Version classiqueVersion mobile

Normative Experience in Internet Politics

Françoise Massit-Folléa
Cécile Méadel
Laurence Monnoyer-Smith

Chapter 4. Towards a Typology of Internet Governance Sociotechnical Arrangements

Romain Badouard, Francesca Musiani, Cécile Méadel et Laurence Monnoyer-Smith

Texte intégral


  • 1 However, for the purpose of this paper, we define openness not as the core feature of the Internet (...)

1The realm of Internet governance (IG) can be read today as an increasingly complex nexus of contents and processes. Constant evolutions in the technical and political architecture of the Internet, and the growing number of actors willing to mobilise in relation to Internet governance issues, only make the articulation in this nexus more complex. Involved stakeholders are not only more numerous, but more heterogeneous in nature, size and scope, ranging from individuals to organisations, from lay participants to professional actors and from institutional to private entities, thereby raising new challenges to traditional mechanisms of international cooperation and decision-making as well as to national policy. Moreover, most organisations and entities involved in the production of norms within the IG realm boast a high level of openness1 and inclusiveness, as a heritage of the web culture. The consequent variety of actors and perspectives needs to be taken into account in any comprehensive discourse on IG-related arrangements, whether the focus is on their internal procedures or on the relations between them.

2This study addresses the proliferation of socio-technical arrangements within the IG realm and aims at clarifying their attempts to include new actors in the decision-making processes related to the Internet as a public good. To do so, it proposes a typology of Internet governance arrangements. By means of a number of relevant test cases, we will elaborate a classification of ideal-types of procedural behaviour, identified as having relevant features or traits in common, and detail some of their institutional and cultural specificities. This typology is proposed as a methodological tool for future research in IG, with the aim of evaluating the suitability of the procedures with respect to their outcomes. However, it should be noted that it does not intend to address the general issue of Internet multilevel governance, and the articulation between those heterogeneous actors.

Theoretical foundations

  • 2 For the purpose of this study, “norm” is defined as the outcome of a procedure gathering a number (...)

3Internet governance is one of the most recent examples of the creative dimension of governance management in a multi-polar world, where many actors claim to have their say in global issues of common concern, along with climate change [Pavan, 2009; Drake & Wilson, 2008]. These actors challenge traditional processes of inter-state cooperation to build international norms2 by demanding participation in a decision-making process which excludes them. A description of the various socio-political dynamics that compose this realm is necessary in order to advance a critical evaluation of the real political significance and impact of Internet governance experiences. This approach holds promise for any scholar interested in governance and multi-stakeholder processes, but also for the actors themselves and their efforts to assess their relative weight and impact on the normative processes they are summoned, or intend, to become a part of.

4The theory of reflexive governance, inspired by Jürgen Habermas and extended by Ulrich Beck on environmental issues [Beck, 1992], has sought to redefine the conditions of norm production in a reflexive society challenged by technological developments. This theory emphasizes the specific role of civil society actors, on the basis that democratic legitimacy in post-conventional society is built through an agreement amongst state and non-state collective actors; those approaches contribute to the functioning of democracy, giving birth to various deliberative processes and experimentations, and modifying the sequence of decision-making. The concept of governance has emerged out of the necessity to describe these new types of actors (private and public entities) and these processes, which intervene in the national and international scene as relevant spokesmen for interest groups or civil society; not to mention that their relevancy could be highly questionable insofar as their legitimacy is the result of the process and not its precondition. Whether this new international governance is a symptom of a liberal turn, which weakens traditional states, or of a new territorialisation of power [Mueller, 2010], which empowers citizens and corresponds to new global stakes, is still a point of discussion among academics.

5In the present paper, we do not wish to discuss the theoretical founding of reflexive governance. Rather, we would like to develop an empirical and methodological tool enabling researchers to address the reality of some of its main claims, such as the participatory dimension of the decision-making process for instance. Rather than addressing the debate with reference to the very concept of “governance”, already the subject of many studies, our focus will remain on the procedural dimension in our attempt to interpret, compare and measure different entities that intervene in global Internet governance. We consider multi-polarity as the critical dimension that puts an end to the monopoly of “traditional” authorities on regulatory activities. The integration of a plurality of actors in decision-making processes allows for less centralised procedures – and in turn, opens the way to an uncertainty of boundaries in the political arena and to a reconfiguration of forms of political engagement and governmental practices. This could be seen as a form of de-politicisation of central authorities and a politicisation of civil society actors involved in decision-making processes.

6In addition, the concept of governance is made operational by means of a set of analytical tools. Today, a wide variety of organisations claim that they have a role to play in Internet regulation, and each of these organisations represents a specific governance model, setting up its own procedures. This study contributes to the analysis of the different ways in which some actors claiming to be involved in Internet governance address the challenge of multi-polarity by adapting their decision-making processes.

7The approach developed in this study is therefore primarily methodological: while the multi-faceted and distributed realm of Internet governance has often been seen as too problematic to analyse systematically, we will develop a set of analytical categories and tools enabling a codification, and comparative study, of norm production processes within the field.

8This methodological approach is linked to the work of post-Habermassian scholars aimed at measuring the quality of debates within deliberative procedures [Stromer-Galley, 2007; Janssen & Kies, 2005; Trenel, 2004; Monnoyer-Smith & Wojcik, 2011], but detaches itself from this body of literature inasmuch as its focus is on formal procedures rather than on the quality of interactions within them. More closely related to our study is Fung’s work on patterns of participation in complex governance processes [Fung, 2006]. The three major dimensions in Fung’s typology of deliberative processes (selection of participants, communicative interactions between actors, and level of authority of the participants) show a direct link with the dimensions of the ideal-types proposed in this study.

Hypotheses and method

9The main hypothesis underlying our work is that procedures in the multi-polar IG realm possess features that are decisive for the outcome(s) reached and the type of norm constructed. Thus, procedures and outcomes are reciprocally and dynamically influenced, even if local practices differ somewhat from what is suggested by the institutions’ status: i.e. main features of the procedure are likely to influence the features of its outcome, and conversely the nature of the norm explains the creative process engaged to build it. For example, the need to build an open standard to be implemented in various contexts by different actors is likely to generate an open production process in which numerous actors are involved.

10For the purpose of this study, we refer to our test cases as “arrangements”, a Foucaldian notion frequently used in academic literature on participation in order to define specific procedures which involve a wide variety of actors, technical devices, institutions, legal norms, etc… which interact in order to produce a norm. In this paper, the term “Internet governance arrangement” will refer to a specific type of actor (e.g. an organisation, generally open to actors from all over the world), identified as relevant in the IG realm, whose action is defined by a specific procedure (which will produce instructions, standards, relevant issues for discussion, consensus, regulations, collective norms…). We choose “arrangement” rather than “organisation” or “institution” in order to signal their dissimilar status, from a sociological, juridical and political point of view.

11Taking into account these two different aspects allows us to point out not only a specific, dynamic procedure or task among the many different ones that an arrangement might perform or be involved in, each with its own features and actors involved. It also indicates a particular moment across time, thereby resolving possible uncertainties linked to the chronological dimension: a point that is crucial for those devices that underwent repeated reforms during their history, e.g. ICANN [Froomkin & Lemley, 2003].

  • 3 For instance, when we code the integration of the IETF normative process as “open to all without c (...)

12IG arrangements generally claim to be both procedural and reflexive, for purposes of efficiency and legitimacy. The aim of this paper is to work on the formal discourses within procedures, so as to study their normative horizon – thus, it relies almost exclusively on formal documents issued by the organisations. Nevertheless, it goes beyond its scope to address the gap between this normative horizon and the corresponding implementation and practices, and therefore, the common contradiction between the public display of rules and their appropriation by actors3.

  • 4 On the definition of procedures, see [Callon et al., 2009].
  • 5 See [Malcolm, 2009].

13The nine arrangements analysed in the paper have been carefully chosen according to the diversity between them – and thus the extent to which their features might be found in, or assimilated to, other features in other arrangements. Aspects taken into consideration were4: purpose of the arrangement (technical coordination, standards development, public policy recommendation)5, geographical scope (regional, national, international), kind of outcome foreseen (standard, norm, law) and actors involved (institutions, citizens, civil society organisations, academic experts, private sector). The procedure selected for examination within each arrangement is deemed as particularly representative of the activities of that specific arrangement.

14Our first analytical grids included elements such as the modalities of selection of actors, the description of their involvement, the ongoing flexibility of the format, the ex post accountability of the procedure. The three steps we have ultimately defined as crucial in the process are the answer to three different questions, aimed at assessing the suitability of the procedure vis-à-vis its outcome:

  1. What is the link between the variety of actors and the extent of their intervention?
  2. What is the link between the choice of a procedure and the kind of norm it generates?
  3. What is the link between the assessment given by actors and the public6 on the procedure, and the actual reflexivity of the device?

15The nine arrangements selected to start the typology are the following:

  • World Wide Web Consortium (W3C): the elaboration of a technical web standard,
  • Internet Engineering Task Force (IETF): the elaboration of a technical standard,
  • Association Française pour le Nommage Internet en Coopération (AFNIC): the elaboration of organisational policies concerning country domain names,
  • Internet Corporation for Assigned Names and Numbers (ICANN): Generic Name Supporting Organisation’s (GNSO) policy development process,
  • Forum des Droits sur l’Internet (FDI): national (French) establishment of a venue for dialogue between relevant stakeholders (closed down in December 2010),
  • Internet Governance Forum (IGF): worldwide annual plenary session for dialogue between stakeholders on a relevant issue,
  • European Network and Information Security Agency (ENISA): establishment of a Working Group on a critical issue for European network security for recommendation to the European Parliament,
  • -International Telecommunications Union (ITU): elaboration of a technical standard.
  • -Internet Society (ISOC): implementation of a programme within a broader initiative.

16This sample offers enough cases to give an idea of the variety of procedures used in IG, and to embrace the diversity of the work undertaken by these actors. However, generalisations made out of the specific cases studied are to be handled with caution; the test cases give accurate examples of the features composing the types we propose, but do not represent ideal-type models in themselves.

17We have also decided not to distinguish in the list of arrangements – and therefore in the analysis – between governmental and intergovernmental institutions as two different types of actors; nor do we make a distinction between public and private external mechanisms of control on the choice of the procedure. Like other choices, these were the result of keeping a balance between a proper diversity of features and the feasibility of the framework construction.

18To flesh out (at least some) features of the procedures of norm production, we focus on their three key moments:

  • Integration: the moment when a certain number of actors and groups of actors gather around a particular IG-related issue.
  • Normative process: the moment when the format of the procedure is chosen and the norm is produced (the degree of codification and enforcement of the outcome are defined).
  • Assessment: the moment when the norm becomes operational, and the procedure is assessed by the actors or the public at large, through various possible means and tools.

19The choices concerning on the one hand the dimensions of the process we focussed on, and on the other hand the sample we selected as representative of the diversity of arrangements in the IG realm, are the result of ongoing tests of the suitability, flexibility and completeness of our analytical framework. The graphs resulting from our categorisation are a first step towards the establishment of meaningful relationships between the relevant dimensions of each step. To build the graphs, we have attributed numerical values to the criteria that compose the different dimensions, summarized in Table 1. We will define the criteria and explain how the axes are constructed for each dimension in the following dedicated sections.

Table 1. Overview of dimensions and criteria composing the typology


20The first graph illustrates the degree of integration of the analysed arrangements: the moment when a certain number of actors and groups of actors gather around a particular IG-related issue. We assess the degree of integration by means of two criteria: variety, the number of actors participating in the process; and openness, their condition of access to the process.


21Openness is defined here as the way in which the integration of actors is linked to more or less stringent modes of selection. Five levels of openness are singled out:

  • Designation by external mechanisms: a supervising unit or a secretariat, not involved in the rest of the procedure, selects the actors or groups of actors to be included in the procedure.
  • Cooptation: actors are included in the procedure thanks to the vote or selection made by already existing actors or members.
  • Open with formal conditions: actors are included in the procedure according to formal requirements (explicitly stated in a founding document or chart and/or enforced), e.g. a specific technical expertise.
  • Open to all with monetary conditions: actors are included in the procedure depending on payment of a membership fee (considered for the purpose of this analysis as a weaker condition than possessing a specific technical expertise).
  • Open without conditions: actors can join the procedure without needing to provide any specific credential or satisfy any specific requirement.

22The criterion of variety refers to the diversity of actors/groups of actors that are present or can participate in the arrangement. Five categories of actors are identified for the purpose of this study:

  • Private sector: enterprises and companies run for private profit, not controlled by national or international political entities.
  • Institutions: governmental and intergovernmental political structures.
  • Citizens: Individuals acting in their name.
  • Civil society organisations (CSOs): Legally-constituted organisations, including Non-governmental organisations (NGOs).
  • Experts: academic or independent researchers providing specific technical or other expertise.

Table 2. Overview of Integration criteria for each arrangement.


23Graph 1 links the variety of actors represented in the device (y-axis, from 1 to 5 according to the number of categories of actors present) to their mode of selection (x-axis, from 1 being the least open – external designation – to 4 being the most open – open without conditions).

24The more an arrangement is located to the upper-right side of the graph, the more the procedure possesses a high level of integration capacity (i.e. gives space for a large inclusion of actors).

Chart 1. Integration.

25The level of openness and involvement of actors is generally high among the examined arrangements. More than half of them involve more than three groups of actors, and the conditions of involvement as set in the formal documents by the actors themselves appear to be quite flexible. Low variety appears to be specific to the most “technical” arrangements (ENISA, ITU and W3C), as if technical competencies rely more on cooptation than the political or multi-approach ones; but IETF offers an interesting counter example.

26The most open arrangement seems to be the IETF process of technical standards development. All types of actors can be a part of it, and they can integrate the procedure without conditions. The IETF process is organised around several working groups (WGs) whose members are connected by mailing lists. To enter one of the WGs, the only condition is to subscribe to the proper mailing list. The subscribers will then integrate the exchanges between members regarding the issues debated inside the WG. After the debate, if it reaches a consensus, it could become a Request For Comments (RFC), i.e. a technical standard awaiting validation by the whole IETF community. So individuals or organisations from civil society, the private sector, or academic research institutions can subscribe on a list and become involved in IETF standard development. As mentioned, however, openness and inclusiveness might be more allowed and encouraged than actually take place.

27Inversely, ENISA and ITU appear to be the least open and inclusive arrangements among those examined. ENISA is an agency established in 2004 by the European Parliament with the aim of improving information and network security in the European Community, and unlike IETF or W3C, its procedures require the establishment of working groups (WGs) of experts, to address emerging issues in network security. However, and again unlike IETF or W3C, the groups are created ad hoc by the external Permanent Stakeholders Group, which also has the responsibility for identifying the relevant issue the groups must deal with. ITU, an agency of the United Nations since 1947, is generally regarded within the Internet community as a complex, hierarchical and highly bureaucratic organisation, features unsuited to its mission of technical standardisation; the participation in the arrangement is open only to representatives of governments, and to some private sector bodies (with associate membership subject, depending on the countries they are located in, to the approval of their national governments). This makes ITU the only arrangement for technical development that is primarily hierarchical.

28In conclusion, it appears that the correlation between the modalities of integration of actors in the procedure and the purposes or goals of the arrangements is weak or absent. The three technical standardisation bodies – W3C, IETF and ITU – bear very different characteristics, both in terms of conditions for participation and of variety of actors involved, varying from total openness to highly hierarchical, government-restricted integration. The same conclusion can be drawn when comparing arrangements coordinating domain names, ICANN and AFNIC, and the venues primarily aimed at creating and supporting dialogue on IG, such as IGF, FDI and ISOC.


29The second graph represents the dimension of the normative process, i.e. the phase in which the format of the procedure is chosen and the norm is produced and made operational, which includes the definition of the degree of codification and enforcement of the outcome.


30The format of the procedure is defined as the act of choosing and setting the specifications according to which the procedure will function. Four different ways of setting this format are identified:

  • External definition: a supervising unit or a secretariat, not involved in the rest of the procedure, defines the format of the procedure.
  • Internal definition with external constraints: the actors included in the procedure (or a subgroup of them) define the format of the procedure, but ground this definition in guidelines or constraints received by an external unit.
  • Internal definition: a subgroup of the actors included in the procedure (selected by the actors themselves) defines the format of the procedure.
  • Collective internal definition: all the actors included in the procedure define its format collectively.

31The criterion production of the norm refers to the extent to which the production of the norm is carried out collectively. Three levels are highlighted in the graph:

  • Independent units coordinated ex post: a number of working units (e.g., study or working groups) carry out the norm production and work independently; the different contributions are then harmonised/ made a coherent whole by a separate, ex post coordinating body.
  • Units working in common by consensus: each working unit, albeit recognisable as separate from the others in its tasks, iteratively seeks the consensus of other units during the entire duration of the normative process.
  • Consensus at large: in addition to the previous one, the consensus of the public at large is iteratively sought during the normative process.

32The impact of the outcome is defined as the degree of codification and possibility of enforcement of the outcome of the normative process. We distinguish three different ways in which the norm can be adopted:

  • Recommendation/proposition: the outcome is agreed upon and codified in written form, but has no legal value nor is it enforceable.
  • Adoption with reservation: the outcome is agreed upon and codified by its intended recipients, but its enforcement depends on subsequent actions or debates.
  • Imposition: the outcome is codified and enforced by the appropriate internal or external authority.

Table 3. Overview of Normative Process criteria for each arrangement.


33These three criteria are linked in a hybrid histogram/curve graph (the curve being the third dimension). Graph 2 compares each arrangement analysed according to the format of procedure chosen, the norm production itself, and the impact of the outcome.

Chart 2. Normative Process.

34This graph permits both intra- and inter-device comparisons, and makes it possible to identify those devices for which a high degree of collectiveness in procedure choice and norm production does not correspond to a high impact on decision – i.e., from a point of view of democratic construction of norms, for which the potential does not deliver.

35For instance, we can see that the format of the procedure and the way the norm is produced within the W3C and the IETF are based on the establishment of a consensus among the different actors, and the link between production and decision remains strong. Instead, while the way the norm is produced within the IGF is also supposedly based on reaching as large a consensus as possible, the link between production and decision is quite weak. In this last procedure, the involvement of actors and the establishment of a consensus produce less codified outcomes in terms of impact on the decision, while a larger correspondence between the two aspects can be observed for those actor/procedures in the graph that relate to the production of technical standards.

36In conclusion, this graph seems to show an inverse relation between the impact of the norm produced within the arrangements, and the strength of their link to the political sphere. For example, the IETF and the W3C exhibit very weak links to the political, but succeed in producing norms with a strong impact, while establishing and maintaining a collective production of such norms. Norms produced by the ITU and the ENISA, arrangements bearing a political structure and producing a technical outcome, have a degree of impact that is relatively weak, while their normative process is not collective.

37Thus, it appears that the impact of the outcome is not correlated with the extent to which the production of the norm is collective. Each case has its own combination between collectiveness of the process, format of the procedure and impact of the outcome.

38The determining factor in this appears to be the link of the arrangement to the political sphere: the arrangements producing expertise or advice destined for political bodies have a weak impact. The reason for this might be that the political sphere does not recognise in these arrangements a decision-making power in its own right. A delegation of power, or a more inclusive re-definition or reconfiguration of power arrangements, does not actually take place, with “traditional” authorities still holding control of procedures that “matter” in terms of impact strength. The graph might imply, as a consequence, the reluctance of traditional powers to give up some of their prerogatives and authorise a delegation of power – in short, reluctance towards multi-polarity. This point will be considered again in the concluding remarks.


39The third graph represents the dimension of assessment, i.e. the phase in which the actors of the normative process and/or the public at large provide feedback on the procedure that has originated the norm (now operational), and this feedback is taken into account within the arrangement. We aim to understand how, what and who can provide a feedback and how this feedback is taken into account.


40The feedback criterion seeks to define what parties, direct or indirect, of the arrangement are allowed to provide through official means a feedback or assessment on the arrangement itself. Three different cases are identified:

  • No feedback allowed: No official channel for feedback is provided by the arrangement.
  • Feedback by actors allowed: The possibility of feedback is officially provided only to the actors that have already taken part in the normative process.
  • Feedback by actors and public allowed: The possibility of feedback is also officially provided to the actors that have not been directly involved in the normative process.

41Accountability is defined as the modality in which the arrangement itself takes into account and re-elaborates the received feedback in its subsequent endeavours. Four levels are singled out:

  • Not formalised: There is no tool or procedure, official or unofficial, for consideration of the feedback.
  • Advice/debate: There are tools that allow the feedback to be debated by internal and external parties to the arrangement.
  • Recommendation: The feedback is codified and submitted as a specific recommendation to the arrangement.
  • Request: Consideration of the feedback by the arrangement is formalised and granted by an official procedure, instrument or device.

Table 4. Overview of Assessment criteria for each arrangement.


42We link these two dimensions in Graph 3, relating the feedback (from 1 being the weakest degree of feedback to 4 being the strongest) to the accountability (from 1 being the weakest degree of accountability to 4 being the strongest).

43The more an arrangement is located on the upper-right part of the graph, the more the possibility of assessing the procedure properly is granted, as the arrangement is both transparent and receptive to the advice coming from both the actors and the public.

Chart 3. Assessment

44All the arrangements considered allow at least partial feedback through official means; wider differences among arrangements are to be found in the ways in which this feedback is re-elaborated.

45The arrangement that scores best on both dimensions is, once again, the IETF. During the development of technical standards among the IETF, actors receive constant feedback. After the consensus is reached in a WG, the chair exposes the outcome to the other WGs, posting it on their mailing list. Any member can expose his argument in order to invalidate the work done by the WG. If the whole community judges that the argument is relevant, then the WG will have to fix the problem raised. When the community approves the work done, then the outcomes are published in the form of a RFC (Request For Comments) on the IETF website. If any individual interested in the issue wants her voice to be heard, then she can subscribe to the relevant mailing list to expose her argument, and the work done is again the subject of discussion. The adaptability of the development of technical standards among the IETF appears to be granted, since the actors involved and the public receive a feedback during and after the procedure. The actors, like the public, can express their view on the norm produced, and the consideration of this feedback is formalised and granted the IETF status, i.e. the way the norm is produced can evolve according to the feedback from the actors and the public.

46In the other body for standard development that we have inserted in the graph, the W3C, the return to the actors is granted once the work on a recommendation is completed, but it is less clear how the returning accountability is established. It should, however, be taken into account that there is an important dimension of ongoing merit check during the definition of the standard itself, which might partially account for this lack of clarity.

47The arrangements whose mission is the coordination and management of critical resources are, yet again, scoring relatively poorly in terms of accountability. This might be traced back to their developing a hybrid way between technical and political, or maybe to the fact that their target of regulation is more specific in nature, time or space, a “microcosm”.

48Thus, this graph also allows a conclusion in terms of multi-polarity at the specific level of assessment. With the exception of the FDI, the arrangements close to the traditional political spheres (whether they are primarily technical or not) generally fail to give any guarantees that feedback will be taken into account, unlike those arrangements having absent, weak or more ambiguous links to the political spheres, and whose target of regulation are what we have called microcosms.


49This section presents a fourth, and concluding, graph regrouping the three indicators expressed in the previous sections into a single indicator, with the aim of obtaining information about every arrangement’s performance vis-à-vis the levels of integration, normative process and assessment. We first present the data and calculations on which the graph is built; then, the observation of this graph allows us to draw an overall typology that simultaneously takes into account the different moments of norm production.


50For each of the three moments of norm production (integration, normative process and assessment), we sum the values of the different criteria, in order to obtain a general figure. For the “integration” moment, we sum “openness” and “variety of actors”, i.e. the resulting figure takes into account the barriers to entry and the number of actors involved: the higher the figure, the better the performance of the arrangement in terms of integration.

51For the “normative process” moment, we sum “format of the procedure”, “production of the norm” and “impact of the outcome”, i.e. we take into account not only the degree of participation in the normative process, but also the impact of the outcome produced. Here, the higher the figure, the better the arrangement has succeeded in combining a high degree of collaboration during the normative process with a strong impact of the outcome.

52For the “assessment” moment, we sum the “feedback” criterion with the “accountability” one: the higher the figure, the better the arrangement combines the listening of the feedback given by a wide variety of actors with an official way of taking it into account.

53The three figures obtained for each arrangement are then expressed in the form of a percentage, by comparing it to the maximum possible figure that can be obtained.

  • 7 It seems fit to recall here that the criteria are defined according to formal documents authored b (...)

Table 5. Overview of performance evaluation for each moment of norm production and arrangement.7

Concluding discussion

Chart 4. Overall performances

54Three ideal-types of arrangements can be identified based on the indicators contained in Graph 4, and for the available cases.

55The expert type. This type is characterised by a weak level of integration of actors in the process, but a medium or high degree of collectiveness in the work among concerned actors during the normative process. The way the norm is produced has a high impact on the form assumed by the norm itself. The opportunity for feedback is granted to actors (albeit not to the public), therefore closely concerned by the assessment phase as well.

56This type is the less “collective” at all moments of norm production. However, a consideration of what arrangements fit the profile (ENISA, W3C, ITU) may indicate a suitability of this feature to their main purpose: i.e., the production of technical norms for a specific audience, or more generally, the regulation of specific domains in which the influence of political authorities might be on who participates in what, but cannot entail the final outcome (for the ENISA, network security recommendations for political bodies; for the W3C, technical standards for enterprises of the Web, for the ITU, technical standards at the frontier between the Internet and telephone networks). This type shows the reciprocal influence between normative process and nature of outcomes hypothesised [supra].

57The open type(s). This type’s core feature is a widely inclusive gathering of actors in the integration phase. However, from this departure point there are two subtypes, depending on how the high collectiveness granted in this first phase translates effectively into the following two moments. For a first subtype (represented by the IETF), the collective work performs effectively and determines a strong impact on decision and assessment; for a second subtype, including the FDI, the ISOC and the IGF, the initial openness does not translate into the production of a norm with a strong weight, or into an inclusive and noteworthy assessment. The reciprocal influence between normative process and nature of outcomes can once again be observed here, in an especially interesting way: since the very creation of the FDI and the IGF is the answer to a request by some groups of actors to be more involved in IG decision-making processes, their main purpose seems to have remained the mere involvement of these actors, but the subsequent (and more important for the actual decision-making) steps do not perform as effectively. Instead, while the IETF is also intended as an answer to all the stakeholders in the domain who wish to participate in standards production, its primary purpose is and remains the production of effective standards, which is reflected in the indicator of performance (as a link between collectiveness and impact on decision) for the normative process. In this case, as well, the difference between the IETF and the three other examples appears to be the relation of the arrangement to its political environment: the world of technical standards is ultimately a self-managed microcosm, while the world of dialogue and debate initiators/facilitators is primarily a political one, trying to take over some prerogatives of traditional powers – an attempt, however, which still has a long way to go before the goal is reached.

58The management type. This type is characterised by a high level of homogeneity in the indicators across the three moments of norm production. It includes procedures that are initially open to all kinds of stakeholders, but exhibit relatively closed ways of involvement (e.g. cooptation); have good performances in terms of normative process and assessment, but do not provide any formal channel nor guarantee for the actors’ feedback to be taken into account. The two test cases that fit this type are the ICANN and the AFNIC, the organisations that undertake tasks of critical Internet resources coordination – issues that on one hand require technical expertise, due to their specificity and complexity, but are, on the other hand, a matter of concern for a variety of actors and for the public at large, thus entailing the necessity of giving as many voices as possible the opportunity to be heard. This is reflected in the strong representation of different actors vs. restrictive modalities of implication during the integration phase, and in the large possibility for feedback vs. the lack of guarantees of its consideration during the assessment phase. Our hypothesis appears to be, once more, confirmed.


59The arrangements involved in Internet governance taken into consideration in this study are generally open and inclusive (at least formally). A wide category of actors is involved in the production process, and the search for consensus seems to be an important way to reach the expected outcomes (with the possible exception of what we have called the “expert type”). This general scenario, however, does not erase the internal contradictions fleshed out by our intermediate discussions. Some procedures, appearing to be open during the actual normative process, are in fact heavily influenced by specific units or categories of actors during the integration phase; others, involving numerous actors and grounding their work in the search for consensus, do not deliver in terms of their collective potential when the impact of the produced norm remains weak, or the feedback provided by the groups of actors concerned fails to be harnessed during following endeavours.

60The variability between arrangements may ultimately be traced back to the kind of norm produced: the impact is much stronger among the development of technical standards, while it is weaker for those devices whose main purpose is political. This indicates a reciprocal link, which is observed in all three ideal-types, between the nature of the norm and the way it is produced: if the procedure shapes the norm, the procedure is also shaped in order to reach specific features related to the nature of the norm, in a process of interactive co-construction.

61Maybe, the crucial drive behind the shaping of IG arrangements is the extent to which the involvement of all interested parties is a pre-condition for the norm to be meaningful and useful. There is a demand by the actors concerned by IG issues to be a part of the process, but there is no formal link established in political processes between the variety and quality of voices heard, and the relevance of the norm – which raises interesting questions concerning the actual democratic potential of such arrangements; inversely, for the establishment of technical standards, the need to involve relevant actors at all stages of the normative process is crucial to the usefulness and usability of the resulting norm.

62In conclusion, it appears that multi-polarity is indeed at work in the Internet governance realm: what this study has ultimately shown is a variance between the statuses of the different poles involved at different degrees in the governance of the Internet. Some of them are interested in specific, more easily definable environments, such as technical standardisation; as this has rarely been the field of authority of traditional powers, these arrangements are filling a gap deriving from the swift rise of new regulatory needs, and can enjoy their position of autonomy in the IG field, even a relatively short time after their entering the IG arena. Other arrangements seek to occupy poles for which competition with classical political authorities is harsher; so, even while they seek large-scale support by the public during the normative process, the impact of the norms they produce is weaker, as those same traditional authorities have, in fact, delegated the amount of power they have to these new arrangements. Ultimately, what our typology shows is, in fact, the set of challenges for Internet governance multi-polarity “at work”.


1 However, for the purpose of this paper, we define openness not as the core feature of the Internet’s technical architecture (along with for example decentralisation and egalitarianism), but as the trait of a political device that entails the possibility of a variety of actors to access and participate meaningfully in the normative process.

2 For the purpose of this study, “norm” is defined as the outcome of a procedure gathering a number of actors – the normative process – and is usually expressed in the form of standards of conduct or behaviour, defined in terms of rights and obligations. Norms are endowed with different degrees of formality and coercion, depending on the nature of the procedures that originate them and the enforcement mechanisms that are envisaged. Norms constitute a reference for future practices in the specific domain they address – this applies to both technical standards, determining new uses, or a socio-political meaning, implying new rules of behaviour.

3 For instance, when we code the integration of the IETF normative process as “open to all without conditions”, we are aware that integrating an IETF working group requires time, money, and a good knowledge of English and that, in practice, people engaged in the process have a specific sociological profile. On this specific case, see [Pouzin, 2009].

4 On the definition of procedures, see [Callon et al., 2009].

5 See [Malcolm, 2009].

6 We define “public” as the set of individuals that are interested in a normative process and its outcome. However, they are involved in the process through non-institutionalised means, as mere individuals, and do not qualify themselves as actors during the process.

7 It seems fit to recall here that the criteria are defined according to formal documents authored by the actors themselves. Thus, these final figures express the normative horizon of the arrangements, not their degree of integration, collaboration and assessment “in practice”.

Table des illustrations

Légende Table 1. Overview of dimensions and criteria composing the typology
Fichier image/jpeg, 515k
Légende Table 2. Overview of Integration criteria for each arrangement.
Fichier image/jpeg, 235k
Légende Chart 1. Integration.
Fichier image/jpeg, 188k
Légende Table 3. Overview of Normative Process criteria for each arrangement.
Fichier image/jpeg, 322k
Légende Chart 2. Normative Process.
Fichier image/jpeg, 163k
Légende Table 4. Overview of Assessment criteria for each arrangement.
Fichier image/jpeg, 255k
Légende Chart 3. Assessment
Fichier image/jpeg, 186k
Légende Table 5. Overview of performance evaluation for each moment of norm production and arrangement.7
Fichier image/jpeg, 137k
Légende Chart 4. Overall performances
Fichier image/jpeg, 211k


© Presses des Mines, 2012

Licence OpenEdition Books

Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search