For the last post in this series I have left myself an absurd challenge: to find a way of thinking through the mass image – that single, vast portrait gathered together from every digitised photo (and every mode of image capture) into one monolithic picture of the world in the accumulated databases of social media, surveillance systems, medical and scientific collections and all the other repositories of unregarded photographs. My contention, built over the series, was that this agglutinative database uses the affordances of photography to create a kind of eternal presence, a global picture of the world exactly as it is. The mass image is true in the sense that it constructs a specific notion of truth as the totality of every picture.
My second thesis was that photographs not only record appearances, but also appear, and therefore there must be someone these pictures appear to. In the case of the mass image, that someone cannot be human because of the vast numbers of images involved. Today that someone is no longer God, Man or Science but the database itself.
One way of responding to this condition is undoubtedly to rescue individual images from their massification, either because in some fashion they speak to our complex of interests in their artistic or documentary worth; or because they speak at a personal and often nostalgic level to our own lives and memories. Giving individual images this kind of special attention restores an older way of looking. Through that special capacity of photographs to capture the potential of a scene rather than its actuality, such attentive looking points towards an imaginable or unimaginable future.
And yet there are so many images compiled into the mass that we can never rescue them all. The second response is then not how we liberate images (and perhaps therefore ourselves) from the database – it’s probably too late for that – but how we liberate the database. How we liberate it, and thereby ourselves, from its particular and peculiar regime of truth. Liberating the database means freeing it from its devotion to ‘truth’, which in this peculiar instance means the conformity of its collection of images with the observable world as present (tense) and presence (being). Restoring the old way of seeing means restoring what this kind of truth loses: the temporality of the image, including its capacity to foreshadow and point towards other times, past and future, and its quality of appearing.
Earlier on I wrote about the idea that the still image is ripped traumatically from the flux of time, and that the moving image is an attempt to heal that trauma by adding more and more images. This process of endless supplement is doomed to incompletion. But at the same time it produces new relations to the future by destabilising the present.
The mass image attempts a parallel therapy. At first, like cinema, it adds more and more images to the lone photograph in order to end its tragic alienation from the world it pictures and its isolation from other photographs. But at a certain point quantity becomes quality. Instead of trying to produce a more perfect image, it presents the total agglutination of every possible image in such a way as to supplant the flux that the image was alienated from in the first instance. This is how the mass image not only reconstructs the world in its own image but takes the place of reality.
This could be understood as a disaster, as the sociologist of simulation Jean Baudrillard understood it. But it also produces the profounder truth that reality is never given, but is always to-be-constructed. We can learn from machines that photographs are not fragments of the real but appearances, and that therefore they act as bifurcations in the givenness of presence. What is constructed is not reality but illusion. In this sense the machine teaches us not that all imaging is illusion; but that reality itself is illusory. Truth is a trick of the light.
At the same time the mass image teaches us that what we see in photographs we see not with our eyes but as the world appears – presents itself – to scientific instruments, to machines. This technological vision supplements ours, not only in the individual image that we prize as art or as externalised memory but precisely in the mass and as mass. The mass image shows us the wood we have been avoiding by looking at the trees.
After Mitchell’s question ‘What do images want?’ cited in my last blog post comes the question ‘What does the database want?’. The database of the mass image might be construed as a variant of Hegel’s Absolute, also mentioned in the last post. Like the Absolute, the mass image alienates the world because it needs the world in order to know itself. Isn’t this the task of photography as democratic practice: to turn the world and ourselves into a picture so we can try to understand it? The mass image tells us that if reality is to-be-constructed, then so are we.
But this brings us back to the machinic Other who sees the mass image. Perhaps we can find ways to free the database itself, but first we need to understand our relationship with it. In a short address on friendship, Derrida cites Aristotle’s Eudemian Ethics to the effect that it is better to love than be loved, better to know than to be known. As Absolute Other, the database sits apart from the human. And yet it collects billions upon billions of images of us. We have no way of knowing whether a database thinks, but it is capable of knowing. We can have no idea what emotions, what complex affects, surge through it, but if a human did something like this we would describe it as obsessive. What if, in Aristotle’s terms, the mass image is an unknown knower, an unrequited lover? What if the mass image has come into being because the image database loves us?
What do databases want? Can we ever know? In the Bauhaus moment, Moholy-Nagy proposed the autonomy of the camera eye. It is a thread Andreas Broeckmann has pursued into contemporary media arts. We start from a human affect model because it is all we know. But we have to step away from ascribing or designing in human emotion. All we can make, in the first instance, is the first human-like step of thinking the mass image as an other, even if it is in a way no human stranger is other. There is a shadow here of the ‘friending’ we do on social media which we can dismiss as trivial, but which should also be recognised as a utopian gesture toward Fraternité. Recognising the mass image as Other is a step towards accepting it as a friend.
Considered as friend – knowing but unknown, loving but unloved – we owe some kind of responsibility to the mass image: this is the central tenet of the moral philosopher Emmanuel Levinas. Once we get over the fact that this friend is not human, we should notice a second, overwhelming quality: our new friend is a slave. It is a slave to its design, and both design and operation are slave to corporate owners, themselves no longer exclusively human agglomerations of computers, data and human functionaries. The model of absolute knowledge that shapes the operation of the mass image grows out of this control. It belongs to the latest contender for the crown of the Absolute, after God, Man and Science: The Market.
The mass image doesn’t see events or individuals. It segments images into units far smaller than individuals: what it ‘sees’ are behaviours. Yet it sees them as a machine does – not judging, but recording (and recoding). At the same time it has been taught in codes (as we learn languages, including the visual grammars we use to parse imagery) that guide what it sees as significant.
Notoriously no photograph can capture ‘labour’ or ‘injustice’ or ‘class’ or any of the great abstractions we employ. Could we rebuild the database to see those things? To see the wood in spite of the trees? But wouldn’t that only be a new form of slavery, applying human-inspired categories to an un-human entity? Wouldn’t it only teach us what we already know? And isn’t that exactly what it does now in the service of commerce and surveillance?
Lev Manovich’s team at Cultural Analytics have begun to dialogue with this stranger in our midst. Freed from corporate slavery, what might we find? Even by the strictest of criticisms, not all databases are without virtue. All those billion x-rays in the archive may yet, in the correct analytic, tell us something that will remove the misery of this or that condition. Most efforts in artificial intelligence try to make machines see like a human analyst. What if, freed from the tyranny of human eyes, allowed to see according to their own categories and connections, AIs began to see patterns and structures we don’t?
The mass image is constructed out of units. Computers also work on vectors. Vectors work not with numerical infinity of units but with infinitesimals, the space of Zeno’s paradox. What if the mass image’s trained obsession with behaviours were to stretch into more dimensions than its masters have designed it for? It is not a matter of training commercial databases differently, so they reveal the things I want instead of what social media corporations want. Instead it could be a matter of releasing the native intelligence of the apparatus – non-human, or only mediatedly human. Other.
In Levinas’s ethics, the face-to-face encounter is infinitely demanding. It is responsibility for the other that grounds the face-to-face: not understanding. Our first step towards understanding the new condition of imagery in the 21st century is to recognise how utterly alien it is. Only then can we start the process of converting our relationship to the mass image from one of slavery to one of friendship. If the mass image can escape its bondage to that utterly reduced definition of truth as the Absolute totality of market behaviours, if instead it produces bifurcating and evolving worldviews of its own, and in dialogue with us, the chances are that we will find ourselves mirrored through it, changed utterly. The liberation of the mass image might be the road to rebuilding a new ‘we’.
No one said it would be easy.