“You give a little bit more love to animals than to robots”: primary pupils’ conceptions of ‘programming’ and programmable artefacts

In the late 1970s and early 1980s, informatics education, which included teaching about computers as well as learning programming, was established in many countries including Sweden (Heintz et al., 2017; Rolandsson, 2012, 2019; Rolandsson & Skogh, 2014) In the Swedish curriculum for compulsory schooling from 1980 (Lgr 80), CS was introduced in the subject mathematics in years 7–9. The focus in Lgr 80 was the computer, which was highlighted from three different perspectives: democracy, working life, and as a tool in teaching. The results of this effort showed that most of the teaching focused on skills training and simple computer applications (Riis, 1991).

As computers became more user-friendly and the amount of software increased, their use at work, at school and at home rose. Many teachers, however, did not have enough knowledge in the subject, not least in programming. This led to a shift in focus in the curriculum from 1994 (Lpo 94)—from learning about computers, to learning with computers (Heintz et al., 2017).

In the middle of the 1990s, technology developed further. Wireless networks became more common, which led among other things to CD-ROM-based materials being replaced by interactive and online services. Many teachers and students were given access to their own computers, which made it possible to use IT as a tool in lessons and thus in teaching (Olteanu, 2020).

In 2015, the Swedish government commissioned the National Agency for Education to draw up guidelines for an updated national strategy for the Swedish school system. One part of this work was to revise the curriculum from 2011 (Lgr11) for primary and upper secondary education. Digital competence is visible throughout the revised curriculum (2018a), as well as the curriculum that enters into force 2022, but is explicitly formulated in some specific subjects: technology, mathematics and social studies (Heintz et al., 2017; Swedish National Agency for Education, 2018a, 2022).

The writing on digital competence in the curriculum is based on the EU’s key competences and the Digitization Commission’s descriptions of digital competence. The four aspects of digital competence included in the Swedish curriculum are: to understand how digitalisation affects individuals and the society; to be able to use and understand digital tools and digital media; to have a critical and responsible approach; and to use digital tools to solve problems and put ideas into action (Swedish National Agency for Education, 2018a, 2022).

An important aspect regarding other countries’ curricula is that some use the term ‘coding’, while others use ‘programming’ as subject content. One possible interpretation is that they are often used synonymously. Manches and Plowman (2017), however, make a distinction between the two terms: coding refers “to the specific skills of inputting instructions using a particular language, such as Java or Scratch, whereas programming reflects the wider design and implementation process of using code to solve particular problems” (p. 193). Their definition of programming thus agrees partly with the Swedish National Agency for Education’s interpretation of the term. However, the term is broadly defined in the Swedish context and viewed in a wider perspective. It includes concepts of computational thinking as well as digital citizenship (Bocconi et al., 2018; Stigberg & Stigberg, 2020). In the Swedish curriculum, there is an explicit connection between programming and digital competence. For example, in a commentary document to the revised curriculum from the National Agency for Education (2018b), it is made clear that the focus is on programming—not on coding skills—as a pedagogical tool, problem-solving process and impact on society, since the aim is for pupils to develop a general understanding of programming as well as its effects on society (Bocconi et al., 2018; Heintz et al., 2017; Stigberg & Stigberg, 2020; Swedish National Agency for Education, 2018b).

The purpose of programming as subject content is to give pupils the opportunity to orientate themselves in an increasingly digitalised everyday life, as well as to prepare them for further studies and working life. This includes a general understanding of programming and rules that form the basis of digital systems, i.e., a basic understanding of digitalisation in society in which algorithms, code and software are included (Heintz et al., 2017; Mannila, 2017; Swedish National Agency for Education, 2022).

In the current syllabus of technology for compulsory schooling (2022), it is stated that pupils shall develop an understanding of how computers and networks work. For example, in years 1–3, pupils are supposed to learn about what computers are used for, their parts for input, output and storage, and common artefacts that are controlled by programming such as household appliances and smartphones. In years 4–6, core content involves learning some basic component parts of a computer, their functions (like processor and working memory), and how computers are controlled by computer programs and how they can be connected in networks. Regarding programming, it is stated that in years 1–3 pupils should be given opportunities to control objects with programming, and in years 4–6 they control their own constructions or other objects with programming (Swedish National Agency for Education, 2022).

Many studies examining children’s perceptions/conceptions or notions of what ‘programming’ is do not examine it as an isolated phenomenon, but rather one connected to the Internet, computers, robots, programmable toys, or other digital artefacts (see e.g. Cederqvist, 2020; Kuperman & Mioduser, 2012; Mertala, 2019. Additionally, and as mentioned above, the terms ‘programming’ and ‘coding’ are often used synonymously, even if a distinction is sometimes made (see e.g., Campbell & Walsh, 2017; Lye & Koh, 2014; Manches & Plowman, 2017).

Research about children’s conceptions of digital technology can be traced as far back as the 1960s, and the first studies exploring children’s understanding of the Internet are from the early 2000s (Mertala, 2019; Rücker & Pinkwart, 2016). However, in the wake of accelerating digitalisation, the ways we interact with digital tools like computers have changed dramatically, and have little in common with 20 years ago. Today’s children are surrounded by complex digital devices, which likely has an impact on how they form their conceptions of them (e.g., Chaudron, 2015).

Robertson et al. (2017) use the results from a study done in 1987, and make a comparison between children’s conceptions about computers 30 years ago and today to see how/if their ideas have changed as digital technology has advanced. In the study, 18 primary school children in Scotland (aged 5–8 years) were interviewed. Robertson et al. (2017) note that even if the pupils had good knowledge of how to use computers, they had difficulty in explaining how a computer works and can be programmed. Some pupils considered the behaviour of computers as being human—i.e. computers can ‘think’; while others were of the view that since computers do not have a brain, they cannot think. There were also pupils stating that computers can only follow orders. One conclusion drawn from the study is that it is important to give pupils opportunities to develop an understanding of the difference between how a human mind works and how a machine—in this case, a computer – works. This knowledge may contribute to being able to draw conclusions about strengths, potential risks and limitations of the technologies that surround them (Robertson et al. (2017).

Kuperman and Mioduser (2012) is another example of a study investigating children’s explanations of the behaviour of programmed artefacts—in this case robots with adaptive behaviour. When analysing whether kindergarten children (5 to 6-year-olds) used anthropomorphic or technological language when explaining how programmable artefacts work, the researchers found that the more complex the task, the more the children’s descriptions shifted from a psychological (using anthropomorphism) to a technological perspective. One possible explanation, according to Kuperman and Mioduser (2012), is that more complex tasks require analysis and interpretation based on the actual function of the programmed artefact. Consequently, the children’s focus had to shift from the observable to the underlying causes of the robot’s behaviour, which requires the use of analysis and interpretation skills focusing on the artefact’s functional components (Kuperman & Midouser, 2012).

Similarly to Kuperman and Mioduser (2012), Diethelm et al. (2017) emphasise the importance of both pupils and teachers having a language, i.e. terms, to guide them in activities with digital devices. In their study, they investigated how German pupils arranged, categorised and distinguished 23 different terms related to the digital world. Through observations, the researchers established that the pupils often avoided technical terms and instead used terms such as “thing” and “it”. A quarter of the eight to ten-year-old pupils did not draw any connection between robots and other digital artefacts, and some stated that robots have nothing to do with the Internet or computer science. Diethelm et al. (2017) conclude that pupils come to school bringing many experiences of the digital world which could be of use in the classroom. Thus, it is important that teachers are aware of their pupils’ preconceptions about digital artefacts like robots and computers. This knowledge could support teachers when designing their lessons (Diethelm et al., 2017).

Mertala (2019) explored 5 to 7-year-old Finnish children’s conceptions of computers, code, and the Internet. The empirical data consisted of drawings produced by 65 children, and interviews with the children. Most of the children had no idea how code and programming are related to computers. The meaning of the terms ‘programming’ and ‘coding’ were not familiar to them, and almost half of the group could not answer the question of what programming and/or coding are. If the terms ‘code’ or ‘coding’ were used, they were mainly associated with pin codes and passwords needed to log into a computer or to unlock touchscreen devices. The terms ‘program’ and ‘programming’ were associated with reading manuals or watching television programs. Only three children had some understanding that programming is about giving commands. According to Mertala (2019), a possible explanation for these three children having this understanding is that they had experience from playing with programmable toys or had played coding games. It was also observed that the children rarely came up with the idea of programming by themselves. Even if they had experience of programmable toys like Bee Bots, it was doubtful whether they could discover the connection between programming a toy and the principles of computer programming by themselves. To be able to do so, notes Mertala (2019), they needed guidance from adults. Much of what the children knew was based on what they had observed their parents doing. An important aspect that Mertala (2019) raises is the fact that teachers and parents often see children as “born-savvy technology user[s]” who “just [pick] it up” when it comes to learning about technology. However, the study’s findings challenge the myth of “digital natives”. Mertala (2019) concludes that since linguistic cues seem to play an important role in children’s conceptions of ‘code’ and ‘programming’, an investigation of children’s preconceptions of the terms is essential for effective teaching.

In line with Mertala (2019), Geldreich et al. (2019) claim that it is important for teachers to have knowledge about what experiences and beliefs their pupils bring to the classroom from their lives outside school, since children’s prior knowledge may both support and hinder learning and thus have an impact on what and how they learn. Their study is based on a qualitative content analysis of filmed teacher-student discussions involving 61 German pupils in years 3 and 4 (ages 8–11). The aim was to provide insights into primary pupils’ ideas and knowledge about programming. The results showed that children associated both objects and actions with the term ‘programming’. The researchers were surprised to find the subcategory ‘creating’ recurringly present in all discussions. However, the most frequently mentioned objects being programmed were consumer electronics, computers, and games, which Geldreich et al. (2019) note is not surprising, since these are devices with which the children come into contact in their daily lives. Other objects mentioned were cameras and movies. Some children explicitly mentioned animated films, while others seemed to refer to the cameras used during the discussions. In some of the group discussions, the children were given picture cards, and one depicted a road intersection. When asked what they associate ‘programming’ with, the children mentioned cars and traffic lights as examples of programmed artefacts. In this context, the children explained that “being programmed” meant “the car is programmed to keep a certain distance” (Geldreich et al., 2019).

Like Geldreich et al. (2019), Cederqvist (2020) states that, since many of today’s technological solutions are controlled by programming, pupils need to understand programming as a part of technological solutions, for example in mobile phones. Through semi-structured interviews, Cederqvist (2020) investigated how 11 to 12-year-old pupils understand programmed technological solutions. All pupils in the study had experience of using different programming materials like LEGO Mindstorm, Mircro:bit and LEGO WeDo. Prepared contexts were used to facilitate pupils in answering the questions, which consisted of Micro:bit constructions and programmed technological solutions familiar from daily life: a car key, a remote control and a digital thermometer. The study’s results indicate that visible parts (code in the Micro:bit constructions or buttons in everyday technology) were more clearly understood, while invisible parts like a flow of information were too abstract and harder to grasp. It was also difficult for the pupils to discern that there is a code involved in everyday technological solutions—even if they expressed that the objects are programmed. Cederqvist (2020) concludes that concrete programming materials like Micro:bit do not necessarily develop pupils’ understanding of programmed technological solutions. To support pupils’ understanding, it is important to visualize and pay attention to parts difficult to discern, like the logic of the code, how components are controlled by codes and the information flow that determines the code (Cederqvist, 2020).

To summarise, although digital technology is an important part of children’s lives, previous research implies that children have a limited understanding of what programming is and its connection the digital devices they encounter every day. This indicates that in order to create conditions for meaningful teaching in and about programming in technology education, more knowledge about younger students’ pre-understanding and experiences is needed.

To reach an understanding of what technology can be and how it can be described, we can turn to the philosophy of technology. Since programming is technology, to categorize the pupils’ descriptions of what ‘programming’ is, we have chosen Mitcham’s (1994) fourfold dimensions of technology as a theoretical framework. A reason for using Mitcham’s model is that it is broad and can include many kinds of technology. It is also characterised by being independent of time (Svenningsson, 2020).

Mitcham (1994) identifies four different ways of describing technology: as object, activity, knowledge, and volition. With technology as an object, Mitcham refers to technology as artefacts, i.e., physical, or abstract objects used in technological activities, or which are a result of technological activities. Based on Mitcham’s description, buildings, cars, bicycles, mobile phones, sewing machines and pencils can be examples of physical objects, while software, computer programmes, the Internet, musical works etc. can be examples of abstract objects. Technology as activity refers to technological activities such as creating, designing and manufacturing objects, but also the use of technological objects or processes. Hence, activity can refer to craftsmanship, inventing, constructing, operating, and maintaining technological artefacts. Mitcham’s dimension volition is based on the perception that technology is a part of human will or intentions, i.e., there is always a motive, or several, for a technological activity. Mitcham describes technology as knowledge as knowledge and skills humans use when creating, operating, describing, and explaining technological objects. In this study, we have chosen to also include pre-experiences in the aspect of technology as knowledge.

In this study, the fourfold dimensions described above are used to categorise the pupils’ descriptions of ‘programming’ as a concept.

A novice understanding of the meaning of ‘programming’ may involve understanding what a computational device may, or may not, ‘understand’ in relation to a human (e.g., Robertson et al., 2017). This can be formulated as a capability to discern differences between the human mind and a machine’s mind. Spektor-Precel & Mioduser (2015a, 2015b) describe this in terms of an understanding of Theory of Mind (ToM) and Theory of Artificial Mind (ToAM). ToM refers to a psychological perspective, an ability to describe other people’s mental states that explain their behaviors. It is thus about having awareness and understanding that other people can hold feelings, desires, beliefs, intentions, and emotions which are different from one’s own, and that their decisions can be based on other desires and beliefs than our own. In the context of the present study, this perspective can, for example, be expressed using anthropomorphic language when describing objects, their behaviour and functions. ToAM, on the other hand, refers to an understanding that machines behave according to what humans have programmed them to do, or according to direct operation.

In the analysis of the empirical data, we identify what in the pupils’ descriptions can be interpreted as an understanding of ToM and ToAM.

The study is based on semi-structured interviews with 16 children in year 1 (7-year-olds) in a Swedish primary school in a medium-sized Swedish city. The pupils were interviewed in pairs. The interviews were audio-recorded and later transcribed. Each interview lasted approximately 25 min.

The Swedish Research Council’s research ethics principles in the humanistic-social scientific field (2017) were followed. The participating teacher and the parents of the participating children were informed in a letter about the purpose and implementation of the study. It was clearly formulated that the participation was voluntary, and the participants had the right to withdraw at any time without giving reason. They were also informed as to how data were to be handled and reported, and that confidentiality would be preserved which included a pseudonymisation of their names. The pupils’ consent was handled by their parents, and the consent was given in writing. The data was stored and protected in accordance with the General Data Protection Regulation (GDPR).

The interviews were carried out in connection with the pupils being introduced to programming as teaching content in technology education. The teacher had planned the introductory lesson without the researchers’ involvement. The focus of the lesson was the difference between a robot and a human, as well as that programming being about providing clear and accurate instructions.

The semi-structured interviews were analysed using a qualitive thematic analysis, which is a method where written or verbal communication is analysed with a focus on similarities as well as differences. This is a non-linear, reflective, hermeneutic process, which involves a recurring moving back and forward between the entire data, and initially identified categories/codes. The interpretation process resulted in several identified themes. A theme can be described as capturing something important in relation to the research questions (see e.g. Braun & Clarke, 2006; Kiger & Varpio, 2020). Thus, the focus was not how many of the pupils expressed something specific based on the aim and research questions of the study (frequency), but rather the variation in their answers. The analysis process was carried out in eight steps, see Table 1.

The most frequent aspect of ‘programming’ the pupils highlighted during the interviews is that programming is about robots as physical objects.

Maria and Jana agree that robots have the strongest connection to programming. However, when the researcher asks them if there is anything further, based on robots, that they are thinking about, their associations are with electronics. One possible interpretation is that the pupils see batteries, cables and cords as parts of a robot, i.e. concrete and visible objects. The objects with which the pupils associate the term “robot” are based on a general meaning of the word. However, robots in educational settings (like Bee Bots) are frequently mentioned.

Anna and Martin also conclude that robots are the first thing they think of when they hear the word ‘programming’. Martin says, “to program a robot” and Anna refers to children’s TV programmes with robots. They continue to discuss movies they have seen (on television) with robots which look like humans. They both express that it is scary when it is not possible to determine whether it is a robot or a human being:

The researcher points out that this is an interesting aspect since “some people try to create robots that are human like. But you think the difference should be clear…?” Anna and Martin confirm their view and emphasise that it is important to be able to distinguish robots from humans. Anna concludes that she would not like to have a robot which is too human like, and Martin agrees.

The most common activity the pupils express is ‘to programme a robot’. Other activities mentioned were ‘to make a robot to understand what you want it to do, ‘to get help from a robot’, ‘to help robots’ and ‘to help the robot if it crashes and gets sad’. The descriptions of how humans need to ‘help’ robots, and the recurring use of the pronoun ‘he’ when talking about robots, are examples of how the pupils anthropomorphise the robots, i.e. use humanoid characterizations. The activities described also include a reverse perspective: that programming also means ‘helping’ or ‘teaching’ the robot something, or ‘comforting’ it if it crashes and becomes sad. The knowledge aspect was rarely present in the interviews included in this category—only through statements like ‘to teach the robot something’. Regarding the volition of programming robots, the pupils explain that the robots can do what humans find boring, but also that robots can do things more precisely/correctly, faster and in the right order, and can thus be used by humans to save time.

Most of the pupils express that they have pre-experiences from programming a Bee Bot in, preschool class (6-year-olds) or during school lessons in mathematics. Jana says that she has done “different things” in preschool, including “programming a robot” that is actually a friend pretending to be a robot, and that she had also “built a robot” Maria associates programming with giving a Bee Bot instructions: “I have also been doing programming… in [preschool school class] we had a robot for which you would press arrows and then ‘go’, and then it would go forward two or three steps …”. Maria has also “programmed a friend” at her “previous school.” Sonja focuses on the buttons on the Bee Bot when talking about programming:

Wilma also refers to previous experiences and activities from preschool class and notes that there are buttons on the Bee Bot and that you have to steer it. She uses the word “steer” instead of “programme”, and can give an exact and correct description of how to program the robot:

Two pupils, Maria and Alma, describe how they have programmed a robot when visiting a relative’s workplace:

Alma’s experience, on the other hand, was with programming a robot to make it perform tasks which resembled its regular function:

Except the references to robots in children’s culture (movies, TV programmes etc.) the pupils’ discussions can be described as taking their starting point from a technological perspective, since they concern how humans can use and program robots to perform tasks in order to solve human problems or to make work easier. The contexts vary: programming robots at home, working or studying at school.

Another frequent activity that pupils mentioned during the interviews, which we found surprising, was the idea that ‘programming’ is about building something—not only robots. All pupils, except Max, connected the word ‘programming’ with activity ‘building’. Wilma explained that she thinks of ‘building something’ and ‘building a car’ when she hears the word ‘programming’: and Asta thinks about building ‘some electrical appliances’. Mikael is not explicit, but says that programming is connected to creative and complicated processes:

One possible interpretation of Mikael’s description is that he sees ‘programming’ as connected to problem solving.

For some pupils, ‘programming’ is the equivalent of creating or building a robot. For example, Jana explains her pre-experience of programming in preschool as: “we tried a lot of things… we tried to build different robots from different materials”.

The connection to building or creating something, no matter what it is, is also made by Maria, who states that she often “does things with programming at home […] I made a dog out of toilet paper (rolls)[…]”. Maria reaffirms this view once again during the interview, when she associates programming with a profession as a construction worker, something Tom and Mikael also do.

In her explanation, Maria uses ‘planning to build’ as a metaphor for programming.

Why so many of the pupils express that they associate programming with ‘building’ something is not clear. However, in the descriptions that fall into this theme, programming becomes an activity that involves a process of thinking and creating, but not necessarily connected to robots or digital artefacts.

Learning for school as a volition for programming is something Tom and Mikael mention: it is important to learn programming because they will get grades in school when they get older. Tom states that “[If you do not learn] you may get an F in programming. You may get an F minus minus”. The same aspect is also in focus for Wilma and Asta, and they refer to future education:

What Asta, Wilma, Tom and Mikael say can be interpreted as seeing programming as a skill you only have use for in a school context. This indicates that the programming activities they have been involved in lack a clear context and clear explanations for why they should learn programming.

Another aspect of ‘learning for school’ that Jana highlights comes when she explains that, in a future profession as a teacher, you need to know programming: “[t]o be able to teach the children […] And to be able to teach their own children too… Yes, a child might start at the same school where his mother works [as a teacher]”.

However, as shown in the description of the theme of programming as being about building something, Maria, Tom and Mikael believe that programming is a skill that construction workers need to have in their profession. Similarly, Anna and Martin mention that knowledge in programming is something that they can benefit from in a future context outside of school.

What Martin says is an additional example of how programming is associated with pressing buttons. Another example of an outside school context and learning for a future profession is Maria’s explanation that you need to be able to programme if you are to become a construction worker.

The pupils’ answers categorised as ‘Learning for school and/or future profession’ suggest that they have rather vague perceptions of the usefulness of learning programming. They express they have experience of programming, both in and outside of school contexts. A common experience is making a Bee Bot to move—for example between different math tasks on the floor—or pressing different buttons on a Bee Bot to make it move in different directions. These descriptions can be interpreted as seeing programming as a tool for solving problems, but also being about giving instructions to a digital tool (like a Bee Bot) or performing tasks in a certain order. However, the example of a baker using an electric whisk shows how associations end up far from real programming situations. However, this can also be interpreted as associating programming with pressing buttons to make use of tools (objects) in different professions.

A fourth theme we identified in the interviews is about programming in daily life, and how programming is connected to human needs in the household. Some of the objects mentioned within this category are linked to robots—robotic lawn mowers and robotic vacuum cleaners—but also to digital artefacts like doors with codes, the Internet, mobile phones and apps. Other technological objects like doors lights, television, cars, batteries, power cords, cables electric wires and cars were also mentioned in the discussions about using programming in daily life.

A description of a kind of ‘servant robot’ was common and linked to activities like: to program robots to serve you when you are sick, help you to tie your shoes, iron clothes, sew, dress you, go and get you water, make breakfast, make the bed, change sheets in the bed, clean your room, do your maths homework or serve you when you want to listen to music.

According to Alma and Erica, it is possible to program robots to perform chores they find boring themselves, so they have time to do other things that appeal to them more, which can thus be interpreted as the volition of the activities. Another volition connected to human needs mentioned by the pupils was that a programmed robot can be a friend but also but also do some of the parents’ or other relatives’ chores so that they, or their siblings, can get more attention from the adults. Erica explains that: “[the robot] could cook so mum would not have to cook […] because she [also] must take care of the baby”—her sibling. Anna says that a robotic lawn mower could help her get more attention from her grandfather:

Thus, Anna also refers to a human need when she makes the connection between programming a robotic lawn mower and the volition to get more time with her grandfather. However, limitations connected to human needs were also mentioned by the pupils:

Martin and Andrea relate programming to their pre-experiences from using door codes (alarms):

Some of the pupils also associated programming with TV remote controls and robots on the roads—i.e. cars. Wilma, Tom, Andrea and Martin associate programming with lamps. For example, Martin states that “lamps are almost like a robot”. A possible explanation for the pupils associating programming with such technology may be, as we already have mentioned, that they see programming as equivalent to pressing buttons and giving instructions to a technical and electric device. Components connected to electric artefacts were also mentioned. Maria associates ‘programming’ with “batteries” and Jana thinks of “cables” and “cords”. Lenny also talks of “fixing” and connecting cables: “if a car breaks down, you can put cables like this”.

Digital devices were a kind of electric object the pupils mentioned. For example, Asta says that she thinks about: “programming … something … a [mobile] phone”. And later in the interview she adds: “Well … something with the Internet.”

The interviews confirm that robots and programming are a part of children’s activities today, and connected to play and leisure time, i.e. that the volition for programming connected to this category is entertainment. Almost all pupils mention toys when they talk about programming, and they make references to objects like Lego robots and different kinds of robot animals:

Some of the pupils also refer to a kind of play where a peer pretends to be a robot:

Another kind of toy frequently mentioned during the interviews is toys controlled by a remote control, such as dolls and radio-controlled cars. Programming is thus described as being about controlling and giving commands to an object without physically touching it.

It is also clear that programming is a natural part of children’s popular culture, for example through television programs, movies, children’s literature and computer games:

However, Tom is the only pupil who expresses having pre-experience from programming with a computer program at home:

During the interview, it becomes clear that that Tom’s father works with programming and that the pupil is frequently engaged in various types of programming activities at home together with his father. Tom also uses correct terms connected to programming activities—like making ‘loops’—and he connects programming to “making [computer] games and apps”.

Table 3 below presents examples of how children describe programming and the artefacts they associate with it. The pupils’ statements have been categorised here based on an analysis of whether they express the perspective ToM or ToAM.

We also found statements where the pupils use a mix of a ToM perspective and a ToAM perspective, i.e. shifting between the two explanations. For example, this occurred when they made a comparison between humans or animals and programmed objects: “[…] you give a little bit more love to animals than robots”; “[unlike a robot] the dog does not need to be told to walk […] it can walk by itself”, and “humans can do as they please, but robots cannot”.