Josh Redstone
Institute of Cognitive Science, Carleton University, Canada
Social robots exploit human-like behaviors so that people might form emotional bonds with them. Ostensibly, this bonding is an empathic response from the person toward the robot. However, as Catrin Misselhorn points out, it’s conceptually problematic to say that people empathize with social robots, for the social robots of the present arguably don’t possess human-like emotions. To address this concern, she proposes that: (1) empathy with robots can be conceived of in terms of an interaction between perception and the imagination that she calls “imaginative perception” [henceforth: “IP”]; and (2), that a failure of IP might explain negative emotional responses toward robots, such as the “uncanny valley” phenomenon. In this paper I’ll analyze these elements of Misselhorn’s work, toward the end of making sense of empathy with social robots.
Firstly, by outlining some different constructs of empathy, viz.: affective empathy (e.g. sympathy, emotional contagion) and cognitive empathy or “theory of mind,” I’ll show that IP only concerns affective empathic phenomena. Since some researchers hold that intact cognitive empathic abilities are necessary to experience the full range of affective empathic phenomena, I’ll make the case that cognitive empathic phenomena also have an explanatory role to play where emotional responses toward social robots are concerned. Secondly, I’ll explore the role that the imagination plays in IP, with the aim of showing that IP is better conceived of in terms of misperception, rather than as something that is “imaginative.” I shall conclude with some remarks concerning how this alternative to IP is also helpful for explaining other emotional responses toward robots, such as the aforementioned uncanny valley phenomenon.
About the author: Josh Redstone
Julia Knifka
Department of Philosophy, Karlsruhe Instiute of Technology, Germany
Inspired by C. Breazeal’s interaction paradigms – readability, believability and understandability –, the focus of this contribution lies on the question of understandability in a human-robot-interaction. The ability to understand each other is the foundation of any social interaction. The way we perceive the other as an actor in relation to ourselves (1st person vs. 2nd person perspective), facilitates our understanding. The notion of understanding will be discussed and transferred onto the possibility of a human-robot-interaction. To this extent, the socio-phenomenological approach of Alfred Schütz on human-human interaction will shed light on the possibilities and problems of a human-robot interaction. The Schützian Phenomenology of Life-World concentrates on the social aspects of our day-to-day experience within the life-and common-sense world. The life-world as such, is not only the fundamental ontological category of human existence (Schütz) but is a socially constructed world. Within this life-world, we intersubjectively relate to each other. The foundation of this relationship is formed by observations, a human being makes concerning the actions of the other being. These observations are influenced by the “natural attitude” towards our fellowmen, which means humans project their living bodies onto everything they encounter and relate to it intentionally. This is not restricted to humans, but it can include animals, plants, artifacts and thus robots. The sociable robot, as envisioned by Breazeal, should achieve acceptance by understandability. However, an interaction between a human and a robot, which is based on mutual understanding, appears to be asymmetrical and para-social in its nature.
About the author: Julia Knifka
Maria Grazia Rossi, Manual Bastioni, Pierluigi Graziani,
MakeHuman Project, University of Urbino, University of Cagliari, Italy
In this paper we provide a philosophical interpretation of the uncanny effect, a term used by Masashiro Mori in to describe the unpleasant sensation, often a sensation of rejection, which an human subject feels when observe quasi-human subjects in which something does not seem to be in the right place. Specifically, we consider the uncanny effect as a specific case of recognition failure, that is the not willingness to interact with each other in an empathic way. By proposing an analogy among the human-human recognition failures and human-robot recognition failures, we argue that these two kinds of recognition failures can be explained by using the same underlying psychological mechanisms. In the first part we clarify what the uncanny effect has to do with the problem of recognition. In the second part we offer a philosophical explanation of the uncanny effect and support the hypothesis that the feeling of revulsion that defines the uncanny valley is linked to the functioning of the emotion of disgust. By examining the role of disgust we argue that recognition (the willingness to interact with each other in an empathic way) has a primarily perceptual nature: it is a recognition that is achieved by using automatic information processing, unconscious, obliged. Using the contemporary literature on disgust we argue that the reaction of repulsion typical of this emotion is relevant not only when our body is in danger, but also when the sanctity and purity of the social order are in risk. Our proposal is that the same sense of repulsion is helpful to explain the way in which we interact with humanoids that inhabit the UV. In this way, we insist on the idea that the disgust emotion has a key role in managing the interpersonal dimension between humans and between humans and humanoid agents.
About the authors: Maria Grazia Rossi, Manual Bastioni, Pierluigi Graziani
Ryuji Yamazaki
Department for Culture and Society, PENSOR group, Aarhus University; Hiroshi Ishiguro Lab, ATR, Japan
The purpose of this paper is to consider the potential degree of sociality of social robots with focus on the notion of empathy. Social robots are designed to interact with people as if they were human—that is, such robots are designed not only to appear human in terms of physical shape but also to exploit various biological mechanisms that trigger anthropomorphizing reactions in humans (Breazeal 2002). Humans often react with empathy for robots when they interact with them (Rosalia et al. 2005). A central question is whether such empathic reactions by humans are justifiable from a conceptual and ethical point of view. In this talk I will mainly address the conceptual strand of this question, leaving ethical considerations for the end; I will investigate whether empathy with robots is appropriate or misapplied relative to our current concept of empathy.
In a first step I clarify what is in the scope of empathy, i.e., what it is that we relate to when we have empathy with someone. Often empathy is described as the ability to perceive and share the feelings of others. Since robotics engineers have, in fact, constructed systems that seem capable of experiencing or feeling something like pleasure and pain, this account of empathy would seem to apply to robots. In a second step I discuss the question whether such ‘sharing’ of feelings requires that we reflectively attribute mental states to the object of empathy; exploring this path I show how the Theory-Theory and the Simulation Theory of the ascription of mental states would handle empathy. Alternatively, one can argue that the sharing of feelings is an emotional state that we attain interactively and rather immediately; this is the perspective of Interaction Theory and I argue that this view gives us the more plausible account of empathy. In a third step I will show that Interaction Theory can be fruitfully linked to classical phenomenological studies on feelings, moods, emotions, and intersubjectivity, i.e. bodily interlacement of selfhood and otherness.
About the author: Ryuji Yamazaki
