For an autonomous existence of images: an archaeological perspective

Monitoring and interpreting an increasing number of images has become part of people’s daily lives. These images trigger a complex process of relations that can result in direct human or non-human actions over people, over services or over the very space. As part of a broader and widespread mediascape, the repertoire of images, its organization and connection to multiple devices and huge databases, make interpretation processes much more complex and beyond our reach. When arranged in a network, technical devices do not need to follow a logic narrative of facts. Paradoxically, they contribute to the construction of all possible narratives. This work proposes, from an archaeological perspective, that the intentionality of images, especially those that are produced and circulate in digital environment, is the symptom of a contemporary episteme that delegates to objects not just a functional autonomy, but also one of existence and of description of the world. The multiplicity of digital images makes of them Beings that exist beyond the human and that constitute a kind of continuous phenomenological machinic process, an awareness of the self and of the other.


Introduction
In Thomas Elsaesser (2018) work, we are invited to think about cinema from an archaeological point of view and as part of a larger set of communication practices born and developed over the last two centuries. Elsaesser tries to present new possible methodologies for the study of moving images, different from those classically adopted by Film Studies.
The first major question that arises for those unfamiliar with the perspective of media archaeology is how to renew the approach to an object so thoroughly studied by more hermeneutical biases. In relation to both form and content, traditional analysis has focused on producing some kind of rational connection between films and something external to them. Film's meaning is therefore revealed through an act of interpretation -centered on narrative or on cultural issues related to the film as a social phenomenon. The interpretation and its resultant meanings depend on an external agency, which is that of the spectator, the critic or other human beings using intellectual mechanisms to produce human-centered readings of film discourse.
Existential phenomenology (SOBCHACK, 1992) presented the possibility of an equal existence of both the body of the spectator and the body of the film. Although, in this case, the production of meaning by the co-existence of multiple bodies transcended hermeneutic interpretations, the results of consciousness investigation is still originated and dependent on one privileged conscience from a specific sentient body. Little has changed in the subject-object relation and its inherent hierarchy.
As part of a broader and widespread mediascape, present in virtually all moments of everyday life, images can now be read and interpreted by non-human devices. Invisibly, the repertoire of images, their organization and connection to multiple mechanisms and infinite databases, make interpretation processes much more complex and beyond our reach. Each day, images manifest a more evident autonomous consciousness.
According to Vivian Sobchack (2011), media archaeology is an undisciplined discipline, which tries, through different methodological approaches, to deal with discourses originated from the materiality of media itself. Thus, archaeology must be concerned with how objects shape thinking and our being in the world. Archaeological bias recognizes a sort of agency from the objects, something that has not been regarded as important by classical humanities. On the other hand, for authors like Ernst (2011), there is no reason to look for meanings from a cultural point of view. Media (and he refers specifically to the material aspects of things) are themselves the very meaning.
For an autonomous existence of images: an archaeological perspective Therefore, we start from a type of narrative built from the multiplication of images. Understanding, based on the considerations of Bruno (2012), the permanence of digital traces of any action in digital environments, especially on the Internet, as potentially recoverable, we are faced with a virtually infinite set of images. The act of "reading" and "interpreting" all of these images exceeds human capacity, especially in the case of large collections of data compiled through different search engines. How to work within this chaotic scenario?
Based on Hayles' (2012) proposals on the relationship with large databases as new tools for the humanities and from the expansion of the notion of agency by Latour (2013;2005), we will think to what extent this proliferation of images brings us closer of a "truth" about the facts or simply relativize their very existence.
This work proposes, from an archeological perspective, that the intentionality of images, especially those that are produced by and circulate in digital environment, is the symptom of a contemporary episteme that delegates to objects not just a functional autonomy, but also one of existence and of interpretation of the world.
The multiplicity of digital images makes of them Beings that exist beyond the human condition and that constitute a kind of continuous phenomenological machinic process, an awareness of the self and of the other.

Actor-network theory and digital traces
For Latour (2005), sociology traditionally refers to the "social" as a set of stable relationships that could explain the ongoing processes through which beings relate.
The essentialist character of this conception would not be concerned to understand the flow of action between objects, limiting itself to display them as a result of larger structures: Market, Capital, Society. The relative permanence of relationships assumed in this perspective is problematic for not taking into consideration the determining features that "every" involved element has in the composition of association networks.
Thinking of social as "associations" rather than as "social ties" allows a greater clarity on the composite character of society, whether relations may be transient or more permanent. The number of actors participating in those processes, according to this proposal, should increase. "When we act, who else is acting? How many agents are also present?" (LATOUR, 2005, p. 43). The task of explaining how the social is weaved becomes more complex due to the inclusion of all sorts of agents, human or nonhuman.
Technical apparatuses are a key part in building these networks. Describing them and trying to understand the relationships that arouse, however fleeting they may be, is part of a dynamic process. Networks and collectives are thus constructed, mapped, reconfigured, and mapped again in an endless movement. The "social" is that what emerges from these changes. The fact that when we think of this type of configuration we make previously invisible relationships (and objects) visible is implicit in the adoption of these concepts.
The invisibility of objects is due to their accommodation in certain spaces.
When they do not represent any kind of innovation or do not demand new forms of interaction, the existence of such objects is taken as something natural. Thus, the agency that such objects play seems not to come from them, but from a higher instance.
When proposing innovations, artisans, technicians, or scientists renew their visibility (of the objects) and their role of mediation. In addition, the distance to these objects -either through space or time -nurtures their qualities of innovation. The strange arrangements proposed by old apparatuses displaced in time may be disruptive, at least shortly, for contemporary technology users. Familiarity or indifference towards the object can make it disappear from the horizon of our expectations. The failure or lack of efficiency of objects also endows them with a power of agency not previously considered. When a particular process is interrupted by a "technical failure", the object transforms itself from an intermediary into a mediator. Its status changes from something that is just completing a gap between actions into an effective action proposer -an "actant". This mediation can become a hindrance of great proportions the more objects are linked together. A failure of a server or of a computer network can result in the compromise of basic services such as electricity supply in a large area of the country. A defect in a car engine can result in huge traffic jams. Each technology brings within itself a "state of crisis" in which it was generated. This condition tends to be unnoticed as long as we get used to it. The dynamics involved in its development can be retrieved through archives and collections. We impose silence to technical objects when we take them as natural effects of social circumstances. We treat technology as something desirable, but at the same time, as a set of complicated methods about which there is nothing to be asked, just to be made use of. "How have the Moderns managed to miss the strangeness, the ubiquity, and yes, the spirituality of technology? How could they have missed its sumptuous opacity?" (LATOUR, 2013, p. 210). Latour suggests that in the very field of fiction we can find narratives about new relationships with objects. To give them personality or power over the human, as it has been common practice in fantastic and/or science fiction literature, is to recognize a deeper type of mediation.
To think about digital traces, as proposed by Bruno (2012), is to put into For an autonomous existence of images: an archaeological perspective question the state of visibility and invisibility of things. The trace would be a "quasiobject", located "on a threshold between presence and absence; visible and invisible; Surveillance mechanisms have also been routinely used in order to help governments scan through the vast material found online and to develop the most sophisticated tools of repression. Facial recognition, for instance, helps finding and arresting suspects of criminal activities. The flow of traces, many of them in the form of images, is gigantic. The database thus constructed can only be "read" and analyzed "at distance" by machines. To frame and look at things have become now a much more complex activity.

Database and narratives
For Hayles (2012), the conformation of increasingly complex databases is both an inevitable phenomenon and an opportunity to broaden the humanities field of research. In general, the hard sciences have adapted more easily to new working methods from digital tools. However, the humanities, still attached to a textual and interpretative approach form, struggle to adapt to the new context. Symptomatically, Manovich (2013) identifies the software as an engine of contemporary societies. New technologies operate at different time scales and through cognitive processes different from that of humans. Hayles (2012) believes that a conceptual model for understanding the new interactions between humans and technical devices is needed. Close to that proposed by actor-network theory, Hayles asserts that Objects are seen not as static entities that, once created, remain the same throughout time but rather are understood as constantly changing assemblages in which inequalities and inefficiencies in their operations drive them toward breakdown, disruption, innovation, and change. Objects in this view are more like technical individuals enmeshed in networks of social, economic and technological relation, some of which are human, some nonhuman. (HAYLES, 2012, p. 12).
Such interactions are just possible from a co-evolution of humans and devicesa technogenesis. Hayles starts from the theory proposed by James Mark Baldwin in the late nineteenth century, that when genetic modifications occur in a particular group or population, they are stimulated and accelerated through environment reengineering processes. This epigenesis has become more intense due to environmental conditions than to genetic factors.
Thus, the way we interact with digital technologies necessarily imply changes in our physical and mental processes. New models of cognition have been studied by neurophysiology, neurology and by the cognitive sciences, pointing to a great plasticity of the human brain. The increasingly intense contact with screens of computers and other digital devices has resulted, according to Hayles, in a decrease of what she calls "close reading". This kind of reading, characterized by a deepening in the content accessed and by an "interpretation" of an unattainable meaning of what is read, would be responsible for most of the cultural capital that literary studies produced so far. These, however, would go through a moment of crisis, being unable to follow the development of new forms of reading. Such practices mediated by new devices would differ not only because of the support on which one reads, but by the way brain and the environment are integrated. New arrangements of the body, the senses, and hypermediatic contexts do not support the close reading format.
The increase of information on the Internet has made available an amount of texts far superior to those existing in printed form. The hyper reading revealed itself as a necessary practice in these new environments. The rapid association and grouping of information according to themes and topics for quick and superficial researches is one of the strategies to cope with this quantity of data. The automation of processes via software works as an initial filter, preparing for a more deep reading.
Anyway, filtering would be in itself a form of reading: a machine reading endowed with an extended memory.
Manovich (2001)   Besides its ability to cope with endless processes of grouping and indexing digital objects, the database works from a principle of self-containment. The user does not need to leave the database to describe it. It would be impossible to operate within the logical systems of computers unless it worked that way. The executable commands comprise objective actions that separate, organize and list the elements of a previous list. There is no action outside this perspective, or better said, out of this paradigm. The association in networks can expand the contents of the databases, but does not change its closure character. In the case of narratives, there is always something out of the set of previously planned actions. Narrative depends on the subjective construction processes involving text and reader. Reading, in the case of databases, is a machine reading, unable to deal with the inexplicable, the ineffable, the unconsidered or the impossible.
Narratives are temporal constructs that can bear different chronologies according to each of its levels (text, history and fable). The order in which the events are narrated is fundamental to the narrative, and the relationship between different temporalities produces a complexity that the database form cannot anticipate. Its main (and almost exclusive) function would be the visualization of information in a spatial manner, from the traditional arrangement in rows and columns "to the more complex n-dimensional arrays and spatial forms that statisticians and data analysts use to understand the stories that data tell" (HAYLES, 2012, p. 179).
Data analysis in itself would build a strange kind of narrative -in general, a quantitative one -about the events. What has become a common practice is to delegate to numerical speculation the production of non-hermeneutical meanings for certain social phenomena.
Close to the machine Bruno (2018) alerts us to the impossibility of "seeing" the city in its different scales. When we talk about "seeing" something of such magnitude, we hardly refer to an exclusively human act, especially if we have the intention to perceive it in its entirety. Images have always mediated both the need to apprehend a broader amount of space and the possibility of somehow storing that "space". Countless camerasespecially those connected to some type of database ( 238). Her interest, in this specific case, falls on the surveillance mechanisms of public spaces, such as the Rio de Janeiro Operations Center (COR) or monitoring systems in places such as subway stations that can "read" passers-by behavior, anticipating unexpected actions like suicide attempts. The hundreds of monitoring points scattered throughout the city produce a map not only of physical spaces, but also of the bodies and attitudes of individuals who circulate in these spaces. The network constituted by the integration between devices (with facial recognition capacity) and large databases is a fundamental element for the successful functioning of smart cities (a very popular notion nowadays). In 2017, the Japanese company Hitachi announced the development of a tracking camera system using artificial intelligence (AI). The system can perceive and differentiate more than one hundred individual characteristics such as sex, skin color, hair color, clothes, mannerisms etc. from each person in a large group.
Using this technology, it will be possible to detect a suspicious individual or a lost child using information from eye-witness accounts to detect a person fitting that description from public security cameras set up in large facilities or city areas. (HITACHI, 2017).
For an autonomous existence of images: an archaeological perspective The use of AI or machine learning in the production and analysis of these images has become a tool for not only State control, but is also spreads as a common mechanism for security and individual protection. Companies, specialized in monitoring apparatuses, offer a wide range of products. For example, IntelliVision website, at its initial page, presents us a short description of its services: IntelliVision is a market leader in AI and Deep Learning-based video analytics and video cloud software. IntelliVision solutions provide actionable insights and intelligence for security and monitoring in Smart Home, Smart Building, Smart City and Smart Retail applications, and Advanced Driver Assistance Systems (ADAS) for driver and road safety in automobiles. (INTELLIVISION, 2020).
Monitoring and interpreting an increasing number of images has become part of people's daily lives, even of those who are not aware of the existence of such cameras.
These images trigger a complex process of relations that can result in direct human or non-human actions over people, over services or over the very space. Anyway, these digital processes inevitably produce traces, which are stored as data for later use.
The use of the term "intelligence", when talking about technologies, refers to the fact that, due to abilities that certainly differ from human ones, these devices are able to expand their repertoire of information and refine the actions taken from there.
The fact that this whole process is obscure for humans does not change the fact that it constitutes a cognitive capacity with a certain autonomy: an apprehension of the world through data, which is opaque for human consciousness. Attempts to compare machine consciousness with human consciousness, as if both were corresponding processes, must also be taken carefully since the two cognitive processes presuppose different logics. If, for existential phenomenology or neurophenomenology (MERLEAU-PONTY, 2006;ANDRIEU, 2006), human consciousness is the result of the material conditions of the body and their relationship with the environment, this bodily materiality is dispersed and much more subtle in the coupling between the multiple elements of informational networks. Machines' consciousness is produced by the communication between the points of its network's rhizomatic structure, relativizing space, as we will see in the next section, and treating it through its own internal logic. Machine's perception of reality may initially resemble that of its creators, but, in fact, it produces a distinct ontology of things.
Ross Goodwin and Gene Han, in a VICE journal online article (2015), demonstrate the camera they created, which, through Machine Learning, would record different people's images and attempt to describe them in "human" terms. When confronting these images with different databases, the camera would try to guess their jobs, what they were doing at that moment, their mood, and other personal information. Goodwin says: "We wanted to create an entity with its own sense of social awareness, its own eyes, and an ability to communicate with humans, albeit with some glitchiness that underscores the limitations of the current technology" (apud VICE, 2015).
In that test, the camera performs at least two operations. The first one is to record images and compare them to all other similar images organized in databases.
The second, perhaps even more difficult, is to translate, albeit in a lamely way, into a spectrum of human competences, the results of the first operation. By asking the machine to describe "in English" what it could see, Goodwin is actually subsuming its perception to a restrict horizon within a human capacity to describe the world.
Nevertheless, for the machine, "reality" doesn't work exactly that way.
Goodwin in particular is interested in how machines "perceive" humans -[...] his last project featured a camera that would send the images it took to a crude artificial intelligence that then attempted to describe it in English. It was funny, and, for a second at least, it opened a channel between the human and mechanical mind. (VICE, 2015).
Media archaeology seeks precisely to describe the world through this nonanthropocentric perspective, a perception of things themselves. This is paradoxical, of course, since the knowledge produced, after all, is destined to a human audience. The archaeological exercise, here, serves as an alternative that can reveal a redistribution of agencies in certain social systems. It is useful in confronting us with the fact of how far we are from an apprehension of the concrete world at a time when it is practically impossible to avoid technological mediation.
Elsaesser (2018) understands the digital shift as responsible for the adoption of archaeological bias, especially in the audiovisual field, as a practice and as an analytical concept. The widespread digitalization of production and circulation processes relativized the functions of various machines, practices, industrial processes, economic activities and personal habits, suppressing the differences between them and proposing a common praxis. In doing so, it led to the erasure of part of the history of media and accelerated the process of obsolescence of supports and devices.
The changes in technological, social and cultural realms needed to be thoroughly understood and, for this, the historical character of the techniques, their organization in layers of time, needed to be highlighted. The bridge between disused and more current practices was a way of attributing to the new digital media a certain origin, even if they proposed strange and non-existent functions in previous analog media.
By inserting new media into a kind of genealogy, it would be possible to analyze them more comfortably within contemporary cultural contexts. With this non-continuous For an autonomous existence of images: an archaeological perspective organization of different times, the feeling of an inexorable evolution of technologies can be questioned: "There was a distinct need for explanation, a review of the layers of time to accommodate all these transformations, but also to reconsider the very nature of the change itself" (ELSAESSER, 2018).
Being "close to the machine", in the words of Ernst (2011), means understanding the functioning of devices, referring to a more pragmatic sphere of things' possibilities in a technological regime.

Creating narratives about "real" events
The conversation between different digital networks -where there is no human intervention -also generates digital traces. These can remain indefinitely dormant or may be recovered somehow. However, there is no material evidence of their existence: neither from the actions of the machine, nor human ones. Although the idea of transparency is commonly associated with computers, they do not work quite that way. Rather, behind the interface that allows the access to information we have layers and layers of codes with which we are incapable of communicating.
Understanding the code as logos, in the words of Chun (2011), is to perpetuate the idea that the software would be able to "create" things. Automatic programming, by freeing the programmer from direct compilation and repetitive tasks, allowed programs to produce other programs, creating a higher-level language. There would be, therefore, a greater control over the final product. However, this delegation of responsibilities decreased the effective control of developers over the machine, making the process more opaque.
As databases work more endogenously than refer to external factors, the relationship between information and "reality" seems to be weakened. The informational construct meets the needs of logical commands determined by search engines, data mining etc. The formulation of "truths" about this or that pattern of the database is a narrative determination that, as we have seen, is of no interest to the machine.
Therefore, "reality" is an artificial notion, which requires a rearrangement of prior knowledge. Once images and sounds come to us mediated by multiple layers of software, they present us with a paradigm, which lacks an ordering syntagm. An ontological rupture regarding the relation between the proliferation of images and their connection to concrete reality faces us. As for analog images, the fact of someone not being present at the moment of their production does not make them less real or less concrete. Likewise, the fact that there is no physical record on digital media forms does not make digital images less true. Chun (2011)  At the same time, these images have no circumstantial existence indices.
We cannot rely on a strip of negative film to prove a fraud: there are no laboratory procedures and no revelation chemical processes. The verification of authenticity in digital images is harder. But, like most processes mediated by computers, digital images resemble or act like transparent media, providing access to "reality" rather than a "dirty window" dealing with the opacity of the visible.
By thinking this way, we could assign a visibility of electronic processes in contrast to obscurantist analog forms. In their struggle to abolish noise, digital media make clear the commitment uniquely to the recorded object (a figure without a ground) and not with the interference from the machines. Noise jeopardizes the privileged view of the viewer, hence it would depart him/her from reality. This position ignores how mediated digital images are, excluding part of the experience considered as unimportant, but, at the same time, compromising a comprehensive apprehension of things. The control of images is so unavoidable, that, paradoxically, it seems that recording them is a quasi-direct perception, not conditioned by material determinants. It is actually a simulacrum of transparency.
For an autonomous existence of images: an archaeological perspective

Conclusions
Ubiquitous information systems and the abundance of digital images are important aspects of new media. As in the case of computational processes, these media intend to be transparent, but work through a peculiar logic. Understanding the number of dispersed images on the Internet as a paradigmatic set upon which search engines do not produce any subjective interpretation is to multiply the possibilities of narrative concatenation of the same images. Following the arguments of Chun, software, interfaces, computers, and digital imaging, work ideologically. Digital technologies and its extensions in society work with a logic of causality -a system ordained by codes, which, although not accessible, propose a blind faith in their own reason. The truth is not an accessible one, but we believe it to be functional.
The truth of the images is born within them. It is related to the network of associations created between devices, files, computer users and spaces. The associations are so ephemeral and fragile as vast are the nodes produced. Any narrative generated from the database of images is part of a larger and virtual repertoire of possibilities. There is no chance that any of them can work assertively as a representation or documentation of the "real." As we have seen, the truth of the images does not imply a direct relationship with the "real" of the world. The essentialist meanings, criticized by Latour for an excessive emphasis on large structures, facilitate the narrative construction since they attribute effects of cause and consequence to the relationships between objects. When arranged in a network and, mainly, when made visible by the organization in databases, technical devices do not follow the narrative logic of the facts. At the same time, they contribute, paradoxically, to the construction of all possible narratives.
The "truth" of images does not imply a direct relationship with the concrete world. Any narrative generated from the image bank is part of a larger and virtual group of possibilities. These aim to attribute meaning and verisimilitude to the facts through an a priori of impersonality of technologies. Moreover, following Latour's premises, the "imperfection" of devices puts them in a state of eternal visibility.
The intentionality of images, especially those that are produced and circulate in digital environment, is the symptom of a contemporary episteme that delegates to objects not just a functional autonomy, but also one of existence and of description of the world. The multiplicity of digital images makes of them Beings that exist beyond the human and that constitute a kind of continuous phenomenological machinic process, an awareness of the self and of the other.