Negotiating Boundaries through Play

Siân Lindley Microsoft Research Cambridge

What are you like when you play? Are you light-hearted, or are do you take it seriously? Are you mischievous, a bit sneaky perhaps? I would guess that the answer to these questions is, “it depends” – you might show different sides of yourself when you play chess, football, a musical instrument or a computer game; you might be different again when you play make-believe or ‘tag’ with a child. Yet the word ‘play’ can be applied in all of these instances. So what do they have in common?

There are many different characterisations of play. For example, Jussi gives an overview of some of those used by psychologists earlier in this volume. What I want to do in this chapter is give a brief account of how play is characterised in Huizinga’s classic text, Homo Ludens (1950), and use this to unpack the playful ways in which myself and some of my colleagues used a prototype mobile phone application, Pickpocket. Central to Huizinga’s argument is the idea that play is characterised by implicit rules and occurs within a set of boundaries. He draws on metaphors such as the playground, playtime, and a magic circle to suggest that play is separate from the world. Consequently, when people are ‘just playing’, they can act differently; they can adopt different personas. In Huizinga’s view, play allows for a certain freedom within a set of boundaries. Salen and Zimmerman (2004), in their book on computer games design, have drawn on this idea to describe play as ‘free movement within a more rigid structure’.

In a recent paper (Lindley et al., 2010), we used this concept to understand why the design of a situated messaging device, Wayve (see the spread on p. 118), engendered playful behaviours. We suggested that constraints bound up with the device, and the social context in which it was placed, provided a sort of structure within which acts of expression could be freely undertaken. Wayves were plugged in at home, and the messages they displayed were expected to be seen by an audience of family and friends. Thus, Wayves were separate from more serious spheres, such as the workplace. Furthermore, their gradual interweaving with acts of playfulness meant that eventually, simply by dint of using them, one could indicate a playful mood. Concepts such as ‘the magic circle’, ‘the playground’ and ‘playtime’ seemed apt in interpreting how this device was used. It seemed to permit the ‘free movement within a more rigid structure’ that Salen and Zimmerman describe.

In this introduction, I want to look more closely at how a mobile phone application, Pickpocket, was used as a device for play. Like Wayve, Pickpocket is not explicitly a game, but a technology around which playfulness emerged. Unlike Wayve however, Pickpocket is mobile. Furthermore, it was given to a group of colleagues to use, and that use was undertaken almost entirely at work. I want to use this example to think more carefully about how technologies that are not so clearly disassociated from work and the ‘non-serious’ are appropriated in ways that are playful. In particular, I will attempt to explore how boundaries are negotiated through play, rather than providing an overt frame for it. Through doing so, I hope to highlight the need to underpin this type of negotiation when designing for play.

Before I say much about how Pickpocket was used, I should briefly describe it. Put simply, Pickpocket is a mobile phone application designed around the metaphor of pickpocketing material goods. Users can put image files in a ‘pocket’ on their mobile phone, and these pockets can be ‘snooped’ by their ‘buddies’. Effectively, this means that people who are friends through the application can look in each other’s pockets to see the images that are in there. Further, and as you may have already guessed, buddies can, if they are quick enough, take image files from each other. However, this cannot be done in secret. The act of snooping triggers an alert on the phone of the pocket being snooped. An attempted theft triggers a further warning message, following which the owner of the image has three seconds to tap a ‘Stop Thief’ button in order to hold onto their image. Otherwise, the file is deleted from their phone and copied to the pocket of the ‘thief’. Running totals for the number of snoops and thefts (both successful and prevented) are displayed on a ‘Stats’ board.

Evidently, there are game-like features here, but the scoreboard was the least viewed interface within the system, and users confirmed that they were not interested in scoring points for stealing or preventing thefts. Instead, the six co-workers, plus two remote colleagues, developed their own, playful, conventions through the application. It is worth saying up front that I was amongst the co-workers in question. So this is a personal reflection on the way that the technology was used.

However, the experience was interesting as well as fun. It provides a good illustration of how play is a means of negotiating boundaries, and how through this negotiation, players come to learn more about one another. In what follows, I will consider how myself and my colleagues made ourselves ‘part of the game’, how strategies regarding what was considered ‘good’ and ‘bad’ pickpocketing evolved, and how the negotiation that underpinned this framed the way that the game unfolded, and was framed by the wider context in which it developed.

“It’s almost kind of a honeypot …
you’re attracting the hit”

Let’s think again about the idea of a ‘magic circle’ of players. The participants in our deployment of Pickpocket certainly could be described in this way. They had dedicated Windows 6.5 mobile phones running the application, and they were listed together on the Buddies screen, along with a profile picture. Not only could they snoop in each other’s pockets and steal from each other; they could also send text messages and add images to each other’s pockets. In short, they were explicitly defined as a group, and had a unique means of connecting with each other.

However, being a ‘buddy’ did not make one a ‘player’. Through our interactions with the application, and with each other through it, we demonstrated that we wanted to be a part of the game. Signalling a willingness to participate was done through the curation of a pocket of images that were of some value to others, and that were therefore worth stealing and would need to be defended.

Various strategies were adopted to ensure that this was the case. Some of my colleagues placed images in their pockets that they thought would be interesting to others, with the expectation that they would try to take them. This was even explicitly advertised on occasion; for example, one colleague formed a collection of “wacky plants” with one of his co-workers in mind, and then placed one of them in this person’s pocket, as a way of “tantalizing” him into visiting his own. Another strategy that was sure to attract attention was to take an image that was known to be highly prized by someone else. This inevitably resulted in them trying to take it back, and provided the opportunity for defense. A final approach was to add a photo of a buddy to the pool. None of us enjoyed seeing ourselves in someone else’s pocket, so finding photos of ourselves typically led to repeated efforts to take those images and then delete them. As one of us noted, placing this kind of content in a pocket was “a honeypot … you’re attracting the hit”.

Having content that was prized by others was a way of signalling involvement in Pickpocket. Rather than attending to the Stats board, the game was played by an implicit value system that underpinned our use of the system. It was desirable to own content that was desirable to others, because by doing so you could attract their attention. In essence, it was a way of issuing a challenge.

“It is not fun stealing from grandmothers because what is the challenge there?”

Having a pocket of prized images was one way of demonstrating success at Pickpocketing. But like many games, there was more to Pickpocket than point scoring, even if this was done implicitly. Just as important was the way in which images were acquired, this being a demonstration of skill. ‘Good’ steals were underpinned by two factors, which emerged as the deployment progressed. They related to firstly, a sense of challenge and secondly, being able to witness the reaction to a successful steal.

Early in the deployment one of the most common strategies was to wait for someone to be away from their phone before picking their pocket. Those of us who worked in the same office space would watch for when someone was at the printer or coffee machine. Those of us who were in neighbouring offices even listened out for the sound of heavy typing. When we knew a Pickpocket phone had been left unattended, or we believed someone to be distracted, we would quickly try to pick their pocket. However, this rapidly became seen as too easy a strategy – it was considered much more skilful to take an image when that steal could technically have been stopped. Relatedly, stealing from people who were considered to be “particularly on the ball” was considered “a victory”. Thus approaches to playing Pickpocket evolved. We started to explicitly focus on taking images that could, in theory, be defended. This sentiment was summed up nicely by one of my colleagues, who observed, “Real pickpockets will say it is not fun stealing from grandmothers because what is the challenge there?”

Related to this was the possibility of seeing the reaction to an image being stolen. Snooping someone’s pocket whilst they were in view not only heightened a sense of challenge but also enabled telling exchanges of looks. The snooped-upon could communicate in a glance that ‘they knew’ you were ‘in’ their pocket. Being able to witness the reaction to a successful steal became an important part of the game for those of us who were on-site, and even when steals were performed across offices, co-workers often came together afterwards to discuss what had happened. More broadly, we discussed examples of particularly calculating steals, such as an occasion when one of my colleagues phoned another on their office phone, whilst surreptitiously picking their pocket.

“The angel grows horns”

In addition to this growing sense of what made a ‘good’ steal, there also seemed to be an inherent understanding of what made a ‘bad’ steal. While we took great pleasure in taking our colleagues’ favourite images, making them run to their offices to fetch their phones and antagonising them by keeping photos of their faces in the image pool, this is not to say that there were no boundaries in the way that Pickpocket was played. In contrast, various strategies were considered to be inappropriate. For example, it was considered ‘bad form’ to completely empty someone’s pocket when they were making a cup of coffee, or to interrupt them when they were obviously very busy. Participants who were particularly ineffective (or disinterested) pickpocketers even found themselves the recipients of gifts. Far from emptying their pockets, we would add images to them to try to keep them in the game.

This understanding of the boundary between acceptable and unacceptable pickpocketing developed as the deployment progressed. Devious tactics were adopted warily at first, and then blatantly. Picking someone’s pocket whilst in a meeting, or during a presentation, for example, was something that we were initially unsure about. Was this an appropriate thing to be doing, whilst we were supposed to be working? By the end of the trial, we were all doing this, often in a manner that was completely lacking in restraint. In one meeting, one of my colleagues pinned down another’s hands, whilst stealing an image from her, so that she was unable to defend her pocket. This was something of a turning point in the way the game was played, making meetings an appropriate venue for pickpocketing. The launching of sustained attacks, where players bombarded each other in an attempt to take an item, is another example of a particularly annoying tactic that was not engaged in early in the trial but that became increasingly popular. Like pickpocketing during meetings, it was the type of behaviour that, once deemed acceptable, was embraced.

We might then think of Pickpocket as a game with few set boundaries, but many implicit margins that were negotiated by the players themselves. It was a legitimate outlet for certain personality traits, such as competitiveness and sneakiness, to emerge, but having access to the application did not mean that players immediately demonstrated either of these. Instead, as one of my co-workers mentioned, “the angel grew horns”; we gradually tested boundaries, and built on our knowledge of what was deemed acceptable to different individuals within the group to decide which lines could be crossed.

“You need to figure out where someone’s level is”

This negotiation was one reason as to why Pickpocket lent itself most readily to co-present use. As already noted, the sense of being present through the game was supplemented by conversations about it, often occurring immediately after some steal had taken place, and instances of collocated play. Relatedly, effects of “social cascading” were felt to occur, in which multiple players would engage in a period of play at once. This would often begin with one player snooping another, and that player counter-snooping or snooping someone else. The spread throughout the group gave a sense of simultaneous involvement that could result in those of us on-site being drawn together as a group to discuss what had just happened. These opportunities for conversation were a way of understanding how people reacted to different strategies and the targeting of different images. Further, interactions through the application were a “nice excuse” for more general conversation, and a means for us to get to know each other differently. We came to understand who was competitive and mischievous, for example, traits that are not typically portrayed in our workplace.

So what of the remote players? All players had an online presence, and could show through aspects of game play that they were attending to one another. Remote players often took part in the cascades of snoops that rippled through the group as a whole, and they sometimes became aware of joint actions of co-present players, sending messages such as, “your feeble collaboration will come to no good”. However, the co-workers’ sense of getting to know one another did not fully extend to the remote players, especially if they didn’t already know them prior to the trial. This view was echoed by the off-site participants, one reporting that it felt a bit “bizarre” to pick the pocket of someone that he didn’t know well, and the other noting that an important part of using the application was being able “to figure out where someone’s level is”. Face-to-face “feedback” was vital for this. So in general, the game was played more and sustained better within the co-present group than with the remote players. The fact that the application was framed by a wider process in which on-site participants got to know each other better meant that their initial feelings of apprehension were gradually overcome and they could really play.

“A free activity standing quite consciously outside ‘ordinary’ life”

So far I have described an example of a game-like system that was playfully appropriated in ways that were not necessarily anticipated by its designers. Participation in the ‘game’, and the way that it was played, were aspects of use that were negotiated between players as they grew increasingly familiar with the application and with each other. Rules and boundaries emerged and gradually evolved through interaction between players, rather than being designed into the system.

If we return to the idea of play as “free movement within a more rigid structure” (Salen and Zimmerman, 2004), we can see that in many ways our use of Pickpocket can be characterised in this way. It occurred within precise limits of time and place and engaged a ‘magic circle’ of players. For those of us in the same office space at least, Pickpocket was used by a group of people made visible through their access to a dedicated device, whilst they were at work, and during office hours. So despite the differences in the technologies and the circumstances of the deployment, we could apply a similar analysis to Pickpocket to that we put forward when trying to understand why Wayve underpinned such playful behaviour (Lindley et al., 2010). Resonating with the notion of “free movement within a more rigid structure”, both systems have salient qualities that frame use. With Pickpocket, a restricted circle was built into the technology; with Wayve, location was fixed due to the form of the device. This structure, a direct consequence of design decisions, is complemented by a wider set of boundaries that relate to the context in which the systems are used. The fact that Wayve was situated meant that it was associated with a certain group of people; friends and family. The fact that Pickpocket was used with a certain group of people meant that it was used principally at work, during office hours. In both cases, metaphors such as playtime and the playground are fitting.

One of the reasons why this structure is so important is so that play can be recognised for what it is; it makes it obvious to everyone when someone is ‘playing’. Huizinga’s characterisation of play as a “free activity standing quite consciously outside ‘ordinary’ life” (Huizinga, 1950, p. 13) is crucial to the idea, illustrated here, that play is a safe haven where people are free to express themselves. Indeed, with both Wayve and Pickpocket, simply by dint of using the technology people communicated a wish to be playful, to not be taken too seriously. Yet it is also clear from our analysis that Pickpocket’s status as a ‘safe haven’ was negotiated rather than ascribed. What made a good steal was bound up with an understanding of the group, and the individuals within it. What was stolen, and how this was done, mattered. Furthermore, the evolution of strategies represented a testing, or pushing, of boundaries. This was especially important, given that the workplace in question was not typically a site where competitiveness or mischievousness was seen. These character traits were readily expressed through the deed of pickpocketing, and so the negotiation of boundaries was often done not through conversation, but through taking action and then witnessing the fall-out. This meant that although the game was mobile and could be played asynchronously, the testing of boundaries was largely predicated on synchronicity and togetherness. No wonder then, that Pickpocket was used almost exclusively at work, and that it was the on-site players who felt that they had grown to know each other better through the application. Through being together, they could get a sense of what was appropriate, funny or sneaky, and what was annoying or distracting.

The above demonstrates how Pickpocket can be positioned as a safe space for experimentation and expression, albeit one that needed to be negotiated. However, it would not be true to say that this play was separate from the rest of life. The boundaries that were settled amongst players were also framed by the context in which the game was played: the workplace. Having a Pickpocket phone on display served as a signal to other players, but one that was quickly noticed more generally. Further, the fact that the game was an increasingly potent undercurrent in meetings or at lunchtimes meant that it became ever more visible to non-participants. Such behaviour was described as “quite funny to watch” by those not involved in the deployment, and often attracted requests for ‘Pickpocket phones’. However, it was also described as “excluding” and “unbearably cliquey” by colleagues who had to endure silliness during meetings, which they had no means of partaking in. Additionally, the backdrop of work influenced what was considered appropriate pickpocketing. When, during a meeting, one of my colleagues felt that he had “overstepped the mark” by pinning down his colleague’s hands away from her phone, this was not only bound up with how she would interpret his action, but also how it would be perceived by others in the meeting. So while Pickpocket did, to an extent, provide an environment in which learning can be undertaken without fear of negative social consequence, the fact that not everyone was part of the ‘magic circle’ meant that the ‘safe haven’ had limits; it was bounded.

Design for play?

In the above I have presented an account of how a mobile application, Pickpocket, was playfully used by a set of colleagues, both collocated and remote. The play that I have described can be understood as bounded, with these boundaries being formed both through the design of the technology itself and negotiation amongst players regarding what was appropriate pickpocketing. Further, these boundaries emerged in a wider social context that also needs to be understood; in this case, the workplace. Play is not separate from the rest of life, but is framed by and interweaves with it.

This raises some interesting questions about how we can design to support playfulness. Myself and my colleagues (Lindley et al., 2010) have argued previously that technologies for play should allow activities to unfold in a separate sphere. Following our analysis of how Wayve was used, we suggested that providing simple but expressive tools, which support reciprocal use, and which provide a clear structure whilst enabling social conventions to develop, could underpin playfulness. The above analysis of Pickpocket use suggests that these recommendations should be extended to consider more carefully how boundaries are negotiated by players themselves. The fact that Pickpocket was principally played at work, drawing on characteristics such as competitiveness that were unusual in this context, meant that people were cautious in their use of Pickpocket in a way that we did not see in the field trial of Wayve. There was no clearly demarcated separate sphere; the overlap with work and the uncertainty surrounding what was appropriate meant that boundaries were in flux, being negotiated through the play itself.

This suggests that design for play should not follow the premise that play is always understood to be ‘safe’, nor that it can be made so simply by providing a clear structure for it. Even where there are obvious boundaries, such as a restricted circle of players, differentiated by ownership of dedicated gaming devices, the play that emerges is influenced by those who provide the backdrop for the game, and an uncertainty surrounding how others within the game will react. Play in this case was bound up with supporting the negotiation, pushing and even crossing of boundaries. It was this that made Pickpocket a way to get to know others better, this that supported the evolution of the game, and this that kept it interesting and fresh.

So how can we design to support this? Being able to act differently, and understand how others respond to this, was crucial to the play that emerged through Pickpocket. The fact that the application encompassed a set of rather blunt actions meant that it was important to know how interactions through it were perceived. Feedback was essential, but was not supported by the application, and so Pickpocket brought people together at work, and play with off-site participants gradually dropped off. Of course, these implications are related to the fact that Pickpocket was a game played with (or against) others. But they suggest that design for play requires the development of boundaries that are pliable, that can be pushed. This may be especially important when play is set against a more serious backdrop, or when it involves connecting with others in new ways. When people are ‘just playing’ they can act differently but this requires delicacy. A framework for this type of exploration, be it of oneself, of others, or of one’s environment, is a framework for play. This need not be done by providing a clear set of boundaries, but it is central that boundaries can be formed and recognised. We might say that with Pickpocket, rather than boundaries framing play, the negotiation of boundaries was play.
Lindley, S.E., Harper, R. and Sellen, A. (2010). Designing a technological playground: A field study of the emergence of play in household messaging. In Proceedings of the 2010 SIGCHI conference on Human Factors in computing systems (CHI 2010), 2351-2360.
Huizinga, J. Homo Ludens. A Study of the Play Element in Culture. Beacon Press, Boston, 1950.
Salen, K. and Zimmerman, E. Rules of Play. Game Design Fundamentals. MIT Press, Cambridge MA, 2004.

Pickpocket was developed by Paul Dunphy at Newcastle University’s Culture Lab and Stuart Taylor at Microsoft Research Cambridge. They, alongside Helena Mentis, Richard Harper, Tim Regan, Tom Simpson and Alex Soojung-Kim Pang, were instrumental in guiding the reflections noted here.

In the next section of this book, we consider a number of systems that aim to evoke play. This may, at first glance, seem trivial. People are playful. However, designing to elicit playfulness is not as simple as it seems. The subsequent pages present various technologies, some that were explicitly designed to support playfulness and some that were not. For some of the former, playfulness surfaced in unexpected ways. For example, the family archiving system on page 84 was turned into a toy box for virtual dinosaurs by one of its users. In other cases, playfulness was difficult to find, even when sought after. We see this in the technology designed to support playful web wandering, presented on page 100, which clashed with people’s expectations that search should be efficient and purposeful. In the spreads that follow, we will introduce a number of technologies, and invite you to consider how they might spur plei-plei.