ABSTRACT. In Reading Machines, Stephen Ramsay argues that some in the digital humanities have seen the marrying of computation and interpretation as a one-way relationship, an application of computational methods in order to introduce “scientific rigor” into literary study. However, Ramsay argues that the relationship between reading and machines should be mutually transformative, allowing the insights of close reading to inform the practices of distant reading. Interpretive methods already deform, sort, and manipulate texts by way of human reading procedures, and these methods should be paired with rather than opposed to machine processes. This panel draws inspiration from Ramsay to address how we read with machines, not in the sense of using machines as tools for interpretation but instead in the spirit of reading alongside machines, collaborating with and learning from them.

ABSTRACT. This stream of panels exhibits research from a multidisciplinary team of scholars and practitioners addressing themes of plant intelligence. We have been engaged in an ongoing set of conversations and collective readings considering questions of how plant intelligence is modelled and mediated and whether the capacities of plants challenge received understandings about plants and/or media. In recent years, experimental research on plant intelligence has utterly transformed not only how we understand plants but also how we understand memory, perception, sentience, movement, cognition, and intelligence. Because plant intelligence occurs without the physiological correlates associated with intelligence in animals (such as sense organs, neural networks, brains), it has spurred a thorough reconsideration of some of the ruling assumptions of the cognitive sciences. These trends open up paths not only for thinking the intelligence of plants and their ecological networks themselves, but also for how we attempt to generalize these notions in approaching other forms of non-human cognition, living or otherwise.

The three panels bring the interests and approaches of our team to bear on questions of non-human minds and and forms of intelligence as posed by this year’s SLSA conference. In “Plant Intelligence/Human Intelligence,” Vicki Kirby, Isabel Kranz, and Thomas Lamarre address concerns of anthropomorphism head-on, exploring the affordances of anthropological, literary, and philosophical approaches in conjunction with scientific research on plant intelligence, and asking what these approaches mean for the interplay of human and plant modes of thought. In “Information, Inscription, and Transcription with Plants,” Vera Bühlmann, Christina Jauernik and Fabian Puttinger, and Zach Yost study issues of perception and signification which are raised by acts of sensation and acts of recording performed both by and about plants, and they consider how material and semiotic processes shed light on embodied notions of cognition. Finally, in “Grounding Plant Intelligence: Ecology, Land, and Vegetal Life” Michael Fisch, Yangqiao Lu, Jun Mizukawa, and Adam Nocek ask how the agencies of multi-species ecological networks and entangled technologies work to collectively shape their environments and communities, thinking through cases of landscape re-engineering and guerrilla planting in post-3.11 Japan, the cross-territorial practice of Chinese artist Mao Chenyu, and vegetal architecture and design.

ABSTRACT. The ubiquity of gendered, particularly feminized, AIs (Alexa, Siri) raises serious ethical concerns, alongside future policy considerations, about the role that these AI assistants may have in reinforcing sexist and dehumanizing divisions of labor, particularly as many typical tasks assigned to these assistants (scheduling, reminding, caring) are thought of as stereotypically “feminine.” To date, no significant empirical interventions have been carried out on this topic. Our series of studies has begun to shed some light on this important topic by building off of related work in the domain of human-robot interaction (Kuchenbrandt et al. 2014). 420 native-English speaking adults (49.8% self-identified as women) were recruited via Prolific and randomly assigned to one of four conditions in our 2x2 between-subjects design. Each condition asked participants to imagine that they had been paired with either a feminized- or masculinized-AI (Nera or Nero) to help them with a feminized- or masculinized-activity (caring for a sick relative or planning a vacation) with four component tasks (e.g., managing finances, scheduling appointments). Activities and tasks were normalized on a number of dimensions including gender-typicality and complexity in a prior study. Participants were asked to assign between one and three tasks to the AI. They were also asked to rate the AI on a number of dimensions, including competency and personality metrics. We hypothesized that there would be a series of at least nine interactions between AI-, participant-, and task-gender. Initial results support this complex picture. We found, for instance, that participants are more likely to assign tasks to feminized- versus masculinized-AIs (t(416) = 2.82, adjusted p < .05), and this effect is pronounced for male-participants assigned to feminized-activities (t(415) = 2.82, adjusted p < .05). Furthermore, participants less familiar with AIs are more likely to assign feminized tasks to a feminized- AI (t(411) = 3.24, adjusted p < .05). Additional findings and further steps are also discussed.

ABSTRACT. Artificial intelligence technologies have become overwhelmingly gendered female: first in popular culture and science fiction narratives, then in real technological artifacts, and then looped back in to commercials and newer science fiction narratives. This has been observed and commented on readily, but how do we make sense of this gendering? How does this gendering come to matter in the sociotechnical reproduction of real AI artifacts? This paper takes up the ELIZA program, often touted as the first chatbat, to make sense of the genealogy of gender and natural language AI models. ELIZA was created in the Project MAC lab at MIT under the supervision and coordination of Dr. Joseph Weizenbaum. This program forms the early foundation of natural language processing technology, and was also readily gendered female. This paper draws on the framework of play as a mediating force between reality and fantasy to make sense of the gendering of AI technologies as a masculinist enactment of an unconscious fantasy of the perfect woman.

ABSTRACT. The Game Studies Stream at SLSA 2022 brings together a range of scholars, creators, and developers to present their work on digital and analog games. This year's stream features four sessions:

GAME STUDIES 1: Making the Jump: New Approaches to Platforming Games GAME STUDIES 2: Ludoidentities GAME STUDIES 3: Ludonarrativities GAME STUDIES 4: Works-In-Progress Roundtable

ABSTRACT. For SLSA 2019, we organized an art exhibition called "At the Margins: Experimental Engagements in Science, Literature, and the Arts". Participants responded to our initial prompt — What is the location of your experimental engagement? — by providing an image, a caption, and a quotation derived from a source of their creative inspiration. We formatted these materials into a set of 42 posters which represent a cross-section of the “A in SLSA.” We are now expanding this creative experiment into an anthology project that will offer deeper reflections on issues of shared concern to the exhibition artists. It will include material derived from in-depth follow-up conversations with a selection of the artists, accompanied by additional documentation of their creative work. Interspersed with these edited conversations will be images and texts from the other artists, reformatted for the book. The book will be organized into themes that highlight the creative practice overlaps we observe in the group, tentatively titled Obscure Terrains, Extra-human Agencies, Collective Engagements, and Psychic Machines. For SLSA 2022, we propose a roundtable event where we will invite the selected artist-authors in attendance — many of whom are SLSA regulars — and the larger SLSA arts community to consider our themes and our lines of interrogation of these themes. (If accepted, we will also encourage attendance by the artist-authors.) The roundtable conversation will inform the editorial direction of the project, and be integrated into the book as further creative content. The roundtable will be facilitated by at least one of the anthology editors (Jesse Colin Jackson and Antoinette LaFarge) and one of the Brill Experimental Practices series editors (Stephan Besser and Christoph Brunner).

ABSTRACT. In Carmen Maria Machado’s story “Inventory” (2017), the unnamed narrator keeps lists—of her teachers, jobs, “trees that begin with m,” spices, and more. These lists order her life and give shape to her mourning for loss during an unnamed global pandemic. In fact, the entire story is structured as a chronological list of narrator’s sexual encounters, each section opening with an anonymized reference to past lovers (“One man.”, “One woman.”, and so on). It becomes clear at the conclusion of the story that her lists are not only a technology for remembrance in times of crisis, but also a literary form that produces meaning by implying a narrative of connection between the items given.

My paper reads Machado’s short story alongside popular uses of lists to catalogue loss during both the COVID-19 pandemic and climate crisis. While scholars have argued that the list as form is a provocation to the reader (Ian Bogost, Alien Phenomenology), a catalogue of loss (Ursula Heise, Imagining Extinction), or a manual that contains and organizes possible actions (Atul Gawande, The Checklist Manifesto), I argue that lists are a literary form that stand in for ecological interconnectedness during crises. Lists insinuate myriad possibilities of connections between the items—ecological, ontological, analogical, or otherwise. In times of crisis, these possible connections are disrupted. The list, I show, not only gestures to lost connections, but becomes a substitute for those connections, a type of literary ecology that I call a “list/ecology.”

ABSTRACT. Based upon ethnographic research in South Africa, and drawing from feminist technology studies and queer ecologies, this paper considers how precision agriculture is potentially surveilling and optimizing both plants and Black Bantu-speaking women farmers’ labor in ways that reinforce hierarchies of knowing and being. Precision agriculture involves sensors for tracking soil and livestock, and satellites, planes, and drones for overhead sensing of crops. While such technologies provide some opportunities for smallholder women farmers, they may also emerge as a new—yet familiar—system of “data colonialism” that engenders appropriation and control over their labor and knowledge, evoking legacies of colonial extraction and exploitation (Couldry and Mejias 2019). In heralding AI-based sensors as telling more accurate and true stories about plants and crops, precision agriculture implies that humans and plants tell less precise stories. This paper provides understandings into South African men’s and women’s different practices and decision-making related to agricultural production and the adoption of relevant technologies. At the same time, it challenges conventional understandings of nature as raw material by attending to plants as storytellers and models of collectivity and communicative cooperation that contests hierarchies of difference (Gibson and Gagliano 2017; Mortimer-Sandilands and Erickson 2010). Plants are teachers, holders of knowledge, and guides to understanding more-then-human worlds with unique modes of communication, articulation, and a language all their own (Kimmerer 2013). It argues that attention to gendered and vegetal ways of knowing offers ways for challenging algorithmic culture and its hierarchical ordering of algorithms as ultimate producers of knowledge.

Couldry, Nick, and Ulises A. Mejias. 2019. The Costs of Connection: How Data is Colonizing Human Life and Appropriating it for Capitalism. Stanford, CA: Stanford University Press.

Gibson, Prudence, and Monica Gagliano. 2017. "The Feminist Plant: Changing Relations with the Water Lily." Ethics & the Environment 22, no. 2 (2): 125-146.

Kimmerer, Robin Wall. 2013. Braiding Sweetgrass: Indigenous Wisdom, Scientific Knowledge and the Teachings of Plants. Minneapolis, MN: Milkweed Editions.

Mortimer-Sandilands, Catriona, and Bruce Erickson. 2010. Queer Ecologies: Sex, Nature, Politics, Desire. Bloomington, Ind.: Indiana University Press.

ABSTRACT. Computer-assisted writing and visual art have been developing since the earliest days of computing, and members of various aesthetic movements and collectives have found algorithmic and aleatory means of producing avant-garde literary text (whether poetry or prose) and visual art. Within the last several years, major advancements in machine learning and cloud computing have made the tools of artificial intelligence more accessible to lay users, and creative applications of large language models like OpenAI’s GPT-3 have led to computational poetry projects such as David Jhave Johnston’s ReRites poems and prose experiments like Ross Goodwin’s 1 the Road (2018). The surreal and hyperreal images created by generative adversarial networks (GANs) also have their uses both in artistic contexts and as well as in lending credibility to dubious social media accounts. The latest examples of images generated from text prompts (usually with VCGAN+CLIP) are impressively evocative. But despite these advances in computational literary text and art, the medium of comics remains curiously resistant to artificially-intelligent authoring. In this paper, I argue that the unique spatio-temporal semantic structures between comic panels relies on an irreducible complexity that is neither text nor image but imagetext. Examples of computational comics so far avoid the challenge of what Thierry Groensteen calls the “arthrology” of the comics page in one of three ways.  By providing frames and templates that a program makes use of (John Pound’s Ran-Dum Comics), by making the panels the content (KUMO Collections’ “Elongation” anthology or my book of empty panel layouts, An Arthrogram) or be eschewing panels altogether (“The AI-Made Comic Book #TAIMCB 1: the State of Art of the Future of Comics” by Adam Niman or “Untitled Harsh Noise Graphic Novel” by Cementimental). By discussing the limitations and provocations of these examples and more, I intend to show that the general absence of truly computational comics offers more evidence for the unique formal alchemy that underlies the medium of comics.

ABSTRACT. In 2010, Amazon Mechanical Turk’s crowdsourced translation of Melville’s famous novel into Japanese emoticons was published to mixed acclaim. In her article on this book, Emoji Dick, Lisa Gitelman argues that it illuminates “a ludic contact zone between human intelligence and algorithmic processing, between text and image under the sign of computation” and ultimately “between literature and whatever the fate of the literary may be in an ever more digitally mediated and data-described world” (2). Highlighting the difference between visual and linguistic forms of media, she suggests, is futile since large scale data comprehension seems to lend itself more easily to visual media than writing per se. Gitelman furthers her assessment of Emoji Dick as a work that allows us to “consider the legacy of East/West contact as much in terms of media history as in terms of cultural exchange, [and] international trade” (16) by putting emphasis on “media” over traditional “cultural” and economic terms. Her reference to East/West contact as a “media” zone raises interesting questions about Orientalist thinking in Melville’s era: How might Emoji Dick be successful as a global code in a more equitable way than the old East/West binaries imply? My paper argues we must account for the West’s reappropriation of the emoji (since it originated in Japan), and asks specifically whether this approach to translating English literature is entirely free from deploying the Oriental as a means to complement Global English. I address the significant discrepancy that while emoji is understood to be global Unicode, Queequeg’s signature, termed a “queer round figure” in Melville’s novel, as well as his tattoos, render the character impervious/immune to universal significations.

ABSTRACT. Countless readers – whether autistic or allistic, academic or lay – have at some point diagnosed a fictional character as autistic. The fraught ethics of these recognitions are not the subject of this paper (and neither are the problematic debates over autistics' capacity for theory of mind, let alone basic empathy). Instead, I ask: why is this diagnostic game so easy to play? My method is to invert Ato Quayson's notion of "aesthetic nervousness," which occurs when "the dominant protocols of representation within the literary text are short-circuited in relation to disability." Quayson's analysis is important, but as an autistic reader, I've found that his admitted focus on "a nondisabled reader" can only get us so far. As such, I am particularly interested in "aesthetic familiarity," which I define as the phenomenon of autistic readers finding kinship with fictional characters who are not explicitly autistic. I focus on several stock characters of science fiction (e.g. the alien, the mad scientist, the android, the mutant, the AI, and the mind-reader) to argue that Darko Suvin's genre of "cognitive enstrangement" might be better understood as an attempt to reckon with the enstranging cognitions of neurodivergent individuals. Furthermore, the ubiquity of potentially autistic tropes throughout literature (e.g. the holy fool, the hermit, and the mute) implies that autistics have not only always existed, but have also always had an outsized impact on literary culture.

ABSTRACT. A curated screening of contemporary video art by eight artists who are using generative imaging and/or exploring concepts of computer-assisted creativity as a form of 21st century expression.

....................................................................................

For this year's theme “Reading Minds: Artificial Intelligence, Neural Networks, and the Reading Human.”, we wish to present time-based artwork from artists who research the meaning of data, but who also morph and sculpt its materiality to create new forms of imagery that incorporate video, sound, and literature. While some of the selected works use purely individual meditations on data, others use collective image-making modes and/or generative gan-based techniques to form their works.

We, Dakota Gearhart and Derek G. Larsen, whom are video artists in our own right, wish to share our ongoing research into this evolving genre of computer-assisted creativity as it relates to video with the SLSA community. We are particularly interested in screening selected works that re-image data's numerical structure as visual facades that engage immeasurable sensations, like beauty, intuition, and identity.

The national and international artists we are contacting for the screening and their websites are listed below. We wish to show one video from each artist, which would make our proposed screening about 35 minutes long. A Q & A will be held after, with one or two artists present to take questions.

Ian Cheng http://iancheng.com

Refik Anadol https://refikanadol.com

Sophia Crespo https://sofiacrespo.com/

Yaloo https://yaloopop.com

Claudia Hart https://claudiahart.com

Casey Reas https://reas.com

Sarah Rothberg https://sarahrothberg.com

Marina Zurkow http://o-matic.com

ABSTRACT. The cultural dimensions of juxtapositions of science and literature are arguably nowhere more urgent than in consideration of the Anthropocene. Reception theories of ecological imaginings show that ecocritical literature leads to heightened ecopolitical climate consciousness and awareness (Schneider-Mayerson, Malecki et al). In the humanities classroom, how do we manage these emotional influences and affective responses in ways that contribute both efficaciously and ethically at intersections of geological and human timescales, which may seem overwhelming? Similar challenges arise in teaching students about surveillance, algorithmic bias, and other topics at the intersections of technology, science, and the humanities. How might humanities scholars understand the meaning, nature, and strategic value of hope in an increasingly dystopian world and disrupt the prepackaged narratives of capitalist constructions and military-energy regimes? This panel/roundtable considers a range of theoretical and pedagogical approaches to the question of how our fields of study might develop concrete strategies to help people (including our students) understand the enormity and complexity of these problems while simultaneously equipping them with ways to respond more productively than with despair. We believe that this project showcases how the humanities, and particularly literary studies and feminist approaches to science and technology, provide vital intellectual tools for facing the challenges of the world today.

ABSTRACT. In 1987, researchers at General Electric pioneered a method for generating computer graphics from medical scan data that featured an underlying language of faceted cubes. I wanted to make this seminal computational procedure — now known as Marching Cubes — into something people could build with. I translated the algorithm into 3D printed construction units that permit people to act out its logic. I also created a user’s guide: input any 3D scan or model, and a custom computer script outputs assembly instructions. Each assembly is unique, and created collaboratively: together, we perform the computer’s process. We also perform the fabrication process: in aggregating the units, layer by layer, we mimic the 3D printer that made them. Sometimes, we simply play: with humans doing the work, strict procedural logic is optional. Participants in assembly events embody the algorithm’s procedure and demonstrate the visual language the algorithm imposes on its world. Extended interaction begins to reveal the language’s syntax patterns and representational limitations. Taken together, these evolutionary behaviors constitute an inverted form of machine learning, where human collectives interpret and improve a computational method, expanding the possible results towards more creative and meaningful ends.

Note that there is an interactive Art proposal associated with this paper, entitled “Marching Cubes: an Inverted Machine Learning Workshop.” The two proposals can be considered mutually exclusive — neither is reliant on the other.

ABSTRACT. With the advent of artificial intelligence, deep learning, and neural networks there is no denying the profound impact of these technologies in changing the way that data within research is transcribed, quantified, and understood. Yet, the language and routes of understanding and processing AI remain inaccessible to the general public -filled with language and explanations that make little to no sense without years of coding experience. This leads to an air of confusion that lurks around the idea of artificial intelligence – the uncertainty chips away at the impressiveness. Furthermore, the challenge to understand the language leads to a sense of stagnation and a lack of questioning of the ways in which these technologies are used. How can we determine the effectiveness and reproducibility when the route to understanding requires learning a whole new language? How can we trust these things that make no sense? It is no surprise that this feeling of unknowingness permeates the air around us causing us to question the limits of our humanity when machines become involved. At what point do we allow ourselves to be overcome by the fears of technology. By examining the language and the literature surrounding machine learning, I seek to demystify and find the source of this complication. Is there a route to making these technologies available to the understanding of all? These questions become essential to think about as we turn towards a future in which we rely on artificial intelligence.

ABSTRACT. Machine Learning bias against Black and Indigenous people has been recognized in all domains of public and private service, including criminal justice, healthcare, and finance. These biases have direct material and fatal consequences. Many ML researchers have endeavored to develop methodological ‘fixes’ to the data collection, classification, and connection processes that influence these biases. Some scholars insist the concern begins not with the mundanity of data collection and classification but in the very nature of the questions ML systems are being used to answer. In this paper, I turn to the physical infrastructure of ML systems to explore the entanglement between the epistemic and environmental toxicity of data archival and cloud service delivery. Drawing on Ingrid Burrington’s investigations of the Internet’s infrastructure, Rob Nixon’s concept of ‘slow violence’ and environmental racism, and David Zeitlyn’s explorations of archives as projects of colonialism, I interrogate the compulsion for data collection itself as an expression of neocolonial entitlement - a practice which produces toxicity in both the ecological and epistemic domains. Finally, I engage with Indigenous Scholar Roxanne Dunbar Ortiz’s work on “unforgetting” as a guide for “uncollecting” – a practice that intervenes on colonial hoarding through divestment. I invite readers to struggle with me over the complex differences between ‘forgetting’ and ‘uncollecting’ in the context of archival practices. Drawing from Sunaura Taylor’s ‘Disabled Ecologies’, I imagine ‘uncollecting’ as a form of agential forgetfulness that serves as a criptechnoscientific intervention on the archive which moves memory from hegemonic repositories into interdependent collective social networks.

ABSTRACT. In a topological conception, space is conceived of as a network of paths between machines or nodes produced by machines. Human-rock interaction framework uses topological space to create materially-charged and affirmative ways for humans to relate to stone, rock and mountain bodies. In the framework, the body is a thought object to think about the world, and through the acknowledgement of the bodily existence of stone, agency and materiality of inanimate beings gain a new meaning. This is ultimately a non-Cartesian way of thinking about the mind, and therefore subverts Western mind-over-matter thinking by extending a hand into Indigenous ways of becoming. Mobility and movement is reconsidered as markers of animacy and therefore being alive, through the lens of plate tectonics and geology by showing how rocks are more loquacious than we think they are. The framework interprets stone as a person and a body as a pragmatic strategy to look at human-nonhuman interaction from a materialist and affirmative perspective. Only through stone bodies are we able to read our pasts, and this pasts include a vast majority of our ornaments, ritual and domestic objects, musical instruments, and material culture in general, as an endless source of information on the constitution of personhood.

ABSTRACT. Proposal for a workshop

In this workshop, participants will work together to produce an interactive, virtual 3D environment composed of photographic fragments and audio recordings of participants' immediate, physical surroundings. Over the course of the workshop, participants will use their phones to capture audio and photographic source material and work together to compose this material into a 3D scene using the Unity game engine and the Community Game Development Toolkit (see below). Participants will then collectively explore the resulting environment, charting a network of chance encounters with disembodied audio-visual fragments of participants’ surroundings as they move through the virtual space. Each play-through of the scene will yield a unique path through this collaboratively produced collage, representing a unique, virtual composition of participants’ collective experience of space, material and sound.

The Community Game Development Toolkit

The Community Game Development Toolkit is a set of tools that make it easy and fun for students, artists, researchers and community members to create their own visually rich, interactive 3D environments and story-based games without the use of coding or other specialized game-design skills. Building on the popular 3D game design engine Unity, the toolkit provides intuitive tools for diverse communities to represent their own traditions, rituals and heritages through interactive, visual storytelling.

In order to quickly create vibrant, visually rich scenes without the use of 3D modeling, the toolkit draws on creators’ own photos, collages, drawings and sound recordings to create objects and textures in 3D space. This technique allows creators to bring their own visual references and sensibility into the game environment and makes creative experimentation rewarding and fun for creators who may have no prior experience in 3D modeling or even visual art. The toolkit also provides a set of game components that make it easy to add many types of interactivity to games without the use of code including mechanisms for moving between scenes to create interactive visual narrative, simple non-player characters and interactive text.

Development and previous presentations using the Community Game Development Toolkit

In the summer of 2022, The Community Game Development Toolkit will be developed as part of an NSF REU at the Visualization and Virtual Reality Lab at Hunter College, CUNY. Previously, the toolkit was developed with the support of the Motion Computing / SloMoCo microresidency in 2021. The toolkit has been presented in collaborative workshops and presentations at the Society for Literature Science and the Arts conference, the iDMAa Conference, Winona State University; the Show Don’t Tell Symposium, as part of Culture Push, New York; and the New Media Caucus Showcase at the College Art Association Conference. The toolkit is currently taught at Baruch College, CUNY; City Tech, CUNY and Parsons School of Design, The New School.

Further information about the toolkit can be found at https://www.daniellichtman.com/wold-building-workshop//

Technical specifications

Participants can be in person, or connect via Zoom. Participants will use their own phones to capture images and audio, and use their own laptops. No further equipment is required.