Ever since Internet use has spread among the public at large, the same controversy regularly flares up: does this new communication device enhance democratic debate? The discussion got a new lease of life with the advent of the blog and more broadly the web 2.0 application which allows web users to express themselves even more easily than before. The Internet, unlike radio or television, places the broadcaster and the receiver on the same plane, hence it is, on the face of it, the ideal tool for a participative democracy wherein the citizen could intervene on a regular basis in the public debate. In this paper, I propose to investigate how this matter was first addressed, as the new technology took off, and what is the current situation. Does Internet duplicate traditional media concentration or does it give a voice to new actors? Does the new cyberspace enhance democratic deliberation or does it further a Balkanisation of public opinions? Is Internet ultimately in symbiosis with the new patterns of civic engagement? Fifteen years after networking became accessible to the public at large, it would be no luxury to take stock.
A virtual Agora or a cacophony?
In the early nineties, the Internet was often hailed as a new virtual Agora . In the first book publicizing the new technology , journalist Howard Rheingold compares at length the Web to the Habermassian public sphere. He saw in it a device liable to revitalise democracy. This political vision of the Internet would be taken up by numerous authors, not least Al Gore, then Vice-President of the United States, in his address to the International Telecommunications Union . This was to be a major factor to recommend the new technology.
However academics observing online community behaviour were soon to dispute this point of view, forums often play host to abuse exchanges (flame wars) as users virulently cling to their own views. For Mark Poster , online debates do not meet the standards of the public sphere, namely of a debate between equals where rational arguments prevail and where a common position is sought. Internet only matches the first feature. Web users can indeed exchange on an equal footing but whether this exchange is reasoned is another matter. The debate does not move towards the elaboration of a common position but rather splinters into many contradictory viewpoints. This splintering is further reinforced by the fact that users’ identities are hazy and shifting. Not only do contributors use aliases and create for themselves a virtual identity but they may change this identity or have several.
This coexistence of identities, which was studied by Sherry Turkle  appears to be one of the major causes of this difficulty online communities find in arriving to a common viewpoint. In real life, the diverse traits of a person are unified by being bound in the same body; when dealing face to face, each speaker perceives the complexity of the other and can bank on that complexity to reach an agreement. By contrast, virtual communities foster the multiplicity of adamant viewpoints rather than flexibility .
Some Communities of interests less homogenous than people think
These academic papers, emanating essentially from psychology and psycho-sociology departments, appear formally to write off online public debates. Yet forums, chats or discussion lists exchanges remain an important Internet activity. Surely, users do not access these virtual spheres for the sole pleasure of insulting each other or assuming another persona! Online communities have been identified by the Internet founders as “common interest communities” . It’s an easier way to find people liable to share such and such interest of ours than in real life. This exchange does not touch on the whole of a person’s life but some elements of their personality relating to some leisure pursuit as well as more private matters such as illness, family issues… The exchange will be intense but restricted to one personality element. We are then in the presence of so called “instrumental intimacy”. These communities, very numerous on the net and often enduring, might exchange experiences or expertise[For a typology of these communities, see Michel Gensollen, "Information goods and online communities" in Eric Brousseau and Nicolas Curien (eds) Internet and Digital Economics, Cambridge University Press, Cambridge (UK), 2007]. They have been studied mainly by economists and sociologists.
The “epistemic communities” are a particularly interesting object of observation as they bring together diversely skilled participants. Free software communities are a fine example of this. They bring together developers, expert users and beginners. The former advise the two other groups, but conversely, the latter provide developers with information on the software’s faults and gremlins. Although exchanges are thus in the main balanced, with the “experts” mediating between the beginners and the developers, the community may yet be confronted to log jams. These communities can only survive if they are regulated. In the instance of Debian, a Linux sub-community that Nicolas Auray  studied, many procedures have been set up, to arrive at a base format for the texts sent, manage messages reporting bugs, enable correction cycles, select aspiring developer… A permanent discussion list, the Debian Policy, is the compulsory channel through which to receive and debate new proposals. Elaborate voting systems are provided. We have here a direct democracy moderated by technological aptitude and the knowledge of past history.
Many other examples spring to mind in the field of Health , reviews of cultural products… In all those cases, communities of interest represent a realm where Internet can operate as a genuine hub for exchange and productive public debate. It is also worth noting that these communities are less homogenous than has often been supposed. Experts and beginners rub shoulders in a constructive manner.
Consumers and Citizens
The discussion on the place of the Web in the democratic debate got a new lease of life after the publication of American law professor Cass Sunstein’s book entitled Republic.com . Two points in this book, which gave rise to extensive controversy, deserve closer attention and I shall address them in turn. First the tendency Internet has shown to model citizen sovereignty on consumer sovereignty; and second the fact that political debate on the net essentially brings together users of similar views.
What fundamentally distinguishes the Internet from earlier media is the way this technology makes it possible to offer the user customised information. Not only can consumers find much more easily a product meeting their exact demand but they can even put together their own newspaper. A concept that IT futurologists like Toffler, Gilder or Negroponte could only dream of in the 80s and early 90s is currently available to web users through personalised portal services such as Netvibes and RSS feeds which make it possible to get constantly updated information. In this case of figure the consumer is sovereign and chooses what he pleases. Cass Sunstein rightly reminds us that political sovereignty is radically different. “Citizens do not think or act like consumers” . Political democracy emanates from a deliberative government. Political choices do not always match a person’s personal interests but those of the community. Public opinion is arrived at through debate, exchange and deliberation.
Conversely, many 90s Internet thinkers, whose theories have often been close to those proposed by the libertarians, think that political choices should increasingly be managed like economic choices. Viz Louis Rossetto, the founder of Wired magazine, a major source of critical thinking on the web, who thought that the new information network would permit the suppression of the state education system, leaving each family free to find the education set-up they thought best suited to their children.
But these provocative views found a more sophisticated expression in a 1996 article published by two jurists under the title “The new Civic Virtue of the Internet”. For David Johnson and David Post, the crucial difference between cyberspace and ordinary space is the absolute mobility of its users. They can change site in no time at all, thus controlling the information providers’ power. This mobility, rather than voting is now the vector through which people can express collective preferences. “In the online context, the check against sysop tyranny is not "one person, one vote" but, rather, the ease of exit. And there is reason to believe that the combination of decentralized rule-making by means of (1) the unilateral actions of sysops to define online spaces and (2) the unilateral decisions by users to join or leave such spaces will arrive at a good solution to the collective-action problem .” We have thus arrived at a framework where the citizens’ opinion is expressed in the same way as the consumers’: through the dismissal of the product or website which has ceased to suit. Public opinion is not achieved on the basis of a range of debates and exchanges; instead, at a moment’s notice citizen-consumers can vote with their feet or more precisely with their mouse.
This thesis, much in tune with early Internet Utopias, was robustly disputed . Several authors stressed the fact that this was tantamount to removing all State regulation. In his answer to his critics, David Post clarified his position on regulation: “We don’t need ‘a plan’ but a multitude of plans from among which individuals can choose, and ‘the market’, and not action by the global collective, is most likely to bring that plenitude to us.” 
At a time when Internet is much more regulated by governments, this debate may seem a thing of the past but it is worth quoting in so far as it shows that there was indeed, at the onset of the web, a distinct blurring of consumer and citizen roles, one which has endured, as Sunstein correctly appreciated.
Communication Medium and Democracy
The second threat to democracy Sunstein anticipates from the web lies with the fact that political debates are liable to take place between people of similar views. This thesis which has been partly echoed by Azi Lev-On and Bernard Manin , has, unlike the former, given rise to extensive discussion.
Nobody disputes the fact that, thanks to the Web, citizens potentially have access to richer information than before, and that they can take part in many debates. But opinion is divided on whether users turn only to sites or forums representing opinions close to their own or on the contrary are given the opportunity, thanks to the net, to confront opinions differing from it. In other words does the Internet hold back or reinforce deliberative democracy?
For Sunstein, the wellspring of democratic public expression is free speech in the parks, demonstration on the public highway. In this way citizens find themselves unintentionally confronted to viewpoints other than their own; they become aware that other views exist. The other democracy-founding constituent is shared experience, which provides some sort of “social glue”. The major media focus public attention on a few showcase programmes in contrast with the Internet, which is liable to balkanise political discourse. There, Sunstein converges with a media sociology trend that has taken a close interest in “media events”. These are to be read as symbolic acts ritualised by means of ceremonial dramaturgy, the object of which is to enable society to restate its allegiance to its fundamental values .
Contrary to these mass media, Internet is not a player in these great collective rituals. Truth to say, the nature of these rituals is changing, though not as a result of the emergence of the web. For the great television debates that governed public life are increasingly replaced by “infotainment” programs. Are the many talk shows, where politicians show up more and more often, to be considered essentially as places where political issues are shun? Or should they be seen on the contrary as the occasion to glimpse at the candidates’ personal qualities, today a key factor in political choice? Or isn’t it the case that, anyway, as television programming discriminates less and less between the private and the public spheres such an evolution is inescapable ?
As always at its onset, and before it goes on to modify society in due course, a new technology fits in with existing developments. In a world where private and public talk get more and more intermingled it is hardly surprising to find Internet busy offering its users the interconnecting of their exchanges. The web also comes about against a background of major media diversification, when the punter is faced with a much broader offer than hitherto.
In order to assess how web users respond to this proliferation of public expression, it behoves to scrutinize the way this information realm is structured. Is it concentrated or splintered? Is it open to new actors? Are there connections between sites?
The concentration of online information
The Web proffers in huge quantity information both rich and profuse. A study on the 2005 European Constitution referendum for instance, numbered more than 300 websites addressing the question. Similarly a 2003 American survey arrives at 1700 sites treating of the abortion issue . It will be clear that comparing the Web to mainstream media is not straightforward since the offer is of a different type. For, unlike press or television material, web information remains available forever. Suffice it to say here that the Internet has today become a major media contender.
Does this wealth of online information breed a wide-ranging diversity of information intake? Not remotely: citizens focus on a few sites hailing from the traditional media world and deemed to provide diversified and quality information. An American survey conducted on the 2004 campaign  shows that about half (48.5%) of users consult major media sites (31.5%) or Internet portals (17%), with CNN.com alone netting 20% of the demand. This concentration of the online readership has grown a great deal since, in 2000, major media sites and portals only pooled a quarter of the public.
This growing concentration of the readership can partly be put down to a specific Internet device, that of hyperlinks. For many web users get to a site via links. The American survey quoted above studied web use on six political issues  and shows that the main site gathers, depending on the subject, between 7 and 53% of the links set by other sites, the 10 main sites gather between 36 and 38% and the first 50 between 70 and 95%. We thus arrive at a statistical break down known as Power Law whereby a tiny number of sites soak up the bulk of the links and a very large number get very few.
This study further confirms the key role played by search engines in retrieving information for users. It is a well-know fact that Google classes the sites suggested in response to a search according to the number of hyperlinks leading to these sites. The authors of this survey rightly speak of Googlearchy to typify the phenomenon they have observed. Coincidentally, they had observed in another piece of research that there was a strong correlation between the number of links leading to a site and the number of visits. Ultimately, and contrary to what early theorists thought, the web, like other media, fosters high concentration. This “winner take all” situation has been spotted by a number of economists who see in it the explanation to the success of such high tech enterprises as Amazon and Facebook.
An Open Space
Does this concentration of websites imply that the Internet is no different from other media? Obviously not. Internet is a space where it is easier to produce information than anywhere else, where the obstacles to entry are lower. Many opinions are voiced on the net, which had not found an outlet in the mainstream media, or if so, with great difficulty. This was the case during the European Constitution referendum. Whereas the “yes” camp had ready access to the mainstream media, the “no” camp, which, for the most part, did not belong to the leading political organisations made extensive use of the web to represent their views. According to Guilhem Fouetillou’s research , two thirds of the web sites dealing with the referendum supported the “no”. In this instance, the Web became the obvious communication channel for those ill served by mainstream media. Coincidentally, the web is also laid open to minority - nay fringe - opinion trends, notably negationist and racist of different shades. The same polarisation around a few sites that is almost consubstantial with the web can be observed in these realms.
Beyond the political representation of existing groups (media, political or ideological organisations…), the Web also gives individuals the possibility to voice their opinion via self-publication sites or blogs. Contrary to what has sometimes been said, political blogs amount to little more than a fraction of the Blogosphere, at best a few tens of thousands of the two millions active blogs in France. In the typology of blogs and their public established by Dominique Cardon and Héléne Delaunay-Teterel , the political blog makes up a specific category which distinguishes itself by the fact that, unlike other blogs (diaries, teenagers mags intended for their gang), the proposition is detached from the person of the proponent whose identity is known; the object is to exchange views in the public sphere (frequently mentioning messages from other blogs in their own) directing to numerous external links, in particular to media or political party websites. Finally the blogrolls, that is the list of links towards favoured blogs is longer. We have left a type of discourse never divorced from the person behind it and aimed at a restricted group to enter the public sphere. Reader’s comments, which are displayed, like in all blogs are more numerous and longer 
A closer analysis of political blogs  reveals different categories of authors: national or local political figures, professional observers (journalists operating within their paper’s framework, independent journalists, communication advisers), or ordinary citizens. The object of the blog may be to comment on political life, to promote a political figure or organisation, to rally citizens in the context of elections or social problems. Political blogs evince the same concentration of links and readership as had been observed regarding websites . Must we infer from these several observations that political blogs are rather a minority interest involving only a few citizens and with a very concentrated readership? The influence of blogs on public debate is rarely direct; it is most of the time relayed by major media. An American study  shows that the best-known blogs are read and quoted by journalists. They have the power to challenge or indeed to influence the agenda. The Dan Rather episode, when the CBS star anchor’s false reporting of GW Bush’s military history was exposed by blogers, bears this out. Blogers have become the new newsmakers. Blogs produced by professionals, especially journalists, exhibit an entirely novel type of writing. Donald Matheson refers to it at as "knowledge-as-process" as opposed to "knowledge-as-product" , in which the author states his varied sources, discusses them, offers many hyperlinks. Thus this information model challenges its reader and invites comments and debates.
Mapping the informational online space
The range of websites or blogs research mentioned here shows that the online information space is strongly focalised around sites which have become influential, they are the most read, the most quoted, the most discussed. Though, in many cases, they represent existing media or political actors, new actors have also emerged (the example of the “noniste” French sites during the European referendum is a case in point). As ever, innovation has, to some extend, reshuffled the cards. But this topography of the web needs to be studied further. Are these sites, unequal though they be, structured in a fragmented way, by opinion affinities, or do they, on the contrary, smooth the way of web users around the opinions sphere? We are currently short of data on users’ movements in the political realm ; but it is nevertheless possible to study the way publishers steer web surfing, the extra information they direct to, via the hyperlinks they set up for their readers.
Research  conducted on American blogs during the 2004 presidential election  show that although the links are many (14 on average), they are in their majority (91%) turned towards their own side. It is however worth noting that the most read blogs are the least inwards looking (85%). The study mentioned above on the European referendum yields fairly different results since “intracommunity” links are weaker all around: 76% for the “no” sites and 53% for the “yes” side. Their greater openness can be put down to the specificity of this debate. It so happens that two political parties, the Socialists and the Green, split on the issue during the campaign, leaving in place the many links which existed prior to the “referendum divide” . Another idiosyncrasy is owed to the fact that the “yes” sites connected more with institutional and major media sites  broadly favourable to them, whereas the “no” sites had to look to their own for reference information. These very specific features of the debate on the European Referendum thus account for the gap existing between French and American surveys.
All considered, the research on websites and blogs concentration and that on hyperlinks structure yields two results permitting to lay Sunstein’s case to rest. Internet, in spite of its diversity, is focalised around a restricted number of sites. Generalist sites are usually the work of journalists, either as the offshoot of on an existing news provider such as CNN or as an array of press agency dispatches, as is often the case for news sites on major portals, or even in the case of new press ventures such as Rue89 (Street89). In line with non-specialised media, they offer a fairly diversified range of news and comments. However, given a much more manageable initial investment than is the case for traditional media formats, the Internet has also assisted the development of a new, issue driven media. These “issue driven websites” or specialised blogs exhibit sub-communities similarly focalised around a few sites. In this field, which is patently the most innovative Internet feature, the dearth of these sites’ links towards other sections of opinion may represent a genuine risk of public space Balkanisation, of opinion communalisation. But to size up this danger, an other internet feature needs to be scrutinized: that of debates between users which take place in particular in forums, chat rooms or discussion lists . This is by definition where online deliberative democracy is at.
A survey Jennifer Stromer-Galley conducted with the members of three online discussion groups makes it possible to judge whether the exchanges take place within homogenous (like-mindedness theory) or heterogeneous communities. There can be no doubt that Internet lets you meet people who think like you. This is indeed the community of interest theory that is at the core of the web’s development. This demand for contacts with like-minded people seems the stronger as people find themselves isolated in their natural environment (offline). If nothing else, this goes to show that Internet impact cannot be analysed independently from real life context.
Conversely many people surveyed value diversity. They extol the merits of meeting online people who are different from themselves, socially, geographically but also people who don’t think like them. These web users find in this way an audience before whom they can state their views. Sometimes, these divergent opinions might shock or offend them but the encounter is rather seen as an opportunity to clarify one’s views, to fine-tune one’s arguments. Such opportunities are rare in real life. Indeed, Wyatt and Katz’ research on political discussions  show that these tend to take place at home or at work and occur mostly between people with whom disagreements are not frequent.
A generalisation of Stromer-Galley’s monographic study may lead to the conclusion that Internet fosters public debate. However, if the debate is intentional, users will have decided on the specific sites they will turn to; whereas participation to public online debate is often less intentional than has been supposed. A study conducted in the United States on the site Slashdot, designed for IT buffs, show that political debates hosted on that site during the 2004 presidential elections drew a lot of interest. Representing less than 5% of the whole, these political discussions were the liveliest, drawing more frequent comments – on average 35% more . It was likewise observed in France that, during the latest presidential campaign, the political debate was also significant on leading non-political sites such as doctissimo (health site) or hardware (IT site). This must be seen as equivalent to the chance encounters Sunstein sees as the origin of American deliberative democracy.
But it is conceivable to use the Internet in a more proactive way. Within the Electronic dialogue project launched during the 2000 presidential elections, a sample of US citizens were asked to take part in 60 political discussion groups that went online live on a regular basis. These debates patently increased voting and local involvement in public life[Vincent Price et Joseph Cappella, “Online deliberation and its influence: The electronic dialogue project in campaign 2000”, IT&Society, 2002, Vol. 1, n° 1, pp. 303-329.]]. For all that this is a limited experience, it can nevertheless serve as a model for the organisation of public debate, the “dialogical” democracy studied by Michel Callon, Pierre Lascoumes and Yannick Barthe . Internet may in this case advance the setting up of processes liable to enrich public debate and stimulate Citizens’ involvement .
The Internet and Political Commitment
To close this survey of the role of the Internet within the democratic process, it behoves to check the role IT networks play in party political activity. In a context in which political involvement is mostly on the wane (fall in support, reluctance to “sign up”), the web has set up a tool suited to new involvement formulae, be it in traditional parties or in new campaigning structures. The experience of Temps reels, the Socialist Party (PS)’s virtual arm is worth observing. The original intention was to try and bring sympathizers together, disregarding traditional geographic embedding. But this particular subgroup, and here lies its most innovative feature, has also become an essential PS forum on information and communication technologies, where the debate is open not only to the subgroup members (60) but to other party members (70) and to “correspondents” (100)  who contribute though they are not members. Thus within its own competence framework, Temps reels has organised a new debating model reaching beyond the scope of the organisation. The debate is lively since 60% of members participate.
This new form of virtual involvement was at the heart of Ségolène Royal’s site Désirs d’avenir. There too, the object was to call on expert knowledge and evidence coming from the grassroots, beyond the party. It got a mixed response . On the one hand the readership was low (125.000 unique visitors a month, that is 10% of left daily Libération’s site readership), on the other, the campaigning output was impressive (in October 2006, 45000 messages had been posted). However the most innovating factor resides in the campaigning activity shift from oral communications, actual meetings to the design of standardised forms, (style sheets, website architecture). Political involvement now calls on other skills; it is dictated by procedure.
Obviously the model itself should not be imputed to the Internet medium. For evidence of this, just compare Désirs d’avenir to François Bayrou’s campaign site. In the latter case, activist participation is completely different, the object is no longer to make points or deliberate but to express feelings, react to the campaign. The users who intervene are newcomers to political action who need to comfort each other and to share in a common experience .
Whilst major political structures do begin to get the hang of the web, this tool was first used by the latest arrivals on the campaigning scene, and whose organisational resources are low. In this instance, Internet is not adjunct to political commitment; it becomes one of its key instruments. The antiglobalisation movement Attac is a fine illustration of this case of figure. Originally, the Internet was intended as a way to reciprocate information among activists. But it was soon to become the channel for horizontal debates independent from the leadership within the organisation. These debates, which, as is often the case on the net, actively involve only a small minority, suddenly acquire a high visibility as the number of messages more than make up for the want of active participants. This “strength in numbers” Gabriel Tarde refers to in L’Opinion et la foule (Opinion and the Crowd) can bring about shifts in the balance of power within the organisation . Some of these discussion cheerleaders thus achieve a new power to call into question. They can get the Association to include some items that matter to them on the agenda.
Thus, the web is unquestionably a tool that enabled Attac’s grassroots activists to be heard, that brings about a reduction of the distance between leadership and grassroots. There is however a risk, namely that the most vocal people online become self-appointed representatives even though they have not been elected, as they draw part of their legitimacy from the faculty to be active, indeed ubiquitous in this new sphere. In an organisation that has had to handle many internal conflicts, it is plain to see that the relationships between Internet and democracy are shifting to say the least.
But the further one gets from an activism firmly structured around lasting organisations, the greater the proximity one finds between the network pattern of the web and that of some late arrivals to political activism. Fabien Granjon has referred to a “structural affinity” between the Web and the “anti-globalisation” movements . This movement, which builds up through the regular high points provided by the World Social Forums, is an activists’ transnational network. Information must be circulated, synthesized, sometimes translated, before being fed to a network which has no centre. This is therefore quite unlike the Désirs d’avenir model, where Internet is the means to re-centralise, or the Attac model (the ideological closeness not withstanding) where there exists constant tension between the elaboration of positions through voting and through the executive on the one hand and through the Internet on the other. In the altermondialist neo-activism of the Social Forums, there is a political project attuned to the Internet model, a fresh attempt at a reticular democracy is being evolved which ties in with civil society’s involvement. Pierre Rosanvallon observes for his part that one of the major characteristics of the new social movements is their pursuit of a watching, denouncing and recording agenda which underpins “counter-democracy”. He further observes a few pages on that the web is ideally suited to this type of activity .
Like other media tools, the Internet bears the mark of what notions its designers might have had of social and political communication. Theirs was a pioneering innovation socially as well as technologically. They felt and experienced these new forms of social relationships that have become known as “networked individualism” and which typify both people’s private and professional life . This social model also features in the latest types of political activism where individuals choose their modus operandi, but always in coordination with others. The net has thus a strong homology with new modes of democratic involvement and deliberation. Admittedly, not all web users, far from it, visit political news and debate sites, nevertheless new information devices have appeared and new debating formats are being set up.
Novel though it is, the Internet reproduces some features of earlier media. There is strong public concentration on a few determining sites. They provide the seedbed of shared information necessary to social cohesion. Internet is a network, a web where different data are connected one to another. This geography of hyperlinks is less open than had been supposed. There is therefore a danger that the web be more like an archipelago than a “Superhighway”. However this dispersion of the news providers also exists in other media.
All in all, the Internet per se does not have a negative impact on democratic deliberation. It has, to some extent, been shaped by society’s distinctive features but it also offers genuine opportunities for new – manifold, reticular – democratic formulae wherein citizens do not settle for just electing their representatives but can also discuss, watch and judge their actions.