A few hours spent on Chatroulette allow to take part in situations as diverse as: readings, concerts, acrobatic performances and miscellaneous forms of cultural show-and-tells, lectures on specific topics (photography with a photographer, capitalism with a Chinese student, etc.), many mise-en-abîmes on the topic of discussing in Chatroulette, competitions in which incongruous objects are shown to the camera (stuffed animals, kitchenware...), dances with hysterical swimsuited people, dinners with families gathered around the table, fool plays with female american students offering a strip for a strip and whose male friends can be heard guffawing off-screen, photoflash games and so on. Also, we access videos of wankers displaying all kinds of dimly lighted solitude, drunk teenagers sprawled on sofas, not caring for discussion and taking our image as an object for their own conviviality, embedded videos2 going from political slogans (most of them being feverishly anonymous racist or nationalist messages) to porn and gore clips (the man who breaks a bottle as he is pushing it inside his anus, or the famous “2girls&aCup”, etc.). And lastly, we regularly access to black screens...

Over all, there are two types of events occurring on Chatroulette: first, interactions between Web users through videos, sometimes augmented with text messages, interactions which constitute a shared experience and a performance that is co-produced by – and staged for – themselves. Second, moments setting the user in the position of a solitary observer, as far as nobody is responding or following on what he is himself doing in front of his webcam. This last type reassembles the images, flickering for just a second, of users automatically hitting the “next” key, the indistinct series of those we are ourselves discarding and all the embedded clips, frozen and black screens... Indeed, we are only catching a glimpse of each one of these images, but they are adding together with an effect of recurrence, and if they are reducing the possibility of having a singular interaction with the users behind them, they eventually exist as a generic type of image, as quasi-percepts. These images are not hiding anything but, quite the contrary, they are showing that they are not showing their source, and while they are stating the fact that a face-to-face interaction will not be possible, they are not pretending that nobody is behind them, observing.


The publicity made around Chatroulette very soon after its creation has produced, within print and online medias, quite a large amount of immediate commentaries from journalists as well as a noticeable number of sociologists. One surprising fact is that these commentaries have regularly mentioned the existence of perverts to name those we for our part called wankers3. Whether they are described with humor or animosity, perverts have been elevated to the rank of a descriptive category without a word of explanation. The definition for this category is only perceptible as it commensurates with adjoining categories: Chatroulette would feature such a proportion of males, such a proportion of females and such a proportion of perverts. In other words, users are to be found on the Internet that are neither male or female but essentially perverts4. This way of talking about users of the system is quite caricatural ; yet it replicates within other descriptions, which in theory should be free from any kind of moralism as they are listing types of images instead of users. In this regard, commentators5 are differentiating webcams showing “identifiable persons” from webcams showing “genitals”. Yet the proportions of male and female users of Chatroulette are only based on the listing of “identifiable persons”, excluding 5 to 8% of “genitals” that, to say the least, look very much gendered! Each time commentators have to deal with the distribution of gender, the third category appears: thus there would be non-sexual men, non-sexual women, and something sexual that would not be a person with a gender but not insignificant enough to be simply ignored. The problem is plain: one identical visual object, which is identified by every commentators, is alternatively designated through a medical category that characterizes a person or through a semiotic category, which here, however, does not clearly states its name: pornography. Sex panic among commentators...

Why was the choice not made of describing the contents of Chatroulette with a taxonomy that featured a category for “pornographic images” ? Then the “genitals” or perverts could have been listed with the webcams displaying pornographic pictures, embedded video clips such as the one of the man and his broken bottle, or even with webcams showing explicit and humorous references to the sexual uses of Chatroulette6, as well as videos of young women flashing their breasts, etc. Such a reasoning, however, would have had the effect of blurring the obvious: on Chatroulette, it is possible to interact with wankers, and this is obvious enough to be a problem for the commentators who have been on the website and have made it the blind spot of their accounts. Interacting and watching porn do not manage be compatible in their mind.

Our sexual culture clearly differentiates two ways of understanding sexual practices outside the Web: the ones we are involved in (sex with others, the “real thing”), and those we are only viewing/watching whatever we would also be doing (pornography, which is a solitary thing). The distinction between the two was plain and simple before the Web came up, because pornography was the only one supported by a documentary nature. But the Web has made this evident distinction to falter as it is hosting both interactive and solitary sexual practices, as well as what would pass for a more neutral, non-sexual type of communication. As a result, the issue is moving from technics to semiotics: on the Web, pornography is only designating a type of document, it is a label that marks the sexual dimension of a picture even before we are able to access the picture itself. Traditionally, webcam chats have always been documented with a vocabulary describing altogether the contents, the actions and the users to whom access is granted: a simple nominative contact list of users we already know, a database of profiles coming with a categorial search engine and a history, or even a system of tag-words and indexes if users do not know each others already. In that case, the vocabulary gathers categories that are related to the social representations of relations, identities, roles and sexual practices7, and this regime is effective from the fact that users know that they do not know each others, they have come to attend or perform a sexual act that has been advertised. In contrast, without such a semiotic infrastructure, where does the sexualization of an online interaction would start: at the visible erection ? At the sexy eye contact ? At the exhibition of a naked body ? Half naked ? A light summer dress ?

Chatroulette does not play that game, there are no profiles, no tags or “contacts”, not even image captions that would inform us on what we are looking at and what to expect from the next video. On Chatroulette, there is no priority given to the semiotics of writing, indexing and hypertext links, but to the video image and to the audiovisual performance only. The website does not create any kind of indexical data8, and therefore no information about images, people and visible actions. At the same time, its very minimal interface does not provide any meta-discourse regarding the contents: no slogan about friendship, fun or love like some dating sites would and even less slogan advertising for “porn inside”. The homepage does not say anything about what is going to happen, and the interface is leaving users (and researchers) free to find any signification they want in the encounters they experience and the images they see9. Thus the website neighbors uses of the Web as well as desires and motivations that, until then, were clearly distinguishable, because they were distributed in classes of websites that were themselves “naturally” distinguishing from each other by advertising their contents (as dating sites, porn sites, social networks, personal pages, etc.). Chatroulette appears to be twice mean on the informations describing the contents a priori, and this makes it impossible to consider as neither a social network10 nor a pornographic website. Thus when the researcher happens to face about a hundred different images of wankers over an hour, he has to find within his own ressources ways for interpreting sexual situations he might not had chosen for an object of study, and about which nature nothing was advertised a priori.


One blogger has presented a proper report of his experience on Chatroulette instead of yet another count11. The writer's account comes with quite a big series of pictures taken during the session he has organized with the intention of viewing women's breasts on his screen. In order to do this, he has played various roles, and therefore produced various types of video images, of which he eventually presented the effects. One of his tactics was to show a sign reading “tits for France”, and this was the one with which he affirms he had the most positive answers to his request. Among the young women agreeing to the display, are mentioned one young man lifting his shirt as a way to signify his inability to satisfy the blogger's request, and one curious character with an old man's face and a prominent bosom (Surgery ? Hormones ? Artificial breasts ?). Both these men have responded to the sign game, showing their “tits” and are appearing in the article, but they are not listed among the wankers. If the blogger comments on the performance of the first writing “that one could have spare me”, and the second one with a “what the FUCK!”, he however provides the reader with the possibility of confronting a personal reading of these pictures with the author's own interpretation, and of understanding to what extent these comments are a reaction to what has marked and affected him while looking at his screen. In this article, the author holds a typical heterosexual position and openly comments on his experience from this place.

Further, he also recounts his experience as an “incidental wanker”, so as to “see what these thousands of wankers might see, from the other side of their screens”. He recalls “two gay guys wanting to watch him masturbating”, women judging his penis too small and, finally, a few displays of breasts rewarding his commitment to journalism. Through such an intervention, this blogger shows that accounting for sexual uses of Chatroulette is correlated with the evolution of interactions rather than the sight of the first image, women are only undressing according to the negotiation they engage in. Yet, as he writes about this part of his experiment, the blogger hides the images from his own webcam and does not show the screenshots of the men he has discussed with, and who were mentioned in his article. Why these disappearances ? It does not matter in the end, as long as his affects are manifested in his account, as long as he signifies the margins of his personal integrity, of his prudishness, of his courage too... This might be how he allows the reader to understand what were, for him, the differences between “facing...”, “interacting with...”, “consulting documents about...” and “being in relation with...”, or even “being seen by... without seeing” and “being able to account for...”. This blogger has accepted to be held12 in possibles roles and in interactions for which he could not control all of the consequences. From a methodological point of view, the affects of the observer are part of the issue, as much as his own reflexes are related to them, as well as his habits, without which the sexual situation would not be more singular than any other. They are also an issue in the ways he reacts, in his adaptations, his boredom looking at the screen, or even in the inventions he makes in order to resist the indigestible aspect of some interactions. So, the observer has to work upon his discourse, to tire it out and locate the consequences of his practice for his knowledges, that is: for the descriptive categories used for the account as well as for his abilities as a user/actor13. If it is utopian to consider that one should be completely transparent to himself in order to achieve the production of an account (reflexivity is especially limited when it comes to sexuality), it is about being able to present – and represent to oneself – how the opacities of one's online uses and practices are likely to alter the resulting account. Quite a task the perverts counters have neglected to perform.


Every Internet user has his own ressources, his own feelings, his own exhaustions, etc. This is a limitation for the possibility of accounting when the user happens to be a journalist or a researcher, but as we just saw, it is not so much of an interference and the reader knows how to cope with it. Any researcher working on the Web knows that, on it, he might find sexual documentation under one form or the other, and that the Web provides with tools of some use for sexual exchanges, which might be destined to stay online or might move offline. Despite that fact, studies on interactive apparatuses are, most of the time, focusing on regions of the Web that are sexually neutralized, regions that might be regarded as “fleur bleue” and romantic14, and which preserve pornography, without any explanation, as exterior to their object, as a fixed exteriority of no reason for them to question. Thus, they nourish the idea that the Web would be technically and “naturally” a sexually neutral medium, in which the occurring interactions are never conceived as sexual15. It is possible to formulate the hypothesis that this methodological black hole is primarily made of a technicist knowledge and passion that is overestimating the integrity and identity of the technical object. This object is awarded the ability of intrinsically producing interactions (which by default, are neutral). Conversely, its technical essence degrades when corrosive pornographic contents make their apparition, contents which are only dedicated to some obscure and dangerously non-interactive practices.

From Foucault, we know that a whole economy of confessions, but also a whole apparatus of locations and specific conditions allows to talk about one's sexuality, not to repress it but to produce it intrinsically as a knowledge16. Following Kosofsky-Sedgwick, we also know that silence and ignorance take part in these apparatuses as performances by acting actively on the circulation and formulation of affects, desires and identities, noticeably the hetero- and homosexual ones17. It is intriguing not to find, in the studies and comments about Chatroulette, the tools provided by these two authors, especially the one idea that sexuality, even beyond the existence of the category of pornography, is an issue of the exertion of powers, or in other words a specific configuration of interactions, wherein discourses of authority and truth about one's sexuality are intimately articulated with the commitment to the acts. All the commentators that are focusing their efforts on the category of perverts or its euphemized versions are producers of a knowledge on the sexuality online, and all of them have been participating in the (re-)production of categories of reading, understanding and experimentation of the apparatus. Parallel to their underlining of the sexual contents, they are rejecting the issue on the perversity of Internet users or neutralizing it by narrowing it down to a question of specific type of image (the framing of “genitals”), rather than questioning their own half-neutral-half-pornographic conception of the apparatus, which is already a sexualized conception. Doing so, they remain as systematically silent on the issue of the way they are using their own experiences of interactions and their own affects within their accounts, and spare themselves the epistemological investigation that Chatroulette is serving on a platter: the opportunity of not being able to confuse or distinguish a neutral apparatus and a sexual one anymore.

Indeed, the problem that Chatroulette is posing is not, primarily, a problem of classification of the visible, but that of the repertoires of interaction, of which embedded or fixed videos, inactivity or immobility are plainly part, among many other forms. The eye of the semiotician tends to renew, sometimes subtly, the distinction between neutral images and pornographic images, the first category only being the negative of the second and focusing on that one alone silently signifying its technically problematic aspect, while pretending to ignore it. From this point of view, the issue of wankers is not so much connected to the solitary or perverse nature of their sexuality and the contents they are producing, rather than to the fact that they are using a medium that is supposed to be interactive and that through it, they are making commentators to call upon antagonist ressources: for the latter, acknowledging the image of the wanker, being able to describe it according to its sexual modality is using a ressource that is constitutive of a culture of pornography, this is, to some extent, responding to the summons for participating according to an identical sexual modality. This modality becomes incompatible with the very idea that the apparatus is interactive, as far as it would have to be sexually neutralized, as it is neutralized in the idea of an objective scientific investigation. According to cases, interlocutors are not facing identical media objects, because they are not using identical knowledges. In other words, the pervert is liable of a technical misuse, and he interferes with normal uses, the communicational and interactional ones, disregarding the obvious fact that his interlocutor is not playing that game. Moving away from Chatroulette and looking at many other types of websites, we would observe that there is a lot of non-sexual exchanges going on in some pornographic webcam sites, that most of the interactions in dating sites are not leading to an offline encounter, or even that it is possible to balance the discourses of romantic sentimentality and those of explicit sexuality (on gay and lesbian websites, for example, where the figure of the pervert is almost non-existent, as opposed to romantic websites), the moral integrity of the technical apparatus is fading, and the tragedy of being exposed to the contamination of online perversion turns out to be a much lighter problem.