6.4.2. Consent

What has yet to be examined are the processes to be employed to ensure that the bureau’s discussions actually are adequately deliberative, thereby fulfilling the democratic principle of consent which underlies its legitimacy. This is even more important in respect of the plenary body, for which the principle of consent cannot be fulfilled by making it proportionally representative.

A starting point is for the strengths and deficiencies of the IGF’s existing processes for deliberation to be identified. Its strengths can be stated in three points, each of which relates back to earlier discussions of the features of deliberative democracy:[1]

The main deficiencies of the IGF as a forum for democratic deliberation can also be stated in three points:

The first two of these deficiencies have already arisen for consideration elsewhere. A summary of the strategies required to address the first deficiency, as to the IGF’s capacity to capture the participation of all affected stakeholders, was given at Section and will be revisited under the heading of inclusion below.[4] As to the second deficiency, the importance of the IGF’s policy-setting role and the structures necessarily to facilitate its performance were considered in the two preceding sections of this chapter.[5] This leaves the third deficiency, as to the IGF’s incapacity to engage in deliberation towards the end of achieving a rational multi-stakeholder consensus, to be dealt with here. Democratic deliberation

This last incapacity is illustrated by the fact that when the IGF or its dynamic coalitions have developed policy positions, these have tended to reflect the prevailing views of their dominant members, rather than emerging from a process of deliberation between equals in which the preferences of all stakeholders are considered and balanced.

This is evident at several levels, including those of the open consultation meetings, the Advisory Group, the annual plenary meetings and the workshops and dynamic coalitions:

Instilling a more deliberative quality into the IGF’s processes is therefore a project to be undertaken across multiple institutional layers. Furthermore, it is to extend within each of these layers from the earliest phase of discussion—that of agenda-setting—through to the final phase in which its output passes through to the next layer (or falls back to the previous one).

As will shortly be seen, the specific processes best suited for adoption at a given layer of governance within the IGF, or at a given phase within each layer, will differ. Processes will also differ markedly between those applicable to participants present in person, and those participating online.[8] But despite this variance in detail, the underlying features of most institutionalised frameworks for democratic deliberation, as examined in Chapter 4,[9] are common and relatively simple:

Any of the large-scale structures for democratic deliberation that incorporate these features, including the 21st Century Town Meeting, citizens’ assembly, consensus conference and speed dialogue, have the potential to be applied directly to the IGF’s plenary body. Since speed dialogues came close to being trialled for the Rio IGF meeting, and have been successfully employed by the ITU in an analogous context, these seem the most natural choice of method to improve the deliberative character of the IGF’s plenary meetings.

As Henry Judy of the American Bar Association explained at the May 2007 open consultation meeting, the strengths of the speed dialogue format include the following:

First, it introduces a large number of people to one another who might otherwise not have spoken to one another. It is a great networking tool, and it stimulates networking, and thus it would strengthen the multiparticipant orientation of the forum. Secondly, it is a great equalizer. The great and the small are at the same table and must listen to one another. Third, it forces people to speak crisply. You do not have time for the extended use of diplomatic code, euphemisms, and circumlocution. Fourth, it is useful for synthesizing the state of opinion and emotion in the group.

The reference here to “synthesizing the state of opinion” of the group is noteworthy. Ordinarily, speed dialogues are not used at the final stages of decision-making; unlike, for example, the 21st Century Town Meeting in which the closest to a rough consensus position that the groups achieves is put to a formal vote at the end of the meeting. However for the plenary body of the IGF, which as Nitin Desai is fond of noting has no defined membership, vote-taking is out of the question. Instead, it is for the bureau, with the assistance of the facilitators of the speed dialogue, to assess the state of the group’s progress towards consensus following a speed dialogue session, as a preliminary stage to the deliberation and more formal decision-making phase that is to occur within the bureau. More will be said of this process of assessment under the following heading.

As originally scheduled for Rio, there was to have been an Athens-style moderated panel presentation and a speed dialogue session for each of the four main themes of the meeting, taking place one after the other. However, it was the speed dialogue session which was to have been held first, and the panel presentation second. More consistent with deliberative democratic principles would be for the order to be reversed, so that the panel presentation could provide the background of balanced briefing material upon which participants in the speed dialogue would begin their deliberations.

The panel presentations in turn should be built upon the written submissions contributed by stakeholders and dynamic coalitions in advance of the meeting, which provide a more diverse base of briefing material than that which can be provided by a necessarily limited group of panelists (around twelve for Athens, and six for Rio). Whereas the selection of briefing material is in most deliberative democratic models a matter for the group’s facilitators in consultation with stakeholders, there is precedent in the case of the consensus conference for the partial devolution of this function to the group itself. This would also be so in the case of the IGF, in that whilst written submissions would be received and published without moderation (as is the case already), it would fall to the bureau to draft a synthesis paper summarising all of the contributed perspectives in a factually accurate and balanced manner.

This synthesis paper would then be translated and distributed to those attending the meeting, both by its advance publication on the IGF’s Web site and by its inclusion in the materials received by each delegate upon registration. This did not occur for the Athens meeting, when the synthesis paper was published by the Secretariat only after many participants had already departed for the meeting, and in neither Athens nor Rio was the paper distributed to participants or referred to during any of the plenary sessions.

Doubtless, there will be challenges in implementing speed dialogues at the IGF. As Henry Judy summarised the drawbacks of speed dialogues:

First, it depends on each table having a reporter who can represent the views at the table in the summary in a skillful and disinterested manner. It is not easy to find a Markus Kummer for each table. Second, it requires a high degree of prior planning and instruction on the part of the organizers as well as a high degree of compliance on the part of the participants. Otherwise, it can become a confusing and unproductive exercise in herding cats, if I may use the English expression. I have heard it said that the likelihood that the technique will be successful is directly proportional to the tendency of the group to start its meetings on time. Third, it tends to work less well as the group becomes larger.

To some extent these difficulties, which largely accord with those discussed in Chapter 4,[10] may require the IGF to become better resourced so as to enable it to attract a skilled team of facilitators, along with sufficient translators for each of the round-table groups (though resourcing constraints were not given as a reason for speed dialogues being dropped from the Rio programme).

However they will also require the inculcation of shared norms such as trust, equality and cooperation that are necessary for the success of any consensual or deliberative democratic decision-making process.[11] Whilst these cannot be developed instantaneously, for a core of IGF participants (including many of the members of civil society’s CS-IGC and the private sector’s ICC/BASIS) such norms have been in the process of development since the first PrepCom of WSIS in 2002. Whilst this group is only a narrow segment of the present-day IGF, it could catalyse the development of a broader culture of cooperation.

Beyond this, the development of shared norms to reinforce the deliberative process can only come from the initiation, repetition and refinement of that process, as stakeholders within the group learn to understand and trust each other, and the group as a whole builds up its social capital.[12]

The use of speed dialogues as a framework for democratic deliberation has been put forward above only for the plenary body. For smaller groups within the IGF such as the multi-stakeholder bureau, dynamic coalitions and open fora, different techniques may be required. (For present purposes, the plenary body in open consultation can also be considered as one of these smaller groups, since the number of members in attendance is of approximately an order of magnitude smaller than at the annual plenary meeting.)

Provided that it incorporates at least the three main features of the deliberative democratic frameworks identified above—the use of background briefing, the guidance of a moderator or facilitator, and the pluralism and equality of the group—there is no need to be prescriptive of the precise method by which democratic deliberation is institutionalised within the IGF’s smaller subcommittees.

In particular, insights from models of small group democracy such as that of Gastil,[13] those from the study of deliberative democracy such as the citizens’ jury,[14] and those of consensus in small groups such as the Consensus Workshop,[15] are all potentially applicable. It is a feature of the latter that defining the process to be followed forms the group’s first item of business; although to bootstrap the group into a form capable of deliberating upon its own processes (let alone anything else), these must at least initially be specified by hierarchical means, such as a constitutional document if one exists, or by its chair, or through standing rules previously established by the group.

Without detracting from this latitude on matters of detail, some broad guidance for the processes to be adopted by the open consultation meetings, the workshops and dynamic coalitions, the open fora and the multi-stakeholder bureau does flow from the findings already made, particularly given that the structural relationship between these bodies requires their deliberative procedures to be coordinated to some degree. Taking these in turn: Assessing consensus

One of the roles of the bureau that has received only cursory attention so far is that of assessing the consensus of the plenary body, along with that of dynamic coalitions or workshops submitting reports on their activities. Part of the difficulty of this endeavour lies in the lack of a universally accepted measure of consensus, given that even within the Internet governance regime, working definitions range from “general agreement” as in the case of APNIC,[23] to unanimity (though allowing for abstention) in the case of the W3C.[24]

For present purposes, the definition of consensus from Johnson and Crawford that was adopted in Chapter 4, and that is broadly consistent with deliberative democratic theory, will be accepted as appropriate for the IGF: that “opposition to a particular policy is limited in scope and intensity (or is unreasoned), and opposition does not stem from those specially impacted by the policy.”[25]

It is noteworthy that the application of this definition to the IGF would not require full consensus or unanimity. This limits the capacity of individual stakeholders with limited interests in an issue to exercise disproportionate power over the decision-making process, whilst still preserving the rights of stakeholder groups to veto any measure that is against their interests as a whole, through the institutionalisation of this right within the bureau. At the same time, the definition looks not only to the number of stakeholders in dissent on a particular issue, but to the directness of their interest and the strength of the reasons for their position.

Although conceptually sound, the definition is very much a subjective one, which raises the prospect that the bureau might find consensus where it wishes to find it, and fail to find it where it does not. The appropriate response is to build in mechanisms of accountability to counteract this risk.[26] Some possible ways of doing so can be drawn from the experience of the use of consensus in other organisations of Internet governance. Two of those examined in Chapter 4 were ICANN and APNIC, which are also amongst this chapter’s examples, and which will be revisited again here in turn.[27]

It will be recalled that ICANN assesses the consensus of the community upon proposed new policies by means of a formal Policy Development Process. Taken as a process intended to facilitate the formation of consensus, the PDP does not provide a successful model for the IGF to follow. However what it has achieved, by requiring that the level of consensus upon any proposed new policy be fully documented, is to address criticisms that ICANN’s board had previously ruled upon the existence of such consensus without any adequate factual basis for doing so.[28] The principle that may be drawn from ICANN’s example is therefore that the most transparent and accountable way for the achievement of consensus to be assessed is through a process that is formally and openly documented.

An alternative and contrasting model is that of APNIC, which is a much more culturally homogeneous organisation than ICANN with a less tumultuous history. As described above, its process for establishing the achievement of consensus on a new policy is predicated upon the subjective judgment of the Chair of the Open Policy Meeting, but this is subject to various checks and balances: that the proposal first have been developed within the relevant SIG, then tabled four weeks ahead of the meeting, then achieved consensus at that meeting both amongst SIG members and in plenary session, then survived an eight week comment period, before finally receiving the majority approval of APNIC’s Executive Council.

The most appropriate model for the assessment of consensus by the IGF’s multi-stakeholder bureau is likely to be a hybrid of those of ICANN and APNIC. Like ICANN (only more so) the IGF is politically and culturally heterogeneous, and therefore the thorough documentation of any consensus claimed to have been achieved by its plenary body at an annual meeting or open consultation will avoid the same suggestions of partiality being made of the bureau that some have made of the Secretariat and Advisory Group.[29] The same applies to dynamic coalitions, who as noted above, ought in like manner to document the consensus that they have reached on any recommendations being forwarded to the bureau for presentation to the plenary body.

Furthermore, as in APNIC’s case, it is also desirable for the achievement of consensus to be confirmed at more than one level. The procedures already put forward in this chapter ensure that this is so, as any consensus of the plenary body will be reconfirmed by consensus of the stakeholder groups within the bureau before it becomes a formal recommendation of the IGF. Similarly any consensus of a dynamic coalition (if accepted as such by the bureau) will have to be confirmed by the plenary body, before being reconfirmed by the bureau if it is to become part of the IGF’s formal output.

The example of APNIC also teaches that any consensus reached at a plenary meeting should remain subject for a short period to the input of those who were unable to participate in that meeting. This practice will be discussed further when considering the ways in which to accommodate online participation in the IGF’s processes.[30]

On this basis, a suitable initial process for the assessment of consensus by the multi-stakeholder bureau (subject, of course, to refinement through open consultation) would incorporate the following elements:



See Section


See Section 6.4.


See Section


See Section 6.4.4.


See Section 6.2 and Section 6.3.


See Section


Third World Network, Statement for the Preparatory Session for the Internet Governance Forum (2007)


Though for present purposes, processes for online deliberation will be left aside, to be revisited under the heading of inclusion at Section 6.4.4.


See Section and compare Section


See Section, Section and Section


See Section and Section


See Section 4.3.2.


See Section


See Section


See Section


See Section


See Section


See Section


This option will be explored further under the following heading.


See Section


See Section


See Section


See Section


W3C, Process Document (2005), para 3.3


Johnson, David R & Crawford, Susan P, Why Consensus Matters: The Theory Underlying ICANN’s Mandate to Set Policy Standards for the Domain Name System (2000)


See also Section below.


See Section 4.4.3.


See Section 2.1.3.


For example, by the Third World Network at the February and May 2007 consultations.


See Section 6.4.4.