The aim of this position paper is to propose a reflection on how to account for and investigate the many ways in which interaction between robots and humans requires co-ordination, negotiation or reformulation of meanings emerging in the ongoing interaction. Towards this aim, we argue, a perspective shift may be needed: to frame interactions as social engagements whose meaning can only be understood by the standpoint of those participating in it. We first present research on psychological benchmarks and design patterns for sociality in the HRI field. Then we provide arguments for including a new interactional element currently missing in the literature: participatory sense-making processes. As we will argue, such elements can be conceived and operationalised both as a relational benchmark as well as an interactional pattern, therefore proving useful for HRI research.
Over the past decades research into human interaction with robots (HRI) has flourished, mostly
thanks to both continuous advancements in AI technology and a germination of research practices and
reflections from disciplines such as philosophy, neuroscience and cognitive science. One of the most
uncontroversial findings emerged from HRI research is that people behave socially with robots in
several ways and to different degrees, depending on an array of robot-related aspects such as, among
others, human-likeness, conversational capacities, quality of a robot’s emotional display. The aim of
this position paper is to propose a reflection on how to account for and investigate the many ways in
which interaction between robots and humans requires co-ordination, negotiation or reformulation of
meanings emerging in the ongoing interaction. We argue that a perspective shift is needed and
propose to frame interactions as social engagements whose meanings are created and understood
online by the participants. How can our proposal be applied to the research on human-robot
interaction? First, by clarify how the situated moves of humans and robots influence and cause each
other within dynamical, sequentially coordinated processes. Second, by designing research paradigms
that are based in the practices of human sense-making in real social encounters. We suggest that both
strategies can open the way to new conversational, participatory or self-reflective, practices adopted
and explored within HRI. Modelling HRI on participatory sense-making would mean abandoning the
paradigm of individual cognition as the foundation of HRI and focusing on what individuals achieve
together, thus putting the emphasis on the relation and its interaction dynamics. Towards this aim, we
will first present research on psychological benchmarks and design patterns for sociality in the HRI
field. Then we will provide arguments for including a new interactional element currently missing in
HRI: participatory sense-making [
© 2020 Copyright for this paper by its authors.
Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR Workshop Proceedings (CEUR-WS.org)
The relationship between humans and social technologies has been the core of a
longestablished epistemological debate on the nature of autonomous robotic agents, but also and above all
on the future directions this relationship may take, with its social, cultural and economic
consequences. Since the early Macy Conferences on Cybernetics (1946-1953), the interest in
exploring the nature of human-robot interaction (HRI) has been growing with the purpose of “to
better understand human relations, and to design future collaborations between humans and
machines” [
At the beginning of the 1990, a major shift in the field was advanced by the Computers Are
Social Actors framework (CASA) [
Much of the growing CASA-oriented body of research has been guided by the underlying question: what are social robots, in the eyes of a human being; and which physical, psychological, communicative features contribute to make them so? In the attempt to address this question, Kahn and colleagues [15] have proposed a first distinction between ontological and psychological explanatory claims about HRI. The ontological claim focuses on what the humanoid robot actually is, while the psychological claim focuses on what functions people attribute to humanoid robotic agents, how and to what extent they do so under which circumstances. As it was originally posed by Hofstadter and Dennett [16], the ontological question mainly turned around the very concept of self-consciousness and its access by computers or artificially created agents. Since then, extensive debate in cognitive science and AI has focused on whether artificial agents, as we can conceive of them today in material and structure, are capable of consciousness. At the same time, the psychological aspects of HRI have received just as much attention, reflected in studies investigating how to make socially competent robots and what their competences would afford for in social encounters with humans. Noteworthy, in some cases drawing a clear line between what an artificial agent is, and what it is perceived to be from a human’s eye might be difficult. Partially because the ontological nature of any living system or agent determines its social affordances, and how those are perceived by those interacting with it. So, for instance, if we know that dogs are animals, and we know that animals cannot talk the way we do, then we should not expect our house dog to respond to us when we ask it: “do you want some food?” (ontological claim). And yet, this knowledge does not prevent us to keep interrogating the poor dog as if it had human-like communicative properties (psychological claim). The same may be equally true with artificial agents, with the exception that in this case their ontological nature is somehow established by, and therefore bound to, their creators (that is, by human beings). And despite all humans’ attempts to make them autonomous systems, most of the current artificial agents are designed to model, and appropriately respond to, humans’ psychological, social, cognitive, motor, affective capacities (as the CASA framework shows).
An initial step towards addressing and reflecting upon both these claims with sharper lens is to establish psychological benchmarks [15] of the things that are foundational for human interactions, and how to translate them into testable interactional scenarios. The scenarios were defined in terms of design patterns for sociality [17, 18].
In 2007, Kahn and colleagues outlined a non-exhaustive list of nine benchmarks intended as key aspects characterising fundamental categories of social and moral interactions between humans. They defined benchmarks as “categories of interaction that capture conceptually fundamental aspects of human life, specified abstractly enough so as to resist their identity as a mere psychological instrument (…) but capable of being translated into testable empirical propositions” [15] and identified as autonomy, imitation, intrinsic moral value, moral accountability, privacy, reciprocity, conventionality, creativity, and authenticity of relation. The overarching aim of pinning down human interactional benchmarks and transpose them into HRI equivalents was to advance the current understanding of what makes human-robot interactions successful considering the entirety of possible human relational experiences, from social to moral and thus began “to explicate what an interactional theory (as opposed to a use model) could look like for HRI” [15]2.
Drawing upon their psychological benchmarks framework, Kahn and his colleagues started to explored the micro-dynamics of HRI by identifying recurrent interactional design patterns occurring between children and a social robot (namely, Robovie) in several lab-based social situations and settings. A preliminary set of eight design patterns was systematically identified and coded: initial introduction, didactic communication, in motion together, personal interests and history, recovering from mistakes, reciprocal turn-taking in game context, physical intimacy, and claiming unfair treatment or wrongful harm. As this was a preliminary proposal, the authors specified that further combinations, hierarchical integrations and specifications between and within patterns would be just as likely as desirable Kahn’s work was precious in combining the theoretical tenets of human sociality with a data-driven approach to implement empirical categories derived from the situated observation of interactional patterns between humans and social robots.
Among the many aspects covered by the work of Kahn and his colleagues, we think there is one missing: the situated and dynamic coordination and negotiation of meanings, intentions, expectations and interpretations that human beings experience on a daily basis in their social contexts. To exemplify, consider transition between one interactive sequence and another – e.g., meeting at the restaurant moving from greeting ritual to the issue of deciding who sits where - requiring co-agents’ online coordination to make sense of the other’s action and enable the continuous unfolding of action sequences. We take this to be an instance of participatory-sense making, and in what follows we propose to expand the current landscape of design patterns of HRI to include dynamical and situated processes of shared sense-making as a central feature of meaningful engagements between interactants.
With the aim of extending the current landscape of interactional possibilities between humans and social robots, we will start by discussing two theoretical frameworks of social interaction and consider what contributions these frameworks may offer to the field of HRI. The first account is the second-person approach to engagement, which defines social interactions as the embodied, situated meaningful engagement between persons. Within this framework, interactors are active participants engaging in situated, open-ended, experience-dependent action dialogues, sustained by a reciprocal understanding and coordination of their embodied, mutual presence. This presence is construed as interactants address and acknowledge each other intersubjectively as a “you”. Engagement, in this approach, captures the qualitative aspect of social interactions [19] as it constitutes the very process by which interactant’s reciprocal, direct experience of understanding and knowing each other unfolds. It also reflects the way this experience is described in everyday language (e.g., “being in sync with someone”). “Experientially, engagement is the fluctuating feelings of connectedness with one another, including that of being in the flow of an interaction” [20].
The second account is enactive philosophy, according to which
“Social interaction is the regulated coupling between at least two autonomous agents, where the regulation is aimed at aspects of the coupling itself so that it constitutes an emergent autonomous organization in the domain of relational dynamics, without destroying in the process the autonomy of the agents involved” [21] 2Please refer to the original article by Kahn and colleagues [15, 17] for detailed descriptions of benchmarks and design patterns. Full reference to the articles can be found in the list of references.
Enactivism considers sociality in its broadest form as the meaningful engagement between
(embodied) subjects. Engagement occurs along a life-mind continuum based on six mutually
supporting, operational concepts: autonomy, sense-making, embodiment, emergence, experience, and
participatory sense-making [
Participatory sense-making
Under this view, ‘sense-making’ becomes ‘participatory’ insofar as social dynamics, such as dynamical coordination of actions and affects, take over individual sense-making [21]. Enactivism defines participatory sense-making as “the coordination of intentional activity in interaction, whereby individual sense-making processes are affected and new domains of social sensemaking can be generated that were not available to each individual on her own” [21]. Making sense with and of others occurs along a participation continuum which involves the experience of dynamically coordinating, negotiating or repairing of meanings in interaction by means of actions, gestures, bodily postures, verbal communication or other multimodal resources. As humans, we engage in participatory forms of sensemaking since birth, when we first start construing our social experience with others in situated, dynamically coordinated ways [20]. As we make sense of the world by moving around in and with it, and coordinate our movements with others in interaction, we literally participate in each other’s sense-making activities.
Engaging in participatory sense-making processes may occur within situations of interactional uncertainty, e.g., where an individual's course of actions unexpectedly is halted or changed by someone else’s move. To make the interaction moving forward, individuals need to mutually coordinate their actions or moves, even within a very short time frame. Individual sense-making is thus transformed into a participatory effort of making sense of the situation together. De Jaegher and Di Paolo offered an illustrative example of this, imagining two people walking in opposite directions along a narrow corridor.
“In order to get past each other, they must adopt complementary positions by shifting to the left or to the right. Sometimes the individuals happen to move into mirroring positions at the same time creating a symmetrical coordinated relation. Due to the spatial constraints of the situation, such symmetry favours an ensuing shift into another mirroring position (there are simply not so many more moves available). In this way, coordinated shifts in position sustain a property of the relational dynamics (that of symmetry) that all but compels the interactors to keep facing one another, thus remaining in interaction (despite, or rather thanks to, their efforts to escape from the situation)” [21].
Can the notions of engagement and participatory sense-making contribute to expanding our current knowledge and research practice in the field of HRI? We believe so, in (at least) two ways: first, by proposing a new theoretical perspective on how researchers conceive of robots’ and humans’ participation in social encounters, both from an ontological and a psychological perspective; second, by offering examples of potential situations in which humans and robots may need to negotiate or coordinate meanings while interacting. Both these avenues may prove useful for designing robots capable of participating in real-life situations in smoother and well-coordinated ways. We now present both these arguments. Endorsing a second-person account to interactions implies to see them as social experiences whose meaning can only be understood from the standpoint of those participating in it [19]. In this perspective, interactants’ perception of the co-agent (be it a robot or a human) would be informed by their affective experience of the other as a co-interactor; that is, the experience of being acknowledged, responded to and moved emotionally by the other [24, 25, 18]. In a way then, a second-person approach to HRI should be concerned with the question of what can this engagement bring to its participants? Clearly, here the ontological issue of what a robot is to the human’s eyes is settling. What definitional criteria should we use to evaluate robots as agents – autonomy, selfconsciousness, moral standards? How can we account for the many intersubjective possibilities of engagement with social robots along the “ontological continuum”?3
As it was presented above, in the human context participatory sense-making involves coordinating our sense-making activities to achieve social understanding through the joint generation and transformation of meanings in interaction. Endorsing participatory sense-making in HRI would implicate a deep shift in ontological as well as psychological perspective. First, in this view robots are seen as agents whose contribution to the interaction process is as valuable as that of the human. This might lead humans to adapt their social and interactional skills to fit the robot’s capacities, and not (as it is usually assumed) the other way around, opening up to new conversational, participatory or selfreflective possibilities and modalities. Importantly, it would mean thinking about design in terms of the joint contribution of robots and humans as participants of equal standing and acting together as a system. Second, and related to the previous point, a more constructivist view in HRI can lead to designing new interaction patterns, taken from real life practices of sense-making in social encounters: exploiting negotiations, misunderstandings, repairs, coordination, or misalignments.
To develop HRI in a more dynamic and fluent direction, perhaps letting pragmatic heuristics guide the interaction, certainly would be a daunting task. People in social robotics long have studied real life interaction for inspiration, but the problem is the technology: Roboticists have not been able to design (build) technology that can interact like humans, e.g., engaging in multimodal communication along several temporal scales during one and the same interaction episode, acting spontaneously while responding adequately to contextual variation. The most popular humanoid robots such as Nao and Pepper cannot be programmed to do this. Technology development is piecemeal. One pertinent question is what functions are basic to such open and responsive human behaviour that also can be programmed in existing robot platforms. Another question concerns learning: what algorithms to use and how to increase behavioural complexity. Modelling social interaction by extending models of individual cognition has not been a very successful strategy in advancing human on-line interaction with social robots. Modelling HRI on participatory sense-making would mean abandoning the paradigm of individual cognition as the foundation of HRI and moving from the individual agents to joint action and what individuals are producing together, viz., putting the emphasis on the relation and its interaction dynamics. In fact, in recent years we have seen what might be considered a breakthrough in multimodal design in HRI including new touch-based approaches to emotion communication and attempts at modelling the entire social setting of on-going interaction [26, 27, 28].
Additionally, there is a particular aspect of human participatory sense-making processes that we
think may be relevant for building better models of interactions with robots: the successful
coordination of sense-makers in interaction, be they humans or machines, depends on the specific
interactional resources afforded by and to its participants. This means that “although agents can have
different individual potentials for entering into an interaction, this potential is not fixed and can be
modulated by actual interactions.” [
But what does it take for humans and robots to coordinate and negotiate actions and meanings? Imagine the following scene: A robot accompanies an elderly lady to go shopping. After leaving the supermarket, the robot and the elderly lady have to cross a very trafficked road using the zebra crossing. The lady leans on the robot's arm, which slowly guides her towards the zebras until both are almost on the curb. A car is fast approaching and the old woman can only make small steps. The robot must choose what to do. In modelling a human-like decision-making process, the robot might lean its head forward in an attempt to gauge the speed of the car, or to understand if the driver has acknowledged their presence and is intending to stop. In the best scenario, the robot will try to use all available information (the car speed, the lady’s walking pace, the drivers’ face orientation) to take a decision about what do to. Clearly the driver might as well do the same: try to understand if the two intend to cross or if they will wait for the car to pass over. Most importantly, how this interactive sequence might end up largely depends on the coordination of considerations and actions (meanings) of all actors involved. That is, on a shared effort of making sense of how to go on. The question is then: what kind of technology needs to be implemented to afford social robots the capacity to directly 3 For participatory sense-making with materials see: Brinck 2018 [29]; for emotional engagement with materials see: Brinck & Reddy 2020 [30]. perceiving their own and other’s actions, movements, affective displays and to act accordingly and in coordination with the other? These issues are addressed, for instance, in research on autonomous cars and is key to technology development in the area. Raising the question in the context of HRI stimulates thinking in new directions and will lead to other answers. Some attempts towards simulating shared sense-making situations have been made. To illustrate, Satake et al. [7] developed a technique for pro-active approach and tested it in a series of empirical studies where the robot approached people who were passing through a shopping mall. Once the robot successfully approaches a person, it will identify the person’s reaction and provide a timely response by coordinating its body orientation. In comparison to previous simple approaching methods, Satake et al. found that their technique demonstrated a significant improvement in approaching performance, as reflected by the visitors’ reactions. Overall, the social situations just sketched can prove to be helpful standpoint to construe a new design pattern in HRI. Implementing and consolidating interactions where humans and robots coordinate or adjust their sense-making can be done to add a new design pattern embedding the ways in which robots and human beings behave in similar situations.
We think that Kahn and colleagues' distinction between psychological and ontological claims in the
field of HRI as essential to reach a deeper understanding of the interactional dynamics between
humans and robots but needs to be understood in a new context. We have suggested that a theoretical
shift in how we conceive of social interactions as oriented to the enactivist and second-person
approaches will have important implications for both claims, and perhaps show their irreducible
interconnection. Our proposal in this article is to extend the landscape of HRI using the concepts of
engagement and participatory sense-making as psychological benchmarks and interactive design
patterns. Specifically, including participatory sense-making both as a conceptual psychological
benchmark and a candidate interactional design pattern may clarify the ontological-psychological
continuum. One way to do it is to use micro-observational and ethnomethodological analytical
methods [
This work was funded by the Wallenberg AI, Autonomous Systems and Software ProgramHumanities and Society (WASP-HS). WASP-HS is an initiative by the Wallenberg Foundation for research in humanities and social science in AI and autonomous systems. [7] S. Satake, T. Kanda, D. F. Glas, M. Imai, H. Ishiguro, N. A robot that approaches pedestrians.
IEEE Transactions on Robotics, 29(2), 508-524 (2013). [8] S. Stange, S. Kopp. Effects of a social robot’s self-explanations on how humans understand and evaluate its behavior. In 15th ACM/IEEE International Conference on Human-Robot Interaction (HRI), 619-627 (2020). [9] H. R. Lee, S. Šabanović, E. Stolterman. How humanlike should a social robot be: a user-centered exploration. In 2016 AAAI Spring Symposium Series (2016, March). [10] C. Balkenius, B. Johansson. Almost Alive: Robots and Androids. Frontiers in Human Dynamics, 2 (2022). [11] M. Siegel, C. Breazeal, M. I. Norton. Persuasive robotics: The influence of robot gender on human behavior. In IEEE/RSJ International Conference on Intelligent Robots and Systems, 2563-2568 (2009). [12] P. H. Kahn, T. Kanda, H. Ishiguro, B. T. Gill, J. H. Ruckert, S. Shen, R. L. Severson. Do people hold a humanoid robot morally accountable for the harm it causes?. In Proceedings of the seventh annual ACM/IEEE international conference on Human-Robot Interaction pp. 33-40 (2017). [13] S. Finkelstein, E. Yarzebinski, C. Vaughn, A. Ogan, J. Cassell. The effects of culturally-congruent educational technologies on student achievement. Proceedings of Artificial Intelligence in Education (AIED), July 09-13, Memphis, TN (2013). [14] A. Moon, S. Rismani, H. F. M. Van der Loos. Ethics of Corporeal, Co-present Robots as Agents of Influence: a review. Current Robotics Reports 2 (2) (2021) 223-229. https://doi.org/ 10.1007/s43154-021-00053-6. [15] P. H. Kahn, H. Ishiguro, B. Friedman, T. Kanda, N. G. Freier, R. L. Severson, J. Miller. What is a Human? Toward psychological benchmarks in the field of human–robot interaction.
Interaction Studies, 8(3) (2007) 363-390. [16] D. R. Hofstadter, D. C. Dennett. The mind's I: Fantasies and reflections on self and soul. Basic
Books, New York, NY 1981. [17] P. H. Kahn., N. G. Freier, T. Kanda, H. Ishiguro, J. H. Ruckert, R. L. Severson, S. K. Kane.
Design patterns for sociality in human-robot interaction. In Proceedings of the 3rd ACM/ IEEE international conference on Human robot interaction, pp. 97-104 (2008, March). [18] A. Sauppé, B. Mutlu. Design patterns for exploring and prototyping human-robot interactions. In Proceedings of the SIGCHI Conference on Human Factors in Computing Systems pp. 1439-1448 (2014). [19] V. Reddy Why engagement? A Second Person Take on Social Cognition, in: A. Newen, L. De Bruin, S. Gallagher (Eds.), Oxford Handbook of 4E Cognition, Oxford University Press, Oxford, 2018, pp. 433-452. [20] V. Fantasia, H. De Jaegher, A. Fasulo. We can work it out: an enactive look at cooperation.
Frontiers in psychology, 5, 874 (2014). [21] H. De Jaegher, E. Di Paolo Making sense in participation: An enactive approach to social cognition. Emerging communication, 10 (2008) 33. [22] H. De Jaegher, E. Di Paolo, S. Gallagher. Can social interaction constitute social cognition?.
Trends in cognitive sciences, 14(10) (2010) 441-447 [23] H. De Jaegher. Loving and knowing: reflections for an engaged epistemology. Phenomenology and the Cognitive Sciences, 20(5), (2019) 847-870. [24] V. Reddy, P. Morris. Participants don’t need theories: Knowing minds in engagement. Theory &
Psychology, 14(5) (2004) 647-665. [25] I. Brinck, C. Balkenius Mutual Recognition in Human-Robot Interaction: A Deflationary
Account. Philosophy & Technology, 33 (2020) 53–70 [26] S. Bråten. On being moved: From mirror neurons to empathy. John Benjamins Publishing, 2007. [27] K. Altun, K. E. MacLean, Recognizing affect in human touch of a robot. Pattern Recognition
Letters 66 (2015) 31-40. [28] R. Andreasson, B. Alenljung, E. Billing. Affective touch in human–robot interaction: conveying emotion to the Nao robot. International Journal of Social Robotics 10 (2018) 473–491. [29] A. Hayashi, L. K. Rincon-Ardila, G. Venture. Improving HRI with Force Sensing. Machines, 10 (15) (2022). [30] I. Brinck. Empathy, engagement, entrainment: the interaction dynamics of aesthetic experience.
Cognitive Processing. 19(2), (2018) 201-213. [31] I. Brinck, V. Reddy. Dialogue in the making: emotional engagement with materials.
Phenomenology and Cognitive Science, 19, (2020) 23–45. [32] I. Brinck, C. Balkenius. Mutual recognition in human-robot interaction: A deflationary account.
Philosophy and Technology (2018) 1–18. [33] I. Brinck, C. Balkenius. Recognition in Human-Robot Interaction: The Gateway to Engagement.
In Joint IEEE 9th International Conference on Development and Learning and Epigenetic Robotics (ICDL-EpiRob) (2019) 31-36. doi: 10.1109/DEVLRN.2019.8850691.