Dancing with the Nonhuman

Dancing with the Nonhuman: A Feminist, Embodied, Material Inquiry into the Making of Human-Robot Relationships Petra Gemeinboeck Rob Saunders Swinburne University of Technology Centre for Transformative Media Technologies Melbourne, Australia University of Applied Arts Vienna Media Theory Vienna, Austria [email protected] Leiden University Leiden Institute of Advanced Computer Science (LIACS) Leiden, The Netherlands [email protected] Figure 1: Machine Movement Lab (MML) project snapshots from 2015 to 2022; with dancers Tess de Quincey, Linda Luke, Audrey Rochette, Felix Palmerson, Arabelle Frahn-Starkie. Photo: P. Gemeinboeck ABSTRACT Conference on Human-Robot Interaction (HRI ’23 Companion), March 13ś 16, 2023, Stockholm, Sweden. ACM, New York, NY, USA, 9 pages. https: //doi.org/10.1145/3568294.3580036 We propose that feminist reconceptualisations of agency and difference could dramatically expand our possibilities for both relating to robots in social scenarios and designing them as social agents. A performative approach to human-robot interaction favors the artefact’s relational, participatory capacities over representational attributes to explore the meaning-making potential of human-machine couplings rather than the predefined meaning of an individual robotic agent. We discuss the feminist concepts of intra-action and diffraction and explore how they could expand our understanding of the workings of the interference patterns that characterize human-robot relationships. Our collaborative Machine Movement Lab project serves as a case study to look at the situated enactment of the subjects and objects that shape our human-robot relationships through the embodied lens of performance-making. 1 INTRODUCTION Creative performance-making practices have a history of developing new kinds of agential relations, where the making of the performance relies on a kindling of imaginative alliances between human and nonhuman co-performers [16, 21, 40]. While collaborations with performance practices can still mostly be found on the fringes of robotics research, they are providing a productive arena for experimenting with new pathways to engage audiences with new socio-technical encounters [33, 34] or evaluating a robot’s communicative capacities [35]. But performance-making practices can be more than a place for staging relationships; as a place of research and experimentation they can offer valuable material and embodied perspectives into the making of subjects and objects. The latter is also a core concern in Science and Technology Studies (STS) and feminist practices of new materialism [7]. Feminist reconceptualisations of subjects and objects as being mutually constituted [3, 44] and agency as a material enactment [2, 3] offer alternative and more differentiated starting points for understanding how robots become social agents [19]. This paper introduces our collaborative, transdisciplinary approach bringing together creative robotics, performance-making and new materialist concepts to explore alternative pathways to both robot design and how we imagine our relationships with social machines, beyond modelling them after human-human relationships. When it comes to our relationships with robots, we tend to imagine them in our image to narrow the gap between us and them, often concealing the nonhuman behind a humanlike veneer. Feminist new materialism understands matter as active, dynamic, and inventive, moving between łthe animated and automated, bodies and environmentsž [7]. Looking at human-robot interaction CCS CONCEPTS · Human-centered computing → Interface design prototyping; · Computer systems organization → Robotics. KEYWORDS agency, diffraction, embodiment, intra-action, movement, performativity, performance-making ACM Reference Format: Petra Gemeinboeck and Rob Saunders. 2023. Dancing with the Nonhuman: A Feminist, Embodied, Material Inquiry into the Making of HumanRobot Relationships. In Companion of the 2023 ACM/IEEE International This work is licensed under a Creative Commons Attribution International 4.0 License. HRI ’23 Companion, March 13ś16, 2023, Stockholm, Sweden © 2023 Copyright held by the owner/author(s). ACM ISBN 978-1-4503-9970-8/23/03. https://doi.org/10.1145/3568294.3580036 51 HRI ’23 Companion, March 13–16, 2023, Stockholm, Sweden Petra Gemeinboeck & Rob Saunders through a new materialist lens highlights the participatory capacity of nonhuman matterings and how subjects and objects co-constitute each other [44]. Our arts-led approach investigates human-robot interaction (HRI) as a practice concerned not only with bringing forth human-machine relations but also the making of subjects and objects and, with it, the configuring of boundaries between human and nonhuman entities. According to feminist views, the boundaries between humans and machines are not ‘naturally’ given but, instead, are enacted through both material and discursive practices and materially, socially and ethically matter [3, 44]. In this paper, we seek to propose and demonstrate that a feminist rethinking of human-robot relationships can serve to materially engender more-than-human notions of sociality arising from difference-in-relation [19]. To do this, we will take a closer look at two key concepts from feminist new materialism and how they can expand our understanding of human-robot relationships: intra-actionÐunderstanding agency as an enactment, and diffractionÐforegrounding the relationality of differences. Agency, as reconceptualized by Karen Barad [2003], is not a quality or property that an individual agent can hold but, instead, is a material enactment as part of the world’s ongoing reconfiguring. Both, Donna Haraway [1992, 1997] and Barad [2007] employ the phenomenon of diffraction as a figurative lens to foreground and attend to łthe relational nature of differencež [3, p. 72]. In our practice, we work with patterns of diffraction and interference to explore the making and unmaking of subjects and objects. As we acknowledge the mutual constitution of subjects and objects and how agency is enacted in the dynamics of the encounter, human-robot interaction is no longer about an exchange of separate entities, designed and controlled by and for humans. We thus understand human-machine relationships as a complex ecology, where the human and nonhuman are always already entangled: a more-than-human encounter. ‘More-than-human’ here rejects the dualist thinking that preserves the hierarchical separation of subjects and objects, human and nonhuman entities, and attests to social agency being enacted in-between. Our performance-based approach proposes an alternative path for exploring human-robot relationships by investigating how we can bodily resonate with something strange, more-than-human and to make meaning with entities that dramatically differ from us. MML’s proposition is that becoming corporeally entangled with the machine artefact and its different material-spatial qualities and affective potential opens up new modes of empathy and meaning-making. Embodied forms of empathy, we believe, are key to meaning-making with social machines without relying on fake emotional facades (e.g., a humanlike face). The following sections introduce two of the mapping methodologies we have developed, Performative and Relational Body Mapping (PBM and RBM). 2 BACKGROUND In HRI practices, the otherness of machines is often seen as an obstacle for pleasant interaction and successful communication and there is widespread agreement that it requires softening or masking [6, 10, 24, 41]. HRI design approaches thus frequently invest into bridging the ontological gap between humans and robots [25] by fabricating homologies between what are essentially separate entities, marked by profound differences and łdeep asymmetriesž [44, p. 11]. According to Lucy Suchman, however, humanlike robot designs may appear to break down rigid subject-object boundaries but instead commonly serve to echo and reaffirm the differences and hierarchies that this boundary is founded on [45]. From a creative perspective, this mimicking approach poses restrictions not only on what a robot could be but also what relationships we could have with them. Feminist influences in HRI practices or creative robotics projects that seek to materialize feminist concepts beyond the theoretical work in STS are still rare. Pat Treusch’s technofeminist intervention into robotic collaboration brings the sociomaterial dynamics of knitting to the robotics laboratory to contribute an important embodied mode of caring to the practice of coboting [46]. Treusch employs diffraction as a methodological tool for developing and analysing her robotic knitting intervention and the material entanglements and knottings it brings to the laboratory as a practice of łsituated co-engineeringž [46, p. 63]. The interactive performance Babyface by Kate Ladenheim and Amy LaViers brings together both feminist ideas and performancebased strategies to materially perform and make tangible the feminized tropes prevalent in social technologies [38]. Drawing on Haraway’s Cyborg Manifesto [26], the machine-augmented dance performance renders łthe spectacle of woman and machinež [38] a physical experience, using techniques from choreography, robot design, and character development. Notions of Haraway’s becoming-cyborg and performance-as-research also characterize Elizabeth Jochum et al’s work with exoskeletons and the łintimate, physical alliances with human bodiesž [33, p. 2] they produce. The interdisciplinary team develops a humancentered, situated approach, grounded in embodied experience, to counter rigid gender assumptions engrained in robotics practices. Marco Donnarumma develops performance works with artificially intelligent body technologies that seek to move beyond the cyborg paradigm, arguing that such human-machine embodiments 1.1 Machine Movement Lab Our discussion will revolve around a case studyÐour own arts-led, collaborative Machine Movement Lab (MML) project, which seeks to demonstrate the potential of feminist thinking for materially reimagining how we can relate to robots. MML develops embodied, performance-based inquiries into the situated enactment of humanrobot relationships (Figure 1). It brings together the practices of creative robotics, choreography, dramaturgy, and machine learning, grounded in concepts from enactive cognition. The creative research project harnesses dancers’ creative movement expertise and kinesthetic awareness to scaffold the relational capacities of non-humanlike, abstract robots based on movement qualities and the spatial dynamics they effect. Dancers work with costumes, standing in for a robot-in-the-making, in embodied and entangled ways to probe into the shape’s different material affordances to find new, more-than-human movements with them. They usually inhabit or wrap themselves around the robot shape to kinesthetically feel into this other, machinic body and to bodily learn about the world from this more-than-human perspective. 52 Dancing with the Nonhuman HRI ’23 Companion, March 13–16, 2023, Stockholm, Sweden 3.1 Intra-action: Agency as Enactment are łforms of co-dependence rather than pairing of two different thingsž [14]. Elena Knox’s creative, performance-based research practice with gynoid robots brings to life the ethics of artificial ‘life’, including questions of ageing, sexuality, and self-preservation [36]. Her work not only exposes fetishist representations of the female body but also explores these complex questions of care and responsibility from the robot’s perspective [37]. It could be argued that most robotics practices consider a robot’s (social) agency to be an intrinsic attribute; something that the robot can be given or programmed with based on its features and behaviors. From a feminist new materialist perspective, however, agency is a matter of enactment [2, 3], unfolding in the dynamics of exchange as participants (human and nonhuman) intra-act, co-constituting each other [2, 5, 44]. A performative approach to human-robot interaction design thus shifts the focus from representation to performativity: our understanding of the situation of encounter and the robot’s capacity to participate in the relationmaking have a greater effect on the social experience than a robot’s appearance or predefined personality. Barad’s new materialist account of performativity is about the mattering of bodies and thingsÐtheir processes of materializationÐas well as the mattering of our practices (e.g., that of robotics)Ðhow they differentially enact boundaries that give rise to, for instance, subjects and objects or the separation of mind and body [2]. Countering dualist thinking, Barad’s agential realism rethinks the construct of interaction, which assumes two separate, pre-existing entities; intra-action, on the other hand, acknowledges how subjects and objects emerge in the dynamics of the encounters [3, see also 44]. 2.1 The Potential of Movement Our Machine Movement Lab project looks at the potential of movement for situated meaning-making and corporeal empathy across human and nonhuman embodiments. According to Naoko Abe [2022], artists’ creativity and imagination can give rise to a paradigm shift in robot movement design by seeking movement capacities that are specific to a machinelike robot morphology rather than replicating human movement. Abe characterizes robot-specific motion as being designed and generated based on a robot’s material and morphological features, including its electro-mechanical abilities [1]. Robotic movement design so far has largely focused on locomotion, human(like) non-verbal communication, or imbuing machines with an expressive character or personality [1, 24]. For the latter, movement serves as a medium for łaccurately expressing the robot’s purpose, intent, state, mood, personality, attention, responsiveness, intelligence, and capabilitiesž [31, p. 91]. More experimental approaches to motion design for machines can often be found in theatrical or dance performance practices. The performance work Grace State Machines by Bill Vorn, Emma Howes and Jonathan Villeneuve [2007], for instance, explores questions of kinaesthesis and perception in a dialogue between a dancer and a machine performer. The artists sought to propel the conversational exchange between human and abstract robot through the empathy produced by the complex movements that the articulated machine is able to perform [48]. Jochum and her collaborators [34, 35] have shown that strategies adopted from choreography, theatrical performance practices or traditional puppetry can inform creative solutions for robot motion design. LaViers et al. [2018] explore tools and techniques from choreography and somatics for expanding expressive robotic systems. Catie Cuan et al. [2018] explores how a dance performance and audiences witnessing humanśrobot interaction being performed can transform their perception of a robot, framed łthrough the lenses of power, valence, and acceptabilityž [8, p. 255]. Gil Weinberg’s Shimon, an interactive robotic marimba player [32] demonstrates a motion-centric approach to robot design, which successfully combines mechanical looks and gestural movement to propel improvisational performances with human band members [31]. 3 3.1.1 Agential Enactment in HRI. We argue that this reconceptualization of (social) agency and, with it, the meaning-making that we associate with social agents, being enacted as part of the intra-actional encounter, dramatically expands our possibilities for designing robots to participate in social scenarios. Frank Hegel et al. [2009], for instance, state that łwith a functional designed robot it is impossible to express human facial expressions and consequently emotional displaysž [30, p. 173]. However, if the enactment of meaning and affect do not rely on representational expressions or predefined emotional displays, we can begin to imagine rich, social encounters with machinelike or łfunctionalž robots. The social and affective potential of movement, sound or other relation-making modes (e.g., dramaturgical strategies) for rendering artefacts relational is already well-known to artists. The Table: Childhood (1984ś 2001) by Max Dean, Raffaello D’Andrea and Matt Donovan, for example, generates surprising relational dynamics between a table and audiences [15]. The familiar artefact is made strangely evocative through situated movement qualities that afford the table to perform intention and attentiveness. A performative approach to human-robot interaction thus favors the artefact’s relational, participatory capacities instead of representational attributes and looks at the meaning-making potential of human-machine couplings rather than the predefined meaning of an individual machine agent. 3.2 Diffraction FEMINIST CONCEPTS FOR RETHINKING HUMAN-ROBOT RELATIONSHIPS Diffraction as a methodological tool can attune us to the effects of difference that characterize our more-than-human relationships. The representationalist approach of reflection, in contrast, serves to keep us and the world at a distance. Humanlike robots present us with a superficial reflection of ourselves and thus increases the distance between us and the machine beneath. Whereas diffraction traces the effects of difference and, according to Haraway [1992], In the following, we will discuss two feminist concepts that are core to our practice: intra-action [2, 3] and diffraction [3, 27]. The remainder of the paper then explores how they materially mobilize the diffractive processes that our MML project engages in, across the lab, dance studio and the gallery space. 53 HRI ’23 Companion, March 13–16, 2023, Stockholm, Sweden Petra Gemeinboeck & Rob Saunders is ła mapping of interference, not of replication, reflection, or reproductionž [27, p. 300]. In line with Barad’s agential enactment, a diffractive approach calls łinto question the givenness of the differential categories of ‘human’ and ‘nonhuman’, examining the practices through which these differential boundaries are stabilized and destabilizedž [2, p. 808]. Diffraction as both a tool and a practice can make manifest the destabilization and stabilization of boundaries [47] and serve to collapse the distance between subjects and objects and bring them into resonance. extend into and embody nonhuman entities [12] can spawn alternative, more-than-human approaches to understanding human-robot relationships. Movement creation and its dynamic qualities and connection-making, empathy-inducing capacities in tandem with different modes of embodiment thus is the engine driving our performative investigation and material, aesthetic explorations. 3.2.1 Diffraction in MML. Rather than juxtaposing humans and machines or making them appear to be the same, MML seeks symbiotic possibilities based on difference patterns that render the boundaries between subjects and objects more elastic. Diffraction and patterns of interference are thus not only figurative concepts but serve as a methodological tool for łattending to and responding to the effects of differencež [3, p. 72]. The entanglement of bodies and things (as explored in following section) maps their effects of difference similarly to Barad’s description of superposition: Our approach exploits one of the most significant social characteristics of robotic artefacts: that we can bodily resonate and relationally make meaning with their embodied dynamics and the spatial configurations they produce. Movement in MML is harnessed as a dynamic, relational, and generative force, producing the trajectories along which meanings are made and unmade. Attending to the potential of movement dynamics and how they unfold through łspatial, temporal, and energic qualitiesž [42, p. 49] allows for meaningand relation-making in the encounter itself. Understanding movement, both performatively and from the perspective of phenomenology and embodied cognition, we bodily participate in the generation of meaning, łoften engaging in transformational and not merely informational interactions; [we] enact a worldž [13, p. 39]. As we enact and experience meaning through movement, we also make sense of other bodies by resonating with them and their movements; we bodily resonate with the affective qualities or environmental affordances produced by movement [17]. The two methodological stages of MML that we will take a closer look at below, Performative and Relational Body-Mapping (PBM and RBM), explore how the relational, enactive potential of movement can scaffold a robot’s ability to participate in the dynamic, social processes of the encounter; movement here is understood as a mattering, diffractive phenomenon itself, generating interferences and entanglings that give rise to affects and meanings. 4.1 Movement as a Dynamic, Relational, and Generative Force ł... waves can overlap at the same point in space. When this happens, their amplitudes combine to form a composite wave form [and] the resultant wave is a sum of the effects of each individual component wave; that is, it is a combination of the disturbances created by each wave individually. This way of combining effects is called superposition.ž [3, p. 76] Similar to the transferences we witness in the diffraction or interference of waves, mechanisms of composition, divergence, intermingling, superposition, etc., also shape our material, choreographic experiments for investigating human-nonhuman entanglements (see Figures 2, 4 and 5). The difference patterns emerging from the more-than-human corporeal alliances that our performancemaking promotes allow us to witness relationships-in-the-making and their evolving effects of difference-in-relation. 4 CASE STUDY: MACHINE MOVEMENT LAB A box of a size roughly reaching to the mid of your thighs, swivels on one of its corners, lingers as if hesitating, then promptly twists up, wiggles around and precariously teeters on one edge before it bumps onto the ground and comes to a sudden halt. All the while, these whimsical moves produce planes that apparently look at youÐlike facesÐbut they come about, pivot and shift as swiftly as this ‘box’ changes its trajectory. Figure 2: Dancer Audrey Rochette being entangled with the PBM prototype costume, 2019. Photo: P. Gemeinboeck 4.2 Performative Body-Mapping (PBM) MML experiments with material renderings of the concepts of agential enaction and diffraction to render subject-object boundaries more tangible, inhabitable, and reconfigurable. To do this, our approach embraces the differences and asymmetries between humans and machines and investigates creative strategies for reimagining our relationships with them. It is a collaboration with dancers, choreographers, AI researchers, engineers, and numerous materials (from cardboard, PVC tubes, plywood to aluminum framing, motors, motor controllers, cables, cable binders, and software programs), across robotics labs, dance studios, fab labs, and gallery spaces over the past seven years. At its core, MML revolves around the idea that the generative potential of movement and our human capacity to MML started with a series of material encounters, in which we collaborated with dance performers to bodily investigate the performative potential of a wide range of material shapes (e.g., cylindrical, cuboid, cubic, tetrahedral) made of qualitatively different materials (e.g., stiff, elastic, springy). To put them into motion and explore their relational potential, we asked the dancers to inhabit and activate them from the inside, rather than puppeteering them from the outside. We mostly experimented with simple geometric shapes as we found that they made it easier to make sense of movement qualities and how they transform the afterfact’s shape and its affective potential. 54 Dancing with the Nonhuman HRI ’23 Companion, March 13–16, 2023, Stockholm, Sweden Our motivation for entangling dancers with a potential abstract machine shape was driven by a curiosity to explore how this intimate material, intercorporeal [17] collaboration could advance our understanding of a particular robotic embodiment and the kinds of relational behaviors it could engender. Our goal is not to create a robot capable of simply imitating a human wearing a robot costume. Rather, the aim of this entanglement is to enable dancers to bodily extend themselves and their kinaesthetic awareness into the different body of an abstract shape, with its unique spatial-material affordances and social potential (see Figure 2). Both, mechanical design and the learning of motion behaviors can then be informed by this hybrid interior perspective: due to the different morphological makeup of humans and nonhumans involved, the movement behaviors resulting from this enmeshment are neither human nor nonhuman. Instead, they are shaped by the human dancers’ movement knowledge and sociocultural background, the material affordances and constraints of the costume, as well as the emergent possibilities that this close collaboration brings about within specific choreographic and/or dramaturgical contexts. Over the next three years, our initial exploratory experiments evolved into our PBM methodology for designing robots, which utilizes a wearable costume as a low-fidelity prototype standing in for the shape of a becoming-robot and its spatial-relational affordances. Figure 2 shows our first PBM prototype costume with the regular shape of a cube (75x75x75cm). The plain, omni-directional geometry of a cube was interesting to us because it does not imitate a living entity and offers a blank canvas, we found, for the dynamics unfolding in this performer-costume entanglement. The PBM costume [22] serves to: cube performer (Figure 3). The cube performer and its mechanical makeup was designed based on more 15 hours of motion capture recordings of the performer-costume entanglement [22]. 4.2.1 Robot behavior derived from more-than-human interference patterns. PBM both relies on and exploits the differences between human and nonhuman embodiments and the interference patterns they can effect. The performer-costume entanglement serves to bring forth a material conversational exchange between these two very different embodiments, allowing the human performer to step into the shoes of a cubic performer. This corporeal, material intermeshing produces a wide range of unique interference patterns that we capture in the form of kinetic dynamics that the cube performer (robot prototype) can learn from in the machine learning stage. The motion data provides the biases and constraints that shape the robot’s improvisational skills and, with it, dynamic, relational capacities. We believe that the underlying machine learning system learning from these hybrid patterns, both more-than-human and more-than-machine, can render the robot both unique, machinelike and readable for human participants. The latter is based on the idea that making sense and meaning in the affective exchange of the encounter relies on embodied forms of empathy [22, 19, 9, see also, 12] that can be induced by such hybrid motion patterns without relying on humanlike appearances. So far, we tested our idea in two participatory studies and feedback from participants has indicated support for this proposition [22, 23]. Gemeinboeck [2022] provides a few situated, material accounts of the experimental process and how the juxtaposition of human body and cubic costume can bring about a productive superposition. The next stage in our methodological development builds on PBM to explore the agential possibilities of encounter itself: Relational-Body Mapping (RBM) probes into human-robot relationships, the process of meaning-making and how it enacts subjects and objects. (1) provide dancers with an embodied insight into the morphological characteristics of a specific robot embodiment; (2) support the development of a repertoire of movement qualities that are specific to the robot’s morphological form; and (3) capture of movement data that a robot resembling the costume’s shape can learn from, with little or no translation. PBM combines the ideas of a theatrical costume and demonstration learning from human-robot interaction research, enabling dancers to ‘step into’ and inhabit the other, more-than-human embodiment of the becoming-robot in the form of a costume and to learn to kinesthetically extend into and move with it. PBM thus sidesteps the technical challenge of łrobot programming by demonstrationž [4], which cannot account for the differences between the programmer’s human embodiment and that of the learning robot [11]. The PBM costume affords dancers to bodily feel into the asymmetric relational potential of this different embodiment and to resonate with them, to the extent to which their proprioception transforms to encompass their cube entanglement and its movement and position in space. We deliberately do not work with narratives or the expression of (human) emotional states; instead, performers sometimes use mental imagery of nonhuman dynamics or materials to guide the reconfiguring of their bodies and finding of new movement patterns, such as the image of a pressure-cooker or the fluid quality of water. The movements resulting from this performer-cube entanglement are then captured to inform the improvisational machine learning of our cube robot prototype, the Figure 3: Cube performer, robot prototype, at the Games as Performing Arts Festival, AMATA, Falmouth University, UK, 2018. Photo: P. Gemeinboeck 4.3 Relational Body-Mapping (RBM) RBM expands PBM and the single performer-cube entanglement to seek more complex, nested entanglements, e.g., a (human) performer with a (robotic) cube performer; a performer-costume with a cube performer; or a (human) performer with a performer-costume and a cube performer, and so on (Figure 4). RBM focuses on the relational potential of the performer-costume entanglement by exploring how performance-making processes 55 HRI ’23 Companion, March 13–16, 2023, Stockholm, Sweden Petra Gemeinboeck & Rob Saunders can be about mingling (as in meld or unifying transformation) or mangling (disconnect or deformation); both are of interest from an aesthetic-affective viewpoint and have a different effect on the formation of subject-object boundaries. The boundary-shaping effect is fluid and quickly changes based on different movement dynamics and spatial alliances. Importantly, the interference patterns are not controlled by the dancers; meaningful intra-actions require us to bodily listen to and collaborate with the artifacts’ transformational material capabilities. Figure 4: Relational Body-Mapping (RBM) seeks nested performer-cube entanglements; with dancers Arabelle FrahnStarkie and Felix Palmerson, 2022. Photo: P. Gemeinboeck can contribute to reimaging our relationships with machines. How can we probe into the making of subjects and objects and render the boundaries that delineate them more porous? Rather than juxtaposing humans and machines or making them appear to be the same, we seek more-than-human possibilities to relate with machinelike robots based on difference patterns. Drawing on techniques from choreographic and dramaturgical performance-making, RBM builds on feminist conceptions of bodies and things as always already inherently porous, relational, and reconfigurable. Importantly, we do not seek to choreograph the encounter by fixing configurations or predefining meanings. Rather, choreography and dramaturgy provide us with the tools to imagine or notice interference patterns and attend to the emergence of new relations and any meanings they may acquire. The RBM entanglement serves to: (1) study a range of embodied, exploratory scenarios with machinelike robot morphologies and their potential to enact meaningful, more-than-human configurations and exchanges; (2) develop participatory, experiential scenarios with machinelike robot morphologies that allow participants to explore possible relationships or take the perspective of the machine; (3) capture movement and sensor data that can expand a robot’s skills for improvising motion behaviors in relation to other actors. Figure 5: Stretching the SubjectśObject Boundary; with dancers Arabella Frahn-Starkie and Felix Palmerson, 2022. Photo: P. Gemeinboeck Our performance-making practice develops choreographic-dramaturgical strategies for producing and attending to the emergence of diffraction and interference patterns as the dancers’ bodies extend toward, align with, bend with, intersect with or stretch across the cubic artefacts and/or each other. Haraway reminds us that diffraction patterns map the effects of differences rather than the differences [26]; in our practice they trace the continuous making and re-making of subjects, objects, and subject-object hybrids. Needless to say that from both a feminist and a performance-making perspective, human bodies do not necessarily align with the boundaries of a subject, nor do nonhuman things with the bounds of an object. On the contrary, subjects and objects are not pregiven, they are fluidly enacted and continuously re-enacted in the encounter, propelling the evolution of more-than-human relationships. The following outlines three of the most common interference patterns that we have witnessed and experimented with so far, characterized by the degree of entangledness and emergent effects of difference. 4.3.2 Stretching the subject-object boundary to become a gateway. In our experiments we found that there is a hybrid state in which the dancers dwell on and attend to the threshold of the boundary between their body and that of the object or thing (Figure 5). Rather than a barrier, the boundary between assumed subject and assumed object stretches to become a gateway for a body becoming-thing (object) or a thing becoming-body (subject) or something in-between the two. For relationship-making, this boundary stretching and dwelling not only expands the kinds of relationships that we can experience but also allows for exploring the hybrid affordances that such intra-actions bring about. 4.3.3 Nested interference patterns. When more than one performercostume entanglement intra-act with one another we witness more complex, nested patterns of interference. These often result in a continuous series of attachments and detachments between bodies and things, and body-things, and aligning of edges, corners and body parts that disassociate again, only to reconfigure themselves in a new constellation superposing the corporeal and geometric. We found that a core transformative potential of these nested interference patterns is to render the cube artefact tentacular, i.e., capable of extending toward or being connected to other bodies and/or things. Haraway [2016] speaks of 4.3.1 Spatial superpositions. This interference pattern occurs in both PBM and RBM and is about carefully exploring and attending to how the dancers’ bodies and the cubic artefacts interfere and the kinds of spatial and affective relations this effects. Interfering here 56 Dancing with the Nonhuman HRI ’23 Companion, March 13–16, 2023, Stockholm, Sweden łtentacular ones [and how they] make attachments and detachments; they make cuts and knots; they make a difference; they weave paths and consequences but not determinisms; they are both open and knotted in some ways and not othersž [29, p. 31] This tentacular capacity not only emerges in nested performercostume entanglements but also characterizes the relational-affective capacity of the cube performer (robot) intra-acting with a dance performer or performer-costume entanglement (see Figure 6). 4.3.4 Performance-making as immersive low-fidelity prototyping. Looking at human-robot relationships through the lens of performance-making and attending to these more-than-human interference patterns assists with better understanding some of the inner workings of the performer-cube entanglement with regards to its relation-making capacity. The latter is about exploring and appreciating our (human) capacity to extend toward or empathize with other, nonhuman bodies and how they intra-actively participate and extend back toward us. The performance-making practice also informs our development of participatory scenarios and performancebased prototyping with both robot costumes and robot prototypes (e.g., cube costumes and the cube performer). The RBM costume thus opens up a form of inexpensive, immersive low-fidelity prototyping, allowing rapid scenario development and embodied forms of perspective taking to facilitate bodily engagement and participatory design processes based on intercorporeal resonances and empathy [18], perspective-taking, and interaffectivity [17]. According to J.D. Zamfirescu-Pereira et al. [2021], such low-fidelity prototyping can broaden participation and lead to greater diversity in human-robot interaction design. Figure 6: Rehearsal of Dancing with the Nonhuman, with dancers Felix Palmerson and Arabella Frahn-Starkie, Tin Sheds Gallery, Sydney, 2022. Photo: K. Mah 5 DISCUSSION: FROM HUMAN-ROBOT INTERACTION (HRI) TO HUMAN-ROBOT EXPERIENCE (HRX) Shifting from a representationalist to a performative view, humanrobot relationships are not about interactions between separate entities with predefined, built-in agencies and fixed boundaries but instead open up to dynamic experiences, where both human and nonhuman participants play a part in the meaning-making process of the encounter. We coined the term Human-Robot Experience (HRX) to emphasize the performative characteristics of our practice: foregrounding material embodiment and kinesthetic, embodied empathy to attend to human-robot relationships as specific experiential situations that give rise to the differential enactment of subjects and objects. The core event propelling Dancing with the Nonhuman, for instance, is an unfolding of intracorporeal intraactions, producing interference patterns that transform both the human bodies and artefacts, with regards to their resultant shapes, agential possibilities, and their meanings. Each iteration of the performance produces different interference patterns and different kinds of shapes, agencies and meanings. HRX proposes that we need to position ourselves in the middle of the encounter to explore the potential of these specific experiential situations by attending to the ongoing dynamics that a particular unfolding scenario makes possible. Importantly, such a performative reconfiguring of human-robot interactionÐhuman-robot intra-actionÐalso frees the machine to participate as a machinelike participant, no longer requiring a humanlike veneer. HRX not only embraces but also seeks to aesthetically exploit the effects of difference between humans and machines. The aesthetic potential of our practice, we believe, lies in the transformative combination of asymmetries arising łfrom the very different grounds of being and acting that animate human and machine participantsž [43, p. 18]. The affective potential of these physical and dramaturgical more-than-human alliances flows from the seemingly dissonant intercorporeal resonances that they effect. Rather than serving to make the strange look more familiar, aesthetics here is about rendering the differences between humans and robots more relational. 4.3.5 Performance score for Dancing with the Nonhuman. As part of our RBM development, we are generating a semi-improvisational performance score for Dancing with the Nonhuman to open our process up to the diverse embodied perspectives of audiences and potential stakeholders and engage them in questions of corporeal empathy and intercorporeal resonances with machinelike robots. The first iteration of this score is for a semi-improvised dance performance with two (human) performers, two cube artefacts (costumes) and one cube (robot) performer (Figure 6). At the time of writing, we are yet to perform this work in public. As we strive to collapse the distance between subjects and objects, we also seek to render the boundary between performers and audiences more porous. To facilitate this, Dancing with the Nonhuman is to be performed in gallery spaces, rather than on stage, and includes intro and outro stages, in which the performance site is gradually un-/established. During these transitional stages, audiences are encouraged stay inside or join the performance space and get up close with the artefacts, including the (robotic) cube performer. At this stage, we are mostly interested in exploring whether audiences are willing and able to bodily imagine themselves as being a part of this more-than-human ensemble. Performance events will include an audience survey, which engages audience members in reflections on their ability to make sense of and/or empathize with the social alliances expressed in the form of spatial alignments and relational configurations. 57 HRI ’23 Companion, March 13–16, 2023, Stockholm, Sweden Petra Gemeinboeck & Rob Saunders Within the HRX framework, performance-making becomes a mode of generative-diffractive inquiry into the re/enactment of boundaries as part of the dynamic configurings unfolding in the encounter. It is about carefully attending to possible re/configurings of human performers and nonhuman artefacts; how they matter, couple, interfere, and łundo and redo each otherž [12, p. 61], and how their differences in-relation give rise to meanings beyond traditionally assumed human-nonhuman boundaries. This notion of intercorporeal communication as an interference pattern builds on the ethological concept of ‘embodied empathy’ [12], which refers to the reciprocal yet asymmetrical attunement occurring in interspecies communication. Diffraction not only serves as a figurative lens but also shapes our material process by allowing us to attune to the specific, material effects of different bodies and things. Entanglement, interference, and superposition thus become corporeal-material tools in our dramaturgical/choreographic process. The hybrid couplings that our HRX practice provokes challenge some of the hierarchical dualisms that are still engrained in robotic design practices and, with it, our socio-technical visions of our future by reorienting our representational focus toward the performative potential of the encounter. We introduced our collaborative, transdisciplinary Machine Movement Lab (MML) project, which develops a diffractive approach to imagining human-robot relationships based on performance-based inquiries into the relational enactment of subjects and objects. Our creative practice builds on Barad’s new materialist conceptions of agency and difference and aesthetically puts to work the difference patterns at play in our relationships with robots. We discussed two feminist concepts, intra-action [2, 3] and diffraction [3, 27], that foreground human-machine differences and are core to the workings of the subject-object interference patterns that characterize human-robot relationships. Our two methodologies, PBM and RBM, promote creative, embodied, and participatory strategies for diffracting robot design and look at the making of human-robot relationships through the lens of performance-making. Our HumanRobot Experience (HRX) framework emphasizes material embodiment and kinesthetic, embodied empathy to attend to human-robot relationships as specific experiential situations that give rise to the differential enactment of subjects and objects. Utilizing costumes as low-fidelity robot prototypes and a performance-based approach to investigate our ability to bodily resonate and empathize with a particular robot morphology, we believe, can widen, and diversify the opportunities for involving researchers, practitioners, and potential stakeholders in robot design. Techniques from dance and concepts from feminist new materialism, as demonstrated in our work, not only expand our design toolbox but open up a more horizontal playground for the making of human-robot relationships. 5.1 Embodied Low-Fidelity Co-Design A diffractive approach means to design with an encounter, as opposed to designing for an encounter. It means to position ourselvesÐ as the designersÐin the middle of the encounter and to form alliances between bodies and materials and their relational, affective capacities and aesthetically attend to possibilities as they emerge. Diffracting the making-of encounter and, with it, human-robot relationships, also means that we need to attend to the politics of who gets to define and design our relationships. Designing robots and imagining the relational scenarios in which we encounter, work, or live with them thus requires a much wider participation of potential stakeholders, most of whom are currently excluded from the making of our socio-technical futures. In our next research stage, our diffractive performance-making practice will serve as a platform for developing embodied participatory HRX strategies for playfully enacting relationships with low-fidelity robot prototypes. Following PBM, these playful relational enactments, when recorded, can inform the mechanical design and machine learning of robots. We believe that embodied play and empathy with lowfidelity prototypes could open up new and more diverse possibilities for intercorporeal meaning-making in human-robot relationships by promoting robot designs with uniquely machinelike embodiments and expanded, more-than-human behavioral repertoires. 6 ACKNOWLEDGMENTS The research presented in this article has been partly supported by several funding bodies: the Australian Government through the Australian Research Council (DP160104706 and FT190100567); and the Austrian Government through the Austrian Science Fund (AR545); and the EU Framework Programme (FP7) European Research Area Chairs Scheme project (621403). We would like to thank our collaborators for their contributions to this research: Roos van Berkel (Eindhoven University of Technology, NL); Maaike Bleeker (Utrecht University, NL); Katrina Brown (Falmouth University, UK); Arabella Frahn-Starkie (Melbourne, AU); Rochelle Haley (University of New South Wales, Sydney, AU); Lesley van Hoek (Rotterdam, NL), Sarah Levinsky (Falmouth University, UK); Linda Luke (De Quincey Co., Sydney, AU); Dillon McEwan (Sydney, AU); Kirsten Packham (Sydney, AU); Felix Palmerson (Melbourne, AU); Marie-Claude Poulin (kondition pluriel, Montreal, CA; University of Applied Arts Vienna, AT); Tess de Quincey (De Quincey Co., Sydney, AU); Audrey Rochette (Montreal, CA). SUMMARY REFERENCES [1] Naoko Abe. 2022. Beyond anthropomorphising robot motion and towards robotspecific motion: consideration of the potential of artistÐdancers in research on robotic motion. Artificial Life and Robotics 27 (2022), 777ś785. https://doi.org/10. 1007/s10015-022-00808-0 [2] Karen Barad. 2003. Posthumanist performativity: toward an understanding of how matter comes to matter. Signs: Journal of Women in Culture and Society 28, 3 (2003), 801ś831. https://doi.org/10.1086/345321 [3] Karen Barad. 2007. Meeting the Universe Halfway: Quantum Physics and the Entanglement of Matter and Meaning. Duke University Press, Durham, NC. [4] Aude Billard, Sylvain Calinon, Rüdiger Dillmann, and Stefan Schaal. 2008. Robot programming by demonstration. In Springer Handbook of Robotics, B. Siciliano and In this article, we argue that feminist reconceptualisations of agency and difference could greatly expand our possibilities for both relating to robots in social scenarios and designing them as social agents. A performative approach to human-robot interaction design favors the artefact’s relational, participatory capacities over representational attributes and looks at the meaning-making potential of human-machine couplings rather than the predefined meaning of an individual machine agent. 58 Dancing with the Nonhuman [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] HRI ’23 Companion, March 13–16, 2023, Stockholm, Sweden [25] Andrea L. Guzman. 2020. Ontological boundaries between humans and computers and the implications for human-machine communication. Human-Machine Communication 1 (2020), 37ś54. https://doi.org/10.30658/hmc.1.3 [26] Donna J. Haraway. 1991. Simians, Cyborgs, and Women: The Reinvention of Nature. Routledge, New York, NY. [27] Donna J. Haraway. 1992. Promises of Monsters: A Regenerative Politics for Inappropriate/d Others. In Cultural Studies, L. Grossberg, C. Nelson, and P.A. Treichler (Eds.). Routledge, New York, NY, 295ś337. [28] Donna J. Haraway. 1997. Modest_Witness@Second_Millennium.FemaleMan_Meets_OncoMouse: Feminism and Technoscience. Routledge, New York, NY. [29] Donna J. Haraway. 2016. Staying with the Trouble: Making Kin in Chthulucene. Duke University Press, Durham, NC. [30] Frank Hegel, Claudia Muhl, Britta Wrede, Martina Hielscher-Fastabend, and Gerhard Sagerer. 2009. Understanding social robots. In Proceedings of the Second International Conferences on Advances in Computer-Human Interaction. IEEE, New York, NY, Cancun, Mexico, 169ś174. https://doi.org/10.1109/ACHI.2009.51 [31] Guy Hoffman and Wendy Ju. 2014. Designing Robots with Movement in Mind. Journal of Human-Robot Interaction 3, 1 (2014), 91ś122. https://doi.org/10.5898/ JHRI.3.1.Hoffman [32] Guy Hoffman and Gil Weinberg. 2011. Interactive improvisation with a robotic marimba player. Autonomous Robots 31 (2011), 133ś153. https://doi.org/10.1007/ s10514-011-9237-0 [33] Elizabeth A. Jochum, Louis-Philippe Demers, Bill Vorn, Evgenios Vlachos, Paul B. McIlvenny, and Pirkko L. Raudaskoski. 2018. Becoming Cyborg: Interdisciplinary approaches for exoskeleton research. In EVA Copenhagen 2018ÐPolitics of the MachinesÐArt and After. 1ś9. [34] Elizabeth A. Jochum and Damith Heath. 2016. Robot Choreography: Performance Paradigms for Experimental HRI Setups. In Proceedings of Robophilosophy/TRANSOR 2016. IOS Press, Washington, DC, 86ś88. https://doi.org/10.3233/ 978-1-61499-708-5-86 [35] Elizabeth A. Jochum and Todd Muphey. 2014. Programming Play: Puppets, Robots, and Engineering. In Routledge Companion to Puppetry and Material Performance, D. Posner, C. Orenstein, and J. Bell (Eds.). Routledge, New York, NY, 308ś321. [36] Elena Knox. 2016. Degrees of freedom. Fibreculture, Special Issue on Creative Robotics 28 (2016). https://doi.org/10.15307/fcj.28.204.2017 [37] Elena Knox. 2019. Gynoid Survival Kit. Queer STS Forum 4 (2019), 21ś48. https: //queersts.com/wp-content/uploads/2020/01/Forum-4-2019-21-48-Knox.pdf [38] Kate Ladenheim and Amy LaViers. 2021. Babyface: Performance and Installation Art Exploring the Feminine Ideal in Gendered Machines. Frontiers in Robotics and AI 8 (2021). https://doi.org/10.3389/frobt.2021.576664 [39] Amy LaViers, Catie Cuan, Catherine Maguire, Karen Bradley, Kim B. Mata, Alexandra Nilles, Ilya Vidrin, Novoneel Chakraborty, Madison Heimerdinger, Umer Huzaifa, Reike McNish, Ishaan Pakrasi, and Alexander Zurawski. 2018. Choreographic and somatic approaches for the development of expressive robotic systems. Arts 7, 2 (2018), 1ś21. https://doi.org/10.3390/arts7020011 [40] Andrew Lepecki. 2012. Moving as Thing: Choreographic Critiques of the Object. October 140 (Spring 2012) (2012), 75ś90. https://doi.org/10.1162/OCTO_a_00090 [41] Eleanor Sandry. 2016. The potential of otherness in robotic art. In Robots and Art: Exploring an Unlikely Symbiosis, H. Damith, K. Christian, and Stelarc (Eds.). Springer, Singapore, 177ś189. https://doi.org/10.1007/978-981-10-0321-9_9 [42] Maxine Sheets-Johnstone. 2012. From Movement to Dance. Phenomenology and the Cognitive Sciences 11 (2012), 39ś57. https://doi.org/10.1007/s11097-011-9200-8 [43] Jackie Stacey and Lucy Suchman. 2012. Animation and Automation: The liveliness and labours of bodies and machines. Body and Society 18, 1 (2012), 1ś46. https: //doi.org/10.1177/1357034X11431845 [44] Lucy Suchman. 2007. Human-Machine Reconfigurations: Plans and Situated Actions (2nd ed.). Cambridge University Press, Cambridge, UK. [45] Lucy Suchman. 2011. Subject Objects. Feminist Theory 12, 2 (2011), 119ś145. https://doi.org/10.1177/1464700111404205 [46] Pat Treusch. 2021. Robotic Knitting Re-Crafting Human-Robot Collaboration Through Careful Coboting. De Gruyer, Berlin. https://doi.org/10.1515/ 9783839452035 [47] Iris van der Tuin. 2014. Diffraction as a Methodology for Feminist OntoEpistemology: On Encountering Chantal Chawaf and Posthuman Interpellation. Parallax 20, 3 (2014), 231ś244. https://doi.org/10.1080/13534645.2014.927631 [48] Bill Vorn, Emma Howes, and Jonathan Villeneuve. 2007. Grace State Machines. https://billvorn.concordia.ca/robography/GraceState.html [49] J.D. Zamfirescu-Pereira, David Sirkin, David Goedicke, Ray LC, Natalie Friedman, Ilan Mandel, Nikolas Martelaro, and Wendy Ju. 2021. Fake It to Make It: Exploratory Prototyping in HRI. In Companion of the 2021 ACM/IEEE International Conference on Human-Robot Interaction (HRI ’21 Companion). ACM, New York, NY, 19ś28. https://doi.org/10.1145/3434074.3446909 O. Khatib (Eds.). Springer, New York, NY, 1371ś1394. https://doi.org/10.1007/9783-540-30301-5_60 Elizabeth Broadbent. 2017. Interactions with robots: The truths we reveal about ourselves. Annual Review of Psychology 68 (2017), 627ś652. https://doi.org/10. 1146/annurev-psych-010416-043958 Claudia Castañeda and Lucy Suchman. 2014. Robot visions. Social Studies of Science 4, 3 (2014), 315ś341. https://doi.org/10.1177/0306312713511868 Rebecca Coleman, Tara Page, and Helen Palmer. 2019. Introduction. MAI: Feminism & Visual Culture, Special Issue on Feminist New Materialist Practice: The Mattering of Methods 4, Summer 2019 (2019), Article 1. Catie Cuan, Ishaan Pakrasi, Erin Berl, and Amy LaViers. 2018. CURTAIN and Time to Compile: A Demonstration of an Experimental Testbed for Human-Robot Interaction. In Proceedings of the 27th IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN). IEEE, New York, NY, Nanjing, China, 255ś261. https://doi.org/10.1109/ROMAN.2018.8525520 Luisa Damiano and Paul G. Dumouchel. 2020. Emotions in relation. Epistemological and ethical scaffolding for mixed human-robot social ecologies. HUMANA.MENTE Journal of Philosophical Studies 13, 37 (2020), 181ś206. https: //www.humanamente.eu/index.php/HM/article/view/321 Kerstin Dautenhahn. 2013. Human-robot interaction. In Encyclopedia of humancomputer interaction, (2nd edition), M. Soegaard, R.F.D. Dam, and K. Aarhus (Eds.). Interaction Design Foundation, Aarhus, Denmark, 2283ś2366. Kerstin Dautenhahn, Chrystopherm Nehaniv, and Aris Alissandrakis. 2003. Learning by experience from othersÐsocial learning and imitation in animals and robots. In Adaptivity and Learning: An Interdisciplinary Debate, R. Kühn, R. Menzel, W. Menzel, U. Ratsch, M. Richter, and I. Stamatescu (Eds.). Springer, New York, NY, 217ś421. https://doi.org/10.1007/978-3-662-05594-6_17 Vinciane Despret. 2013. Responding bodies and partial affinities in human-animal worlds. Theory, Culture & Society 30, 7ś8 (2013), 51ś76. Ezequiel A. Di Paolo, Marieke Rohde, and Hanne De Jaegher. 2010. Horizons for the enactive mind: values, social interaction, and play. In Enaction: Towards a New Paradigm for Cognitive Science, J. Stewart, O. Gapenne, and E. Di Paolo (Eds.). MIT Press, Cambridge, MA, 33ś87. https://doi.org/10.7551/mitpress/9780262014601. 003.0003 Marco Donnarumma. 2017. Beyond the Cyborg: Performance, attunement and autonomous computation. International Journal of Performance Arts and Digital Media 13, 2 (2017). Josée Drouin-Brisebois. 2018. Caught in the act: the viewer as performer. ABC Art Books Canada Distribution. Peter Eckersall, Helen Grehan, and Edward Scheer. 2017. New Media Dramaturgy: Performance, Media and New-Materialism. Palgrave Macmillan, London, UK. Thomas Fuchs. 2016. Intercorporeality and Interaffectivity. In Intercorporeality: Emerging Socialities in Interaction, C. Meyer, J. Streeck, and S. Jordan (Eds.). Oxford University Press, Oxford, UK, 3ś24. https://doi.org/10.1093/acprof:oso/ 9780190210465.003.0001 Thomas Fuchs and Sabine C. Koch. 2014. Embodied Affectivity: On Moving and Being Moved. Frontiers in Psychology 5, 3 (2014). https://doi.org/10.3389/fpsyg. 2014.00508 Petra Gemeinboeck. 2021. The Aesthetics of Encounter: A RelationalPerformative Design Approach to Human-Robot Interaction. Frontiers in Robotics and AI 7 (2021). https://doi.org/10.3389/frobt.2020.577900 Petra Gemeinboeck. 2022. Difference-in-relation: Diffracting human-robot encounters. Matter: Journal of New Materialist Research (Special Issue on Prospects for a New Materialist Informatics) 3, 1 (2022). https://doi.org/10.1344/jnmr.v3i1.38958 Petra Gemeinboeck and Rob Saunders. 2016. The Performance of Creative Machines. In Cultural Robotics: First International Workshop, CR 2015, IEEE ROMAN 2015, Kobe, Japan, August 31, 2015. Revised Selected Papers, Jeffrey T.K.V. Koh, Belinda J. Dunstan, David Silvera-Tawil, and Mari Velonaki (Eds.). LNAI, Vol. 9549. Springer, 159ś172. https://doi.org/10.1007/978-3-319-42945-8_13 Petra Gemeinboeck and Rob Saunders. 2018. Human-Robot Kinesthetics: Mediating Kinesthetic Experience for Designing Affective Non-humanlike Social Robots. In Proceedings of the 27th IEEE International Symposium on Robot and Human Interactive Communication (RO-MAN). IEEE, Nanjing, China, 571ś576. https://doi.org/10.1109/ROMAN.2018.8525596 Petra Gemeinboeck and Rob Saunders. 2019. Exploring social co-presence through movement in human robot encounters. In Proceeding of the AISB 2019 Symposium on Movement that Shapes Behaviour. Falmouth, UK. http: //aisb2019.machinemovementlab.net/MTSB2019_Gemeinboeck_Saunders.pdf Petra Gemeinboeck and Rob Saunders. 2022. Moving beyond the mirror: relational and performative meaning making in humanśrobot communication. AI & Society 37 (2022), 549ś563. https://doi.org/10.1007/s00146-021-01212-1 59