In this work we present and test a RAG-based model called RAGGAE (i.e. RAG for the General Analysis of Explanans) tested in the context of Human Explanation of Robotic BehaviorS (HERBS). The RAGGAE model makes use of an ontology of explanations, enriching the knowledge of state of the art general purpose Large Language Models like Google Gemini 2.0 Flash, DeepSeek R1 and GPT-4o. The results show that the combination of a general LLM with a symbolic, and philosophically grounded, ontology can be a useful instrument to improve the investigation, identification and the analysis of the types of explanations that humans use to verbalize - and make sense of - the behavior of robotic agents.

Para responder adecuadamente a la cuestión referida en el título, es necesario aclarar antes de nada los diferentes significados que el término “ética” puede asumir dentro del debate moderno. Surgen de este término, si se mira con atención, diferentes significados según el nivel elegido y la perspectiva adoptada. Intentamos, seguidamente realizar una sintética (y necesariamente esquemática) enumeración de los aspectos y de las principales cuestiones en juego. Aquí tenemos las diferentes perspectivas a partir de las cuales es posible profundizar el concepto de ética:...

Research in education has shown that making errors is essential for learning. However, it is not always clear what constitutes an error, for example in a problem-solving task. Different teachers may have different ideas about what constitutes an error, and these different ideas may differently shape their educational choices. Theoretical literature identifies two typologies of programming errors: (1) errors in the code that cause the robot to behave unexpectedly, and (2) stylistic inaccuracies in the programming rules that are recognised as such regardless of the robot’s behaviour. This paper aims to analyse how teachers conceive of programming errors in educational robotics. Nineteen experienced educational robotics teachers were interviewed about their conception of errors and the most frequent errors pupils made in their classes. We asked the teachers to reflect on their experience and consider the role of errors in the learning process. The thematic analysis identified two main themes concerning errors in educational robotics, each with three sub-themes. The first theme, Definition of Error, includes behavioural errors, stylistic errors, and errors as learning tools: errors not only signal unexpected robot behaviour or stylistic issues in the code, but also provide opportunities for students to reflect, experiment, and engage in problem-solving. The second theme, “Locus” of the Error, distinguishes errors related to software, hardware, or interactions with the physical environment, highlighting how robot behaviour can be influenced by multiple factors. Overall, these findings suggest that understanding both the type and origin of errors can support more targeted teaching and transform errors into valuable learning opportunities.

An increasing number of studies are attempting to determine, through quantitative experimentation, whether people adopt an intentional stance towards robots. These studies mainly use questionnaires in which participants are asked to choose between mentalistic and non-mentalistic descriptions of robotic behaviours portrayed in pictures. While these methods are extremely interesting in their attempt to operationalise Dennett's theoretical constructs, they only capture one aspect of the intentional stance: the attribution of mental states to robots. They neglect the question of whether participants also attribute rationality to the system. Consequently, they are not well equipped to analyse how people form expectations about the behaviour of the robots they interact with, which is crucial for studying the dynamics of human-robot interaction. There is indeed no reason to deny that laypeople might occasionally attribute mental states to robots while believing that they can act irrationally or model the decision-making processes of the system in terms devoid of any reference to rationality. Building on these considerations, this article reflects on an emerging area of research in human-robot interaction from a philosophical perspective, identifying a potential limitation that could be overcome by referring to psychological literature on the attribution of rationality to humans.

Individual differences play a fundamental role in shaping people's attitudes towards robots. Among them, culture is one of the most deeply ingrained in individuals throughout their growth. Thus, given the crucial role of culture in shaping individuals' development, the present study investigated whether, and to what extent, individuals' cultural profile modulates their tendency to anthropomorphise robots, i.e., the likelihood of attributing human-like traits to them. Specifically, we focused on middle school students as one of the main targets of technological artefacts, including robots, in the near future. To this aim, we asked our sample (N = 85) to fill out a set of questionnaires about their cultural profiles beyond their nationality. Then, we adapted two well-established paradigms in experimental and social psychology to capture two dimensions of their anthropomorphism towards robots: (1) the Implicit Association Test (IAT), which allowed us to measure the strength of the association between the concept of "robot" and the attribute "anthropomorphic"; and (2) the Cyberball ball-tossing game, which measures participants' tendency to socially include robots. Results showed that individuals' cultural profile, operationalised as participants' scores at the cultural questionnaires, modulated their tendency to anthropomorphise robots, yet differently based on the kind of task (IAT vs. Cyberball). Interestingly, only some cultural values significantly affected participants' anthropomorphism towards robots, thus indicating that culture is a multi-faceted concept that deserves attention in future studies in the field of human-robot interaction.

This pilot study explores the tendency of people with autism spectrum disorder (ASD) to mentalise a non-humanoid robot. The study involved eleven children aged 7 to 10 years with ASD. A small non-humanoid robotic vehicle (CoderBot) was used. The children participated in a structured experiment involving true belief (TB) and false belief (FB) tasks. The tasks were designed to assess children’s tendency to mentalise the robot, using an adapted version of a classic Theory of Mind change-of-location paradigm, with verbal and non-verbal response modes. Results indicate that five participants used mentalistic terms referring to CoderBot in both TB and FB tasks and two participants in one of the tasks (FB and TB respectively), suggesting a mentalistic interpretation of the robot's behaviour. The terms employed fall into various categories including reference to the robot’s perceptual capacities, beliefs, cognitive capacities and states, and intentional agency. The results indicate FB attribution to CoderBot for six participants and TB attribution for four participants. The methodological insights of this study, including the integration of verbal and non-verbal responses and the helping paradigm, suggest promising methodological directions for future research.

Occasionally, in science, models are used to stimulate other systems rather than to perform surrogative reasoning. More specifically, in what is called surrogative stimulation, a model is used to stimulate a focal system in order to learn how it would respond to the system represented by the model. This article proposes a methodological reconstruction of the surrogative stimulation strategy and addresses the so-called model evaluation problem in relation to it. It is argued that in order to be adequate for surrogative stimulation, the model must provide stimuli that are similar to those provided by the target system, and a tentative definition of 'stimulus similarity' is offered. It is also argued that whether the model and the target system are similar in this sense is a question that depends not only on the context and interests of the modeller, but more crucially on facts about how the focal system works. Representative examples are taken from ethorobotics and social robotics, but the analysis made here is not intended to be applicable only to these areas of research. While much remains to be learned about this emerging use of models, the analysis undertaken here aims to offer a preliminary methodological reconstruction that may be useful for future studies.

It has often been argued that people can attribute mental states to robots without making any ontological commitments to the reality of those states. But what does it mean to ‘attribute’ a mental state to a robot, and ‘to make an ontological commitment’ to it? It will be argued that, on a plausible interpretation of these two notions, it is not clear how mental state attribution can occur without making any ontological commitment. Taking inspiration from the philosophical debate on scientific realism, a provisional taxonomy of folk-ontological stances towards robots will also be identified, corresponding to different ways of understanding robots’ minds. They include realism, non-realism, eliminativism, reductionism, fictionalism and agnosticism. Instrumentalism will also be discussed and presented as a folk-epistemological stance. In the last part of the article it will be argued that people’s folk-ontological stances towards robots and humans can influence their perception of the human-likeness of robots. The analysis carried out here can be read as promoting a study of people’s inner beliefs about the reality of robots’ mental states during ordinary human-robot interaction.

Culture plays a fundamental role in shaping how individuals perceive and accept technology, influencing their willingness to adopt it. Social robots, as interactive agents, are subject to cultural variations in acceptance, yet existing Human-Robot Interaction (HRI) literature predominantly equates culture with nationality, overlooking more nuanced frameworks. This paper advocates for a non-geographical approach to culture, grounded in Hofstede's model, specifically focusing on Uncertainty Avoidance (UA). The proposed model posits that UA exerts a moderating influence on the relationship between Perceived Control (PC) and social robot acceptance, such that the positive effect of PC is more pronounced for low-UA individuals and less so for high-UA individuals. This speculative analysis introduces novel perspectives for future empirical research, challenging conventional methodologies in the domain of cultural HRI studies.

Anthropomorphism is the tendency to attribute human-like characteristics to nonhuman agents, including robots. In the context of Human-Robot Interaction (HRI) research, it is relevant to understand what factors are at play in modulating individuals' anthropomorphism towards robots. This literature review addresses whether and how people's culture, which we identified as a potential factor of interest, affects their tendency to attribute anthropomorphic traits to robots. Moreover, we sought to determine whether the presence (or absence) of a relationship between culture and anthropomorphism towards robots varies as a function of i) the definition of both culture and anthropomorphism and ii) methodological factors, such as the measurements of culture and anthropomorphism adopted in the reviewed studies, as well as participants' and robot's characteristics. In most of the studies we reviewed, we observed a relationship between culture and anthropomorphism, i.e., individuals' cultural profile significantly affects how and how much they attribute anthropomorphic traits to robots. However, the directionality of the relationship is not consistent across studies. Furthermore, there is a small number of reviewed studies that showed a lack of relationship between culture and anthropomorphism towards robots. Although our findings do not vary as a function of the theoretical and methodological factors we identified, results are mixed, probably due to the large variability in those methods. The review contributes to extending current knowledge regarding the impact of individuals' culture on anthropomorphism towards robots, and provides suggestions towards a more controlled and rigorous investigation of the phenomenon.

The panel investigates the attribution of mental states and cognition to robots from a philosophical perspective, taking into account epistemological, ethical and technological (design) dimensions. These interconnected dimensions are explored through four talks. The first talk lays the groundwork by analyzing the different styles people may adopt to model the mind of robots. On these grounds, the second talk focuses on the role that emotion attribution to robots has in shaping our interactions with social robots. The third talk deals with robots’ decision-making capabilities in the context of social assistive robotics, with an eye to ethical implications. The fourth talk closes the panel, investigating how an enactive conception of intentionality impacts both our understanding of human-robot interaction and the design of robotic interfaces and architectures.

Prior research in human robot interaction (HRI) has largely focused on whether users adopt an intentional stance towards robots. We propose a methodology to investigate people’s explanations of robotic behavior that emphasizes finer-grained distinctions between explanations. The study also explores how users explanations change according to the modelling of a form of social competence in the robot by means of a computational cognitive architecture. Findings offer initial insight into how different explanatory strategies resonate with users.

Under what circumstances do we attribute a mind to AI systems? And, in this case, how do we think their mind works? Answering these questions is crucial to inform the design of safe and trustable AI, to inform research on the ethical, social and legal issues raised by the increasing presence of AI systems in everyday life and to investigate how they can be used as tools to study human and social cognition. This work proposes a philosophical reflection on the possible structure of people's mental models of AI systems. We distinguish between two possible styles of modeling that people may adopt in everyday contexts. Both involve the attribution of mental states and cognitive abilities to the AI system, even though they differ from one another in some relevant aspects. One modeling style is akin to folk psychology and relies on the attribution of beliefs, desires, and other propositional attitudes to the system. The other, which we will refer to as folk-cognitivist, is more akin to the account of the structure of the mind that characterizes classical cognitive science. These modeling styles correspond to different classes of mentalistic stances that people may adopt when they interact with AI systems in ordinary contexts.

Purpose The purpose of this paper is to explore the phenomenon of organizational unlearning with a focus on challenging path dependence and its implications on the organizational change field. By generating a taxonomy of unlearning definitions and examining the dimensions, actors and processes involved, the authors aim to offer a holistic understanding of organizational unlearning and its potential applications for organizations facing ambiguity and uncertainty. Design/methodology/approach This conceptual paper draws the literature on organizational unlearning to map existing definitions and categorize them into a comprehensive taxonomy to propose a model focused on the outcomes. Findings The findings highlight that organizational unlearning involves the three main organizational dimensions (micro: individuals; meso: groups; macro: organizations) and that the macro-organizational perspective represents the best fit for the concept. Furthermore, the authors’ argue that the most appropriate process for understanding the unlearning phenomenon is through challenge, as it allows the questioning of the present and facilitates critical reflection. Finally, applying organizational unlearning to path dependence concept, the authors’ posit that organizations can overcome negative transfer effects and build new awareness to reinterpret their dependencies in light of environmental changes. Originality/value This study contributes to the literature on organizational unlearning by providing a comprehensive taxonomy of definitions, clarifying the dimensions, constructs and processes involved. The integration of challenging path dependence with organizational unlearning offers a novel perspective on the potential for organizational change field. The paper’s findings have practical implications for organizations striving to survive and develop in uncertain environments.

Robots have been often used as models to theorise on the behaviour and cognition of living systems. Quite recently, a novel approach using robots for epistemic purposes has emerged. In what is called here interactive humanoid robotics, humanoid robots are not used as simulations of human beings (as in classical cybernetics and biorobotics), but as tools to provide predictable, manipulable, and controllable "social stimuli" to humans-i.e., as tools to stimulate, rather than simulate, living systems. By analysing how humans react to the stimuli delivered by humanoid robots in various conditions and having various characteristics, one obtains empirical evidence to theorise about the dynamics of humanhuman interaction. The structure and validity of this experimental approach has received little attention by philosophers of science so far. This article aims at paving the way for a methodological analysis of it, by (1) circumscribing the class of research questions that can be sensibly addressed using interactive humanoid robots, (2) outlining the structure of the experimental procedures adopted in the field, and (3) reflecting about the validity of the approach, with a particular focus on the impact of biomimicry and of the (folk-psychological) theories that humans, perhaps implicitly, formulate and use during interaction with humanoid robots.

Computer simulations are widely used for surrogative reasoning in scientific research. They also play a crucial role in engineering, more specifically in the design of new robotic systems, yet the nature of this role has been little discussed so far in the philosophy of technology literature. The main claim made in this article is that the notion of surrogative reasoning is central to understanding how computer simulations can serve the purpose of designing new robots. More specifically, it is argued that computer simulations can support two forms of surrogative reasoning, which are called model-oriented and prediction-oriented, whose inferential structure is reconstructed to some extent. And it is argued that, when computer simulations are used to design new robots, they are distinctively used in the model-oriented way. By unravelling the structure of the computer simulation-supported methods adopted in robotic design, this article may contribute to a finer-grained understanding of the epistemic processes involved in technological research.

In this paper, we ask one fairly simple question: to what extent can biorobotics be sensibly qualified as science? The answer clearly depends on what ‘science’ means and whether what is actually done in biorobotics corresponds to this meaning. To respond to this question, we will deploy the distinction between science and so-called technoscience, and isolate different kinds of objects of inquiry in biorobotics research. Capitalising on the distinction between ‘proximal’ and ‘distal’ biorobotic hypotheses, we will argue that technoscientific biorobotic studies address proximal hypotheses, whilst scientific biorobotic studies address distal hypotheses. As a result, we argue that bioroboticians can be both considered as scientists and technoscientists and that this is one of the main payoffs of biorobotics. Indeed, technoscientists play an extremely important role in 21st-century culture and in the current critical production of knowledge. Today’s world is increasingly technological, or rather, it is a bio-hybrid system in which the biological and the technological are mixed. Therefore, studying the behaviour of robotic systems and the phenomena of animal-robot interaction means analysing, understanding, and shaping our world. Indeed, in the conclusion of the paper, we broadly reflect on the philosophical and disciplinary payoff of seeing biorobotics as a science and/or technoscience for the increasingly bio-hybrid and technical world of the 21st century.

Robotics and Philosophy of Science What relationship holds between robotics and philosophy of science? Can robotics research contribute to research in philosophy of science? Conversely, can the results achieved by philosophers of science contribute to the progress of research in robotics? This article will deal with both questions based on the distinction between the so-called philosophy of general science and the philosophy of the particular sciences. It will draw relatively pessimistic conclusions about the possibility that robotics research shed light on questions pertaining to the philosophy of general science, and more optimistic conclusions about the potential contribution of robotics research to the philosophy of particular sciences. The normative aspect of philosophy of science is consistent with the possibility that this field of research, taking robotics as its object of study, provides guidelines and regulative methodological principles for engineers and roboticists. Through these considerations, this article attempts to explore a possible field of interaction between philosophy and scientific-technological research, with reference to a discipline that is assuming an increasingly predominant role in many aspects of daily life.