the first critics to point this dilemma out in relation to Foucault was probably Nicos Poulantzas (Poulantzas, 2000(Poulantzas, [1978]]; Jessop, 2004).I should also mention here that Poulantzas has rightly argued against the Foucauldian view for reducing power to its exercise and for downplaying the continued significance of repression in modern societies.
It is one thing to emphasise that power is not only repressive, and a very different and incorrect thing to imply that power has nothing or almost nothing to do with repression.
The problem is even more acute than that.Not only is it that on Foucault's radical account of subjectivity it is hard to imagine how a subject capable of resisting could emerge, but we can also ask why such a subject, if he/she were somehow to emerge, would want to struggle?If the subject's body is a blank slate, completely formed by power, it will not in any way collide with the functioning of power.As Lena Gunnarsson (2013: 12) put it in a different context: "We cannot make sense of from where socially constructed power structures derive their oppressively constraining effect on people without a notion of constraints that are not socially constructed [i.e. are a part of human nature], such as the need for recognition that underpins our vulnerability to other people's views".
No needs or interests will clash with the exercise of power since all of the subject's needs and interests are completely created by power itself; to say otherwise is to admit that human bodies are not "totally imprinted by history" and "infinitely malleable".But if there is no conflict of interests in the exercise of power there is also no reason for resisting it.
It is, therefore, not surprising that Foucault himself later moved away from this "ultra-radical" concept of subjectivity.If in his genealogical period the individual "is not the vis-à-vis of power" and is, instead, "one of its prime effects" (Foucault, 1980: 98), this position was reversed in the late 1970s and early 1980s, just before his death.He noted that when he "was studying asylums, prisons, and so on" he insisted "too much on the techniques of domination" (Foucault, 1993: 203-204).He went on: "But, analysing the experience of sexuality, I became more and more aware that there is in all societies, I think, in all societies whatever they are, another type of techniques: techniques which permit individuals to effect, by their own means, a certain number of operations on their own conduct, and this in a manner so as to transform themselves, modify themselves.[...] Let's call this kind of techniques a techniques of technology of the self" (Foucault, 1993: 203; emphasis added).
It seems that the later Foucault endowed subjects with a great deal of autonomy and even allowed room for something like a minimal human nature.With this he opened himself up to the possibility of a different notion of power.Earlier he denied, consistent with his anti-humanist stance, any autonomy to the subject and so refused the classical concept of power.Power could not be conceptualised as something possessed and exercised by subjects over other subjects.Quite the reverse, subjects themselves were seen as being possessed and exercised by power.But later, moving away from the "ultra-radical" notion of subjectivity, Foucault (1983: 217) said something very different about power: "Let us not deceive ourselves: if we speak of the structures or the mechanisms of power, it is only insofar as we suppose that certain persons exercise power over others".So, for him the term ʻpower' designates relationships between partners [...], an ensemble of actions which induce others and follow from one another [...], a mode of action upon the actions of others" (Foucault, 1983: 217, 221).No longer were subjects seen as mere effects of power, "docile bodies".Instead, power "is exercised only over free subjects, and only insofar as they are free.By this we mean individual or collective subjects who are faced with a field of possibilities in which several ways of behaving, several reactions and diverse comportments may be realised" (Foucault, 1983: 221).As he put it in an interview of 1980 in more detail: "What does it mean to exercise power?It does not mean picking up this tape recorder and throwing it on the ground.I have the capacity to do so -materially, physically, sportively.But I would not be exercising power if I did that.However, if I take this tape recorder and throw it on the ground in order to make you mad, or so that you can't repeat what I've said, or to put pressure on you so that you'll behave in such and such way, or to intimidate you -well, what I've done, by shaping your behaviour through certain means, that is power" (Foucault, 1988: 2).This conceptualisation is much more like the classical sociological notion of power (and domination), and if my preceding arguments hold Foucault was right to accept it.

The differences between domination and exploitation
With the definitions of power and domination in place we can now turn to the oft-used but rarely clearly or at least usefully defined concept of exploitation.It was noted above that domination is a subset of power, so domination is simply one form of power.Now I want to add that exploitation is itself a subset of domination, i.e. exploitation is a particular form of domination.More specifically, exploitation is just that form of domination which involves -besides the usual undermining of objective interests and either personal or impersonal force which are characteristic of domination -the extraction of unpaid surplus labour from the targeted agents.These three elements are, then, the key elements of exploitation: (a) the undermin-ing of agent's objective interests, (b) personal or impersonal force, and (c) the unremunerated extraction of surplus from the agent.Together they provide sufficient conditions for exploitation (on this see also Reiman, 1987;cf. Kymlicka, 2002: 177-190 andVrousalis, 2013).
Such a definition is more nuanced than the classic Marxist concept of exploitation.According to the classic Marxist definition, exploitation occurs wherever unpaid surplus labour is extracted from immediate producers, regardless of the presence of force and how this extraction process impacts the interests of those involved.Therborn, for example, explicitly rejected the category of (objective) interests because it ostensibly rests on "unwarranted and untenable assumptions" (Therborn, 1980: 5).He went on to define exploitation very broadly: "The concept of 'exploitation' in historical materialism refers simply to the separate appropriation of surplus labour; in other words, to the fact that one category of economic agents works more than is necessary for their own reproduction and that the fruits of their surplus labour are appropriated by another".There is a good reason for rejecting this broad and otherwise virtually ubiquitous Marxist concept of exploitation.This is so because, as mentioned in the introduction to the article, it generates false positives.It categorises as exploitation those phenomena which are under no reasonable stretch of imagination actually exploitation, as happens in the case of handicapped people who receive welfare benefits which are extracted without remuneration from workers through taxes.The definition I offered avoids this problem.To give more examples: according to my definition, the mere act of giving gifts need not be exploitative, yet on the classic Marxist view it has to be.It is, of course, true that the act of giving a gift involves the transfer of unpaid surplus labour and its appropriation by the agent who receives the gift.But this is not necessarily exploitation.It becomes exploitation only when the gift-giver is forced, either personally or impersonally, to offer gifts and when the act of giving nontrivially hurts his/her objective interest in material wellbeing, autonomy, etc.Similarly, when a poker player loses a large sum of money, this situation need not be viewed as exploitative even though an appropriation of unpaid surplus labour (in the form of money) has occurred; it might also be that, if the sum is high enough, the loser's objective interests are greatly diminished.But as long as the play was not a result of impersonal or personal coercion, it is not reasonable to call it exploitation.
It was said above that exploitation is not synonymous with domination but is a particular form of domination.What distinguishes exploitation from domination (as opposed to mere power) is precisely Therborn's "appropriation of surplus labour".So, on the one hand, the relations between a slave and a slaveowner, a serf and a landlord or a worker and a capitalist are all relations of exploitation.On the other hand, the relation between society's prevailing religious group and the minority religious group, wherein the first, for example, fundamentally restricts the democratic rights of the latter, or the relation between Early Modern European colonisers of North American and its indigenous population, are both relations of domination (the last example is from Wright, 1995).The sociological implications of these two different forms of power -exploitation and domination -are not the same.Exploited agents have a much better bargaining position than those who are dominated.The exploited possess an important amount of potential power in relation to their exploiters.This is so because, as Erik Olin Wright (1995) put it, the exploiters need their exploited subjects -they are dependent on them for their own material benefit -while the dominators do not care about their dominated subjects.A capitalist who loses his or her workers, or a feudal lord who loses serfs, cannot generate income which she or he otherwise gets from them.This is of a serious concern to the exploiter which means that their own interests compel them never to completely ignore the pleadings and pressures of their exploited subjects.It is not only the fact that they do not want to lose their subordinates that compels them not to completely ignore the subordinates' requests.It is also, and more importantly for sociological analysis, that when subordinates exert pressure over them, they immediately feel that pressure on their own pocketbook.
In contrast, the prevailing religious group which dominates the minority does not care at all if the latter vanishes, as the European colonisers of North America did not care about what happened with the indigenous population.In fact, dominators usually want the dominated to be gone.This is the main difference between relations of domination and relations of exploitation.As Wright (1995: 92) noted: "Genocide is thus always a potential strategy for nonexploitative oppressors.It is not an option in a situation of economic exploitation because exploiters require the labour of the exploited for their material well-being.It is no accident that culturally we have the saying, 'the only good Indian is a dead Indian', but not the saying 'the only good worker is a dead worker.'The contrast between South Africa and North America in their treatment of indigenous peoples reflects this difference poignantly: in North America, where the indigenous peoples were oppressed (by virtue of being coercively displaced from the land) but not exploited, genocide was the basic policy of social control in the face of resistance; in South Africa, where the European settler population heavily depended upon African labour for its own prosperity, this was not an option".From this it follows, as was said, that the dominated have a much weaker bargaining position in relation to their dominators than do the exploited in relation to their exploiters.An important consequence of this fact is that when the dominated struggle against their dominators they can be successful in their struggle only if they manage to find a way to make themselves felt by the dominators, to make them care, to make the dominators' ignorance of the dominated costly.As the struggle for democratic rights reveals (Therborn, 1977;Tilly, 1978;Fox Piven and Cloward, 1979), the dominated have always been aware of this and have usually realised their demands by building on the latent power of the exploited, usually in the form of strikes, sit-downs and other collective actions which can be immensely costly to the exploiters (as well as those dominators who are indirectly dependent on the exploiters' power and revenue) and thus make them sensitive to the demands of the movement.

Conclusion
The popular Foucauldian view of power which, in contrast to the classical sociological one, construes power as that which creates subjects instead of subjects being those who use power is at least doubly unsatisfactory.It succumbs to methodological holism according to which subjectivity is a mere ruse of power retaining no causal autonomy, and it renews the myth of facts and values being separated by an unbridgeable chasm thus banishing the normative dimension outside the purview of social science.This myth and the consequent value-neutrality of social sciences create another problem for the Foucauldian concept of power: they make it unable of differentiating between power and domination. 6n fact, as we have seen, there are subtle but important ontological and analytical differences between the categories of power, domination and exploitation.Power is most usefully and convincingly conceptualised as a capacity (whether exercised or not) of an agent or a group of agents to influence other agents in such a way that the latter do something they otherwise would not have done, or not do something they otherwise would have done.The levers of power are either personal, impersonal or both: they can refer to the structural positions of agents, the resources and rights they possess or have access to, sheer physical force, intimidation, agendasetting, status, offers of awards and benefits etc. Foucault himself recognised the usefulness of this concept of power, at least to an extent, after his "genealogical" phase.When power is negative, i.e. when it hurts the objective interests of the targeted agents, in a nontrivial manner, it becomes domination.Such domination becomes exploitation if it involves unpaid transfer and appropriation of surplus labour of the dominated.