Coda

Post-cinema was first associated with the omnipresence of moving images, digital cameras, and screens. We are indeed surrounded by digital images in public spaces, streets, squares, trains, subways, and so on. These images appeal to our senses and engage us as consumers. Yet we also use our digital devices to record films and videos ourselves, and to send these latter as digital packages into the world via various digital platforms. Filmmaking has become ubiquitous, gaining independence from the studio system. Films are made all over the world. Even Hollywood has multiplied into Bollywood, Nollywood, and Wellywood. The clear separation between producers and consumers has become increasingly unstable. As Malte Hagener, Vinzenz Hediger, and Alena Strohmeier write, this instability poses new questions for film and media studies: ‘Who makes films and under what circumstances? Who controls the circulation of images and controls censorships?’11The State of Post-Cinema: Tracing the Moving Image in the Age of Digital Dissemination, ed. by Malte Hagener, Vinzenz Hediger, and Alena Strohmaier (London: Palgrave Macmillan, 2016), p. 3. Their response to these questions is that the previous conceptual Beginning of page[p. 182] distinctions and hierarchies no longer work. In their book The State of Post-Cinema, the working hypothesis they put forward is that,

the hierarchically employed dichotomies that have long guided the study of films — for example, center/periphery (that is, the film culture of urban centers such as Paris, Berlin, and New York vs. the rest of the world), first/second and third cinema, theatrical/non-theatrical, auteur cinema/non-artistic forms and uses of cinema, professional/amateur, and also experimental avant-garde/mainstream cinema — no longer apply to the current state of moving-image culture.22Ibid.

‘If’, the authors continue, ‘moving images, like water, always find a way to spread, then the question is how they influence the spaces they reach and the spaces they use in between.’33Ibid. I agree with this claim, yet I add that we have to deal today with even more: we are also simultaneously surrounded by digital cameras, images, and screens that are not there to address human eyes, but instead to collect and transmit data. The cameras may be surveillance cameras, thermal imaging cameras, or infrared cameras. They register our movements and generate images that store and process information; they are part of a ubiquitous computing network and a comprehensive process of big data collection. But at the same time, we also actively produce data and metadata ourselves when we record, store, process, and send our moving images and films using digital devices.

In other words, post-cinema is currently embedded in a ‘phase shift of advanced automation’, which Beth Coleman,Beginning of page[p. 183] author of Race as Technology (2009) and Technology of the Surround (2021), sees as paralleling the emergence of ‘environmental immersive technologies that surround although [are] often invisible.’44Beth Coleman, ‘Ten Statements of Technics’, in Technics: Media in the Digital Age, ed. by Nicholas Baer and Annie van den Oever (Amsterdam University Press, 2024), pp. 28–30 (p. 29) <https://doi.org/10.1515/9789048564569-004>. These technologies include virtual reality, augmented reality, generative AI, and machine learning. This shift is accompanied not only by an acceleration of datafication, but also by a change in the socio-technical relationship: machine-learning based algorithms are supposed to be able to predict and identify behaviour more accurately due to their ability to process large amounts of information, and thereby also to shape our conduct. Yet, as Alexander Campolo and Katia Schwerzmann, among others, have persuasively argued, models of machine learning are trained by ‘examples’, that is by using data that has been turned into examples. The term ‘example’ is, as they emphasize, not congruent with that of ‘data’: ‘Instead, it designates the complex assemblage by which data is aggregated, formatted and processed so that norms can emerge.’55Alexander Campolo and Katia Schwerzmann, ‘From Rules to Examples: Machine Learning’s Type of Authority’, Big Data & Society, 10.2 (2023), pp. 1–13 (p. 11) <https://doi.org/10.1177/20539517231188725>. Campolo and Schwerzmann show in detail how these examples elicit norms in an implicit or emergent manner to make prediction and classification possible. They use the concept ‘artificial naturalism’66Ibid., p. 1. to characterize the tensions that result from this ambiguity between data and norm, and ask the following question:Beginning of page[p. 184] ‘What can it mean to be governed according to machine learning’s type of authority and its implicit normativity?’77Ibid., p. 11.

However, this is not the only problematic aspect of the techno-social, that is, media-anthropological relationship of immersive environmental technologies. In addition, machine- and data-based predictions of the future can only ever be an extension of the past. As Beth Coleman rightly reminds us:

In the face of the institutional legacy of hegemony, engagement and activism are mission critical. If we do not speak out against the forecasting of the past as the future, we won’t have a future.88Coleman, ‘Ten Statements’, p. 28.

The stakes are high: how is it possible to hold on to an open future under the conditions generated by environmental immersive technologies, which are simultaneously intertwined with neoliberal capitalism’s increasing aggressivity and anti-democratic bent? This question corresponds with another: how is it possible to stand up to control and surveillance? The answer cannot be via generative AI, the environmental immersive technologies described, or the techno-social relationships they stand for.

In this book, I argue together with Deleuze that a new kind of resistance can be found in the aesthetic function of the cinematic image. I argue further that queer post-cinema is inventing precisely this new kind of resistance. The aesthetic function of the cinematic image, in contrast to the social function of the visual, is characterized by a ‘supplement’ that refers to the qualitative nature of time. Deleuze describes it as ‘a kind of gap with a still virtual audience, so you have to play for time and preserve the traces Beginning of page[p. 185] as you wait’.99Gilles Deleuze, ‘Letter to Serge Daney: Optimism, Pessimism, and Travel’, in Deleuze, Negotiations, 1972–1990, trans. by Martin Joughin (New York: Columbia University Press, 1990), pp. 68–81 (p. 73); my translation. This ‘kind of gap with a still virtual audience’ cannot be implemented in the algorithms of machine learning models. For it doesn’t exist in passing and calculable time, but in time that lasts and coexists. This lasting and coexisting time is the time of a virtual, open future.