The term consensus has a democratic ring to it. Perhaps it is this democratic aura that can account for the popularity of the term, but - as so often with popular terms - this also creates quite some confusion regarding its meaning. Ask ten different people how they would define consensus and you are likely to get at least five different answers. There is clearly no consensus on the meaning of consensus.

In itself, this would not be a problem, but after having spent years analyzing processes of consensus decision-making, I increasingly notice how people use the term consensus in manipulative ways, as an apparently innocent term that promises emancipatory decision-making. We all know that much abuse can and has been done in the name of consensus, but we are often unable to pinpoint exactly where the problem lies. 

I argue that this is because we don’t understand consensus. We have no clear idea about what it really means. Even our own notion of consensus becomes rather vague once we start thinking about it. Try it. Stop reading for a moment and see if you can come up with a concise definition of what you mean by consensus, one that you are still happy with after you've written it down. - If, in contrast, I asked you to write down your definition of voting, you would have much less difficulty and what you wrote down would probably correspond with what most other people wrote down. There is a clear consensus on the meaning of voting.

Why is this? Because we have all been taught the rules of voting at an early age and - even more important - we’ve been practicing those rules over and over again for years. Voting is common sense and that makes it such a powerful practice to resort to in times of conflict, when we want to settle things rather than increase the trouble by discussing the meaning of consensus. Voting also has a democratic ring to it, although it does not sound quite as democratic to say "we reached a clear majority on that" as it is to say "we reached a clear consensus on that". But for pragmatic reasons, we often accept the creeping sense of violence that accompanies majority rule. It’s still better than tyranny! And it allows us to avoid reaching consensus on the meaning of consensus. 

Some people do, of course, take the more difficult and winding road of consensus, even when there is no consensus about what this means. These are the people whose decision-making processes I've been studying: social justice activists across Europe. At first, I followed the activists' self-image that taking decisions by consensus, as they do, is very special, and, most importantly, very democratic. But I soon came to reflect more on what consensus actually implies. 

Around 100.000 people converge on the centre of Madrid, commemorating the 15M indignado movement. Demotix/David Osuna.. All rights reserved.

No worries! My aim here is not to write another one of those confessions where people realize all the downsides of consensus decision-making. My aim here is in many ways more trivial, but - I would claim - nevertheless much more interesting: In the following I will first provide a clear distinction between the logics of voting and consensus. I will then show that the general logic of consensus easily translates into a variety of different practices of consensus decision-making, each of which gives a different meaning to the term consensus. I will explain why it is crucial to distinguish between these practices and that whenever someone speaks of consensus we should ask what kind of consensus they are actually talking about. Finally, I will argue that both consensus and voting cannot solve the problem of hegemony which has accompanied modern political thought for centuries. The question that I will contemplate at the end is therefore how decision-making in assembly publics can be (and already is) made non-hegemonic rather than counter-hegemonic. 

The logics of voting and consensus

In voting, everyone expresses their will/opinion (in the form of a vote), which means that voting produces certainty regarding the distribution of preferences among the constituency. The decision rule is applied to this distribution to determine the collective decision. For example, in the case of a simple majority vote, the option with the most votes becomes the collective decision. Or if a super-majority is required (e.g. a 2/3 majority) the proposal is either accepted or rejected, based on the number of positive votes it received. 

Because of this close link between the voting logic and various majority rules, the two are often conflated. But voting does not per se imply majority rule. Voting is also necessary to achieve unanimity, i.e. to achieve a situation where everyone agrees. If a decision is to be taken unanimously, we need to determine whether everyone agrees and the only way to do this is to ask everyone for their preference, which is the same as voting (a vote can also be "cast" orally or by some gesture which is registered in order to produce the preference distribution).

This is not the case in the logic of consensus. In order to reach consensus, it is not necessary that everyone expresses their will or opinion. The distribution of preferences remains unknown. - But what is the decision rule, then? The decision-rule is that no one disagrees. To be more precise: consensus means that no one expresses dissent. Consensus is therefore quite different from unanimity (although those who say "there is consensus" often like to imply that "there is unanimity") since all we know in the case of consensus is that no one opposed the decision. We don't know how many actually agree. It may well be possible that most participants simply don't care and see no reason to prevent the others from going ahead with whatever it is they want to do. Or - as I will show in a minute - they may very well disagree or even want to utter their dissent, but never get the chance to do so.  

Phillippe Urfalino uses the term "apparent consensus" instead of consensus. Firstly, because consensus often suggests general agreement while all we can observe is the absence of voiced disagreement. General agreement (alias consensus) therefore merely appears to exist. Secondly, the term apparent refers to the way we observe consensus: it simply appears in front of us, as it were. It appears when we observe the silence, i.e. the absence of dissent. A voted decision, in contrast, does not simply appear but has to be actively produced by aggregating the votes and comparing the preference distribution with the requirements of the decision-rule. I agree with Urfalino, but I want to stick to the commonly used term "consensus" precisely to make it clear that I am talking about the very same consensus that we commonly talk about, not some "weaker" form, which merely appears to be consensus. Differently said: consensus is always apparent and the distinction that we need to be clear about is that between unanimity and consensus. Unanimity belongs to the logic of voting, consensus belongs to the logic of - well - consensus.

Having clarified this distinction between different logics of decision-making, let us move on to the actual practices, i.e. what people actually do when they make collective decisions in a meeting or an assembly. I will skip the part about voting practices since we all know how voting works and there is not much new about what I could observe in this domain. The crux in voting is not so much in the actual practice of voting but in designing the decision-rule. But that is a different story.

The practices of consensus

To acknowledge that consensus is not the same as unanimity does not mean that consensus is deficient. Rather, understanding the logic of consensus allows us to critically investigate how consensus is achieved in practice. In other words, knowing that consensus requires a situation where dissent is absent, we can ask how such a situation is produced in practice. Based on my observation of decision-making processes in about 200 meetings and assemblies, I distinguish four types of consensus: imposed consensus, acclaimed consensus, hasty consensus, and considerate consensus. 

In an imposed consensus, the facilitator (or perhaps more adequately: the chair) of the meeting simply claims that consensus has been reached and directly moves in to the next topic or closes the meeting. Participants who feel that this claim to consensus is not legitimate (either because they intended to voice their dissent of because they sense that others wanted to do so or perhaps did but where ignored) they will have to be quite articulate in order to make their objections heard. In fact, they actually have to disrupt the meeting and demand that the observation of consensus be annulled. I say "disrupt the meeting" because there is no explicit slot for them to voice their concerns; the meeting has already moved on to a new topic and to annul the consensus requires that the ongoing conversation is interrupted and suspended until the previous issue has been settled. Anyone who has been in such a situation knows that we tend not to make this disruptive step unless the point is really important to us or we feel that we have strong support from other participants. Whether consciously or not: facilitators know that people don't want to find themselves in the position of the quibbler or unreasonable screwball and they often use this constraint to make the meeting move on - not necessarily to the disadvantage of the participants.

In an acclaimed consensus, there is an explicit slot for participants to express their views, but its official purpose is not to express dissent but consent. Participants are asked to confirm that they agree with "the consensus" (e.g. "Does everyone agree?"). The assumption is that, yes, everyone agrees. A "No" is clearly the less preferred option (implied in the wording of the question and how it is articulated). But perhaps more importantly, the fact that such a question prompts participants to actively consent means that dissenters can easily get lost or be - deliberately or not - overheard and consequently ignored. If a dissenting voice is not immediately recognized as such, it becomes even more difficult to subsequently disrupt the meeting in order to claim dissent, than in the case of imposed dissent, where the disruption may be seen as having some evident legitimacy due to the absence of an explicit slot for dissent. Although such a slot is more or less absent also in the case of acclaimed consensus, this absence is not so evident, since a slot does exist and it is not so clear why this is not really a slot for dissent.

In a hasty consensus, in contrast, participants are asked if there is anyone who disagrees. A frequent wording in the meetings I observed was, "Is there anyone who cannot live with this decision?" thus providing an explicit slot for dissenters to, "speak now or forever remain silent", as it were. In this practice consensus appears as a stretch of silence and by observing this silence as the absence of dissent, participants determine that consensus exists. In this practice, those who have previously voiced concerns will find it easy to insist on their objections if they wish, but opening up new concerns may be somewhat more difficult at this point since the question, especially when someone is asked if they “cannot live” with the decision, still exerts pressure on possible dissenters by suggesting that they should not voice their dissent unless they really have a big problem with what is going to be decided. Although the situation is more balanced than in the previous two types of consensus, there still is an expectation that people should think twice before breaking the silence. For this reason, the silence often remains rather short, in order not to "provoke" anyone to speak up. Calling this practice "hasty" may seem somewhat unfair, but the term may become more plausible in comparison with the following, less hasty consensus practice.

Considerate consensus, finally, comes close to establishing unanimity with the means of consensus, even though this is strictly speaking not possible, as we have seen above. I compare considerate consensus with unanimity because participants are not only given the opportunity to express dissent, but they are even encouraged to do so in order to make sure that no one is silenced. Although this is not the same as asking everyone to state their preference, considerate consensus probably comes as close to observing agreement by everyone as is possible within the logic of consensus. The facilitator encourages people to raise concerns, perhaps even re-stating previous concerns to see if they have been sufficiently addressed, or people who look unhappy or seem to be frowning may be personally addressed and encouraged to put their concerns on the table. The aim is to discuss the concerns and objections that people may have and to find ways to integrate them into "the consensus" so as to make it acceptable to all. In this practice, consensus is only established when a significantly long stretch of silence has been observed despite having made it easy to break it. 

15M in Madrid's Plaza del Sol. Demotix/Humberto Rolens. All rights reserved.

I should be careful not to rhetorically blur the distinction that I flagged as so crucial in the previous section: there is no way how even the most considerate consensus can be the same as unanimity. The difference between the two is not one of degree, i.e. of quantity, but of quality. But the point of looking at how consensus is actually established in practice is to see that despite the fundamental difference  in logic, consensus and voting share a problem that may be more evident in voting but which - it seems - is also unavoidable in consensus: there is always an element of coercion. It becomes more subtle with each of the four types of consensus, but even in the practice of considerate consensus, there is almost inevitably a certain pressure on dissenters to eventually remain silent, even against their will. In the final section, I want to explore the conditions of this inevitability and I will argue - as the disclaimer ‘almost’ already indicated - that coercion can in fact be avoided, but that this requires overcoming what Richard Day calls the "hegemony of hegemony".

The problem of hegemony 

Hegemony is a struggle for dominance that relies on coercion as well as consent. According to Antonio Gramsci, a group seeking "supremacy" must, on the one side, lead the masses that recognize and accept the group's superiority (and therefore give their spontaneous consent to this leadership); on the other side, there are always some people who simply refuse to understand what is best for them - they it must be coerced, disciplined, subjugated, or even liquidated. Or to use Day's succinct formula: "Hegemony = Dictatorship + Democracy". 

In his book Gramsci is Dead, Day shows us how the two big theoretical traditions of western political and social thought in the last two centuries, liberalism and Marxism both theorize emancipation and human progress as necessarily relying on both consent and coercion: "Despite their many differences, modern liberals and modern Marxists share a fundamental assumption about the necessity of hegemony" and he goes on to show that hegemonic tendencies even remain dominant in the new and newest social movements since the 1960s when they seek to establish various forms of counter-hegemony. Yet, he also observes that activists and political thinkers are increasingly exploring alternative paths: "What is most interesting about contemporary radical activism is that some groups are breaking out of this trap [of reproducing hegemony through counter-hegemony, CH] by operating non-hegemonically rather than counter-hegemonically". As Day acknowledges, this is easier said than actually done, and it is to this debate that I intend to add some thoughts by scrutinizing practices of decision-making in assembly publics.

Celebrating the 15M anniversary in Madrid. Demotix/Alberto Sibaja Ramirez. Some rights reserved.

In the case of majority voting, the coercive aspect is quite evident since the majority gets to dominate the minority. Voting, like modern liberal thought, emphasizes the role of individual autonomy and the need to integrate them into the moral community by giving them an equal say in collective decision-making. Consensus, in contrast, focuses on the collective and emphasizes the importance of its unity. The role of the individual here is to consider how important his or her (dissenting) position is in relation to the collective. Because an individual veto can, in principle, prevent the others in the collective from deciding what they want, dissenters carry a huge responsibility on their shoulders so that they are required to at least justify their position in the light of generally acceptable norms. Difference is valued at certain stages of the decision-making process, but once a sense of possible consensus emerges, minority dissenters who cannot give convincing reasons will increasingly be perceived as a problem and consequently be silenced or ignored.

This points to an aspect of collective decision-making in assembly publics that consensus and voting have in common and which is ultimately responsible for their hegemonizing character: the decision is made by all for all, i.e. every participant is expected to comply with the decision. Differently said, assemblies usually assume the existence of one community (or the need ro create one) which governs itself through the decisions it takes. This corresponds to the universalizing and totalizing nature of hegemony which is aimed at "effects that are (1) felt over an entire social space, usually a nation-state, and (2) are expected to occur across a wide spectrum ... of social, political, cultural, and economic structures and processes" (Day 2005: 65). If we apply this insight to the smaller social space of an assembly public, this means that in order to reduce the oppressive character of collective decisions, regardless of whether these are made through voting or consensus, we need to find ways to secure an exit option for dissenters. In the case of assembly publics, this means that:

• - the option not to decide a particular question as a collective should never be ruled out: the decision to decide collectively is also a collective decision;
• - the decision-making collective, however conceived, can always be reconfigured by breaking it up into smaller units: integration is not a value in itself;
• - the acceptance of a veto should ultimately not depend on reasons given: a veto is a veto.

These points are up for discussion and I would like to briefly explain each one. If an assembly reserves the possibility to decide not to decide a particular issue, the pressure on dissenters is considerably reduced. This does not mean that the discussions of the assembly were in vain or that the issue in question cannot be addressed at all. As mentioned in the second point, the issue can always be taken up by members of the assembly at a more local level, i.e. in smaller groups who might decide to solve the issue in different ways rather than in the one and only way that the assembly might have decided. This includes the possibility of "the assembly minus one dissenter" deciding whatever they want to decide (which would mean that the dissenter does not have to comply with that decision). This possibility removes the fear of the majority that an "unreasonable" dissenter will block their progress. Consequently, there is no reason to pressure dissenters to justify their veto, especially not in the terms of the majority. While it is surely more productive for both sides to engage in a discourse on good and bad reasons, the acceptance of the dissenter's reasons cannot be the condition for the acceptance of the veto. 

If these suggestions for a non-hegemonic logic of collective decision-making in assembly publics are put into practice, this does not automatically replace the logic of hegemony with what Day calls "the logic of affinity guided by groundless solidarity and infinite responsibility", but surely some clarity about how consensus decision-making in itself is not emancipatory, is not a bad idea. Marianne Maeckelbergh, author of “The Will of the Many” has shown that the alter-globalization movement as well as the occupy movement are actively experimenting with post-hegemonic forms of decision-making – taking a closer look at their practices and experiences is a promising way forward for the necessary reinvention of democracy – in the small scale of assembly publics as well as in the large scale of society.

This article is part of the Creating Publics, Opening Democracies partnership, funded by the Open University. Read more about openDemocracy's editorial partnerships programme.