A Peer-Reviewed Journal About MACHINE FEELING

A Peer-Reviewed Journal About

MACHINE FEELING

Mitra Azar Daniel Chávez Heras Michela De Carlo Iain Emsley Malthe Stavning Erslev Tomas Hollanek Rosemary Lee Carleigh Morgan Carman Ng Irina Raskin Tiara Roxanne Rebecca Uliasz Maria Dada Tanja Wiehn Brett Zehner

Christian Ulrik Andersen & Geoff Cox (Eds.)

Volume 8, Issue 1, 2019 ISSN 2245-7755 1 Contents

Christian Ulrik Andersen & Geoff Cox Editorial: Feeling, Failure, Fallacies 4

MAKING SENSE

Iain Emsley Iteracies of Feeling 10

Irina Raskin Machine Learning and Technoecological Conditions of Sensing 20

Maike Klein Robotic Affective Abilities 34

(UN)BEING 47

Brett Zehner Machines of Subjection: Notes on A Tactical Approach to Artificial Intelligence 48

Maria Dada Queering Global Information Systems 58

Tiara Roxanne Digital Territory, Digital Flesh: Decoding the Indigenous Body 70

Rebecca Uliasz Assemblages of Desire: Reappropriating Affective Technologies 82

FEELING GENERATORS 95

Carman Ng Affecting Reality: Intersecting Games, Trauma, and Imaginaries 96 Malthe Stavning Erslev I forced a bot to read over 1,000 papers from open-access journals and then asked it to write a paper of its own. Here is the result. Or: A quasi-materialist approach to bot-mimicry 114

Michela De Carlo Synthetic Bodies and Feeling Generators 128

Tanja Wiehn (Un)Predictable Acts of Data in Machine Learning Environments 142

SEEING THINGS 155

Mitra Azar POV-matter, Cinematic POV and Algorithmic POV between Affects and Umwelten 156

Daniel Chávez Heras Spectacular Machinery and Encrypted Spectatorship 170

Tomasz Hollanek Non-user-friendly: Staging Resistance with Interpassive User Experience Design 184

Rosemary Lee Operative Image: Automation and Autonomy 194

Carleigh Morgan Calculated Error: Glitch Art, Compression Artefacts, and Digital Materiality 204

Contributors 218

A Peer-Reviewed Newspaper Journal About_ ISSN: 2245-7755

Editors: Christian Ulrik Andersen and Geoff Cox Published by: Digital Aesthetics Research Centre, Aarhus University Design: Mark Simmonds, Liverpool CC license: ‘Attribution-NonCommercial-ShareAlike’ www.aprja.net EDITORIAL

FEELING, FAILURE, FALLACIES

Christian Ulrik Andersen & Geoff Cox

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. EDITORIAL

capture of everyday styles, expressions, Feeling preferences, sentiments, and so forth – the very means that Williams alludes to. Digital culture has become instrumental for If, in general, machine learning appears capturing and managing what Raymond to lack an affective dimension, then in what Williams would once have called “struc- ways are we to understand its resolute and tures of feeling”. The journal issue A Peer- concerted pursuit of this? What old registers Reviewed Journal About Machine Feeling of processing culture and organizing time, alludes to this, and points to a material space and power does it build on? What analysis of aesthetics and culture, including potential new sensibilities and structures of its technical and social forms, and in the way feeling may arise in such normalized registers that this concept was originally employed of our habits? What new cultural and social as an acknowledgment of the importance of forms and practices emerge in the coming the hard to capture dimensions of everyday together of machine learning and structures life. Styles, expressions and sentiments are of feeling? In each their own way, the authors always in flux, yet Williams, and others after in this journal explore these questions. him, have with this term argued that they are grounded in cultural history and specific everyday situations. In developing a critical and analytic understanding we should therefore Failure turn our attention to changes in language, style, aesthetics and those social forms To capture moments when new patterns of which are active in the present, but not yet experience emerge, when people start to fully formed or captured by a conceptual or think differently, when new sensibilities arise scientific knowledge framework. Taking their will first and foremost depend on a large set point of departure from Williams, Devika of training data — sound, text, biological Sharma and Frederik Tygstrup write: data, and more that can be used for image recognition, sentiment analysis and more. At We recognise the facts of cultural a more general level, these datasets absorb life once they are established and all kinds of social and cultural production; institutionalised, but we tend to miss they seek to absorb every moment that peo- those moments when new patterns of ple start to think, act, sense, and experience experience emerge, when people start phenomena in new ways. to think differently, when new sensibili- There is a certain paradox in this. As ties arise, when habits swerve. (4) pointed out by Matteo Pasquinelli, machine learning is a paradigm of intelligence that This journal issue further explores fails to provide a methodology of failure. this line of thinking, and more specifically What people generally refer to as artificial responds to the current developments in intelligence and machine learning is merely machine learning and the ability of technolo- a statistical mapping of correlations in the gies to capture and structure feelings and dataset. Because of this, machine learning experiences that are active, in flux, and in will reduce the least common structures in the present; for example, in the ways that the dataset, simply in order to reduce calcu- automated experiences of seeing and read- lation costs. Consequently, machine learning ing begin to produce knowledge through the is not a sign of cognition, but of compression

5 APRJA Volume 8, Issue 1, 2019 as a means to efficiency, which on the other Zehner); feeling generators (Malthe Stavning side is also a loss of diversity; failure does Erslev, Michela De Carlo, Carman Ng, Tanja not exist. In this, he claims, machine learning Wiehn); and seeing things (Mitra Azar, Daniel seems much more aligned with a history of Chavez Heras, Tomasz Hollanek, Rosemary optical lenses who operate by resolutions Lee, Carleigh Morgan). and diffraction. This is what he calls statisti- There is more than a hint of Williams cal cinema. This problem of generalization or (and his cultural materialism) across these “regression towards the mean” is mathemati- positions in recognition of the ways that cal but not without political consequences. certain ideas (such as affect theory and machine learning) achieve hegemonic status. We, as contributors to this journal issue, all feel/felt the weight of history and privilege Fallacies here, not least as the workshop leading to the publication was held at the University What then is the role of researching digital of Cambridge where Williams himself once culture and machine feeling? On the one taught. The setting for our (and his) work is hand, to follow Williams and capture the clearly an important issue if we take struc- “habits that swerve” seems to be relegated tures of feeling seriously and recognise that to corporate research institutions that seek to the contents of a journal such as this are a align calculation costs and statistical resolu- consequence of a wider factors that include tion; institutions that perform the statistical actual work, social relations, and place of spectacle of contemporary digital culture. On production: “it is a trivial fantasy to suppose the other, could researching machine feeling that these general and pressing conditions be regarded as an interrogation of the fail- are for long or even at all separable from the ures of machine learning; or, even providing immediate and the personal”, as Williams a methodology of failure that machine learn- puts it (Culture and Materialism 222). Herein ing otherwise lacks? lies the tension between received forms and This kind of research could take differ- lived experience, of structures of feeling. ent shapes. For one, it might address the implied inclusions and exclusions that are at Thanks to all authors as well as further play in the politics of research, such as the contributors to the workshop (Anne intersectional feelings of race, gender, and Alexander, Alan Blackwell, Anja Breljak, class. It might address the emotionalisation Jennifer Gabrys, Kristoffer Gansing, of not only politics and a people born to Leonardo Impett, Matteo Pasquinelli, Søren feel (which seems to be intrinsically related Pold, Winnie Soon, Magda Tyzlik-Carver, to the statistical spectacle), but also of re- Martin Zeilinger), a collaboration between search itself and how it links to subjective transmediale festival, Aarhus University, patterns of experience. The contributions and Cambridge Digital Humanities Learning to the journal resonate with this approach Programme. and expose some of the fallacies at work in research processes once feelings are en- We dedicate this issue to the memory of gaged. The subsections of this journal reflect Sascha Pohflepp. this problem: making sense (Iain Emsley, Maike Klein, Irina Raskin); (un)being (Maria Dada, Tiara Roxanne, Rebecca Uliasz, Brett

6 EDITORIAL

Works cited

Pasquinelli, Matteo. “The Undetection of the New.” Machine Feeling. University of Cambridge, 14 Jan. 2019. Keynote at workshop.

Sharma, Devika and Frederik Tygstrup. “Introduction.” Structures of Feeling: Affectivity and the Study of Culture, edited by Devika Sharma and Frederik Tygstrup. Berlin: De Gruyter, 2015. Print.

Williams, Raymond. “Structures of Feeling.” Marxism and Literature, Oxford: Oxford University Press, 1977, pp. 128-135. Print.

––– Culture and Materialism: Selected Essays By Raymond Williams, London: Verso, 2005. Print.

MAKING SENSE Iain Emsley

ITERACIES OF FEELING

Abstract

Computational readings of culture allow us to pose new questions or create new cultural forms supporting new forms of critical thinking and reading. Yet the machine may not be able to identify some of the qualities, such as emotion, that might be central to the question raised. Using the Next Rembrandt project as a case study, this paper suggests an approach to consider the medium as the site of meaning making in digital culture and how this affects critical practice using Raymond Williams, David Berry and Jacques Derrida. In the first part, I consider the idea of reading with machines and how this might be considered within the medium. The second part uses iteracy to find meaning in the models and how this might reveal new critical paths through readings of the image. The final part presents a reading of the digital object itself and how these can be used to create a space for meaning to come into being. Through this, the article raises questions about critical techniques for understanding the material object in distant reading methodologies as ongoing research.

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Iain Emsley: ITERACIES OF FEELING

Computational readings of culture allow us This suggests two critical responses to pose new questions or create new cul- that I will explore in this paper. The considera- tural forms supporting new forms of critical tion of structures of feeling requires not only thinking and reading. Yet the machine may human reading but also technical reading not be able to identify some of the qualities, itself using models to understand the digital. such as emotion, that might be central to the Reflecting on Hayles’s sense that print is question raised. The advantages of distant shallow but code is deep, I suggest that this reading, such as scale, may be tempered by develops cyborg reading, where the “reader a realisation of what may be missing. necessarily is constructed as a cyborg, Using the Next Rembrandt project as a spliced into an integrated circuit with one or case study, the aim of this paper is to sug- more intelligent machines” (85), a technique gest an approach to consider the medium as to interpret the medium’s discourse. I want the site of meaning making in digital culture to develop this through the way that reading and how this affects critical practice. In the digital culture means reading with machines first part, I consider the idea of reading with to understand the data. machines and how this might be considered I want to develop this reading as an within the medium. The second part uses experimental process as well as consider- David Berry’s iteracy to find meaning in the ing the materiality of computational culture. models and how this might reveal new criti- Berry’s iteracy, “the ability to read, write and cal paths. The third part presents a reading understand processes” (190), is a key to un- of the digital object itself and how these can derstanding the artefact and to interact with be used to create a space for meaning to its relocation of epistemology. This uses both come into being. cultural hypotheses and evidence to test how Raymond Williams’s structures of feel- the data is being created, so placing a hu- ing suggest a way of beginning to think about man meaning into the process. Building on this new understanding. I build on this con- Berry and Fagerjord’s call that “culture […] ception that “a cultural hypothesis, actually is materialised and fixed in forms specific to derived from attempts to understand such material digital culture”,(142) I want to think elements and their connections in a genera- about how the presentation of the final form tion or period, and needing always to be re- reveals and hides the metamedium nature turned, interactively, to such evidence”(133) of the digital, capable of transforming exist- by situating the elements and connections ing media and creating new media and within a digital reading. Computational read- technologies. ing derives features from the data based on human thought and interpretation of the hypothesis, either in the construction of algorithms or labelling of data. Once identi- Next Rembrandt fied, the features may then be analysed or combined to create new structures and In this section, I want to consider the Next elements. Qualitative feelings such as emo- Rembrandt[1] and to think about how it is tions become uncertain elements that the read with machines. The project was an ex- quantitative seeks to understand through periment to create a new picture from a read- models. I will suggest that a critical reading ing of Rembrandt’s portraits, shown in Figure of the digital object reveals ways in which the 1. The Next Rembrandt is an algorithmically human might be understood and to suggest generated image by a partnership of J. Walter a critical practice. 11 APRJA Volume 8, Issue 1, 2019

through new models of space, location, and artistic models to make it visible to the human. Culture needs to be read through computational remediation. Yet the work appears to capture human aspects, such as emotion. Emotional reactions may have alternative meanings in the technical world. They may be seen as signs of both com- Figure 1: Image of the Next Rembrandt. Photo: J. Walter Thompson Amsterdam. mercial and personal engagement or a data point in a model to suggest new content or Thompson Amsterdam, ING, Microsoft, to try to refine models of how to understand TU Delft and the Mauritshuis. Using high humans. Even these emotional points can be resolution 3-dimensional scans and digital limited to a recognised and constrained set versions of 346 of Rembrandt’s portraits, of emotions. We need to consider the con- machine learning algorithms identified the texts they exist in. Emotional markers may key points to be created from 150 Gigabytes be read through sentiment analysis or from of generated graphics. Using a mix of cloud a reaction on a page by a machine looking and physical machines, the algorithms were for engagement or a commercial opportunity. tested and run in parallel. The rendering of Or is it learning how to recreate our reality the portrait took 500 hours on machines and through a numeric system? a 3-dimensional printed picture is created from 148 million pixels. The computational aspects both hide themselves behind and are bound into the impression of paint. Reading with machines This impression is an imitation made up from a digital reading of the drops of paint We need to read with machines to begin and the intentional layering to be recreated accessing this culture and understanding its in 3-dimensions for the physical portrait. The new forms. created form echoes an earlier tradition to The visual layer invites a human read- present itself as a singular object. A closer ing of the image and to infer the emotional reading suggests that Finn’s assertion that states represented in the eyes and the wistful “code can be magical” (5), where the code is mouth. Within the given boundaries, the alan agent in changing the world and the mind, gorithms create an image using an intensely is at play here. The layered paint sections hint close reading of colour palettes and shades at algorithms used to create them as does but can only imitate emotion. A reading of the the gaze in both editions of the painting. The object as a visualisation provokes questions attempts at authenticity in both digital forms about how the representation is considered, provoke a desire to read it more closely, to either as a close reading of the portraits so understand the entwined cultures, but the enhancing a trait that it reads or whether the machine resists traditional cultural readings algorithm has created it? to fully tease apart the layers. Authenticity Having considered using machines becomes inauthentic, unless it is read with a to generate and process the data, I want machine or a machine in mind. to think about how we might begin to think As a digital object, a machine is re- with the machine and considering how they quired to read and render the cultural data might generate knowledge. At the very least,

12 Iain Emsley: ITERACIES OF FEELING we must admit that we need to read with the from the evidence, to which the evidence machine. Born digital culture, like the Next is returned once the process has tested it. Rembrandt, cannot be read by a human; it I would suggest that this understanding is requires computational remediation. Such art based in computational materiality. alters our critical relationship with machines The project is a close reading of por- and raising a crucial role for questioning the traits conducted by machines to create a medium itself as site of cognitive practice machine-readable data set. The algorithms through remediation. identify relevant parts of information. The Using a machine to write data suggests resulting image requires computational re- that it is required to read and remediate it, so mediation to convert the numeric world into a using it as part of the interpretation through human readable one. In a very real fashion, the models encoded into the process. It may we can only read the image with a machine. be mediated through visualisation or sonifica- This echoes Adorno and Horkheimer’s cul- tion processes, providing another area that ture industry that broadcasts and replicates needs to be understood. Instead of reading itself, where we rely on technology to create data, we read presentation models that af- and remediate culture. fect hypothetical models as a strategy of not reading (Clement; Moretti, “Conjectures in World Literature”). This practice accepts that the quantity of information cannot be read Reading with machines at a close level, by humans but that broad patterns can be viewed through machines. We begin to read with machines and to un- In cultural terms, this builds on Moretti’s derstand how both sides form and contribute concept that “distance is not an obstacle, to digital culture. but a specific form of knowledge: fewer By this, I mean that we need to consider elements, hence a sharper sense of their not only the interface and how that creates a overall interconnection” (Moretti, Graphs, reality but how we can use any given options Maps, Trees 1). These abstractions, allowing or even access to the algorithms to consider the reading of patterns over specificities, are the logics at play. Reflecting on the roots of digital structures used to support interpreta- iteracy as iteration, I want to think about how tion or remediation. Machine interpretation it can be used to repeat a process, perhaps may also be fuzzy and not show outliers or with alterations, to allow the algorithm to emerging patterns if they are too slow and be the point of interaction. Through making long, suggesting that the subtleties of emo- changes, user meaning can be given to the tion may be aggregated through counts into machine to continue hypothesis testing. clusters of readings at the machine level. When Next Rembrandt was being created, Although brought together as one im- algorithms were repeatedly run in parallel. age, the picture is a series of algorithmic The repetition of these processes provides observations. Each of these is a specific a space for the human thought to enter the form of knowledge gained through the distant process and realise the potential of Ramsay’s reading and pieced together through other algorithmic criticism (33) to reconceive both forms of knowledge. The machine, through the form and criticism’s logics in a playful its learning algorithms, uses a hypothesis form. Tweaking the parameters and repeat- to test its understanding and creates an im- ing the process not only reveals the process age. Its understanding of a structure is taken through which the picture is made but also

13 APRJA Volume 8, Issue 1, 2019 allows humans into the iterative loop and interrogate the structure in its new tradition. realise the hypothetical nature of the work The newly created digital reading raises through experimentation. This site of interac- questions as to which tradition is being de- tion moves human cognition into the machine veloped. The artist becomes the subject of so as to embed the concept of thinking with the work rather than the creator, though he is the machine and its models. elided from the public gaze. As well as read- The materiality of computation being and showing the new structures of feeling, comes more apparent through these acts. the object itself is not part of Rembrandt’s tra- Researchers defined that the resulting dition. Aside from the ownership questions, person was Caucasian, male, wearing black the newly created picture exists within an clothes with a white collar and between thirty alternate context. Material questions about and forty. This suggests a machine logic whether it can be considered as a work of that might recognise the image but requires new media art or data visualisation and what guidance through wider cultural nuance that its relation to Rembrandt’s oeuvre might be? might be either difficult to model or statistically Is it a creation or an analytical work? Digital insignificant. This is translated into a model reading of the portraits detaches this from the or set of constraints. The limitations of the paintings and creates a new tradition through machine’s cultural understanding become algorithms and processing power. more visible as does the imposition of human Infrastructural questions can be raised. values into the reconstruction algorithms. One project partner, ING, fund cultural insti- Taking an active stance in considering how tutions, such as the Rijksmuseum, enabling the computational both reads and writes the cultural institutions to remain open. However, data reveals not only different meanings but it might also be read as patronage. A culture suggests new critical practice. industry arising from the financial ability to Using machines raises questions about support human endeavour and the physical culture. Are new cultural forms appearing: infrastructure. JWT Amsterdam also paid for ones that can be appreciated by both ma- the physical version of the painting to be cre- chines and humans? What forms of culture ated. This provokes further questions about may arise from this? Who owns the created the relationship between digital and physical form? Is a new culture industry being created artistic culture. The digital has the potential through the use of social media or infrastruc- for writing, assuming the protected mode is ture companies to create cultural forms? off. An act of execution and change, writing I want to take a brief pause to consider is a permissioned act within computation. the critical theoretical response to this posi- The machine owners may grant or deny the tion. At one remove, the process of creating permission for non-owners to write any data the model of the image reduces the human without a visible infrastructure. As such, the to a set of constructs, such as average width culture shown is one where the non-corporate between the eyes, which is then broadcast entity is deemed lower and granted read only to the viewer. The digital can reproduce the permissions. image in a variety of forms from the same Patronage can be rethought through underlying data and the results of the im- the computational. The scale of the data pro- perfect structures are encoded in this view. duced by the project as well as the amount Benjamin’s assertion that “reproduction de- of processing power needed to run the facial taches the reproduced object from the domain recognition and rendering processes on of tradition” (215) can be operationalised to such a large digital object suggests that new

14 Iain Emsley: ITERACIES OF FEELING platforms are required. The infrastructure is and printed in minor layers. The computa- provided by Microsoft in this instance. Both tional re-presents the paint medium as an of these require financial and computational abstraction that requires a deep reading to power so renewing patronage as a computa- understand the artifice of a natural process. tional form. This embeds such providers into Paint drops become composed rather than the warp and weft of digital culture. Without a accidental. As well as remediating the data return to reading with machines and consid- into new forms, new tools and conceptual ering their logics, we return to a transmit only processes are required to understand the culture. Technical specialisations create the materiality of the object and how these fits in conditions for a read only culture, so project- with existing traditions. ing their dominance into a cultural sphere. A consideration of the image involves its The rendered figure, through its created methods of creation. From here, we need to class and social standing, perhaps points to extend Hayles’s notion of thinking through the the role of the creative partners in creating network to consider the physical machines, the image. such as printers, and materials involved. The face is a construction of models The printed image was not only made on a and aggregations rather than being read and 3-dimensional printer but through layers of interpreted from a sitter. Critical questions printing substrate, though both are controlled remain about whether this project was a safe by the file made from the image. As well as space to develop and use facial recognition encouraging us to read in different ways, we and reconstruction algorithms. This project need to think about the techne itself and how creates an ethical safe space to reflect on this supports an epistemological reading. these algorithms within a known set of biases, Iteracy’s root as literacy provokes ques- ones defined in the underlying data set and tions of how one might read or listen to the the parameters given for the reconstruction. results as abstractions and patterns. The I want to raise these as critical issues to be act of interacting with the process embeds a able to interrogate the created image and the human element in part of it, suggesting that assumptions that give it focus. the object being read comes from thinking The born digital image is both remedi- through a network. Next Rembrandt may be ated into a human readable image and ren- read as an image but to understand it, one dered as an artefact that I want to think of as needs to consider new practices of reading reborn analogue. This latter form, the printed and meaning making. In many aspects, image, continues to challenge the concepts of this is a technically demanding reading. tradition. Through being made into a physical We might feel the sadness and warmth in object, it is placed into a museum setting that the sitter’s eyes or the slightly worn look the funders support as well as what might derived from the way the light plays on the be considered an old media cultural setting. features and through the layers of paint. I The image is also available digitally and can contend that we are inside an interpretational be copied and reproduced. This latter tradi- loop, reading the evidence supplied to us tion that is represented is one driven by the from a hypothetical model encoded into the technological medium as something that can process. The machine uses aggregations of be easily shared at minimal effort. What the the models and the data to create a new set physical print elides is that it was printed from of data points derived through a model. The the digital file. Generated from the digital file, underlying algorithms create a numerical ‘paint’ is calculated through machine learning reading, themselves bound within what the

15 APRJA Volume 8, Issue 1, 2019 limitations of the algorithms and the hard- but it shows a need for a critical practice to ware. Its surface is a visualisation, where determine where the model might come mappings mediate the numerical data into a from and how it is represented. As the data new point, which humans perceive as colour is being rendered, the model’s values are at a location. From a human perspective, we being applied through the processes. The note the stylistic similarities, the attention to reconstructive stage shows the machine’s detail in the style and the emotion in the face. iteration as it mapped the facial features to There is a disjunction here between the two proportions until it achieved the final image. readings that reveals the need for conceiving Through testing the image, the algorithms about how this can be critically approached. are testing themselves. Using this, one might Read together, these data points begin to read the intention behind the models that are suggest the underlying logics, such as the shown and to understand the two readings position of light, as we move from a macro- to available – the numerical and the rendered – a microanalysis. With the assumption that we and to probe its limitations. are unaware of the human provided limita- A key point is Williams’s issue with the tions, the reading can interrogate how the specifics of what constitutes an element in machine reads the data to project a model of discourse is further problematised through its understanding. translation and encoding required for the I want to turn to models as an integral machine to understand them as hypothetical part of these computational structures. constructs. A new discourse is created from McCarty echoes Weizenbaum in consider- the results, which require reading when it has ing computational systems as dependent on bene returned to the evidence from whence it the models given to them to understand a came. The underlying computer model both conception of the world. The use of Artificial makes and is made from the translation. Intelligence to create data sets and models This alters the location of epistemology from raises questions of who is the designer the reading and interpretation to within the and whose world is being created? The computational. A necessary consequence model’s structure of an element rests on is a potential change of the location of the how the designer or implementer translates element’s negotiation. and transcodes the element into their work Whilst it may happen as part of wider as well as the model’s purpose. The model cultural discourse, it is happening within the itself requires critical consideration of what is algorithms and their models of the world. As being modelled and what is being presented discussed, human intervention can help to through the computational. mould the uncertain elements into an appre- The use of the machine suggests that ciable form through a combination generated the computational materiality needs to be from iterate readings. considered for what is being modelled and presented but how and the values that lie within the processes. Weizenbaum’s consideration that the “symbolic recreation of [the The digital object designer’s] world” (18) may be read in two as pharmakon ways. Firstly, the model and its associated processes reflect their purpose and process. Having considered the image and its con- Secondly, the medium affects the object sequences for realising the digital, I want through its own limitations and understanding to focus on the digital object itself. Having

16 Iain Emsley: ITERACIES OF FEELING discussed potential ways of making mean- the layers. Taking Manovich’s conception of ing, I want to illuminate the material that the metamedium (101-102), critical practice reshapes both it- and ourselves as the site of becomes a tool of and about the medium. cognitive practice. This practice, as shown above, does not I see the pharmakon acting as dis- necessarily need to use computational tools course. Derived from Plato’s view that writing to be reflexive but can also be theoretical “will make them remember things by relying by bringing the object into a different being. on marks made by others, from outside Having suggested that the digital object is themselves, not on their inner resources” a pharmakon, I want to extend the reading (69), Derrida suggests that it “acts as both through Derrida’s différance and the use of remedy and poison” (Derrida, Dissemination play that it reveals. 73) and injects itself into discourse. At one As human and machine discourses mix, level, text provides a discourse that can be they reconstruct their own context into new read and shared, yet it also removes the discourse. The ontotheological message of ability to query the underlying discourse and a machine reading data becomes one with remakes it in its own symbols. This imitation the potential for multiple meanings. As the of practices suggests that there is a compu- model is read and processed as “a sort of tational cognitive economy where only those writing” (Derrida, Of Grammatology 56), its who can create the tools to understand the form is recontextualised, moving from an im- digital object may interpret it. I suggest that age through numerical models to become its by exposing the pharmakon, one can bring own grammatology. Realising that the object different tools to understand it. This sug- is made up of these changes recognises the gests an alteration how we think the digital différance, the gap created between the signi- affects writing. Where Plato’s writing loses fier and the signified when the computational both access to memory and the underlying elides itself. Although based on a learned discourse, the object is central to both as the aggregate set of elements, like the colouring locus between humans and machine cogni- and the geometries involved, a human read- tive practices. It both creates and transforms ing may infer emotion into Next Rembrandt’s the cultural forms, acting as memory and eyes or face and realises a human part of the discourse to express them. language. Where machine process may be The Next Rembrandt image is a medi- limited in their qualitative meaning making, cine in its form. Without reflection of its mate- human readers may recognise the possibili- riality, hidden by the (in)authentic surface, it is ties of the elements that exist as a series of a poison. Where the textual medium removes interpretative gaps that expose the potential access to an oral discourse through remedia- for new critical readings of the image. tion, the digital can be remediated into dif- Alternate considerations might be ferent media though the original language is brought to the eye as a form of play. Derrida’s computational. It can be accessed using tools consideration of play as a de-centring of and with permissions. Understanding that meaning within bounds allows critical logics the digital can be presented in different ways, to be reconfigured. The new readings al- such as a born digital or a printed picture, low for interpretation to take place through creates the space for a critical gap to appear. experimentation, continuing the suggested Even using machines and programming lan- move from a digital culture that projects its guages, one has to acknowledge the trans- meaning. Iteracy and play work together lations and transcoding to converse across with the pharmakon to create new contexts

17 APRJA Volume 8, Issue 1, 2019 and movements between remedy and poi- new cultural forms that are returned to the son. Different structures of feeling might be evidence that it came from. By understand- viewed as evidence to be interpreted. Our ing these processes as a mix of human and readings work to remake meaning within the machine discourse, we can think about how boundaries through combining human and to both interpret and interact with them. machine meanings. Doing this, we begin to Iteracy encourages not only a different form recontextualise the medium as a site of cog- of reading but also critical engagement with nitive practice where readings combine and the underlying discourse, so considering recombine. I want to use reading with ma- the medium as the site of cognitive practice chines to think about critical practice beyond where discourses mix and create interpreta- the interface and at the level of the medium. tive gaps. The claiming of the computational Reading with machines supports as a metamedium provokes the need for new methodological changes that are hinted at by practices of making meaning that consider Williams. The element’s existence and its in- the medium. These theoretical considera- terpretative possibilities as part of an emerg- tions are the subject of ongoing research into ing discourse is problematised through this the digital object as a core concern in distant process. The remediation of the images in reading methodologies. different media suggests that the presented Rather than seeing the digital mediation evidence is a poisoned reading that is guided of cultural forms as a machine-driven pro- by the models of presentation. Translating a cess, I contend that considering them within feature into a series of technical languages the medium opens up new forms of critical to create a new model and element alters the interpretation and techniques that use the discourse and its specificities. It is only by revealed discourse. From this we understand looking for the imperfections in the surfaces that computational structures of feeling be- that we are able to begin a reading that cri- come imperfect structures of feeling. tiques these discourses. By taking part of the presented data, hypotheses can be formulat- ed and tested. Meaning can be interrogated by altering parameters and questions to test the new way of thinking and interpretations, while recognising that the structures of feeling may be made of other structures. The evidence that we are examining for clues is made of other evidence and hypotheses that is generated from the machine.

Conclusion

Using Williams’s definition of structures of feeling as cultural hypotheses, this paper argues that they might be seen models of thought that are translated into computational models. The evidence is used to generate

18 Iain Emsley: ITERACIES OF FEELING

Moretti, Franco. Graphs, Maps, Trees: Notes Abstract Models for Literary History. London. Verso. 2007. [1] Next Rembrandt, https://www.nextrembrandt.com/. Moretti, Franco. “Conjectures in World Literature,” Moretti, F. Distant Reading. London. Verso, 2013. Works cited Plato, Waterfield, Robin. (trans). Phaedrus, Oxford. Oxford University Press. 2002. Adorno, Theodor and Horkheimer, Max. Dialectic of Enlightenment. London. Verso. Ramsay, Stephen. Reading Machines: 1997. Towards an Algorithmic Criticism. Chicago and Springfield. University of Illinois Press, Berry, David M., Critical Theory and the Urbana. 2011. Digital. London. Bloomsbury. 2014.

Weizenbaum, Joseph. Computer Power Berry, David M. and Fagerjord, Anders. and Human Reason: From Judgment to Digital Humanities, Cambridge. Polity, 2016. Calculation. San Francisco. WH Freeman & Company. 1976. Clement, Tanya E., Steger, Sara, Unsworth, John, Uszkalo, Kirsten “How Not to Read A Williams, Raymond. Marxism and Literature. Million Books,” 2008, http://people.virginia. Oxford. Oxford University Press. 1977. edu/~jmu2m/hownot2read.html.

Derrida, Jacques, Johnson, B (trans.) Dissemination. London, Bloomsbury, 2016.

Derrida, Jacques, Spivak, Gayatri Chakravorty (trans.) Of Grammatology (40th Anniversary Edition). Baltimore. John Hopkins University Press, 2016.

Finn, Andrew. What Algorithms Wants: Imagination in the Age of Computing. Cambridge, Massachusetts. MIT Press, 2017

Hayles, N. Katherine. “Print Is Flat, Code Is Deep: The Importance of Media-Specific Analysis”. Poetics Today. 25 (2004): 67–90. https://doi.org/10.1215/03335372-25-1-67.

Manovich, Lev. Software Takes Command. London. Bloomsbury. 2013.

19 Irina Raskin

MACHINE LEARNING AND TECHNOECOLOGICAL CONDITIONS OF SENSING

Abstract

In what way can machine learning be understood as a computational mode of sensing? How does the practice of making sense take place in the context of developing machine learning applications? What assumptions and conflicts are constitutive for that very process of sensing? Bringing case studies from machine learning into conversation with theoretical work primarily by Erich Hörl, Luciana Parisi, Wendy Hui Kyong Chun and Karen Barad, this article reflects on the re-configuration of sense in the course of the expansion of media-technology. It questions how computational expressions become relatable as well as the mechanisms for encapsulating the capacity of sensing for determining purposes.

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Irina Raskin: MACHINE LEARNING AND TECHNOECOLOGICAL ...

of performativity (cf. Chun, Barad). From Introduction the perspective of performativity, the critique against an instrumentalization of computa- The expansion of media-technology leads to tional sensing for governing reality cannot an extension of the remediation of artifacts be exclusively addressed in regards to its and signs. It proliferates the evolvement of partaking in the determination of sense. machinic programs into relatable environ- Rather, it asks for taking into account the ments for sensing. Furthermore, it takes measurements for evaluating and shaping part in a re-configuration of sense and the the process of determination. This implies conceptions of how meaningfulness is con- to direct one’s attention towards the appa- stituted. As argued by German philosopher ratuses and its infrastructure that sustains Eric Hörl, under a media-technological algorithmic sensing, towards questioning condition of sensing, a “technoecological” how one relates to the expressed sense as form of sense-making is disclosed. This well as the accountability of computational emphasizes the operative dimension rather sensing. than the representative function of signs as information. Sensing, here, on the one side highlights the interlacing between sensory, cognitive and affective fields within sentience Sensing and its significance for intelligibility. On the other side, the term sensing stands for a Sensing is feeling and thinking. It is an ex- primarily relational capacity through which perience that constitutes surfaces of entities boundaries between sense-making entities and at the same makes time their boundaries are enacted. At the same time, media- questionable and negotiable. It is a capacity technological devices become commodified which cannot be isolated. Relations are its tools for exploiting sensing by encapsulating very substance. Take for example the very the capacity of meaningful articulations into a trivial but existential experience of sensing calculated determination of sense (cf. Hörl). hunger. When I feel hungry, I know that I Thus, media-technology enables a govern- have to eat. If the hunger just emerges, I ing of reality by instrumentalizing sensing for might be able to wait for a while. If the sense determining purposes. of hunger is more intense, I might become By taking into consideration a case dizzy, unfocused, moody, not able to hold on study of machine learning, this article re- to a clear thought. If I wait for too long and gards the materialization of information via let the hunger expand, the first bites might machine learning as a process of sensing. It cause sickness rather than delightful relief, focuses thereby on the premises of connec- even though it is a well-known feeling of swal- tivity and the program’s ability to ‘generalize’ lowing more or less pulpy food. Sometimes I Generalization describes an algorithmic know that I need to eat something, but I am processing, which relies on the abstraction too nervous to feel hungry. Fortunately, I am of information by gaining structuring models in a privileged situation, where I have access from data. It points to a transformation of the to resources and it is my decision how I deal concept of computation and the epistemol- with hunger, I can choose when, what and ogy it engenders (cf. Parisi). Further, the how to eat, I do not suffer from hunger. article discusses this computational mode Why do I elaborate on the sensation of of sensing in regards to different concepts hunger here? It seems to be a good example

21 APRJA Volume 8, Issue 1, 2019 for how feeling and thinking, knowledge and audio-visual recordings or drawings, model- perception, cognition and affect intertwine in lings, simulations etc. — may not only bring complex ways; for how the same sense may distinct phenomena into the realm of atten- undergo very different nuances and states; tion, making them detectable for perception for how this sense is generated by multiple and cognitively knowledgeable, they are also entities on different conglomerating scales; tools for intervening in the process of world- for how organism and its environment are in- ing. The procedures of sense-making are not terdependent and make sense of each other; just means to establish truths about reality or for how an ‘automatic’ metabolism is a part storages for representations of the world, but of me and can as well become a conflict for are rather partaking in processes of individu- my ‘self’; for how due to reflective or intuitive ation. They are interfering in ontogenesis by knowledge I can act upon and modify it, but affecting “spacetimematter’s” (Barad) intel- never fully control it; and for how the word ligibility and sensibility. This is a crucial, as ‘hunger’ subsumes a wide range of different, this approach stresses the power of artefacts singular intensities. Nonetheless, terming engendered by sense-making practices as this phenomenon with one word enhances well as the limits of sensing: reality can nei- a relatability and provides further means for ther be fully grasped, understood, perceived sharing or differentiating an experience. In in its distinct parts nor as a whole. Though other words, it is a way to illustrate the insep- this is not because the means of sensing arability between knowledge and aesthetics, would have to be improved, made more ad- or intelligibility and sensibility that comes into equate or sufficient, but because reality will play when I refer to ‘sensing’. have been already re-shaped by its means Sensing emphasizes a material- of sensing. Thus, reality will have become a semiotic (cf. Haraway 11) understanding different one, once sensing has transformed of inhabiting the world. Instead of being it. Moreover, from this point of view, sensing’s conditioned by a teleological meaning or a capacity remains particular no matter how transcendent subject, sensing is determined ex- or intensive it might become. Borrowing by the relations of its materialization. This from Karan Barad, sensing can be compared does not imply that meaning becomes obso- to the practice of agential cuts. lete, rather significance is an indispensable aspect of becoming. Everything that is, has [A]gential cuts are at once ontic and to be meaningful. Also, the mode of being is semantic. It is only through specific decisive for what it is, with the consequence agential intra-actions that the bounda- that “there is no single world in which all liv- ries and properties of “components” ing beings are situated […], there are series of phenomena become determinate of ‘worlds-for’” (Thrift 465)[1] that interrelate and that particular articulations with each other. To understand sense this become meaningful. In the absence way, first, accounts for a non-representative, of specific agential intra-actions, affective, pre-cognitive or “nonconscious” these ontic-semantic boundaries are (Hayles)[2] knowledge that is inherently ac- indeterminate. In short, the apparatus tive within material ontogenesis. Second, the specifies an agential cut that enacts a diverse cultural operations for making-sense resolution (within the phenomenon) of of the world are at the same time methods the semantic, as well as ontic, indeter- for worlding. Artefacts — e.g., ranging from minacy. (ibid. 148) oral narrations, reports, measurements, to

22 Irina Raskin: MACHINE LEARNING AND TECHNOECOLOGICAL ...

For Barad distinct entities are primar- relations rather than essentially stable and ily conditioned by an “ontic and semantic self-contained meaningful relata are the only indeterminacy” so that they cannot be taken decisive criteria for rationality, with the effect for granted but are temporal materializations that “signs are no longer seen primarily as within a phenomenon. She argues against representative but as operative entities” (ibid. the assumption that knowledge is produced 19). A technoecological culture of sense thus by an interaction of essentially separable stresses that prior to being representations, unities and suggests instead that distinctions signs are operators that directly act upon the are the result of “intra-actions” within mate- relations they express. rial agency (cf. ibid. 132-185). So, agential Additionally, as elaborated by Luciana cuts are the generative effect of intra-action Parisi, this conceptual shift is accompanied processes that transform an onto-epistemo- and influenced by the transformation of tech- logical indeterminacy into a temporarily de- nological apparatuses as such. According terminate separability. Here, indeterminacy to Parisi, multi-sensorial, algorithmic, is equated to an immediacy which refuses automated and networked devices for data any direct access. Any form of determination processing can no longer be understood is understood as a process of mediation, as means for communication or tools for constituting itself through in- and exclusion transmission, but have to be addressed as of possible onto-semantic materializations, machines for prehension[4] that “expose a whereby exclusions are the constitutive nonsensuous mode of feeling irreducible to matter of indeterminacy’s potential (cf. ibid. the split between the mental and the physi- 179). Though it seems that from this point of cal, the rational and the sensible“ (Parisi view, disentanglement is an absolute impos- Technoecologies of Sensation 182). Besides sibility and is transferred into the realms of bringing forward an entanglement between the unthinkable, it also draws attention to the thinking and feeling, concretization and aspect that detachment inherently partakes abstraction, technoecological conditions of in any act of sensing. sensing and sense-making disclose processes immanent to worlding that pass above and below cognitive and sensory perception, thus inherently expanding the realms of sentience Ecologizing and fundamentally modifying its configuration. “Because media no longer mediate (at Following Erich Hörl, this understanding of least not primarily) our senses; rather, they sensing can be described as an “ecologization mediate — insofar as ‘mediate’ is at all still of thinking” (Hörl 1), feeling and perceiving the right term — sentience itself, and they — a “shift from signifying to technoecological do so in the overwhelming majority of cases sense.” (ibid. 4). Due to the implementation before any occupation ‘we’ can have through of media-technological devices on one side and at the level of our sensory organs.” and the expansion of the concept of ecol- (Hansen 373)[3] This makes automatized ogy which underwent a denaturalization and commodified modes of sensing in forms on the other, Hörl states a reconfiguration of mediatechnological devices important of the “culture of sense” (ibid.) towards a instruments for an environmentally acting fundamental relational conception. Within power that “[operates no longer] through per- the realms of a technological condition, fectly integrated circuits of communication,

23 APRJA Volume 8, Issue 1, 2019 but through a new interlocking of distinct ratios, in the form forcefully wielded by milieus of information sensing.“ (Parisi, the mathematics of power.” (Hörl 8) Technoecologies of Sensation 182) Though the ecological conception of From this point of view, algorithmically, sensing came into matter in the course of an automatized sense-makers in the form of ubiquitously present mediatechnology, and computational media-technology are pro- therefore through the infiltration of computa- moting a rational epistemology, because tion into processes of sense-making, Hörl the methods of calculation are based on states that the mathematical way of relating pre-determining axioms. Driven by desires fosters an epistemology and politics that fun- induced by cybernetics and capitalism, algo- damentally opposes the epistemology and rithmic automatization of sensing processes politics of relational sensing. enable new ways of governing reality. Hörl marks a difference between a technoeco- It [mathematics] only knows of logical culture of sense and a computational extensive vectored relations between sensing carried out by commodified media- pre-given terms, terms that always technology. In the following section, I want to precede the relation, terms that are, examine this opposition by roughly exploring but do not become. The “dominance of the processes of sense-making in a case the mathematical” reterritorialized rela- study of machine learning. tions whereas the counter-knowledge of recent anthropological work in particular deterritorializes relations and drives the elaboration of a real relation Learning ecologism. (Hörl 8) Though machine learning programs are Hence, for Hörl mathematical proce- based on neuro-scientific hypotheses which dures are operations that genuinely seal are implemented into an architecture of capacities because they pre-determine rela- algorithmic networks, they have become tions. He argues that more than just models for the cognition of living beings. They have become the at- “For today, we find ourselves at a tempt to transform computation into a field very specific point in the history of of sentience, to integrate it into the capacity relationality that brings out the ques- of thinking-feeling, establishing an artificially tion and the problem of relationality built, partly automatized, yet not autonomous much more radically than ever before: mode of sensing. relational technologies and an algorith- Artificial neural networks are only one mic governmentality reduce, regulate, of the possible architectures to maintain control, even capitalize relations to computational operations that are subsumed an enormous extent, and precisely under the term machine learning.[5] They are in so doing, become essential to the basically up to several millions simultane- form of power of Environmentality. […] ously interconnected algorithmic units. Active There is, in other words, a neoliberal- in different locations, they are used to detect capitalist destruction of the relation regularities of data. So, they are tools to or- [Bezug], a reduction of relations to ganize material by extracting information from calculable, rationalizable, exploitable data without necessarily having a pre-given

24 Irina Raskin: MACHINE LEARNING AND TECHNOECOLOGICAL ... evaluation system coded to determine what detect features by recognizing patterns in a is supposed to be meaningful information. In one-to-one ratio, i.e. they are able to signal, other words, these programs are supposed if data repeats. The researchers test the to develop ‘their own’ semantics by process- relation between single directions and a net- ing data. The algorithmic units within artificial works ability to generalize. Single direction neural networks are themselves organized refers to the phenomenon of ‘selective’ cy- by principals of correlation, recursion and berneurons which caught a lot of attention in repetition[6]: in a mise en abyme like struc- a previous experiment: the activity of several ture, algorithmic units are layered one after cyberneurons is said to express a selectivity the other in such a way that the output of that matches with semantic concepts of what processed data automatically becomes the the data is supposed to show. For example, input for the next one. It is only a question there are cyberneurons which with a high of hardware resources and programmer’s probability detect cat faces, others human choice how many layers are integrated faces or human silhouettes, while they are within one network. The more layers are se- being inactive towards other kinds of depic- quenced, the ‘deeper’ the machine learns. In tions (cf. Le). Starting from the observation of the case of programs for detecting features those “easy to interpret neurons” (Morcos and from digital images originating from different Barrett), the researches investigate the role contexts, the algorithms gain their sensibil- of the other “confusing” (ibid.) ones, which ity for formal similarities and differences so far make up the majority of an algorithmic between the images by applying probability network, for generalization. For example, in calculation on the pixel’s appearances.[7] the case of the ‘confusing’ cyberneuron, it is The resulting numeric value is equated with equally active towards an image of a giraffe, a state of activity or inactivity of the affected a house and a hamburger and equally inac- cyberneuron towards the processed image. tive towards an image of a dog, a plane and [8] So, when a cyberneuron reacts actively a cat (ibid.).[10] While the in/activity of some towards images, it is regarded as an expres- cyber-neurons seem to indicate that they sion for the detection of a similar feature or are sensible towards data in a meaningful pattern between them and when it reacts way, the majority of the cyber-neurons make inactively, it indicates that the images do not connections between images in a way that is have any meaningful correlations. rather puzzling, because their responses to In 2018, Google’s company DeepMind the datasets seem to remain random — the published the paper On the Importance algorithmically conglomerated patterns of of Single Directions for Generalization information do not make any semantic sense (cf. Morcos). Generalization stands for a to the human mind. What kind of patterns or “structure-finding network” (Morcos 3), which features do those images have in common? means that the network is capable to learn Is there a ‘hidden’ sense, a pattern, a rela- a model that structures the processed data. tion which cannot be perceived by sensory Networks capable to generalize well are organs, cannot be conceptualized by cogni- more likely to predict regularities for inputs tive thinking? How to address this lack of which have not been part of the training set. comprehension or miscommunication? Hence, their numeric value signals refer- By deleting diverse constellations of ences according to a structuring model.[9] ‘selective’ as well as ‘confusing’ cyberneu- Contrarily, for memorizing networks match- rons, the researchers tested their influence ing is the measurement for learning as they on the overall performance of the network

25 APRJA Volume 8, Issue 1, 2019 to generalize (cf. Morcos and Barrett) and appears to be meaningful in a comprehend- concluded the following: First, the ‘confusing’ ible or ‘confusing’ manner. In this specific or seemingly indecisive cyberneurons are case, where the artificial network’s capacity not less important than the ‘selective’ ones. to generalize is tested, the practice of collat- Second, the cyberneurons that have an ‘easy ing and abstracting information is inscribed to interpret’ selectivity towards previously axiomatically into the interconnected struc- unknown data (i.e. images that were not part ture of probabilistic processing. of the data training set) are “more resilient” The conditioning of the computational to deletion than networks that are only ‘se- mode of sensing via machine learning, which lective’ towards already calculated data (cf. seeks to install an algorithmic capability to Morcos and Barrett). Thus, the capacity to generalize, seems to attest to the epistemo- detect features is not exclusively dependent logical shift stated by Parisi: on the seemingly high degree of selectivity to be found in isolated algorithmic units. The The training of algorithms becomes ‘confusing’ cyberneurons are not malfunc- more similar to an articulation of pro- tioning. Rather the experiments’ results hint cedures by means of which algorithms towards their significance for the capability of not only learn to think, but above all the network to abstract structures or models learn how to gain meaning from the in data. Thus, it seems that this perceived conceptual infrastructure associated randomization of data indicated by the ‘con- with the granularity of data. Learning fusing’ cyberneurons is a concomitant of the here coincides with the knowledge network’s ability to generalize. The research- of how hypotheses are generated, ers asssume “that highly class selective units whose indeterminacy in regards to may actually be harmful to network perfor- its results expands the possibilities mance” (Morcos 10). to extend the search for and from Though the study doesn’t resolve the meaningful information.” (Parisi, Das reasons for the observed causality between Lernen lernen oder die algorithmische the ‘confusingly’ acting cyberneurons and a Entdeckung von Information 103)[11] functioning generalization of machine learning applications, it does make an argument for Here, what appears to be knowledge- the acknowledgment of being connected and able is not given. The functioning of rea- making connections as a profound principle soning is not defined as a reproduction of of acting intelligently. The act of connecting symbolic information that has been already seems to be decisive on many levels: it is predetermined to be meaningful. In other crucial for the design of the experimental ar- words, significance is defined by a specula- rangements as well as the scope within algo- tive process of abstracting similarities out of rithmic processing. The numeric response is relational differences that can be found in regarded as a way of the network to connect data. Thus, this marks a transition where the with the data and it is simultaneously a way design of machine learning applications is for the researches to connect to the network, not concerned with what to learn, but “learn- namely for assessing what and how the net- ing how to learn” (Parisi, Reprogramming work has learned. Hence, being connected Decisionism 4). Parisi outlines that “[c] and making connections here coincidences ybernetic instrumentality replaces truth as with the generative quality of making sense, knowledge with the means of knowing, and disregard whether the numeric response announces a metaphysical dimension of

26 Irina Raskin: MACHINE LEARNING AND TECHNOECOLOGICAL ... machine knowledge originating from within for building taxonomies, for organizing and its automated functions of learning and pre- structuring by modeling and thus providing diction.” (ibid.) The fact that that DeepMind’s new information. researchers favor networks which learn by Yet, at the same time the case study also generalization rather than memorization con- shows that there are conflicting measures firms this epistemological shift. Here, learn- coming into play regarding the evaluation of ing as a practice is not about reproducing the cyberneurons’ in/activity. The computed already known information, but about gener- numeric values are unavoidably exposed to ating connections as the basis for producing the comparison with already established tax- new information. In the face of computational onomies which function as means to evalu- devices that proliferated into archives “sav- ate the ‘rightfulness’ of the calculated meaning us from the past, from repetition through ing. This is why some of them are regarded repetition” (Chun, Programmed Visions 157), to be ‘confusing’ and others to be ‘easy to educational skills such as learning by heart interpret’. Hence, though the computational become less relevant. So, while computers processing might exceed a representational seem to have relieved humans from the bur- culture of sense, it is implemented by the den to train their mind to become a storage researchers’ ascriptions, i.e. their under- for stable knowledge, the computational de- standing of what a meaningful detection is, sign works on transforming the computer into serves as an evaluation of the networks’ ca- a machine that is simultaneously an archive pacity to generalize and this understanding and its registrar. is especially representational. This applies to The case study described above the ‘easy to interpret’ cyberneurons’ in/activ- shows that integrating the means of learning ity which matches with semantic concepts how to learn into computational processing of the images’ depictions, as well as to the is accompanied by the introduction of a ‘confusing’ ones, whose in/activity is eventu- technoecological culture of sense into the ally assessed by their ability to sustain the realms of computation. The cyberneuron’s networks’ function. Moreover, images which in/activity towards the data is determined in make up the set of data are put here into a specific way: Though the regularities to the role of representational signifiers. The be detected are not pre-inscribed into the whole procedure of machine learning here program, the very conception of the architec- includes various remediations which are not ture of the algorithmic network is based on addressed by the researchers. Though the the assumption that to learn means to make perceived randomization of meaning, which connections. Each cyberneuron’s in/activ- is indicated by the ‘confusing’ cyberneu- ity becomes a new speculative information rons, is a point of interest for researching, that arises out of the process of connecting it does not become the entrance point for what appears to be contingent. Instead of questioning the processes of remediation, having a pre-given schema that determines assumptions about the detectability of pat- the cyberneurons’ in/activity, it is this very in/ terns or the premise of connectivity. Instead, activity that becomes the schema for sensing it is problematized as well as resolved by the in a connective infrastructure. Irrespectively concept of generalization, which at least in of whether the in/activity appears to be this case becomes a method for maintaining ‘easy to interpret’ or ‘confusing’, it becomes a representational order of things—a way of a mean for operating with and through the ensuring a general equivalence between sig- data—a mean for un/detecting regularities, nifiers and signified, a general equivalence

27 APRJA Volume 8, Issue 1, 2019 of meaning between different artifacts and emergence of software as a concrete entity diverse modes of sense-making. and commodity” (Chun, On “Sourcery,” or Code as Fetish 303). Drawing upon Judith Butler’s understanding of performativity, Chun argues against conceptualizing code Expressing as a merely machinic expression:

In order to account for a computational What is crucial here is: first, code that sentience in machine learning applications, succeeds must be citations — and the processing of numeric values has to be extremely exact citations at that. conceived as a performative act. Regardless There is no room for syntax errors; of whether the in/activity of the cyberneurons second, that this iterability precedes is evaluated to be meaningful in a repre- the so-called subject (or machine) sentational way or whether the in/activity is that is supposedly the source of the understood from a technoecological point of code; and third, and most importantly, view accounting for an excess of meaning, an entire structure must be in place in it points towards a sensing of data within order for a command to be executed. worlds-for data. Evidently, it is fundamental to This structure is as institutional and consider the in/activity as a mode of sensing political as it is machinic. (ibid. 322) in order to be able to relate to and modulate its operability. Though, it is important to notice Applying Chun’s argument to the in/ that the machinic mode of sense-making, the activity of the cyberneurons, computed specific computational sensing within the numeric values cannot just be addressed as network, derives not only from the program expressions of the artificial neural network. as such or a genuine ‘execute-ability’ of One has to take into account the social and the algorithms. Rather, an entire apparatus political infrastructure, where these materi- sets structures and conditions potentials for alizations are embedded in and which render how and what is to be expressed through its capacity of becoming an expression at all. numeric values — this concerns for exam- Recurring to Barad’s notion of agential cuts, ple the required material infrastructure, the to address the in/activity as a mode of compu- programming of code, the labeling of data, tational sensing implies the following: on the the digitalization of images, the generation of one side, it stresses that the discursive and power, the transmission of electronic signals the material dimensions within what emerges etc.[12][13] So, a lot of work has to be done as a machinic expression are inseparably by human and non-human labor to ensure entangled. Thus, every materialization — in a computational sentience and even more this case the cyberneurons’ in/activity — is work to navigate the machinic expressions already inherently political and social. On of sense into desired directions. To conceive the other side, these materializations are the in/activity of cyberneurons as an enun- regarded less as the result of assembled hu- ciation of algorithms alone would result in a man and non-human workforces, but more misconception comparable to the one Wendy as events that temporarily (re-)produce the Hui Kyong Chun has worked in her study of boundaries between human and machinic the performativity of code: a “conflation of labor. Moreover, they have the capacity to instruction with its product — the reduction reformulate the relations that constitute the of process to command — that grounds the agents at work. Though Barad’s concept of

28 Irina Raskin: MACHINE LEARNING AND TECHNOECOLOGICAL ... performativity differs to the one of Chun, it too the conditions of production that sustain their raises attention towards the infrastructure or effectiveness. Nonetheless, they are imbed- apparatuses of expressions: ded in those cultures of sense-making, which shape the arrangements of the program’s ap- In an agential realist account, perform- paratuses and influence how one relates to ativity is understood not as iterative the sensing. From this point of view, numeric citationality (Butler) but as iterative values cannot just be regarded as signifiers intra-activity. Intra-actions are agentive, referring to patterns, features or semantic and changes in the apparatuses of concepts. Rather they are signs for the effec- bodily production matter for ontological tiveness of a complex assembly that weaves as well as epistemological and ethical a computational mode of sense-making into reasons: different material-discursive the realms of sentience by simultaneously practices produce different material implementing a technoecological culture of configurations of the world, different sense into the machinic infrastructure. difference/diffraction patterns; they do not merely produce different descriptions. Objectivity and agency are bound up with issues of responsibility Conclusion and accountability. Accountability must be thought in terms of what matters The expanding evolvement of media-tech- and what is excluded from mattering. nological devices does not only transform (Barad 184) concepts of computation but also brings forward a further dimension of the interlac- From this perspective, even a repre- ing between sensory, cognitive and affective sentational culture of sense cannot just be fields within sentience. The implementation regarded as means for merely depicting the of media-technology introduces new environ- world, but rather has to be addressed as a ments for sensing and re-configures modes specific way of intervening into reality —a of sensibility and intelligibility. From this point specific mode of worlding. In regards to the of view, the reductionist and quantitative above described case study of machine characteristics of applied mathematics do not learning, a representational logic is applied per se encapsulate capacities for sensing. in particular as a measurement to modulate Though computational methods of sensing the expressions in forms of numeric values differ from other ones such as writing, touch- into desired articulations in order to channel ing, hearing, smelling, thinking, feeling etc., the programs capacity into an instrument for they neither genuinely oppose them in terms the (re-)production of restrained meanings. of their tendency towards determination of Whereas this specific way of re-configuring meaning, nor can they be executed exclu- the process of materialization relies on the sively within the realms of computation. As account of conceiving the cyberneuron’s in/ determination is inherently part of any mode activity as a responding expression, which of sensing and a condition for un/becom- allows the machine and its apparatuses to ing, it is so to speak an immanent cruelty of become a relatable milieu for sensing. The worlding. So, a critique against an instrumen- artificial neural networks are situated in the talization of sensing for exploitative means realms of probabilistic procedures and they cannot just address logics of calculation or are insensible for cultural connotations or capturing, but also has to take into account

29 APRJA Volume 8, Issue 1, 2019 the apparatuses and their measurements which create and sustain computational pro- Notes cedures for instrumental means. Thus, this does not imply that the problematic of de- [1] Thrift is referring here to Jakob von termination as such becomes meaningless. Uexküll’s concept of “umwelten”. On the contrary, as it highlights that every made connection simultaneously points to a [2] Hayles coins the term “nonconscious” detachment, the determining process asks in order to describe a mode of thinking that to be further problematized. Therefore, it traverses cognition, but is not executed directs one’s attention to the measurements consciously. It is an automatically enacted which are incorporated in sense-making and decision for interpretation of information, demands a continuous questioning of what which is pervasive in life forms as well as kind of world the determinations make im/ technical systems (cf. Hayles). possible. It raises the awareness for the ethical dimension within sensing — because [3] Translated from the German publication each connection goes in hand with separa- by the author. tion, learning with unlearning, expression with muteness. [4] The term “prehension” has been suggested by by Alfred North Whitehead. It describes a registering or comprehending mode of existence that is intrinsic to all organic and inorganic forms of perception and thinking (cf. Whitehead 57ff). For the relevance of Whitehead’s philosophy for Parisi’s thinking see also Was heißt Medienästhetik? (44-49).

[5] In his publication on the Machine Learners: Archaeology of a Data Practice Adrian Mackenzie gives an overview of the different operations that are assembled under the term ‘machine learning’. He examines the consequences machine learning has on forms of knowledge production, critical thought and strategies of power. Notably, by machine learners he “refers both to humans and machines or human- machine relations” (Mackenzie 6) and therefore rather to practice that is situated in specific “accumulations of settings, data and devices” (ibid.).

30 Irina Raskin: MACHINE LEARNING AND TECHNOECOLOGICAL ...

[6] It is interesting to note that this resem- [10] This example refers to the explanatory bles what Félix Guattari, who himself was graphic provided by the researchers (cf. influence by cybernetic theory, described as Morcos and Barret). ‘machinic’: an affective mode of thinking that proliferates non-pre-given, irreversible and [11] Translated from the German publication singular enunciations, which result into an by the author. excess of meaning by assemblies that are organized through recursion and connectiv- [12] For instance, Andreas Sudmann ity (e.g. cf. Guattari). At this point it should emphasizes that the reason for the recent be also mentioned that Hörl develops the popularization and proliferation of machine notion of the technoecological culture of learning applications is neither primarily to sense amongst others in close reference be found in more elaborated algorithms nor to Guattari’s idea of ecology as well as his the enlargement of data training sets, but notion of non-significant heterogenesis the parallel organization of fast GPU- or of meaning which is apprehended to be TPU-chips (cf. Sudmann 63, 69). machinocentric (cf. Hörl 13-21). [13] See e.g. the work published in the [7] This depiction of how machine learn- context of Data & Society (https://datasoci- ing processing of images is arranged by ety.net) that provides insightful research on engineers and how it works on a computa- social consequences as well as conditions tional level is quite simplified here. It is to for mediatechnological industries. be said that there are different parameters for designing such a program and that there are further aspects such as regulatory measures (e.g. batch normalization) that Works cited shape its operability. Nonetheless, for the context of this paper, I want to emphasize Aplaydin, Ethem. Machine Learning. MIT the premises of connectivity as well as the Press 2016. attributes of the network’s architecture such as correlation, recursion and repetition in Barad, Karen. Meeting the Universe regards to their partaking in the configura- Halfway: Quantum Physics and the tion of an algorithmic sensing via machine Entanglement of Matter and Meaning. Duke learning. University Press 2007.

[8] If it tends towards zero it is regarded Chun, Wendy Hui Kyong. “On ‘Sourcery’, as inactive and if tends towards one it is or Code as Fetish.” Configurations. John regarded as active. Hopkins University Press, vol. 16, no. 3, Fall 2008, pp. 299-324, DOI: 10.1353/ [9] According to Alpaydin, the network’s con.0.0064. capability to generalize is the main feature that marks machine learning’s capacity to —. Programmed Visions: Software and govern information. He states: “This ability Memory. MIT Press, 2011. of generalization is the basic power of machine learning; it allows going beyond the training instances.” (Alpaydin 42)

31 APRJA Volume 8, Issue 1, 2019

Guattari, Félix. “Über Maschinen.” Ästhetik Parisi, Luciana. “Das Lernen lernen und Maschinismus. Texte zu und von Félix oder die algorithmische Entdeckung von Guattari. Merve, 1995, pp.115-132. Information.” Machine Learning: Medien, Infrastrukturen und Technologien der Hansen, Mark B. N. “Medien des 21. Künstlichen Intelligenz. Transcript, 2018, pp. Jahrhunderts, technischer Empfinden 93-113. und unsere originäre Umweltbedingung.” Die technologische Bedingung. Beiträge —. “Technoecologies of Sensation.” zur Beschreibung der technischen Welt. Deleuze, Guattari & Ecology. Palgrave Suhrkamp, 2011, pp. 365-407. Macmillan, 2009, pp. 182-199.

Haraway, Donna. Modest_Witness@ —. “Reprogramming Decisionism.” Second_Millenium. FemaleMan©_ e-flux journal #85, Oct. 2017, Meets_OncoMouse™_Feminism and e-flux.com/journal/85/155472/ Technoscience. Routledge, 1997. reprogramming-decisionism/.

Hayles, N. Katherine. Unthought: The Power — and Erich Hörl. “Was heißt of the Cognitive Nonconscious. University of Medienästhetik? Ein Gespräch über die Chicago Press, 2017. algorithmische Ästhetik, automatisches Denken und die postkybernetische Logik Hörl, Erich. “Introduction to General der Komputation.” Zeitschrift für Ecology: The Ecologization of Thinking.” Medienwissenschaft. Medienästhetik. General Ecology: The New Ecological Diaphanes, I/2013, pp. 35-51. Paradigm. Bloomsbury Academic, 2017, pp. 1-65. Sudmann, Andreas. “Szenarien des Le, Quoc V. et. al. Building High-level Postdigitalen. Deep Learning als Features Using Large Scale Unsupersived MedienRevolution.” Machine Learning: Learning, 12. Jul. 2012, arxiv.org/ Medien, Infrastrukturen und Technologien pdf/1112.6209.pdf. der Künstlichen Intelligenz. Transcript, 2018, pp. 55-73. Mackenzie, Adrian. Machine Learners: Archaeology of a Data Practice. MIT Press, Thrift, Nigel: “From Born to Made: 2017. Technology, Biology and Space.” Transactions of the Institute of British Morcos, Ari S. et. al. On the Importance of Geographers. Vol. 30, No. 4, Dec. 2005, Single Directions for Generalization. 22 May pp. 463-476. Republished in: Non- 2018, arxiv.org/pdf/1803.06959.pdf. Representational Theory. Space, Politics, Affect. Routledge, 2008, pp. 153-170. Morcos, Ari and David Barret. Understanding Deep Learning through Whitehead, Alfred North. Prozess und neuron deletion. 21 May 2018, deepmind. Realität: Entwurf einer Kosmologie, com/blog/understanding-deep-learning- Suhrkamp, 1987. through-neuron-deletion/.

32 33 Maike Klein

ROBOTIC AFFECTIVE ABILITIES

Abstract

Within both popular media and (some) scientific contexts, affective and ‘emotional’ machines are assumed to already exist. The aim of this paper is to draw attention to some of the key conceptual and theoretical issues raised by the ostensible affectivity. My investigation starts with three robotic encounters: a robot arm, the first (according to media) ‘emotional’ robot, Pepper, and Mako, a robotic cat. To make sense of affectivity in these encounters, I discuss emotion theoretical implications for affectivity in human-machine-interaction. Which theories have been implemented in the creation of the encountered robots? Being aware that in any given robot, there is no strict implementation of one single emotion theory, I will focus on two commonly used emotion theories: Russell and Mehrabian’s Three-Factor Theory of Emotion (the computational models derived from that theory are known as PAD models) and Ekman’s Basic Emotion Theory. An alternative way to approach affectivity in artificial systems is the Relational Approach of Damiano et al. which emphasizes human-robot-interaction in social robotics. In considering this alternative I also raise questions about the possibility of affectivity in robot-robot-relations.

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Maike Klein: ROBOTIC AFFECTIVE ABILITIES

Making sense of encounters with ‘emotional’ robots

At GV Lab in Tokyo, I met a robot arm that was equipped with a PAD emotion program combining the values of pleasure, arousal, and dominance to constitute an emotion that is expressed by a movement. Through interaction via different sensors, the robot executed different behaviors. If, for instance, I entered the robot’s ‘personal space’, as detected by a distance sensor, the robot executed an ‘emotional’ movement, as you can see in the image below:

Figure 1: Robot arm. Source: https://www.youtube.com/ watch?v=q1DO4PBSA6M&feature=youtu.be.

I encountered the humanoid robot ‘Pepper’ in a shopping mall: it was standing in the corner, overwhelmed by ordinary noise. In the media, Pepper has been advertised Figure 2: Pepper, emotional robot. Image by author. as “the world’s first emotional robot” (Singh). This doesn’t seem very convincing. Most of Finally, I encountered Mako, the robot the time, if there are no technical problems, cat that I built at GV Lab in order to learn first- passers-by did not pay attention to it. It hand what machinic affectivity, and building a did not seem to me that Pepper has much machine in general is all about. When does emotionality that humans typically react to. the machine start to interact, to be a robot, Perhaps Pepper would have been more and to be affective? Mako is an Arduino- interesting if it would have raised its voice or based small device equipped with distance just gone somewhere else to avoid being ig- and touch sensors for interaction. Moreover, nored. Contrary to the robot arm, as you can it can express itself: by text through an LED see in the pictures, Pepper is a humanoid display, by movement through a servo mo- robot that has a face, changes its voice, and tor, and by noise through a piezo buzzer. so on. Its outer appearance is intended to be It has neither emotion programing running, cute and to evoke positive emotions. nor a capacity for changing facial expression.

35 APRJA Volume 8, Issue 1, 2019

Thus, it is a very basic robot. Nevertheless, it elicited confusion in humans that were not What was going on in these due to any malfunction. First the display says encounters? Various kinds the neutral but welcoming message “what a of affectivity! nice day”, then if the human approaches, the messages “touch me” and “do not approach” are shown; and if the human touches, the Besides everyone’s own feeling of and for robot expresses a loud piezo beep and the affectivity, there exists a variety of definitions message “go away”. for affective phenomena in living beings that one can choose from, making the phenomena hard to grasp. To follow the works of Colombetti, Deonna and Teroni, I use the notion of ‘affectivity’ to subsume phenomena like emotions, feelings, moods, primordial affectivity, sentiments, or affect. The various definitions of affective phenomena in living beings come with different theoretical frameworks. In philosophy and psychology, these are notably emotion theories, each emphasizing different aspects of emotions. For instance, take the distinction between non-cognitivist and cognitivist emotion theories: When William James writes about emotion, he means the occurring bodily changes and their felt experience. For him, emotions without a bodily component are “cold” mental states (James 189). In contrast to this, Martha Nussbaum claims that emotions are cognitive “judgments of value” and the possibly occurring bodily changes are just their byproduct through physically imitating the cognitive processes (Nussbaum 194). These theories exemplify antagonisms and exclusions: Nussbaum’s theory is far from an embodied perspective and excludes animals and children up to a certain age. ‘James’ theory’ (or the ‘James-Lange theory’ that was later developed) would possibly Figure 3: Mako, robotic cat. Image by author. be difficult to apply on systems that do not have bodies of flesh and blood. This example emphasizes that there is no uncontroversial definition of emotion or affective phenomena in living beings. In other disciplines, affective phenomena in artificial systems have been studied extensively (e. g. Suchman, Picard, Dautenhahn et al., Marsella et al., Boden).

36 Maike Klein: ROBOTIC AFFECTIVE ABILITIES

When we look at robots and engi- In this paper, I reflect upon the emo- neered applications, we often find emotions tion theoretical implications to affectiv- – machines modeled over emotional expres- ity in human-machine-interaction, having an sions, emotions evoked in humans through academic background in practice-oriented human-machine-interaction, and even emo- philosophy and a practice in creatively tional robots. Due to the lack of consensus exploring technology. I will briefly introduce regarding the definition of emotion, from an two of the commonly used emotion theories emotion-theoretical perspective, the theo- that here shape the emotion theoretical retical basis of machinic emotional abilities discourse from the Western tradition: Russell has to be based on a working definition. and Mehrabian’s Three-Factor Theory of For several years now, in disciplines like Emotion (the computational models derived Affective Computing and Social Robotics, from that theory are known as PAD models) computer scientists and roboticists have and Ekman’s Basic Emotion Theory. My goal applied (mostly psychological) emotion in this text is not to provide an exhaustive theories (e. g. Ekman and Friesen, Russell overview or detailed analyses of either emo- and Mehrabian), and taken emotion theories tion theories/models or artificial systems that as a foundation of their programming and include affective abilities. Rather, my goal engineering (e. g. Bennett and Šabanović, is to raise questions and initiate discussion Rincon-Ardila et al.). With no clear definition about the application of emotion theory to of what an emotion is, however, it is difficult to robots and the complexities of assessing the choose which theoretical framework to take ostensible affectivity of robots. and how to translate the (more or less) wordy During the three encounters I described theories into numbers. Besides, although above, I was confronted with three differ- there are many different theories of emotion ent ways of modeling affective abilities into and affectivity, most emotional programs and machines: internal, external, and relational machines depend on just a few theories that (Damiano et al. 8). As shown in the chart be- are limited in describing emotions in general. low, each way of modeling comprises several Thus, what is this thing called emotion that in features I experienced during my encounters the end comes out of the machines? with the robots.

37 APRJA Volume 8, Issue 1, 2019

Depending on the numerical value of each Emotion theories as bases of the three dimensions, they form or explain for emotion models a different emotion. The statistical methods used and the resulting significance of the theory is a topic open for further discus- If the outer appearance of a machine sion. This theory, however, has been further includes emotional features, for instance developed in various ways since 1977, with displayed emotions, or emotion recognition Russell’s ‘Core Affect Theory’ as its most technology, the work of Ekman and his col- popular contemporary spin-off. Today, PAD leagues is typically used as the theoretical is still popular when equipping robots and basis (e. g. Bennett and Šabanović). In 1976, virtual characters with emotions. For this, Ekman and Friesen provided a system to the PAD values are mapped as vectors in a divide facial movements into “action units”, three-dimensional space. Thus, an emotion which can be a movement of one or more program can be coded and added to the ro- muscles (and one muscle can be part of bot’s other programs in the operating system. more than one action unit). Ekman, Friesen Depending on external or internal stimuli, for and their colleagues isolated each muscle instance through incoming data from an ul- movement in their own faces and observed trasonic distance sensor, the robot’s emotion video recordings and photographs in order to changes internally and produces, depending make sure that every facial expression con- on its physical features, an emotional behav- sisting of one or more action units is unique. ior as outcome (e.g. Rincon Ardila et al.). They called the resulting catalogue of facial A we have seen, in robotics-influenced expressions ‘Facial Action Coding System’ emotion research, the external/social/ex- (FACS). Moreover, Ekman continued the pressive (see the BET example) and internal/ ideas of Darwin with his work on basic emo- individual/regulative aspects of emotion (see tions: over many years and several studies, the PAD example) are distinct. Damiano et he identified six basic emotions (happiness, al. discuss this critically. External emotional surprise, fear, anger, disgust, contempt) that features are often referred to as simulations are expressed by unique facial expressions. of emotions, whereas an internal emo- According to Ekman’s studies, humans from tion generating mechanism (as in emotion various cultural backgrounds can equally models) would lead to genuine emotions. identify “at least five” of these six emotions Therefore, these two distinct ways of creat- (Ekman 551). ing affective abilities have also been rated as If a machine includes non-visible emo- true/ eal vs. false/fake emotions (Damiano tional features like regulative mechanisms et al. 8). According to these verbal distinc- that consider the environmental input through tions, it seems preferable to have regulative sensors and human-machine-interaction emotion mechanisms rather than visible or internal changes, such as changes in emotional expressions — but why? Because temperature or the former emotional state, this could provide a real-world-correlate to the Three-Factor Theory of Emotion is often our imagination? used as the theoretical basis (e. g. Rincon As Damiano and Dumouchel point Ardila et al.). This theory is sometimes bet- out, this way of thinking is deeply Cartesian ter known as ‘Pleasure-Arousal-Dominance’ and exclusive (Damiano and Dumouchel 6). (or PAD). For Russell and Mehrabian, The phenomenon of emotion is split up into emotions can be captured and described in a Cartesian construct that is related to one terms of pleasure, arousal, and dominance.

38 Maike Klein: ROBOTIC AFFECTIVE ABILITIES matter, either ‘body’ or ‘soul’, which reflects emotions as events that are individually, modeling on the ‘outside’ or ‘inside’. Besides, internally, and thus covertly generated, the binary distinctions are in some cases not and that then we can expressively com- comprehensible (e.g. the true/false rating: municate to others — i.e., the very how can an attempt to model affective abili- conception of emotions which legiti- ties into artificial systems be ‘true’ or ‘false’?) mates robotics to distinguish between and seem to be misleading. How could we the internal and the external aspects of tell which human emotion is real or fake, if emotions and empathy (Damiano et al. we go beyond evolutionary or basic emotions 8). that are necessary for survival? What if we look at ‘higher’ or social emotions? In these They call this approach a “relational concep- cases, we could possibly measure whether tion of emotions”. the person smiling at us is smiling with a In this theory, affective phenomena Duchenne smile – or possibly, we cannot can emerge from a relation that includes detect anything in the emotions of others living beings or social robots, everyone and and have to trust on what the person reports everything that is no (mere) tool and capable verbally about their emotion. Emotions can of interaction. This could happen in an inter- surely be ‘artificial’ also in humans (Stephan action with Mako, the robot arm, or Pepper, 310) in the sense of true/ eal vs. false/fake. depending on the properties of the relation Moreover, the distinction between between the robot and the interacting agent. external and internal affective features of Thus, if we imagine a human-robot interac- a robot goes against understanding it as tion involving one human being and one an integral agent. Of course, the possible robot, we have to think about three aspects behavior range of robots is much less exten- of affectivity: What is going on affectively in sive than the human behavioral range, but at the human during the interaction with the ro- least in the intended interaction, the robot is bot? What is going on affectively in the robot an integral agent within its individual limits. during the interaction with the human? What With an approach that focuses on is emerging affectively from the relation and interaction and relation between interacting what does this do to the respective interac- systems, Damiano et al. suggest one way to tion partner? This holds also for interactions make the binary distinctions between external of (two or more) living beings and interactions and internal emotional features obsolete. At of living beings and non-living entities. What the same time, they do not exclude mechani- about a robot-robot relation? cal systems of a certain complexity from the possibility of having emotions. According to Damiano et al., interacting agents do not simply exchange Affectivity in a robot-robot relation information about their supposedly pre-defined and individual emotional A good example for both thinking machines states, [they rather] mutually define— as more than mere tools and machinic af- co-determine — their emotions during fectivity among machines is the installation their ongoing interactions. […] [This Nintendogs of the artist Fabian Kühfuß. His view] requires us to abandon the tra- installation captures perfectly the fascination ditional philosophical understanding of (coming from science fiction) and absurdity

39 APRJA Volume 8, Issue 1, 2019

(coming from scientific reality) of the ques- affectively loaded meaning to this scenario. tion of whether a robot can have genuine What can this example tell us from a emotions that go beyond sharp definitions relational point of view? There clearly is an or are more than the intended outcome of a interaction between two machines. Plus, relation. The artist combines a Nintendo DS there is a human observer that does not take console that runs the game Nintendogs with part in the interaction. Is there, however, a motorized device that moves the console something affective going on between the so that a pencil can touch the virtual dogs. Nintendo DS and the touch-pen device? Affective in the sense typically applied to living beings? Affective in the sense of other possibly affective entities? Moreover, who or what is feeling something in this relation? Is this even important, as we cannot always see or detect what other human beings feel or if they are able to feel an emotion at all? In any case, in the Nintendogs example, tasks and goals have been fulfilled successfully. The touch-pen device fulfilled its task to touch the display where it could perceive the puppy (thus, it stroked). The puppy displayed the behavior the game designers and programmers intended for the case after having Figure 4: Fabian Kühfuß, Nintendogs (2017). Courtesy of the artist. been stroked. The human observer enjoyed and interpreted the artwork. The purpose of this game is to raise and educate a baby pet dog. One of the possible actions is to stroke the dog, usually an affective action between two living beings. In Affectivity ≠ affectivity his installation, Kühfuß transfers this action into a robot-robot-relation. This work raises There is obviously an affective difference at least two questions: ‘Can machines have between artificial systems and humans. The leisure activities, too?’ and ‘What is machinic difference becomes, for instance, evident in affectivity?’. the spontaneity and goal-orientation of the For the virtual dog, it does not make interacting agents and their emotion range any difference who or what strokes it. For that still clearly separates machinic affective the motorized device, it makes no difference abilities from those of living beings. There what it touches with the attached touch-pen. is something interesting about confusing In both cases, the result is the same: the dog behavior, reacting not as expected, not cre- is stroked, the touch-pen touches. For play- ating the ‘perfect’, goal-oriented, faultlessly ing the whole game, however, the machine designed user experience. It creates some would need more features that enable it to other kind of relation where humans need execute all the other necessary steps. Does to engage in another way because they are the stroked virtual dog ‘feel’ good within its somehow challenged. Moreover, there is affective spectrum? Intuition is, however, that something interesting (and relieving) about the human interpretation adds the specific not being useful, not having to be useful, not

40 Maike Klein: ROBOTIC AFFECTIVE ABILITIES needing to serve a certain purpose, not being psychologists found the emotion theories instrumentalized, like machines always are, of their computer scientist collaborators too because they are built to serve as tools, even old-fashioned (Broekens 8). We should keep if social robots have an ambiguous status in mind here that if the aim is to model af- (Damiano and Dumouchel 2, 3). fective abilities in artificial systems, there are Moreover, in these examples, my own limited possibilities of translating the wordy reactions to the robot’s affective abilities theories into a relatively simple and at the were, besides some aspects of a recipro- same time more complex model and finally cal relation, more like meta-reactions to the into numbers. If we, however, accept that machinic affective abilities. For instance, if there can be adequate emotion definitions Mako tells me “Go away!” after having told that may not fully hold for a human being me to approach, I think this is rather funny. (as well as that emotion definitions made Besides, it is not boring (at least the first time for humans may not hold for other kinds of you try the robot out), because it reacts not systems or even children – as we have seen as expected and does not fulfill any higher in the brief distinction between cognitivist purpose (that we, maybe, expect in a ma- and non-cognitivist emotion theories), we chine). If Mako was a human being, I would can claim for logical reasons that a ‘genuine’ be irritated, confused, or even concerned. So emotion comes out of an artificial system if was the woman who tried Mako out during a an emotion theory is translated and modelled lab visit – she was irritated and confused by into that system and if there is an outcome Mako’s behavior. that results from the emotional program. With However, if we aim to facilitate un- this view, we would at the same time avoid derstanding and cooperation between ‘speciesism’. As already indicated, machinic roboticists, computer scientists, psycholo- emotions may be very different from emo- gists, and the humanities, we should be tions of other systems – but not only from open to adding definitions and theories from those. As there are many different artificial technological fields to the many (imperfect) systems and different emotion theories that emotion theories we already debate in phi- are used to model their emotions, many losophy, psychology, and other disciplines. If different behaviors and mechanisms can be we want to understand the work roboticists understood as emotions. and computer scientists are doing, and if we aim to collaborate in reflecting and developing mechanical affective abilities, we should accept the emotion definitions, It’s all about imagination theories, and models from other disciplines, like computer science and (social) robotics, No matter which theory is used to model af- as specific emotion theories that are possibly fective abilities in artificial systems, in many able to explain emotions with their specific cases there will be human beings interacting limits (all emotion theories have these limits, with these systems. For instance, robots with they are simply different for each and every emotional abilities are used for therapeutic theory). This means, we should include them settings with autistic children. Among oth- as equal candidates for emotion theories ers, Cabibihan et al. provide evidence that that potentially explain emotions within liv- autistic children prefer to interact emotionally ing beings, too. This will help us to avoid with robots and that this can help to facili- problems such as those that occurred when tate the interaction with other humans, too.

41 APRJA Volume 8, Issue 1, 2019

One of the main goals of equipping artificial Thus, as human beings have the ten- systems with affective abilities is to facilitate dency to anthropomorphize, they will likely human-machine-interaction. This can be compare the outcoming emotional reactions very useful in industrial settings where the to human emotional reactions. Furthermore, worker is obliged to work with a robot that depending on e.g. the personality or informa- is very boring or that the worker does not tion and / or education about artificial systems, understand very intuitively. In such cases, human beings may have a completely differ- the amelioration of working conditions is ent understanding of affectivity in general, possible. Another possibility of human-robot- and of what artificial systems are capable interaction is found within a capitalist context. of. Apart from the scientific discourse, the For instance, especially in Japan but some- main sources of information about this topic times also in Europe, and as already briefly are, besides one’s own affectivity, media and described in one encounter above, the robot science fiction stories that sometimes tend to Pepper can be spotted in sales or customer converge with each other. Herbrechter sug- service environments, for instance in shop- gests a new media genre resulting from the ping malls, airports, and karaoke bars. convergence of fiction and facts: “Science The crucial point in all of these machinic Faction” (Herbrechter 101). As a result, one varieties and human-machine-encounters is urgent question is how to separate unrealistic imagination. According to a study of Heider ideas of machinic affective abilities from what and Simmel, even the simplest shapes are is actually happening in science to finally already anthropo- or at least bio-morphized break with the perceived mysteriousness of (Heider and Simmel 246). Humans ascribe artificial systems due largely to human imagi- intentions to the simplest moving forms nation (Sharkey and Sharkey 12, 18). even though they know that they do not have them. Moreover, humans attribute affectivity to simple shapes (de Rooij et al. 2). With a more complex design, the possible Acknowledgements ways of bio- or anthropo-morphizing a thing increase in scope. The human expectations Funded by the Japan Society for the of this technology rise and the moment of Promotion of Research (JSPS) to support deception becomes longer and more dense. participation in the JSPS Summer Program The ‘uncanny valley’ graph shows various 2018, made possible by a research stay at intensities of anthropomorphism (Mori 99). GV Lab at Tokyo University of Agriculture It is highly controversial for at least two rea- and Technology. sons: 1) There is much empirical evidence for and against it that cannot be true at the same time (e. g. Misselhorn; Bartneck and Ishiguro); and 2) It implies a strong normative dimension that holds the ‘healthy body’ as the ultimate ideal. Nevertheless, the uncanny valley is used (almost?) always as a reference in (social) robotics research and the modelling of artificial agents, avatars, or movie characters.

42 Maike Klein: ROBOTIC AFFECTIVE ABILITIES

Damiano, Luisa and Dumouchel, Paul. Works cited “Anthropomorphism in Human–Robot Co-evolution.” Frontiers in Psychology, vol. Ahmed, Sarah. “Happy Objects.” The Affect 9, 2018, DOI:10.3389/fpsyg.2018.00468, Theory Reader, edited by Gregg, Melissa and Seigworth, Gregory D., Duke University Darwin, Charles. The Expression of the Press, 2010, pp. 29-51. Emotions in Man and Animals. Dover Publications, Inc., 2018. Bartneck, Christoph and Ishiguro, Hiroshi et al. “My Robotic Doppelgänger.” The 18th Dautenhahn, Kerstin, et al. “KASPAR - A IEEE International Symposium on Robot Minimally Expressive Humanoid Robot and Human Interactive Communication for Human–Robot Interaction Research.” Toyama, Japan, 2009, pp. 269-276. Applied Bionics and Biomechanics, vol. 6, no. 3, 2009, pp. 369-397, Bennett, Casey C. and Šabanović, DOI:10.1155/2009/708594. Selma. “Deriving Minimal Features for Human-Like Facial Expressions in Deonna, Julien A. and Teroni, Fabrice. The Robotic Faces.”.International Journal of Emotions: A Philosophical Introduction. Social Robotics, 6 (3), 2014, pp. 367-381. Routledge, 2012. DOI:10.1007/s12369-014-0237-z. de Rooij, Alwin, et al. “Abstract Expressions Boden, Margaret. AI: Its Nature and Future. of Affect.” International Journal of Oxford University Press, 2016. Synthetic Emotions, vol. 4, no. 1, pp. 1-31, DOI:10.4018/jse.2013010101 Broekens, Joost. “Modeling the Experience of Emotion.” International Journal of Ekman, Paul and Friesen, Wallace Synthetic Emotions, vol. 1, no. 1, 2010, pp. V. “Measuring Facial Movement.” 1-17. DOI:10.4018/jse.2010101601. Environmental Psychology and Nonverbal Behavior, vol. 56, no. 1, 1976, pp. 56-75. Cabibihan, John-John, et al. “Why Robots? A Survey on the Roles and Benefits of Ekman, Paul. “Are There Basic Emotions?” Social Robots in the Therapy of Children Psychological Review, vol. 99, no. 3, 1992, with Autism.” International Journal of Social pp. 550-553. Robotics, vol. 5, no. 4, 2013, pp. 593-618. DOI:10.1007/s12369-013-0202-2. Ekman, Paul. “Basic Emotions.” Handbook of Cognition and Emotion, edited by Colombetti, Giovanna. The feeling body. Dalgleish, Tim, and Power, Michael J., Affective science meets the enactive mind. Wiley, 1999, pp. 45-60. MIT Press, 2014. Fong, Terrence, Nourbakhsh, Illah, Damiano, Luisa, et al. “Towards Human- and Dautenhahn, Kerstin. “A Survey of Robot Affective Co-evolution Overcoming Socially Interactive Robots.” Robotics Oppositions in Constructing Emotions and and Autonomous Systems, vol. 42, no. Empathy.” International Journal of Social 3, 2003, pp. 143-166, DOI:10.1016/ Robotics, vol. 7, no. 1, 2015, pp. 7-18. S0921-8890(02)00372-X. DOI:10.1007/s12369-014-0258-7. 43 APRJA Volume 8, Issue 1, 2019

Heider, Fritz, and Marianne Simmel. “An Rincon Ardila, Liz et al. “Adaptive Fuzzy Experimental Study of Apparent Behavior.” and Predictive Controllers for Expressive The American Journal of Psychology, Robot Arm Movement During Human and vol. 57, no. 2, 1944, pp. 243-259, Environment Interaction.” International DOI:10.2307/1416950. Journal of Mechanical Engineering and Robotics Research, vol. 8, no. 3, forthcom- Herbrechter, Stefan. Posthumanismus. ing 2019. Eine kritische Einführung. Darmstadt: WBG, 2009. Russell, James A. and Mehrabian, Albert. “Evidence for a Three-Factor Theory James, William. “What is an Emotion?” of Emotions.” Journal of Research in Mind, vol. 9, no. 34, 1884, pp. 188-205. Personality, vol. 11, no. 3, 1977 pp. 273- 294. DOI:10.1016/0092-6566(77)90037-X. Kühfuß, Fabian. Nintendogs. Video installation. Nintendo DS, microcontroller, stepper, Russell, James A. “Core Affect and aluminum frame, software, 2017. the Psychological Construction of Emotion.” Psychological Review, Marsella, Stacy et al., “Computational vol. 110, no. 1, 2003, pp. 145-172., Models of Emotion.” A Blueprint for Affective DOI:10.1037/0033-295x.110.1.145. Computing-A sourcebook and manual, edited by Scherer, Klaus R. et al., 2010, pp. Sharkey, Noel and Sharkey, Amanda. 21-46. “Artificial Intelligence and Natural Magic.” Artificial Intelligence Review, vol. 25, Minsky, Marvin. The Emotion Machine: no. 1-2, 2007, pp. 9-19. DOI:10.1007/ Commonsense Thinking, Artificial s10462-007-9048-z. Intelligence, and the Future of the Human Mind. Simon & Schuster, 2006. Singh, Angad. “’Emotional’ Robot Sells Out in a Minute.” CNN Business, 23 Jun. 2015, Misselhorn, Catrin. “Empathy with Inanimate https://edition.cnn.com/2015/06/22/tech/ Objects and the Uncanny Valley.” Minds pepper-robot-sold-out/index.html. and Machines, vol. 19, 2009, pp. 345-59. DOI:10.1007/s11023-009-9158-2. Stephan, Achim. “Zur Natur künstlicher Gefühle.” Natur und Theorie der Emotion, Mori, Masahiro. “The Uncanny Valley.” edited by Stephan, Achim, and Walter, IEEE Robotics & Automation Magazine, Henrik, mentis, 2003, pp. 309-324. vol. 19, 2012, pp. 98-100. DOI:10.1109/ MRA.2012.2192811. Suchman, Lucy. Human-Machine Reconfigurations. Cambridge University Nussbaum, Marta C. “Emotions as Press, 2007. Judgments of Value and Importance”. Thinking about feeling: Contemporary phi- losophers on emotions, edited by Solomon, Robert C., Oxford University Press, 2004, pp. 183–199.

44 45

(UN)BEING Brett Zehner

MACHINES OF SUBJECTION: NOTES ON A TACTICAL APPROACH TO ARTIFICIAL INTELLIGENCE

Abstract

As the big data revolution ramps up, we are drawn to online platforms that modulate political identity far removed from so-called liberal politics (Cheney- Lippold 2011, 165). There are two ends to the extreme. We have seen the return of white supremacists on supposedly democratic networks while on the ‘back end’ of computational culture, algorithms de-subjectify users for proprietary gain. In the broad sense here, subjectivity is an individual’s relation to themselves. However, machine learning occupies a powerful position within the logics of capital by shifting the site of identification into a digital sphere (165). With the widespread use of machine learning practices, abduction creates an overall “sensibility to change and alter events” (Mackenzie 2013, 402). By abstracting concrete social practices into dta vectors, machine learners measure, forecast and modulate human behaviors. Put simply, machine learners have become some of the most potent social inscription devices today. It is within this context that my dissertation asks — how does the recent ubiquity of machine learning affect how we wield political subjectivity?

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Brett Zehner: MACHINES OF SUBJECTION

As humans feed affect, thought, and extreme, we have seen the rise of white sociality into algorithms, algorithms supremacists propagating through networks feed back into what used to be called that segregate public opinions. Yet, on the subjectivity. This shift is what has given ‘back end’ of computational culture, machine way to a post-representational politics learning algorithms de-subjectify human us- adrift within information space. ers for proprietary gain. Capitalism doesn’t — Hito Steyerl care if you’re a fascist, a passivist, or even a bot; so long as it can extract behavioral information from your actions to be packaged and resold by its advertisers. As Cheney-Lippold 1. Machines of subjection points out, machine learning shifts the site of identification into the “measurable, digital For the past two decades, fields of knowledge- sphere” (165). Between the front-end user production that utilize statistics have adopted interface, and the back-end logics of com- machine learning as their primary mode putation — machine learners are embedded of operation (Mackenzie, “Programming within the powerful contradictions of capital- Subjects” 434). Due to the advances of ist logics. computational technology, machines can Amidst this seeming contradiction, the now be programmed to find patterns in large concept of subjectivity may be an unhelpful datasets. ‘Machine learners’[1] recursively category. ‘Enlightenment Man,’ the Cartesian use patterns to infer correlations, essentially subject divided between mind and body, the hailing new performative judgments on the rationalist ‘view from nowhere’ — these euro- world. Adrian Mackenzie goes so far as to centric notions of subjectivity are founded on claim that we now live within a regime of the measuring functions of coloniality and the predictivity characterized by computational technological organization of capital brought practices that rely less on verification than to bear on the individual.[2] Still, technologi- inference and abductive reasoning. With the cal imaginaries have also been mobilized to widespread use of machine learning practic- trouble hegemonic notions of subjectivity. es, abduction creates an overall “sensibility Donna Haraway’s feminist subjects, for in- to change and alter events” (402). By ab- stance, dethrone the “god-tricks” of scientific stracting concrete social practices into data rationalism through situated technopolitical vectors, machine learners measure, forecast practices (Haraway, 1988). The notion of and thus modulate human behaviors by es- technological subjection, or perhaps more sentially scripting performatives. Put simply, accurately, the notion of de-subjectivization, machine learners have become some of the occupies a set of complex problems that most potent social inscription devices today. garner closer attention. It is within this context that I ask – how does Now emerging scholarship at the in- the recent ubiquity of machine learning affect tersection of identity and machine learning the production of subjectivity? has opened new pathways of research in As the big data revolution ramps up, digital cultural studies. Healy and Fourcade much attention has been drawn to online observe that the state used to be the only platforms that modulate political identities apparatus with the technological power “situated at a distance from traditional liberal to track its subjects. However, this is no politics and removed from civil discourse” longer the case (Fourcade and Healy). The (Cheney-Lippold 165). On two ends of the recent ability for machine learners to track

49 APRJA Volume 8, Issue 1, 2019 online users’ digital footprints, or their “data automation of decision theory that set the exhaust,” marks an important moment for stage for the first machine learning program what Shoshona Zuboff calls surveillance to be utilized for economic and military plan- capitalism. Every action a user performs on ning by the RAND Corporation in 1955. Tung- a digital system is considered a signal to be Hui Hu advances this historical analysis by analyzed, packaged, and subsequently fed mapping the topography of power relations back into the system. The quantity of user within ‘cloud computing’ where decisions are data is much more important than quality. As distributed across networked assemblages. long as an action online can be converted He argues that borders seem to be out of into data, it can be utilized in predictive date conceptions at the foundation of the behavioral models. Zuboff explains that no sovereign subject (14). The Tiqqun collec- online action is too trivial to be aggregated, tive in their cybernetic hypothesis posit that repackaged, and sold again (79). “Facebook traditional class divisions and social conflict likes, Google searches, emails, texts, pho- no longer cut through the middle of society, tos, songs, geo-location, communication but through the middle of each of us. What patterns” are all considered lucrative data to is troubling is that the production of subjec- marketing firms and myriad other companies tivity seems to be no longer about creating (79). Though let’s be clear. Surveillance “people of substance” but of turning each capital is not merely a social media concern. person into a “fleshless envelope, the best The algorithmic bias of machine learners possible conductor of social communication” stems from a long line of quantitative racism (18). And most recently, in the Trump era, and surveillance (Browne). The targeting of Luciana Parisi links the de-personalization the poorest members of society continues, of machine learning systems to the rise only now it operates through various forms of of post-truth politics. Here indeterminacy data surveillance and predatory credit scor- and the unknown “push automated cogni- ing (Fourcade and Healy 31). Zuboff argues tion beyond knowledge-based systems” that technique supplants authority, and that (“Reprogramming Decisionism” 10). What we “discipline and control produce a certain ultimately find within the political subjection knowledge of human behavior independent (and de-subjectivation) of machine learners of consent” (81). In this extractive logic, we is a brutal instrumentalism based more on see an impersonal form of subjection at the mechanical functions than on ideological heart of surveillance capital. New forms of content. power emerge alienating persons “from their Now, despite considering machine own behavior while producing new markets learning’s effect on social identity,[3] the of behavioral prediction and modification” above scholarship on machine learning has (75). left open an opportunity for rigorous scholarly The liberal idea of the rational decision- attention to de-subjectivation. For instance, maker then seems to unravel as a locus of John Cheney-Lippold asks: “What does the power relations. In The Control Revolution, banality of competing for a job interview historian of technology James Beniger de- using machine learning to predict future scribes the automation of decision theory in friendships say about subject formation” the 1930s. “Any decision tree of finite length (8)? This line of questioning still focuses can be duplicated by a finite automaton, on subjection at the level of performatives thereby equating the question of decidability and self-awareness. This limited viewpoint with that of computability” (64 ). It was the imagines the subject merely as a ‘user’ who

50 Brett Zehner: MACHINES OF SUBJECTION is always already ideologically ‘hailed.’ Even practices engaging technology as always though data analysis seems to aggregate our being wrapped up in power relations. The most intimate habits, surveillance remains activist ethos of tactical media has been mo- automated and deeply impersonal as it bilized in various registers by Rita Raley, The bypasses individuated modes of subjectivity Critical Art Ensemble, and Beatriz da Costa. and signifying semiotics. Both digital media In each case, performance is considered a studies, if focusing merely on identification mediating process that enacts technological through computational performatives, is apparatuses. Each usage of the term tactical limited in offering any new insights into the media is dependent upon the specific set of forces at play in our present moment. technopolitical relations that the practition- I argue that the acceleration of predic- ers hope to intervene within. In the case of tive techniques and impersonal forms of Beatriz da Costa and Kavita Philip, their tacti- control require a more robust consideration cal biopolitics replaces the term ‘media’ out of de-subjection. Along these lines, tactical of a consideration of specific technoscientific machine learning would have two goals. forms of knowledge production modulating First, the goal would be to update theories the possibilities of life (da Costa and Philip). of subjection, and de-subjection, for the pro- Jussi Parikka, in a similar manner, mobilizes liferation of machine learning devices with a a geological imaginary to intervene within keen attention to practices that bypass clas- extractive environmental politics and digital sical definitions of the subject. And second, culture. In this vein of mapping a specific set to provide an analysis of social practices ex- of technopolitical relations, the title of this ternalized into the technologies of machine essay mobilizes tactical media in the service learning. We must describe and experiment of exploring the temporal regimes of machine with certain tactical media concepts that learners. The title also borrows directly from undergird machine learning today – scenario Saidiya Hartman’s Scenes of Subjection. planning, training, and prediction. I hope to expand on scenographic modes of subjection by drawing on the cybernetic imaginary to elucidate forms of technological de-subjection at the heart of identity politics. 2. A tactical media My initial hypothesis is that through so-called approach to machine new media regimes, old forms of subjection learners mutate through the new technopolitical conditions that arise. I look to unearth the technologies of subjection as they traffic through Methodologically, a tactical media approach the digital sphere. to machine learning must be situated within To provide insight into the processes of the perspectives of media philosophy as well de-subjection I rely on two threads of critical as the practice of media arts. The legacy of theory. The first mode of critical theory that I tactical media (and its forebears in 1960’s utilize comes from post-autonomist marxism intermedia, conceptual, and performance and its theories of signification and subjec- art practices) informs much of my project tion within the late capitalist technosphere. combining the fields of digital media with I find it useful to consider the performative performance studies. For instance, tactical statements and decision architectures of ma- media was outlined in the late 1990s by chine learners via the philosophy of language David Garcia and Geert Lovink as a set of found in the work of Michel Foucault, and

51 APRJA Volume 8, Issue 1, 2019

Félix Guattari. Especially useful is Foucault’s Although the post-autonomist critique concept of the dispositif — the structural yet of technology is quite useful in understand- mutable union between institutions, subjec- ing both processes of subjection and de- tivity, and discourse. He defines the dispositif subjection within the logic of computational as an autonomous technique which exists capital, there is still the problem of the spe- “on the other side of juridical and political cific historical and material contexts in which structures of representation” (Foucault 40). machine learners are situated. One must The dispositif is a mechanism of capture, wonder if Guattari’s exploration of de-subjec- both material and discursive, which directly tion can find a more radical usage today. In manages the experience of everyday life. this manner, and concerning de-subjection, Guattari further identities two dispositifs the second strain of critical theory we must of power that operate in a contradictory man- engage with is queer-of-color-critique. ner. On the one hand, we face systems of We must turn to the negative iden- social subjection. Social subjection catego- tity politics that refuse to validate, affirm, or rizes us with assigned identities — it gives strengthen forms of subjectivity presently us a gender, a race, a profession — a posi- produced under capitalism. Recent antago- tion of symbolic representation. However, nistic positions and pessimisms are powerful the production of an individuated subject is not because they have to do with identity “but also coupled with a different process that because they have to do with the “mundane proceeds though desubjectivization. Guattari radicalism of the desire to de-subjectivize defines this process asmachinic enslavement all categories” (Menon). Queer-of-color which dismantles the individuated subject, critique has long grappled with processes consciousness, and representations, acting of de-subjection as a crucial step in forming on both pre-personal and supra-individual minoritarian collectivity. Disidentification op- levels. In machinic enslavement, the indi- erates “in and against dominant ideologies” vidual is no longer instituted as an “economic while refusing assimilation. subject” or a “citizen.” She is instead consid- Hortense Spillers’ theorization of the ered “a gear, a cog, a component in financial flesh and the body is important to consider and various other institutional assemblages” in relation to present data practices which (ctd. in Lazzarato, Signs and Machines 25). quantify human behaviors. Spillers positions For instance – advertising focus groups the distinction between body and flesh as the stopped using questionnaires long ago in central difference “between captive and liber- favor of measuring biometric response to ated subject positions”. For Spillers, the body stimuli such as taste tests or eye tracking. is possessed by an individual who is the sole Capitalism is so successful because it oper- owner of their selfhood. Yet for a captive, ates heterogeneously at the intersection of as in the case of chattel slavery, the body social subjection and machinic enslavement. is reduced to flesh. This flesh is exposed We are all caught in a double bind between to violence without protection from legality, performative individuation and the dissolu- equality, or democracy. In fact, Jasbir Puar tion into our dividual parts, unknown to our- writes that the violence of capital is legiti- selves. Guattari’s critique (which I extend to mated through the right to maim. Outside of the analysis of machine learners) is of critical representation, vision, or ideology, the flesh theories that deal only with language and/or records the primary narrative of the horrors recognition while ignoring de-subjectivizing of liberal humanism. It is the suffering of the processes and their non-representational flesh which exceeds white coloniality semiotics. 52 Brett Zehner: MACHINES OF SUBJECTION and acts as transgenerational memory, I consider these positions, structure vs. highlighting the ways black bodies remain empiricism, not as antagonistic but as a cou- as flesh. Spiller’s political antagonism is an pling of strategy and tactics. Jared Sexton’s attention to the memory of the flesh existing structures of social death measure the terms on the side of the de-subjectified, the already of total struggle while Moten’s empirical fu- outside, beyond the limits of the subject or gitivity finds its escape lines on the shifting the law. grounds of everyday survival. Put simply, Of course, the right to maim emerges Afropessimism both refuses social subjec- in the data practices of predictive policing tion, while escaping machinic enslavement. that quantifies black behaviors which are Guattari and Afropessimism both critique the correlated to racist databases. All of this subject as always already commodified and digital magic is merely a weapon to legitimate inscribed (albeit in different ways) by violence police horrors in the streets. Yet, as Spillers and exploitation. Afropessimism seems more claims of the memory of the flesh. How can relevant in its call to desubjectify not to reap- we reclaim the memory of our data exhaust? propriate or celebrate already existing posi- Again, de-subjection and disidentifications within racial capital. Perhaps it is here tion hold tactical power. Whereas optimism that data exhaust and its nefarioius uses can and the will to produce new subjectivity may be a site of real struggle. Data exhaust is look to the redeemable of the world, pes- used to expose and reconstitute subjects to simism takes up its position as an attempt new techniques of power. Facebook strikes, to channel the forces of the outside. We can sabotage of data sets used to predictively see, for instance, a glint of optimism in the police entire populations, and adversarial work of Fred Moten — in the gap between network attacks all become viable options in the flesh and the law. This is where Moten political struggle. finds ungovernable fugitivity. For Moten, con- Although the tactical potentials of ma- tinuous movement exceeds the subjection of chine learning are emerging as we speak, social death. Blackness for afropessimists one brief example we can gesture towards however, cannot escape social death. The is Zach Blas’ project Facial Weaponization distinction here between optimism and pes- Suite. Intervening directly in biometric sur- simism is a difference in ways to refuse the veillance systems, it “protests against biom- measure of racial capital. Moten proposes etric facial recognition by making ‘collective movement and evasion. Afropessimism how- masks’ in community-based workshops that ever sees no such possibility. are modeled from the aggregated facial data The methodological divide here is nu- of participants, resulting in amorphous masks anced, yet crucial. Afropessimism attempts that cannot be detected as human faces by to disarticulate the real object from the object biometric facial recognition technologies.”[4] of knowledge. The analytical task Sexton The masks are subsequently used for public suggests is to move from a measure of em- interventions and performances. What is so pirical experience to the structure of political provocative about Blas’ project is the gesture ontology. Afropessimism claims that the em- of obfuscation — a digital act of refusal of pirical existence of racialized exclusions are individual subjection achieved not through “in danger of entering the discursive record escaping systems of capture but by turning as transcendental truths”. The task is clearly surveillance systems toward a disaggregated to resist any empiricism which may play back collective subject. Facial Weaponization Suite into the measure of white supremacy. peers back into machine learning systems in

53 APRJA Volume 8, Issue 1, 2019

virtually unrecognizable to their engineers. What emerges is what Luciana Parisi calls the “alien rule” of algorithmic ubiquity:

Far from making the rational system of governance more efficient, this new level of determination forces governance to rely on indeterminate probabilities, and thus to become confronted with data that produce alien rules. These rules are at once discrete and infinite, united and fractalized. (Contagious Architecture 11)

Of course, the tautological empiricism of machine learners is problematic as they can be used to reinscribe prejudiced data to justify social segregation (Mackenzie, “The Production of Prediction” 441). However, for those political struggles not interested in recognition but that are more invested in functional power and the right to opac-

Figure 1: Zach Blas, Facial Weaponization Suite (2011- ity; perhaps there is an opportunity offered 14). by machine learners to turn their alienating weapons against systems of exploitation their own likeness, alien aggregates, algorith- (Coulthard; Glissant). mic approximations of the crowd, statistical average identities that ironically protect the individual identities most at risk from exploitation. Blas’ machine learners operationalize data through the excessive overload of computational measure, illustrating the potentials of direct digital struggle. The politics of machine learning are not yet entirely clear. What is clear is that machine learning needs endless supplies of data. Any data will do. And increasingly that data can be unstructured. What is perhaps most interesting here is that the processes in which machine learners operate are becoming less understandable to the designers engineering their functions (Fourcade and Healy 11). In instances where there is no initial hypothesis, no pre-existent model, machine learners experiment in ways that are

54 Brett Zehner: MACHINES OF SUBJECTION

Notes Browne, Simone. Dark Matters: On the Surveillance of Blackness. Durham: Duke [1] The generalized practice of machine University Press, 2015. learning encompasses many techniques of predictive modeling that are used to classify Brunton, Finn, and Helen Nissenbaum. events and things into stable categories. Obfuscation: A User’s Guide for Privacy and Some of these techniques include linear Protest. Cambridge: MIT Press, 2015. regression models, Bayesian classifiers, and k-nearest neighbors. Decision trees, Cheney-Lippold, John. 2011. “A New deep belief networks, and neural networks Algorithmic Identity: Soft Biopolitics and the however are the most interesting in terms Modulation of Control.” Theory, Culture & of subjection. The research on machine Society 28.6 (2011): 164-181. learning is evolving, seemingly on a week to week basis. Coulthard, Glen. Red Skin, White Masks. Minneapolis: University of Minnesota Press, [2] It is useful to remember here what Gilles 2014. Deleuze, and Lewis Mumford before him, were keen to observe: that technologies are Da Costa, Beatriz, and Kavita Philip, eds. social before they are mechanical. Tactical Biopolitics: Art, Activism, and Technoscience. Cambridge: MIT Press, [3] For our purposes here social identity and 2010. subjectivity are one and the same. Social identity is considered a given coordinate in a Deleuze, Gilles, and Félix Guattari. A state-based system of categorization: race, Thousand Plateaus: Capitalism and gender, name, social status. Schizophrenia. Minneapolis: University of Minnesota Press, 1987. [4] See http://www.zachblas.info/works/ facial-weaponization-suite/. Deleuze, Gilles. ”Postscript on the Societies of Control.” October, vol. 59, pp. 3-7.

Critical Art Ensemble. “Tactical Media at Works cited Dusk?” Third Text, vol. 22 no. 5, 2008, pp. 535-548. Aarons, K. “No Selves To Abolish: Afropessimism, Anti-Politics and the End of Ernst, Wolfgang. 2016. Chronopoetics: the World.” Metamute, 29 February 2016, The Temporal Being and Operativity http://www.metamute.org/editorial/articles/ of Technological Media. Lanham, MD: no-selves-to-abolish-afropessimism-anti- Rowman and Littlefield International. politics-and-end-world Michel Foucault. 1999. Abnormal: Lectures Beniger, James. The Control Revolution: at the Collège de France 1974-1975. Trans. Technological and Economic Origins of Graham Burchell. New York: Picador. the Information Society. Cambridge, MA: Harvard University Press, 1986.

55 APRJA Volume 8, Issue 1, 2019

Fourcade, Marion, and Kieran Healy. Lazzarato, Maurizio. 2014. Signs and “Seeing Like a Market.” Socio-Economic Machines: Capitalism and the Production of Review, vol. 15, no. 1, 2016, pp. 9-29. Subjectivity. Los Angeles: Semiotext(e).

Garcia, David, and Geert Lovink. “The ABC Mackenzie, Adrian. “The Production of of Tactical Media.” Nettime, 16 May, 1997, Prediction: What Does Machine Learning https://www.nettime.org/Lists-Archives/ Want?” European Journal of Cultural nettime-l-9705/msg00096.html. Posting to Studies, vol. 18. No. 4-5, 2015, pp. 429-445. mailing list. Mackenzie, Adrian. “Programming Subjects Ghamari-Tabrizi, Sharon. The Worlds of in the Regime of Anticipation: Software Herman Kahn: The Intuitive Science of Studies and Subjectivity.” Subjectivity, vol. 6, Thermonuclear War. Cambridge: Harvard no. 4, 2013, pp. 391-405. University Press, 2009. Menon, Madhavi. Indifference to Difference: Glissant, Édouard. Poetics of Relation. Ann On Queer Universalism. Minneapolis: Arbor: University of Michigan Press, 1997. University of Minnesota Press, 2015.

Guattari, Félix. Soft Subversions. Los Mumford, Lewis. The Myth of the Machine: Angeles: Semiotext(e), 1996. Techniques and Human Development. New York: Harcourt, Brace & World, 1967. Haraway, Donna. “Situated Knowledges: The Science Question in Feminism and the Muñoz, José Esteban. Disidentifications: Privilege of Partial Perspective.” Feminist Queers of Color and the Performance Studies, vol 14, no. 3, 1988, pp. 575-599. of Politics. Minneapolis: University of Minnesota Press, 1999. Hartman, Saidiya. Scenes of Subjection: Terror, Slavery, and Self-Making in Parikka, Jussi. A Geology of Media. Nineteenth-Century America. Oxford: Oxford Minneapolis: University of Minnesota Press, University Press, 1997. 2015.

Hu, Tung-Hui. A Prehistory of the Cloud. Parisi, Luciana. “Reprogramming Cambridge: MIT Press, 2015. Decisionism.” e-flux journal, no. 85, October 2017, https:// Lazzarato, Maurizio. “Exiting Language: www.e-flux.com/journal/85/155472/ Semiotic Systems and the Production of reprogramming-decisionism/ Subjectivity in Félix Guattari.” Cognitive Architecture: From Biopolitics to Noopolitics. Parisi, Luciana. Contagious Architecture: Architecture and Mind in the Age of Computation, Aesthetics, and Space. MIT Communication, edited by Hauptmann D. Press, 2013. and Neidich W. Rotterdam, Netherlands: 010 Publishers, 2010, pp. 502-521. Pasquinelli, Matteo, ed. Alleys of Your Mind: Augmented Intelligence and its Traumas. London: Meson Press, 2015

56 Brett Zehner: MACHINES OF SUBJECTION

Raley, Rita. Tactical Media. Minneapolis: University of Minnesota Press, 2009.

Spillers, Hortense J. “Mama’s Baby, Papa’s Maybe: An American Grammar Book.” Diacritics, vol. 1, no. 2, 1987, pp. 65-81.

Hito Steyerl. “Proxy Politics: Signal and Noise,” e-flux journal, no. 60, December 2014, http:// www.e-flux.com/journal/60/61045/ proxy-politics-signal-and-noise

Tiqqun. “The Cybernetic Hypothesis.” The Anarchist Library, https://theanarchistlibrary. org/library/tiqqun-the-cybernetic-hypothesis. Retrieved on May 29, 2010 from cybernet. jottit.com

Zuboff, Shoshana. “Big Other: Surveillance Capitalism and the Prospects of an Information Civilization.” Journal of Information Technology, vol. 30, no. 1, 2015, pp. 75-89.

57 Maria Dada

QUEERING GLOBAL INFORMATION SYSTEMS

Abstract

What’s the relationship between GIS and the political subject? In an effort to address this question, this paper traces the movement from the map to GIS. The map is shown to be the performative utterance of the state, one that supports its national discourse and narrative. GIS, on the other hand, is shown to be a device of neoliberal governmentality, its non-representational economic practices, divided discourse and subjectivities. Despite the seemingly hopeless situation surrounding GIS, however, certain simulation and modelling practices are attempting to construct subjectivities out of economic neoliberalism’s fractured narratives. They do this by reading meaning into otherwise mathematical datasets and models. These practices could form a basis for queering GIS.

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Maria Dada: QUEERING GLOBAL INFORMATION SYSTEMS

but rather a disruption and displacement of Introduction the map by the model. In fact, most applications that later become the digital map didn’t The field of cartography as the academic have a map to begin with. They were created study of the history and meaning of mapmak- in order to forecast population information ing is in decline, or “is dead” (“Cartography for the user by city officials, planners and Is Dead” 4), according to Denis Wood, one businesses. The so-called maps, such as of the foregrounding figures of critical cartog- the OXAV and SYMAP were complex and raphy. It is being ingested into what might, had their own symbols with an accompanied on the surface, seem like a continuation of user manual that explains how they were to the discipline but in fact is not. The practice be interpreted. None had a drawing of the taking over from cartography is that of digital terrain or land. modelling. The following paper extends this dis- Nonetheless, many critical cartogra- course by showing the relationship between phers including Wood, find no reason to subjectivity and GIS, a relationship that is bemoan this loss. They might even celebrate missing from various accounts of critical GIS it. “Thank God” (“Cartography Is Dead”), which centre on critiquing statistical model- proclaims Wood (4). Contrary to what one ling for its alleged positivism as does Stan might imagine, the map is a relatively recent Openshaw, for instance, in his 1991 article, “A practice dating back only to the 1500s. Its view on the GIS crisis in geography”. Rather lineage coincides with the emergence of the than interpret digital modelling as strictly a disciplinary sovereignty and the state’s right quantitative method, as Openshaw regarded to political violence, what political theorist it, this paper reveals its performativity, one Archille Mbembe would call necro-political that is remarkably different from that of the power. map. The two are similar in that they not The map in that sense is a performative only describe but rather construct territories, utterance of state territory. Without a map, understood as extension of sovereign power, the state would not be conceived of as a however with modelling, these territories thing, a map-able object with borders and are no longer bound to the land, they are no edges, “state borders are brought into being longer strictly spatial but rather penetrate the through mapping” (The Power of Maps 45). psychology, behaviour and even molecules The map becomes the icon or as historian of those subjected to its power. and political scientist Benedict Anderson Digital media theorists such as Alexander claims the logo of the state and this icon with Galloway and Bernard Stiegler, following its definite borders erases the lineage of its Gilles Deleuze in the article “Postscript to the construction. Societies of Control,” have interpreted this As critical geographers such as Paul A. territorial permeation by digital media more Longley and Matthew W. Wilson affirm, it is generally as a cause for the disarmament not that the map was transformed into a digi- of the political subject, the subject’s endless tal map but rather that digital modelling as a division into manipulable units of data or practice cannibalised the remnants of a dying code. Deleuze himself sees the individual tradition for its own gains (Longley; Wilson). subject transformed into a dividual, endless Accordingly, the trajectory of the move from units of data subtracted from individuals and mapping to Global Information Systems their bodies, he explains, “Individuals have (hereafter GIS) is not a linear progression become ‘dividuals,’ and masses, samples,

59 APRJA Volume 8, Issue 1, 2019 data, markets or ‘banks’”(5). In The Exploit also includes Foucault’s work on biopolitics Galloway and Eugene Thacker also suggest and political economy. Foucault centres that the political subject of digital media is much of his later publications and lectures on divided into atomic units that make up part the genealogy of power, or what he calls the of a larger network. Paul B. Preciado, on the knowledge-power nexus. In Discipline and other hand, sees this type of control modulat- Punish he shows the movement of power ing subjectivities at a molecular level through from what many, including political theorist pharmaceuticals. GIS could, therefore, be Achilles Mbembe, have referred to as necro- seen to follow this logic and further dismantle political power; the power of the sovereign the political subject into units of data through king to take away life as a form of punish- its use of statistics and cognitive psychology. ment for transgressing his law, to regimes of Contrary to this reasoning, however, this pa- discipline or disciplinary society, where con- per shows that GIS or digital modelling in fact trol is no longer based in the threat of death. offers a way to potentially unearth a radical Instead, disciplinary society is engaged in political subject. the self-disciplinary techniques of discourse, With that in mind, this paper is divided the institutions and surveillance mechanisms into five sections. Section one begins with the similar to the architecture of the panopticon problem set out by Deleuze in “Postscript to as described by Jeremy Bentham. the Societies of Control.” It relates his article However, according to Deleuze, to Michel Foucault’s work on neoliberalism in the institutions that Foucault describes in The Birth of Biopolitics, setting the scene for Discipline and Punish, are in crisis or in section two where the link between Deleuze perpetual need of reform. They have been and Foucault’s ideas around subjectivity are replaced by “a new monster” (Deleuze 444). brought closer to the economic practices of Disciplinary control is no longer positioned digital modelling through the work of Philip at the institutional level but has rather been Mirowski. internalised by each subject who as a result The remaining three sections delineate is no longer a subject but a ‘dividual’, a term the movement from mapping to GIS within Deleuze shares with Félix Guattari. In other the fields of cartography and geography more words, it is no longer the architecture of the specifically. The key difference between GIS, school, the barracks, the prison that keep us which is largely based in non-representa- from misbehaving. Control society works at tional economic practices, and maps, which modulating subjectivities at a more granular is based in narrative, is outlined in section level through mechanisms that theorists, five. Finally, the conclusion touches on the including Galloway, have interpreted to re- ways in which GIS could potentially reinstate semble those of digital media. a form of political subjectivity and retain a Actually, Foucault, in his lectures on The critical dimension. Birth of Biopolitics, delivered around thirteen years prior to the publication of “Postscript to the Societies of Control,” was beginning to touch on some of these ideas. His account, Deleuze after Foucault described in the remainder of this section, offers an alternative interpretation to Deleuze’s Deleuze’s article “Postscript to the Societies notion of the dividual which informs its rela- of Control” references and extends Foucault’s tionship to digital modelling and GIS. work, primarily Discipline and Punish, but He dedicates the lectures to highlighting

60 Maria Dada: QUEERING GLOBAL INFORMATION SYSTEMS the multiplicity of shifts within power or the art beyond the control and interpretation of any of government discussed above, the move- sovereign, man or subject. ment from punitive sovereignty, or necro- Now as it happens the most effective political techniques of power, where the state form of rationality, which is used in order assumes the role of what Foucault describes to calculate and make sense of the self- as “a cold monster” to what he describes as limitation of governmental reason, is political a more ‘reasonable’ form of power. economy or the supposed natural laws of Under necro-political art of government the self-regulation of the market. Foucault the king was able to punish and kill while be- continues, ing answerable to no-one but the divine laws of God. Breaking the divine laws of God would the intellectual instrument, the form of force the sovereign to step down. However, calculation and rationality that made as the mode of power shifts so too do the possible the self-limitation of govern- laws that govern it. If the necro-political king mental reason as a de facto, general is only accountable to his subjects in relation self-regulation […] is political economy to the divine laws, the sovereign of govern- (13). mental reason is not accountable at all but rather limited by nature. In fact, as he himself admits, all of Under what Foucault refers to as the Foucault’s final lectures on biopolitics need ‘reasonable’ raison d’état the sovereign has to be understood through the lens of political to negotiate their power with that of nature economy and its tools such as economics. It and its laws. The paradox, of course, is that is the intellectual apparatus born out of the one cannot reason with nature. It is in a raison d’état to enrich the state against its sense the condition and the limit of rationality. enemies. Therefore, the laws of nature are supposedly Political economy, nonetheless, imposed on the state. The latter is, of course, determines the success or the failure of the fallacy that Foucault is exposing in the government but does not illegitimate it. lecture. Governments can simply be mistaken by In other words, the laws of nature, ignoring the new laws of nature, the laws which are imposed on the state of the raison of the market. A bad governor is not wicked d’état, operate differently to the laws of God. but ignorant. Ignorance does not dissolve a Foucault explains, government. The relationship between truth and self-limitation, however, is not about wis- To say that there is a de-facto limitation dom of rule such as that of the Machiavellian of governmental practice means that a prince. In place of the wisdom of the prince, government that ignores this limitation governments rely on economic experts will not be an illegitimate, usurping “whose task is to tell the government what in government, but simply a clumsy, truth the natural mechanisms are of what it is inadequate government that does not manipulating” (17). do the proper thing. (10) It is the judgement of governmentality on success in opposition to legitimacy that Put differently, breaking with the internal pacifies the political subject and turns them limitations of governmental reason will not into dividuals in Deleuze’s terms. The issue, render it illegitimate because these limita- the reason governmentality is no longer tions are no longer juridical. Natural laws are judged for its legitimacy, and what troubles

61 APRJA Volume 8, Issue 1, 2019

Foucault, is that sovereignty and the law statistics, or modelling contained in any are no longer set in relation to citizens as economic work” (Mackenzie 7). subjects. The market, and its supposed natural laws, are now the medium between In fact, there wasn’t a particular sovereign and subject. dominant form of economics. Mirowski in- The only means for the sovereign to sists, “there was no dominant orthodoxy in govern its subjects is through the market and American economics prior to World War II, its economic experts. However, economics, although the indigenous strain of American the discipline of political economy that aims Institutionalism held some key strategic out- to understand the market remains agnostic posts at Columbia and Wisconsin” (Mirowski, to narrative, meaning and representation. It, Machine Dreams 190). Institutionalism was therefore, reduces the subject into a market of the view that institutions played a major actor at best or a multitude of divided econom- role in shaping the markets and encouraged ics units in Deleuze’s view, the dividual. The the broader understanding of their role in subject, through an economic understanding such a process. of the market, is nothing but a multitude of After the second world war in the 1950s cogs in the system never united under a the rise of the American economic model of single rebellion against the king for instance. laissez faire and the increasing availability of data fortified the link between mathematical modelling and the market. The relationship between the two fields was also influenced by The emergence of the burgeoning field of Operations Research digital modelling (hereafter OR) and the impact of the cold war’s reinforcement of technical innovation. Mirowski’s claim is that mathematical The next section will look at the emergence and later digital models developed during the of digital modelling within economics as second world war fuelled the highly special- neoliberal governmentality’s means of mak- ised new discipline of OR. OR is regarded ing sense of, and therefore governing, the as the predecessor to most computing dis- market by translating each of its elements ciplines. It is influenced by early inventors of into computable units of data for use in math- the computer such as Charles Babbage. It ematical modelling or statistical mechanics. was mostly invested, however, in the analysis According to Philip Mirowski, prior to and management of market-based decision- the Second World War the rational choice, making, including but not limited to rational mathematical model-based economics that choice theory, a system for simulating or engulfs our current economic system was not modelling social and economic behaviour the dominant discourse. Donald Mackenzie within a market or market-like system, how agrees, market actors make market-based decisions. OR is often referred to as decision science or Economics had developed in the management science. eighteenth and nineteenth centuries It is key in spawning academic disci- predominantly as what the historian of plines such as game theory, cybernetics, economics Mary Morgan calls a “ver- cognitive science and even artificial intel- bal tradition.” Even as late as 1900, ligence all of which employ some form of “there was relatively little mathematics, digital and mathematical model.

62 Maria Dada: QUEERING GLOBAL INFORMATION SYSTEMS

Modelling allows these disciplines, and control borders. Most maps prior to 1500, decision science more generally, to make including the oldest map that remains in economic sense of any element in a market existence, the clay “Babylonian world map” or market-like system regardless of what that dating back to the 6th century BC, were element is or how small it is. In other words, created for cosmological speculation rather digital modelling is a means of determining than territorial redistribution. The map as a the state of a market-like system through measure and distribution of resources didn’t mathematical non-representational methods, begin until after the 1500s with examples of methods that are not based in narrative or the Habsburg emperor Phillip II of Spain who meaning but rather elements, actors, cogs commissioned surveys of his various pos- and perhaps dividuals. sessions in differing territories. The next few sections will look at how As a matter of fact, very few maps have the map, more specifically, as an apparatus survived from the Greek, Roman or Medieval of governance, has been transformed as a era. There are a lot of descriptions of maps consequence of digital modelling, how global and how to create them, including Ptolemy’s GIS comes to take over from the map as the Geography and the various different suc- new apparatus for a new style of governance, ceeding comments on it. However, what one that is based in non-representational we know as the Geography was more often economic principles. referred to as the ‘cosmographia’. Ptolemy and his commentators such as the medieval scholar Al Khawarizmi intended to use maps in order to speculate on the known world. Necropolitics and the map They did not survey it with the aim of dividing it up and creating zones. This section will explore the significance of In a sense, as Scott proclaims, the map, and consequently the discipline of cartography, as an apparatus of necro- The premodern state was, in many political power, the power of the sovereign to crucial respects, partially blind; it knew take away life as a form of punishment for precious little about its subjects, their transgressing his law. It will show that necro- wealth, their landholdings and yields, political regimes are interested in maps in their location, their very identity. It order to enforce taxes, voting patterns and lacked anything like a detailed ‘map’ of population control and management. Maps, its terrain and its people. (Scott 2) in that sense, are the performative utterance of sovereign space as the playing field for That is not to say that there weren’t governmentality and power. any map-like drawings conceived of to man- As mentioned in the introduction and as age particular problems such as plans and critical cartographers such as Denis Wood drawings of cathedrals. Many of such draw- and James Scott have made evident, the ings served as a form of inventory but none map is a relatively recent apparatus dating surveyed the many details of the land as the back only to the 1500s. Its lineage coincides topographical maps by the time they were with that of the sovereign and state power. completed in the 20th century. They were Prior to the 1500s few maps were created more interested in the plan of a restricted in the vein in which they were drawn under area for a specific use. In many cases when necro-political rule, to assign territory and map-like drawings did exist, such as the

63 APRJA Volume 8, Issue 1, 2019

Arabic naqshah, these graphical representa- representation of the earth’s surface, hiding tions were not referred to as maps. In that the performativity of the process of its own sense, they are more akin to paintings. The creation. In this sense, maps are an onto- structure of the modern map is fairly recent logical claim of the existence of the state. and it coincides with its paradigm of use. The next section will delineate further the Contemporary maps, that demarcate relationship between maps and the territory territory, didn’t begin to appear until after they demarcate. the 1500s. Most heads of state around that time continued along the direction of mapping space, infrastructure and land under the sovereign’s control. For instance, Jean- Territory Baptiste Colbert, a minister of home affairs under Louis XIV, ordered the surveying and The geographer Stuart Elden dedicates his mapping of the whole of France in 1663. monograph The Birth of Territory to show- The most extensive cartographic pro- ing that the notion of ‘territory’ refers less ject occurred in France after the end of the to the land but rather more generally to the conflict between Spain and France following measure and extension of sovereign power. the Treaty of the Pyrenees in 1659, a treaty As such its meaning is mutable and based on that results in a joint commission to set the the varying forms of sovereignty that appear boundary between the duelling states. throughout history. He claims, The boundary between Spain and France was instated as the first official boundary Territory is not simply land, in the in Europe. Other notable boundaries were political-economic sense of rights of the result of the cartographic work of the use, appropriation, and possession Cassini family over four generations, the first attached to a place; nor is it a nar- trigonometric map regarded as a topographic rowly political-strategic question that land survey. Not to mention the fortification is closer to a notion of terrain. Territory of the Sébastien le Presetre de Vauban comprises techniques for measuring country. Any institutional history book would land and controlling terrain. Measure point to multiple examples from European and control—the technical and the history and beyond of sovereigns ordering legal— need to be thought alongside the surveying of their territory. Mapping was land and terrain. (Elden 322-323) a key proponent of what Foucault would call disciplinary state sovereignty. The technical that Elden is referring What is being proposed here is that to is synonymous with mapping techniques the lineage of the emergence of the modern which, as I will show, later become model- map coincides with the lineage of disciplinary ling techniques taking over from the map’s society and necropolitical power. The bigger form of measurement. Maps allow a certain, claim, however, is that the map is an artefact, representational grasp of the materiality of a mode of writing, technology, that brings the nature, its mountains, deserts and tundras, state’s territory, and therefore the extension not to mention the way maps were used to of its power, into being. In other words, the impose divisions on the colonised. They de- map is responsible for the state’s existence lineate and sustain territory through national and vice versa. The state then goes on to state narrative. affirm the map by insisting that it is a mere

64 Maria Dada: QUEERING GLOBAL INFORMATION SYSTEMS

At the moment, however, the state’s fields such as psychology and biology have stronghold on the map is weakening because become more akin to ‘mapping’ than car- the structure of the state and its institutions, tography. The model, in its economic sense as Deleuze professes, “is in crisis” (Deleuze but also in a wider sense of mathematical 444). The state has not been eradicated, as modelling, expands the notion of territory such, but rather qualitatively transformed due without excluding previous formations of it. to global geo-economic conditions and the The difference between the two is a matter neoliberal governmentality that has emerged. of temporal and geometric scale and the way State territories have been reconfigured in that neoliberal governmentality operates at response to global trade influence. These these varying scales. contemporary conditions do not abolish or Modelling is not restricted to physically confine state territories but rather produce observable phenomena such as Newtonian new state spaces that are entangled in trade physics and geometry. For instance, the relations and new forms of competition. The weather can be modelled in what is referred institutional questions that concern the state to as real time. In the same vein, modelling, no longer converge and in that respect as Neil stretches to cover many aspects of social and Brenner makes clear in New State Space it political life such as voting patterns, criminal might be misleading to speak of ‘the state’ offending patterns, the tax value of homes, as such. Indeed, this is Henri Lefebvre’s bus routes, bike paths as well as consumer point when in the 1970s he discusses the preference. And yet it doesn’t exclude things ‘explosion of spaces,’ a concept then only like the modelling of farm land, roadblocks or in its infant stage. The institutions, regula- other infrastructure. Mapping, on the other tory agencies and markets that comprise the hand, operates only at the Newtonian scale, state are no longer easily demarcated and in the observable and representational, and that sense are somehow in crisis. encompasses areas interpretable through The representational scalar vocabular- signification and language, signs, semiotics ies of the map have been ill equipped to etc. The next and final section will trace the describe the new geo-economic interde- lineage of GIS to show its links to economic pendencies, interdependencies that have practices. come to demand a new style of governance where the market and economists intervene at every level. Consequently, the discipline of cartography becomes more detached from GIS the practice of mapmaking in the traditional sense of drawing maps with pens, paper The story of the digital map in the1960s sheets and hand drawn projections. There is coincides with the emergence of computer a decline in cartography in favour of a more modelling techniques, social econometrics economics-based and consequently non- and the infiltration of these practices into representational model and this logic runs the field of geography. However, creating parallel with the restructuring of territory and maps with computers in those days required perhaps the state altogether. sophisticated graphical mapping applications In the digital era map making is more which didn’t mature until much later. Even readily referred to as global information sys- before their advent, however, computers tems (hereafter GIS), surveying, city plan- were still modelling data for urban analysis. ning or real estate planning. Even drawn out

65 APRJA Volume 8, Issue 1, 2019

In other words, the applications that later became the digital map didn’t have a map to begin with. They were created in order to forecast population information for use by city officials, planners and businesses. One of the first geographers to lay the grounds for the digital map was a scholar named Howard Fisher from Northwestern University. He founded the Laboratory for Computer Graphics and Spatial Analysis. Fisher began as an architect and then came to setting up a company that adapted factory methods to the creation of prefabricated houses only to see his company fail with the pressures of the Great Depression. It was at Northwestern that Fisher ap- pointed programmer Betty Benson to develop the Synagraphic Mapping System (hereafter SYMAP). Tensions were present between Figure 1: A map produced by Carl Steinitz while at MIT the mid-century cartographic community and in the mid-1960s called “The Principle Local Activity of Fisher’s new practice of spatial analysis. a Place.” However, for those wanting to see modelling enter the discipline of geography, the map funded by business groups such as the Ford was seen as a vehicle that would enable Foundation. The so-called maps, such as the geography to rise and become a science. OXAV and SYMAP were complex and had Quantitative geography was quite dif- their own symbols with an accompanied user ferent from its qualitative counterpart as it manual that explains how they were to be had more in common with that of econom- interpreted. None had a drawing of a terrain ics or economic geography. So much was or land. clear with the recruitment of William Warntz It was only through the work of Warntz, as associate director of the Laboratory of his experience as an economic geographer Computer Graphics and Spatial Analysis from and his work on urban simulation routines, his previous post as an economic geographer not to mention the remainder of the money at Penn. He ended up working on the urban from the Ford Foundation that SYMAP’s simulation routine called METROPOLIS drawings began to resemble geographical which used SYMAP to create an animated maps. And even when this occurred it was cartography of Lansing, Michigan. only in order to simplify the display and read- One of the first so-called digital ing of population data so that a layman would maps was used to map urban blight in the be able to interpret the results. Washington city of Spokane. The map looked The point is that much of what is nothing like a geographical map but was referred to as GIS is based in the mathemati- rather a graphical representation of popula- cal modelling techniques that come out of tion in the aid of control. Most of the research economics and various forms of market- in early digital mapping which was later to based decision theory, techniques that are become the now extensive field of GIS was largely quantitative, non-representational

66 Maria Dada: QUEERING GLOBAL INFORMATION SYSTEMS and adhere to a particular style of govern- they lack the capacity to generate narrative. ance. And yet, GIS, like its predecessor the Studies by computational media pioneer Fox map, remains the performative utterance of D. Harrell show that there are alternative a territory that can only be known through ways to engage computational modelling. the model itself, a model that is supposedly With the help of computational and algebraic non-representational. As Michel Callon and linguistics, Harrell aims to derive meaning Donald Mackenzie claim, despite being able from what is otherwise viewed as structural to make do without the representational and numerical datasets, datasets that drive idiom, models remain performative within the many computational models. world that constructs them. For example, with the Advanced Identity Representation (hereafter AIR) vatar Platform Harrell constructs a system that reveals patterns in various modelled virtual Conclusion identities. AIRvatar helps reveal the meanings behind a system, and consequently its A remarkable difference between maps and biases, of which model and dataset crea- models is that maps as descriptors of the tors may or may not have been aware. As earth’s surface contain graphical visualisa- a platform it has been instrumental in the tions that rely on imaging hermeneutics and discovery of statistical patterns of race and the application of signification or meaning. gender discrimination in video games. Simply put, they are comprised of signs, the Harrell has also been looking into lines of the borders as signifier to the territory alternatives to economic modelling in social as signified. The interpretation of the map is media, models that do not assume all ac- a function of power and society as it relates tors to be motivated by economic decisions. to the way the map is drawn. Nonetheless, In the online interactive game Chimera: its value as a representation creates a kind Gatekeeper Harrell constructs a dataset of regulative fiction where the becoming of that attempts and maintains the fluidity of state territory is sustained socially through the user’s identity in relation to the changing national narratives. context of the interactive narrative. In other words, the national discourse What Harrell’s experiments show is that constructs and maintains the identity of the algebraic linguistics could be used to read state as that which is acted out by the per- meaning into the so-called dividual elements formative utterance of the map. The question of data that models produce and manipulate. of resistance becomes one of manoeuvring Doing so would enable a type of critique or through normative frameworks. resistance to the territorialisation by various Digital models, on the other hand, may model structures. A map describes only what have simulations, executions of the models, is seen on the surface of the body of sover- that resemble maps or take the form of rep- eign territory. Models on the other hand ex- resentations and visualisations but primarily tend their performativity deep into the crust, they are mathematical entities, quantifiable tracing ecologies and patterns wherever they and statistical. They divide their subject into may be found. However, by reading meaning manageable units that are not primarily into these patterns and datasets there is no representational. reason why critical GIS practitioners could However, and this is what the paper not continue to construct an extensive critical will conclude on, this does not mean that discourse and practice.

67 APRJA Volume 8, Issue 1, 2019

MacKenzie, Donald, et al. Do Economists Works cited Make Markets?: On the Performativity of Economics. Princeton University Press, Brenner, Neil. New State Spaces: Urban 2008. Governance and the Rescaling of Statehood. Oxford University Press, 2006. Mirowski, Philip. More Heat than Light: Economics as Social Physics: Physics as Callon, Michel. The Laws of the Markets. Natures Economics. Cambridge University Blackwell, 1998. Press, 1995.

Callon, Michel, et al. Market Devices. Mirowski, Philip. Machine Dreams: Blackwell, 2008. Economics Becomes a Cyborg Science. Cambridge Univ. Press, 2006. Elden, Stuart. The Birth of Territory. University of Chicago Press, 2013. Scott, James C. Seeing like a State How Certain Schemes to Improve the Human Deleuze, Gilles. October: the Second Condition Have Failed. Yale Univ. Press, Decade, 1986-1996. Edited by Rosalind E. 2008. Krauss, MIT Press, 1997. Wilson, Matthew W. New Lines: Critical GIS Foucault, Michel. The Birth of Biopolitics and the Trouble of the Map. University of Lectures at the College De France, 1978- Minnesota Press, 2017. 1979. Translated by Michel Senellart, Palgrave Macmillan, 2011. Wood, Denis, and John Fels. The Power of Maps. Routledge, 1993. Lefebvre, Henri. The Production of Space. Translated by Donald Nicholson-Smith, Wood, Denis. “Cartography Is Dead (Thank Blackwell, 2009. God!).” Cartographic Perspectives, no. 45, 2003, pp. 4–7., doi:10.14714/cp45.497. Longley, Paul A. “The Academic Success of GIS in Geography: Problems and Prospects.” Journal of Geographical Systems, vol. 2, no. 1, 2000, pp. 37–42., doi:10.1007/s101090050027.

Openshaw, S. “A View on the GIS Crisis in Geography, or, Using GIS to Put Humpty- Dumpty Back Together Again.” Environment and Planning A: Economy and Space, vol. 23, no. 5, 1991, pp. 621–628., doi:10.1068/ a230621.

MacKenzie, Donald A. An Engine, Not a Camera: How Financial Models Shape Markets. MIT Press, 2008.

68 69 Tiara Roxanne

DIGITAL TERRITORY, DIGITAL FLESH: DECODING THE INDIGENOUS BODY

Abstract

Western Indigenous cultures have been colonized, dehumanized and silenced. As AI grows and learns from colonial pre-existing biases, it also reinforces the notion that Natives no longer are but were. And since machine learning requires the input of categorical data, from which AI develops knowledge and understanding, compartmentalization is a natural behavior AI undertakes. As AI classifies Indigenous communities into a marginalized and historicized digital data set, the asterisk, the code, we fall into a cultural trap of recolonization. This necessitates an interference. A non-violent break. A different kind of rupture. One which fractures colonization and codification and opens a space for colonial recovery and survival. If we have not yet contemporized the colonized Western Indigenous experience, how can we utilize tools of artificial intelligence such as the interface and digitality to create a space that de-codes colonial corporeality resulting in a sense of boundlessness, contemporization and survival?

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Tiara Roxanne: DIGITAL TERRITORY, DIGITAL FLESH

For to survive in the mouth of this Indigenous body and the removal of (colonial) dragon we call America, codes this body is tied to. We must therefore we have had to learn this first and vital confide in the biologic and the historical and lesson – that we were thereafter, enter the digital. Simply put, we never meant to survive. Not as human must go beyond the flesh. beings. By going beyond the flesh, we enter the — Audre Lorde, “The Transformation digital. This is an attempt to de/reprogram of Silence into Language and Action.” the Indigenous/coded body by entering a digital territory, one that is made possible via the interface. The interface is the lacuna between human and virtual worlds. Such a Introduction lacuna situates the Indigenous body outside of colonial/physical territory. It disentangles Aztecas del norte, mojados, Indigenous territory and makes boundlessness possible peoples, First Nations People, mestizas, for the Indigenous body to inhabit. This is Redskins, Indians, Native Americans, digital territory. This is where one embodies Natives, savages, minorities, at risk peoples digital flesh. Since the contemporization of or asterisks peoples are some names or Indigeneity is not possible within its current codes the Indigenous body is subjected to colonial paradigm, I am speculating on the using settler colonialist language. The settler radical possibility of colonial recovery within names the Indigenous person or body which a posthuman digital framework. codifies and marginalizes. Not only does AI learn from these colonial pre-existing biases that codify and mar- ginalize, it also re-inscribes the notion that Indigenous body, Natives no longer are but were. As AI codes indigenous borders Indigenous bodies according to its colonial input, it also classifies these communities The body is a biological figure that identifies into a marginalized digital data set, the aster- and is identified by the space that surrounds isk, the code. As AI codes the marginalized it. It encompasses dimensionality and is en- Indigenous body, it reproduces historical cased within a dimensional structure. All bod- erasure of Indigeneity which necessitates an ies live in spaces that dwell within borders. interference. A non-violent break. A different However, spatiality for the Indigenous body kind of rupture. One that destroys the settler is both territorial and historical, a byproduct colonialist triad and interrupts AI bias and of colonialism, a designation of territorial promotes survival. acquisition and forced migration. The body Here, I summon the source from where that was colonized will always be colonized, Indigenous subjectivity originates by return- more specifically, the Indigenous body (of ing to the body and the land this body inhab- the West). The Indigenous body, however, is its, by breaking the boundaries it is bound by subject to colonialism and more specifically, and begin to speculate on the notion of digital settler colonialism, a term used to describe territory and possibly even digital flesh. the colonialist relationship between the This kind of return to, and rupture of, Indigenous peoples and the colonizer. The the Indigenous biologic is one of ontological concept of Indigenous is inspired by Audra abstraction: One which focuses solely on the

71 APRJA Volume 8, Issue 1, 2019

Simpson who writes in her book, Mohawk Moreover, “a border is a dividing line [where] Interruptus, that “‘Indigenous’ is embedded the prohibited and forbidden are its inhabit- conceptually in a geographic alterity and ants” (Anzaldúa et al.). Borders separate the a radical past as the Other in the history of settler from the Indigenous where the settlers the West” (7). Indigenous peoples are pre- “make Indigenous land their new home and colonial peoples with a narrative that is geo- source of capital” and the Indigenous are graphically, cosmologically and ontologically pushed out (Tuck and Yang 5). This record tied to their land (within Central and Northern of geographical domination is a fundamental Americas, for instance). Their relationship to colonial classifier, also known as “settler land and identification as such starts with ter- colonialism,” one which occupies and estab- ritory which carries a polyvalence regarding lishes the Native land through erasure (Tuck ancestry, origin, spirituality and so forth. et al. 5). Furthermore, this spatial circum- Specifically, the Indigenous body scription reattributes the Indigenous’ overall refers to the biologic, social, and political experience in and of the world. By framing colonized Indigenous person of the West. the Indigenous body between physical and Again, Simpson writes, “indigeneity is political structures and by claiming their land, imagined as something entrapped within the settler erases Indigenous identity and the analytics of ‘minitorization,’ a statistical history. model for the apprehension for (now) racial- This total migration of force pushes ized populations ‘within’ nation-states” (211). the Indigenous body into a space of wilder- The Indigenous peoples are minoritized and ness, the forbidden and the prohibited, the colonized. According to Simpson, “Because erased – a ghost territory. This demand is a ‘Indigenous’ peoples are tied to the desired process of naming or anti-naming the body territories, they must be ‘eliminated’; in that is forced out of their homeland. To name, settler-colonial model, ‘the settler never or take one’s name away, determines an en- leaves’” (19). Indigenous peoples had their gendered locality, i.e. coding the body, which land stolen and repurposed within the settler is an “ordering of matter around a body” (qtd. colonialist structure, one which assumes in Hanson). As the Indigenous are coded, Indigenous erasure. Thereafter, spatiality for their body is degenerated from embodied the Indigenous body is both territorial and corporeality to mere flesh. Hortense Spillers historical, a byproduct of settler colonialism, reminds us about the division between body a designation of territorial acquisition and and flesh, she writes forced migration. In this way, space develops as a ges- [...] the distinction as the central one ture of colonization where borders mimic this between captive and liberated subject- “system of dominance,” and subjugates the positions. In that sense, before the Indigenous body (Osterhammel 4). Such a ‘body’ there is ‘flesh,’ that zero degree system aims to create a space of segregation of social conceptualization that does where the Indigenous are territorially, socially not escape concealment under the and politically trapped. brush of discourse, or the reflexes of When “borders are set up to define iconography (67). the places that are safe and unsafe, to distinguish us from them” (Anzaldúa 25), the The body that is subjected to, imprinted Indigenous body is claimed not only by the upon, named or coded is done so according settler but also by the borders that surround it. to its flesh. Where Frank B. Wilderson might

72 Tiara Roxanne: DIGITAL TERRITORY, DIGITAL FLESH refer to the “presence of the body” (“The Inside-Outside of Civil Society”) in reference Indigeneity: Body memory to Spillers’ notion of ‘flesh’, the Indigenous and flesh memory body who is re/moved and named suffers a similar antagonism. A history grounded in the removal and eras- The Indigenous loses their identity ure of Indigenous culture, identity and bodies as well as their sense of belonging to their therein, encapsulates memories passed homeland. And since “flesh is the fundamen- down through generations of misplaced tal indifference between body and world” and coded bodies, is carried through body (Hansen xi), the Indigenous people suffer memory and flesh memory. from this codification process done so by Traumatic events the body experi- the settler. Again, Spillers’ notion of flesh ences are passed down as bodily memories, exemplifies this codification, as the flesh is encompassing a corporeal memory archive positioned and held in line with ‘captivity’ also known as body memory. Recent stud- (67). The flesh that is named and marked is ies on epigenetics displayed in the article, imprisoned accordingly. “Trauma May Be Woven into DNA of Native The marginalized space or territory binds Americans” insist that “our genes carry mem- the Indigenous body within borders, icono- ory of trauma experienced by our Indigenous graphically and geographically. Furthermore, ancestors” (Pember). Meaning that “trauma the settler names the Indigenous according experienced by earlier generations can influ- to their flesh which is a codified identification ence the structure of human genes, making process that further marginalizes the body. them more likely to ‘switch on’ [negative] Thereafter, the Indigenous body is referred responses to stress and trauma” (Pember to as, but not limited to the following names et al.). More importantly, these traumatic or codes; Aztecas del norte, mojados, experiences influence gene structures which Indigenous peoples, First Nations People, are physically and psychologically revealed. the mestizos (people mixed of Indian and They are expressed symptomatically; two Spanish blood), minorities, ‘at risk peoples’ examples are Post Traumatic Stress Disorder or ‘asterisks peoples’, “meaning they are rep- (PTSD) and depression. For the Indigenous resented by an asterisk in large and crucial body, the suffering remains unresolved by data sets” (Tuck et al. 23). This codification proxy — to be Indigenous is to always be [in] of naming a community of bodies or an the trauma body. That traumatic body mem- individual body dehumanizes and colonizes ory lies dormant in the peripheral nervous the body being named/anti-named. It is this system waiting to be triggered by reminders codifying that then serves as a placeholder of the trauma event whether experienced for machine learning systems which conceive personally, secondarily or genetically. and reproduce colonization of the Indigenous Another form of memory that is exhib- body, commonly referred to as AI bias. ited via the body is what Alikah Oliver called flesh memory. She defines flesh memory by first quoting the definition of flesh in the American Heritage College Dictionary and flesh memory later in her own words:

73 APRJA Volume 8, Issue 1, 2019

flesh (n): 1. the soft tissue of the body compartmentalization is a fundamental of a vertebrate, consisting mainly of behavior AI undertakes. As AI grows and skeletal muscle and fat. 2. the surface diversity is tackled through the non-binary, or skin of the human body. or rather, against the universal, we fall into a trap of re-colonization, or digital colonization. flesh memory: 1. a text, a language, Two terms that digital colonization a mythology, a truth, a reality, an draws from are data colonialism and digital invented as well as literal translation of colonialism. Data colonialism “combines the everything that we’ve ever experienced predatory extractive practices of historical or known, whether we know it di- colonization with the abstract quantification rectly or through some type of genetic methods of computing” (Couldry 1). And memory, osmosis, or environment. 2. digital colonialism is “a quasi-imperial power the body’s truths and realities. 3. the over a vast number of people, without their multiplicity of language and realities explicit consent, manifested in rules, designs, that the flesh holds. 4. the language languages, cultures and belief systems by a activated in the body’s memory. (4) vastly dominant power” (“Resisting Digital Colonialism”). Both are hegemonic digital re- These somatic experiences for both inscriptions of historical colonization. More body memory and flesh memory are activated specifically, each use and integrate methods and felt as present when past traumatic or of data collection via algorithms and machine violent memories arrive or are triggered. For learning systems which creates a general both types of memory, they require the body data identity stripping away any form of indi- and the history of that body, a basis of ontol- vidual or body. ogy. A history grounded in the removal and Furthermore, this kind of data collection erasure of Indigenous bodies encapsulates serves as a type of surveillance which Gary T. these types of memories. Marx calls “the new surveillance” (206). This Since trauma is an inherent part of new form of surveillance is divvied up in ten the Indigenous experience, both biologi- sections, a few of which point directly to the cally (body memory) and ontologically (flesh sharing of data, the storing and compressing memory), there is no way out. Recovery from of data and specifically that “data collection is colonization and trauma is challenged ac- often done without the consent of the target” cording to the body’s situatedness. And since (Marx 218). Identity is not only generally AI learns through the colonial paradigm, it based on the data that is collected, it refuses is also re-colonizing and traumatizing the an ontological perspective. It disavowals the Indigenous body; thereby, digital colonization body, the being, the historicity one’s body and artificial intelligence bias are also crucial carries in body memory and flesh memory is to critically integrate. dismissed, overlooked. In other words, the data refuses to acknowledge the marginalized body, i.e. the Indigenous body in the margins, whilst AI bias re-marginalizing it which, in turn, is digital colonization. And since data collection has Since machine learning requires the nothing to do with beingness or the bodily and input of categorical data, from which AI because the historicity is so much a part of the develops knowledge and understanding, Indigenous experience, and the experience

74 Tiara Roxanne: DIGITAL TERRITORY, DIGITAL FLESH of living, more generally, it continues to erase awakening, not a resistance or recalibra- history. It persists in colonizing and thus is tion, because, remember, computers do not how I am determining digital colonization forget. through data collection systems. Such data In order to disrupt the pre-existing co- collection systems, like the algorithm, are lonial input of AI, the Indigenous body must taught via machine learning, for example, to interrupt their own subjectivity which relies collect and produce categories of identifica- heavily on history and territory. Herein lies tion which further reduces the identity of the the importance of ontological abstraction. Indigenous body to a code. The experience of trauma, whether it be Therein lies a danger: the codification displacement or otherwise, such as ances- process of AI engages in biases that classify, tral genocide or any other kind of violence categorize and codify the Indigenous body against one’s body, is ontological because even further. And because AI learns from it is implicitly biological and being-oriented. pre-existing biases and collects data based And abstraction allows for a different kind of on these biases which further marginalizes, experience or beingness to arrive. it is not only re-colonizing, it is erasing what Abstraction here is supported by Sylvia has not yet been contemporized. Wynter’s notion of autopoiesis, a term used to AI is learning to perceive the world describe “subjects given over to death within based on its colonial input, and is acting as a a certain regime of being human/ human disembodied surveillance that re-categorizes knowing” (Hantel 3). The subjects who are bodies based on general data collection. “given over to death” (Hantel 3) are liminal in Since AI codes and thus digitally colonizes their colonized state of being. Fundamentally, through multiple factors such as AI surveil- a liminal subject is a colonized human being lance systems and data collection, I want who is forced to be within a mode of constant to meditate on the question: can AI provide survival. a space for the Indigenous body to digitally Wynter’s notion of the “liminal subject” reorient? derives from abstraction. It strikes a chord when expressing the body as biologic, autopoietic and perhaps even represented as a multispecies. Firstly, her liminal sub- Digital territory, digital ject characterizes the subject as being on flesh “the threshold of a new world in the midst of cultural ritual” (qtd. in Hantel 69). In this way, we can understand Wynter’s liminal Settler colonialism, AI surveillance and data subject in terms of Indigenous culture and collection compartmentalizes the Indigenous ritual practice. Perhaps the liminal subject is body which paralyzes it to a constant state formed through an abstraction which allows of colonization: “I cannot decolonize my for corporeal overrepresentation through body.” There is no way out of this body, this means of survival. trauma, this memory. There is nowhere to go. This survival is exhibited through ritual. Now more than ever, with such embedded By returning to ritual, Indigenous peoples social, political and digital hierarchies, the unify through memory and tradition and Indigenous experience is at risk of historical return to the cosmological. However, for erasure. The intermingling of each sphere colonized / Indigenous people, it is important produces a great need for disruption and to discover means of survival in a world

75 APRJA Volume 8, Issue 1, 2019 that was meant for its antithesis. It has never been safe to practice ritual, even on Digital lacuna: ‘given land’. Indigenous peoples are killed off The interface through historical mediums of representation as well as through technological representa- By inserting the already codified body into a tion. Meaning that this is a kind of ‘death’. virtual and boundless landscape, Indigenous This death, or data-digital erasure, forces the peoples disrupt their own subjectivity and Indigenous body to find new ways of survival. corporeality as well as contemporize their And because the Indigenous body is bodies as memory systems and flesh. This stuck within a colonized world, territorially break in the sphere opens a space for and digitally, it is important to imagine the rearticulation. ‘other-worldly’ to veer from settler colonial- The interruption is corporeal contem- ism. In this way, the Indigenous re-imagines porization and survival. This is possible their subjectivity through overrepresentation. because datafication refuses an ontological For example, “the liminal subject assumes a acknowledgement. Though the Indigenous structural role at the limit of the overrepre- would not re-enter their territory and claim it sentation of Man, indexing an outside to our back, as if it is even possible, the Indigenous current descriptive statement by their very would need to enter a technological posthu- existence and paradoxical survival [...]” (qtd. manist framework, the virtual, the digital via in Hantel 70). Thereafter, as a liminal subject, the interface. the Indigenous is given opportunity to recover However, before entering a borderless and possibly even reclaim, contemporize and cartography, as made possible through ma- survive. chine learning, it is important to distinguish This overrepresentation of the liminal the differences between architectural and subject, within a technological framework, virtual spaces. As it is experienced in the is envisioned as digital flesh. However, prior physical world, architecture manipulates the to speculating digital flesh more elaborately, body to move through space and thus the it is necessary to first understand the space body forms an understanding of itself, creates the body needs to enter before the idea of meaning-making and applies knowledge to digital flesh is even possible. and of the world. Galloway writes on account The Indigenous body must discover of Deleuze, “that one should not focus so a sense of boundlessness that gives way much on devices or apparatuses of power to subjective interruption, therefore, the they mobilize, that is more on the curves Indigenous body must reterritorialize where of mobility and force,” further explaining, “each one of his [their] organs, his [their] “these apparatuses, then are composed of social relations, will, in sum, find itself [them- the following elements: lines of visibility and selves]re-patterned, so as to be re-affected, enunciation, lines of force, lines of subjectifi- over-coded as a function of the global re- cation, lines of splitting, breakage, fracture” quirements of the world” (Guattari 10). To (qtd. in Galloway 18). By applying these arrive at a space of reterritorialization, the apparatuses to the Indigenous experience, a liminal subject (as overrepresented) is placed different landscape is possibilized, one that outside of the parameters it is bound by. This is not so manipulated or reduced to by AI. kind of reterritorialization implies the need for The landscape portrayed here derives a new landscape. from architectural technics, a term used to describe the technological space architecture is

76 Tiara Roxanne: DIGITAL TERRITORY, DIGITAL FLESH growing into, most specifically the interface. It becomes digital flesh. As it is placed in ac- The interface is more than an infrastructural cordance with the interface, its meaning and space. It is a threshold, a space of mediation identification extends beyond the boundaries between body and world, both physical and of the embodied — the human/corporeal virtual. A gap. A lacuna. It is cartographic flesh and its nervous system. plasticity. “The interface is not something that The Indigenous body is no longer hu- appears before you but rather is a gateway man flesh: it is a digital body. And by invocat- that opens up and allows passage to some ing Francis Bacon’s notion of force, where the place beyond” (Galloway 30). It is in the body serves as a mediating horizon between space of passage between the physical and self and interface, the Indigenous body seeks virtual spaces that is the break or disruption to reestablish a grounding where experience or rupture. The passage is the interface. This develops as a somatic relationship between liminality between the physical and the virtual self and virtual, self and digital, a boundless embodies movement though it is actualized space which delineates traumatization, i.e. as an interface. colonization, by identifying the body beyond By arriving into a space that is not the corporeal and symbolic flesh into the named as imperial or colonial as such be- space of the digital. The corporeal body here cause it is its own structure outside of the serves as Bacon’s notion of force within a body — a moving and/or malleable struc- digital landscape. ture — it destabilizes normative corporeal Here, I lean on Galloway where he thought, that which identifies the body as writes on two kinds of spatial digitality which corporeal flesh. are flat digitality and deep digitality, he writes, Here, the Indigenous body specu- “Flat digitality results from the reduplicative lates the possibilities of de/reprogramming multiplexing of the object” (68). And then he beingness. A gesture of de-coding. The reviews deep digitality: Indigenous body does not resist or protest digital colonization but navigates through it Deep digitality results from the by entering digital territory via the interface. reduplicative multiplexing of the Within the digital space, the Indigenous body subject. Instead of a single point of is then “over-coded,” as Deleuze writes (qtd. view scanning a multiplicity of image in Galloway 18). However, the complexities feeds, deep digitality is a questionof of subjectivity greatly evolve regarding eth- a multiplicity, nay an infinity, of points ics here. Meaning, for example, subjective of view flanking and flooding the interruption between a natural world and a world viewed. These are not so much non-natural or virtual/digital world refers to matrices of screens but matrices of multifarious meanings that transcend em- vision. (Galloway 68-69) bodiment and require refusal. Thereafter, deep digitality is the kind of digital and territorial unfolding the Indigenous body would become into or through, as a gesture Non-being & survival of reterritorialization, rupture and force. This is not to say that this decentering The Indigenous body becomes something of corporeality, this disembodiment, is a else. It becomes something other than only positive move toward transcendence, it is data. The Indigenous body becomes digital. only stating that the interface possibilizes a

77 APRJA Volume 8, Issue 1, 2019 different kind of subjectivity, perhaps what territory and become or start to embody digi- Daniel Colucciello Barber refers to as non- tal flesh. Both digital territory and digital flesh being and the “no-thing” in reference to his are made possible through machine learning interpolation of difference (“The Creation of and computation and enter a deep digital ter- Non-Being”). He first discusses the notion of ritory. This kind of embodiment of subjectivity being and writes, “being — or the possibility or beingness perhaps is the non-being or the thereof — grounds itself not through its own ‘no-thing’ Barber (dis)assembles. coherence, but through an enactment of Of course, one must lean into the power that is staged by anti-black violence” metaphor here and think radically about the (Barber et al.). Barber continues further by Indigenous biologic and about digitality as elaborating the existence of beingness or immanent where machine learning facilitates “non-being” as the refusal of beingness and what the Indigenous body becomes within the “no-thing,” he states: digital territory as digital flesh, the multispecies, the liminal subject, the body that Difference antecedes both positive refuses colonization, negates beingness and being and negative being [...] In other welcomes contemporization and survival. words, difference is not between op- This is not a sim character or machinic posed beings but in itself, autonomous extension of oneself. It disrupts subjectivity from and antecedent to ever being or and mediates beingness outside the limits of thing; difference is real, but precisely general data collection. Outside of the flesh. as a matter of non-being. Its reality is It is unknown digital territory. And because it not the being of the thing, it is no-thing. is unknown, it characterizes the uncharac- (par. 13 et al.) teristic, the non-being or the ‘no-thing’, an immanent adventure within the digital, one In this way, we can apply non-being that looks back at colonialism and machine and Barber’s definition of no-thing to the learning and enters the digital lacuna, mak- digitally incorporated Indigenous body. The ing recovery and survival imaginable. Indigenous body acquires its own power or self-reclamation through difference and refusal. In this way, the Indigenous body refuses its colonial subjugation, or present- time beingness, by becoming or embodying the no-thing, as made possible through the interface, a deep digital lacuna between natural world and virtual or digital world. For the Indigenous body to enter a disembodied digital landscape is to perform a potential for survival and even contemporization. Hereafter, the Indigenous is not in recognition of Self (or beingness) within a geographical, political and colonial structure, the coherence here remains outside the body. Thereafter, the refusal of colonization and codification gesticulates the Indigenous body to enter deep digitality terrain or digital

78 Tiara Roxanne: DIGITAL TERRITORY, DIGITAL FLESH

Elizondo, Virgilio P. The Future Is Mestizo: Works cited Life Where Cultures Meet. Boulder, CO: U of Colorado, 2000. Allen, John. What Is the Future of Artificial Intelligence? Reference Point Press, 2017. Fuchs, Thomas. “Chapter 1. The Phenomenology of Body Memory.” Body Anzaldúa, Gloria. Borderlands La Frontera: Memory, Metaphor and Movement, The New Mestiza. 3rd ed. San Francisco: edited by Sabine C. Koch, Thomas Fuchs, Aunt Lute, 1987. Michela Summa and Cornelia Müller, John Benjamins Publishing Company, 2012, pp. Barber, Daniel Colucciello. “The Creation of 9–22. https://doi.org/10.1075/aicr.84.03fuc. Non-Being.” Rhizomes: Cultural Studies in Emerging Knowledge, no. 29, 2016. https:// Galloway, Alexander R. The Interface Effect. doi.org/10.20415/rhiz/029.e10. Cambridge, UK: Polity, 2012.

Bergson, Henri. Matter and Memory. trans. — . Laruelle: Against the Digital. University Nancy Margaret Paul and William Scott of Minnesota Press, 2014. Palmer. London: G. Allen, 1912. Guattari, Félix. Schizoanalytic Bombay, Amy, Kim Matheson, and Hymie Cartographies. trans. Andrew Goffey. New Anisman. “Intergenerational Trauma.” York: Bloomsbury, 1989. Journal of Aboriginal Health/Journal de la santé autochtone, vol. 5, no. 3, 2009. Hantel, Max. “What Is It Like to Be a Human?: Sylvia Wynter on Autopoiesis.” — . “The Intergenerational Effects of Indian philoSOPHIA, vol. 8 no. 1, 2018, pp. 61-79. Residential Schools: Implications for the Project MUSE, doi:10.1353/phi.2018.0003. Concept of Historical Trauma.” Transcultural Psychiatry, vol. 51, no. 3, 2013, pp. Hansen, Mark B. N. Bodies in Code: 320–338. Interfaces with Digital Media. New York: Routledge, 2006. Couldry, Nick, and Ulises Mejias. “Data Colonialism: Rethinking Big Data’s Husain, Amir. Sentient Machine: The Relation to the Contemporary Subject.” Coming Age of Artificial Intelligence. Television and New Media. Sage, 2018. Scribner Press, 2017. LSE Research Online. http://eprints.lse. ac.uk/89511/ Maccormack, Patricia. Posthuman Ethics: Embodiment and Cultural Theory. Deleuze, Gilles. Bacon. New York: Routledge, 2017. Bloomsbury, 2013. Marcus, Mitchel P. A Theory of Syntactic Deleuze, Gilles, and Félix Guattari. A Recognition for Natural Language. MIT Thousand Plateaus: Capitalism and Press, 1986. Schizophrenia. Translated by Brian Massumi, University of Minnesota Press, 2014.

79 APRJA Volume 8, Issue 1, 2019

Maurice Merleau-Ponty, Phenomenology Wilderson, Frank B. III, Samira Spatzek, of Perception, trans. Colin Smith. London: and Paula von Gleich. “The Inside-Outside Routledge & Degan Paul, 1962. of Civil Society: An Interview with Frank B. Wilderson, III.” Black Studies Paper, vol. 2, Marx, Gary T. Undercover: Police no. 1, 2016, pp 4-22. Surveillance in America. University of California Press, 1988. Wynter, Sylvia. “Unsettling the Coloniality of Being/Power/Truth/Freedom: Towards the McPherson, Stephanie Sammartino. Human, After Man, Its Overrepresentation Artificial Intelligence: Building Smarter — . An Argument.” CR: The New Centennial Machines. Twenty-First Century Books, Review, vol. 3, no. 3, 2003, pp. 257–337. 2018.

Ogden, Pat, Kekuni Minton, and Clare Pain. Trauma and the Body: A Sensorimotor Approach to Psychotherapy. New York: W.W. Norton, 2006.

Oliver, Akilah. The She Said Dialogues: Flesh Memory. Smokeproof Press, 1999.

Oliveira, Arlindo L. The Digital Mind: How Science Is Redefining Humanity. MIT Press, 2017.

Osterhammel, Jürgen. Colonialism: A Theoretical Overview. trans. Shelley Frisch. Princeton: M. Wiener, 1997.

Pember, Mary Annette. “Trauma May Be Woven into DNA of Native Americans.” Indian Country Media Network, 3 Oct. 2017.

Spillers, Hortense J. “Mama’s Baby, Papa’s Maybe: An American Grammar Book.” Diacritics, vol. 17, no. 2, 1987, pp. 64-81. DOI: 10.2307/464747.

Tuck, Eve, and K. Wayne Yang. “Decolonization is Not a Metaphor.” Decolonization: Indigeneity, Education & Society, vol. 1, no. 1, 2012, pp. 1–40.

80 81 Rebecca Uliasz

ASSEMBLAGES OF DESIRE: REAPPROPRIATING AFFECTIVE TECHNOLOGIES

Abstract

This paper considers the operations of affective technology within contemporary technocapitalism through affect theory. It is argued that affective technologies enter into power arrangements with political and corporate interests, altering an acting bodies’ affect — in the Spinozan definition, the “capacity to affect and be affected” — within social and political life. Affective computation uses machine learning techniques to ‘capture’ and quantify affective intensities in data form, automating a normalizing logic of division and categorization that classifies bodies, emotions, and objects. Affective technologies invoke what Luciana Parisi called “automated decisionism,” where machine learning processes digitize incomputable states in order to impose a self-rationalizing logic structure that regulates a user-subject’s actions (Parisi, “Reprogramming Decisionism”). Affective technologies exert biopolitical control over users through quantified logics of division and devaluation. It is suggested that affect might simultaneously operate as an analytic lens to speculate on whether collective affectivity and political agency might be reclaimed through using these technologies. The following concludes with an engagement with Deleuze and Guattari’s “assemblages of desire” to suggest that affective technologies might produce other micropolitical arrangements that increase user agency as social and political subjects.

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Rebecca Uliasz: ASSEMBLAGES OF DESIRE

affects using machine learning techniques. A tale of two plushies Affective technologies are increasingly ubiquitous to our everyday operations, social In 2013, British supermarket chain Tesco relations, and consumer habits. The ‘age contracted with an American tech startup of artificial intelligence’ forms a networked company to install facial scanning cameras tangle of affective technologies that expand in 450 petrol stations. The motive for doing from the virtual realm of our social media so: selling a toy penguin. ‘Monty the Penguin’ feeds to the aura of ambient technologies was both the title of a heartwarming adver- that pepper each room in our homes. The tisement for a John Lewis plush, and the goals of affective technologies vary — they Christmas gift fad of 2014, selling out stock might be sold as commodities that promise only a few hours after the ad premiered. Back easier and more seamless user interaction, in the US, at the Sunny View Retirement while simultaneously mining troves of user Community in California, a white harp seal data that might be leveraged by corporate named Paro snuggles and coos at elderly or state interests. Affective technologies may patients with dementia. The robot seal has take the form of anthropomorphic robots, been designed specifically for the calming or cute fluffy Christmas presents. They effect it has on its holders and goes for about also might operate subliminally behind the 5,000 USD on eBay, although a number of screen, in order to quantify knowledge about lesser knock-off Paros can be spotted on a user that might be used to sell them future department store websites. A seemingly Christmas presents, or perhaps even sold to arbitrary pair apart from their shared cuddly other agents who might use this knowledge demeanor, Monty and Paro have a similar to justify sometimes violent means of regula- story of origin, although they went on to lead tion over certain bodies. But what kind of user quite different lives. The development of information is extracted by these processes, both Monty’s branding campaign and Paro’s exactly, and by whom? What do affective engineered persona are in fact, the resulting technologies purport to calculate? artifacts of the techniques of ‘artificial emo- This paper argues that affective tech- tional intelligence’, also known as ‘affective nologies appropriate social relations in the computation’. Affective computation is an service of capital. Through techniques of umbrella term for an interdisciplinary set of quantification, affective technologies extract sciences organized around the interpreta- data from user-subjects that is then lever- tion, codification and stimulation of human aged for profit. It is argued that affective technologies participate in what Deleuze and Guattari call an “assemblage” — or arrangement of “bodies, actions and passions, an intermingling of bodies reacting to one another” that “are necessary for states of force and regimes of signs to intertwine their relations” (Deleuze and Guattari 71). Such assemblages, it is argued, are imbedded with power operators that — to quote Spinoza — alter a bodies’ “capacity to affect and be affected” within political society. Within con- Figure 1: PARO the Seal. Credit: PARO Robots USA. temporary technocapitalism, affect — which

83 APRJA Volume 8, Issue 1, 2019 is here defined by a biological, social, the name of optimization — ‘truth’ becomes political or technical body’s capacity for ac- equivalent to efficiency, which under capital- tion — is made commodity through technical ism means whatever is the most productive instrumentalization. Affective technologies of capital (Foucault 19). Framing affective provide a means for the biopolitical control computation as a power operation suggests of users — a term which is here considered that it regulates operations and human synonymous with ‘subjects’ — by those who relations in accordance with the interests of control privatized algorithms and massive capital through the extraction, quantification databases, constraining users to certain pos- and datafication of affective information. If sibilities of action through standardization. affective computation aims to commoditize Is it possible, then, for affective technolo- social relations, might framing them as power gies to be critically deployed, or must they operators within an assemblage suggest be dismissed as irrevocably engrained with what Brian Massumi calls a social “potential oppressive logics of division and devaluation for re-relating with a difference” for the user- of user-subjects? Framing affective technolo- subject (Massumi 54)? Following Deleuze gies as participants in an assemblage per- and Guattari, are there other aspects of mits a critical analysis of their deployment, a this “assemblage of desire,” new affective necessary lens in questioning whether such relations that fall outside of regulatory and technologies might allow for other modes of predictive capacities exerted by the power subject expression with the toolset granted assemblages performed through affective by technocapitalism. technologies? Contemporary scholarship on affect emphasizes the analysis of everyday modes of being and feeling as a linkage between the ‘micro-political’ (how a certain body ex- An atlas of emotions periences a political context) and the ‘macro- political’ (how a certain political assemblage We know that face recognition technology is arranges bodies across society). Normative deployed across the globe in order to surveil, forms of being in society are given though police and regulate algorithmically marked what Deleuze, following Foucault, calls a bodies, but we are beginning to realize “power arrangement” — a formation of insti- artificially intelligent programs may be used tutional and State powers that constitute “the to capture emotions — where ‘emotion’ is whole social field” (Deleuze 123). Power ar- registered as the discrete cognitive states rangements act across micro and macro po- that effect human communication — as well. litical registers in what Deleuze and Guattari As of early 2019, Amazon, Apple, Facebook, call “assemblages of desire” — relationships Google, IMB, Microsoft and other powerful between social subjects, territories, technolo- platform corporations are developing and gies, and institutions — that are in constant rolling out new technologies that utilize what flux and recomposition (125). Affective tech- is known as “emotion AI,” or “artificial emotion nologies are automated participants in a par- intelligence” (McStay 2). Other smaller scale ticular power arrangement that exerts control tech startups offer clients customizable pack- over users by constructing norms through ages for data collection, including a variety statistical standardization. Technological nor- of biometric sensors, cameras, microphones, malization, Foucault explains, is an economic and multi-modal software. An emotionally operation that produces knowledge effects in intelligent technology might make use of high

84 Rebecca Uliasz: ASSEMBLAGES OF DESIRE detail cameras and other imaging sensors computing’ — an engineering practice deal- in order to measure certain muscle expres- ing with machines that ‘have emotions’. sions, generating a representation of an Following the cybernetic dream of bringing emotion based on a model programmed into together the mind and the machine, affective the device. Image and scanning techniques computation has intervened into the broader would ‘capture’ emotions based on micro- umbrella science of artificial intelligence by scopic movements of a human face, which staking its claim — the human mind is always are discretized and assumed to be universal embodied, and humans have emotions, for the sake of calculability. therefore, a more functional intelligence Artificial emotional intelligence takes machine might have the ability to detect plural forms, some designed to imitate human and respond to emotional states. Corporate empathy back to the user (harkening back descriptions of affective technologies often to ELIZA, the original virtual Rogerian psy- use verbiage that collapses the terms ‘af- chotherapist), while others are programmed fect’ and ‘emotion’, which affect theorists to assess user’s dominant emotional state like Brian Massumi assert have important and trigger certain prompts based on their political distinctions. For Massumi, affect is calculations. We might find these tech- a proto-political and pre-subjective ‘charge’ nologies deployed by companies to monitor that is always in flux, whereas emotion is consumer response in order to assess user “the way the depth of that ongoing experi- engagement and dynamically alter advertise- ence registers personally at a given moment” ment content, contributing to what has been (Massumi 4). We might detect each other’s elsewhere called the ‘emotion economy’. emotions through cognitive and social cues, [1] Emotional intelligence might tap into our like the tone of a voice combined with the consumer desires, subliminally determining expression of a face, but a smile or grimace the future choices and actions we will take. can’t be assumed to imply the same informa- In this sense, artificially intelligent emotion tion universally. While emotion is bound to technologies make a wager on our (yet) the individual subject, the concept of affect unlived desires, feeding off of the affective allows emotion to have political implications surplus of our data exhaust. The cybernetic because affect arises through encounters ideal of systemic control extends itself into between a multiplicity of actors — it is neces- the virtual realm of the future through emo- sarily linked to social relations. Emotions are, tional artificial intelligence, where it steers in a sense, reductions of affect, crystalized our bodies through the inhuman logic of determinations of the “capacity to affect and capital. The consumer within the emotion be affected” that communicate something economy is subject to what Luciana Parisi about an affective encounter. The digitiza- calls an “alien reason” — or a computational tion of affects, as it is automated by affective form of automated reasoning that feeds off of computation, erases the distinction between contingency in order to produce new levels emotion and affect, constraining a users’ of determination — the machine not only capacity to act in accordance with their pre- knows, but brings our future actions into be- codified affective states, intentionally remov- ing (Parisi, “The Nanoengineering of Desire” ing any consideration of contextuality . 86). Affective technologies aim to digitize More recently, there has been a rapid and programmatically engage with human increase of tech companies engaging with affect. Affect is a force or intensity that — the relatively new science of ‘affective Melissa Gregg and Gregory Seigworth

85 APRJA Volume 8, Issue 1, 2019 explain — arises in the relations of bodies, whether those bodies are human, nonhuman, machinic, or conceptual (Gregg and Seigworth 1). It is an “ever-gathering accre- tion of force relations” (2) that structures the conditions for a body’s knowing and being in the world, through framing what is “felt to be real” (Massumi 54). Affect creates a “temporal contour” (Stern 62) that at once evades “received psychological categories” (Massumi 27) while also reorganizing the Figure 2: Paul Ekman’s Facial Action Coding System sensations and instincts felt in everyday life chart from FACS manual. (Bertelson and Murphie 148). It performs what Erich Hörl calls the “technoloecologi- computation — if a human limbic system is zation of sense,” where phenomenological replaced by extra-human sensors, she asks, experience is constrained by the affordances what types of affective communication might of the interface (Hörl 5). Affective technolo- a computer enable through its own ‘emergent gies programmatically enact a range of com- criteria’? putational techniques to enframe the norms Contrary to the opinion that computers of user experience, installing what Massumi should be completely ‘rational’ or logical calls a “politics of conformity” (57). machines, Picard proposes that an affective The scientific field of affective computa- dimension within computation might lead tion implements a number of techniques in computers to be better decision makers. order to standardize data and produce the Because, she claims, human cognitive intel- norms around which it operates. The science ligence is so bound to the material processes was coined by MIT Media Lab director and of the limbic system, a smart machine would scholar Rosalind Picard in a 1995 white pa- be able to naturally recognize and express per of the same name. “Affective computing,” emotions in their interaction with a human or “computing that relates to, arises from, or through taking in environmental sensory influences emotions,” makes use of Antonio information and responding appropriately. Damasio’s experimental neurobiological An affective machine might make use of high research to provide a framework for an detail cameras and other imaging sensors “emerging criteria” of emotions in computers in order to measure certain muscle expres- (Picard 1). It allies itself with Damasio’s “so- sions, generating a representation of an matic marker hypothesis,” which proposes emotion based on a model programmed into that emotions arise in the limbic system and the device. are later cognized, implying that affect is Affective computation makes psy- pre-conscious and pre-subjective (2). Higher chologist Paul Ekman’s ‘Facial Action Coding decision and learning processes require that System’ an executable program, creating a an affect is recognized, generalized, and mapping of the human face that corresponds labeled so that it might be used to prompt to a predetermined “atlas of emotions” decision, trigger action, and feed back into (Picard 5). Eckman created the FACS be- the homeostatic system of the cognitive tween 1972-78 based on his research on agent. Picard’s interest lies in applying this what he called “micro expressions,” claiming twofold physical-cognitive theory of affect to it to be the “first and only comprehensive tool

86 Rebecca Uliasz: ASSEMBLAGES OF DESIRE for objectively measuring facial movement. rights, such as in the case of the infamously Eckman proceeded to develop tools using the proposed Verizon media console that would FACS for clients as broad as the TSA, FBI, use microphones and sensors in order to CIA, health care providers, the Dalai Lama, constantly survey its ambient environment and the makers of Pixar’s Inside Out.[2] It for emotional cues and adjust television ads was appropriated by the Picard to create a accordingly. device that would learn and evolve over time Affective computers designed through to create a more fluid and ‘natural’ user ex- Affectiva extend the human perceptive ca- perience. Analytics mined from social media pacity of the user — where humans register platforms might be fitted to models derived each other’s emotions through verbal and from FACS that enable affective intelligence nonverbal communication, sensor technolo- to infer the way a users’ emotional disposi- gies enable a direct codified registration of tion effects their browsing habits. Thus, an a pre-conscious affective state through the ‘affect’ is produced in an affective computer sensory capture of microscopic, unseen and and constructed as a universal marker of unheard information —the machine knows an ‘emotion’ — an object of scientific truth, what we are ‘feeling’ better than we do. The a sort of ‘emojification’ of a human feeling. technologies employed by Affectiva explode The body of the user becomes collapsed out of the interface into the surrounding envi- into a one-dimensional data point. A smile, ronment in order to form a digitally expanded a click, a ‘like’, are all equivalent codified nervous system, of which human users exist representations that can be amassed in as the mere fleshy embodiment of the logics a database, assessed for patterns, made of capital. As a mixture of techniques for codi- into calculable models, and extrapolated to fying affective states, affective computation produce further information in accordance enables the creation of new markets through with the objectives of the programmer. the quantification of minute action, like the A 2015 New York Times feature on a flicker of a smirk, or the clenching of the jaw. spin-off company that came out of Picard’s If affective computation appears to be an MIT Lab, called Affectiva, makes explicit automated expansion of capital, a power ar- connection between so-called emotionally rangement that constrains the conditions for intelligent machines and the “glimmer of an subjectivation — the possibilities for a social emotion economy.” Affectiva, formed by subject to realize its “capacities to affect and Picard and partner Rana el Kaliouby, devel- be affected” — is it a worthwhile exercise ops custom software for clients that provides to speculate on whether such technologies facial and voice analysis in order to gather might play a role in alternative liberatory analytics that companies can use to under- regimes, explicitly outside of the domain of stand consumer emotional engagement. capital? Can affective technologies increase, This technology has been used by CBS, rather than restrain and regulate, human Millward Brown, AOL, IBM, and eBay among social and political capacities? Perhaps others in order to assess user engagement this becomes more complex than a critique and dynamically alter advertisement content of whom is deploying the techniques of af- through technologies that perform what is fective computation when it is claimed that called sentiment analysis (Khatchadourian). standardization and normalization of social Several of Affectiva’s client contracts have relations is a codified function of these tech- caught the interest of the public eye, raising nologies in and of themselves. concerns around the violation of privacy

87 APRJA Volume 8, Issue 1, 2019

was the question of what actually drives An ocean of affect these systems (Halpern 143). The existential questions — what is human will, and why do The development of affective technologies humans act irrationally? — become techno- follows the cybernetic aspiration of bring- logical questions of circuit design. How then ing together the mind and the machine. might we build a machine that adjusts for ‘Cybernetics’, or the “the scientific study of contingency in order to regulate output in the control and communication in the animal and name of efficiency? the machine,” was terminologically derived Luciana Parisi traces this bio-informatic from the Greek term for ‘governor’, ‘helms- phase of capitalism, where cybernetic theo- man’, or ‘steersman of a ship’ (Weiner 11). In ries are modeled on the biological processes the heyday of the Macy Conferences, cyber- within a body (Parisi and Goodman 136). netics obsessed over the idea of making a Norbert Weiner’s model, based on the idea machine that would mimic the human mind, of feedback or “the property of being able to turning to psychological models of human adjust future conduct by past performance,” intelligence provided by Freudian psychoa- takes a body as a system (whether it be nalysis. Part of the “dream of self-organizing a machine or living organism) that may systems and autopoietic intelligences be controlled and regulated via its inputs produced from the minute actions of small, and outputs (Parisi and Goodman 136). stupid, logic gates,” as Orit Halpern calls Bioinformatic capital subsumes a machinic the speculations of the Macy Conferences, assemblage through adjusting the inputs and outputs correspondingly, not though cutting off the affective flow of information, but by op- timizing the assemblage to perform the most efficiently. Moving beyond the cybernetic regime of interaction control, “affective capitalism” creates a future feedback effect — it is “a parasite on the feelings, movements, and becomings of bodies, tapping into their virtuality by investing preemptively in futurity” — exemplified by the finance technologies that typify contemporary global capitalism (164). Contemporary technocapitalism feeds forward into the future, determining the “not yet come” forms of cultural representation though bidding on the actions of social subjects. Our branded and media saturated environment has learned to capitalize on the affective states of users, defining social user- subjectivation as a consumer rather than a communal experience. In the age of affective computation, which computes human affect in order to Figure 2: Logic gates modelled by McCulloch and Pitts “Logical Calculus for Neural Activity.” produce evolving and complex affective data, Credit: McCulloch and Pitts. the questions of the Macy Conferences fall

88 Rebecca Uliasz: ASSEMBLAGES OF DESIRE short in addressing the social effects of ubiq- technologies can be typified as part of a uitous affective technologies. If it is claimed particular power arrangement, where rather that affective capitalism has subsumed all of than merely simulating or producing ‘affect’ social function, what chance could there be they are regarded as normalizing opera- for other types of subjectivation processes tors on the conditions of possibility within a that do not align with cybernetic paradigms structure of feeling. In other words, affective of bio-informatic control? What would it mean technologies operate on the sensorium of to attempt to reinstall what Massumi calls a everyday life in order to enforce normalized “politics of affect,” or what Bridget Bargetz constraints on the actions and decisions of names “a political grammar of feelings”, that user-subjects. Affective technologies are not emphasizes that shared affective dispositif is neutral, but rather, are prosthetic extensions necessarily a project of politics? The final part of logics of division and devaluation of hu- of this essay will engage affective technology man life for the benefit of capital interests. with contemporary theory on the politics of With an emphasis that affective technologies affect derived from the Spinozan definition come encoded with bias, we might begin to of affect — the power to “affect and be af- ask what kinds of knowledge they produce, fected” — in order to question whether affec- and if they might be deployed to produce tive technologies have potential to activate more equitable socio-technical relations. Is the capacities of a user-subject in modes that it possible to reclaim affective technologies increase their political and social agency. towards other machinic vectors of subjectiva- Massumi’s “politics of affect” consid- tion that do not simply service the ‘emotion ers power and affect together insomuch economy’? as they affect desire, or the potential of an A beneficial function of affective individual to become otherwise. For Deleuze technologies can be found in their ability to and Guattari, desire is an affirmative and strengthen human to human communication, productive force that mutates and trans- facilitating new means for social relations. forms matter, linking biological, technical, Consider the case of Paro the seal, where social and economic bodies in an energetic affective technologies are used to enable mechanic assemblage (Parisi 12). Desire is patients with dementia to more comfortably never given, but both realized through prac- relate to their environment and their care tice and affected by power relations. In other providers. In a similar vein, Picard’s original words, power arrangements delimit and re- interest in developing the tools to provide duce assemblages of desire within specific better education to children with autism is societal, political and historical regimes of based in the idea that these technologies representation and sensation. Affect, here, might allow educators and autistic students “acts in the nervous system not of persons to more clearly understand each other. In but of worlds” (Berlant 14) to frame what situations where affective technologies are Raymond Williams calls a “structure of deployed to intensify social relations, rather feeling” or shared historical organization of than alienate user-subjects through reducing culture and the elements contained within it them to statistically regulated consumers, (Williams 53). Affective technologies, when it serves to induce a different type of “mic- deployed by technocapitalism, claim to ropolitics” — what Guattari calls the partial produce affective capacity, expanding the techniques of power that produce beliefs, possibility of what a (user) body is and what desires, and sense of self on a social level it can do. In their actual deployment, affective — that remain open-ended and productive of

89 APRJA Volume 8, Issue 1, 2019 unexpected subjective capacities (Deleuze machines and humans alike, turning us into and Guattari 213). Invoking the notion of the addicts or avoidants, leaving us wanting ‘assemblage’, the complexity of all social re- more, feeling depressed, or changing the lations for Deleuze and Guattari always con- way we come to recognize emotion within tains potential for subjectivity to be remade others and ourselves. Recognizing the ways differently. What new types of usership could in which emotional artificial intelligence as a exist if affective technologies were used to technique of power is key in acknowledging create more transparent interfaces between the way that such technologies have the user and machine, or user and user? ability to automate the political agency of a Despite their possible use to increase user, and how they might activate this user connection, improve social communication, otherwise. and empower users, it should not be forgotten that affective computation is part of a cybernetic legacy that is specifically designed to operate on the future through prediction and regulation. Recalling the story of Monty the Penguin, we might see artificial intelligence become quite good at knowing what gives us that ‘heartwarming feeling’, enabling client companies of these technologies to adjust their products and campaigns accordingly. In a much more sinister vein, security cameras might draw conclusions about the affective states of targeted subjects to jump to unjust predictions about their future actions and intentions. Imagine a world in which hidden facial scanners serve as evidence and justifi- cation of discriminatory policing practices, for example. Imagine that the technology exists to make this possible, and imagine that its implementation is a matter for political and ethical guidelines, or lack thereof. Just as affective technologies learn from the data they capture from embodied subjects, they also have the ability to shape and transform the emotional states of users in an affective feedback loop. Consider studies on the linkages between social media and dopamine levels — tech companies are master manipulators of our biochemical reward pathways, with enormous insight into the forms of interaction, layouts, colors and designs that will get us hooked on that feel-good rush of interaction (Haynes). Amplification of universal affective codes shape the way we encounter

90 Rebecca Uliasz: ASSEMBLAGES OF DESIRE

Face++ Cognitive Services. 209. December Notes 2018.

[1] The term wascoined by Richard Yonck, Foucault, Michel. The Birth of Biopolitics. frequent blogger for Affectiva, and self New York: Palgrave Macmillan, 2008. proclaimed futurist in a 2017 online article titled “Welcome To The Emotion Economy, Goddard, Michael and Jussi Parikka. Where AI Responds To — And Predicts — “Unnatural Ecologies.” Fibreculture Journal: Your Feelings,” which first appeared on the Unnatural Ecologies 17 (2011): 1-5. website of the major media branding agency Fast Company. Gregg, Melissa, and Gregory J. Seigworth. The Affect Theory Reader. Durham: Duke [2] Found in Paul Eckman’s “Timeline of University Press, 2010. Achievements,” on the Paul Eckman Group website. Guattari, Félix. Molecular Revolution: Psychiatry and Politics. New York: Penguin, 1984.

Works cited Guattari, Félix. “Regimes, Pathways, Subjects.” Crary, Jonathan and Sanford Angerer, Marie-Luise, Bernd Bösel, Kwinter. Zone 6: Incorporations. New York : and Michaela Ott. Timing of Affect: Zone Books, 1992. Epistemologies, Aethetics, Politics. Zürich: Diaphanes, 2015. Hörl, Erich. General Ecology: The New Ecological Paradigm. London: Bloomsbury Bargetz, Brigette. “Mapping Affect: Academic, 2017. Challenges of (Un)timely Politics.” Timing of Affect. n.d. 289-302. Halpen, Orit. “The Neural Network: Temporality, Rationality, and Affect in Berlant, Lauren. Cruel Optimism. Durham: Cybernetics.” Timing of Affect. n.d. 119-144. Duke University Press, 2011. Hansen, Mark. “The Ontology of Media Deleuze, Gilles, and Félix Guattari. A Operations, or, Where is the Technics in Thousand Plateaus: Capitalism and Cultural Techniques?” ZKM Zeitschrift Schizophrenia. Minneapolis: University of für Medien - und Kulturforschung (2017): Minnesota Press, 1987. 169-185.

Deleuze, Gilles. “Desire and Pleasure.” Haynes, Trevor. “Dopamine, Smartphones Lapoujade, David. Two Regimes of & You: A Battle for Your Time.” 1 May 2018. Madness: Texts and Interviews 1975-1992. Harvard University, The Graduate School of New York: Semiotext(e), 2006. 122-134. Arts and Sciences. March 2019.

Ekman, Paul. Facial Action Coding System Kaliouby, Rana el. “Why Monty Waddled into Manual. 1987. Document. First Place.” Affectiva. 19 December 2014. affectiva.com.

91 APRJA Volume 8, Issue 1, 2019

Khatchadourian, Raffi. “We Know How You Panayotov, Stanimir. “To Engineer the Time Feel.” 19 January 2015. The New Yorker. by Other Means: Interview with Luciana November 2018. https://www.newyorker. Parisi.” 17 August 2016. Figure/Ground. com/magazine/2015/01/19/know-feel. November 2018. https://figureground.org/ interview-with-luciana-parisi/. Lazarrato, Maurizio. Signs and Machines: Capitalism and the Production of Parisi, Luciana. Abstract Sex: Philosophy, Subjectivity. Semiotext(e)/ Foreign Agents , Bio-technology and the Mutations of Desire. 2014. London: Continuum, 2004.

Leprince-Ringuet, Daphne. “Social Media Parisi, Luciana and Steve Goodman. has Totally Warped the Way You Think “Mnemonic Control.” Clough, Patricia and About Happiness.” Wired. 30 January 2019. Craig Willse. Beyond Biopolitics: Essays on wired.co.uk. the Governance of Life and Death. Durham : Duke University Press, 2011. 163-176. Lovink, Geert. “Sad by Design.” Eurozine. 10 January 2019. eurozine.com. Parisi, Luciana. “Computational Logic and Ecological Rationality.” Hörl, Erich and Massumi, Brian. Politics of Affect. James Burton, eds., General Ecology: Cambridge: Polity Press, 2015. The New Ecological Paradigm. London: Bloomsbury, 2017. 75-100. Massumi, Brian. “The Market In Wonderland: Complexifying the Subject of —. “Reprogramming Decisionism.” e-flux Interest.” Timing of Affect. n.d. 321-338. journal #85. 2017.

McCulloch, Warren S. and Walter Pitts. “A Parisi, Luciana. “The Nanoengineering Logics Calculus of the Ideas Immanent in of Desire.” Giffney, Noreen and Myra J Nervous Activity.” Bulletin of Mathematical Hird. Queering the Non/Human. Ashgate: Biology 1990: 99-115. Aldershot, 2008. 283-311.

McCulloch, Warren S. Embodiments of PARO Robots USA. PARO Therapeutic Mind. Cambridge: MIT Press, 1965. Robot. 2014. March 2019.

McLuhan, Marshall. Understanding Media: Peters, John Durham. The Marvellous The Extensions of Man. Cambridge: MIT Clouds: Towards a Philosophy of Elemental Press, 1994. Media. Chicago: University of Chicago Press , 2015. McStay, Andrew. Emotional AI: The Rise of Empathetic Media. Los Angeles : SAGE, Peterson, Hayley. “A British Supermarket 2018. Chain Is Installing ‘Creepy’ Face Scanning Cameras to Track Consumers.” Business Microsoft Azure. Cognitive Services Face Insider 5 November 2013. businessinsider. API. 2019. com.

92 Rebecca Uliasz: ASSEMBLAGES OF DESIRE

Picard, Rosalind. Affective Computing. MIT Media Laboratory Perceptual Computing Section Technical Report No. 321. Cambridge: MIT Media Laboratory, 1995.

—. Affective Computing. Cambridge: MIT Press, 2000.

Robinson, Martin. “Toy Penguin in New John Lewis Christmas Advert is a Hit!” Daily Mail 6 November 2014. dailymail.co.uk.

Sensum. “Chose Your Sensors for Empathetic AI.” 5 December 2018. sensum. co.

Stern, Daniel N. The Present Moment in Psychotherapy in Everyday Life. New York: Norton and Company, 2004.

Weizenbaum, Joseph. “ELIZA - A Computer Program for the Study of Natural Language Communication between Man and Machine.” Communications of the Association for Computing Machinery 9 1966: 36-45.

Williams, Raymond. Marxism and Literature. Oxford: Oxford University Press, 1977.

Yonck, Richard. “Future Reflections on the Developing Emotion Economy.” Affectiva. 2 February 2017. affectiva.com.

Zuboff, Shoshana. The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. PublicAffairs, 2019.

FEELING GENERATORS Carman Ng

AFFECTING REALITY: INTERSECTING GAMES, TRAUMA, AND IMAGINARIES

Abstract

This essay examines how digital games shape human affective repertoires and envisioned dynamics with nonhuman agents such as robots. Entanglements among humans, machines, and technologies impact essential issues in the historical present: from surveillance, climate change, cultural heritage, art, to the elicitation, habituation, and capturing of feelings. Approaching digital games as frontiers of such entanglements, this essay expounds dynamics among gameplay, affects, and gamic materiality through a case analysis of Nevermind (Flying Mollusk), a trauma-themed independent psychological thriller game with affect-sensing technologies. Discussion explores how the game can generatively engage with lived experiences and discourses of grief and trauma; and the relationality among individuals, structures of feelings, and stigmatization. Anchoring the essay is an argument that digital games represent and operate with fundamental tenets of posthumanism, communicating meaning across affective and semiotic dimensions, bodies, machines, and sociocultural contexts. This essay emerged from an ongoing project on affective semiotics and social impact game design, in connection with a transnational research project on human-robot interaction supported by the European Research Council.

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Carman Ng: AFFECTING REALITY

dynamics contributes to work on several Introduction fronts. Affect research in the past decade focused on distinguishing the phenomenon Pivotal to comprehending “structures of from emotions, problematizing the longstand- feeling” in Raymond Williams’ theorization ing emphasis on individual intentionality, is an intellectual openness toward exploring cognition, and categorical emotions. Recent dynamic experiences, expressions, and so- accounts emphasize the social relationality cial forms consistently in flux, in the present, of affects (von Scheve), cementing the focus and immune to claims of alleged finality. on affectivity as processual, transpersonal, Structures of feeling concern interanimating socioculturally constituted, and emergent dynamics between lived experiences and across bodies, including technological cultural expressions: how the latter shape systems. and express emergent ways of being in Yet, it remains unclear how the affect- the world; and develop understanding of inspired focus on social interactions may the emergent nature of lived experiences. constructively engage with trauma, which Digital games encapsulate such dynam- straddles individual and social realms. ics on both micro and macro scales: in the Recovering from trauma entails awareness moment-to-moment process of gameplay; of one’s emotions, triggering events, coping and the medium’s interactions within par- mechanisms, and available sources of sup- ticular technocultural contexts and media port, healthcare, and intervention. Difficulties ecologies, all pertinent to the materialization war veterans face in overcoming post- of artistic and design practices, transmedial traumatic stress disorder (PTSD) illustrate relations, and surrounding, media-shaping, tensions within such a matrix; challenges social discourses. Thus, it is productive to confronting marginalized social groups are explore machine feeling through affective likewise indivisible from such factors as race, digital games, which detect and dynamically gender, social classes and the cross-gener- respond to players’ affective states. ation ramifications.[1] Trauma-themed digital Intersecting affects and emerging tech- games thus provide a gateway to critically nologies, this essay emphasizes the shaping engage with the often unspoken aspects of of affective repertoire. The concept, as I traumatic experiences, as well as systemic propose, explores spectra of human capaci- factors that enforce contemporary regimes of ties to feel, express, and regulate feelings, silence and stigmatization, by interweaving informed and potentially expanded, refined, design, gameplay, narrative, technology, and or enframed by technological facets of lived complex affects. experiences. Affective repertoire stems from On affects, the concept of repertoire perceiving this malleable range of affects is under-theorized. This may be due to the and associated reactions, which may then seeming incongruence in pairing affects and support individuals to consider and deha- repertoire. The former has been characterized bituate certain responses for future affective as precognitive and nonrepresentational; the encounters. Oriented toward posthumanism, latter, culminated from learning and curation, this conceptual tool aims to untangle how concerns competence, contexts of use, and technological designs prevalent in specific components of identity, as in the example mediated encounters, environments, and of linguistic repertoire.[2] Conceptualizing sociohistorical contexts, incubate feelings affective repertoire is an attempt to ques- and bodily intensities. Knowledge of such tion the assumed incompatibility, inviting

97 APRJA Volume 8, Issue 1, 2019 inquiries into agency. Proceeding from inten- argument that, with cogitative design and tion to actions capable in effecting change, narrative, the medium can harness affective agency is essentially constrained by social computing technologies for enhanced game- factors beyond individual control. That said, play and potential intervention. Finally, the trauma does not necessarily eclipse agency, essay affirms a future-oriented perspective, when one mindfully engages with feelings, positing a tripartite research methodology to thoughts, and aspects of any experiences engage digital games as incubators of aes- that might be uncertain, destabilizing, or thetic potential, complex affects, and visions otherwise habituated. As Shaun Gallagher of human-technology interaction. observes, the body crucially constitutes the mind, meaning, and communication. A critical orientation toward agency, embodied cognition, and trauma thus builds in parallel with Posthumanism | digital affective repertoire and resilience. As efforts games | affect in articulating ontologies of affects and emotions expand, exploring bodily sensations Images of human-nonhuman relationships as indexical of affects, emotions, and mobi- percolate the mediasphere. Their eclecticism lization of behaviors in digital games brings manifests across cinema (e.g. Blade Runner complementarity. Identifying patterns among 2049, Ex Machina, The Matrix), videogames game design features and activated affects (e.g. Metal Gear Solid, Deus Ex), television enable a rethink about the experimental role (e.g. Westworld, Humans), and experimental of digital games and their aesthetic, techno- art that interfaces the body with prosthet- logical, and sociopolitical importance across ics, networked systems, and biotechnology a l l e g e d c o n fi n e s o f t h e i n t e l l e c t a n d f e e l i n g s . (e.g. Stelarc). Understanding the diversity This essay presents excerpted analyses of such visions, technological innovations, from an ongoing research project on digital and cultural production carries importance, games and human-robot interaction (HRI), especially upon our understanding of the as illustrative of changing social realities and “nonhuman turn.” Conceptualized in the contemporary concerns. These range from 2015 eponymous book (Grusin, vii), the porousness between real and virtual worlds nonhuman turn involves intersections among to ethical quandaries regarding artificial intel- human and nonhuman entities (e.g. bodies, ligence. The driving premise is that digital technologies) in tackling issues of the 21st games are shifting our affective capacities, century, including terrorism and climate sci- eliciting various affectivities while informing ence. As intellectual inquiry, it decenters the our understanding of the posthuman condi- unified human subject through the notion of tion. Readers first find a contextualizing the nonhuman, finding resonance in affect overview of posthumanism, digital games, theory, animal studies, cognitive sciences, and current developments in affect-centered and new media theory, to name a few rel- game analyses. The section outlines the evant fields of study. need to articulate the meaning-making logic Similarly re-assessing the symbiotic of digital games as prominent cultural forms relationships among humans, technologies, and posthumanizing artifacts of procedural and nature, posthumanism challenges social and multimodal complexities. Then, readers categories and dichotomies with technosci- find analysis of an independent psychologi- ence, inviting philosophical discussion on how cal thriller game Nevermind, in support of the

98 Carman Ng: AFFECTING REALITY technology fundamentally constitutes the hu- and such technologies as virtual reality man condition (Haraway; Nayar; Hauskeller, (VR) interfaces, increasingly applied in non- Philbeck, and Carbonell 3). The archetype of entertainment contexts, including education cyborg, at once organic and mechanic, has (Gee, de Freitas & Maharg) and military inspired various schools of thought with the training.[3] It matters to critically engage with potential to steer human development on the the medium in design, gameplay, analyses, scale of civilization. A telling example is the and pedagogy. Digital game criticism broadly techno-utopian discourse of transhumanism, involves three trajectories: formalist, which which embraces technological augmentation, explores the aesthetics and form of games; human-machine singularity, and freedom af- social, which considers the medium in rela- forded by “anthropo-technologies” (Kurzweil; tion to cultures and histories; and integrated Sloterdijk). Carrying a more complex outlook approaches, which combine practice and than the anti-humanism in transhuman- design (Jagoda 213-215). This extends into ism and the work of Haraway, posthuman an expanding network of research areas, humanity centers on creating sustainable methods, and foci, including philosophy, digi- human-nonhuman futures (Braidotti 55-104). tal humanities, media and cultural studies, Discussions of posthumanism and platform studies, ethnography, psychology, digital games began from ideas such as and political economy. Established in 2001, narrative, representation, and player-avatar game studies has observed the development relations. The scope has since expanded of concepts and analytical frameworks on to how unconventional game forms and capacities of games to foster “critical play” automatic gameplay challenge notions of (Flanagan 1-17) and function as, for instance, subject and object; all concerned with the “allegorithm” (Galloway 83) and ethical daily entanglements of humans, technology, systems (Sicart). Despite insufficient discus- increasing automation, and environments sions of posthumanism and games beyond (Fizek et al.). In this context, I propose ap- representation, the expanded approaches proaching digital games as posthumanizing and concepts indicate a growing field of aca- encounters. During play, meaning unfolds demic inquiry targeting a fuller understanding across technical materiality, bodies, real and of games and their social influence. virtual worlds (Keogh 14-17; Leino), exempli- A vital aspect to digital game play and fying the distributed and emergent charac- research that is gaining traction is affect. In teristics that define posthuman subjectivities Playing with Feelings, Aubrey Anable argues (Hayles, How We Became Posthuman; that digital games construe a most significant “Reconfiguring the Posthuman”). From this art form of the 21st century, allowing players baseline, it is feasible to consider games to rehearse specific affective states benefi- beyond representational and technological cial for contemporary life (e.g. how to relate terms, exploring their influence on individuals to work and failure). Her approach highlights (in subjectivities, feelings, and worldviews), the cultural embeddedness and gendering of societies (shifts in discourses and practices media artifacts, attending to game types and surrounding games), and how such knowl- engagement often overshadowed by main- edge informs ways of designing and critically stream games, including indie games, art engaging with new media. Digital games games, and casual games. In her argument, are among the fastest growing media with affect-mediating processes among players, ubiquitous presence, economic viability, devices, machines, and code — what is char- congruent progress with affective computing acterized as posthumanizing in this essay

99 APRJA Volume 8, Issue 1, 2019

— form a contemporary structure of feeling. and artificial intelligence. This knowledge, It intersects with, for instance, diminishing I suggest, invites human interlocutors to work-play boundaries, where types of casual ponder ways to critically design, engage, and games can mitigate what contemporary work research emerging technologies. culture may lack (involvement, identifiable outcomes, pleasure); and yet commodify affect through in-game microtransactions, resembling capitalist labor. Whereas, games Unravelling gamic that foster frustration may guide players in materiality understanding and handling failures. Numerous inquiries examine such Digital gameplay experiences are gestalts multi-level interactions among games, (combinations of parts) involving procedural, players, and changes in social systems. In semiotic, and algorithmic elements. A mere support, I propose posthumanizing affective dissection of these units does not capture semiotic operations as an orientation to build how games mean and elicit feelings. Yet, methods and vocabulary that justly examine this dissection is essential to developing a meaning-making in games. It may likewise theoretical language to understand interac- complement research on machine feeling. tions among affects and meaning at play. Discourses on machine learning fuse with For this purpose, procedural rhetoric and debates about artificial intelligence and robot socio-semiotic multimodality present pro- ethics, foregrounding concerns like social ductive perspectives. Procedural rhetoric effects of algorithmic biases (O’Neil), as examines how computational media convey well as expanded moral and legal respon- persuasive messages through mechanics sibilities when autonomous machines factor and simulation (Bogost 5, 14, 28-29). Rooted into romantic relationships, healthcare, and in social semiotic theory that views culture warfare (Lin, Jenkins, and Abney). Against as sets of inter-related semiotic systems this backdrop, my work on games and HRI (Halliday; Halliday and Hasan), multimodality explore aspects of machine learning that are has investigated how media artifacts, experi- perhaps overlooked. In Nevermind, affective ences, and interactions as sign-complexes computing enhances dynamic qualities of communicate through multiple resources gameplay and motivates player reflection termed modes (e.g. visuals, language, of the narrative and their management of sound, music, haptics); the usage of which emotions. Ethical dimensions of machine is regularized by communities of sign users, learning, such as collection of actual interac- sociocultural contexts, and therefore underly- tion data for training mass-produced social ing political, economic, and ideological forces robots, are explored in the collaborative HRI (Kress; Jewitt, Bezemer, and O’Halloran). research phase. Lastly, combining corpora Multimodal research has explored, for and annotation tools with automated analysis instance, film (Bateman and Schmidt), components forms a trend in empirical multi- interactions (Norris), comics, experimental modal research, paving way for larger-scale literature (Gibbons), art (O’Toole), math- studies of games. These three facets, from ematical discourse (O’Halloran), movement games as designed experiences, ethics of in space (McMurtrie), and digital platforms data collection and use for machine learning, (Jewitt). That said, among studies approach- to changing research methods, explore on ing digital games as discourse (e.g. Aarseth; different scales the rising influence of games

100 Carman Ng: AFFECTING REALITY

Ensslin; Gee, Unified Discourse Analysis), the foundational meaning-making mecha- multimodal research remains nascent. nisms, while clarifying common interpretive 2019 saw the publication of two books that ambiguity (e.g. “medium” as intermediary in examine games as persuasive and ludonar- communication versus “(mass) media” as rative artifacts from a multimodal perspective understood from an institutional lens, 103). (Hawreliak; Toh); and the first collection on Secondly, discourse semantics contextual- videogame discourse and linguistics (Ensslin izes and outlines the range of sensible inter- and Balteiro) appeared in May 2019. Weimin pretive possibilities for particular multimodal Toh presents a four-level ludonarrative combinations (116-121). Thirdly, “canvas” model that, pursuing a similar trajectory as introduces an analytical perspective and tool my earlier work to map out meaning-making to delineate intersections between the sub- units and dynamics in games, identifies con- ject and means of communication. It refers to nections between, for example, gameplay any bearers of perceivable and interpretable and narrative as “modules”: interacting material regularities, be they analog, digital, communicative systems realized by various unfolded physically in time, and resulted from modes and “elements,” such as game rules technological processes (86-88).[5] This and mechanics (34-47). Jason Hawreliak focus on communicative form and intent con- proposes conceiving procedurality itself as nects with prior discussions of the “transmis- a semiotic mode to highlight how games sion” and “semiotic” components of media communicate through processes, not solely (Ryan 1-40); and enables an informed fixing representation (80-94). While this attempt of of analytic focus, by systematically “slicing” reframe reinforces complementarity between each communicative situation into various game studies and multimodality, it injects canvases and sub-canvases.[6] unwarranted ambiguity into core concepts Dependent on genre, mode of game- (e.g. mode and affordance) and calls for play, and context, game analyses involve empirical analyses, to prevent collapsing a range of canvases and analytical units fundamental strata in theorizing multimodal (e.g. narrative, events, and mechanics). The meaning-making.[4] gameplay interface and player enactment My interest to integrate multimodal- possibilities form my material, analytical foci; ity with digital game criticism lies in its and on the social dimension, connected phe- empirical support to systematically tease nomena and civic discourses, for instance, out the intersemiotic relations and interplay how digital games are embedded in and of elements in digital games as affective may problematize the military-entertainment processes. Stressing meaning-making as complex. To make data analyzable, tran- sociohistorically-situated media practices, an scription is a necessary first step. It involves empirical multimodal approach offers three transcoding complex data into an inspect- main insights. They are the re-construal of able, manipulable form, commonly as tabula- the notion of media, centrality of discourse tion intersecting analytical units in rows and semantics, and the analytical concept of the information conveyed in columns (Baldry “canvases” (Bateman, Wildfeuer & Hippala). and Thibault), such as shot analyses in film Firstly, their theorization re-considers media studies. as historically stabilized sites that use se- Figure 1 shows how, previously, I used lected semiotic modes according to commu- analytical software to annotate and visualize nicative purposes. This emphasizes the need findings from digital gameplay data, in an for interdisciplinary import into examining attempt to identify the immersion-shaping

101 APRJA Volume 8, Issue 1, 2019

Figure 1: Example of multimodal annotation of digital game with analytical software “Multimodal Analysis Video” (Ng 20-22). effects co-realized by gameplay mechanics biofeedback technology with gameplay, and audiovisual aesthetics. Such fine- centered on psychological trauma in con- grained and structured analyses aim to content and puzzle-solving in form. Nevermind solidate empirical research of complex multi- engages with three significant strands that modal phenomena. In addition to a range of have emerged from the development trajec- relatively well-recognized software, such as tory and discourses of affective technologies. ELAN, ATLAS.ti, and NVivo, computational Firstly, it highlights a changing focus initiated approaches to multimodality have begun by affective computing, since conceived by to incorporate machine learning and deep Rosalind Picard in 1997, namely, a re-frame learning.[7] It is foreseeable that algorithms of emotions from static, universal human and automatic processing may support hu- faculties to dynamic processes that unfold man annotation and thus empirical research moment to moment in gameplay. Secondly, of data across levels of multimodal complex- it evidences a contemporary pursuit for ity, scale, and methods. immersion-based innovations, such as VR experiences. These strands in turn dialogue with possibilities and (ethical) questions regarding serious applications of games, Playing with trauma machine learning, as well as connections among bodies, feelings, and technologies. One advancement in digital game design is My focus is on how digital games afford the application of affect-sensing technolo- opportunities to engage the often silenced gies that comprehend and respond to play- and stigmatized aspects of trauma, in both ers’ emotions. To explore the multilateral discourse and lived experiences. meaning-making processes in digital games, Memory is at the core of Nevermind. I turn to Nevermind, a game designed with Narratively motivated by the recovery and application potential in public health contexts, organization of traumatic memories, the such as therapy, with clinical trials in planning gameplay involves exploring the psyche (Flying Mollusk, “Therapeutic Applications”). of psychological trauma patients, puzzle- [8] The independent thriller game integrates solving, discovering, and sorting memory

102 Carman Ng: AFFECTING REALITY photographs into a coherent account of a world and gameplay difficulty adapt to the traumatic event. Patients’ subconsciousness player’s states of stress (fig. 3), Nevermind are often portrayed aesthetically as twisted, amplifies how digital games constitute disturbing, and surreal (fig. 2). Five playable posthuman subjectivities. On the one hand, cases have been released, tackling topics it complicates the layering of human affects from child abuse, post-traumatic stress dis- and machinic cognition in micro-gameplay order (PTSD), to LGBTQ identity. Gameplay moments. This opens up a common notion lasts on average four hours and includes six of gameplay as input-output feedback loop, narrative stages: orientation, development, to consider ways that the medium organizes disequilibrium, crisis, climax, and denoue- affectivity by dynamically intersecting player ment.[9] Here, I zone in on the initiating case action, design, and levels of algorithm-based Client #251 to discuss: (a) how meaning thinking (e.g. inferring player emotion by con- unfolds across mechanics, plot, and multi- trasting facial expression data). On the other modal combinations during gameplay; and hand, it gestures toward increased attention (b) intersections between digital games and to games as serious applications to address therapeutic interventions, leading into the current issues (e.g. psychological wellbeing final discussion on investigating relations be- and healthcare). tween affective experiences and procedural- semiotic patterns in gameplay.

Figure 3: Biofeedback technology in Nevermind. © Flying Mollusk.

Client #251 explores suicide and the witnessing of traumatic events. Analyses identify three connecting motifs. The first mo- Figure 2: The subconscious landscape of Client #251. tif concerns financial and marriage difficulties © Flying Mollusk. between the female client’s parents; sec- Experiencing Nevermind involves the ondly, the patient’s witnessing of a traumatic automated perception of the machine. The event and resulted guilt and self-blame; and 2016 VR edition uses biofeedback technol- thirdly, her father’s suicide. I refer to these ogy to detect players’ physiological and emo- three motifs as M, W, and G respectively. tional states, eye movement, and modulates From the opening cutscene, players learn gameplay difficulty accordingly. Physiological that the patient was informed by her mother biofeedback tracks changes in, for instance, since childhood that her father had died in heart rate and pulse as indicators of stress, a car accident. However, gameplay and the anxiety, fear, and psychological arousal. de-briefing cutscene (i.e. pre-rendered cin- Emotional biofeedback concerns detecting ematic sequence) reveal that her father shot players’ facial expressions.[10] As the game himself in her presence.

103 APRJA Volume 8, Issue 1, 2019

Figure 4. Cohesive connections in Nevermind: Client #251 (excerpt).

To understand how meaning traverses development, players find clues to the par- gameplay mechanics, narrative motifs, and ents’ failing marriage. Washed-out marriage audiovisual representations, a useful means and family photos (M) and a safe-unlocking is to map out their cohesive connections puzzle (combination: milk, gun, and sor- (fig. 4).[11] Situated in the middle is the row) (M1) imply tenuous family dynamics gameplay-grounding emotion and triggering and violence, in contrast with a subsequent event: the patient’s guilt and having spilt milk memory photo that presents a false, ideal- as a child. These horizontally connect to the ized marriage (M2, marked with a dashed significant narrative motifs identified (M, W, line). The patient’s guilt from milk-spilling is G). Vertically, as gameplay progresses, indi- likewise introduced, first as an accusatory vidual motifs accrue significance by referen- message “You Spilled,” written in red, shaky, tiality, recursive representations, gameplay handwriting-resembling font on the mirror, mechanics, and contradictory information. that cues the correct safe combination. The Such representations may take the form of motif then recurs as milk cartons (printed player-maneuverable and contextual objects, with guilt-centered texts and nutritional la- diegetic sounds, and sound effects that con- bels) and a memory photo, at development stitute the game world.[12] For example, a (beginning at 00:09:13 and 00:19:17) and sound of gunshot coupled with fade-out for a disequilibrium (00:13:35) respectively. Such gameplay segment at the parents’ bedroom recurrence forms a discourse semantics that (00:06:15) multimodally signals a transition cues players the sensible interpretations from orientation to development; simultane- and co-occurs with a build-up of affects and ously suggesting gun violence in the death of emotions. While the spilling of milk denotes the patient’s father.[13] a micro-level, aggravating incident in the pa- The case first connects the motifs of tient’s childhood, it connotes an overarching parental issues, milk-spilling, and guilt. In sensitivity of guilt, self-blame, and anger.

104 Carman Ng: AFFECTING REALITY

Gameplay mechanics gradually re- stages, creating a prosody in both content contextualize this guilt-grounded sensitivity, and affects (e.g. calm, anxiety, disturbance, by uncovering the symbolic dimension of a and shock). Puzzle-solving concentrates seemingly mundane yet narratively moti- in development as scaffolds to access the vated mechanic tied to the motif of witness- patient’s buried memories; disequilibrium, ing: teacups arranging. First appeared as crisis, and climax then focus on navigating contextual objects in orientation, teacups mazy spaces, often coupled with disturb- take on increased puzzle-solving poten- ing audiovisual aesthetics. In presenting a tials in development and denouement. To funeral service in the patient’s childhood retrieve the final memory photo, the player home, denouement forms a poignant stage places teacups to guide water into a burial intersecting the motifs of memory, inner and ground (W2). Eyes on the three water mills outer life (the emotional tension of which is peel open, decreasing the violence and fear described in the pre-gameplay cutscene), formerly associated with seeing/witnessing, witnessing, and potential closure, with the respectively portrayed as aggressive funeral recurrent puzzle-solving mechanic of teacup attendees and female faces with tears and arrangement. cavernous mouths who visually follow player This initial analysis suggests a reflec- movements (W and W1). It connotes wak- tive quality between the gameplay mechan- ing up to the truth, visualized in the graphic ics of Nevermind and practices relevant to memory photo (G1) and the monochromatic mood management, trauma processing, and visual of suicide in the de-briefing cutscene psychotherapy. In particular, the gameplay (G2). processes of navigating the clients’ subcon- Collectively, the motifs of Client #251, scious, collecting, and organizing memory along with the audiovisual aesthetics, photos share a similar focus with therapeutic gameplay mechanics, and space, create an practices, in building awareness of and vo- atmosphere that oscillates among suspense, cabulary to process emotions. As expanding surrealism, turmoil, and calm. From the research literature postulates, digital games, choice of landscape, gameplay mechanics such as the apocalyptic The Walking Dead (e.g. jigsaw puzzles and teacup arrange- (Telltale), invite critical dialogues between ment), to various surreal representations, game studies and trauma studies to explore they orient to particular aspects and stages trauma in games via “interreactivity,” em- of the narrative, which then structure the pathy, and complicity (Smethurst & Craps; gameplay experience. Similarly, shifts in Smethurst). Videogame series such as gameplay environment modulate the rhythm the Japanese, stealth-based Metal Gear and narrative levels in gameplay. As Michael Solid (Kojima) and Max Payne (Remedy Nitsche illustrates, game spaces evoke Entertainment/Rockstar Games) illustrate six narratives by inviting player perception, prevalent motifs of PTSD, including trauma- interaction, and interpretation. The patient’s identity relations (Bumbalough and Henze subconscious landscape is comprised of 15-33).[14] Nonetheless, the focus of these private, public, and fantastical spaces, from studies remains on representation. A major idyllic gardens, site of traffic accident, to bi- value in examining Nevermind, I contend, zarre and distressing locations not conform- lies in how the game design narratively and ing to real-world logics. Analyses observe procedurally aligns with specific principles a concomitant complication in spaces and and techniques in trauma therapy and resil- gameplay mechanics across the narrative ience-building. These include acknowledging

105 APRJA Volume 8, Issue 1, 2019 events, mindfulness, affect regulation, integral to such an emergence. Through a and approaching resilience as continuum case study of Nevermind, the essay inter- (Southwick et al), aligning with my proposed sects game studies and multimodality to concept affective repertoire. In addition to examine a key phenomenon in this cultural highlighting the expanding applicational moment: affective digital games. Analysis scope of digital games, Nevermind gestures illustrates how cohering narrative motifs, toward a convergence of social phenomena gameplay mechanics, audiovisual aesthet- and discourses pertinent to the technological ics, and affect-sensing technologies enables capture and structuring of feelings. It embod- a form of metaphoric play akin to stages of ies four co-emerging entities or dimensions: processing trauma. Hence, one aim of the the medium of digital games; the developing essay has been to further understanding of technology of machine learning; fluctuat- digital games as complex systems involving ing affects; and shifting regimes on trauma affects, multimodal semiotics, proceduralism, and mental health, toward increasing de- and contexts; with the potential to strengthen stigmatization. The fact that digital games one’s affective repertoire for engaging themselves are emerging technologies, in with complex affects and contemporary my view, renders them specifically suitable to challenges. engage visions of human-technology interac- In the current mediasphere, two trajec- tion, new sociocultural norms, and practices tories seem to be forming. On expression, as a result of machine learning. mutual influences among designs of digital games, interfaces, and virtual/augmented reality technologies (e.g. Meta AR headsets) signal intersections among media and visions Future directions of future human-technology interactions and experiences. On content, representations of The digital present is affective, unfold- human-nonhuman dynamics in mainstream ing, and propelled by human-nonhuman games have expanded, suggesting a shift relationships. In this essay, I have explored from war-focused posthumanism (e.g. MGS digital games as posthumanizing encounters games) to include portrayals of human-robot

Figure 5: Research methodology. 106 Carman Ng: AFFECTING REALITY affective bonds and conflicts (e.g. Detroit: difficult feelings and issues. For design and Become Human). Such robotic imaginaries technology-focused research, examining form the research focus connecting my work (dis)connections among meaning-making and the collaborative ERC-supported project units in games, intended affective response, on HRI and emerging technologies entitled as well as the operative and reactive ac- Emotional Machines: The Technological curacy of affective computing software may Transformations of Intimacy in Japan contribute to developing affective artificial (EMTECH, 2017-2022). Our inquiries aim to intelligence in digital games and media. articulate dialectics between imaginaries and The still emerging phenomenon and realizations of human-robot relationships, cultural narratives of affective interactions specifically processes and effects of forming with robots and digital technologies carry affective bonds with robots, digital devices, ramifications across automation, social in- and networked technologies. timacies, and war. Confronting the com- Digital games evidence a multidimen- plexities involved demand an engagement sional emergence: in media, affects, human- with diverse sociopolitical issues, robust technology interaction, social discourses, research, and designs that explore beyond and research methodologies. Always user-friendly, technological solutions. In the fluctuating, affects are nonconscious bodily continuing conversations on (post)human- intensities that underlie thought, behaviors, nonhuman developments, critical play and and yet elude human observation (Massumi; research will inform our participation with Stewart). They manifest physiologically perspective, intention, accountability, and (e.g. micro facial expressions, pulse, skin openness to engage the associated, diverse, conductance) and as a central constituent and potentially difficult feelings. to meaning-making and behavior, they are yet amply examined from a multimodal lens. Thus, I propose to integrate affect theory, corpus-based multimodal game analyses, with players’ biophysical and interview data to empirically expand knowledge on the medium, toward game designs that foster empathy and mental health (fig. 5). The triangulated data is expected to complement ethnographic findings from EMTECH on interactions with digital technologies and robots in homes and public spaces. As importantly, this addresses a methodological need to incorporate discourse analyses for a textured, discursive, view of affects as embodied practices entangled with contexts and social relations (Wetherell). Such data triangulation may also support future research on affective repertoire. Through episodic engagement with affective digital games, researchers and participants may evaluate if, or how, digital gameplay modulates ways of encountering

107 APRJA Volume 8, Issue 1, 2019

[2] Emotional regime, a concept in the Notes history of emotions literature proposed by historian William Reddy, concerns [1] A developing and debated field called dominant forms of emotional expression epigenetics explores how changes in and thought in specific periods and environment, effects of stress and trauma cultural contexts (124-129). The notion is could change the expression of genes, with under-explored in both studies of emotion transgenerational effects. Approaching the and affect. See Reddy, William M. The nature/nurture link, studies are exploring Navigation of Feeling: A Framework for the what epigenetic mechanisms (e.g. changes History of Emotions. Cambridge University in RNA molecules and DNA methylation) Press, 2001. Repertoire, as a linguistic might signal the “inheritance” of trauma, phenomenon, refers to flexible and dynamic if any. For popular versions debating the use of language oriented toward social plausibility of trauma- transmission, see action and contexts. Margaret Wetherell’s Henriques, Martha. 26 March 2019, “Can approach toward affective practice and the Legacy of Trauma be Passed Down the description of repertoires (135, 138) would Generations?” BBC Future, http://www.bbc. prove helpful. See Wetherell, Margaret. com/future/story/20190326-what-is- epige- Affect and Emotion: A New Social Science netics; and Carey, Benedict. 10 December Understanding. Sage, 2012. For information 2018, “Can We Really Inherit Trauma?” on conducting discourse analysis via the The New York Times, https://www.nytimes. approaches of critical discourse analysis com/2018/12/10/health/mind-epigenetics- and discursive psychology, see Phillips, L. genes.html. and Jørgensen, M. W. Discourse Analysis Also, readers can find an overview of as Theory and Method. Sage, 2004. epigenetics research and post-traumatic stress disorder in Zannas, Anthony S.; [3] The Entertainment Software Association Provençal, Nadine; and Binder, Elisabeth reported in 2018 that the digital game B. “Epigenetics of Posttraumatic Stress industry contribute a $36 billion consumer Disorder: Current Evidence, Challenges, spend in the United States (2017) and a and Future Directions.” Biological GDP of above $11.7 billion. Sources on the Psychiatry, 78 (5), 2014, pp. 327-335, htt- military-entertainment complex, or “militain- ps://doi.org/10.1016/j.biopsych.2015.04.003. ment,” include Stahl, Roger. Militainment, Rachel Yehuda and scholars conducted a Inc.: War, Media, and Popular Culture. New small-scale study of offspring of Holocaust York: Routledge, 2010; and Der Derian, survivors, which is allegedly the first James. Virtuous War: Mapping the Military- demonstration of epigenetic change caused Industrial-Media- Entertainment Network. by preconception parental trauma. The Routledge, 2009. study received much scrutiny, including criticisms of its conclusions, suggesting [4] The general orientation to semiotic mode that further research is necessary. See adopted by Jason Hawreliak differs from Yehuda, Rachel et al. “Holocaust Exposure that pursued in multimodality, specifically the Induced Intergenerational Effects on FKBP5 empirically-driven approaches increasingly Methylation.” Biological Psychiatry, 80(5), strengthened in state-of-the-art multimodal 2016, pp. 372- 380, https://doi.org/10.1016/j. research. The latter, pursued by such schol- biopsych.2015.08.005. ars as John Bateman, Janina Wildfeuer, and

108 Carman Ng: AFFECTING REALITY

I, argues for more discriminating accounts [9] The average time of gameplay is of semiotic modes, emphasizing the role informed by the website “HowLongToBeat”: of discourse semantics and resistance of howlongtobeat.com/game.php?id=29412. assumed alignment between semiotic and sensory modalities. Detailed theorization of [10] Nevermind uses Affectiva Affdex semiotic modes can be found in Bateman, technology to detect and measure view- John A., “The Decomposability of Semiotic ers’ facial expressions. The cloud-based Modes.” Multimodal Studies: Exploring solution can identify 7 emotions and 20 Issues and Domains, edited by Kay L facial expressions, based on a database of O’Halloran and Bradley A. Smith. Routledge, 40,000 advertisements and 7.7 million faces 2011, pp. 17-38. analyzed. For details, see www.affectiva. com/product/affdex-for-market-research/. [5] For example, a classroom communication scenario may be segmented into eight [11] For details on building cohesion chains canvases, from interaction between the for audiovisual media, such as film, from teacher and blackboard, pupils’ use of the perspective of functional and systemic books, to pupil-to-pupil interactions. linguistics, see Tseng, Chiao-I, Cohesion in Film: Tracking Film Elements, Palgrave [6] For an overview of conducting multi- Macmillan, 2013. This analytical form has modal research and the identification of likewise been applied to comics and graphic multimodal slices, see Bateman, Wildfeuer, novels. In this essay, I select a visual-based and Hippala, Ch. 7, § 7.1.1 “Media and their format of communication to ensure clarity canvases” and §7.1.2 “From canvases to and accessibility for a broader readership. analyses.” [12] For details on categorization of game- [7] In a recent, politically significant study, play objects and its application in conducting researchers combine multimodal analysis gameplay analyses, see Ng, War and Will, with natural language processing, computer chapters 3 and 4. vision, and machine learning to examine the spread and re-interpretation of ISIS propa- [13] Timestamps of gameplay are informed ganda and images via digital networks. See by the author’s gameplay experience Tan, Sabine; O’Halloran, Kay L.; Wignell, and approximate average extracted from Peter; Chai, Kevin; and Lange, Rebecca. playthroughs streamed on such websites as “A Multimodal Mixed Methods Approach for YouTube. Examining Recontextualisation Patterns of Violent Extremist Images in Online Media.” [14] The six themes common to the portray- Discourse, Context & Media, 21 (March al of PTSD in popular videogames identified 2018), pp. 18-35, https://doi.org/10.1016/j. include: how characters build trauma dcm.2017.11.004. into their identity; PTSD interference with personal relationships; representations of [8] Nevermind originates from a 2012 trauma through nightmares; self-medication graduate research project at the Interactive as coping mechanism; personification of Media Program at the University of Southern PTSD through villains; and how trauma California. catalyzes digital gunplay.

109 APRJA Volume 8, Issue 1, 2019

Entertainment Software Association. Works cited Essential Facts about the Computer and Video Game Industy: Sales, Demographic, Anable, Aubrey. Playing with Feelings: and Usage Data. 2018, www.theesa.com/ Video Games and Affect. University of wp-content/uploads/2018/05/EF2018_ Minnesota Press, 2018. FINAL.pdf.

Aarseth, Espen. “Ontology.” The Routledge Ensslin, Astrid. The Language of Gaming. Companion to Video Games, edited by Mark Palgrave Macmillan, 2012. J.P. Wolf and Bernard Perron. Routledge, 2014, pp. 484-492. Fizek, Sonja, Paulo Ruffino, Seth Giddings, and Justyna Janik. “Video Games and Bateman, John Arnold, Janina Wildfeuer, Posthumanism.” Panel at DiGRA 2018: The and Tuomo Hiippala. Multimodality: Game Is the Message, July 24-28, 2018, Foundations, Research and Analysis – A Universita di Torino, Italy. Problem-Oriented Introduction. Walter de Gruyter, 2017. Flanagan, Mary. Critical Play: Radical Game Design. MIT Press, 2009. Baldry, Anthony, and Paul J. Thibault. Multimodal Transcription and Text Analysis: Flying Mollusk. “Therapeutic Applications.” A Multimedia Toolkit and Coursebook with Nevermind, 2016. nevermindgame.com/ Associated On-Line Course, Textbooks, and wellness/. Surveys in Linguistics. Equinox, 2006. Gallagher, Shaun. How the Body Shapes Bogost, Ian. Persuasive Games. MIT Press, the Mind. Oxford University Press, 2015. 2007. Galloway, Alexander. Gaming: Essays on Braidotti, Rosi. The Posthuman. Polity Algorithmic Culture. University of Minnesota Press, 2013. Press, 2006.

Bumbalough, Mathew, and Adam Henze. Gee, James Paul. Unified Discourse “Infinite Ammo: Exploring Issues of Post- Analysis: Language, Reality, Virtual Worlds Traumatic Stress Disorder in Popular Video and Video Games. Routledge, 2014. Games.” Emotions, Technology, and Digital Games, edited by Sharon Y. Tettegah and Gee, James Paul. Good Video Games + Wenhao David Huang. Elsevier, 2016, pp. Good Learning: Collected Essays on Video 15-34. Games, Learning and Literacy. 2nd ed. Peter Lang, 2013. De Freitas, Sara, and Paul Maharg. Digital Games and Learning. Continuum, 2011. Gibbons, Alison. Multimodality, Cognition, and Experimental Literature. Routledge, EMTECH. Emotional Machines: The 2012. Technological Transformations of Intimacy in Japan. Project website, www.geschkult. Grusin, Richard. The Nonhuman Turn. fu-berlin.de/en/e/emtech. 2017-2022. University of Minnesota Press, 2015.

110 Carman Ng: AFFECTING REALITY

Halliday, Michael A. K. Language as Social Keogh, Brendan. “Across Worlds and Semiotic: The Social Interpretation of Bodies: Criticism in the Age of Video Language and Meaning. Edward Arnold, Games.” Journal of Game Criticism, 1(1), 1978. 2014, http://gamecriticism.org/articles/ keogh-1-1/. Halliday, Michael A. K., and Ruqaiya Hasan. Language, Context, and Text: Aspects of Kress, Gunther. Multimodality: A Social Language in a Social-Semiotic Perspective. Semiotic Approach to Contemporary Oxford University Press, 1989. Communication. Routledge, 2010.

Haraway, Donna. The Haraway Reader. Kurzweil, Ray. The Singularity is Near: Routledge, 2004. When Humans Transcend Biology. Viking, 2005. Hauseller, Michael, Thomas D. Philbeck, and Curtis D. Carbonell, eds. The Palgrave Leino, Olli Tapio. ‘‘Untangling Gameplay: Handbook of Posthumanism in Film and An Account of Experience, Activity and Television. Palgrave Macmillan, 2015. Materiality within Computer Game Play.’’ The Philosophy of Computer Games, vol. Hawreliak, Jason. Multimodal Semiotics and 7, edited by John Richard Sageng, Hallvard Rhetoric in Videogames. Routledge, 2019. Fossheim, and Tarjei Mandt Larsen, Springer, 2012, pp. 57-75. Hayles, N. Katherine. “Reconfiguring the Posthuman.” Comparative Literature Lin, Patrick, Ryan Jenkins, and Keith Abney. Studies, 41(3), 2004, pp. 311-316, http:// Robot Ethics 2.0: From Autonomous Cars muse.jhu.edu/journals/cls/summary/ to Artificial Intelligence. Oxford University v041/41.3hayles.html. Press, 2017.

Hayles, N. Katherine. How We Became Massumi, Brian. Parables of the Virtual: Posthuman: Virtual Bodies in Cybernetics, Movement, Affect, Sensation. Duke Literature, and Informatics. University of University Press, 2002. Chicago Press, 1999. McMurtrie, Robert James. The Semiotics of Jagoda, Patrick. “Videogame Criticism Movement in Space: A User’s Perspective. and Games in the Twenty-First Century.” Routledge, 2016. American Literary History 29, Spring 2017, pp. 205-18, https://doi.org/10.1093/alh/ Meta Company. 2017. www.metavision.com. ajw064. Nayar, Pramod K. Posthumanism. Polity Jewitt, Carey, ed. The Routledge Handbook Press, 2014. of Multimodal Analysis. 2nd ed. Routledge, 2014. Norris, Sigrid. Analyzing Multimodal Interaction: A Methodological Framework. Jewitt, Carey, Jeff Bezemer, and Kay Routledge, 2004. O’Halloran. Introducing Multimodality. Routledge, 2016.

111 APRJA Volume 8, Issue 1, 2019

Ng, Carman. War and Will: A Multisemiotic Smethurst, Tobi. Playing with Trauma in Analysis of Metal Gear Solid 4, 2017. PhD Video Games: Interreactivity, Empathy, dissertation. Perpetration, 2015. Ghent University, PhD dissertation. Nitsche, Michael. Video Game Spaces: Image, Play, and Structure in 3D Worlds. Southwick, Steven M., George A. Bonanno, MIT Press, 2008. Ann S. Masten, Catherine Panter-Brick, and Rachel Yehuda. “Resilience Definitions, O’Halloran, Kay L. “The Language of Theory, and Challenges: Interdisciplinary Learning Mathematics: A Multimodal Perspectives.” European Journal of Perspective.” The Journal of Mathematical Psychotraumatology, 5, 2014, doi: 10.3402/ Behavior, 40 (Part A), (December 2015), ejpt.v5.25338. pp. 63-74, https://doi.org/10.1016/j. jmathb.2014.09.002. Stewart, Kathleen. Ordinary Affects. Duke University Press, 2007. O’Neil, Cathy. Weapons of Math Destruction. Broadway Books, 2016. Toh, Weimin. A Multimodal Approach to Video Games and the Player Experience. O’Toole, Michael. The Language of Routledge, 2019. Displayed Art. 2nd ed. Routledge, 2011. Von Scheve, Christian. “A Social Picard, Rosalind. Affective Computing. MIT Relational Account of Affect.” European Press, 1997. Journal of Social Theory, 21(1), 2017, pp. 39-59, https://doi.org/10.1177% Ryan, Marie-Laure. “Introduction.” 2F1368431017690007. Narrative across Media: The Languages of Storytelling, Marie-Laure Ryan, ed. Wetherell, Margaret. “Affect and Discourse University of Nebraska Press, 2004, pp. – What’s the Problem? From Affect as 1-40. Excess to Affective/Discursive Practice.” Subjectivity, 6(4), 2013, pp. 349-368, doi: Sicart, Miguel. Beyond Choices: The Design 10.1057/sub.2013.13. of Ethical Gameplay. MIT Press, 2013. Williams, Raymond. Marxism and Literature. Sloterdijk, Peter. “Anthropo-Technology.” Oxford University Press, 1977. New Perspectives Quarterly, 31.1, 2014, pp. 12-19. https://doi.org/10.1111/npqu.11419.

Smethurst, Tobi, and Stef Craps. “Playing Ludography/media with Trauma: Interreactivity, Empathy, and Complicity in The Walking Dead Blade Runner 2049. Directed by Denis Video Game.” Games and Culture, Villeneuve, Sony Pictures, 2017. 10(3), 2014, pp. 269-290, https://doi. org/10.1177/1555412014559306. Cage, David (Director) and Simon Wasselin (Designer). Detroit: Become Human. Quantic Dream. 2018.

112 Deus Ex (series). Ion Storm/ Eidos Montréal/ Square Enix. 2000 - present.

Ex Machina. Directed by Alex Garland, A24, 2015.

Flying Mollusk. Nevermind. Steam. 2016.

Humans, created by Sam Vincent and Jonathan Brackley, AMC Networks, 2015 - present.

Kojima, Hideo. Metal Gear Solid. 1998 - 2015. Assorted videogame platforms.

Max Payne (series). Remedy Entertainment/ Rockstar Games. 2001 - 2012.

The Matrix. Directed by The Wachowski Brothers, Warner Brothers, 1999.

The Walking Dead. Telltale Games. 2012.

Westworld, created by Jonathan Nolan and Lisa Joy, Warner Brothers, 2016 - present.

113 Malthe Stavning Erslev

I FORCED A BOT TO READ OVER 1,000 PAPERS FROM OPEN ACCESS JOURNALS AND THEN ASKED IT TO WRITE A PAPER OF ITS OWN. HERE IS THE RESULT. OR, A QUASI- MATERIALIST APPROACH TO BOT-MIMICRY Abstract

The article develops an approach for close reading of auto-generative writing agents (i.e. bots). It introduces the concept of bot-mimicry (a practice of writing in a bot-esque style), and argues that bot-mimicry inherently entails that reader and writer alike imagine a conceptual (fictional) bot which could have written the text. As such, it investigates the concept as a fruitful way of engaging with cultural, aesthetic and political conceptions and imaginaries surrounding bots. Furthermore, and through an example reading of the “Olive Garden tweet”, the paper develops, introduces and applies a quasi-materialist approach, where seemingly immaterial elements such as implicit conceptual bots are considered through a framework inspired by materialist media theory from the fields of software studies, media archaeology, and electronic literature.

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Malthe Stavning Erslev: I FORCED A BOT TO READ ...

defined by continually more advanced text- Bot or not? generation systems, it is increasingly viewed as a democratic concern to do so (Laquintano my ukulele is not a baby / please do and Vee; Ferrara et al.). Accordingly, the not reply to this maybe / we did not find developers of one of the latest and most it on the internet / the ukulele advanced text-generation systems, GPT-2, highlight its policy implications precisely be- Consider the poem above: was it written by cause of such democratic concerns (Radford a human or by an auto-generative writing et al.). Importantly, the point here is not to agent (a bot)? Chances are you will guess scapegoat the bots for our democratic is- ‘bot’ — at least according to the statistics sues, but to recognize that changes in the of the website bot or not, an online Turing online textual landscape calls for the devel- test for poetry (cf. Laird and Schwartz; opment of more nuanced, fine-tuned, and 67% guess ‘bot’). In fact, it was written by critical reading skills, specifically to navigate a human, Aaron Koh. There is apparently a an auto-generative situation. One important somewhat shared feeling that it reads as an aspect of contemporary text-generation auto-generated text — maybe it has to do is a multitude of tech-narratives, reinforc- with the choice of words, the apparent lack of ing cultural conceptions of text-generation semantic content, the not-quite-right rhythm technologies. of the verses. In any case, this poem, ukulele, The point of this paper is to develop mimics the style of a bot; it is an example of an approach to investigating and critiquing what I will here call bot-mimicry: a practice such cultural conceptions as expressed in of writing in a bot-esque style. The bot or not narratives. The paper takes as its point of website contains numerous examples of the departure a poem which was not written by blurred line between human-written and bot- a bot, but which reads as though it was. At generated poetry. This blurred line evidences one level, it may seem that the poem ukulele that a complete distinction between human- is simply a remediation of well-known ways written and bot-generated text is difficult of troubling the idea of human creativity by if not impossible to uphold in practice. It is referencing machinic processes, such as not controversial to claim that no written text those famously practiced by avant-garde is the product of pure human creativity, but movements including dada and OuLiPo. always already entails ‘technical’ aspects, in- While this may in part be true, I nonetheless cluding plagiarism, remix, reference to fixed argue that there is something more at play grammars, usage of predefined structures, when humans write texts that are supposed etc. (Goldsmith). Likewise, there is always to be read as bot-generated texts without at least some human involvement in any consciously involving any formal logical bot-generated text, if not in editing/curating system. This writing (i.e. bot-mimicry) is the results, then in building and selecting necessarily based on cultural conceptions of data-sets, and not least in programming the text-generation technology in general, which generative software (this holds true even for are then written into the texts in question. so-called ‘unsupervised’ systems). In other words, reading bot-mimicry-texts Still, we may be inclined to maintain an, allows us to study shared conceptions con- albeit troubled, difference between texts pri- cerning bots precisely because they are not marily written by humans and those primarily actually written by bots: the writer and reader generated by bots. Indeed, in a time partly alike are required to (often implicitly) imagine

115 APRJA Volume 8, Issue 1, 2019 a conceptual bot which could have written the text in question.

A quasi-materialist approach

Figure 1: @KeatonPatti’s Olive Garden tweet (Patti). The theoretical grounding of the paper is primarily based on materialist approaches, specifically the field of software studies The case of the Olive (Fuller; Cox, McLean, Ward) along with related perspectives situated in media archae- Garden tweet ology (Wardrip-Fruin), electronic literature (Cayley), and interface criticism (Andersen Since early 2018, Twitter user @KeatonPatti and Pold). Seeing that bot-mimicry concerns has popularized a style of tweet in which non-existent, and as such immaterial, bots he claims to have ‘forced’ a bot to watch (as will be elaborated, I read such bots as over 1,000 hours or episodes of (often pop fictive), it may seem counter-intuitive to apply cultural) video content and then ‘asked’ it to a materialist approach. Though the actual auto-generate new, similar, content. Though text which hints at the bot could be studied @KeatonPatti is not the only one writing in materially, the imagined bot itself may, at this style (the style is now recognized as a first, seem less appropriate for such inquiry. meme by KnowYourMeme; Caldwell), this Nonetheless, I argue that such an approach paper focuses on a specific tweet by @ is not only possible, it is necessary: we need KeatonPatti, posted on June 13th 2018, to study the imagined bots present in a mul- which parodies commercials for the Italian- titude of cultural contexts with the same rigor themed restaurant chain Olive Garden as the actual bots which the imagined ones (see illustrations 1-3). This specific tweet mimic. To this end, I aim to develop what I is chosen because it is the (to date) most call a quasi-materialist [1] approach: a frame- viral tweet in this style; it has at the time of work for applying rigorous materialist theory writing gained ~326.000 likes and ~120.000 to imagined (fictive) entities, in this case bot- retweets (Patti). The tweet also sparked mimicry. The paper takes on a specific case, quite a few reactions on and off Twitter, and the quasi-materialist approach will be including the online magazines Futurism developed in dialectical relation to the case, and Gizmodo, both focusing on the ability where various exemplar frameworks from dif- to discriminate between human-written and ferent fields are brought into consideration, bot-generated text. These articles referred while continually referring back to the case to a series of tweets by @JanelleCShane, as the grounding for the approach. who argued that @KeatonPatti’s tweets were “100% human-written with no bot involved,” and stating that she “wish people wouldn’t present these fakes as bot-written,” though she also found at least some aspects of the tweet “pretty darn funny” (Shane; Shane is considered to be an expert on auto-generative writing and is known for her experiments

116 Malthe Stavning Erslev: I FORCED A BOT TO READ ...

Figure 2: First half of the screenplay attached to the Figure 3: Second half of the screenplay attached to the Olive Garden tweet (Patti). Olive Garden tweet (Patti). with neural network-driven text-generation or recurrent neural networks). Rather, the on http://aiweirdness.com). Taking the wide implied bot relates to auto-generated text range of reactions to the Olive Garden tweet, in general, an amalgamation of a multitude along with its viral status, into account, the of text-generation techniques and the style tweet provides an exceptionally fruitful case. they generally write in. In this case, reading bot-mimicry does not rely on cultural conceptions relating to a single technique (though it might in other cases), but rather on cultural The case as meta-parody conceptions of artificial intelligence/machine learning (AI/ML), and auto-generative text in Reading through the responses to both @ general. As such, the tweets also become KeatonPatti’s and @JanelleCShane’s tweets, somewhat platform-specific to Twitter. one gets the sense that only relatively few The tweets are situated in a context people are actually tricked into thinking that where generative text is commonplace, the Olive Garden tweet was written by a bot. often in the form of so-called Twitter-bots (cf. Many reference the fact that @KeatonPatti Flores), but also represented in the popular is a known comedy writer, who e.g. writes predictive keyboard-based narratives by e.g. for the parody newsmedium The Onion. Botnik Studios (Botnik Studios). Twitter-bots It seems, then, that the comical aspects of are in fact so common that they are viewed the tweet are not at all reliant on the reader as a problem by some, and action has been believing that the bot is real. Rather, I argue, taken towards limiting the presence of auto- the tweet contains two closely connected mated bots on Twitter, or at least to make it jokes — it is a two-fold parody: both a parody possible to locate bot-driven accounts auto- of Olive Garden commercials and a kind of matically (Siddiqui, Healy, Olmsted; Davis et meta-parody of text-generation bots in gen- al.). Most users of Twitter are used to seeing eral, specifically those common on Twitter. auto-generated content, and many of them I argue that the implied bot is not have a somewhat technical understanding of inspired by a single text-generation tech- how Twitter-bots (and auto-generative text in nique (such as Tracery grammars, markov general) works (which also shows in many chains, predictive text keyboards, word2vec, of the responses calling @KeatonPatti out

117 APRJA Volume 8, Issue 1, 2019 for not actually involving a bot in the writ- their reading of it alongside science fiction ing process). Indeed, it seems likely that @ provides an opportunity to “point to a series KeatonPatti’s many bot-mimicry-texts would of themes that illuminate contemporary im- not have been successful outside of Twitter. aginings of the relationship between science, As such, @KeatonPatti relies on his readers technology, and society” (ibid. 773). Though being used to reading these kinds of texts — @KeatonPatti’s tweet hardly illuminates this is virtually necessary in order for them to ubicomp research, this approach inspired appreciate the tweet’s meta-parody. by Dourish and Bell is equally fruitful when Though one would arguably still be applied here. able to find the screenplays funny without In the present quasi-materialist study appreciation of the meta-parody, the parody of @KeatonPatti’s implied bot, the concept of Olive Garden commercials changes when of the diegetic prototype provides a valuable the reader is aware of the meta-parody: the perspective. The term diegetic refers to that relation between the text, the platform, and which is part of a story. A diegetic prototype the output becomes negotiable, and the is a prototypical technology embedded in a reader engages in a creative act of combin- story as a way to communicate or explore ing the reading of the parody and that of the possibilities and dangers connected to wide- meta-parody. My approach is primarily con- spread implementation of these (yet fictional) cerned with the meta-parody, which relates technologies, as is fairly common within to the implied bot and its alleged generative science fiction (Kirby). With the perspective process. of design fiction, Julian Bleecker has shown how the line between ‘science fiction’ and ‘science fact’ is blurred — how the diegetic prototypes known from e.g. Stanley Kubrick’s Reading the implied bot 2001: A Space Odyssey or William Gibson’s Neuromancer have played major roles in both technology development and discourse The implied bot as diegetic (Bleecker). prototype At this point, it is not entirely clear how @KeatonPatti’s implied bot can be viewed as a diegetic prototype. As underlined by Joshua I read the Olive Garden tweet as fiction, Tanenbaum, diegetic prototypes only work maybe even a kind of science fiction. It con- when they are embedded as part of a story tains two stories – the story present in the (Tanenbaum) — and to what extend does screenplay and a meta-story of its genera- that apply to @KeatonPatti’s implied bot? On tion. When reading the tweet as fiction, I in one level, the Olive Garden tweet is obviously part follow Paul Dourish and Genevieve Bell a little story (it takes form as a screenplay), who have studied the interesting relation be- but in addition to this, I argue, it contains a tween science fiction and ubiquitous comput- meta-story which relates to the generation ing (ubicomp) research by reading ubicomp of the screenplay. The screenplay relies on research alongside science fiction (Dourish the readers’ understanding that it was written and Bell). They inquire into the collective by a (fictional) bot. This understanding is of imagining that shapes much of ubicomp re- course derived from @KeatonPatti’s brief search. Their argument is not, however, that intro, written in the tweet, where the screen- ubicomp research is science fiction, but that play is attached as two images.

118 Malthe Stavning Erslev: I FORCED A BOT TO READ ...

Importantly, @KeatonPatti’s intro is arguably not diegetic to the screenplay. It is The implied bot as imagined not part of that story, but is rather a paratext, generative system i.e. a text that is part of the work at hand, but is not part of the story per se, like the text In Noah Wardrip-Fruin’s reading of written on the back of a book. Still, I argue Christopher Strachey’s 1952 Love Letter that the implied bot is diegetic to a different Generator, situated as media archaeology, story, the meta-story in which the person Wardrip-Fruin views the generator, “not as a Keaton Patti (played by @KeatonPatti) de- process for generating parodies, but as itself veloped, trained, and initiated a generative a parody of a process” (Wardrip-Fruin 316). bot, which then outputted the screenplay in The love letter generator uses simple pre- question. Seeing that all these steps are en- written grammars and a fairly small database tirely fictional, I read them as a meta-story in sampled from a thesaurus to generate almost which the implied bot functions as a diegetic rambling expressions that somehow mimic prototype. The relation between the story love letters, but in Wardrip-Fruin’s reading, (the screenplay) and the meta-story (the gen- the letters themselves are “not really the erative process) is the same as the relation interesting part of the project” (306). What between the parody of Olive Garden and the is interesting to Wardrip-Fruin is the genera- meta-parody of text-generation bots which tor’s data and processes, and in his reading was outlined above. To reiterate, my focus of these elements, the love letter generator lies on the meta-story and the meta-parody, is viewed as a parody of mainstream love which is where the study of the implied bot letter-writing activities. This relation between can be conducted. the outputted love letters and the parody of In my approach to reading the implied the letter-writing process is similar to that of bot, this view of the bot as a diegetic proto- the screenplay and the meta-story (and thus type is the point of departure. The reading of the meta-parody) within the Olive Garden @KeatonPatti’s tweet as fictional and the bot tweet. The generative process of the love as a diegetic prototype is fruitful in that it al- letter generator is in itself the parody in ques- lows for an approach to the bot as a concrete tion, which maps onto the meta-story of the entity to be studied through a reading of the Olive Garden tweet — the meta-story and the story in which it occurs. meta-parody both also refer to the implied The implied bot is, of course, only bot’s generative process. Thus, we turn to implied. This makes it particularly difficult to the study of generative software in general study thoroughly, even when viewed as a in order to further develop a quasi-materialist diegetic prototype. In the following, I review approach to our case of bot-mimicry. selected perspectives from the fields of me- One immediate issue is, though, that dia archaeology, software studies, electronic the software per se is arguably an integral literature, and interface criticism, in order to aspect of generative art (Cox). As such, the clarify how one might study any generative reading of generative art entails a close read- bot. In reviewing these perspectives, I seek ing of the system’s generative process (see to identify concrete methods and techniques Wardrip-Fruin’s reading). A close reading to apply in the reading of @KeatonPatti’s of data, processes, and code is of course implied bot. impossible in our case – these simply do not exist. Still, we may be able to study aspects

119 APRJA Volume 8, Issue 1, 2019 of the (implied) generator without having access to its (imaginative) technical elements. Consider Alex McLean’s generative work forkbomb.pl (cf. Cox; Cox, McLean, Ward; see figure 4). A ‘forkbomb’ is a computer program which forks (copies itself) continuously until the system crashes, and are usually very simple programs (they often require only a single line of code). McLean’s work consists of a few more lines than that: in addition to being a simple forkbomb, it also generates a visual output while the program is executing; the visual output is a binary pattern which glitches as the program forks and the system crashes. This visual output Figure 4: “Output and program script from Alex has been read by generative art practitioners McLean’s forkbomb.pl” (Cox 2). The output differs on and scholars Geoff Cox, Alex McLean and different machines, hence the idea of a ‘watermark’. Adrian Ward as “a ‘watermark’ of the proces- sophisticated algorithm which is seemingly sor and operating system” (Cox, McLean, able to discern what is to be considered stage Ward n.p.). While Cox, McLean, and Ward directions, which lines are being spoken by maintain the importance of considering the whom, and so on. This is impossible (or at code of generative works, they include a least very unlikely), and is one aspect which focus on execution as equally important. hints that @KeatonPatti’s tweet should not be The aesthetic appreciation of a generative taken seriously, but rather be read as a joke. work is here considered partly dependent But it also hints at certain tech-narratives on an appreciation of what the work actually that computers are able to extract any kind generates. There is a clear notion that the of information from any datatype. Indeed, output bears an imprint of its generator’s there seems to be a narrative that computers technical elements — even those beyond the are able to extract a kind of essence from source code (e.g. the operating system). In given input data, and this narrative is clearly the case of the Olive Garden tweet, is it pos- reflected in @KeatonPatti’s tweet. Another sible to get a sense of the ‘watermark’ of @ ‘watermark’ of the generator is the ability to KeatonPatti’s implied bot by considering its maintain a structurally cohesive narrative alleged output? throughout the screenplay, including keep- The ‘watermark’ that Cox, McLean, and ing track of characters. Again, though this Ward discern from forkbomb.pl is enlightened is not entirely impossible, it is a distinctively by a knowledge of what the system is doing sophisticated aspect of the screenplay. Both ‘behind the scene’ — i.e. what kind of process these aspects were also brought up by @ is executing. In our case, @KeatonPatti’s JanelleCShane in her critical reading of the short explanation of how the screenplay tweet, but where she read them as fallacies, was generated hints at some aspects of the I read them as parts of the meta-parody, and processing which is allegedly happening in as hints at the tech-narratives at play in the the Olive Garden tweet. Firstly, the relation implied bot. These fairly technical aspects of between inputting video material and the bot the implied bot refer to how the bot processed outputting screenplays hints at an extremely the inputted Olive Garden commercials.

120 Malthe Stavning Erslev: I FORCED A BOT TO READ ...

The implied bot’s transformation of corpus into output

So how can we scrutinize the data, or the ‘corpus’ in natural language processing terms, with a quasi-materialist approach? Let Figure 5: An example of how Cayley inquires into us consider a concrete example of a reading Google’s database. In the example above, it is shown of a corpus by considering the output, situ- how the resulting ‘originality’ changes over time, and ated within the field of electronic literature. depends on which mirror of Google’s database he arbitrarily accesses (Cayley n.p.). In his essay, Writing to be found and writing readers, John Cayley writes with(in) the In relation to the Olive Garden tweet, Google search engine, and through his writ- this is important as it allows us to scrutinize ing practice, he investigates the system’s @KeatonPatti’s idea of a corpus text and corpus (Cayley). His writing technique is how this is processed by his implied bot by relatively simple: he uses the Google search reading the output. One striking insight into engine to look up sequences of words from the relation between corpus and output is a source text, and re-writes the text based the bot’s ability to be creative — i.e. to create on the results — or rather, the new text something new that did not exist in the corpus is based on various concepts relating to text. This is, as of yet, impossible to do with which sequences of words are not found in any text-generation software (or, indeed, at Google’s database. In one example, Cayley all using AI/ML [2]). Accordingly, this feature searches for the longest sequences within a was among the most frequently highlighted predefined string of words which do not get by critical readers of @KeatonPatti’s tweet, any results, letting the poem take form ac- where e.g. the ideas of ‘Gluten Classico’ or to cordingly (see figure 5). These sequences of ‘eat Italian citizens’ were taken as concepts words are, then, technically original, at least which arguably could not have originated in comparison to what has been indexed in any Olive Garden commercial. Thus, the by Google’s ‘spiders’. Cayley’s technique is bot seems to have the ability to not only ex- interesting to our case in that it is a striking tract suspiciously exact data from the video example of how engaging with output text content fed to it, it is also seemingly able to can inform an understanding of a system’s synthesize new and highly creative concepts corpus text. Cayley’s writing engages directly from this data. So, taken together, these with a corpus, exploring it by querying into it. ‘watermarks’, along with the relation between Furthermore, Cayley’s reading demonstrates corpus and output, hint that the implied bot is an engagement with the way this corpus extremely sophisticated. In the following sec- is organized and processed: his results tion, I move beyond this fairly simple reading vary depending on which Google server he and aim to consider a more nuanced under- arbitrarily accesses — something which he standing of the tweet, as well as to provide discovers through his writing practice. a framework for conducting political critique on the cultural conceptions and technology imaginings written into the implied bot.

121 APRJA Volume 8, Issue 1, 2019

The implied bot’s political tendency

As a final aspect of the present reading of @KeatonPatti’s Olive Garden tweet, the field of interface criticism will frame a political critique as well as a general reading of implied bot. In their latest book, The Metainterface, Christian Ulrik Andersen and Søren Pold explore how the concept of ‘the Figure 6: Aarseth’s model of “The Textual Machine” interface’ has changed from being something (Aarseth 21). located in a specific place (e.g. desktop computers) to being something ever-present, denotes what is typically called ‘the reader’, ever-connected, and seemingly immate- who in Aarseth’s approach is situated as an rial – as shown by Andersen and Pold, the integrated, constructive, part of the text/ma- material reappears gradually as aspects of a chine, and not as purely receptive, situated metainterface (Andersen and Pold). In their outside of it. Put briefly, the operator may numerous analyses of exemplar artworks, read into (and in doing so, also reconstruct) one common aspect is that the artworks in the text/machine (the implied bot) by con- question are viewed as self-exploratory; sidering the medium and verbal signs (the they are “a material exploration of [their] own tweet and screenplay). Though the tweet, the technological means of production” (24). screenplay, and even the Twitter platform are Andersen and Pold conduct political criticism also parts of the text/machine, my argument of these technological circumstances by ap- is that the implied bot is integral to the text/ plying a focus on the Benjaminian concept of machine in the operator’s engagement with tendency. Tendency here refers to a deeper the it, since the tweet and screenplay are political tendency as materially embedded in somewhat nonsensical without the addition the technological conditions of production, of the implied bot to the text/machine. Thus, which is revealed in and can be leveraged by I argue that we, situating our reading within artistic production. Andersen and Pold’s interface criticism, may In Andersen and Pold’s work, such approach the tendency of the imaginary bot political tendency is explored through fo- by considering it as part of the text/machine. cusing on various types of interface-critical In the case of the Olive Garden tweet, I artworks, and their approach is to analyze argue that the tendency of the tweet revolves various (artistic) interfaces. Their approach is around a seemingly counter-intuitive dy- in part inspired by Espen Aarseth’s concept namic between docility and autonomy. In @ of cybertext (Aarseth), and his model of what KeatonPatti’s words: “I forced a bot to watch he calls the texual machine, which includes [...] then asked it to write” (Patti; my emphasis). three aspects: operator, medium, and verbal These words give insight into the kind of gen- sign which are brought together in the text/ erative process that @KeatonPatti imagines. machine, any of these aspects and can only The bot may be forced to repeat a somewhat be defined in its relation to the other two typical computer-task (i.e. processing large (21; see figure 6). Note that ‘the operator’ amounts of data), but it has to be asked to

122 Malthe Stavning Erslev: I FORCED A BOT TO READ ... perform a typical human-task (writing being a both gives the computer magic abilities traditionally ‘creative’ act). The content of the while at the same time reinforcing an idea screenplay reflects this dynamic as well, as of complete user control (Chun). Thus, @ the screenplay seems to be both seemingly KeatonPatti’s tweet exhibits a tendency random and at the same time strikingly accu- which is not reserved for AI/ML, but which rate in its depiction of Olive Garden (and their applies to cultural conceptions of computers commercials). Two examples of this are the in general, yet this narrative is arguably only concepts of ‘lasagna wings with extra Italy’ amplified when relating to AI/ML, as one cri- and ‘unlimited stick’. Both these concepts terion for successful AI/ML is that the system seem somewhat randomly generated as a exhibits relatively high degrees of autonomy result of the computer ‘forgetting’ what it was while still remaining controllable. writing and thus combining elements that do not usually belong together, as is typical for much auto-generative text. At the same time they both parody the menu selection A quasi-materialist ap- at Olive Garden, one referring to the highly proach to bot-mimicry Americanized version of Italian food served at Olive Garden (lasagna wings being a mix In order to briefly sum up the quasi-materialist of the Italian dish lasagna and the American approach applied here, the individual aspects hot wings-concept), while ‘extra Italy’ is of it, outlined above, are here put in context added to make the Olive Garden experience to one another. The approach is considered appear more authentic — though the attempt relevant for practices of bot-mimicry, i.e. fails, as there is ‘more Italy than necessary’. situations where humans write in a bot-esque The other refers to the option to get unlimited fashion. This practice entails the (implicit or bread sticks at Olive Garden, which is then explicit) imagining of a conceptual bot which taken to the absurd in claiming that ‘it is could have produced the written text. When infinite, it is all’. analyzing this non-existent (fictional) bot, These two examples demonstrate the viewing it as a diegetic prototype allows for dual nature of the bot as both a docile ma- studying it by reading the story in which it is chine randomly stitching together unrelated situated. Following the idea of a ‘watermark’ concepts from a source text and at the same of technical aspects of a generative work, time a potent comedic parody of the input- such ‘watermarks’ can be located in the story ted data, referencing concepts far beyond and hint at the imagined technical aspects of those that would be present in the alleged the fictional bot. By considering how one can data-set. This situates the bot as harmless read into a corpus by considering the output, while simultaneously having almost mystic we can then also analyze the imagined pro- powers to extract an essence of a given input cessing of corpus into output. Finally, by fo- and synthesize it into a condensed form. As cusing on the political tendency inscribed into mentioned, I read this dynamic as at least the text in question, we are able to conduct one aspect of the tweet’s tendency, which critique of the work’s fictional conditions of reflects more broad ideological conceptions production, and relate these to contemporary surrounding the development of AI/ML. The conceptions which dominate tech-narratives. bot’s dual position as both docile and mystic echoes Wendy Chun’s reading of the fetish- like ideas of source code as ‘sourcery’, which

123 APRJA Volume 8, Issue 1, 2019

Far from claiming that the readings car- Implications and future work ried out here can enable people to distinguish bots from humans online, the paper provides a different proposition: That bot-mimicry can The quasi-materialist framework explored be employed as a creative and critical way here points to several interesting aspects re- to inquire into technological conceptions, lating to contemporary developments within narratives, and imaginaries. I propose that natural language processing. In an era where facilitating writing experiments and conduct- people continually worry about bots posing ing readings in bot-mimicry-texts may be as humans, one way of coping is to imitate fruitful ways of engaging directly with these and parody these suspected malicious bots phenomena that are otherwise difficult to by exaggerating particular aspects of com- articulate concretely by encapsulating them putational writing. Such imitation can either in narratives and considering them as though be rather convincing (the poem ukulele by they were material entities. Such quasi-ma- Aaron Koh) or openly fake (the Olive Garden terialist experiments may, then, provide an tweet). The relation of these imitations to opportunity to further examine and critique computational (real) bots is dialectic as the these cultural conceptions. A lens through imitations are based on encounters with which such conceptions can be negotiated, (and conceptions of) real bots, while they explored, and potentially challenged: through may themselves take part in exploring bot a practice of bot-mimicry. writing, potentially discovering blind spots. As such, these imitations may then influence the development of computational bots, likely making them yet more difficult to recognize. What makes @KeatonPatti’s tweet interesting in this context is that it is not concerned with tricking the reader, but rely on the reader noticing its being fake to conduct comedic critique of both AI/ML discourse and Olive Garden commercials. With continuing developments within natural language processing to make computationally generated text indistinguishable from humanly written text, the ‘style’ of the Olive Garden tweet — what I call bot-mimicry — is increasingly interesting. This style is not inherent to the ability to computationally generate language, but feeds into technocultural conceptions of bots, including imaginaries surrounding AI/ML as well as robots in general. What is interesting here is not so much if @KeatonPatti and others represent bot-writing accurately, but rather how these writing experiments exemplify, inquire into, and communicate shared cultural conceptions of bots and AI/ML.

124 Malthe Stavning Erslev: I FORCED A BOT TO READ ...

Notes Bleecker, Julian. Design Fiction: A Short Essay on Design, Science, Fact and Fiction. [1] My usage of the term quasi-materialist is Near Future Laboratory, 2009. not related to its meaning within philosophy of mind, where the term relates to the ques- Botnik Studios. Botnik. https://www.botnik. tion of mind-body dualism. Rather, the prefix org/. Accessed 4 Mar. 2019. ‘quasi-’ simply refers to the fact that the material is not there in a traditional sense, Caldwell, Don. “I Forced a Bot.” Know yet my approach is to consider the cases as Your Meme, https://knowyourmeme.com/ though it was. memes/i-forced-a-bot. Accessed 04 Mar. 2019. [2] This claim that it is impossible to generate something new with AI/ML Cayley, John. “Writing to Be Found and technology is in part based on a lecture Writing Readers.” Digital Humanities given by Professor Matteo Pasquinelli. The Quarterly, vol. 005, no. 3, Nov. 2011. lecture was given at Cambridge Digital Humanities Learning Programme, University Chun, Wendy H.K. “On ‘Sourcery,’ or of Cambridge, on January 14, 2019, as part Code as Fetish.” Configurations, vol. 16, of the Machine Feeling research workshop. no. 3, 2008, pp. 299–324. Project MUSE, Pasquinelli argued that one integral aspect doi:10.1353/con.0.0064. which defines the capabilities of machine learning software is “the undetection of the Cox, Geoff. Generator: About Generative new” (which was also the title of the lecture). Art and/or Software Art. 2006. Self- The undetection of the new refers to the way published, http://www.academia.edu/down- machine learning algorithms ‘learn’ — only load/36518561/generator.pdf. Accessed 04 by statistically aggregating the existing data Mar. 2019. of the data-set, and thus nothing entirely ‘new’ (which did not already exist in the Cox, Geoff, Alex McLean, and Adrian Ward. data-set) can emerge in a machine learning- “The Aesthetics of Generative Code.” Proc. based system. Cf. http://matteopasquinelli. of Generative Art, vol. 03, 2000. com/cambridge-ai/ Davis, Clayton A., Onur Varol, Emilio Ferrara, Alessandro Flammini, and Filippo Menczer. “BotOrNot: A System to Evaluate Works cited Social Bots.” Proceedings of the 25th International Conference Companion on Aarseth, Espen J. Cybertext: Perspectives World Wide Web – WWW ’16 Companion, on Ergodic Literature. Johns Hopkins ACM Press, 2016, pp. 273–74. Crossref, University Press, 1997. doi:10.1145/2872518.2889302.

Andersen, Christian U., and Søren B. Pold. The Metainterface – The Art of Platforms, Cities and Clouds. MIT Press, 2019.

125 APRJA Volume 8, Issue 1, 2019

Dourish, Paul, and Genevieve Bell. Patti, Keaton. “@KeatonPatti’s Olive “‘Resistance Is Futile’: Reading Science Garden Tweet.” Twitter, https://twitter.com/ Fiction alongside Ubiquitous Computing.” KeatonPatti/status/1006961202998726665. Personal and Ubiquitous Computing, vol. Accessed 04 Mar. 2019. 18, no. 4, Apr. 2014, pp. 769–78. Crossref, doi:10.1007/s00779-013-0678-7. Radford, Alec, Jeff Wu, Dario Amodei, Daniela Amodei, Jack Clark, Miles Ferrara, Emilio, Onur Varol, Clayton A. Brundage, and Ilya Sutskever. “Better Davis, Filippo Menczer, and Alessandro Language Models and Their Implications.” Flammini. “The Rise of Social Bots.” OpenAI, 14 Feb. 2019, https://blog.openai. Communications of the ACM, vol. 59, com/better-language-models/. no. 7, June 2016, pp. 96–104. Crossref, doi:10.1145/2818717. Shane, Janelle. “@ JanelleCShane’s Response.” Twitter, Flores, Leonardo. Genre: Bot. http:// https://twitter.com/JanelleCShane/sta- iloveepoetry.org/?p=5427. Accessed 4 Mar. tus/1007061610005794817. Accessed 04 2019. Mar. 2019.

Fuller, Matthew, editor. Software Studies: A Siddiqui, Husna, Elizabeth Healy, and Aspen Lexicon. MIT Press, 2008. Olmsted, “Bot or Not.” 12th International Conference for Internet Technology and Goldsmith, Kenneth. Uncreative Writing: Secured Transactions (ICITST), 2017, Managing Language in the Digital Age. pp. 462–63. IEEE Xplore, doi:10.23919/ Columbia University Press, 2011. ICITST.2017.8356448.

Kirby, David. “The Future Is Now: Diegetic Tanenbaum, Joshua. “Design Fictional Prototypes and the Role of Popular Films Interactions: Why HCI Should Care in Generating Real-World Technological about Stories.” Interactions, vol. 21, Development.” Social Studies of Science, no. 5, Sept. 2014, pp. 22–23. Crossref, vol. 40, no. 1, Feb. 2010, pp. 41–70. doi:10.1145/2648414. Crossref, doi:10.1177/0306312709338325. Wardrip-Fruin, Noah. “Digital Media Koh, Aaron. “Ukulele.” Bot or Not, http:// Archaeology.” Media Archaeology: botpoet.com/poem/ukulele/. Accessed 4 Approaches, Applications, and Implications Mar. 2019. (Ed. Jussi Parikka and Erkki Huhtamo), Laird, Benjamin, and Oscar Schwartz. “Bot University of Minnesota Press, 2011, pp. or Not.” Bot or Not, http://botpoet.com/. 302–22. Print. Accessed 04 Mar. 2019.

Laquintano, Timothy, and Annette Vee. How Automated Writing Systems Affect the Circulation of Political Information Online. 2017. Crossref, http://dx.doi. org/10.21623%2F1.5.2.4.

126 127 Michela De Carlo

SYNTHETIC BODIES AND FEELING GENERATORS

Abstract

The main purpose of this paper is to describe emerging forms of art and social practices that arise in the social media era, after the coming together of the self-awareness reflected in online environments and the conscious passivity of individuals to the algorithmic manipulation of desires. Accordingly, what follows is a brief introduction to these new forms of social structures and a description of the elements that shape the perfect projection of ourselves in our online experience, combined with samples of artworks investigating the forms and languages emerging in our social media life.

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Michela De Carlo: SYNTHETIC BODIES ...

other people; but we can also think about Introduction sensory ASMR videos, or about those applications tracing our dream activity or helping Digital media have changed the structure of people to fall asleep. our world, allowing us to live our existence The awareness of a wired existence across different stages and platforms. Yet, opens up the question of self-representation the physical borders of our computational in the online environment. The perfect pro- experiences are still well defined: we don’t jection of ourselves becomes an important surf the web through neural implants yet, issue in our social media life, and exploring and the main shift so far has been from the way in which we design it is the main fo- pressing keys on a keyboard to touching a cus of this essay. But in order to get there, we screen with our fingers. Except for the rare first need to outline the social structure that cases in which it occurs to be necessary for technologies and social media have helped survival, we do not integrate technologies to shape, and the new model of individual into our bodies, but we adapt our bodies to on which this social structure is grounded, the way technologies work. Advances have and to which this perfect projection belongs. been made in the medical field, such as the In this effort, we will rely upon the work of creation of artificial organs and use of robots Benjamin H. Bratton, Zygmunt Bauman, and in transplants. However, from this point of Peter Sloterdijk. view, the implementation of technologies in daily life has followed paths which are far from those imagined by science fiction and the media theory of the twentieth century. The user and the bubble As a consequence, a question arises: If people can live with a 3D printed silicone Across the last decades, with the massive heart, why can’t they have feelings obtained adoption of new technologies in the private through their virtual experience or social sphere of individuals and the global con- media life? They can, and they do. In some nectivity bringing together every single thing cases, social media try to reproduce these we do, we find ourselves confronting a new feelings, Facebook’s ‘reactions’ being a social complexity, that has caused, as a common example: six emoticons that allow consequence, a new, strong need to retrace, people to better express how they feel about rephrase and rethink the borders of the social specific content displayed on their wall — if structure we are living in. compared to the emotional neutrality of the In his book The Stack, sociologist ‘like’ — and that allow the system to better Benjamin H. Bratton considers the form of profile us. It’s a pretty basic approach, but it the stack to describe the changes induced works. by an ever more digitized society, but also Hyper-connectivity and new forms of to re-define a hypothetical geo-political map communication influence our feelings, emo- integrating these two aspects in a dichoto- tions, lifestyle and the way we perceive our mous way: bodies. Applications that improve or mask our appearance have been designed, as well I propose The Stack as a way that we as AI ChatBots that pretend to be the perfect might map political geography, but also boy/girlfriends and virtual environments in for how we understand the technolo- which we can reinvent ourselves and meet gies that are making that geography.

129 APRJA Volume 8, Issue 1, 2019

[…] this figure of The Stack both does detailed portrait reflected back convinces us and does not exist as such; it is both of our individual coherency and efficacy.” an idea and a thing; it is a machine (Bratton 260) If our synthetic representa- that serves as a schema as much as tion is mediated by social filters along the it is a schema of machines. It lets us process of transformation from human to see that all of these different machines User, the system in which we choose to are parts of a greater machine, and insert our image — the Interface depending perhaps the diagrammatic image of a upon the Address, depending upon the City, totality that such a perspective pro- depending upon the Cloud, depending upon vides would, as theories of totality have the Earth — gives back to us, in turn, these before, make the composition of alter- social filters, providing a detailed, persuasive natives — including new sovereignties portrait of our coherence and individual and new forms of governance — both effectiveness. more legible and more effective. As This loop between human, User, reflec- the shape of political geography and tion, User and human can be described as a the architecture of planetary-scale circle, a loop with a positive, self-feeding feed- computation as a whole, The Stack is back. The modal value of this paradigm is the an accidental megastructure, one that reflection. If we combine these thoughts with we are building both deliberately and what philosopher and sociologist Zygmunt unwittingly and is in turn building us in Bauman claims in Liquid Life, writing about its own image. (Bratton 4-5) the accelerated rhythms we are subject to, it’s very likely that the reflection sent back by Bratton idealizes a ‘megastructure’, the system wouldn’t match anymore with the exemplifying a hybrid social model — com- idea of coherence and individual effective- putational and non-computational — with ness to which we were referring when we a histogram composed by different levels generated our image as User. co-dependent on each other, and arranged This variance, although minimal, should vertically one upon the other: Earth, Cloud, be added to another circle/loop. If we keep the City, Address, Interface and User (Bratton two poles (human and User) still, considering 10-11). It’s on the level of the User that I’m them as the two input and output poles and going to focus in the following. For Bratton, keeping the perfect shape of the circle, the the User — a word borrowed from the field sum of all the loops will develop by including of design — is the human being as a sub- the Z axis: the third dimension. The sum of ject that organizes the system they inhabit, all this constant and perpetual variance will shaping it in their own image. Their synthetic produce a spherical shape, a globe. double is shaped by social factors such as The three-dimensional rendering of the micro-economies and psychology. In brief, close circuit described in figure 1 evokes for Bratton the User is not an individual or an the metaphor of the bubble, as it is used by un-individual, but rather a plurality of agents, the philosopher Peter Sloterdijk in Spheres a position within a system; and without this I - Bubbles, as the intimate subjectivity of the system, they wouldn’t have a role, nor an individual: the unit of measurement made essential identity. by the individual basket of experiences and In another passage, Bratton writes: “As interactions of the individual. we human users reflect on ourselves with While Bratton calls this unit of measure- images of quantified digital traces, the richly ment User, placing it at the top of his linear

130 Michela De Carlo: SYNTHETIC BODIES ...

Figure 1: Human-User Perpetual Variance. structure composed of overlapping platforms, The cells of the foams lose the perfect shape Sloterdijk, on the other hand, uses the indi- of the sphere, and even if they are attached vidual sphere as a basis for a model of social to one another, forming an ephemeral net, architecture that, in Spheres III - Foams, he they are not truly connected. coherently describes as a “foam architecture” (Sloterdijk 15): a plurality of spheres combined in a disorganized way — one upon the others, one next to the others. Back to The perfect projection of Spheres I - Bubbles: ourselves

In the foam worlds, however, no bubble Although very different from each other, the can be expanded into an absolutely models of individual outlined by Bratton and centered, all-encompassing, amphis- Sloterdijk are very useful to describe the way copic orb; no central light penetrates we live our social media life, and we expand the entire foam in its dynamic murki- our identity online by designing the perfect ness. Hence the ethics of the decen- projection of ourselves. Both Bratton’s User tered, small and middle-sized bubbles and Sloterdijk bubble do not have a fixed in the world foam includes the effort identity and shape, but they are shaped and to move about in an unprecedentedly changed by the system they are part of (and spacious world with an unprecedent- thus change as they move from system to edly modest circumspection; in the system, from platform to platform). And their foam, discrete and polyvalent games of consistency is not an original condition, but a reason must develop that learn to live final achievement — the result of the recol- with a shimmering diversity of perspec- lection of their “quantified traces” (Bratton tives, and dispense with the illusion of 260). the one lordly point of view. (Sloterdijk This achievement is what I call the 75) perfect projection of ourselves. This perfect projection isn’t just the result of an effort in

131 APRJA Volume 8, Issue 1, 2019

Figure 2: Ryan Trecartin, I-BE AREA (2007). Video, 1 hour, 48 minutes. © Ryan Trecartin, Courtesy Regen Projects, Los Angeles and Sprueth Magers. self design — what I call in the following the artist himself and his collaborator Lizzie ‘virtual representation’; but it also requires an Fitch. Although each character is presented ability to actively and passively employ the as an individual, the fact that they speak tools that the digital realm offers us to feel the same language and that they are often and express emotions — what I call ‘feeling interpreted by the same actors (Trecartin generators’; and a willingness to passively and Fitch among them) enforces the feeling accept the algorithmic manipulation of our that they are different manifestations of the feelings and desires, and to actively engage same identity: I-Be, the main character of with non-human personalities and artificial the movie, of which the narrative outlines the intelligences. “area”, the cluster of his various realities and To introduce these three topics, let’s identities. At the beginning of the movie, I-Be, briefly consider one of the first artworks ever a self-proclaimed clone, “I exist because of to engage with the projection of ourselves in Command V. Copy and paste some guy’s online environments: Ryan Trecartin’s I-BE DNA” (Trecartin 8) — has a conversation AREA (2007). The movie, shot as a linear with his avatar. Here, I-Be explains his avatar narrative but also uploaded on YouTube in ten — who wants to assign him a paper — that it minute segments, famously portrays a group can’t assign anything to him, because “I cre- of young, over-active people with heavy ated you”. I-Be’s avatar is his own online pro- make-up in a colorful, messy set designed by jection, the ‘virtual representation’ of himself;

132 Michela De Carlo: SYNTHETIC BODIES ... but at the same time has evolved into an fragmented into a number of ‘quantified trac- ‘independent avatar’ (IA), an autonomous in- es’ — tweets, likes, comments, photos, vid- telligence who writes papers and has its own eos, sounds; some of them are permanent, emotions. But I-Be refuses to recognize and some others are ephemeral, but all of them accept his avatar’s independence, so far to contribute to shaping a portrait of ourselves. decide to delete it: “You can just go cowboy In her digital painting work, the young some abandoned files in my trash can. Swup Chinese artist Ruby Gloom (1991) reflects drag to the trash, empty it, empty it, I emptied on this by combining these traces into iconic, it. Empty.” (Trecartin 10) convincing portraits. In her series Insta Client (2017-ongoing), Gloom makes 3D portraits of people, drawing inspiration from a selfie that is sent her by the client. These portraits Virtual representation are made to be shared on social networks and be traced thanks to the use of hashtags; By posting pictures, sharing articles and in many cases, they are used by the Insta thoughts, or composing 3D avatars, we are Clients as profile pictures. always trying to create the ideal projection What’s especially interesting about this of ourselves in the virtual realm. Our identity project is the fact that most of the photos the expands beyond the body, and ‘users’ can artist receives — providing a model for her become whatever they want, or just idealize portraits — are not rough, plain photographs, themselves showing only their best traits — but are themselves already manipulated us- like a smooth 3D face with no imperfections. ing other applications, presenting for instance Our virtual representation is usually glittering effects, hearts all over the subject

Figure 3: Ruby Gloom, Insta Clients (2017), 3D renders. Courtesy of the Artist.

133 APRJA Volume 8, Issue 1, 2019 and other kinds of digital filters. In some cases […] in varying degrees they all master the faces are masked by Augmented Reality and practice the art of liquid life: filters. Gloom considers these ‘client gener- acceptance to disorientation, immunity ated’ additions made with other applications, to vertigo and adaptation to a state of as they were part of the face, and paints dizziness, tolerance for an absence of them in her portraits. Without distinguishing itinerary and direction and for an indefi- between reality and make-up, she takes the nite duration of travel. […] Looseness image that she gets, and as a machine she of attachment and revocability of produces a new synthetic 3D version of that engagement are the precepts guiding image and spreads the new ‘selfie’ she has everything in which they engage and to created on social networks. which they are attached. (Bauman 4) Thomas Macho’s facial society, that “continually produces faces” (Belting 295) In order to survive this lifestyle, you comes to mind together with the idea of the need to be able to let things go, to eliminate prominent face described by Macho and the past. Then, Bauman assumes that the Hans Belting as a “blank facial formula” same concept works with identities, which (Macho 121); but in this work, it evolves into means that we have to be able to rebuild a filtered facial formula. Here, the virtual rep- ourselves in an easy and fast way, without resentation doesn’t take off from a point zero the fear to leave the past behind like — a that we can consider the real or natural face story on Instagram, that only last 24 hours. (even if we can wonder if a simple photo can A reference to Ryan Trecartin’s I-BE AREA be considered a natural face), but already would fit well here. In the movie, I-Be deletes from a simulation. The result is a simulation his Avatar IA by sending him to the trash, of a simulation. where he can join his other previous avatars. Another point that’s important to high- No regrets — it will be replaced soon. light is that — by examining the representation In 2009, in his interview “Talking to of identity in the social media era — we don’t myself about the politics of space”, Sloterdijk talk anymore only about a specific shape, as played around this concept too, writing about it could be a human body or a human face, multiple personality in relation to online at least not in an absolute way. The focus is activity: more on the manipulation or the masking of the traditional form, and in some cases on its From my point of view, the multiple absence. For this reason when we consider personality is nothing other than the the virtual portrait, we don’t speak about the individual’s answer to the disappear- body, rather we deal with the self. And this self ance of his real social surroundings, is temporary, transient, unstable, ephemeral. and is thus a plausible response to To explain this shift — from the body the chronic lack of social stimulation. to the self — let’s refer again to Bauman. In The second possibility relates to Liquid Life, he states that the acceleration the modern practice of networking. of our contemporary life forces us into new The horde returns in the guise of an beginnings and consequently new losses, iPhone address book. Close physical repeatedly: togetherness is no longer a necessary condition of sociality. (Sloterdijk)

134 Michela De Carlo: SYNTHETIC BODIES ...

Figure 4: Kate Durbin, Hello Selfie Miami (2015), performance. Courtesy of the Artist.

A work dealing with the ephemerality, There Kate and the other performers walked lightness and detachment of digital identities in the water leaving their smartphones on is Los Angeles-based artist Kate Durbin’s the seabed. Like the short human life of The performance Hello Selfie Miami (2015). In Little Mermaid, the selfie’s identity generated this work, Durbin transformed herself and her along the performance and archived in the girl crew into a kind of ‘kitty-mermaids’ made- mobile gallery metaphorically vanishes with up and dressed with pastel colors. During the foam of the waves. The new temporary Art Basel Miami, Durbin wore and put kawaii identity disappears, letting us imagine a new stickers on the body of half-naked woman beginning. performers. They also wore wigs with unnatural and bright colours. After this masking process, Durbin and the performers — voice- less as Andersen’s Little Mermaid — started Feeling generators to take selfies, with their selfie sticks in their hands, among the artworks of the group show I call ‘feeling generators’ those tools — phone in which Durbin was invited to exhibit, without applications, online experiences, digital ever speaking to visitors. After the shooting simulations — that provoke emotions which session, the performers walked slowly and are close to the ones we feel in our physical solemnly outside the gallery, still ignoring the world, but are born in a virtual context medi- audience while passing through it; and they ated by the use of devices, interfaces and walked to the sea, always with selfie sticks in hardwares; and those tools that allow us to their hands - a new extension of their bodies. share our feelings in the virtual sphere. The

135 APRJA Volume 8, Issue 1, 2019 online projection of emotions becomes in emotions, or share statements upon specific turn a generator of emotions for the feeling of issues on blogs, etc. Some common feeling empathy that it causes in other people. As a generators are characterized by a co-exist- corollary to this definition, we can distinguish ence of both aspects, active and passive. the feeling generators into two different Just think about online sexual gaming, or ap- groups: the passive and the active. plications that are based upon the structure of video games, in which active interaction Passive feeling generators are char- by people with a generative feedback by the acterized by the possibility they offer to feel machine and vice-versa is at the base of the emotions produced during and through our game simulation system. online experience without any active interac- Talking about passive feeling gen- tion on our side: we just have to open an erators, let’s briefly focus on the ASMR application, press play, etc. Some examples phenomenon by examining a recent work by are: the state of anxiety generated by the the French artist Caroline Delieutraz. In her lack of response from a person who’s visibly video Unboxing + Tapping + Whispering with online when you write them; the desire to find Rikita (2017), she investigates the world of out the content inside a box when watching the Autonomous Sensory Meridian Response an unboxing video; the combination of posi- (ASMR) by featuring Rikita, a well-known tive feelings and a distinct static-like tingling young French YouTuber making ASMR sensation on the skin while watching an videos. Here, Rikita unboxes a package, de- ASMR video, etc. scribing in a whispering voice what she finds Active feeling generators are those while unwrapping, one by one, the sculptures which allow us to externalize our feelings from the series Embedded files (2015-2017), online: so, we can use default tools pro- by Delieutraz herself. The sculptural work is vided by social networks to communicate our about the embodiment of our internet habits

Figure 5: Caroline Delieutraz, Unboxing + Tapping + Whispering with Rikita (2017). Video, 48 minutes, 56 seconds. Courtesy of the Artist.

136 Michela De Carlo: SYNTHETIC BODIES ...

Figure 6: Juliette Goiffon and Charles Beauté, Does Anybody Know? (2015-2017). Video 18 minutes. Courtesy of the Artists. and about future archaeology: in the series, pleasure with it. In ASMR videos, objects Delieutraz collected images on the internet, become triggers able to generate relaxing or printed them together with technologies of exciting sensations. As Delieutraz explained common use and trendy objects of that pe- in an interview with Stephanie Vidal: “The riod, and enclosed them in paraffin blocks. object’s value is determined by its potential The final result is something in between a as a trigger” (Vidal). So the objects become time capsule and a future fossil. an input to be processed by the voice or the In other words, if in 2015 — with touch of the YouTuber, and the output is a Embedded files — Delieutraz translated our video that people can easily find online. online experience into a physical reality, in the In addition to the fact that these videos 2017 video she associated to this process a are recorded by people for generating effects whispering soul, asking to Rikita to tap, unbox on other people and then uploaded online and describe the sculptures. In this new step, on mainstream channels such as YouTube the previous process of embodiment loses its — which makes them easily accessible on materiality and reverses back to an ephemeral a user-friendly interface — an important state — an ASMR video on YouTube — while aspect of ASMRs is the intimacy in which the achieving a new sense of aura. The objects audience experience them. An ASMR video made by Delieutraz become new objects of is selected out of the many available online, desire thanks to the sensual voice of Rikita. and experienced wearing headphones or, They gain empathy. These sculptures are no even better, earphones (that allow us to bet- longer the untouchable, precious objects on ter enjoy the binaural recording). The feelings display in a white cube set-up. Even if we generated by these videos may vary from are not actually touching them, we can feel relax to ecstatic tingling, from skin pleasure this sensation with our eyes and over all we to non-sexual orgasm. Although a purely can hear this touch and experience a tingling virtual, mediated experience, filtered by our

137 APRJA Volume 8, Issue 1, 2019

Figure 7: Jeremy Bailey, The You Museum (2015). Net-based project. Courtesy of the Artist. eyes and ears, it’s finally through its effects and the need to share these worries over on our skin — its physical consequences — the internet. Does Anybody Know? also that we can measure the effectiveness of an shows our paradoxical vision of medicine, of ASMR video. its highly technological universe which is at A work exploring the attitude to the same time intrinsically human. This suc- share emotions online through what I call cession of visual and textual points of view the active feeling generators is the video brings a double experience of indiscretion Does Anybody Know? by the French art- and projection on the side of the spectator, ists Juliette Goiffon and Charles Beauté. nourished at the same time by the observa- After spending two years observing and tion of the body and the expression of the studying the behaviours of people on medical human thought. blogs, they selected part of the conversations and statements they considered relevant for their research, and they edited them into a video together with a continuous flow of 3D Manipulation of data, scans of different parts of the body. Each machine learning and AI body part is accompanied by a question, an expression of anguish, a fragment of testimo- “you mean machines are like nies stolen from the medical forums. humans?” This hypnotic experience reveals the I shook my head. “No, not like humans. concerns of our society about medical issues With machines the feeling is, well,

138 Michela De Carlo: SYNTHETIC BODIES ...

more finite. It doesn’t go any further. artwork — as chosen for you by the algo- With humans it’s different. The feeling rithm — randomly appeared alongside your is always changing. Like if you love daily online browsing, on advertising banners somebody, the love is always shifting placed in social networks, newspaper home- or wavering. It’s always questioning or pages, and wherever a commercial banner inflating or disappearing or denying or could be placed. hurting. And the thing is, you can’t do A further purpose of the artist was to anything about it, you can’t control it.” highlight the positive artistic potential in us- (Murakami 120) ing data and advertising tools, as he stated in a 2015 interview with Marc Garrett on Finally, the projection of our self in our online Furtherfield: experience is influenced by the conscious passivity of the individual to the algorithmic Yes, I’d like art to reflect positive social manipulation of personal contents and change instead of reflecting negative desires. The elaboration of our personal market demands. Artists have this tre- information allows machine to calculate our mendous ability and power to commu- preferences during our online experience. In nicate and many are wasting that talent Bratton’s words, what happens is “the capi- pandering to the decorating desires of talized translation of interactions into data the rich and powerful. I understand that and data into interactions” (Bratton 42). This everyone needs to make a living, but mechanism is mainly used by companies to we also have a responsibility as artists better profile our needs and focus our atten- to help make the world a better place. tion to the proper advertising. It’s also used I also don’t see why these two things by social networks to highlight contents that need to be in conflict. may get our interest. As a consequence, the interface we live in becomes a container con- Sloterdijk’s and Macho’s notion of taminated by our preferences, our personal ‘nobject’ might be useful here. Consolidating sphere. Macho’s argument, in Spheres I – Bubbles, The Canadian artist Jeremy Bailey Sloterdijk describes nobjects as identifying exploits this mechanism of data calculation a system of co-realities which, in a manner and advertising banners in his net-based that does not include a comparison, are project The You Museum (2015 - ongoing). literally floating as creatures of proximity On a dedicated website, he created a form in front of an inner Self, who is not facing with a few personal and basic questions, that them, because it is itself in a fetal pre-subject the visitors had to answer. Using the answers state (Sloterdijk 200). A nobject is a being given, which were indicators of preferences, who lives in a parallel reality close to ours an algorithm programmed by the artist select- but who has not yet achieved the status of ed which of Bailey’s artworks the visitor might subject. This nobject condition described by like, in a kind of ad-hoc curatorial selection. Sloterdijk and Macho is very close to our cur- Yet, these artworks were not shown to you at rent perception of Artificial Intelligence (AI). the end of the questionnaire. The experience The idea that one day machines will come to on the site was over once you completed the think and learn like human beings dates back form and sent your data to the elaboration to the 1950s; today, also given to the continu- system made by the artist. What happened ous progress in research, we all expect that next was that your favourite Jeremy Bailey — sooner or later — AI will reach this goal:

139 APRJA Volume 8, Issue 1, 2019 the Subject. We are still waiting. Differently Reuell, Peter. “For Teens Who Feel It from the Independent Avatar conceived by All, A Research-Backed Explanation.” Trecartin in I-BE AREA — who became a The Harvard Gazette, September 21, self-sufficient being with his own intelligence 2018, https://news.harvard.edu/gazette/ and emotions — by now, machine learning story/2018/09/harvard-researchers-exam- systems and AI are using ‘big data’ in order ine-evolution-of-emotion-differentiation/. to make predictions of our future behaviour. They learn from us, and reflect us in a more Sloterdijk, Peter. “Talking to Myself about polite and non-empathic way. the Poetics of Space.” Harvard Design Magazine, No 30 - (Sustainability) + Pleasure, Vol. I: Culture and Architecture, Spring - Summer 2009, www.harvard- Works cited designmagazine.org/issues/30/ talking-to-myself-about-the-poetics-of-space. Bauman, Zygmunt. Liquid Life. Polity Press, 2005. Sloterdijk, Peter. Bubbles - Spheres vol I: Microspherology. MIT Press, 2011. Belting, Hans. Face and the Mask: A Double History. Princeton University Press, 2017. Sloterdijk, Peter. Globes - Spheres vol II: Macrospherology. MIT Press, 2014. Bratton, Benjamin H. The Stack: On Software and Sovereignty. MIT Press 2016. Sloterdijk, Peter. Foams - Spheres vol III: Plural spherology. MIT Press, 2016. Garrett, Marc. “Choose Your Muse Interview: Jeremy Bailey.” Furtherfield, Trecartin, Ryan. “I-BE AREA Transcript.” June 26, 2015, https://www.furtherfield.org/ Electronic Arts Intermix, https://www.eai.org/ choose-your-muse-interview-jeremy-bailey/. titles/i-be-area/supportdocs/35.

Macho, Thomas. “Zeichen aus der Vidal, Stephanie. “Interview with Caroline Dunkelheit. Notizien zu einer Theorie der Delieutraz.” Making Contact, online exhibi- Psychose.” Wahnwelten in Zusamennstoss. tion, 2017. Die Psychose als Spiegel der Zeit, edited by Heinz, Rudolf, Dietmar Kamper and Ulrich Sonnemann. Akademie, 1993.

Macho, Thomas. “Das Prominente Gesicht. Vom face-to-face zum interface.” Alle möglichen Welten. Virtuelle Realität, Wahrnehmung, Ethik der Kommunikationen, edited by Fassleer, Manfred. Wilhelm Fink Verlag, 1999.

Murakami, Haruki. Dance, Dance, Dance. Vintage Books, 2003 [1988].

140 141 Tanja Wiehn

(UN)PREDICTABLE ACTS OF DATA IN MACHINE LEARNING ENVIRONMENTS

Abstract

This paper investigates artistic representations of machine learning and their interventional potential. Taking its point of departure in two works of art, the paper discusses effects of predictability and unpredictability caused by machine learning systems. By thinking through “eventfulness” (Bucher) and “nonconscious cognition” (Hayles) in human and non-human environments, the paper analyzes the potential of artistic practices to question and rethink algorithmic processing. The paper provides a framework in which artwork challenges forms of technological predictability and comes to terms with machine learning as a fundamental cultural practice in its own right.

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Tanja Wiehn: (UN)PREDICTABLE ACTS OF DATA ...

(Alpaydin; Pasquinelli). Machine learning is Encounters with basically a form of programming that learns machine learning from the data provided. As Adrian Mackenzie points out, machine learning is an accumula- tion of techniques derived from mathematics Daily entanglement with the technological (statistics) and computer science and is not a other has ambivalent results. Concerns about fundamentally new technology (Mackenzie). the growing impact of algorithms cast doubt Machine learning consists of data training, upon objectivity and reliability in systems of algorithm learning, and model application. machine learning and artificial intelligence, Each of these basic components are crucial notably when their implementation can have for generating an output that — to put it very strong societal ramifications (Mackenzie; simply — is based on the idea of pattern Wang). Critical investigations are simultane- recognition (Pasquinelli). Moreover, machine ously emerging in research discourses such learning is already in everyday technological as New Media Studies, Internet Studies, and use, operating in the background to recognize Algorithmic Studies to analyze and question faces at border control, to generate credit the belief that these technologies are becom- score rankings, and to provide Facebook’s ing providers of solutions to complex social news feed: it is thus more or less visible and equations. Algorithmic tools are advocated tangible (Mackenzie). as means of avoiding all-too-human glitches and forms of unpredictability caused by a A machine learning system is a sort subjective human intervention, as they seem of nooscope, that is a device to map to be “stabilizers of trust, practical and sym- and perceive complex patterns through bolic assurances that their evaluations are vast spaces of data — what, in digital fair and accurate, and free from subjectivity, humanities, is termed as distant read- error or attempted influence” (Gillespie 179). ing. Each instrument of measurement Yet does algorithmic processing of large and perception comes with inbuilt and amounts of data necessarily guarantee neu- contingent aberrations. As much as the trality? Critical investigations of data process- lenses of microscopes and telescopes ing discuss the numerous issues regarding were never perfectly curvilinear and implementation of machine learning and its smooth, similarly AI systems install potential to reproduce racist or sexist biases logical lenses that condense faults and (Kitchin; Wang). Scholarship has thus begun aberrations (Pasquinelli 4). to look more closely at the notion of the algorithm, the ways in which data is used, and In contrast to the statistical core ele- how these relate to machine learning. As a ments in machine learning, the perception result, scholars are looking into the ramifica- of the technology itself as a form of artificial tions of algorithmic decision-making for cul- intelligence receives a different kind of atten- ture and society, drawing on a diverse set of tion. In their article “Situating Methods in the methodological approaches (Elish and boyd; Magic of Big Data and Artificial Intelligence,” Seaver; Kitchin; Gillespie). The movement danah boyd and Madeleine Clare Elish from algorithms towards machine learning argue for a reframing of data analysis tools is of particular relevance here. These methodologies that moves away from the are subsets of artificial intelligence and are terminology of machine learning and towards thus systems that are able to learn and adapt a computational ethnography. These authors

143 APRJA Volume 8, Issue 1, 2019 problematize not only the faith and reliance composition of machine learning relies upon in data-driven technologies but also question several factors in which human input and the the seemingly magical moments of artificial collection of training data are necessary. As intelligence and their branding in commercial a result, the selection of specific training data environments. They are furthermore referring shapes the very core of the machine learning to uncanny effects caused by machines per- model (Mackenzie; Pasquinelli). As Wang forming as though they were human. These notes here, it is always a question of what effects also arise when algorithmic systems input data is provided in the training sets. perform unintelligible glitches, for instance There is no doubt that algorithmic sys- in showing very poor commercial recom- tems tend to infiltrate and influence more and mendations or advertisements. However, the more aspects of life, in different variations algorithmic result can become even more un- and with varying ramifications. A request canny when the machine is providing results for technical insight into these mechanisms (i.e. predictions) as a basis for decisions that seems necessary at this stage, given the are not as banal or mundane as the display problematic outcomes that machine learn- of commercial advertisements. Machine ing environments can entail. Can technical learning-driven surveillance strategies are, knowledge gain insight into algorithms and for instance, introduced in Western democ- their work within machine learning? It is thus racies with the claim of offering superior far understood that the results of machine security assessments based on predictable learning rely on its various components, information. Data collection and processing including its input data. But how can these can, for example, become a resource for au- systems be understood and managed, thorities to assess the likelihood of potential considering that they are often described as future criminal activities. In her book Carceral fundamentally obfuscated and veiled? Capitalism, Jackie Wang untangles the relationship between algorithms, data analysis, racial discrimination, and their carceral function in the United States. Wang uses the The difficulty of predictive policing tool PredPol to show how knowing algorithm(s) data analysis tools are highly reliant on decisions made about their input data. In Wang’s Algorithms are inert, meaningless example, these human decisions determine machines until paired with databases the areas and neighborhoods that are under on which to function. A sociological particular investigation. “Although data has inquiry into an algorithm must always been conceptualized as neutral bits of infor- grapple with the databases to which mation about our world and our behaviors, in it is wedded; failing to do so would be the domain of criminal justice, it is a reflection akin to studying what was said at a of who has been targeted for surveillance and public protest, while failing to notice policing” (Wang 247-248). Forms of policing that some speakers had been stopped informed by machine learning tactics are en- at the park gates (Gillespie 169). dorsed as a reliable science, notably for the sake of security. These methodologies are Wang’s contribution emphasizes the im- claimed to generate predictability through portance of critiquing the implementation the collection and processing of data (Elish of data-driven technologies. In order to and boyd; Wang). As noted above, the very

144 Tanja Wiehn: (UN)PREDICTABLE ACTS OF DATA ... formulate such a critique, it is necessary aim represents a break from the premature to become knowledgeable about what and conclusion that automated data process- how data is collected. Moreover, as Tarlton ing guarantees reliable information and Gillespie points out in the above quote, the predictability (Wang). This is what Bucher forms of data processing through algorithmic calls “distribution of agency,” acknowledging intervention are another crucial aspect. One algorithms as being products of human- way of approaching this problem is through a non-human environments. “Algorithms are technical understanding of machine learning not given; they are not either mathematical and the mechanisms that algorithms carry expressions or expressions of human intent out. Machine learning is highly complex. We but emerge as situated, ongoing accomplish- have already touched upon some composi- ments. That is, they emerge as more or less tions of the technology very briefly here, technical/nonhuman or more or less social/ with a focus on the importance of input data. human because of what else they are related Matteo Pasquinelli proposes deciphering the to” (Bucher 55). Thinking about machine training sets and their processing within ma- learning systems as being constructed from chine learning. Building upon Kate Crawford’s distributed forms of agency is helpful for dis- and Vladen Joler’s illustration of the Amazon pelling the idea of objectivity within technol- Echo system in Anatomy of an AI System, ogy. I’m following Bucher here in considering Pasquinelli highlights the composition and the embeddedness of algorithms in systems affordances of the different training sets that — technological as well as cultural systems. make the machine learn. He furthermore Thus, in looking at the two following works of approaches machine learning algorithms art, I not only acknowledge the interwoven- and the production of bias in their outcomes ness of algorithms with cultural practices but through the statistical and mathematical also consider these examples as potentially compositions in place. demystifying the magical elements of ma- By looking at these training sets, it chine learning as well as forms of human is possible to visualize the human input exceptionalism (Hayles). and decision-making process in machine learning. Taina Bucher likewise highlights the moments of human input in algorithmic processes. For her, these inputs denote a Machine learning as particular interest in the data, the desired nonconscious cognition outcome and the selection of used data in the first place. The necessary human input in In terms of the difficulties of knowing machine learning can thus be characterized algorithmic processes, I refer to Katherine by a prior interest in the data, a particular set Hayles’ exploration of “nonconscious cogni- of assumptions made about a specific case. tion” to gain entry into the first work of art. A preliminary technical insight is valu- In her book Unthought: The Power of The able for understanding machine learning as Cognitive Nonconscious, Hayles uses con- an environment shaped partly by human and temporary neuroscience, literary studies, partly by non-human agency, as a fundamen- economics, and computer science to work tally posthuman endeavor (Bucher; Hayles towards the idea of a nonconscious cogni- How We Became Posthuman). Investigating tion. Hayles describes all forms of a cognition this state of entanglement of machine learn- beyond consciousness as the nonconscious. ing practices with culture and their societal

145 APRJA Volume 8, Issue 1, 2019

She thereby emphasizes the deep entangle- with other critical posthumanists, like Rosi ments of human systems with technological Braidotti. In exploring Hayles’ concept of cognizers. Drawing upon a Deleuzian and nonconscious cognition, I will read the fol- Guattarian understanding of assemblage, lowing work of art as a cognitive assemblage Hayles develops the term ‘nonconscious’ in of a machine learning environment and will the direction of a cognitive assemblage to attempt to render tangible the intimate entan- account for a wider spectrum of “interactions glement of human and non-human systems. between human and non-human cognizers” Pandæmonium is an artwork by the Berlin- (Hayles 115). based duo PWR studio. It is displayed on the website of the Copenhagen-based Annual The point of emphasizing noncon- Reportt exhibition space and was part of an scious cognition is not to ignore exhibition in January 2018. The piece was the achievements of conscious not the only work shown in the exhibition but thought, often seen as the defining will be the focus here. The artists created characteristic of humans, but rather at an algorithm that runs as a text block down a more balanced and accurate view of the screen when entering the website. The human cognitive ecology that opens text block called Pandæmonium involves it to comparisons with other biological seemingly meaningful text and, according cognizers on the one hand and on the to the artists, refers to a dream sequence. other to cognitive capabilities of techni- The composition and visual language of the cal systems. Once we overcome the artwork Pandæmonium allows me to apply (mis)perceptions that humans are the the notion of nonconscious cognition to the only important or relevant cognizers on automated text code shown in the work. the planet, a wealth of new questions, Seen from afar, the piece could at first issues, and ethical considerations glance be read as a form of computer code. come into view (Hayles 10f.) It is a lively mechanism that is taking over the screen of the device. But this form of a can- Hayles challenges an anthropocentric nibalization of the screen can also be read perspective by deploying the notion of as a pressing statement of mechanical feel- non-human cognizers and is thus in line ings coming to expression. Pandæmonium’s

Figure 1: Screenshot from PWR studio’s Pandæmonium (2018), displayed on Annual Reportt’s website, http://annualreportt.com.

146 Tanja Wiehn: (UN)PREDICTABLE ACTS OF DATA ... reference to a dream sequence and a night- learning environments is demonstrated in the mare brings opposing notions regarding ma- glitches and unstructured associations of the chine learning to the table. The work produces work’s text code. The text code running down a layer at the top of the screen, which making the screen as output becomes the unread- its way relentlessly down to the bottom. able code that structures the system itself It adds a code of automated textual informa- behind the scenes. tion while simultaneously scraping off a layer In her book If... Then: Algorithmic Power of the obfuscated and seemingly unknowable and Politics, Bucher establishes the idea of machine learning mechanism. “eventfulness” within algorithmic procedures. She draws here upon a Whiteheadian notion Pandæmonium is a logistical that focuses on the becoming of an entity, nightmare. Pandæmonium is a rather than on its simple being. In Process pan-computational dream sequence. and Reality, Alfred North Whitehead (1978) Pandæmonium shows a future where suggests that the constitution of a being is digital networking has merged with always related to its process of becoming. fundamental reality. Everything is an Transferring this notion to algorithms, Bucher interface to something else. Everything substantially shifts the question from “what is connected to everything else. algorithms are to what they do as part of spe- Everything is inhabited by autonomous cific situations” (Bucher 49). I thus argue that agents acting according to opaque the progression of the text code, the becom- programming (PWR studio). ing of the piece of Pandæmonium, fosters this understanding of the eventfulness of The “autonomous agents” mentioned in algorithms. It is not an end result of data ana- the above quote evoke the magical elements lytics but is a text code in progress, stretching of artificial intelligence. The provocation lies from the top of the page to the bottom. This in part in the machine’s ability to produce visual operation of the artwork makes the apparently meaningful text, a fear related to text block appear as a form of nonconscious the aforementioned human exceptionalism, cognition. Pandæmonium is becoming, is a the idea of humans being the only creatures form of machine cognition acting out the dis- to which cognition can be ascribed (Hayles). play of its unexpressed desires. Reading the The visual language of the work displays the work through this analytical lens allows us to hidden and uncategorized feelings of the decipher the basic components of machine technological assemblage, simultaneously learning environments, such as their eventful a dream and a nightmare, both human and character (Bucher). The constitution of the non-human. The text block itself entails many work demonstrates the interwovenness of references to bones, metacarpal bones, the human and non-human cognizers. It is the mouth, the human body and its interaction piece acting as if it were unpredictable, as if with an uncertain and ever-changing envi- it were dreaming and creating unstructured ronment. A close reading of the work’s text images and thoughts. The work’s aesthetics block might reverse engineer the choices highlight a break from human-centered belief of the categorizations and tokenizing of an of cognition as exceptionally human (Hayles). input text-corpus. The unintelligible glitches in Pandæmonium, grammatical errors and mechanical failures, have not been erased or corrected. The obfuscated nature of machine

147 APRJA Volume 8, Issue 1, 2019

contest, especially in light of the monopolistic Material acts of power of the big social networks. The artists unpredictability Stephanie Kneissl and Max Lackner created various gadgets to shift the balance between user and algorithm within the technical as- One of the strengths of — but also one semblage. Their machinic instruments do not of the problems with — machine learning actually stop the algorithm, as the name of systems is their mundane yet invisible pres- the artwork suggests, but they change the ence. I noted above Wang’s example of their determination and categorization in the envi- implementation in predictive policing strate- ronment of social media sites on both ends: gies, but they are already in place in services the end of the data input and end of the data used on a daily basis, such as news feeds on output in form of, for example, advertise- social media platforms (Mackenzie). In these ments shown in the continuous becoming of cases, no access is given to how input data the newsfeed. is processed or how algorithms are trained. One means of intervening in the categoriza- We often assume that those systems tion and deciphering the seemingly objective are tools, made to connect and predictability in machine learning is through a inspire us, an infinite playground, an different kind of engagement with the system uncontrolled network that constantly itself. The second work to which I refer in reinvents itself. […] But social media is this paper fosters ideas of interventions in not neutral but highly biased and has algorithmic systems through acknowledge- an agenda of its own, with the goal of ment of their embeddedness in a material us to stay online and share as much environment. Through reading the following as possible. What we see on social work, Stop the Algorithm (2018), I seek to media is decided by algorithms that are rethink unpredictability in machine learning highly subjective, favouring popularity environments as a form of intervention. and mass instead of content. This The processing and capitalization of influences our opinions and thoughts. affects and attention spans are core patterns (Kneissl and Lackner). used within the big social networks such as Facebook and Instagram. This makes In Fig. 2, a small wind wheel is con- forms of manipulation and control of feel- nected to a pencil. Driven by a ventilator, ings through data-driven systems difficult to the pencil swipes through the newsfeed of

Figures 2 and 3: Stephanie Kneissl and Max Lackner, Stop the Algorithm (2018).

148 Tanja Wiehn: (UN)PREDICTABLE ACTS OF DATA ... an Instagram account on an iPhone. In the Rethinking cognition in a post-anthro- second gadget shown here (Fig. 3), the pen- pocentric manner becomes crucial for the cil is combined with other touching devices. second work too. An intervention into closed Connected to wheels, they move and stop systems of data harvesting such as social within the newsfeed – as if something caught networks cannot be easily realized. The little their attention – before scrolling further. The gadgets in the work substitute a human cog- movement of the touch on the screen leaves nizer within the technological assemblage traces of attention within the network and of this machine learning environment. The feeds new data into the network. When the artists deploy interaction with the physical little arms touch the surface, they seemingly conditions of the exhibition room — the wind show an interest in the content, as-if-human. wheels, scrolling pens, and touching devices These first two parts of the composition rely on the apparatus — to enable a form of un- on these procedures to maintain a sense of predictability. This material interaction is used unpredictability: interaction with the physical as a strategy for engaging with and challeng- surroundings take the place of a human cog- ing the newsfeed algorithm. The very idea of nizer in scrolling through the device. implementing material elements for creating The artists aim to create random inter- unpredictability in technological systems is action on the basis of material conditions in not new however. In cryptography, a distinc- the exhibition room in order to trick the algo- tion is made between true randomness and rithm. Tricking in this context means engag- pseudo-randomness (Gennaro). The latter is ing with the newsfeed algorithm through an called pseudo because it is a mathematically unpredictable method of scrolling. In Stop the constructed set of numbers so rendered as Algorithm, the platforms’ algorithms are not to appear random. In contrast, true random- actually stopped but are instead detrained. ness cannot be generated by computers but The work thus seeks to change one of the is often based upon the implementation of fundamental components of machine learn- a physical, material set of randomness (Doi ing environments. and Tadaki). A prominent case is the use of lava As the informational networks and lamps in the creation of true randomness feedback loops connecting us and in encryption systems. Lava lamps create our devices proliferate and deepen, an environment that coincidentally merges we can no longer afford the illusion a mixture of oil, water, and wax. By filming that consciousness alone steers our the lava lamps around the clock, the internet ships. How should we reimagine security company Cloudfare creates true contemporary cognitive ecologies so randomness through “the ever-changing that they become life-enhancing rather than aimed toward dysfunctionality and death for humans and nonhumans alike? Recognizing the role played by nonconscious cognitions in human/ technical hybrids and conceptualizing them as cognitive assemblages is of course not a complete answer, but it is a necessary component (Hayles 141). Figure 4: Lava lamp wall at the internet security company Cloudfare, image from WIRED Magazine.

149 APRJA Volume 8, Issue 1, 2019 arrangement of pixels to help create a superpowered cryptographic key” (Airhart). Conclusion: Machine With these measurements, this random- learning as cultural practice ness creates security keys that cannot be predicted by hackers. Only by implementing There can be no doubting the significance a material component can unpredictability be of gaining insight into the technological introduced. As a result, material randomness operations of data collection, databases, holds the potential to overcome the predict- and training sets for machine learning. ability of statistical reductionism in machine Methodologies of reverse engineering learning environments. Simultaneously, the (Bucher), ethnographic research (Seaver), principle of a form of randomness in mate- and critical code studies (Cox) possess great rial environments is becoming an eventful potential for constructing knowledge about coincidence for a data-driven world. In mo- machine learning technologies. In this paper, ments of data processing, these material I propose including the notion of eventful- acts of unpredictability enable a withdrawal ness and the idea of nonconscious cognition from pre-set categorization and classification of human and non-human environments and debunk the myth of knowledgeable data for describing machine learning systems. analytics. I believe that these concepts present op- As mentioned above, eventfulness is portunities for grappling with the potential a core asset of algorithmic environments ramifications of algorithmic processes. They – from data collection to the learning proce- furthermore simultaneously enable the ques- dures and training of machine learning that tioning of seemingly objective output and results in artificial intelligence (Alpaydin). The emphasize the necessity of human input. chosen works of art demonstrate an algorith- The works of art introduced in this paper help mically and material form of eventfulness us negotiate these concepts. Aesthetic and that is useful for dismantling myths, uncanny artistic representations can contribute to the feelings, and magical elements of machine discourse on machine learning ramifications, learning as well as for demonstrating their highlighting the blind spots of computational character as technological assemblages of determination with reference to algorithms as human and non-human environments. In culture and in culture (Seaver). Moreover, the Stop the Algorithm, it is the manifestation of sustainability of a mere technological insight the notion of eventfulness through interaction is questionable in the light of the ongoing with a material environment that disrupts the development of ever-more complex systems original purpose and capitalization of data and their interwovenness with capitalist and collection. Pandæmonium, in contrast, visu- political structures of oppression and social alizes a process of composition in machine forms of control (Wang). learning. The visual language of the works of

art provide a framework for expanding a vocabulary of machine learning and introduce creative interventions into algorithmic systems. Pandæmonium questions the meaningfulness of computational processing in the text code that it displays on the exhibition space’s website, while the notion of material

150 Tanja Wiehn: (UN)PREDICTABLE ACTS OF DATA ... resistance in Stop the Algorithm emphasizes the limits of engagement with highly complex Works cited technological systems that cannot be easily Airhart, Ellen. “How a Bunch of Lava decoded or dismantled. The discussion of Lamps Protect Us From Hackers.” WIRED possible means of engaging with and for- Magazine, 28 July 2018, https://www.wired. mulating criticism towards technologies can com/story/cloudflare-lava-lamps-protect- be accompanied by discursive and aesthetic from-hackers/. Accessed 27 February 2019. forms that demystify images of black boxes and debunk the hype of artificial intelligence. Alpaydin, Ethem. Machine Learning. MIT Reading machine learning as eventful – as a Press, 2016. concept of becoming that entails subjective categorizations and entails nonconscious Annual Reportt. “Pændamonium.” 17. cognition – transforms technology from being Jan. 2018. http://www.annualreportt.com. a neutral instrument into a cultural practice. Accessed 02 March 2019. Machine learning systems are meant to guarantee a mode of predictability through Braidotti, Rosi. The Posthuman. Polity the mathematical reduction of complexities. Press, 2013. Therein lies potential for the unpredictability of a material embeddedness to recognize the Bucher, Taina. If... Then: Algorithmic Power ramifications of machine learning systems and Politics. Oxford University Press, 2018. and challenge the knowledgeability of their output. Cox, Geoff, and Alex McLean. Speaking Code: Coding As Aesthetic and Political Expression. MIT Press, 2013.

Crawford, Kate, and Vladen Joler. “Anatomy of an AI system. An anatomical case study of the Amazon echo as artificial intelligence systems made of human labor.” 2018. https://anatomyof.ai. Accessed 28 February 2019.

Doi, Norihisa, and Kohtaro Tadaki. “Cryptography and Algorithmic Randomness.” Theory of Computing Systems, vol. 56, no. 3, April 2015, pp. 544–80, doi:10.1007/s00224-014-9545-9.

Elish, Madeleine Clare, and danah boyd. “Situating methods in the magic of Big Data and AI, Communication Monographs”, 85:1, 2017, pp. 57-80, doi: 10.1080/03637751.2017.1375130.

151 APRJA Volume 8, Issue 1, 2019

Gennaro, Rosario. “Randomness in Seaver, Nick. “Algorithms as Culture: Some Cryptography.” IEEE Security & Privacy, vol. tactics for the ethnography of algorithmic 4, no. 2, 2006, pp. 64–67. systems.” Big Data & Society, vol. 4, no. 2, 2017. Gillespie, Tarleton. “The Relevance of Algorithms.” Media Technologies: Essays Wang, Jackie. Carceral Capitalism. on Communication, Materiality, and Society, Semiotext(e), 2017. edited by Tarleton Gillespie, Pablo J. Boczkowski, and Kirsten A. Foot, 2014, pp. Whitehead, Adrian North, David R. Griffin, 167-193. and Donald W. Sherburne. Process and Reality: An Essay in Cosmology. Free Hayles, Katherine N. Unthought: The Power Press, 1978. of The Cognitive Nonconscious. University of Chicago Press. 2017.

Hayles, Katherine N. How We Became Posthuman: Virtual Bodies in Cybernetics, Literature, and Informatics. University of Chicago Press, 1999.

Kitchin, Rob. “Thinking critically about and researching algorithms.” Information, Communication & Society, 20:1, 2017, pp. 14-29.

Kneissl, Stephanie, and Max Lackner. “Stop the Algorithm.” 2018. https://stephaniek- neissl.com/reset-social-media. Accessed 02 March 2019.

Mackenzie, Adrian. Machine Learners: Archaeology of a Data Practice. MIT Press, 2017.

Pasquinelli, Matteo. “The Undetection of the New: An Overview of Machine Learning Limitations and Their Social Impact”, University of Cambridge, UK, 14 January 2019. Keynote at workshop.

PWR Studio. Artists Webpage: https://pwr. site. Accessed 02 March 2019.

152 153

SEEING THINGS Mitra Azar

POV-MATTER, CINEMATIC POV AND ALGORITHMIC POV BETWEEN AFFECTS AND UMWELTEN

Abstract

The essay refers to affect theory as a conceptual toolbox to draw a genealogy of POV (Point of View) that goes from the formation of the first organic POV to the reinvention of POV by the cinematic apparatus up to the latest development of algorithmic POV in machine vision and AI. The essay engages with Bergson’s conviction that there’s no perception without affection, and tests it against a phenomenological, cinematic and machinic notion of POV. To do so, the essay introduces what the German biologist Jacob von Uexküll has called Umwelt — the ecological niche emerging from the affordances between organisms, space, and (when applicable) technology. Furthermore, fundamental categories of both phenomenology and psycho-analysis are put at work in relation to cinematic POV and to the algorithmic POV produced by Generative Adversarial Networks (GANs), which seems to re-invent the relationship between seeing/seen (Merleau-Ponty) and eye/gaze (Lacan). This re-invention confirms the category of Umwelt and affect as markers for understanding the transformation between a phenomenological, cinematic and algorithmic notion of POV.

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Mitra Azar: POV-MATTER, CINEMATIC POV...

Could a machine think? inorganic POV-matter. There exists, in other Could it be in pain? words, a form of inorganic POV-matter char- (Wittgenstein, Philosophical Grammar) acterized by orientation but not by the gap associated with orientation when orientation relates to an organic living agent. Building from these first considerations, I. POV, affects and the essay asks what happens to affect when Umwelten between phe- POV becomes associated to more complex nomenology, cinema and forms of life equipped with a visual sensing apparatus, such as in the case of human machine vision beings. Furthermore, what happens to affect when POV becomes technological and as- In this essay, I propose affect theory as sociated to a visual and technological appa- the conceptual framework for analyzing ratus? Although inorganic POV-matter lacks the various regimes of visibility related to an organic form of perception and affect, hu- the formation of different forms of ‘Point man beings have been able to turn inorganic of View’ (POV). The notion of POV is here POV-matter into a technological apparatus approached as a phenomenological feature capable of harnessing both perception (and that in its fundamental understanding goes by doing so the organic POV-matter associ- beyond the visual connotation it has been ated to it), and the affective gap itself. The usually associated with, and can be defined technological apparatus can be referred to as ‘orientation’ tout court. In this sense, mat- as a POV-apparatus because it is constituted ter is oriented despite the organic/inorganic by POV technologies of vision, which means divide, and can be defined as POV-matter, technologies designed to harness phenom- once framed by the concept of orientation. enological POV. Via analog POV technolo- Orientation manifests already at an inorganic gies of vision such as painting, photography level, right down to the spinning of particles and cinema, the POV-apparatus attempts to inside atoms that function independently harness affection via perception. Via what st from organic POVs and associated affects. Mark B. N. Hansen calls ‘21 century media’ POV-matter takes the form of inorganic POV- (266), POV-apparatus attempts to harness matter and organic POV-matter, and the no- affection directly, and turns into a type of tion of affect is presented as the conceptual media which gains the capability of directly tool to understand the differences between attacking the affective gap at the core of or- the two. ganic POV-matter, thus generating inorganic What happens to inorganic POV-matter algorithmic POV capable of mimicking the since the formation of the first organic POV- functioning of organic POV. matter emerging from the pre-biotic soup?[1] In summary, the essay aims to inves- There, the first proto-stable organic forms of tigate what happens to affect in relation to life produce both an orientation and a gap ap- both the (pre)phenomenological and the pearing between the immediate action/reac- technological (specifically cinematic and tion schema characterizing instead inorganic algorithmic) definitions of POV. By question- POV-matter — as demonstrated by Gilles ing the becoming of different forms of POV Deleuze is his first book on cinema, Cinéma and their corresponding affects, the essay 1: The Movement-Image. The notion of affect aims to engage with the French philosopher as ‘gap’ discriminates between organic and Henri Bergson’s conviction that there’s no

157 APRJA Volume 8, Issue 1, 2019 perception without affection (Bergson 17- 76). To do so, the essay introduces what II. From inorganic POV- the German biologist Jacob von Uexküll has matter to affects, organic called Umwelt (“The new concept of Umwelt” POV-matter and Umwelten 111-123), understood as the affordances generated by the interaction between organisms, space, and – when applicable, such as In a fundamental way, POV can be understood in the case of human beings — technology: as orientation. After the Big Bang, fundamen- “everything a subject perceives belongs to its tal blocks of matter organized themselves by perception world [Merkwelt], and everything producing orientations, technically referred it produces, to its effect world [Wirkwelt]. to as ‘spins’ within electromagnetic fields.[2] These two worlds, of perception and pro- In other words, despite the organic/inorganic duction of effects, form one closed unit, the divide, matter is always orientational: matter environment [Umwelt]” (Uexküll, A Foray is directional, and can, hence, be character- into the Worlds of Animals and Humans 42). ized as ‘POV-matter’. Inorganic POV-matter Furthermore, the mutation of affect in relation indicates that orientation manifests itself to the transformation between a phenomeno- in the inorganic world, and only afterwards logical, cinematic, and algorithmic notion of turns into the orientation of organic POV- POV are finally approached in relation to matter. Organic POV-matter exists from the reversibility between seeing and seen the simplest organisms deprived of a visual (Merleau-Ponty, The Visible and the Invisible sensing apparatus, but yet capable to orient 130-155) and the relation between the notion via other senses (such as the ‘tick’[3]) to of eye and gaze (Lacan 67-79). These fun- more complex living beings developing the damental categories of both phenomenology sense of vision, among them human beings. and psycho-analysis (Merleau-Ponty, The In contrast to inorganic POV-matter, organic Visible and the Invisible 130-155; Lacan 67- POV-matter perceives. Thinking generally of 79), are put at work in relation to a specific matter as POV-matter is not trivial, because form of algorithmic POV produced by Deep it allows us to not only think of orientation Convoluted Generative Adversarial Networks or POV as pivotal ontological and phenom- (DCGANs, or simply GANs), which seems to enological concepts, but also and more re-invent the relationship between seeing/ importantly because it allows to re-articulate seen and eye/gaze. This re-invention invites the agential relation between the organic and rethinking the category of Umwelt and affect the inorganic. In this non-trivial sense, a POV as markers for understanding the transforma- is not something that comes exclusively with tion between a phenomenological, cinematic the human or with the technological. Rather, and algorithmic notion of POV. POV as the production of an orientation has cosmological origins which not only predates the appearance of human POV but also the appearance of the first form of simple organic POVs in the prebiotic soup. What then happens to orientation when orientation becomes embedded into organic living being inhabiting a Umwelt, understood as the ecological niche of a living organism? If inorganic matter is, in a way, always

158 Mitra Azar: POV-MATTER, CINEMATIC POV... inorganic POV-matter, Umwelt is in a way organic POVs, and by doing so they manage always POV-Umwelt, because it always to overlap and ultimately bridle human POVs refers to an oriented subjective experience and their Umwelten. Inorganic POV-matter (although not forcibly human). Every Umwelt becomes technological POV-matter, or POV- “has its own spatial and temporal dimensions” apparatus. A POV-apparatus of analogue (Uexküll, A Foray into the Worlds of Animals technologies (such as painting, photography and Humans 49), and “ [they] intersect in and cinema) is a form of visual governmental- many ways without disturbing each other” ity developed to ultimately harness humans’ (Uexküll, “The New Concept of Umwelt” 117). POV and subsume affection via perception. A As Uexküll further explains: “every action […] digital or algorithmic POV-apparatus (such as that consists of perception and operation im- the one enforced by artificial intelligence and prints its meaning on the meaningless object GANs) bypasses perception, and operates and thereby makes it into a subject-related at the affective gap to predict the emerging meaning-carrier in the respective Umwelt affections and related affordances of organic (subjective universe)” (Uexküll, “The Theory POVs within a given Umwelt. of Meaning” 31). Otherwise expressed, when (according to different evolutionary survival criteria) inorganic POV-matter develops into organic forms that orient themselves in III. Organic POV-matter space, the various forms of POV-matter pro- and affects duce their unique Umwelten — of which the organism’s regime of visibility represent the In this section, I argue that organic POV- visual counterpart. Orientation, or POV — in matter is defined as a form of orientation the form of a perceptive agent — produces co-emerging with the affective gap during Umwelt. At the same time, in turn, Umwelt the formation of the first proto-stable organic produces POV. Furthermore, the formation of forms of life. As mentioned earlier, according the organism’s Umwelt is specifically related to Bergson, there is no perception without to the affective quality of organic POV-matter, affection (17-76). In other words, there is an because it is in the affective gap between intrinsic relation between organic POV and action and reaction that an orientation is pro- affects. Bergson’s statement is adopted by duced, and together with it the beginning of Deleuze, in Cinéma I: The Movement-Image, an Umwelt. In this sense, affect establishes to categorize cinematic images: “The thing itself as the inner engine of orientation, and is the image as it is in itself, as it is related orientation as the inner engine of Umwelt. to all the other images to whose action it Inorganic POV-matter turns into the completely submits and on which it reacts technological and cultural expression of immediately. But the perception of the thing complex organic POVs such as human is the same image related to another spe- beings, and becomes a cultural and tech- cial image which frames it, and which only nological product. In their technological retains a partial action from it” (Deleuze, The instantiation, spins and fields of atoms’ par- Movement-Image 64). ticles are geared towards the construction of From a universe composed by images technologies able to reproduce the regime that “act and react on all their facets and in all of visibility correspondent to the Umwelt of their parts” (Deleuze, The Movement-Image a given organism. Inorganic technological 61), Deleuze unfolds the image-perception to POVs, attempt to mimic the functioning of name a type of image which “only receive[s]

159 APRJA Volume 8, Issue 1, 2019 actions on one facet or in certain parts and interval is not merely defined by the spe- only execute reactions by and in other parts. cialization of the two limits facets, perceptive […] the image reflected by a living image and active. There’s an in-between. Affection is precisely what will be called perception” is what occupies the interval, what occupies (Deleuze, The Movement-Image 62). In it without filling it in or filling it up” (Deleuze, other words, image-perception is the im- The Image-Movement 65). This is how the age reflected by organic POV-matter. This difference between inorganic and organic “image reflected” coincides with the regime POV-matter comes into being. Organic POV of visibility of a given Umwelt for a given produces a gap between the continuous ac- organic POV-matter. Nevertheless, how tion-reaction characterizing the functioning of does inorganic POV-matter turn into organic inorganic POV-matter. The gap is where the POV-matter? action-reaction circuit is interrupted. “Even at Affect — understood as the production the level of the most elementary living beings of a gap between action and reaction — is one would have to imagine micro-intervals. the catalyst for the production of affects and Smaller and smaller intervals between more for the formation of organic POV-matter(s) and more rapid movements” (Deleuze, The and their respective Umwelten. Umwelt is Movement-Image 71). here understood as the selective interaction Once reduced to its essence, the affec- between organic and inorganic POV-matter tive gap can be intended as the figure for the (e.g. a species and their environments, also coincidence between organic POV-matter considered in their inorganic composition). and inorganic POV-matter, or, for simplicity, Orientation is what emerges together with between subject and object. This is what affects from the gap, producing both a POV happens, for instance, when the subject is and a Umwelt. In the affective gap between automatically driven to bring attention to action and reaction POV-matter orients itself their body — such as in the case of hunger. according to the elaboration of the input (ac- Hunger happens in the gap between action tion) from the surrounding environment and (the action of the environment on the subject towards an output or reaction — this circuit in the form of the presence of food) and reac- being what Uexküll refers to when he says tion (the movement towards food by the sub- that “everything a subject perceives belongs ject) because of a moment of self-perception to its perception world [Merkwelt], and every- during which the subject becomes the object thing it produces, to its effect world [Wirkwelt]. of their own attention, before enacting the These two worlds, of perception and produc- appropriate reaction. Self-perception sprouts tion of effects, form one closed unit, the from the affective gap, short-circuiting the environment [Umwelt]” (Uexküll, A Foray into relation subject/object. This is interesting the Worlds of Animals and Humans 42). because cinematic POV produces some- Orientation is the fundamental resolu- thing similar by perceptively overlapping tion of organic POV-matter(s) to their specific the subject (audience/actor/director) with Umwelten — and emerges from the affective the object (camera/screen). By doing so, gap at the foundation of organic POV-matters cinematic POV manages to access the affec- to harness the affordances appearing tive gap at the core of organic POV-matter between organic POV-matter(s) and their or phenomenological POVs. To put it simply, Umwelten. Affects fill the seemingly empty the functioning of cinematic POV mimics the gap between action and reaction defining way the affective gap functions at the level the emergency of organic POV-matter: “the of phenomenological POV, and subsumes

160 Mitra Azar: POV-MATTER, CINEMATIC POV... the affective gap via perception. Generally the seamless overlapping between camera, speaking, analogue POV technologies of actor’s body and spectator’s body, thus vision such as painting, cinema and photog- producing a form of seamless overlapping raphy harness organic POV and the affective between the human and the technological. If gap on the side of perception, aiming at technology and human have been overlap- shrinking the distance between technological ping since ‘hominization’,[4] with cinematic and organic POV. Because of movement, the POV the overlapping is seamless because cinematic figure of POV operates this shrink- of the capacity of cinema to reproduce ing flawlessly and in doing so manages to movement, a quality other medium before short-circuit subject and object by giving to cinema couldn’t attend to. Cinematic POV the audience the illusionary (self-) perception articulates the relation between the specta- of breaking through the screen and of mov- tor’s POV intended as the phenomenological ing inside of it on behalf of the character. As orientation produced by an embodied human a consequence, cinematic POV manages to agent in a physical space and the regime of harness the affective gap via the subject’s visibility produced by the cinematic machine. activity of self-perception which emanates The very collapse and overlapping between from it. Digital and algorithmic POV technolo- the embodied agent’s POV and the regime gies in the form of AI and machine sensing of visibility generated by the cinematic ma- — what Hansen refers to as 21st century chine is the main feature of the cinematic media — instead, attempt to harness organic technics of POV. Cinematic POV harnesses POV and perception directly on the side of the embodied POV on the side of perception, affection. Before looking at the role of the and aims at shrinking the distance between ‘gap’ in relation to machinic POVs, I will now technology and body, or between object and discuss the form that inorganic POV-matter subject. takes once it becomes cinematic apparatus, The possibility of generating the seam- and more specifically, cinematic POV. less overlapping between camera, actor’s body and spectator’s POV gives to cinematic POV the ability to intervene into the affective gap at the core of organic POV by re-articu- IV. POV in cinema and phe- lating the relation between the phenomeno- nomenology: Reversibility logical categories of seeing/seen and eye/ between seeing/seen and gaze. From a phenomenological perspective, one of the main features of human POV split between eye/gaze is that of expressing a “worldly sensitivity” (Hansen 266) visually characterized by the In the field of cinema, POV is an acronym reversibility between the coupling of seeing/ which refers to a type of image that allows seen: I’m seeing the world around me but I’m the viewer to see what the character sees also seen simultaneously by others, and this from the character’s perspective (or orienta- reversibility (together with the reversibility tion) (Braningan 55). POV cinematic images between touching/touched) is what defines simulate the movement of an actor within a my being in the world, my embeddedness space, creating a sense of continuity between into an intersubjective world: “the seer is viewers and what is viewed, as if viewers are caught up in what he sees, it is still himself ‘embodied’ in the images they’re looking at. In he sees” (Merleau-Ponty, The Visible and this sense, cinematic POV images generate the Invisible 139). This coupling is molded

161 APRJA Volume 8, Issue 1, 2019 on touching/touched, for which “I can identify 130-155) between seeing/seen — I’m seeing the hand touched as the same one which will the actor/character moving in the scene and in a moment be touching […]. The body [… I’m seen (perceived) by myself as the very tries] to touch itself while being touched and actor/character. This is the way in which initiates a kind of ‘reflection’” (Merleau-Ponty, POV-apparatus functions — by manipulating Phenomenology of Perception 106). affects via perception. Thus, the immersivity The horizontal relation identified by produced by the figure of cinematic POV is Maurice Merleau-Ponty between seeing/ harnessed to disseminate ideology and seen turns vertical once framed by the enforce consumer’s behaviors. At the same originary asymmetry between the eye and time, the collapse between the eye and the the gaze defined in the context of Lacanian gaze operated by cinematic POV re-arranges psychoanalysis: “I’m seeing only from one the relation between the coupling seeing/ point [an eye, or a POV], but in my existence seen and can produce the emergency of new I’m looked at [by the gaze] from all side” forms of political agency. For example, POV (Lacan 72). Cinema does something pretty mobile phone images recorded in the context interesting to these phenomenological and of social unrest and protests and uploaded psycho-analytical categories: if it seems pos- online (such as in the case of the so-called sible to say that cinema enforces the vertical ‘Arab Spring’), became the available format relation between the eye and the gaze – the for revolutionary subjects to perform their eye being the eye of the spectator and the political agency. The Egyptian Revolution gaze being the director’s “all-seeing” (Lacan has been an experimental ground for such 75) — in the case of the cinematic technic of grassroot emancipatory media practices.[6] POV, eye and gaze collapse into each other. Finally, after having attempted to pro- Thus, POV re-establishes the horizontal vide an understanding of the functioning of reversibility between seeing and seen (in this phenomenological POV and cinematic POV, case between the seeing/seen of the viewer I now turn to how the phenomenological and director via the mediation of the actor). intertwining of seeing/seen and eye/gaze is This reversibility of POV is at the root of its played out in the context of algorithmic POVs. pharmacological nature, of its capacity of be- This will be the focus of the last two sections ing both poison and cure.[5] In mainstream of this article, where I will try to unpack the media, this capacity has been deployed functioning of algorithmic POVs and their to “disseminate ideology” and to enforce interaction with phenomenological POV and “consumer’s behaviors” the way Adorno and affects via two case studies: one referring to Horkheimer explicate in their 1944 Dialectic the “Arkangel” episode in the fourth series of of Enlightenment. The reversibility between Black Mirror, the other to the operationality eye and gaze produced by cinematic POV of GANs (Generative Adversarial Networks). generates perceptive immersion and embodiment beyond the surface of the screen. The audience falls into the screen, embodied into the character’s body, and the spectator’s V. POV and algorithms self-perception — understood as the state emerging from the affective gap the way it has “Now object perceives me” stated Paul Klee been described in the above section III — can in his diaries, as cited by French philoso- be directed by engineering the intertwining pher Paul Virilio in the opening of his Vision (Merleau-Ponty, The Visible and the Invisible Machine (1994), somehow prophetically

162 Mitra Azar: POV-MATTER, CINEMATIC POV... envisioning a world of objects that learn how which technically separates the actors from to see — and to “sense” — the surrounding the spectators.[7] space and the bodies occupying it. New tech- With algorithmic technologies the pro- nologies of vision oriented towards new forms cess looks similar but inverted. They produce of ‘data-veillance’ (Clarke 254-271) such as the overlapping between the human and the machine vision, seem to give technological machine by inserting the machine into the consistency to Klee’s intuition. Moreover, human, and not vice-versa, as in the case these technologies seem to confirm the a- of cinema. To do so, they attempt to access symmetry Lacan locates at the very heart human POV by accessing the very affective of our phenomenological intertwining with gap where it emerges from: first breaching the world, making visible the encompass- through the screen of the body, and secondly ing visual power of the (technological) gaze extracting worldly data beyond the human against the localized and punctual vision of conscious threshold. Thus, machines ac- the (human) eye. This a-symmetry is currently cess vicariously a bodily dimension, while taking new forms that extend the capability of humans are exposed to a quantified version the gaze to all-seeing: for example, tracking of their very affective fabric, which — data- technologies based on AI aim at quantifying fied — contributes to the constitution of new a number of qualitative inputs that go from fa- forms of human-machinic Umwelten with cial features and facial expressions to breath- complex political implications. One of the ing pattern and heart beats — inputs that are most significant change in relation to these the embodied manifestations of the affective new forms of Umwelt, consists in the fact that patterns emerging from the affective gap. By the affordances between the human and the quantifying them, algorithmic technologies surrounding space are technically anticipated points at accessing the very affective gap by a capture which re-defines affordances between action and reaction defining organic as such and which claims to design them in POV-matter — or its emergency as an affec- ways that fulfil the subject’s expectations bet- tive, embodied POV. The technological gaze ter than the subject’s agency could possibly tries, thus, to vicariously access the eye by achieve. This is what happens in relation to accessing the affective body right at the very the creation of POV data-doubles and the moment where it emerges as a POV. In this consequent formation of filter bubbles within sense, new technologies of vision based on social networks based on the anticipation AI — part of what Hansen calls 21st century of users’ affects[8] emerging from the af- media — attempt to locate themselves at the fective gap. “Anticipation made possible by very gap where the formation of a worldly algorithms […] become increasingly active, sensitivity, or perception, emerges. Thus, to the point of displacing or marginalizing algorithmic POV technologies invite us to active directedness” (Hui 144). Algorithmic re-think the notion of affect and Umwelt. In POVs manufacture an “automatic future, in cinematic POV the overlapping between which our selections will be to a large extent, human and technology produces the over- if not completely, predefined according to a lapping between the regime of visibility of an specific schema and index” (Hui 150). This is embodied POV and the regime of visibility what happens in “Arkangel” directed by Jody of the cinematic machine behind the surface Foster — the second episode of the fourth of the screen: the audience feels inside the season of Black Mirror — where a mother screen, inside the cinematic machine, and implants her daughter with a device which breaks through the so called fourth wall allows her to see what she is looking at from

163 APRJA Volume 8, Issue 1, 2019 a POV perspective in real-time. The mother is provided with a user-friendly interface that VI. Towards a phenomeno- gives her the possibility of deleting images logical understanding of from her daughter’s sight stream that she GANs thinks could traumatize her – such as images of conflicts or violence. As a consequence of this technology, the daughter grows in- Are there any other ways in which algorith- capable of recognizing conflicts or violence mic POVs attempt to reproduce the affective and indeed becomes incapable of behaving gap of phenomenological POVs? The op- appropriately in such circumstances. In erationality of GANs provides an exemplary “Arkangel”, algorithms implement a reality- case study of new forms in which algorithmic bubble around children, one perceptively POVs try to approximate the functioning of different from the material reality lived by oth- phenomenological POVs. GANs mimic the ers. In this example, the direct capturing of very intertwining between seeing/seen and affects by algorithmic technologies goes be- eye/gaze to which Merleau-Ponty and Lacan yond perception in the sense that perception refer when pointing to the enworlding of phe- appears here as a malleable context to fully nomenological POVs. GANs are the most engineer, to anticipate and design affects. advanced form of algorithmic simulation of a These algorithmic technologies attempt phenomenological enworlding, as I hope to to anticipate the affordances that define the illustrate during the course of this last sec- relation between organic POVs and their tion. What exactly is a GAN, and why am I Umwelten by designing POV-data doubles saying that GANs provide an example of a retro-actively producing the affective subjects new radical way to attempt the simulation of they’re generated from. Algorithmic POV the affective gap at the core of phenomeno- handles the relation between seeing/seen by logical POV? directly harnessing the gap between the two: GANs are a form of unsupervised ma- anticipating the subject’s affordances and chine learning able to access raw data from projecting back into the subject an algorithmic the world and to build an understanding of POV data-double which informs the way the them without the mediation of any linguistic subject operates within her own Umwelt. This labeling applied by humans, or “mechanical is the only way algorithmic POV technologies turks” (Wikipedia), which tag huge data- can currently and vicariously access affects sets of images and prepare them to train despite their incapability of producing the “supervised machine learning algorithms” gap at the core of organic POVs. In doing so, (Wikipedia). GANs build an understanding the (algorithmic) gaze attempts to access the of raw data by establishing an antagonistic (phenomenological) eye and to control it by relation between two neural networks, one vicariously accessing the affective gap, and generating data (generator), the other dis- bypassing perception. criminating the data generated on the basis of a model (discriminator). In a sense, generator and discriminator constitute each other through an algorithmic exchange that closely resembles both the intertwining of the coupling seeing/seen and the split between the eye and the gaze. Generator and discriminator see each other

164 Mitra Azar: POV-MATTER, CINEMATIC POV... and in doing so establish each other’s (in- POV-matter. In brief, the phenomenological organic and algorithmic) POV, while at the intertwining between organic POV-matter(s) same time enacting the distinctive roles of the is rooted in the affective gap from where self- eye (generator) and of the gaze (discrimina- perception – understood as the coincidence tor). At the same time, a form of technological between subject and object – emerges. The Umwelt appears as the place of the emer- intertwining between generator and discrimi- gency of GANs’ affordances. GANs’ Umwelt nator in a GAN, in contrast, is rooted in the emerges at the intersection of generator algorithmic gap emerging from the mismatch- and discriminator’s affordances, and takes ing between generator and discriminator’s the form of what is technically addressed as POVs. In other words, if in phenomenological latent space. Latent space is defined as the POV the coincidence between subject and space where a “generative network learns object emerges from the affective gap, in to map […] a particular data distribution of GANs the coincidence between subject and interest, while the discriminative network dis- object happens when the gap is closed and criminates between instances from the true generator and discriminator POV coincides. data distribution and candidates produced by This happens when the generator generates the generator” (Wikipedia). The latent space an image which fools the discriminator and can be addressed as an algorithmic screen closely matches the model. where an algorithmic recognition happens — If, according to Bergson, there’s no a recognition based on the interplay between perception without affection, when it comes generator and discriminator’s different algo- to GANs and algorithmic POVs in general, rithmic POVs. This algorithmic recognition we can only metaphorically refer to percep- produces a form of algorithmic perception tion — as much as to POVs and Umwelt. where the complexity of the intertwining Nevertheless, GANs re-invent the relation between embodied POVs is reduced to a between POV, affects and Umwelt, and do task-oriented statistical capability of pattern so by deploying algorithms that mimic the recognition — a feature typical of AI in gen- phenomenological intertwining that is charac- eral, according to Matteo Pasquinelli (2017). teristic of embodied POV. If GANs reproduce The latent space is the algorithmic screen the phenomenological intertwining between where a form of algorithmic gap emerges be- seeing/seen and eye/gaze, other types of al- tween the images produced by the genera- gorithmic POVs access the body by harness- tor, and the model on which the discriminator ing the affective gap from within, constituting has been calibrated, when generator and POV data-doubles and retro-projecting them discriminator’s POVs mismatch. There’s no on the affective subject they have been gen- affect in this gap which emerges with the erated from, bypassing perception (as in the emergencies of the intertwining between case of Arkangel). These forms of intensive generator and discriminator’s mismatching or molecular algorithmic capture of the affec- POVs. Furthermore, the gap is filled once the tive gap, differ from the functioning of earlier generator and discriminator’s POV perfectly media which, instead, operate at a molar overlap – which happens when the genera- level, and subsume affection into percep- tor produces a closer enough version of the tion. Cinematic POV does so by collapsing model capable of fooling the discriminator. the eye of the audience and the gaze of the In this algorithmic intertwining, subject and director via the body of the actor-character. object coincide when the gap is closed, not The common feature between cinematic and when it appears such as in the case of organic algorithmic POV consists in the shrinking of

165 APRJA Volume 8, Issue 1, 2019 the distance between the eye and the gaze. This shrinking happens in parallel to the Notes shrinking between inorganic POV-matter and organic POV-matter. Once the former turns [1] ‘Pre-biotic soup’ is an expression related into technological inorganic POV-matter, it to the unstable state of matter in which turns into a POV-apparatus which tends to chemical compounds were about to gener- operationalize the reduction of the distance ate the conditions for life over the planet between the interface and the body, and as Earth at a certain stage of its evolution a consequence between the (technological) (CERN). gaze and the (phenomenological) eye. This is the common feature between cinematic POV [2] Spins describe electrons spinning around and algorithmic POV, despite the fact that nuclei of protons and neutrons forming the they operate this reduction differently, the for- first atoms some 380,000 years after the Big mer subsuming affection via perception, the Bang (Cern.com). latter subsuming affection directly, bypassing perception. By doing so, the POV-apparatus [3] The ‘tick’ is a famous example empha- produces both new regimes of visibility — sized by both Uexküll and Deleuze. See with related affects and Umwelten — and Uexküll’s A Foray into the Worlds of Animals new regimes of truth. From the differences and Humans, and Deleuze’s L’Abécédaire between a phenomenological, cinematic and de Gilles Deleuze. algorithmic form of POV, my argument moves towards defining the techno-phenomenologi- [4] See Stiegler’s Technics and Time I. The cal conditions for the emergency of both new Fault of Epimetheus. regimes of ‘(post-)truth’,[9] and a new form of visual governmentality, which I refer to as [5] The pharmacological nature of technol- ‘POV-opticon’.[10] ogy is highlighted by Bernard Stiegler throughout his philosophy. See Stiegler’s “Pharmacology of Desire”.

[6] I have been writing about media activism in Egypt during the revolution while being based in Cairo and collaborating with activist collectives. See Azar,“The Revolution will not be Tweeted (?)”.

[7] The ‘fourth wall’ mirrors the a-symmetry Lacan refers to the relation between the eye and the gaze: if the spectator can see the actors behind the fourth wall, the actors act without seeing behind the fourth wall that somebody is seeing them. And yet the seeing of the spectator is driven by the seeing of the actors.

166 [8] I have previously explored this topic in relation to the production of a new type of Works cited selfie aesthetic in a paper published last year in APRJA; see Azar’s “The Algorithmic Adorno, Theodor, and Max Horkheimer. Facial Image (AFI) and the relation between Dialectic of Enlightment. Continuum, 1993. truth value and money value”. Another example of these forms of prehension is a new Azar, Mitra. “From Panopticon to POV- MIT prototype that allow users to control opticon: drive to visibility and games basic functions of a computer through an of truth”. Academia.edu, https://www. ergonomic wearable interface able to record academia.edu/37648284/From_Panopticon_ the micro-movements of the subject’s lower to_POV-opticon_drive_to_visibility_and_ jaw as a way to infer brain activity - the jaw games_of_truth. moves slightly when the brain formulate a decision even without the production of a Azar, Mitra. “The Algorithmic Facial verbal utterance – and before the aware- Image (AFI) and the Relation between ness of the subject: “Electrodes on the face Truth Value and Money Value.” APRJA, and jaw pick up otherwise undetectable 8.1, June 2018, http://www.aprja.net/ neuromuscular signals triggered by internal the-algorithmic-facial-image/. verbalizations”. See Herdesty’s “Computer system transcribes words users ‘speak Azar, Mitra. “The Revolution will not be silently’”. Tweeted (?): Media-attivismo al Cairo durante la post-rivoluzione.” Alfabeta, May [9] See Azar’s “The Algorithmic Facial Image 2012. (AFI) and the relation between truth value and money value”. The paper is forthcoming Bergson, Henri. Matter and Memory. as a chapter of G. Lovink and D. Della Ratta George Allen and Unwin, 1911. (eds), Online Self. Palgrave and McMillan, 2021. Braningan, Edward. POV in The Cinema. A Theory of Narration and Subjectivity in [10] I’ve started investigating the relation Classical Film. De Gruyter Mouton, 1984. between POV, regimes of truth and games of truth in a paper presented at the 2018 Clarke, Roger. “Information Technology and After Post-Truth conference in Barcelona. Dataveillance”. Commun. ACM 31, 5, 1988, See Azar, “From Panopticon to POV- pp. 498-512. opticon: drive to visibility and games of truth” (draft version of the paper can be found on Deleuze, Gilles. Cinéma 1. The Movement- academia.edu). Image. University of Minnesota Press, 1986.

Emmeche, C. “Can robots have an Umwelt”. Semiotica vol. 134 (issue 1/4), 2001, pp. 653-693.

167 APRJA Volume 8, Issue 1, 2019

Deleuze, Gilles. L’Abécédaire de Gilles Pasquinelli, Matteo. “Machines that Morph Deleuze, by Pierre-André Boutang, Logic: Neural Networks and the Distorted interview with Claire Parnet, Éditions Automation of Intelligence as Statistical Montparnasse, 2004. Inference”. Glass Bead, 1: “Logic Gate: The Politics of the Artifactual Mind”, 2017. Foster, Jodi, Dir. “Arkangel,” 2nd episode, Ponsoldt, James. The Circle. STXfilms, Black Mirror series, 4th season. Netflix, Europacorp, 2017. 1997. Spin. Cern.com, https://home.cern/science/ Herdesty, Larry. “Computer system physics/early-universe. transcribes words users speak silently”. MIT News Office, http://news.mit.edu/2018/ Stiegler, Bernard. Technics and Time I: The computer-system-transcribes-words-users- Fault of Epimetheus. Stanford University speak-silently-0404. Press, 1998.

Hansen, Mark B. N. Feed-forward. Stiegler, Bernard. “Pharmacology of Desire: University of Chicago Press, 2015. Drive-based Capitalism and Libidinal Dis-economy.” New Formations 72, 2011, Hansen, Mark B. N. New Philosophy for pp. 150-161. New Media. MIT Press, 2004. Supervised Learning, Wikipedia. https:// Lacan, Jacques. The Four Fundamental en.wikipedia.org/wiki/Supervised_learning. Concepts of Psycho-analysis. Routledge (1973) [2018]. The Early Universe, Cern.com, https:// home.cern/science/physics/early-universe. Massumi, Brian. “The Autonomy of Affect”. Cultural Critique, No. 31, The Politics of Uexküll, Jacob V. “The Theory of Meaning.” Systems and Environments, Part II. Autumn, Semiotica 42(1): 25-82 1982. 1995, pp. 83-109. Uexküll, Jacob V. A Foray into the Worlds Mechanical Turks. Wikipedia. https:// of Animals and Humans. University of en.wikipedia.org/wiki/Mechanical_Turk. Minnesotta Press, 2010.

Merleau-Ponty, Maurice. Phenomenology of Uexküll, Jacob V. “Die Bedeutung der Perception. Routledge, 2002. Umweltforschung für die Erkenntnis des Lebens.” Zeitschrift für die gesamte Merleau-Ponty, Maurice. Eye and Mind: Naturwissenschaft 1.7, 1935, pp. 257-272. Merleau-Ponty Reader, edited by Toadvine, T. Lawlor, L. Northwestern University Press, Virilio, Paul. The Vision Machine. University 2009. of Indiana Press, 1994.

Merleau-Ponty, Maurice. The Visible and Wittgenstein, Ludwig. Philosophical the Invisible. Northwestern University Press, Grammar. Blackwell, 1974. 1968. Print.

168 169 Daniel Chávez Heras

SPECTACULAR MACHINERY AND ENCRYPTED SPECTATORSHIP

Abstract

This paper argues certain types of contemporary computation have a spectacular dimension which is consumed today as magic. Using popular images created through Generative Adversarial Networks (GANs) as a case study, I analyse the conditions of production and consumption of imagery generated through machine learning as a type of popular culture, I then compare this creative use of computing with magic shows and the cinema of attractions of the early twentieth century. This approach combines notions of digital cultural materialism with theories of early film spectatorship to suggest an emergent cultural trend: monstrative global computation as a form of spectacle.

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Daniel Chávez Heras : SPECTACULAR MACHINERY ....

What you are about to witness is not way into the public domain under the broad magic, it is purely science. conceptual umbrella of artificial intelligence, — Robert Angier coming from all fronts, pitched at many different levels of detail, and applied to an equally Techniques are like seeds which bore diverse set of problems, from diagnosing and fruit in the soil of magic. treating cancer (Cruz and Wishart 59-78) to — Marcel Mauss playing Starcraft (Vinyals et al.). Agreement about the sense of wonder produced by these hitherto obscure algorithms is not always explicit, but it is clearly there. We need to talk about My first intuition as to how to investigate GANs the affective powers of machine learning was to look at how it has spilled into the arts. And perhaps the best example of this is the Machines for automatic learning are neither recent wave of enthusiasm for generative creative nor intelligent devices themselves, adversarial networks (GANs). but they are indeed awesome. While attribu- GANs are a type of unsupervised ma- tions of machine agency in most academic chine learning algorithm comprised of two circles are fraught, at least for now, there neural networks pitted to outperform each is undeniable enthusiasm about the pos- other. The idea was first introduced by Jürgen sibilities that these technologies appear to Schmidhuber (1990) and was further devel- enable in and beyond academia. It is there- oped and made popular by Ian Goodfellow fore surprising to find that very little atten- et al (2014). The technique has since engen- tion has been paid to the one aspect about dered several applications, most notably in automated learning where there seems to be synthetic generation of photo-realistic im- wider social consensus, which is the affec- agery (See for example: https://github.com/ tive dimension of these systems: that these nashory/gans-awesome-applications), and machines are awesome. in the process it made Goodfellow somewhat They are awesome in the literal sense, of a celebrity in machine learning circles. in that they can be awe-inspiring, cause The MIT’s Technology Review described feelings of reverential respect mixed with him hyperbolically as “The man who’s given fear and wonder. But why? How can, for machines the gift of imagination” (Giles), example, images produced through these and as of today his original GAN paper has techniques be as emotionally affecting to a over seven thousand citations in Google general audience when the mechanisms for Scholar. Granted, this academic celebrity is their creation are based on notoriously dry far removed from actual celebrity, but still, and emotionless statistics? And why has this for a technical paper this is remarkable: affective quality been overlooked in many ar- “GANs have come from an exotic topic to the eas of scientific machine learning research? mainstream and an exhaustive list of all GAN Machine learning techniques have papers is no more feasible or useful” — sum- captured the imagination of researchers and marises Holger Caesar, who until 2017 main- practitioners in seemingly disparate fields, tained an online list of papers on just this one to the point where news outlets are now machine learning technique (See: https:// struggling to make sense of the cornucopia github.com/nightrome/really-awesome-gan). of literature on the subject, which finds its

171 APRJA Volume 8, Issue 1, 2019

What is also interesting is how artists more than scientists were among the first to whole-heartedly embrace the use of GANs in their practices, and with very successful results too. German artist Mario Klingemann, now a self-styled neurographer, for example, had an interesting but mostly niche career until he started using GAN and GAN-esque algorithms between 2015-16 to produce portraits, or perhaps more precisely, ectypes (See: Floridi 317-321). Public interest in his work grew dramatically thereafter, as evidenced by the artist’s own collection of press clippings, which had increased tenfold by 2018 (See: http://quasimondo.com/) Almost at the same time that year, a painting created by a trio of French students using the same technique auctioned at Christie’s for $432,500 USD. The auctioneer’s website promoted the piece with the ques- Figure 1: Memories of Passersby I by Mario Klingemann. tion: “Is artificial intelligence set to become art’s next medium?” (Christie’s; Cohn). The Whether we like it or not GAN-art elevated price and some clever marketing has gone mainstream, and in the sciences put the technique and its practitioners in the as in the arts, machine learning has never public spotlight, to the delight and dismay of looked so awesome. And yet, descriptions some of its proponents, including Klingemann like the ones above, journalistic articles and himself, who commented disparagingly at editorials in specialised art publications, are the time: frustratingly unhelpful in directly questioning why these images are emotionally affecting To me, this is dilettante’s work, the for the wider audiences they are clearly be- equivalent of a five year old’s scrib- getting, people who for the most part neither bling that only parents can appreciate understand nor care about neural networks, […] But I guess for people who have loss functions or backpropagation. never seen something like this before, Even when artists were quick to adopt it might appear novel and different. these techniques (or perhaps because of (Vincent) it), their critics and them seem to lack the language with which to address the immedi- Klingemann’s own work would go on ate affective quality of the images produced auction a few months later, in early 2019, through techniques like GANs. These limita- this time at Sotheby’s, which promoted his tions, I believe, come from artists and art piece Memories of Passersby I as eliciting: critics persistent understanding of machine “an aesthetics that shocks and disturbs as learning as a medium of sorts: a conjec- much as it appeals, a mix of attraction and tural space that affords stylistic diversity repulsion whose principal effect is to present and the potential for aesthetic experience; a surprising new perspective.” (Sotheby’s) a new material and social surface for artistic

172 Daniel Chávez Heras : SPECTACULAR MACHINERY .... expression. The problem with this view is that produced through some of these machine it very often implies an unquestioned over- learning techniques struck a chord beyond reliance on what Noël Carroll (13) calls the the research communities from whence they medium-specificity thesis: the requirements came. It is also apparent, however, that gen- of differentiation and excellence that suppos- eral discourse about machine learning in the edly afford mediums their autonomy and their arts contributes little to our understanding of unique powers of expression — injunctions why (or even how) this is the case. such as films show, do not tell; games enact, In what follows I propose a different do not show, etc. approach. I suggest that images produced Going by this logic, the discussions through machine learning techniques like about images produced through machine GANs are not awesome because they are learning have focused on issues of original- the differentiable production of an artistic ity, authenticity and authorship, and this avant-garde, but on the contrary, that they preoccupation with finding the unique artistic are awesome because they are recognis- affordances of machine learning and how able en masse; because they are consumed authors deploy it in their practice tends to not as art but as a particular type of popular devolve into claims of different degrees of entertainment. My central argument is that machine autonomy as the defining feature of machine learning is emotionally captivating the medium: the quality that makes imagery not because the machinery is intelligent or produced through this type of machinery creative, but because it is spectacular, and original, and ultimately different from, say, a in this, I argue, the way we consume imagery Photoshop filter. The artists themselves have created through these techniques today has been either incapable of dispelling these much more in common with stage perfor- claims or actively complicit in perpetuating mances in the early twentieth century, with them. And critics too, as Carroll remarks, have the cinema of attractions, and particularly confused history with ontology (13), asking with magic. audiences to endorse an AI style under some questionable assumptions about the nature of the medium, i.e. its alleged deployment of non-human agencies. Anatomy of a magic trick This is not only the case with opportun- ists and outsiders, even well-established art- Around the same time GANs burst into the ists like Klingemann, who publicly and vehe- scene, cinema put our fears and wonder mently rejects claims of machine autonomy, about these technologies on screen in films struggle to defend their turf in terms other like Her, Transcendence, and Ex Machina. than style and mastery of the GAN medium These three films deal with the possibilities and are unable or unwilling to articulate non- and consequences of a synthetic intelligence essentialist views of their production. In late indistinguishable from ours, they are interest- 2018, when Klingemann won the gold award ing and in many ways enjoyable films, but like at the Lumen Prize, the event publicized it as: the headlines of auction houses they also “For the very first time, a portrait created by significantly misrepresent the nuts and bolts a machine has won a major global art prize.” of contemporary machine learning in science (Lumen Prize) and engineering research. From this quick romp through the early history of GAN-Art it is evident that imagery

173 APRJA Volume 8, Issue 1, 2019

A glance at the current flurry of papers like The Prestige (Christopher Nolan, 2006). on machine learning easily confirms that The Prestige portrays the misadven- they are overwhelmingly about applying tures of two rival illusionists in 1890s London, techniques from computational statistics Robert Angier and Alfred Borden, who try (classification, clustering, regression, rule to outperform each other in obsessive and inference) to specific problems, like credit increasingly dangerous ways. Borden de- card fraud detection, playing chess, or style velops a magic trick called The Transported transfer. Progress has been made for the Man, in which he appears to teleport instantly most part by going narrow, rather than gen- to opposite ends of the stage. Intrigued and eral.[1] The recent breakthroughs in the field, frustrated, Angier spies on Borden and tries Daniel Dennett writes, to replicate the trick, first using a double, and eventually commissioning a cloning machine have been largely the result of turn- from American scientist Nikola Tesla. After ing away from (what we thought we much speculation, personal drama and understood about) human thought murder, it is revealed that the way Borden processes and using the awesome performed the transported man was by con- data-mining powers of supercomputers cealing from everyone the existence of a twin to grind out valuable connections and brother, with whom he shared not only the patterns without trying to make them stage but also his wife (eventually driven to understand what they are doing.[2] suicide because of the inconsistent personal- (Dennett 87) ity of what she presumed to be an individual but were in fact the twins). This narrow or weak AI paradigm sug- The plot of the film is structured as a gests that we should not look to science fiction series of flashbacks in which the magicians but to history, and also that perhaps it is not take turns at reading the other’s stolen diary. that films get it wrong, but that we are looking Much like the generator and discriminator at the wrong films. If we want to understand modules in a GAN,[3] their rivalry pushed the machine learning not as synthetic intelligence boundaries of magic, albeit in very different or creativity, but as magic, I suggest we take ways: Borden and his twin accomplish the our cue from a film about magicians, a film illusion by concealing a lifetime of duplicity,

Figure 2: The Transported Man, magic trick - A.

174 Daniel Chávez Heras : SPECTACULAR MACHINERY .... while Algier achieves it through a scientific if we, the audience, believe the person who machine that actually duplicates him. In the vanishes from one place appears to be the film, scientific machinery and social perfor- same person that reappears instants later mance are seen as two constituent traditions elsewhere — that Borden somehow man- of magic; two ways of producing the same ages to travel in ways that defy common magic trick. experience. Semantic echoes aside, I also like this But consider, following Mauss, how example because it allows us to think seri- disbelief is always historically situated. We ously about how magic implicates labour and can imagine for instance how nineteenth technology for spectacular effects. Magic, century audiences would have probably writes Marcel Mauss in his General Theory of found equally incredible that a person could Magic, “is the domain of pure production, ex fall asleep in London one day and wake up nihilo. With words and gestures it does what in Manila the following day, and how before techniques achieve by labour” (175). A magi- air travel became common in everyday life cian, he continues, “does nothing, or almost this too could have easily been construed as nothing, but makes everyone believe that he magical. From this perspective, the illusion of is doing everything, and all the more so since teleportation is only a function of our percep- he puts to work collective forces and ideas to tion of the time needed for the necessary help the individual imagination in its belief” transformations required to displace matter (175). Despite being over a hundred years in space. Consider then, how the illusion of old, Mauss’ anthropological account of magic travelling at the speed of light is profoundly illuminates a forgotten link between tech- connected with the social imagination about nique and showmanship, or in other words, technologies like electricity, radio, the tel- of how magicians play with social expecta- ephone, and indeed the kinematograph, in tions of what is technically possible. In the the early twentieth century.[4] John Cutter, case of The Transported Man, for example, the ingenieur working with Angier (played the magic occurs not because the magician by Michael Caine in the film), at one point cannot be at either end of the stage, but advises the performer: “if you need some in- because he appears to travel this distance spiration, there’s a technical exposition at the at an impossible speed. The trick only works Albert Hall this week. Engineers, Scientists,

Figure 3: The Transported Man, magic trick - B.

175 APRJA Volume 8, Issue 1, 2019 you know […] That sort of thing catches the Echoing electrical technologies of the public imagination.” (The Prestige) early twentieth century, machine learning too The originality of Mauss’ approach was disrupts our relationships with perceived time to show how magic worked by regulating the and labour in powerful ways. Indeed, it is my social interfacing with technologies, his work contention here that the main magic trick per- is not only a description of ritualised magi- formed through machine learning systems cal practices, but an analysis of how these consists in using statistical computation for practices are specifically designed to amplify the compression of time through what Matteo weak collective beliefs and disbeliefs so as Pasquinelli calls “the ideological encryption to render them effective at specific moments of labor within technology” (321). Pasquinelli in history: argues, albeit in a wider context, how classical energy theories of labour[7] have “failed Magic protected techniques; behind to recognise the new forms of technified magic they were able to make pro- labour and technified subjectivities that have gress. […] Magic is linked to science in lost any resemblance to the new labour the same way as it is linked to technol- struggles of the past” (321). If we admit his ogy. It is not only a practical art, it is revision of classical Marxist economics, we also a storehouse of ideas. It attaches can easily see how through vast infrastruc- great importance to knowledge — one tures of planetary computation different kinds of its mainsprings. (175) of subjectivities can be encoded, harvested, packaged and sold back to us, through ma- In my view, some aspects of the cur- chine learning, as instantaneous projections rent technological moment with regards to of “artificial” knowledge or creativity. But of machine learning deeply echo those of the course, there is nothing artificial about these early twentieth century: once again we are, subjectivities, it is our perception that is be- like Angier says in the film, “on the brink of ing surpassed since we cannot yet grasp new terrifying possibilities,” and once again the encryption of labour at a global scale. the boundaries of what is technically possible As with The Transported Man, we could also are softened enough so as to present and understand GAN imagery in these terms, as sell technology as magic. My argument here magic protecting technique: the trick being, is that machine learning is being presented to present the results of encoded subjectivi- to us as a series of magic tricks: instant ties and encrypted labour all at once. retrieval, disembodied cognition, as creative Think for example of the thousands of or intelligent machines, all of which bear the images of European portraits Klingemann clear social hallmarks of the magical: they fed to his Old Masters GAN in terms of en- are deployed as forms of alchemy[5] (with crypted labour, and one can then appreciate the right algorithm you can convert your data how he is compressing a thousand years of into gold), animism (the machine thinks and European portraiture tradition into an instant speaks for itself), divination (big data and of release. predictive analytics), and healing (genome decoding and editing). Symptomatically, corporations who wield these powers even present themselves as overtly magical, even in their nomenclature, think for example of Oracle or Palantir.[6]

176 Daniel Chávez Heras : SPECTACULAR MACHINERY ....

Figure 4: Cover to the box for a Cinématographe.

achievements. In his influential essay on the Spectacular machinery cinema of attractions, Tom Gunning referred to this type of spectatorship as popular exhi- Having dissected the magic trick, let us bition of trick films: come back to our original question: how are general audiences affectively bound to the Nor should we ever forget that in pictures created through machine learning the earliest years of exhibition, the such as GANs? cinema itself was an attraction. Early Like a magic show, for the trick to be audiences went to exhibitions to see successfully carried out the audience needs machines demonstrated (the newest to actively participate with their beliefs and technological wonder, following in the social imagination of what is possible. We wake of widely exhibited machines need to have the disposition to be deceived and marvels as X-rays or, earlier, the and to be amazed; we need to buy the trick phonograph), rather than to watch as entertainment. My claim here is that we films. It was thecinématographe , the consume pictures created through machine biograph, or the vitascope, that were learning today in a similar way to how film advertised on the variety bills in which scholars have characterised early twentieth they premiered, not [LE DÉJEUNER century audiences consumed cinema: not DE BÉBÉ] or THE BLACK DIAMOND as cinema, but as attractions; as spec- EXPRESS. (383) tacular demonstrations of technological

177 APRJA Volume 8, Issue 1, 2019

Figure 5: Cross-section of the Turk.

André Gaudreault, who worked closely practices greatly differed from what was later with Gunning, went as far as to suggest a called cinema: revisionist history of the birth of cinema, a version in which neither Edison in 1890, Between the time of the inven- nor the Lumières in 1895, invented cinema, tion of the basic device (between but only the devices later used for it: the 1890 and 1895) and the period of Kinematograph and the Cinématographe.[8] institution (beginning around 1915), According to the film historian, cinema came kinematography was a wide-open field into existence more than a decade later, in of experimentation. This was when the 1910s, when the conventions of theatre artisanal manufacturers of animated and performance were assimilated into films pictures took various initiatives, almost as institutionalised products with a recognis- all of which tended to modify the initial able narrative form. Gaudreault describes project inscribed, so to speak, in the the twenty years between 1890 and 1910 ‘genes’ of the apparatus (or, if you as a period of “kine-attractography” whose prefer, in the various patents filed by its many inventors. (39) 178 Daniel Chávez Heras : SPECTACULAR MACHINERY ....

Both Gunning and Gaudreault described their own fantasies onto the machine’s raw this cinéma des premiers temps in terms of potential more or less freely. its capacity to show the new techniques: slow Spectacular machines can in this way motion, reverse motion, multiple exposure, be presented as immediately awesome, they and even the close-up, which we now take seduce by showing, rather than persuade for granted as part of narrative cinematic through reflexive absorption. And technolo- discourse, but which, Gunning argues, at the gies for observation are particularly alluring time was seen “in itself [as] an attraction and in this regard for their capacity to create the point of the film” (384). trickery that reveals: distorted ways of see- Furthermore, already in the early ing (like the microscope or the kinetoscope), twentieth century, these machine attractions that simultaneously implicate intellect and — kinematography included — conflated imagination, and that give both the produc- notions of the showable and the knowable tion of knowledge and the creation of fiction under the logic of spectacle. The shows that an immediate, often spectacular visual mani- invited audiences to suspend disbelief, as festation. Viva Paci calls this, in the case of we saw through Mauss, also served as ways early cinema, “the attraction of the intelligent to deploy new technical intellectual regimes. eye” ( 121–38). And there is, of course, a rich older history I argue machine learning too embodies of automatons being exhibited as “intelligent” this double function today: it can be under- attractions[9]: Wolfgang von Kempelen’s stood as a set of observational technologies Mechanical Turk,[10] for example, which was that affords us with spectacular trucages qui presented as a mechanical chess player to révéler. This is, I believe, the best way to impress the Habsburg court in 1770 (Schaffer understand how we currently consume GAN et al. 154), or even John Bowes’ Silver Swan imagery, not as expressions of a medium but of 1872, which Mark Twain described as as spectacular demonstrations of the GAN having “a living grace about his movement itself. and a living intelligence in his eyes” (Twain The advantage of this analytical ap- in Holledge 13). Automata were symbols of proach inspired in the cinema of attractions the enlightenment, writes Simon Schafer, is that it allows us to bypass the idea of they “were both arguments and entertain- AI style altogether, since the attraction of ments, designed seductively to place craft the intelligent eye operates equally on the skill within the setting of power, and to allow gooey portraits of Klingemann or in the the selective entry by that power to the inner photorealistic deepfakes which are created workings of art and nature” (135–36). through the same technique but look entirely From the courts of Europe to the different. What attracts us is not the style, burgeoning urban elites of industrialised but the workings of the machine. And this cities, this model of spectatorship based is not the self-reflexive aesthetic modality on attractions historicise these moments of an art that is questioning and testing the of rapid technological development when limits of its own medium, but a much more societies find the machines themselves as general allure, one that accounts better for mystifying; when technologies have not yet the popularity of this form of picture-making. been tethered to particular usages, captured Furthermore, I want to advance the idea that by specific sectors of society, or institution- this is also a viable way to more generally alised into coherent social discourse, and characterise the current affective resonance audiences are therefore still able to project of machine learning in visual culture: not as a

179 APRJA Volume 8, Issue 1, 2019 tool for representation, but as a magic show contingent on the mystifying process through Notes which computation at a planetary scale can encrypt subjectivities and labour. [1] There is research in so-called strong AI To be clear, I am not suggesting all or General AI, but practice in this field has imagery produced through machine learning been dwarfed into a sub-field in the past is designed as a magical attraction, there decade. Other significant related areas in are of course a multitude of configurations of computer science and philosophy include machine learning systems, processes, inten- computability and computational complexity. tions, and a rich melange of techno-social palimpsests, some of which will surely find [2] First emphasis is mine, second one is new ways to negotiate their way into broader the author’s. areas of visual culture. And similarly, we have to distinguish from the relatively small group [3] “The magic of GANs lies in the rivalry of artisans of technoscience trying to outdo between the two neural nets,” states Martin each other’s tricks in academic machine Giles in the same piece where he calls learning research and the business of big Goodfellow “the GANfather.” data analytics, which is already institutionalised into large corporations mostly concerned [4] This theme of time relativity and how with extending their encroachment in society it has profound effects on social relations at large through sophisticated forms of digital features heavily in other films by Christopher governance and wealth extraction.[11] Nolan, for example in Interstellar (2014), My argument is, rather, about consump- where a cosmonaut and the young daughter tion. I submit that there is at this moment a he left behind on earth live their lives in popular appetite to consume these images different temporalities, and he is later able as the magical results of monstrative global to re-encounter her as an elderly woman. computation, much like there was in the early Or Inception (2010), where a crew of days of kinematography for tricks and effects specialised dream bandits go through with moving pictures, and that besides novel nested dream levels, each with a temporal- ways of creating and analysing imagery, ma- ity relative to the level above and below. In chine learning systems afford us with novel these films Nolan knowingly references the ways of enjoying imagery; they fetishise cal- origins of cinema and seems to be acutely culation and the statistical apparatus that aware that one of the greatest powers of makes it possible, and they turn the datafica- cinema as a technology was to afford us tion of society into its own form of spectacle: with new social understandings of duration. spectaculum ex computatio. I believe we are living the early days of these forms of com- [5] See: http://supercommunity.e-flux.com/ putational spectatorship. Goodfellow may texts/the-alchemic-digital-the-planetary- have invented GANs, but the medium which elemental/. will allow us to enjoy sequencing without continuity, narrative without authorship and, [6] Palantir Technologies is a US software ultimately, presence without subject, has not company specialised in big data analytics. yet been invented. The Palantíri, or seeing stones, are a set of interconnected magical orbs in J.R.R. Tolkien’s Lord of the Rings mythology. They

180 Daniel Chávez Heras : SPECTACULAR MACHINERY .... allow their users to communicate and have and the gear noises served to conceal visions of future or past events across the his presence. Amazon named named its world: “Human Intelligence Tasks” marketplace after this automaton, see: https://www.mturk. The palantír replied to each, but all com/. those in Gondor were ever open to the view of Osgiliath. Now it appears that, [11] Corporations pluck techno-artisans from as the rock of Orthanc has withstood academia whenever necessary, of course. the storms of time, so there the palantír Big tech in effect buys out the results of a of that tower has remained. But alone it more malleable field of experimentation. could do nothing but see small images Once engineers and scientists are turned, of things far off and days remote. Very so to speak, data fencing becomes an issue useful, no doubt, that was to Saruman; that then separates them from their original yet it seems that he was not content. research communities. Marx may have Further and further abroad he gazed, called this the subordination of techno- until he cast his gaze upon Barad-dûr. scientific labour to capital. Then he was caught! (Gandalf, in The Two Towers, Chapter 11.) Works cited [7] Pasquinelli mostly refers to Marxist notions of labour as transformation through Carroll, Noël. “The Specificity of Media in energy, and he argues this view has ignored the Arts.” Journal of Aesthetic Education, the latent productive potential of information. vol. 19, no. 4, 1985, pp. 5–20. JSTOR. doi:10.2307/3332295. [8] These two are similar but not exactly the same device, although they were developed Christie’s. Is Artificial Intelligence Set to almost in parallel, one by Edison in the US Become Art’s next Medium? 12 December and the other by the Lumières in France. 2018. https://www.christies.com/features/A- collaboration-between-two-artists-one- [9] Musée de la Magie and Musée des human-one-a-machine-9332-1.aspx. Automates, in Paris, are adjoining twin- museums for which one can purchase a Cohn, Gabe. “Up for Bid, AI Art Signed single ticket. ‘Algorithm’.” The New York Times, 23 October 2018. https://www.nytimes. [10] von Kempelen’s Turk was initially com/2018/10/22/arts/design/christies-art- presented alongside magic tricks, and artificial-intelligence-obvious.html. travelled through the courts of Europe playing exhibition matches and igniting Cruz, Joseph A., and David S. Wishart. speculation in scientific circles about its “Applications of Machine Learning in mysterious functioning. There was for a Cancer Prediction and Prognosis.” time a strong belief that the Turk operated Cancer Informatics, vol. 2, January through magnetism. After the death of von 2006, pp. 59-78. SAGE Journals, Kempelen, the Turk was uncovered as a doi:10.1177/117693510600200030. hoax: a small man was inside the cabinet,

181 APRJA Volume 8, Issue 1, 2019

Dennet, Daniel C. “The Singularity —An Paci, Viva. “The Attraction of the Intelligent Urban Legend?” What to Think About Eye.” The Cinema of Attractions Reloaded, Machines That Think, edited by John edited by Wanda Strauven, Amsterdam Brockman, Harper Collins, 2015. University Press, 2006, pp. 121–38. http:// www.jstor.org/stable/j.ctt46n09s.11. Floridi, Luciano. “Artificial Intelligence, Deepfakes and a Future of Ectypes.” Pasquinelli, Matteo. “The Automaton of the Philosophy & Technology, vol. 31, no. 3, Anthropocene: On Carbosilicon Machines Sept. 2018, pp. 317–21. Springer Link, and Cyberfossil Capital.” South Atlantic doi:10.1007/s13347-018-0325-3. Quarterly, vol. 116, no. 2, 2017, pp. 311–26.

Gaudreault, André. Film and Attraction: Schaffer Simon, et al. “Enlightened From Kinematography to Cinema. University Automata.” The Sciences in Enlightened of Illinois Press, 2011. Europe, University of Chicago Press, 1999, pp. 126-165. Giles, Martin. “The GANfather: The Man Who’s given Machines the Gift of Schmidhuber, Jiirgen. “Making the World Imagination.” MIT Technology Review, Differentiable: On Using Self-Supervised 2019. https://www.technologyreview. Fully Recurrent Neural Networks for com/s/610253/the-ganfather-the-man-whos- Dynamic Reinforcement Learning and given-machines-the-gift-of-imagination/. Planning in Non-Stationary Environments.” Institut für Informatik, Technische Universität Goodfellow, Ian J., et al. “Generative München, 1990. Adversarial Networks.” ArXiv:1406.2661 [Cs, Stat], June 2014. arXiv.org. Web. http:// Sotheby’s. “Artificial Intelligence and the arxiv.org/abs/1406.2661. Art of Mario Klingemann.” Sothebys.com, 8 February 2019. https://www.sothebys.com/ Gunning, Tom. “The Cinema of Attraction[s]: en/articles/artificial-intelligence-and-the-art- Early Film, Its Spectator and the Avant- of-mario-klingemann. Garde.” The Cinema of Attractions Reloaded, edited by Wanda Strauven, Vincent, James. “How Three French Amsterdam University Press, 2006, pp. Students Used Borrowed Code to Put the 381–88. http://www.jstor.org/stable/j. First AI Portrait in Christie’s.” The Verge, ctt46n09s.27 23 October 2018. https://www.theverge. com/2018/10/23/18013190/ai-art-portrait- Holledge, Richard. “Magic Wrought by a auction-christies-belamy-obvious-robbie- Merlin.” Wall Street Journal, 22 December barrat-gans. 2012. Vinyals, Oriol, et al. “StarCraft II: A New Lumen Prize. “The Butcher’s Son.” 2018 Challenge for Reinforcement Learning.” Winners, 2018. https://lumenprize.com/ ArXiv:1708.04782 [Cs], August 2017. http:// artwork/the-butchers-son/. arxiv.org/abs/1708.04782.

Mauss, Marcel. A General Theory of Magic. Routledge, 2001.

182 183 Tomasz Hollanek

NON-USER-FRIENDLY: STAGING RESISTANCE WITH INTERPASSIVE USER EXPERIENCE DESIGN

Abstract

User-friendly design makes our use of emerging technologies intuitive and seamless, but it also conceals the new solutions’ influence over how we act, think and plan. In this paper, I analyze the logic of our newly developed ‘touchscreen sensibilities’ to speculate on alternative, ‘non-user-friendly’ design practices that, by invading intuitive interfaces, could make the users aware of their reliance on invisible algorithmic operations to learn and to feel. I revisit Žižek and Pfaller’s conception of ‘interpassivity’ to explore its potential as a means of resisting interactivity and inciting consciousness in contemporary speculative design. The critical interface I envision must defamiliarize consumption, prevent participation, and de-frame perception — make the user experience what lack of control feels like, and do so to encourage resistance.

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Tomasz Hollanek: NON-USER-FRIENDLY

I am using my iPad to edit this article. What if design was to pose a challenge The Google doc is available anytime and to the logic of immediacy? Deny the viewer- anywhere, so I can always apply changes, user the power to manipulate the image and jot down new thoughts on-the-go. I use the register reactions to it? Stage an experience same device to access anything online, from that cannot be immediately turned into data? bank accounts and YouTube clips to my What if ‘non-user-friendly’ design had the memories stored as data in the cloud. The potential to transform the existing feedback experience of navigating between these dif- loops into a new system of commentary — ferent modes is completely intuitive. It feels to de-automatize choice? In what follows, I natural. The device belongs to me and so describe the logic of our newly developed does, as it seems, the universe fashioned out ‘touchscreen sensibilities’ and speculate on of data. The device ensures that my sense of alternative design practices that, by invading orientation is closely tied up with an illusion intuitive, user-friendly interfaces, could cause of control. While visualization masquerades ‘cognitive glitches,’ exposing our reliance on as comprehension, touch colonizes space. invisible algorithmic operations to learn and Wandering off course is not possible in this to feel. I suggest there is potential in ‘inter- world with a home button. passivity,’ if it were inscribed into the user’s But immediacy comes at a price as experience of media, to resist interactivity, to user-friendly design that makes our use of incite consciousness and encourage change. new technologies intuitive and seamless also aims to misdirect our attention from what is happening behind the scenes. The industry’s ambition to personalize our experience of Touchscreen sensibilities media via ever more adaptive interfaces might lead to a future in which planning is In the late 2000s, Apple released the original entirely delegated to the machine — one that iPhone, a new kind of apparatus equipped employs AI-enhanced analytics to anticipate with a sensory system of its own — a touch- our needs and desires. This delegation of screen, a built-in camera, an accelerometer, responsibility will likely happen surrepti- a proximity sensor, a gyroscope, and other tiously, as technologies have already begun sensors; a device that could at once display to preempt decision-making without us notic- and register images, connect different users ing. The act of purposeful selection on the across a distance, and react to light intensity, part of the user might eventually become ob- movement, and speed; it was both a screen solete — indeed it might disappear from the — but one that could gaze back at the view- menu of options made available by the future er, respond to his or her touch, heartbeat, UX (user experience) design. If choice is and position in space — and a controller, a essential to autonomy, then the question we remote for executing tasks. The controller- are facing today is whether — in the age of screen seems like the ultimate ‘remediation’ automated decision-making, optimization of (to use Bolter and Grusin’s term), realizing options, and ongoing surveillance of actions our desire for instantaneity and immediacy, — we can still conceive of design strategies mobility and interactivity, manipulability and that allow the user to perceive the technology control — a device that can respond to its at work: to become aware of the algorithmic master’s voice, recognize his or her face, or operations and invisible infrastructures that track its user’s steps. It not only determines are shaping our experience of the world. what and how we see, but indeed how we

185 APRJA Volume 8, Issue 1, 2019

‘calibrate’ our bodies, how easily we orient “intimate sameness” (Ekman 1). Ongoing them in the new, augmented reality. developments in user experience design rely Mark Hansen has recently remarked that on dynamic, fully customizable interfaces the “becoming topological of culture” — the that automatically adapt to the viewer-user’s forging of topological relations among “ele- needs, seemingly responding to his or her ments of worldly sensibility” by contemporary desire before it is consciously articulated. media machines (Hansen 34) — demands With advances in user profiling, a process both our reconceptualization of sensibility of generating statistical models from large and phenomenology (37), as today’s topo- amounts of user data, diverse mobile ap- logical machines “provide artificial access to plications can now predict, and attune their a domain of sensibility that exceeds what hu- messaging to, the users’ sexual orientation, mans can process as sensations” (39). Users political affiliations, or even their menstrual have become dependent on their devices, cycle. As the interface facilitates not only the personal touchscreens and other wearables, consumption of digital goods, but also self- to process signals that no human body part tracking, it invites the viewer-user to become can detect on its own. To sense is to register; self-conscious through the technology; self- to feel immersed in reality is to manipulate tracking, however, serves only as a prosthe- it through swipes, taps and pinches. This is sis of the project, an illusion of individuation how touchscreen sensibilities have become aiming to collect ever more data. the dominant perceptual norm of our time. Bernard Stiegler argues that the A decade after Apple introduced its first contemporary media draw “the time of con- iPhone, my personal touchscreen is always sciousness” into production to manufacture there, always at hand. And yet, it remains our desires. His conception of individuation imperceptible. Because touchscreen sensi- in the age of “hyper-industrial” capitalism bilities necessitate a design that obfuscates revolves around the paradoxical relationship mediation; ‘good’ design must feel intimate between the illusion of personalization and and natural to allow the interface to erase it- the massification of cultural consumption — self and pass as an extension of the organic. the ways in which audiovisual technologies The controller-screen moves with us, control “the conscious and unconscious transforming our perception with its machine rhythms of bodies and souls,” by exploiting vision and optimization of experiences. But the aesthetic and treating consciousness as its influence remains concealed through “raw material” in the process of production what designers call ‘user-friendly design’ (Stiegler, Symbolic Misery 2). Broadcast — the kind of design that makes the use of media, Stiegler argues, function as perva- new technologies intuitive and seamless. sive systems of synchronization, relying on User-friendly in the age of big data profiling temporal objects such as TV programs or means tailored to the individual. Ulrik Ekman songs (objects whose affective potential is argues that in the reality of ubiquitous com- inscribed in their very duration), that stand- puting, even environments begin displaying ardize the time of consciousness to format “intelligent attention” to individuals and social the consumer’s behaviour. While in the era groups: “natural setting turns highly artificial of broadcast media standardization (the syn- as it appears attentive rather than neutral or chronization and ‘averaging’ of individuals) non-caring” — it constantly interacts with the disguised itself as personalization (pervad- viewer-user, responding with a directedness ing the home), in the era of asynchronous “coming not from distant otherness,” but viewing, personalization (the profiling of

186 Tomasz Hollanek: NON-USER-FRIENDLY users and the filtering of information) poses as standardization (foregrounding the ‘social Cognitive glitching media’ features of the design, concealing the algorithmic processes and convincing the Non-user-friendly UX design is not necessar- viewer-user that he or she is seeing what ily synonymous with counterintuitive design. everybody else is seeing). While broadcast Different iterations of counterintuitive solu- media have laid the groundwork for drawing tions are being adopted by online platforms the time of consciousness into production, either to draw the users’ attention to their personalized interfaces of the digital era actions by breaking with prevalent design might complete the project by soliciting our patterns, or, the exact opposite, to manipu- attention on a full-time basis. late their decisions by introducing confusing Although user-friendly experience interface elements. Google Chrome, for design in the era of ubiquitous computing example, has experimented with counterin- seems to realize the promise of control tuitive solutions to warn users about inse- vision, unlimited mobility and haptic imme- cure connections. To proceed to a website diacy, it also becomes a means of capturing marked up as suspicious, Chrome’s user has information about the preferences and habits to click on a grey hyperlink displayed below of users and turning the collected data into the warning message rather than on a blue profit for corporations like YouTube. And as button, ensuring that he or she is taking the the techno-extensions of the human sensory risk consciously and not clicking out of habit. system take on the role of imperceptible in- Other companies (such as low-cost airlines) termediaries between corporate agenda and incorporate confusing, illogical procedures our consciousness, they may inhibit our abil- into their web design to trick users into buying ity to plan. The device may feel personal, but extra services. In any case, counterintuitive it has never been truly mine. design, however unfriendly, prompts the user Alternative design would have to call to interact with the system as it is — some- our new, machine-enabled feelings into thing non-user-friendly design should strive question. It would have to free cognition to prevent. from the mobile ‘frame’ of the controller- Non-user-friendly design would have screen and prove a means of paradoxical to become what Anthony Dunne and Fiona ‘de-framing’ of contemporary perception. It Raby describe as “critical thought translated would have to revert the logic of touchscreen to materiality” (35) — design that combines sensibilities — for the idea of the project to speculation and futurology in order “to re-emerge. Can we conceptualize UX design change reality, not merely describe it” (6). For that reconnects the viewer-user with his or Dunne and Raby, design has become “so ab- her time of consciousness, or — in other sorbed in industry, so familiar with dreams of words — attunes consciousness to the lived industry, that it is almost impossible to dream body? Could ‘non-user-friendly’ design suc- its own dreams” (88). Critique, they argue, cessfully harness the feeling of confusion and must be “a refusal, a longing, a desire” (35). dissatisfaction to raise political awareness, to In their book on speculative design, they cause a cognitive glitch? contend that a project’s potential lies not in:

what it achieves or does but what it is and how it makes people feel, especially if it encourages people to

187 APRJA Volume 8, Issue 1, 2019

Figure 1: Facebook News Feed modified bySafebook (https://bengrosser.com/projects/safebook/).

question, in an imaginative, troubling, algorithmic curation by automatically redact- and thoughtful way, everydayness and ing virtually all content — text, images, videos how things could be different. To be — from the website. After installing Safebook effective, the work needs to contain on Chrome or Firefox, the user is left with contradictions and cognitive glitches. a layout of blanks and omissions, with only Rather than offering an easy way the framework of Facebook’s user interface forward, it highlights dilemmas and intact and recognizable. The user can still trade-offs between imperfect alterna- interact with the website, but ‘liking’ an invis- tives. Not a solution, not a ‘better’ way, ible image by clicking one of the concealed just another way. Viewers can make up reaction buttons invariably proves a shot in their own minds (189). the dark. Safebook defamiliarizes the experi- Design as critique must “invade the eve- ence of ingesting information through social ryday,” (43) and that is why non-user-friendly media, as Grosser’s software takes on the design is not necessarily meant as a mere form of a sui generis AdBlock — targeting all inversion of user-friendliness, but rather an content made available through Facebook invasion of intuitive interfaces that exposes — to suspend direct consumption. The their underlying structures, uncovering the browser extension obfuscates the results of apparatus to reveal an aperture, a way out. Facebook’s personalization to diminish the Various digital artists and designers influence of algorithms over what we see, and have toyed with seemingly unfriendly inter- thus seemingly allows us to take back control face elements to foster critique. Benjamin over what we do with our time online. This is Grosser’s Safebook (2018) serves as a also how Safebook indirectly reinforces the particularly evocative example of the trend: idea that the danger posed by contemporary as a plugin available for download from the technology relates to the users’ compulsive artist’s website, Safebook aims to reinvent tendencies: that more software can lead to Facebook as a space free from persuasive more control or, specifically, self-control, and

188 Tomasz Hollanek: NON-USER-FRIENDLY that the challenge for designers in the age of of cognitive estrangement the designers are the continuous stream is to search for new aiming at. Perhaps in the age of touchscreen design strategies that enable digital temper- sensibilities, only non-user-friendly design ance. Safebook, however provocative, shares based on the performance of technological more with the likes of ScreenTime (one of indifference — enforcing the user’s passivity Apple’s latest solutions that allows the user via an interface usually meant to maintain to schedule time away from the screen or set a continuous interaction — could cause a time limits for individual apps) than it initially cognitive glitch. appears to — realizing the conception of new “When people’s participation becomes design as a means of extending control over someone else’s business,” argues Jonathan any previous design. The ‘unfriendliness’ it Sterne, “the social goods that are supposed stages turns out superficial, as the logic of to come with it can be compromised.” He sug- the interface-as-controller remains unchal- gests that “the bad things that media critics lenged. And since it is precisely the interface, have been saying about passivity” seem ap- not the stream of content, that makes us plicable to contemporary media’s “demands believe we have control over our life online, to interact, to participate.” Active participation Safebook fails to embody non-user-friendli- fails to renew commentary and bring about ness as a means of disturbing the illusion of agency, as interactivity “also encompasses technology that merely serves its master. the ‘agree to terms’ button” (Sterne). Perhaps Grosser does gesture, however, any form of UX design based on interactive towards the idea of software that prevents engagement is complicit in the dreams of the rather than encourages interaction. Instead industry? Perhaps to generate a cognitive of making the content invisible to the user, glitch, to open up “all sorts of possibilities that perhaps critical design could do the reverse can be discussed, debated, and used to col- to bring the concept of non-user-friendly lectively define a preferable future,” (Dunne design closer to fruition: render the user and Raby 6) design speculations in the age of imperceptible to the system. An interface that touchscreen sensibilities should reconsider remains oblivious to the user, not reacting to the potential of passivity, of standardization, touch, voice, or any other well-known com- of the screen that cannot gaze back? mands, could also interrupt the false sense of control that a user-friendly interface aims to generate. Writing about the limitations of cinema, Dunne and Raby point out “it can Interpassive interfaces deliver a very powerful story and immersive experience but requires a degree of passiv- If Sterne is right and interactivity is indeed ity in the viewer” (75), contrasting film with the new passivity, then could ‘interpassiv- speculative objects as invitations for “the ity’ become the new activity in the age of viewer to actively engage with the design touchscreen sensibilities? Interpassivity was rather than passively [consume] it” (90). conceptualized by Slavoj Žižek and Robert Dunne and Raby think primarily of physical Pfaller in the late 1990s to describe the rela- objects, but if we transpose their argument to tionship between a subject and objects that digital environments, the opposite may prove have inscribed in them their own reception: to be true: putting the user in a seemingly they anticipate reactions and thus fulfill their passive position, turning to older patterns of role on their own, supposedly not relying on engagement, may enable the mental process the subject to interact with them. The use

189 APRJA Volume 8, Issue 1, 2019

Figure 2: Visualization of an ‘interpassive’ interface (by the author). of ‘canned laughter’ in sitcoms is one of the having to take an active role in the process most common examples of interpassivity, of selection. And yet, the canned laughter in relating the phenomenon to the delegation the form of incessant recommendations, au- of enjoyment to objects themselves: a sitcom tomated playlists and algorithmically curated episode seems to be ‘enjoying itself’ inde- feeds is not ‘canned’ at all, as the interface pendently of the viewer’s presence. maintains a never-ending exchange of infor- Perhaps the term, as defined by Žižek mation with the user; the user is, ultimately, and Pfaller, is not directly applicable to the necessary component of the system, the contemporary digital environments, since, key (re)source of click-throughs, likes and as Gijs van Oenen points out, interpassiv- other sorts of data. The process of active se- ity “refers more specifically to the period of lection on the part of the user might eventually modernity, when subjectivity is ‘haunted’ by become obsolete in the age of AI-enhanced the expectation of incessant activity” (van personalization — but only because the user Oenen 8). Van Oenen argues that interpas- has participated in a continued interaction sivity, understood in the context of modernity, long enough for the system to predict his becomes a delegation of activity, not passiv- or her needs and desires in the future. The ity — “a delegation necessitated by an acute user-friendly interface employs elements that sense of being overwhelmed by interactive overtly encourage interaction only to enable engagements and obligations” (11). Can we a form of interpassivity — it masks our reli- argue, therefore, that new design solutions, ance on the algorithm and our delegation of as they aim to preempt conscious decision- decision-making to the machine with a seem- making, establish an interpassive, rather than ingly controllable, interactive layer of intuitive interactive, relationship with the user? The software. contemporary viewer-user indeed seems to While interpassivity has been theorized unwittingly delegate his or her activity to the in relation to the modernist notions of activ- automated system to remain passive — to ity and passivity, scholars such as Hagen enjoy the effects of personalization without Schölzel have looked beyond this framework

190 Tomasz Hollanek: NON-USER-FRIENDLY to gesture towards the reinterpretation of content, while denying him or her the option interpassivity as a way of “backing away from of navigating the flood of information with circles of control” (187) in our current political recognizable interface elements. Perhaps, culture of participation. I refer to these theo- instead of redacting the content, a truly retical framings that push Žižek and Pfaller’s non-user-friendly software should obfuscate ideas forward to suggest there is potential in Facebook’s interface design and confront interpassivity-as-resitance realized through the user with an assemblage of images, text contemporary critical design. Most recently, messages, and videos that suddenly begin Alex Gekker has referred to the genre of idle to feel uncontrollable — indeed out of reach. games on YouTube (recorded by gamers Only then would touch fail to colonize data. and uploaded to the platform to be watched Only then would design reveal the ultimate by other users) as a way of examining “the lack of control the user has over what he or system as a whole” that allows the interpas- she experiences through the interface. This sive actor to inspect it “without responding to kind of non-user-friendly software would [its] always-on logic” (219). Building on van make the user feel ignored, invisible — to Oenen and Schölzel’s observations on inter- eventually redirect his or her desire for imme- passivity as a form a resistance, he argues diacy, control and omnipresence, satisfied by for interpassivity’s liberating potential as “an the illusion of user-friendly software, from the alternative to straightforward consumption” manipulable data universe to the real world. (221). Gekker’s optimistic interpretation of The critical interface I envision must defa- consumption of specific YouTube videos miliarize consumption, prevent participation, as a liberating, interpassive practice has its and de-frame perception — make the user apparent limitations — after all, how many experience what lack of control feels like, users who subscribe to gaming channels are and do so to encourage resistance. capable of distancing themselves from the viewing and becoming aware of the system as a whole? — but his and Schölzel’s framing of interpassivity as a way of resisting Staging resistance interactivity proves useful in thinking about alternative, critical strategies of UX design. When a touchscreen — an interactive map in The viewer of interpassive gaming a shopping mall, an information board at an videos might be escaping the demands of airport, a ticket machine — fails to respond interactivity, but this still happens through the to your touch (or turns out a regular, non- user-friendly, interactive frame of the person- interactive screen), it can feel awkward and al screen — there is no cognitive glitch. If we cause frustration; this kind of mismatch of ex- redefine interpassivity as non-participation pectations is something user-friendly design in the face of interactivity, then an ‘interpas- aims to avoid. But designers could harness sive’ interface would have to intentionally the negative feeling to transform dissatisfac- disregard the viewer-user, openly reject his tion into disillusionment, and disillusionment or her involvement, essentially prevent the into distance. Perhaps non-user-friendly user’s participation. While Safebook renders interfaces must appear out-of-order to create the content shared via Facebook invisible to disorder. Perhaps only a device that seems halt consumption, an interpassive version dysfunctional can originate a glitch that of Facebook would aim to do the opposite disturbs the hyper-industrial production line — overwhelm the user with an overflow of where the consumer’s consciousness serves

191 APRJA Volume 8, Issue 1, 2019 as raw material. the illusion of user-friendliness and disen- Stiegler defines films and songs as tem- gages the user from the system. poral objects because they are constituted To call our new, machine-enabled feel- by the time of their passing. The interface is ings into question, non-user-friendly design also a temporal object, but it supplants the would need to replace apparent mastery with looped temporality of cinema with continuous enabling vulnerability. It wouldn’t allow the change and never-ending updates. As the technology to gaze back at the user, respond flux of human consciousness is intertwined to his or her touch, heartbeat, or position in with that of the interface, they remain in a space. It wouldn’t be personalized or interac- reciprocal relationship, in passage together, tive. While user-friendly design conceals the affecting one another and adapting to each influence of the controller-screen, the every- other. Stiegler has recently admitted that day enhancer of sensation, non-user-friendly while the new, interactive screen could be design would revert the logic of touchscreen “a threat, enacted through the mediation sensibilities — without a complete erasure of the fully computational and automated of hardware through software. This kind of system,” it could also “constitute a chance, design would have to elicit a sense of confu- an opportunity to renew commentary, to sion; disorientation would form part of the reconnect with the ‘gloss,’ through a com- experience. pletely rethought hermeneutics” (Stiegler, I am using my iPad to edit this article. The Neganthropocene 173). To live a vita What if there was an app uploaded to my de- activa, he argues, we must hold on “to the vice that, once opened, would not allow me promise of a new hermeneutic epoch borne to navigate the unknown through automa- by these screens” (174). Non-user-friendly tisms and well-known gestures? What if this design could realize that promise by embrac- non-user-friendly software would transform ing interpassivity, uncovering a gap between — even if for a little while — the interactive human and machine feeling. To be deemed touchscreen of my personal device into a successful, the experience of non-interaction classical screen — a non-interactive surface must interrupt the illusion of control and perfor receiving projections — and turn me into sonalization, create an opening that divulges an (inter)passive spectator against my will? our very technicity. Would this kind of non-user-friendly design Speculations on non-user-friendly prove a means of paradoxical de-framing design modeled on interpassivity point to of contemporary perception and make the the potential of the existing touchscreen user aware of his or her own expectations infrastructures to de-automatize choice. The of sensation, shaped by design that appears relevance of the search for critical, alterna- user-friendly? To succeed, non-user-friendly tive UX design practices is progressively design would have to feel intrusive. It would becoming more evident, as the culture of have to make explicit the fact that the device ubiquitous computing moves on to more doesn’t belong to the user, that it doesn’t advanced sensors, AR/VR sets, holographic merely serve its master. Non-user-friendly projections, etc. The aim of this provocation design would leave the user dissatisfied, is to invite both users and practitioners to perhaps even angry. It would demand reconsider the potential of interpassivity in resistance. the age of total interactivity, to imagine a design strategy and design experience that reveals rather than covers up, that disturbs

192 critical and historical reappraisal of interpas- Works cited sive phenomena.” International Journal of Zizek Studies, 2, 2008. https://zizekstudies. Dunne, Anthony and Fiona Raby. org/index.php/IJZS/article/view/80/77. Speculative Everything: Design, Fiction, and Social Dreaming. MIT Press, 2013. Žižek, Slavoj. “The Interpassive Subject. Centre Georges Pompidou, Traverses.” Ekman, Ulrik. “Introduction: Complex http://www.lacan.com/zizek-pompidou.htm. Ubiquity-Effects.” Ubiquitous Computing, Complexity and Culture, edited by Ulrik Ekman, Jay David Bolter et al. Routledge, 2016.

Gekker, Alex. “Let’s Not Play: Interpassivity as resistance in ‘Let’s Play’ videos.” Journal of Gaming & Virtual Worlds 10, 2018, pp. 219-242.

Hansen, Mark. “Topology of Sensibility.” Ubiquitous Computing, Complexity and Culture, edited by Ulrik Ekman, Jay David Bolter et al. Routledge, 2016.

Schölzel, Hagen. “Backing Away From Circles of Control: A re-reading of interpassivity theory’s perspectives on the current political culture of participation.” Empedocles: European Journal for the Philosophy of Communication 8, 2017, pp. 187-203.

Sterne, Jonathan. “What if Interactivity is the New Passivity?” Flow Journal. https://www. flowjournal.org/2012/04/the-new-passivity/.

Stiegler, Bernard. Symbolic Misery. Volume 1: The Hyperindustrial Epoch, translated by Barnaby Norman. Polity Press, 2014.

Stiegler, Bernard. The Neganthropocene, edited and translated by Daniel Ross. Open Humanities Press, 2018. van Oenen, Gijs. “Interpassivity Revisited: a

193 Rosemary Lee

OPERATIVE IMAGE: AUTOMATION AND AUTONOMY

Abstract

Visual media is increasingly impacted by algorithmic approaches to image production, which introduce new modalities into existing notions of the image. Rather than a fundamentally new phenomenon, current methodologies instead expand upon the automation of image production described by previous theories regarding the technological character of the image. The “operative image” (Farocki) acts as a central theory to describe attributes of new forms of visual media engaged with algorithmic processes. Introducing and elaborating on the concept of the operative image, comparisons are drawn between existing notions of the image and new features which result from the use of algorithmic processes in the creation of images. This paper aims to develop an understanding of how algorithmic image production affects defining aspects of images.

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Rosemary Lee: OPERATIVE IMAGE

Introduction Operative Image

The image has undergone a remarkable The operative image is central to understand- transformation over the past few decades, ing algorithmic forms of visual media, as it in large part due to the increasing role algo- departs from previous notions of the image rithmic processes play in image production. which have tended to prioritise the visual Harun Farocki’s notion of the “operative attributes of images. Instead, the operative image” has been especially influential in image considers images in terms of the per- describing attributes of new forms of images formance of spatial procedures. In Farocki’s which estrange the point of view from the words, operative images “are images that human subject and eschew representation in do not represent an object, but rather are favour of the performance of machinic opera- part of an operation.” (“Phantom Images” tions. This draws upon changes in the nature 17) This kind of image is concerned with the of images which had already been ongoing performance of an operation, connected to for many years before they were highlighted the real by enacting a process, rather than by Farocki in 2001, but which reached critical representing something other than itself. mass in the military and governmental use It’s worth noting that the idea of the of intelligent machines and surveillance operative image was inspired by Roland technologies in the 1990s. This research Barthes’ concept of the “image-at-one’s invokes the operative image as a fundamen- disposal,” which he uses to describe the tal concept to understanding the paradigm potential for images and words to function shift toward algorithmic approaches to the in an instrumental (Parisi) capacity: “I ‘speak image. Images are increasingly automated the tree’, I do not speak about it. This means using machines, and more and more often that my language is operative, transitively this is done through opaque systems which linked to its object; between the tree and obscure the process behind the production myself, there is nothing but my labour, that of the image from human oversight. The is to say, an action.” (Barthes, Mythologies automation of visual tasks ultimately raises 146) The operation performed when “speak- questions regarding not only how the image ing the tree,” as Barthes refers to it, is at is to be defined in light of algorithmic image once an act of conjuring, which performs production but also as to the autonomy of a representational function, by bringing to artificial intelligence to produce images. The mind the mental image of a tree by invoking present investigation begins by introducing it by name, but words also function as instru- the concept of the operative image, which is ments, ways of interacting with reality. The then elaborated upon through examination of tree, here, is an implement for performing the the historical context which has led up to cur- concept of a tree. The image, tree, conjured rent image production. Following the themes in the process of using that word is opera- of automation and autonomy, the operative tive in the sense that it is a performative and image is then elaborated upon with regard functional conceptual image of a tree, which to these two historical tendencies in imaging is not fixed. Metaphor allows words to per- technologies. form with a great degree of variability upon the relations between the sensual properties of objects and the objects themselves (Harman). Language, in this instrumental

195 APRJA Volume 8, Issue 1, 2019 sense, grants different access to interacting what appears to be the system’s assess- with the real tree one encounters, and may ment of features in its environment. Different transform those encounters conceptually, as coloured, crudely drawn marks on the video well as ontologically (González Valerio). designate the edges of the road or various In his seminal essay “Phantom Images” obstructions in the path of the vehicle. And in and associated trio of video works, Eye/ scenes which Farocki mentions in “Phantom Machine I-III, Farocki introduces the terms Images,” footage taken by drones navigat- “operative image” and “operational image.”[1] ing autonomously in search of targets is An operative image, he explains, results from alternated with a human operator tasked with the performance of an operation (“Phantom watching the footage and overseeing remote Images” 6). Performing an operation through missile strikes. The contrast between human or as an image, Barthes notes, makes it into and machine vision is highlighted by these an action, rather than an object or mere rep- examples, which point to the autonomous resentation. The operative image, thus, is not quality of performing visual processing tasks to be thought of as necessarily representative automatically by computers, robots and of something else. Rather, it exists for itself, drones. to the degree that it is concerned with the execution of a spatial task, and may not point to something beyond itself. Additionally, the performance of an operative image tends to Automation prioritise the machine as the producer of this kind of image. This takes a radical departure While the process-oriented and non-optical from the representational paradigm, as the aspects of algorithmically-produced images outcome of the performance of an operative are contemporary issues, this shift is also image may or may not be visible to humans. deeply rooted in historical developments The point of view, thus, is shifted from the in the automation of image production. subject’s eye, to being situated in a machinic Automating aspects of the creation of images performance of an operation. The ‘I’, as well, through various techniques and machinery is displaced from the human subject to the has a long history, which has contributed viewpoint of an apparatus. to the context surrounding current forms of Farocki describes machines as pos- image production. In this section, a review of sessing a “sightless vision” reliant on compu- key examples helps to develop a background tational processes, such as the programmed against which to compare current trends in navigation of robots and drones. In his video image production. trilogy, Eye/Machine I-III, viewers are faced Algorithmic procedures have come to with several examples of what he means be a defining aspect of current visual media, by operative image. One scene features a especially due to the amount of visual pro- robot performing tasks autonomously, cutting cessing tasks are now commonly delegated between shots of the robot moving around in to computers. They are encountered fre- a room and shots taken from its point of view, quently, playing a role in the creation of con- highlighting written numbers in colour as if to tent, in determining what is visible to whom indicate the robot’s reading those as salient on the web and in governance through mass- features. In similar fashion, video clips from surveillance. In light of this shift, the nature what appears to be a navigational assistance of the image can no longer be solely under- system are overlaid with markings indicating stood in terms of previous formulations which

196 Rosemary Lee: OPERATIVE IMAGE frame the image as a fixed, visual outcome, to film, or “pre-cinema”, saw the creation of such as a developed and printed photograph. a variety of optical gadgets and machines A digitised version of the same photograph, which activated images in various ways, from for instance, is easily recognisable, but it is illumination to animation. Cinema set the im- a product of drastically different technologi- age in motion through variations of multiple cal conditions, governed by computational images, simulating movement: the “move- processes in lieu of the mechanical, optical ment-image” and the “time-image” (Deleuze). and chemical processes employed in analog Digital images allowed the electronic coding, photography. In this sense, algorithmically- display and circulation of images, and this produced images expand upon existing was pushed even further with the use of the forms of automated production, placing em- internet. Networked images, as Alexander phasis on the execution of formal procedures Galloway points out (94), may be displayed in addition to their optical properties. An on innumerable computers simultaneously, algorithm, it is worth noting, is a “process or adding to the mass-transmissibility and intan- set of rules to be followed in calculations or gibility of the image. Generative art went on other problem-solving operations, especially to consider the artistic potential of employing by a computer” (“Algorithm”). The operative autonomous systems to produce images. image takes a fairly broad interpretation of In addition to the technical modes of the this definition, which is useful as we expand automation of image production previously our approach to image-production processes described, formalising processes of artistic that at first glance may not appear to be al- creation in terms of algorithmic behaviour also gorithmic in the more familiar, contemporary explored the dynamics of human-machine sense of complex computational processes, relations. Rather than a fixed outcome from but instead embody procedural processes image-production processes, the operative toward the execution of an image. image may be performed or it may be trans- The problem that automating pro- coded as sets of instructions. Several artists cesses of image production posed to existing who were early-adopters of using computers notions of aesthetic value in images was in their work also experimented with taking famously wrestled with by Walter Benjamin on a performative role in the production of in his essay “The Work of Art in the Age of images, placing the emphasis on process. Mechanical Reproduction.” Mechanising the Vera Molnár, for example, is known for her production of images enabled multiples to “machine imaginaire,” which implemented be produced quickly via technologies such instructions for the production of visual as the printing press and the photographic outcomes, the artist herself taking on the process, and also enabled the mass-dissem- conceptual role of a computer, one which (or ination of images. The facilitation of the serial whom) computes, performing tasks based on reproduction of images undermined the aura a set of predefined rules (“Image Machine” of the original artwork, which had been a 141-142). Taking on this kind of instrumental mainstay of artistic valuation up to that point. role has been a recurring theme in several Artists including Andy Warhol and later the avant-garde movements in the 20th century, art and design group Superflex have played importantly the Surrealists’ engagement with upon the aspect of seriality, making multiples the concept of automatism. They approached of images to undermine the notion of the the mechanisation of art by advocating that copy as inferior. In other developments in artists relinquish conscious control over the the mechanisation of the image, precursors artistic process so as to arrive at art produced

197 APRJA Volume 8, Issue 1, 2019 by the subconscious mind. Automatic writing, as being critical to their technical character. drawing, and painting led artists to develop Flusser describes “technical images” as methodologies seeking to elude their own those images which have supplanted texts, consciousness, often by employing highly not only those which owe their existence to systematised, rule-based techniques to sur- technical apparatus in a direct sense (7). In render creative control by engaging with ser- addition to their technical mode of produc- endipity and randomness. In many instances, tion, the codification and instrumentation of the artist expressly sought to hand over images also adds to their technical and tex- agency, intentionality, or control to a process, tual character. Ingrid Hoelzl and Rémi Marie machine or system. One of the most famous make a correlation between the algorithmic and influential methods to materialise from nature of digital images and the history of this kind of aleatory approaches (Carvalhais) cartography (99), which shifted from thinking is the “cut-up method” which Brion Gysin is of maps as representations of the world in credited for,[2] a process in which a linear pictures to such a representation instead text would be dismembered at random and taking the form of a data set. Cataloguing the rearranged by the artist, influencing the crea- systematised coordinates marking the location of a new work from the rearrangement tions of geographic features and their relative of an existing one. Conceptual artists such spatial relations as an index of mathematical as Sol LeWitt, Yoko Ono, John Cage and information made it possible for Ptolemy’s Lawrence Weiner have similarly employed atlas of maps, Geographia, to be saved, sets of rules in the creation of their works. transmitted and later reconstructed. The pro- Thinking of the process as a form of machinic cess of transcribing a visual representation, or programmed image-production grasps the in this case, a map, from image to numerical operative property of performing algorithmic data and back into an image allows us to see processes. Implementing rule-based sys- a close parallel in other image processes, tems such as in LeWitt’s instruction-based namely, the digital. Considering this index of drawings, the artist gives directions for the coordinates as a set of instructions or source- construction of the work, which may be ex- code for the reconstruction of the maps, ecuted with some degree of variation. though simplified and analogue, is much like Returning to the importance of text to the instructional aspect of digital images. In the origins of the operative image, as was a similar fashion, the canon of proportions apparent in Barthes, several thinkers have outlined by Vitruvius in his De architectura explored how relations between images and describes representation of the human body texts contribute to their algorithmic qualities. geometrically, as if to function as instructions In his enquiries into what he terms “image- for its reconstruction: “The length of the texts,” W. J. T. Mitchell demonstrates the foot is one sixth of the height of the body; various modes of interrelation between im- of the forearm, one fourth; and the breadth ages and texts. Rather than merely referring of the breast is also one fourth.” (Vitruvius) to reality, as the image functions in represen- In this and the previous pre-digital tational terms, imagetexts consider the inter- example, mathematical formulae and the relation between objects, texts, and images, systematic cataloguing of the internal rela- and their potential to be enacted through tionships within images enabled them to be various forms of mediation. Similarly, Vilém transcribed, stored, transmitted and reiter- Flusser explored textual aspects of images ated. Not only did this enable a great deal of

198 Rosemary Lee: OPERATIVE IMAGE new possibilities, for instance making itera- of that visual data, as well as the procedure tions of an image, it also allowed a degree which governed the end result is obscured. of faithfulness to be maintained within the In situations such as when a camera or copies. other instrument serves as a stand-in, taking the place of the eye, technology enables humans to see in ways impossible to the naked eye, but also steals away other aspects Discussion one expects in an image. What is apparent to human viewers observing input intended To speak of automation begs the question of for machines is that operative images func- autonomy. In this case, one must ask what tion based on different parameters and are the operative image means not only for the not necessarily burdened with any need to production of images by machines, but also communicate with human vision. This quality with respect to the sharing a viewpoint with makes them decidedly different from previ- the inanimate (Virilio 59). In what has been ous conceptions of images. Not only has referred to as the “algorithmic turn” (Ulricchio, the machine been thoroughly accepted as Hoelzl and Marie), what is visibly apparent on a surrogate for the eye, but in some cases, the surface of an image is only one aspect of such as the instances covered by Farocki’s the processes at work in algorithmic media. Eye/Machine, the eye may be dispensed of The information content of a digital image, entirely. A consequence of distancing visual for example, is largely unvisualised, acting perception from the eye through apparatus as the code for its enactment, often through is articulated through operative images as an screens, and the image takes on a more dy- ever more blurry boundary between human namic quality than a static entity. The visible and nonhuman agency. Not principally of the surface of the digital image is subjugated to human, by the human, nor for the human the invisible “subface” (Nake) behind it, for (Zylinska 5), nonhuman forms of images fulfil instance, in contexts involving the automated Virilio’s prediction regarding the automation processing of spatial data, where it may or of perception through cameras controlled by may not be necessary to visualise the end computers (59). The result of this automation result, in looking at the metadata attached of vision, a splitting of the viewpoint with the to an image, or in comparing two seem- inanimate (Virilio 59), entails that these im- ingly identical images which were produced ages are far from being self-evident. Looking using different algorithms. The algorithmic alone is not sufficient to thoroughly grasp processes responsible for what is eventu- what is at stake in the output of algorithmic ally visualised as a digital image may vary modes of image production. As a conse- greatly, whether or not those differences are quence of the operative image, the range of visibly discernible to the human eye. But what may be considered to be an image is especially notable here is that when using expanded to include non-optical, algorithmic algorithmic approaches such as machine processes, prioritising process over the im- learning to generate images based on vast age’s visible qualities. amounts of training data, entire databases The operative image is significant, not of images are subsumed by the resulting only because it alters what, ontologically images. In a sense, such images are similar speaking, may be defined as an image, but to composites merging the numerous images it also extends the role of image produc- which an algorithm was trained on, yet much tion beyond the human to autonomous (or

199 APRJA Volume 8, Issue 1, 2019 semi-autonomous) enactment by machine. Automating aspects of the creative process Conclusion calls into question some of the entrenched value systems surrounding images, namely Much as historical reckonings with tech- authorship (Ward and Cox). The aura of au- nological modes of production such as the thorship remains an enduring issue at stake advent of the printing press, photography in the production of images by machines, as or cinema led to reevaluations of the image, evidenced by the current hype around works the current gravitation toward algorithmic of dubious artistic quality (Obvious) which processes has led to new understandings make claims to the autonomous creativity of of the defining attributes of images. Rather machines. Concerns around authorship are than a fundamentally new phenomenon, never far behind discussion of automating current methodologies instead expand upon the production of images by machine, the the automation of image production which potential of autonomous artistic creation has been in progress for decades and even by machines threatening the “death of the centuries. Reformulating the image as an author” (Barthes). Curiously, Harold Cohen operation which is performed as opposed to alternated between signing images produced the fixed outcome of the creative process, using the artificial intelligence software he the operative image offers an entryway to created with his own signature (Amsterdam rethinking the context surrounding the au- Suite A) and that of AARON (20:28). This tomation of image processes which current lends the artworks a sense that Cohen may media build upon. Developing the concept have either felt conflicted as to his role in cre- of the operative image through an overview ating the artwork, and that there may have of historically-significant theories and exam- been a feeling of competition for authorship. ples, this research aims to develop an under- In a general sense, the images produced standing of how the concept of the operative using machine learning, too, carry with them image contributes to a reevaluation of the a spectre which has haunted technologically- image in light of new modalities introduced engaged images throughout the past century, by algorithmic media. what Andreas Broeckmann calls the myth of the machine as artist. The persistent curios- ity surrounding the creation of autonomous agents which in turn create art relies upon the tradition of conceptually separating science and the humanities. While machine learning enables the automation of certain tasks, it also lends itself to a mystification of the process of image production. Image production by intelligent machines offers new technical and conceptual possibilities, it also brings to light certain existing issues which have persisted throughout the past century, including automation, seriality, transcodabil- ity and human-machine relations.

200 —. “The Machine as Artist as Myth.” Arts, Notes vol. 8, no. 1, February 2019, p. 25. Burroughs, William S. “The Cut-Up Method [1] Farocki uses the two words, operative of Brion Gysin.” The New Media Reader, and operational, for the most part inter- edited by Noah Wardrip-Fruin and Nick changeably in his video work and writing. Montfort. MIT Press, 2003. Other thinkers, including Trevor Paglen and Jussi Parikka have gravitated toward Carvalhais, Miguel. Artificial Aesthetics: operational, but the author chooses to use Creative Practices in Computational Art the former, operative, as it indicates the and Design. U. Porto Edições, 2016, pp. sense of agency expressed by machines in 145-178. the performance of operational images. Cohen, Harold. 20:28. 1985. [2] William S. Burroughs is also known for popularising the practice of the cut-up —. Amsterdam Suite A. 1977. method. Deleuze, Gilles. Cinema 1: The Movement- Image, translated by Hugh Tomlinson and Barbara Habberjam. University of Minnesota Works cited Press, 1986.

“Algorithm.” Oxford English, Oxford —. Cinema 2: The Time-Image, translated University Press, 2019. https:// by Hugh Tomlinson and Robert Galeta. en.oxforddictionaries.com/definition/algo- University of Minnesota Press, 1989. rithm. Accessed 14 May 2019. Farocki, Harun. “Phantom Images.” Barthes, Roland. Mythologies. Editions du PUBLIC, edited by Saara Liinamaa et al., Seuil, 1957. vol. 29, 2004, pp. 12–22.

—. “The Death of the Author.” Image-Music- Flusser, Vilém. Into the Universe of Text, Fontana Press, 1977, pp. 142–48. Technical Images, translated by Nancy Ann Roth. University of Minnesota Press, 2011. Benjamin, Walter. “The Work of Art in the Age of Its Technological Reproducibility: Galloway, Alexander. “Are Some Things Second Version.” The Work of Art in the Unrepresentable?” Theory, Culture & Age of Its Technological Reproducibility and Society, vol. 28, no. 7–8, December 2011, Other Writings on Media, edited by Michael pp. 85–102. W. Jennings et al., Harvard/Belknap Press, 2008, pp. 19–55. González Valerio, María Antonia. “Politics of Machines or the Anthropogenenesis as a Broeckmann, Andreas. “Image Machine.” Confrontation with a Certain Kind of Nature.” Machine Art in the Twentieth Century, MIT Open Hardware/Open Machines, Aalborg Press, 2016, pp. 123–64. University, Copenhagen. 14 May 2018.

201 APRJA Volume 8, Issue 1, 2019

Harman, Graham. “Aesthetics Is the Root of Vitruvius Pollio, Marcus. De Architectura, 15 All Philosophy.” Object Oriented Ontology: A A.D, 3.1.2–3. https://en.wikipedia.org/wiki/ New Theory of Everything. Pelican Books, Vitruvian_Man. 2017, pp. 61–102. Hoelzl, Ingrid. “Postimage.” Posthuman Ward, Adrian, and Geoff Cox. “How I Drew Glossary, edited by Rosi Braidotti and Maria One of My Pictures: or, The Authorship of Hlavajova, Bloomsbury Academic, 2018. Generative Art.” International Conference on Generative Art 1999, Milan 1999. http:// Hoelzl, Ingrid, and Rémi Marie. Softimage: www.generativeart.com/on/cic/99/0399.htm. Towards a New Theory of the Digital Image. Intellect, 2015. Zylinska, Joanna. Nonhuman Photography. MIT Press, 2017. Kittler, Friedrich. “The Finiteness of Algorithms.” transmediale. https://transmediale.de/content/the-finiteness-of-algorithms.

Mitchell, W. J. T. Iconology. University of Chicago Press, 1986.

Nake, Frieder. “Surface, Interface, Subface: Three Cases of Interaction and One Concept.” Paradoxes of Interactivity: Perspectives for Media Theory, Human- Computer Interaction, and Artistic Investigations, edited by Uwe Seifert et al., Transaction Publishers, 2008, pp. 93–109.

Obvious. Edmond De Belamy. 2018.

Parisi, Luciana. ‘Instrumental Reason, Algorithmic Capitalism and the Incomputable’. Alleys of Your Mind: Augmented Intelligence and Its Traumas, edited by Matteo Pasquinelli, Meson Press, 2015, pp. 125–137.

Ulricchio, William. “The Algorithmic Turn: Photosynth, Augmented Reality and the Changing Implications of the Image.” Visual Studies, vol. 26, no. 1 (Mar 2011): 25–35.

Virilio, Paul. “The Vision Machine.” The Vision Machine, British Film Institute/Indiana University Press, 1994, pp. 59–77.

202 203 Carleigh Morgan

CALCULATED ERROR: GLITCH ART, COMPRESSION ARTEFACTS, AND DIGITAL MATERIALITY

Abstract

This paper proposes a reconsideration of the aesthetic category of ‘glitch’ and advocates for a more careful theorisation around indexing — in the sense of both locating and naming — errors of a digital kind. Glitches are not as random as they seem: they are ordered and shaped by computational hardware and software, which impose a mathematical rubric on how glitches visually manifest and set ontological and technological constrains on glitch that limit how digital errors can and cannot be made to appear. Most crucially, this paper thinks about how one particular type of glitch — a compression artefact called a macroblock — can often appear as random, erratic, or unpredictable but is, in fact, materially constrained and visually conditioned according to the principles of computing and computer design. At its core, compression aesthetics can shed light on the operations of algorithms, the structures of digital technologies, and the priorities and patterns which occur as a function of algorithmic manipulation. The randomness, unpredictability, or messiness which glitch studies invokes around the glitch is in danger of overlooking the ways that the material architectures and algorithmic protocols structure the digital glitch by organising, constraining, and given form to its appearance.

APRJA Volume 8, Issue 1, 2019 ISSN 2245-7755

CC license: ‘Attribution-NonCommercial-ShareAlike’. Carleigh Morgan: CALCULATED ERROR

Bodies and machines are defined marked by fluctuating glitch effects. by function: as long as they operate All of these works owe at least one of correctly, they remain imperceptible. their particular stylistic effects to the process — Maurice Merleau-Ponty, known as compression hacking. This paper Phenomenology of Perception (239). examines how compression hacking works as process of algorithmic manipulation and In 2005, Takeshi Murata released a short considers what the artistic practice of com- film called Monster Movie. This film show- pression hacking exposes about the com- cased a swamp creature emerging from position of digital images. The key argument the muck that was noteworthy for its unique is twofold: first, that the particular effects visual effects: blocks of mutating pixels that produced by compression hacking are deter- seemed to burst through the monster’s body, mined by the computational processes and deconstructing its image into a scattered material properties of digital media; second, and murky pixel array. Other glitch art, like that the algorithmic functions that are used David O’Reilly’s Compression Reels and the in compression hacking establish the condi- net 2.0 aesthetics of cyberpunk art collective tions by which compression artefacts can PaperRad, were sparked by similar inter- appear, but that sometimes these compres- ests in exposing the underlying algorithmic sion artefacts remain invisible. Nevertheless, protocols and structuring interfaces of digital the production of compression artefacts as media. In the 2000s, glitch style migrated a result of compression hacking depends from its origins on niche punk-art message not only on certain level of algorithmic boards and underground websites to become functionality, but also on the matter of digital incorporated into commercial music videos technologies: compression artefacts, like all for both Kanye West and electro-pop group glitch effects, owe their various materialisa- Chairlift. Directed by Ray Tintori, Chairlift’s tions to technologies which are not entirely music video for Evident Utensil used glitch dysfunctional. In other words, malfunction is effects to create a visual aesthetic marked borne out of function: a digital error depends by an array of mutating colour blocks that on the enduring functionality of the systems fused the band with their surrounding envi- which give rise it, to make such an error leg- ronment and seemed to rupture the diving ible as out-of-the-ordinary. One corollary to line between the environment staged in the this argument is that the presence of what video — its content — and the colours on appears to be a glitch in a digital image does the surface of the screen — its form. Kanye not always indicate the presence of an un- West’s Welcome to Heartbreak achieved a derlying technological error — a claim which more fastidiously controlled, choreographed unsettles the notion of technological trouble- style of glitch art that combined chromakey shooting and the heuristics underpinning a and green screen techniques to unsettle ‘diagnostics’ of technological failure. the grammars of commercial video edit- To these ends, this paper proposes a ing. Starring West and featured singer Kid reconsideration of the aesthetic category Cudi, the music video depicted the rappers of ‘glitch’ and advocates for a more care- ‘melting’ into each one another, alternating ful theorisation around indexing — in the recognisable fragments of their faces with sense of both locating and naming — errors sequences of digital skids and bleeds that of a digital kind. Most crucially, this paper fractured the representational image and thinks about how glitches — which often transformed it into an unstable landscape appear as random, erratic, or unpredictable

205 APRJA Volume 8, Issue 1, 2019

— are materially constrained and visually descriptor which seems to describe the quali- conditioned according to the principles of ties of compression artefacts themselves computing and computer design. Glitches — e.g. ‘moshing’ conjuring a pixel-based are not as random as they seem; in fact, they modularity and squishiness — it does not are ordered and shaped by computational foreground the technological dimension of hardware and software, which impose a this artistic process as clearly as the term mathematical rubric on how glitches visually compression hacking does. ‘Datamoshing’ manifest and set ontological and technologi- elides the role that the artist plays in refor- cal constrains on glitch that limit how digital mulating the video files to produce visual errors can and cannot be made to appear. compression artefacts, and in so doing intro- This paper will emphasise compression duces confounding variables into the discus- hacking for a few reasons. The first reason sion that this paper does not have the scope is to draw attention to the human labour and to address. Thinking about how to achieve the activity of ‘hacking’ which generates com- compression artefacts as a product of pression artefacts. It also reinforces that this compression hacking means that less tech- artistic practice is achieving by playing with nologically invasive methods for achieving the computational logics of compression — its stylistic effects — so datamoshing done compression hacking requires a modicum of by applying a photo or video filter through computational literacy and is an activity un- the implementation filters via programs dertaken by someone who understands how like Photoshop — can be set to one side. to manipulate the information encoded in im- Although any file format can be compression age or video files. With an emphasis on the hacked, this paper will focus on digital video/ ‘hacking’ of compression hacking, it is clear moving images for two reasons: because this that this paper will not address compression is the format which has mostly received the artefacts or glitches as spontaneously oc- attention of compression hackers. curring: the epistemic frameworks used to diagnose a digital error in the instance of a glitch’s spontaneous occurrence would act as a confounding factor. Simply put, looking On compression algorithms at the glitch generally, rather than at glitch art specifically, one is forced to contend with The law of information processing upholds other variables pertaining to the origin of a that the “fewer states one needs to process glitch. To think of glitch as a homogenous a message, the faster and more efficient the aesthetic form rather than glitch art as a spe- system is” (Kane 220). Data compression cific technological practice — or to begin with follows this law by simplifying how data is compression artefacts rather than compres- stored. The purpose of data compression sion hacking — means grappling with the is typically to optimise storage space or notion that the glitch appears despite there increase data transmission rates, and it is being no known intervention from an outside often motivated by a desire to save both time agent (e.g. artist, hacker, programmer) who and money. Compression algorithms record can testify to a glitch’s cause or represent a only the measurable changes in the image reason for its occurrence. data. As a result, only areas of a moving There is another terminological clarifi- image which describe differential motion or cation to make here. Although the term ‘da- changing luminance values are captured tamoshing’ operates as an onomatopoetical by the compression algorithms (Arcangel).

206 Carleigh Morgan: CALCULATED ERROR

According to this principle, images with fewer In order to dramatically reduce the amount substantial changes from frame to frame are of data that needs to be stored, what is easier to encode. “The whole point of digital captured in a compressed video file is only image compression,” Cory Arcangel writes the difference between the initial, or i-frame “is to be able to reconstruct an image without and the subsequent images, the p-frame, having to send all the data.” Lossless com- (sometimes called the delta Δ frames for pression, as the name might suggest, does this reason). These later frames contain the not lose any information from the original image’s transform instructions of the initial source during any point of the compression or keyframe. The illusion of object motion in (encoding) or decompression (decoding) pro- an image or the appearance that the image cess. In his short treatise “On Compression,” itself is moving is determined by relationship Arcangel develops a very clear analogy that between the p frames and the i-frames. If this captures this method of data-optimization in relationship is thought of as the difference in non-technical terms: motion interpellated over time, “subsequent frames could be described as a catalogue of Let’s say we wanted to send this: pure differentiality” (Levin). In addition to key- ‘aaaaaaaaaba’ and we were going frames and predictive frames, there are also to send it over the phone by voice. b-frames: these are similar to p-frames but As opposed to having to send all a b-frame references the frame both before the information by reading out each and after it. Modifying b-frames leads to more letter one at a time, we could just tell unpredictable results than modifying only the someone ‘9a’s, one b, and one a’ keyframes and predictive frames (Arcangel). and they would know we meant ‘a a In short, compression algorithms control the a a a a a a a b a’ and we have saved behaviour of several kinds of frames. When ourselves a bit of breath. In computer combined, these frames act as a catalogue language it means we have stored all of movement, and therefore are functions of the information using less space. time — they measure the differences in image data from frame to frame. Digital video files are composed of Compression hacking creates a new sequences of different types of frames: merging reference between the elements of ‘i-frames’ or initial frames — commonly an original image frame and the successive called keyframes — “are full representations frame. When compression hacking does of a single frame of a video” (Arcangel). In yield visible compression artefacts, they essence, a keyframe is simply a still image occur as a direct result of ‘playing around’ containing all the colour and luminance data with the relationships between the initial of a particular frame and are typically used frames and the predictive frame to create as reference points by animators. In digital digital images characterised by breaks, folds, animation as in hand-drawn animation, key- ruptures, skids, mutations, and pixelated frames are important for determining where blots. “Macro-blocking, pixelating, checker- and when an animation sequence starts boarding, quilting and mosaicking” (Levin) or stops. Predictive, or ‘p-frames,’ on the are kinds of compression artefacts. These other hand, are reference files that inform descriptors capture how these artefacts ap- the video player of changes to the image’s pear as geometric forms; their behaviours compositional arrangement that have oc- and appearances are visibly linked to or- curred since the previous frame (Arcangel). ganisation of a computational grid arranged

207 APRJA Volume 8, Issue 1, 2019 by Cartesian coordinates, a point that will and decoding process of compression, as important to remember. a sign of the image’s technological corrup- Compression artefacts are made vis- tion. However, simply identifying that there ible through the use of lossy compression are compression artefacts within an image algorithms, whereby some of the information is not sufficient evidence for diagnosing the about an image is ‘lost’, although it is more presence of a technological error within the accurate to classify the lost information as compression algorithm itself. The algorithmic unnecessary, surplus, or disposable. Lossy behaviour of the compression algorithm and compression can occur for a variety of rea- the creation of compression artefacts are, sons, but it is not bad — in fact, in instances necessarily, behaviourally linked—but their of low bandwidth or limited storage space, behaviour is not identical. In other words, im- lossy compression is desirable. Lossy com- ages that appear ‘glitched’ are not always pro- pression removes or replaces the initial key- duced by malfunctioning code. Particularly in frames and/or predictive frames in a video the case of compression artefacts using lossy file. It can also cause the “playback image compression, ‘corruption’ within an image and motion-vector data to distort the result- is a matter of artistic perspective. The next ing moving image with unpredictable results” section will briefly examine how compression (Goriunova and Shulgin 91), but it is worth artefacts fit into longer theorisations about noting that the information lost during the the visualisation of technological failure in compression process is not always detected modern and postmodern culture. by the human eye. Often the loss of information is of no great consequence, but the fact that data is lost during lossy compression means that it is limited in its application: Locating the glitch lossy compression techniques applied to text documents, or “any application where all the As a discipline, glitch studies are a relatively information must remain intact” (Arcangel) new area of academic research that has nev- would render the text file unreadable and ertheless furnished a prodigious amount of unable to be restored to its original condition. scholarship in recent years. Despite the high Despite the economic, temporal, and volume of cross-disciplinary contributions logistical advantages offered by lossy com- to glitch studies — from filmmakers, aca- pression, images or data which undergo demics, programmers, and para-academic lossy compression are frequently thought of practitioners — very little academic work has as a downgraded copy of the original image directed its focus towards the artistic practice or data file (Brown and Kutty 168). These of compression hacking. Perhaps one of the downgraded versions of an image or data set difficulties in charting work on compression are optimised for easy storage and retrieval hacking is due to the fact that glitch studies rather than for visual fidelity or clarity. But the is particularly prone to semantic inconsisten- central role of visuality in contemporary digi- cies, especially because the scholarship on tal culture means that the aesthetic of lossy glitch is often positioned as a history of the compression is often read as one typified by present. Keeping pace with the rapid aes- visual blemish or corruption on the surface thetic transformations ushered in by digital of the image. These blemishes are read as media may pose problems for glitch schol- evidence of a technological error which oc- ars, whose subject of study may fluctuate curred at some point during the encoding as rapidly as the internet and the aesthetic

208 Carleigh Morgan: CALCULATED ERROR formations it produces. For example, some explosions may have been one of the few digital media scholars refer to compression ways that workers were given a glimpse hacking by another name — ‘datamosh- into the structural and operational logic of ing’[1] — or simply subsume compression the machines in their midst. By violently artefacts into larger historical and cultural exploding, industrial machines dramatically considerations of the ‘glitch’ as a general exposed their interlocking mechanisms — aesthetic category. the machinic accident might be understood a While compression hacking can be spectacle of the machine’s operational logic. situated within the domain of glitch as an Error, malfunction, breakdown — these states artistic practice, it also fits into a longer cul- presuppose a stability, a rationality, and order tural and historical practice of theorising the from which the accident can erupt. From this technological accident. The spectre of error teleological perspective, the accident acts — alien ‘glitches’ in a system — haunts this as a necropsy to dissect the malfunctioning long era of the technological, invading every- machine — one need only be reminded of thing from the industrial advances in steam ‘exploded-view-diagrams’ today to consider locomotion to computer science to drone how the accident testifies not only to the warfare. Like the character Wintermute in structure and teleology of a machine, but William Gibson’s Neuromancer, glitches are also how “every technology carries its own frequently conceptualised as ghostly forces, negativity, which is invented at the same time malfunctions that interrupt the normal opera- as technical progress” (Virilo 89). tions of technological systems by seemingly The concern over an unpredictable or emerging “out of nothing and from nowhere,” sinister machinic vital force persists today. As giving viewers “a fleeting glimpse of an visual cultural theorist Carolyn L Kane writes: alien intelligence at work” (Vanhanen 46). “computers and algorithmic systems are pro- This theory traces its roots to anxieties that gressively given authority over human action attenuated the industrial and technological and experience […] yet we have a dwindling shifts demarcating the late Victorian from the capacity to recognize [sic] this” (219). Viewed Early modern period, which were populated from afar, she hypothesises that ‘the entire by stories of the technological gothic: “ghosts history of modern art could be construed as in the machine” depicted the “threat to the a glitch and compression of Enlightenment humans subject posed by an autonomous, epistemology” (Kane 219). In “Datamoshing uncontrollable technology” (Rutsky 125). as Syntactic Form,” Thomas Levin articu- Indeed, before ‘glitches’ came to be lates a theory of compression hacking that known as such, the ubiquity of the unnamed ties it to anxieties produced by “the miscom- accident was a frequent source of terror for munication between sender and receiver”. people of the industrial age who struggled He cites this transcoding error as distinctively to come to grips with the provenance and cybernetic, operating historically as an aes- cause of technological catastrophe. Many in- thetic that “exposes societal paranoia by dustrial technologies did not have monitoring illustrating dependence on the digital and systems, failsafe options, or the emergency fear of system failure [and] with the advent stops. As such, industrial machines were of video sharing sites like YouTube […] the constantly threatening to malfunction — one glitch aesthetic has evolved into a pop culture way of treating the factory explosion is not just artefact” (Levin). Casey Boyle advocates for to read it for its catastrophic effects, but also an art theoretical approach to glitch that em- to see in it a perverse rationality. Machinic braces it as a generative practice — and not

209 APRJA Volume 8, Issue 1, 2019 merely as a materialisation of technological computation, which lies under, over failure — because glitches can “render ap- and in the interstices between the parent that which is transparent by design” modular elements if an increasingly (12). Greg Hainge argues that the glitch computational society. (57) foregrounds “how technology always relies on the successful inclusion or integration of failure into its systems” (27). Perspectives like those of Hainge valorise technological Calculated error failure as necessary to technological function Rather than think of this breakdown in the and recuperate the glitch from its negative as- sense of dysfunction, it can be understood in sociations. One ramification of Hainge’s view the sense of take apart. Although it may seem is that the line between something which is like metaphorical hair-splicing, reconsidering ‘noise’ and something which is ‘signal’ is not what is meant by such a breakdown may be an expression of a technological boundary at a crucial avenue for exploration the aesthetic all. Instead, ‘glitch’ is a phenomenon which features of the digital glitch, which reveals can call into question the very stability and itself at the fault-line between breakdown determinability of the distinction between as an entropic activity and break-down as a signal and noise. structuring principle. Within this larger framework of glitch Consider one particular type of com- theorised as an emergent unpredictability in pression artefact: the macroblock, as shown machine function, it is not surprising to see in the image above. Macroblocking visually compression hacking described as a practice destabilises the representational legibility of which brings to the surface of the image the a digital image while simultaneously rear- operational failures of digital systems. But ranging it into ordered blocks. There are compression hacking is not the result of a small sections where the outline of an object technological accident. To consider com- appears, only to be abruptly cut off by large pression artefacts the result of accident, or to blocks of colour. Macroblocking can unsettle think of the compression artefact as ‘glitched’ the ability of a spectator to apprehend an is to both deny the artistic labour which image as representationally legible, even produces compression hacked images and when traces of recognisable objects remain to misunderstand the relationship between within the image. Macroblocking creates visual and technological malfunction. the impression of a carefully controlled Still, it is possible to situate compres- digital schizophrenia in a moving image: the sion hacking in what David M. Berry calls the shifting location, luminance, and colour of ‘new aesthetic’ (NA) a form of “‘breakdown’ macroblocks combine to create an impres- art linked to the conspicuousness of digital sion of movement that seems to originate technologies” (56). Berry writes: in the screen’s pixels. These pixels seem to scatter, breakthrough, or penetrate the digital We might conclude that the NA is the materiality of the screen, moving with regi- cultural eruption of the grammatization mental precision. Jeff Donaldson describes [sic] of software logics into everyday these kinds of artefacts as “a break from an life. The NA can be seen as surfacing algorithmic flow [whose] unanticipated ap- computational patterns, and in doing pearance is simultaneously frustrating and so articulates and represents the mesmerizing”. unseen and little-understood logic of

210 Carleigh Morgan: CALCULATED ERROR

Figure 1: GlitchTVbot (2019). _Image_[Twitter]/ Available at: https://twitter.com/GlitchTVBot/status/1141287271922110464 (Accessed 03 June 2019).

But a macroblock does not actually Using a lossless compression algorithm consist of migrating pixels. A macroblock, yields no loss of data; whether one reports occasionally called a ‘blocking artefacts,’ is ‘aaaaaaaaaba’ or ‘9a’s, one b, and one a’, a distortion in a compressed image that ap- the information remains unchanged. But if pears as a collection of pixel blocks. It can the example above used lossy compres- occur for a variety of reasons, but what is sion, ‘aaaaaaaaaba’ would be simplified to most important about macroblocking is that it ‘roughly 10 a’s’. A similar process, called relies on the mathematical principles of com- quantisation, occurs to create macroblocks. pression in image processing and does not Instead of capturing all of the detail in a feature the motion of pixels themselves. Let’s particular region of an image using a detailed return to Arcangel’s vernacular description of range of values, the compression algorithm compression for a moment: encodes only a single value for a particular region. If an image undergoes a great deal Let’s say we wanted to send this: of lossy compression, the amount of informa- ‘aaaaaaaaaba’ and we were going tion that is reduced during the compression to send it over the phone by voice. process may be significant; in the case of As opposed to having to send all macroblocking, the information the remains the information by reading out each after lossy compression may only pertain to letter one at a time, we could just tell the average colour of a collection of pixel someone ‘9a’s, one b, and one a’ blocks, thereby rendering an area that was and they would know we meant ‘a a formerly populated by multiple colours and a a a a a a a b a’ and we have saved luminance values into a block of a single ourselves a bit of breath. In computer colour. language it means we have stored all the information using less space.

211 APRJA Volume 8, Issue 1, 2019

Donaldson’s description of the logic of the usage of ‘pixel’ in the colloquial sense, computational failure is of particular signifi- typically used to suggest that they are the cance in understanding how the architecture atomic components of a digital image which of digital devices shape the images which can be decomposed and rearranged, pixels they manifest. His remarks are interesting to are not building blocks. Furthermore, a single reproduce here for their emphasis on the pixel pixel is not mobile — despite the descriptions as a format which is defined by a specific set of swirling, mutating, or bleeding pixels used of algorithmic architectures and confined by to characterise the glitch aesthetic, pixels the material makeup and organisation of the remain fixed, and they are not visible to the computer display: unaided human eye. In fact, a pixel is a unit of data that is “fundamentally ambivalent It is a true type of machine art and to vision” (Baraklianou 306) tethered to its a crude form of artificial intelligence array, and intimately connected with it sur- inasmuch that once an algorithm is let rounding pixels. The pixel is arranged with go to run free, due to the architecture others like it into a pixel array — always a of digital systems, a break from routine geometric formation — and the visual quali- creates an ordering of its own. The ties of this array depend on the behaviours pixel grid of the computer display of pixels which constitute it and surround it; provides the framework and serves as colour and luminance are not determined by the canvas for this manifested algo- the value of a single pixel but are “assigned rithmic hiccup. It’s as if the computer is at a later stage” and depend strongly on freed from its normal task and instead the relationships between pixel elements displays what it [sic] wants, the (Baraklianou 306). Whereas the pixel ele- architecture of electronics giving ment is anchored to the architecture of the shape to sudden random image data. screen and the Cartesian arrangement of (Donaldson) the pixel grid, the appearance of a pixel is relational function. In other words, while the Following Donaldson, let’s explore how pixel is geographically fixed to an (x,y) posi- macroblocks owe their particular appearance tion on the screen, its “function is based on to the design and arrangement of the pixel relational value sets assigned through the grid. Pixel, which comes from a contraction matrix of the corresponding elements around of the words picture (pic, pix) and “either it. This enables designation and manipulation element or cell […] is basically the smallest of point-by-point values in the image, which element of a discrete and non-continuous renders the image mutable” (Baraklianou dataset […] arranged via an address on a grid 307). Hence, macroblocking appears as the location (x,y)” (Baraklianou 305). A pixel is moveable, mutating arrangement of pixels, the matter of which digital images are formed but the material architectures of the pixel grid and the medium through which an image and logics of the compression algorithm act takes shape. A single pixel’s appearance is as boundaries on this visual chaos and keep determined mathematically: electrical signals the pixels fixed in place. The movement of are converted into a “two-dimensional array pixel blocks from one location on the (x.y) of information. A pixel is a register of data plane to another is simply a visual illusion, that, in combination with other pixels in an one which appears due to the way that mac- array […] is a sample encoded in a long set roblocking allows a spectator to observe the of binary codes” (Baraklianou 306). Despite geometries of the pixel grid as an emergent

212 Carleigh Morgan: CALCULATED ERROR property of the compressed image. metaphorically and matter-of-factly, the artists At its core, compression hacking can describe the project as an attempt to “make shed light on the operations of algorithms, hidden the data structure” of digital technolo- the structures of digital technologies, and the gies visible: the original images and moving priorities and patterns which occur as a func- images shared across these peer-to-peer tion of algorithmic manipulation. “Machines,” networks are run through “a transformation writes Liam Young, “see the world through machine,” whose oblique name functions like coded sets of rules. Whether through a a black box, with its refusal to confide in the camera lens, sensor, or scanner, they search specific technological process of translation for particular configurations of data, sets that causes a shared file “to dissolve into of predefined relationships, patterns, and pixels” (!Mediengruppe Bitnik). The language geometries” (125). The characteristics of used to describe Download Finished invokes macroblocking, and by extension other kinds images of technological systems that are im- of compression artefacts, can offer clues penetrable and unknowable except through about the way that computer architectures, the form of their spectacular malfunction. such as the pixel array, are structured. The In this way of thinking, the mystifying com- appearance of a compression artefact is mands and obfuscating structures that allow constrained by the logics of compression, computer technologies to work can only and by the geometries confining computer be visibly foregrounded through an error in hardware, such as the pixel array, to a par- these very commands, a breakdown in these ticular gridded orientation. So while it may very structures. seem like pixels are moving in regimental Glitch artist Rosa Menkman also be- blocks, this is an illusion that is shaped by the lieves that glitch art functions to reveal the material architectures of the pixel grid and obfuscated logics of computer processes. the way that the appearance of a pixel array She considers her practice a political one is a product that is relationally determined, because it interrupts the function of computer mathematically constrained, and materially systems by introducing malfunction into a fixed. “highly complex assemblage that is often hard to penetrate and sometimes even completely closed off” (Menkman 12). Menkman’s perspective on glitch art as a radical critique Detecting glitch of technological determinism is echoed by critics like Hainge. On Hainge’s formulation, In the early 2000s, art collective glitching materialises the ‘noise’ that lies !Mediengruppe Bitnik released Download dormant or unseen within the operations of Finished - The Art of Filesharing. Described digital systems. Some media scholars view by the artists as “an online resource which glitch art as the latest instantiation of the “aes- transforms and re-publishes films from thetic use of discarded and deleted data (i.e. P2P [peer-to-peer] networks and online ar- errors)” (Kane, “Compression Aesthetics”) or chives,” Download Finished is a part digital as a reaction against the impenetrability of performance, part post-structuralist critique computational systems. of the hidden technological protocols and New media scholars like Casey Boyle underlying architectures that give colour, also adopt this viewpoint. Boyle’s “Questions form and shape to digital signal and digital Concerning Glitch” explicitly expands on the noise (!Mediengruppe Bitnik). Speaking both work of Katherine Hayles and Bruno Latour

213 APRJA Volume 8, Issue 1, 2019 to argue that a responsible rhetorical practice question that has so far gone explicitly un- for glitch art would involve understanding “all asked: is macroblocking a glitch? In brief, no. mediation and any glitches as generative To label this compression artefact a ‘glitch’ is and not as errors to be corrected” (Boyle not a perception, but a judgement. That is, 12). In The Wretched of the Screen, Hito to always read compressed images — or Steyerl champions this feature of the glitch, visual indecipherability more generally — as too, calling them the bruises of images that a symptom of technological malfunction is to are “violated, ripped apart, subjected to inter- assign a creative intentionality to the com- rogation, and probing” (5). And as Donaldson pression algorithm, which is in fact indifferent writes: to the representational clarity of the images it produces. It also supposes that the visual The artist’s hand no longer dictates the layer of digital images mimics the behaviour outcome the way it does with con- of the algorithmic one. But the compres- ventional fine art. Instead, conditions sion algorithm has no stake in maintaining are created to bring forth something representational sensibility for its viewers. unpredictable, inasmuch as the set Compression hacking can give rise to “ran- parameters are capable of producing. dom image data” (Donaldson) but it can only use the data available to the compression Compression hacking aligns with these algorithm — and the data can only ever be considerations of the glitch. Compression preserved or lost, never rendered more de- hacking works to distorts the sleek, seam- tailed than its original source. Furthermore, less look of the digital image and to create any ‘chaos’ is bounded by the computational an aesthetic that “allows insight beyond limits of the compression algorithm and the the customary, omnipresent […] computer arrangement of the pixel grid. Finally, the aesthetics” and sheds light on “software’s appearance of macroblocking relies on the inner structure, whether it’s a mechanism of smooth operations of lossy compression; it data compression or HTML code” (Galloway cannot occur without the successfully com- 25). To be sure, compression hacking still pletion of the lossy encoding and decoding requires interpretation: it does not reveal process that is part of the overall process of the operations of the compression algorithm compression. The compression algorithm without some work behalf of the viewer and a must be functional in order to generate modicum of computational literacy. However, macroblocking effects; if macroblocking compression artefacts like macroblocking were to be considered an error, or as signal can draw attention to the computational of one, then its antecedent would not be the conventions by which digital images and lossy compression algorithm. After all, mac- rendered visible and by which “digital spaces roblocks are a product of lossy compression. are organized” (Galloway 25). The glitch If anything, the manifestation of macroblocks is a fissure that allows one to peer into the in an image would testify to the successful hidden operations and invisible structures of completion of lossy completion, not act as an digital technologies: “Whether its cause is in- indicator of its failure. tentional or accidental, a glitch flamboyantly In The Interface Effect, Alexander undoes the communications platforms that Galloway writes that glitch art “recuperates we, as subjects of digital culture, both rely on and even relies on failure to succeed. It is and take for granted.” (Manon and Temkin) primarily a systemic relation” (25). Likewise, These theories gesture to an important Michel Serres, in his meditation on functional

214 Carleigh Morgan: CALCULATED ERROR

‘alongsidedness’ writes, “Systems work images which mimic the appearance of tech- because they don’t work. Nonfunctionality re- nical corruption while not relying on techni- mains essential for functionality” (in Galloway cal corruption at all to produce these visual 25). This perspective, however, does apply effects. to compression artefacts in the narrow case Compression artefacts like mac- being examined here. Although compression roblocks, then, are not materialisations of artefacts may give the appearance of being an underlying technological failure — as glitched, they still rely on the smooth opera- the argument goes within glitch studies — tions of the compression algorithm for their but they do visually simulate the effects of materialisation. Serre’s axiom needs to be a technical failure that has not occurred. modified slightly in this case. As compression Compression artefacts indicate that there is hacking demonstrates, sometimes function- a subtle but significant difference between ality remains essential for the appearance of the visualisation of a technological error and non-functionality. Galloway’s observation can its aesthetic simulation. In a way, compres- be similarly adjusted for compression hack- sion artefacts are a pastiche of glitch style. ing, which can foreground how images that By thinking carefully about how compression appear to expose technological failure relies hacking affects the different strata of a digital on an underlying technological ‘success’ for image, one can see how the relationship their production. One can amend Galloway: between the algorithmic dimension and the compressed images show how glitch art imi- visual dimension of these images are inter- tates failure successfully rather than relying dependent but not behaviourally identical. An on failure to succeed. Ultimately, rather than error in the algorithmic layer does not always resigning compression artefacts to the do- manifest at the visual interface; conversely, main of glitch and its related nomenclatures, the appearance of a visual error is not a reli- glitch theorists should think seriously about able indicator of a technological malfunction. how compression artefacts might depend on In other words, the ‘glitch’ is in need of more precisely the opposite of technological failure careful theorisation: one should not confuse for their materialisation. an aesthetic of technological failure with an aetiology of technological malfunction or conflate the visualisation of a technological error with its aesthetic simulation. Finally, it Conclusion is imperative to keep in mind how much the randomness, unpredictability, or messiness Digital media are optical and algorithmic which glitch studies invokes around the glitch in composition: however, the behaviour of is in danger of overlooking the ways that these two dimensions does not always cor- the material architectures and algorithmic respond. The non-representational character protocols structure the digital glitch by or- and unpredictable behaviour of compression ganising, constraining, and given form to its artefacts trouble a human tendency to col- appearance. lapse the optical and algorithmic dimensions of digital images at the level of the visual: specifically, at the visual interface of digital media, where one can see what’s happening. Compression hacking produces compressed

215 APRJA Volume 8, Issue 1, 2019

Boyle, Casey. “The Rhetorical Question Acknowledgements Concerning Glitch.” Computers and Composition, vol. 35, 2015, pp. This work was undertaken thanks in part 12–29. https://doi.org/10.1016/j.comp- to the Trinity College’s External Research com.2015.01.003 (Accessed 14 March studentship; partially funded by a research 2019). grant from the Faculty of Medieval and Modern Languages; and completed due Berry, David M. Critical Theory and the to the enduring patience of the editors of Digital. Bloomsbury, 2015 this special collection. I have no conflicts of interest to declare. Donaldson, Jeff. “Glossing over Thoughts on Glitch. A Poetry of Error.” Art Pulse. http://artpulsemagazine.com/glossing- over-thoughts-on-glitch-a-poetry-of-error Notes (Accessed 17 April 2019). Galloway, Alexander. The Interface Effect. Polity Press, 2012. [1] See Brown and Kutty; Schultz-Figueroa; Manon and Temkin; Levin; and Kane “Error.” Goriunova, Olga, and Alexei Shulgin. The term ‘datamoshing’ was coined by “Glitch.” Software Studies: A Lexicon, edited internet art collective PaperRad. by Matthew Fuller. MIT Press, 2008.

Hainge, Greg. “Of Glitch and Men: The Works cited Place of the Human in the Successful Integration of Failure and Noise in the Digital Realm.” Communication Theory, Arcangel, Cory. “On Compression.” Cory vol. 17, no. 1, 2007, pp. 26–42. https:// Arcangel’s Official Portfolio Website and doi.org/10.1111/j.1468-2885.2007.00286.x Portal, 2007-2008. http://www.coryarcangel. (Accessed 12 April 2019). com/things-i-made/2007-007-on-c/ (Accessed 02 April 2019). Kane, Carolyn L. Chromatic Algorithms: Synthetic Color, Computer Art, and Baraklianou, Stella. “Pixel.” Philosophy Aesthetics after Code. The University of of Photography, vol. 3, no. 2, 2012, pp. Chicago Press, 2014. 305-309. Kane, Carolyn L. “Compression Aesthetics: Brown, William, and Meetali Kutty. Glitch from the Avant-Garde to Kanye West.” “Datamoshing and the emergence of InVisible Culture, 21, 2014. http://ivc.lib. digital complexity from digital chaos.” rochester.edu/compression- aesthetics- Convergence: The International Journal of glitch-from-the-avant-garde-to-kanye-west/ Research into New Media Technologies, (Accessed 05 July 2015). vol. 18, no. 2, 2012. https://doi. org/10.1177/1354856511433683 (Accessed 02 April 2019)

216 Carleigh Morgan: CALCULATED ERROR

Kane, Carolyn L. “Error as Critical Praxis.” Steyerl, Hito. The Wretched of the Screen. Review of Noise Channels: Glitch and Error E-flux Journal: Sternberg Press, 2013. in Digital Culture by Peter Krapp. Millennium Film Journal, vol. 56, 2012, pp. 7-8. http:// Vanhanen, Janne. “Virtual Sound: www.mfj-online.org/mfj-no-56-fall-2012/ Examining Glitch and Production.” (Accessed 11 March 2019). Contemporary Music Review, vol. 22, no. 4, 2013, pp. 45–52. https://doi. Levin, Thomas Y. “Datamoshing as org/10.1080/0749446032000156946 Syntactic Form: Reading Digital Video (Accessed 05 July 2015). Compression Algorithm Hacking.” Kracauer Lectures in Film and Media Theory. 13 Virilio, Paul. Politics of the Very Worst. December 2011. Goethe-Universität. Semiotext(e), 1999.

Merleau-Ponty, Maurice. Phenomenology of Young, Liam. “An Atlas of Fiducial Perception [1945]. Routledge, 2005. Landscapes: Touring the Architectures of Machine Vision.” Log, No. 36, ROBOLOG, Manon, Hugh, and Daniel Temkin. “Notes Winter 2016, pp. 125-134. https://www.jstor. on Glitch.” World Picture, vol. 6, 2011. org/stable/26323702 (Accessed 08 April http://worldpicturejournal.com/WP_6/PDFs/ 2019). Manon.pdf (Accessed 13 March 2019).

!Mediengruppe Bitnik. Download Finished: The Art of Filesharing, 2006. https://vimeo. com/68077652 (Accessed 28 April 2019). Menkman, Rosa. Glitch Studies Manifesto, 2011. http://rosa-menkman.blogspot.com (Accessed June 13 2015).

Parikka, Jussi, and Tony D. Sampson. The Spam Book: on Viruses, Porn, and Other Anomalies from the Dark Side of Digital Culture. Hampton Press, 2011.

Rubenstein, Daniel, and Katrina Sluis. “The Digital Image in Photographic Culture: Algorithmic Photography and the Crisis of Representation.” The Photographic Image in Digital Culture, edited by Martin Lister. London: Routledge Press, 2013.

Rutsky, R. L. High Techne: Art and Technology from the Machine Aesthetic to the Posthuman. University of Minnesota Press, 1999.

217 APRJA Volume 8, Issue 1, 2019

CONTRIBUTORS

Mitra Azar is an artist and PhD fellow at School of Communication and Culture, Aarhus University.

Daniel Chavez is a PhD fellow at the Department of Digital Humanities, King’s College London.

Michela De Carlo is an artist, teacher and PhD student in Digital Humanities at the Università degli studi di Genova.

Iain Emsley is a PhD fellow in Digital Media at Sussex University.

Malthe Erslev is a PhD candidate at the Department of Digital Design and Information Studies, School of Communication and Culture, Aarhus University.

Tomas Hollanek is PhD candidate at the Centre for Film & Screen, University of Cambridge.

Rosemary Lee is an artist and a PhD fellow at the IT-University of Copenhagen.

Carleigh Morgan is PhD candidate at the University of Cambridge, and a Trinity College External Research Scholar.

Carman Ng is a Postdoctoral Fellow in game studies, multimodal semiotics, and affective science at University of Bremen.

Irina Raskin is a curator and PhD student at the DFG Research Training Group “Knowledge in the Arts” at the University of Arts Berlin.

Tiara Roxanne is a Research Associate at De-ZIM Institut, Berlin, Germany.

Rebecca Uliasz is an artist and PhD candidate at Duke University.

Maria Dada is a Research Fellow in Visual Cultures at Goldsmiths University and at The Digital Anthropology Lab at London College of Fashion, University of the Arts London.

218 Tanja Wiehn is a PhD student at the Department of Arts and Cultural Studies at the University of Copenhagen.

Brett Zehner is a writer, artist and PhD researcher at The Department of Theatre Arts and Performance Studies, Brown University.