Applying “Designerly Framing” to Understand Assisted Feeding as Social Aesthetic Bodily Experiences

2.1 Robotic Research on Meals

Robotic assistance for meals is understood to have a desirable and important role for several stakeholders in healthcare, such as care recipients and caregivers [12, 13, 83]. Some robotic solutions have been investigated especially to prepare meals [11] and others to support people to feed themselves [35, 91]. However, these efforts have not been oriented towards aesthetic and social qualities and a holistic perspective of the meal. Other research in HRI have investigated socially assistive robots, conducting both social and physical assistance, where feeding is a relevant task but not investigated in depth from the users perspective (see e.g., [28, 51, 95]). McColl [67] conducted a study of a social robot used in meals at a healthcare setting to prompt users through the steps of eating, and to monitor food consumption and provide social stimulation. The robot would track the food on the plate and comment on the eating, and the user would feed him or herself. Other HRI research has used cooking scenarios to investigate how adaptability affects the experience of a social robot [97]. Such scenarios can also be relevant for people with disabilities in their arms or hands, as they typically need assistants or other people to cook and serve their food.

There are several products on the market that already support people with limited ability in their arms and hands to feed themselves independently. These are typically referred to as meal-assist robots, eating aids, or robotic feeders. Examples of such eating aids include Neater Eater, My Spoon, Meal Buddy, Mealtime Partner Dining System (for an overview see [27]). These products are steered by a joystick (My Spoon), touch-push (Meal buddy), or by the user’s own movements (Neater Eater). The study reported in this article concerns Bestic; a commercially available eating aid that has a robotic arm with an attached spoon [59]. Bestic can be used with a joystick, STAR touch interface to steer, or touch button (for foot or hand) to initiate two different pre-programmed spoon behaviors supporting a more controlled mode or a semi-autonomous mode. According to the producer, it is designed to be safe, to keep the user in control, and to be small, quiet, and portable [59]. A study of Bestic has previously been conducted with two people with intellectual disabilities and four assistants [19]. This study found that a human assistant was still needed to be present during use, that the spoon got stuck in the mouth, and that the feeding was too slow and resulted in spilling. The study describes how this may affect self-esteem and associate meal situations with discomfort, and also sheds. This study also sheds light on a problematic power relationship between assistants and users with cognitive disabilities. Another study of Bestic conducted by Nickelsen [71] concluded that assistive feeding robotics are insufficient as intended advanced health technologies and argue for more engagement with involved groups such as care providers and users. A more recent study of this eating aid took an actor network theory (ANT) perspective and sorted out users, care providers, technology developers and governmental perspectives [72]. The study concluded that the developer’s overall goal for the eating aid was “usability”, whereas the users’ goals were more oriented towards “self reliance, control and agency”.

2.2 Human-Food Interaction (HFI)

In HCI, one strand of research has explored the relationship between people and technology in meals [14, 47]. Such research can provide relevant implications for robotic solutions, as it can widen our perspective of what meals are about, and how design solutions provide opportunities or research challenges related to meals in general. Grimes and Harper have argued that much research in HFI tends to involve a corrective perspective [36]: trying to help people cook, find recipes, choose healthier foods, or learn about nutrition. They have outlined a number of other areas to pursue, such as pleasure, nostalgia, relaxation, creativity, gifting, and family connectedness. We consider this to open more experience-oriented perspectives on meals. For example, Bell and Kaye describe how family connectedness, social activity, bonding, and enjoyment are intimately connected with how we seek to portray ourselves in everyday situations in relation to food [6]. Research on social aspects of food, eating, and cooking, such as how to support people in sharing recipes, is also related to social connectedness [94, 96]. Other investigations have focused on providing social awareness for people eating in different locations, such as CU-Later, which supports people eating in different time zones [98] and CoDine, which provides a video link and a multimodal sensory transfer to remotely create messages regarding food for people who are eating in different locations [104]. Because HCI as a research field often concerns communication technology, we find that HFI research issues related to meals tend to concern communication opportunities. Still, the HFI research field can provide useful perspectives on the meal for HRI, such as the importance of designing for meals beyond nutrition and corrective perspectives. We find that feeding experiences are more closely related to HRI, as feeding solutions concern physically moving food for a person, more or less autonomously, where social capabilities or social perspectives on the assistive solution is relevant to reflect on.

2.3 Framing (and Reframing) Design

In professional design practice, a core and tacit design skill is to undertake activities that support framing (and reframing) a design situation. Essentially, framing is about the process of addressing the right problem and making sure that the overall question framing the design process is sound, in order to avoid developing unsuccessful solutions. In general, an initial framing of a design challenge is never fully correct but should be changed [16, 52]. In HRI, just as in any design oriented research field, there is a risk that previous types of robotic solutions or stereotyped views of people and their needs, are hindering a holistic understanding of needs related to existing and future solutions. Design fixation is a cognitive fixation, a mental blocking, where previous experiences and existing designs minimize the design space and hinder creative solutions [16, 48]. Crilly [16] argues how designers need to be aware of design fixations, and we believe that this is also the case for HRI research. Previous experiences and knowledge of existing HRI solutions for meal assistance will affect robot researchers’ ability to come up with alternative perspectives of the meal as a design space.

The process of framing concerns getting insights about the design space, both what the problem is about and what possible solutions could be [26]. In this process, framing (and reframing) is about changing the initial assumptions about the problem and possible solutions in order to open up the design space. Kolko describes how “reframing is the act of purposefully shifting the normative frame” (p.5) [52]. Framing (and reframing) has been understood as a key characteristic of design thinking [17, 54, 75, 84]. Based on interviews with designers about framing design briefs with clients, Paton and Dorst describe how framing (and reframing) is related to the reflection-in-action [84] paradigm of design, rather than the paradigm of understanding design as a rational problem solving activity (i.e., [90]). They describe how projects that designers consider to be innovative, were essentially reframed in the design process. They also criticize Schön’s [84] initial perspective of framing, and how he presented framing only as conducting alternative reasoning of possible actions in architecture, while neglecting how frames and framing are situated, practice-based and value-laden. Our work builds on the situated perspective, opening up for reframing situated values (such as usability) that are strong in HRI to foreground other perspectives relevant for design aesthetic and social qualities. Successful reframing can be understood as the opposite of design fixation. Framing can be supported by specific design activities, for example, creating several alternative sketches of possible solutions can support avoiding design fixation [48]. However, without a process that has zoomed out from the normative understanding of the initial problem (or existing solutions) in HRI, it will still be difficult to take a successful alternative understanding of the design space and the innovation space is reduced. Dorst [25] describes a specific type of abductive thinking that designers do and how this relates to framing. His perspective is that a frame is related to a specific perception of the design problem, building on a key thesis as an underlying principle: “IF we look at the problem situation from this viewpoint, and adopt the working principle associated with that position, THEN we will create the value we are striving for.” [25] (p.525). This has also been described as designers’ ability to create generators [54] and generative metaphors for the situation at hand, enabling a shift of concerns from margin to center [102]. This model of thinking suggests how designerly framing applied in HRI can challenge normative assumptions of robotic solutions, opening up the design space from an alternative angle starting from values. Starting from aesthetics and social values in assisted meals then means to both create the working principle and the design objects (e.g., robotic service or other solutions) to fit within the values found in the user research and other forms of data gathering. It also means that the solution that matches the new framing may be a different type of robotic support, or not be a robot, even for something initially understood as a robotic problem.

Specific design methods can direct the focus in a design process. For example, in a Soma design process, methods such as bodymaps direct the focus on the researcher’s own bodily experience [44, 93] as a first person perspective. This can be difficult in situations with assisted feeding, since a disabled body is not the norm and simulation methods may lead to wrong conclusions. There is also the theory of embodied interaction, oriented towards understanding the contextual and situated nature of interaction, yet this theory in fact leaves out the bodily experience per se [93]. In user-centred design there is a risk that people are understood primarily as “users”, and that the process leads to an articulation of solutions developed with a preferred pattern of use and a limited perspective of users rather than opening for alternative interpretations and ambiguity [78]. The designerly framing in this article is not about applying a specific design method or theory to design an AT, but on the design skills needed to create the necessary materials and tools to result in an understanding of the situation at hand; here of assisted meals, and the essence of meals as a social aesthetic bodily experience. Here, the meal becomes the focal point. This is fundamentally different from aiming to understand the experience of a robotic eating aid from users’ perspectives, making the device the central focal point.

2.4 Social, Aesthetic, and Practical Use

The designerly approach applied in this article focuses on how to understand experience in a holistic fashion, rather than from a usability perspective of a specific device. This is related to human-centered design processes, as well as theories of acceptance of both robotic products and AT. When investigating acceptance, the field of HRI has a tradition of lab experiments and quantitative approaches, where models such as Technology Acceptance Model (TAM) and UTAUT are relevant [5, 39, 105]. However, TAM has been criticized as lacking social variables and focusing overly on ease of use, which can be problematic in holistically understanding experiences in a domestic setting [105]. Although UTAUT brings forward several relevant considerations related to experience, such as performance expectancy, effort expectancy, social influence, and facilitating conditions and attitude, it has been critiqued to lack explanatory power [39]. Other relevant research conducted by Broadbent et al. on experiences with robots has focused on one of these considerations, such as how expectations may affect experiences with robots [12]. In turn, our approach is qualitative, design-oriented, and phenomenologically-based, as we try to understand a social phenomenon and an activity from the viewpoint of the person being interviewed. This is related to an experience-centered perspective and pragmatism, to understand meaning-making and bodily rich experiences [38, 43, 92]. Our work is also related to use qualities and experiential qualities, describing experiences of use [1, 3, 30, 61, 73] which are widely adopted concepts to articulate aspects of experience in design oriented research. Qualities have been understood as a design language that explains some aspects of experiences when interacting with artifacts. For example, Arvola analyzed practical, social, aesthetic, structural, and ethical dimensions as quality dimensions that designers were working with when developing a web portal [3] and Löwgren explored different types aesthetic qualities as theoretical concepts (e.g., pliability, rhythm, and fluency) for HCI [61]. A framework on qualities developed by Forlizzi and Ford has distinguished between context, users, and products [30]. Another related framework for the meal, created by Desmet and Schifferstein, categorizes aesthetic experience into different levels such as aesthetics, meaning, and emotions [22], but do not focus on social experience.

In 1956, Greger Paulsson and Nils Paulsson developed a design philosophy in the book “The artifacts’ use and imprints” (“Tingens bruk och prägel”, in Swedish) [76]. One of the authors, Nils Bernhard Greger Paulsson, was an art historian with a strong functionalist ideology, who had a devoted mindset that any form of art is bound to specific sociological and psychological factors. One of his primary concerns was to make art and craft more accessible for the general society (not only for a privileged group of people) and connect artists and industry. In the book, Paulsson and Paulsson [76] reflect on three overall categories of use: social, aesthetic, and practical, to clarify experienced qualities of artifacts and the relation between these as a starting point for design. They argue that all things are made to be used but that there are different types of use [76] (p. 13). Overall, their design philosophy clarifies that “to assess things we have to see them in the situations in which they are used,” such as the following:

According to Paulsson and Paulsson [76], social use is about being with things, and concerns status, acceptance, and our relationship to others. This is not only connected to our body; what we signal, what we do and wear but also what we use. Things we use send signals, which affect our relation to others. We take the stance that social use of things can be understood to concern extrinsic desires. According to Krippendorf [53], extrinsic desires are related to acceptance, family, honor, idealism, independence, social contact, and social status. Csikszentmihalyi [18] describes how things we use can represent our religions, interests, and relationships. Social qualities relate how individuals develop both a concept of self and identity through social interaction with others as expressed in symbolic interaction by Blumer [9], which is also the case for AT, as described by Shinohara et al. [88]. Shinohara and Wobbrock [86, 87] show how social factors affect how disabled people use AT, and how technology presentation, use, and others’ perception influence people’s perception of self-confidence and self-consciousness. Hinohara and Wobbrock [87], even defined social accessibility of AT, arguing how social norms and social interactions affect self-concept and identity, and how context frames social order. Meals have been described as a performance, and as a staged act [74].

Paulsson and Paulsson referred to aesthetic use as something that is observed and appreciated in an object. The Greek word aesthetic is related to“perceive” and “sensitivity” and how something “appeals to the senses”. To Paulsson and Paulsson, aesthetic use occurs when our feelings overrule our perception of the object and contemplation occurs. This suggests that aesthetic use is more related to intrinsic (internal) rather than external motivations. Shusterman writes of how we can increase enjoyment and get a more rich experience if we attend to how we eat food, its texture, taste, fragrance, the movement from the plate to our mouth, how we swallow, and finally the overall sensation when the food enters our body [89]. Dewey, taking a related pragmatic perspective, describes aesthetic journeying as being undertaken for “the delight of moving about and seeing what we see” [24]. This is compared to non-aesthetic journeying, which concerns reaching the destination in the most efficient way, not reaching a significant experience. Aesthetic qualities have been stated to describe engaging and seductive means in interaction [61]. We take the perspective that aesthetic qualities are related to our senses, our perception, and experience of pleasure—as well as feelings of displeasure, discomfort, or disgust related to things. Moreover, we find that psycho-pleasures, such as discovery, knowledge, and cognitive actions that can satisfy the intellect [49], are also related to aesthetic qualities.

Practical use concerns how things are physically handled in an activity, such as hitting with a hammer or making coffee, according to Paulsson and Paulsson [76]. Practical use is associated and understood through practical qualities, such as efficiency, time, and durability representing practical qualities—measurable by testing. Practical perspectives have also been discussed by the philosophers Verbeek and Borgman [101]. Verbeek describes how Borgman takes a practical perspective of the meal when describing how microwave ovens simplify and reduce food-preparation time, and how such technologies change human eating and cooking habits.

The perspectives of social, aesthetic and practical use are still prominent in industrial and product design. Crilly [15] articulates how designers differentiate between social, practical, and aesthetic functions, and how “people use artifacts, while artifacts perform functions” (p. 2). This suggests that some functions may be better performed than others, but the research does not present empirical material that clarifies how this may affect users and their experience.

3.1 Participants

For an overview of participants and sessions see Table 1. All of the participants had long experience of human assistance at meals, due to impairments in their arms or hands. None had cognitive disabilities or problems with swallowing. The interviewees were suggested by the company producing the eating aid Bestic. The experience workshop included one participant, the participant’s personal assistant, an expert designer and a researcher. This participant had not used any eating aid before. All participants had given their informed consent to take part in the study. The national ethical authority was informed about the data collection. Pictures of people have been made anonymous, and approved by them for use. Below we will discuss the background of each participant in more detail.

3.2 Home Interviews

All home interview sessions were focused on the robotic eating aid Bestic. The company producing the eating aid did not commission the research but provided contact with the interviewees. The company had several initial questions that they were interested in, and a product technician participated in several activities to learn more about the user’s experience. This included the initial set-up and preparations of the interview study. Overall, the final interviews included an introduction followed by questions about its usability, the users experience with it and some social questions, which will be further explained below.

3.3 Experience Workshop

In the home interviews, the use of the eating aid was clearly in focus. A professional designer at an independent design consultancy firm (not working with the Bestic company), raised the idea of arranging a workshop to open up a holistic discussion of meal experiences and assisted meals. At this point in the process, the Bestic company was not involved in the set up of the research but still accepted the proposed designerly set up. This setup was done to avoid the eating aid dominating and coloring the overall discussion of meal experiences, which had become apparent in the interviews. Moreover, the workshop would then complement and extend the previous data collection. The attendees (see Figure 1 images (a), (b), (c)) were Elin, her assistant Anna, a professional product designer, and a researcher. Elin had experience of personal assistants for a few years but not of using Bestic. The lack of experience of a robotic eating aid was considered as an opportunity to get a greater understanding of the essential qualities of meals and assisted meals, and a possibility to avoid design fixation. It would have been possible to have one or two participants more with their assistants present at the workshop. Essentially, the limited number was due to challenges in recruiting people with disabilities. However, the set up was considered to be enough as the goal was to contrast and diversify the more device-oriented interviews from a qualitative perspective.

3.4 Analysis

For an overview of the empirical material collection, see Table 1. The empirical material consisted of seven audio-recorded and transcribed interviews made in the participants’ homes, and one recorded and transcribed session from the experience workshop. The transcribed material from interviews and the workshop was treated in the same way. They were sorted into three different categories concerning primarily social qualities, aesthetic qualities or practical qualities. In the process it was found how practical functions gave rise to certain bodily experiences which could be primarily aesthetic or social depending on the situation. Thus, the theoretical framework of social, aesthetic and practical use [76] guided the analysis.

4.1 Social Bodily Qualities

Social use becomes prominent when other people are present. Being social means to be aware of other people and oneself. When a person expresses what affects their presence among other people, these are prominent for social use of artifacts, objects, or devices. Stigma occurs when people feel that they might not be accepted, for example due to their appearance [34].

Eric expressed that being fed by someone else “is not so much fun”. He could handle the eating aid well but would also regularly eat directly from the plate with his lips (even if this strained his neck). He would use the robotic solution at home (with the family) but did not want to use it at a restaurant. He considered that being fed by his mother at a restaurant would attract less attention.

Heidi said that she would never use the eating aid in a restaurant. She felt like she looked like a baby after a meal with the eating aid: “It drops food so I feel like a two-year-old with food everywhere.. when I eat soup I need towels on the table and on myself”. When Bestic spilled, this not only led to discomfort for Oscar, but also for the other family members, according to his mother. If there was too much spilling Oscar’s brother would get nausea and also his mother would feel uncomfortable.

At the workshop, Elin reacted to a picture of a person using an assistant protection, which protects the person if the assistant spills. Elin did not use any assistant protection at the lunch workshop. She stressed that those things were not designed for attending social events and exemplified this by mentioning how this was out of the question at social events, such as parties.

Some meals are festive social events, where people want to show that they know the rules and rituals and fit in. The artifacts we use can support or hinder festive feelings. Elin could drink on her own using straws. However, she considered that straws were problematic because, to her, they would not fit in at all types of social events. The designer sketched a suggestion for bigger straws made of glass or other materials intended to fit with a champagne or another festive glass, and also to feel different against the lips; possibly being something that people without disabilities would like to try.

Being social means to pay attention to other people—which can be problematic in a feeding situation. Elin expressed that a downside of having an assistant at a meal with other people was that she had to direct her attention between several people—to be fed by the assistant and to socialize with others. Erik said, “It is awkward too when I’m being fed, because then I need to tell when I want more”. Thus, this would lead to disrupted conversations and attention, as well as overall annoyance.

Meals are a place where we can share thoughts and reflections with our loved ones. Both Elin and Carl had partners who could not support them with feeding due to their own disabilities. Elin would love to share a meal with her partner alone but her assistant needed to stay to feed her: “Now there are a lot of things that we don’t talk about while eating. Some things are secret”. Carl described his first experience with the eating aid: “It is directly more pleasant to feed yourself than to have an assistant”. His wife explained that “it’s much more private (with the eating aid); you can talk about whatever you want”. This suggests how privacy can be important in social meals. An example of desirable social qualities came up during the workshop with Elin. To Elin, eating crayfish had once been a special event at crayfish parties where people sit together and peel off the skins on the red crayfish, sing songs, and so on. To Elin, that was a shared and very social experience of smells, tastes, sounds, and visuals that enabled a rich meal experience.

4.2 Aesthetic Bodily Qualities

Oscar said, “I think it is more pleasant to eat without Bestic” and that “it is more unnatural than when I get help”. He felt more calm when someone in the family would feed him. Similar to Carl, who also had CP, the spoon attached to Bestic was not flexible enough for the spasticity, and this gave rise to the feeling of being stuck. Moreover, Oscar preferred to eat with a fork, which was not possible with Bestic.

Straws, which were used by all participants, gave rise to a wide range of undesirable aesthetic qualities, several described by Elin. Elin did not like the thin plastic straws from the supermarket; those were difficult to take big sips with and she would get both annoyed and tired when she used them. Elin described how she finally found some useful straws in a store having things for children’s parties. They were thicker and in stronger plastic so she could take bigger sips. Still, she pointed out that the straws were not so nice to use. They were difficult to clean, resulting in a situation where “sometimes when I drink a nice drink there is a taste of old milk in one sip. Awful!” Elin would sip coffee through a straw, but she did not find it as pleasurable as sipping directly from a cup as she used to before her illness. She found that the coffee was often too hot to sip. The designer sketched an alternative solution where a cup was placed in a high stand so it would be possible to tip the cup with the lips only. Elin and her assistant then used a small chinese bowl to physically test and imagine this type of solution.

Meals are part of traditions and culture, and certain dishes are associated with certain traditions or procedures where we eat food in a special way [68]. Elin very much enjoyed seafood. She now missed peeling and eating shrimps in front of the TV like she used to before she had the MS illness. Nowadays, she and her partner (who also needs assistance) had shrimps served ready on a sandwich in front of the TV. She described how this took away from the experience of eating shrimp, and how she did not really appreciate the meal of shrimp when served already peeled on a sandwich. To her, that could have been any other sandwich. Elin also longed to eat crab. Before she had MS, she enjoyed taking out the crab meat with a small fork. This illustrates how some aesthetic qualities that were part of her meal experiences before the illness (i.e., peeling the shrimps, feeling, and smelling them) was changed into a reduced meal experience even though the same food was served.

For someone with physical disabilities, physical strength and motivation may vary throughout the day. Elin’s strength would weaken during the day, but some mornings she could eat alone in bed, which for her was a moment of pleasure. This opened up the possibility of enjoying a meal without assistance, and avoiding the forced social situation that a human assistant can be.

4.3 Practical Functions

Several participants described how they tried to make it easier for the assistant or the eating aid to feed them successfully. Elin, for example, described how rice and minced meat are “good food to be fed”. She said that this was not because she especially liked the taste or the texture of rice or minced meat—but that it was easy food to feed for the assistant. The assistant could then take a little of everything for each bite. For a similar reason, Heidi would have an assistant mix her soups so they would function better with the eating aid and reduce spilling. This suggests how people adapt to the device but not necessarily matching their own food appreciation.

The eating aid required the user to either press one button to activate the spoon’s semi-autonomous movement, or several buttons to manually control the steering. Heidi expressed that “I would like to just look at the food and then it would come up to my mouth”. This suggests that the eating aid demanded physical and mental energy. It took Oscar about 40 minutes with the eating aid, instead of the 20 minutes it would take to finish a meal with an assistant. When Oscar could not get the food quickly enough, this would leave him hungry and eating cold food. This was also Oscar’s reason for not wanting to use the eating aid at a restaurant.“If I am really starving and have been out for several hours, I do not want to use it”. Moreover, he said that he would not use it at meals in school because lunch time was too short. This shows how limitations of practical functions can reduce enjoyable qualities of social and aesthetic use situations, reducing the overall pleasure of meals.

5.1 Addressing Social Use

Specific rituals, and the feelings of connection to others and competence play a major role in meals [65]. A design solution with excellent practical functions (such as an assistant protection) may be useless at a meal that is primarily a social event. This exemplifies a situation where a solution is rejected, entirely based on its social qualities. Sometimes solutions need to have social functions as the primary function, and the practical function is more secondary. An example of this is using a scarf as an assistant protection at a party, to create a successful association with adult clothing and fashion. This perspective is very much in line with how designing for social accessibility requires a holistic design stance, balancing users’ functional requirements with their socio-technical identity, as proposed by Shinohara [88], whose perspective is that something may be “socially usable” or not.

Because the appearance of AT and other products we use is intertwined with the user’s appearance and identity, we need a more holistic perspective in the design process. If a design creates embarrassment or stress and unwanted associations (see more examples in Table 2) it does not matter how useful it is, it may still not be used. Our findings indicate that several participants would not use the feeding robot at a restaurant as it attracted unwanted attention. This finding links to prior investigations of assistant technology use where people with disabilities abandoned the aid as they experienced it as not serving their function or too embarrassing to use [77]. Thus, even if the feeding robot would be perceived as a fairly good functioning tool to enhance independence, it is likely that people would still avoid using it. This suggests that social needs can prevail over both practical and aesthetic use.

Social use is not only about appearance but essentially concerns how well a design solution supports our relationships with people. Related research has shown how assisted feeding may be experienced as a situation that the person being fed cannot get away from [66] and that it involves a power relation with the risk of becoming power over, rather than power to someone [19]. Moreover, it can be a violation of privacy or integrity that a human assistant may cause when someone would like a social meal but not with his or her assistant. Robotic solutions are an opportunity to support independent meals in situations where human assistance is not desired. However, even if robotic solutions may enable more private meals and avoid forced social situations, there might be other privacy or independence challenges, such as if the solution is logging audio information or if it still requires a personal assistant to be present.

5.2 Addressing Aesthetic Use

Aesthetic qualities concern sensory input and perception, cognitive pleasure, and intrinsic motivation. Preferable people would get a rich experience of food, appreciating its texture, taste, fragrance, swallowing, and the overall sensation when the food enters the body [89]. Aesthetic qualities are difficult to measure, as they are subjective and highly individual preferences and experiences of sensory input and perception. It is however well known that the overall setting and environment strongly affect the meal experience and the possibility to get rich sensations from eating [46, 65].

Our preliminary results indicate the importance of considering different types of food in meals, and how they work with the proposed design solution. How different food, drinks and textures, sizes, forms and materials may feel different against the lips or to sip from, can be tested. For a robotic eating aid, not only the material and form but also how its movement is perceived, can be an important aesthetic quality.

5.3 Addressing Practical Functions

An extreme case of a meal where the practical function of being fed (nutrition) is central, is when receiving nutrition intravenously at a hospital. If we instead consider meals with AT, practical functions would rather concern measurable qualities of speed, efficiency, ease of cleaning, and the ability to handle different consistencies without spilling. We found that the participants in the study would sometimes try to simplify the situation for the human assistant or the eating aid by accepting mixed or minced food such as mince and rice that was easier to be fed. However, there is a risk that such adaptations may decrease the overall pleasure of the meal, and thus affect aesthetic qualities. When some practical functions are missing or not functioning well, this typically also leads to undesirable social or aesthetic qualities.

It can be difficult to divide attention and to use verbal communication with a human assistant when having a shared meal. This suggests that using conversation as an interaction type may be problematic to assist in feeding. Even if voice commands can be practical in other situations, this appears to be a less desirable function in social meals.

5.4 Applying “Designerly Framing” in HRI

Categorizing different use situations to primarily concern social, aesthetic, or practical use is an effort to investigate a pragmatic approach to address a more holistic perspective of design—beyond usability. Social use corresponds with Shinohara and Wobbrock’s definition of social accessibility, aiming to support assistant technology developers to be more aware of users’ visions of themselves [88]. However, whereas they only considered the social experience, we consider that either social or aesthetic bodily experiences can be primary at a given point. Whereas skilled professional designers can understand situations where social and aesthetic use is primary, this is less understood and applied in assisted robotics.

Below we extract implications for “designerly framing” in HRI design activities related to meals:

5.5 Methodological Considerations and Future Work

Designerly framing and reframing concerns to be reflective and avoid design fixation in HRI, also fixations on established user centered design methods. We have explored how a framework on social, aesthetic and practical qualities could support reframing assisted meal experiences from a holistic perspective in HRI. Our approach extends other frameworks on qualities of experience related to the use of technology or products (e.g., [22, 31]). For a skilled designer, it is clear that qualities such as aesthetics and social qualities (or other underlying values) may override practical functions and guide the design to other types of solutions than initially considered. However, along with industrial designers’ work, this is not necessarily understood by clients who may fixate on the existing functionality and appearance of their current product. Similarly, we find that there is a risk that researchers intending to use a specific robotic platform, are fixated on a specific functionality or focus on usability in a situation where aesthetic or social needs are primary.

This research is socially engaged, embracing the diversity of bodies and experiences as matters of concern, thus contributing to the critical inquiry needed in HRI (e.g., pointed out by [8, 55, 60, 85]). The rich description of experiences and how they are connected to social, aesthetic and practical qualities of assisted meal experiences supports understanding users’ underlying values in meals and what may lead to rejection and abandonment. As previously described, this research extends and deepens existing research on the abandoning of AT (e.g., [77, 79]) and abandonment of assistive social robots [12, 29, 95]. The designerly framing of the meal clarifies how it can be problematic to study only one robotic device on a long term basis to understand abandonment to get implications for assisted feeding devices. This is contrary to some early work on abandonment (c.f [77]). As designerly framing becomes more understood in the field, similar approaches could be applied in other situations beyond the meal, supporting researchers to explore ways to open a design space rather than fixating early on a specific robotic platform. HRI researchers may have specific values and assumptions on robotic assistance which may not match existing needs and perspectives among users (e.g., [12, 29, 57]). Similarly to Lee et al. [57], we argue for engaging in diverse perspectives of users in order to prevent rejection and abandonment of robotic solutions, and our work contributes to a situated and detailed account of related experiences.