To formulate a theory about a future society both very modern and not dominated by industry, it will be necessary to recognize natural scales and limits. We must come to admit that only within limits can machines take the place of slaves; beyond these limits they lead to a new kind of serfdom. Only within limits can education fit people into a man-made environment: beyond these limits lies the universal schoolhouse, hospital ward, or prison … Once these limits are recognized, it becomes possible to articulate the triadic relationship between persons, tools, and a new collectivity. Such a society, in which modern technologies serve politically interrelated individuals rather than managers, I will call “convivial.” (Illich, 1980)

Introduction

This chapter is an attempt to weave together some strands in my thinking about learning, technology and design. It takes the form of an argument, with illustrations, rather than a report of new empirical research or a description of a specific technological innovation. It will mainly be of interest to those who are concerned about the broad intellectual framing of the fields of practice and inquiry in which technology, learning and affect meet.

The short version of my argument is as follows.

  • We cannot understand affect in isolation from activity – feelings are entwined with personal projects, including the work we do to make sense of ourselves and our world, and to help others make sense of us.

  • Activity is both socially and physically situated. A person’s activity is shaped, in ways that can be subtle and powerful, by the situation in which it unfolds.

  • There is a valuable class of activities and human relationships, which – following Illich – I will label “convivial”. It is important to be able to understand the qualities of situations that afford conviviality and to understand what, if anything can be done to construct and preserve such situations, and strengthen such affordances.

  • In scoping the field of learning, technology and affect, it would be a great mistake to focus on taken-for-granted but obsolescent educational goals and processes. Optimising instruction for nineteenth-century outcomes is not the direction in which we should be heading.

Since most activity results in some learning, we can think of all situations as affording learning. While such a general case might be made, I will focus here on situations in which people are consciously and collaboratively trying to make sense of the world, to co-construct interpretations, to improve and apply ideas for some valued purpose. The challenge then is to understand the qualities of environments that afford convivial learning – at a minimum to satisfy our curiosity as learning scientists, but ideally to a level where we can design and help construct (parts of) such environments.

As with the older body of research and development (R&D) on intelligent tutoring systems, much of the work reported in this book focuses on modelling the changing state of an individual learner: the earlier emphasis being on cognitive states, represented in student models; the present work introducing an affective dimension. This focus on interactions between a computer system and an individual learner is a sensible R&D strategy, insofar as dealing with multiple learners might be thought of as a harder problem, but also one where solutions for individual learners might simply be aggregated to deal with multiple learners. Part of the point of the argument I am trying to develop in this chapter is that we should (also?) be considering other research strategies. The field can make some useful advances by producing clearer accounts of how tools for learning evolve, and how human practices and capacities evolve with them. Such an account cannot be constructed through methodological individualism (reducing all social phenomena to the actions of individuals); we need richer ways of understanding technology, social practice, collective enterprises and their relations with individual thought, feeling and action.

Moreover – thinking about a robust and properly ambitious research strategy for the field – we ought also to be careful to avoid narrow, conventional delimitations of what counts as learning. While my title, and some of the ideas developed in this chapter, draw inspiration from the writings of Ivan Illich – author of “Deschooling Society” – I do not want to underestimate the importance of helping young people achieve the goals set for them within prevailing educational systems. Action in the short term that persuades young people that such goals are delusions, conjured up to suit the needs of organised capitalism, typically injures the already marginalised more than the privileged. Thus the chapters in this book which show how a better understanding of affect can improve educational outcomes in established curriculum areas are – at least in the short term – of practical value. But if we are thinking longer term and more broadly about the constitution of this research field, it seems to me that we ought to be open to taking into account the full range of circumstances in which people learn. Otherwise, our enacted definition of the “science and ­technologies of learning” is too narrowly and partially circumscribed. A broader view also makes sense in terms of practical benefit, since much of what people learn (and value) occurs in the years after formal schooling has ended, and outside the schoolroom. An appropriate meta-level question we should be asking ourselves, from time to time, is whether the technologies we are helping build are primarily aimed at improving the efficiency with which learners achieve outmoded educational goals and, as Illich puts it, introducing “new kinds of serfdom” in the process.

To understand the relations between learning, affect and technology, so that we can make some more informed decisions about promising research directions for the field, it helps to have a sense of how technology is evolving, and to have some ethical principles about how we might best help it evolve. This is tricky work, and likely to involve some embarrassing failures. But the alternative implicit strategy of taking technology for granted simply will not do.

The next section of this chapter defines some of the key terms needed to frame the argument. After that, I explore some of the main issues involved in considering the evolution of convivial learning technologies, illustrating with some fragments of real-world practice.

Key Terms

Learning Environment

In much of the literature on technology-enhanced learning, a learning environment is something created by software and which sits within a computer. For example, one may read of interactive, intelligent and/or adaptive learning environments (Clancey & Soloway, 1990; de Corte, Linn, Mandl, & Verschaffel, 1992; Jones & Winne, 1992; Lajoie, 2000), virtual and online learning environments (Weller, 2007) or immersive learning environments (Peachey et al., 2010). In the broader literature of educational research, “learning environment” takes on a more diffuse set of meanings, and is often complicated by a conflation of objective/physical elements and subjective/mental elements. In such literature, important distinctions are sometimes missed between (say) the assessment demands inscribed in course documentation and students’ interpretations of the best strategies for meeting those demands (Entwistle, 1996; Goodyear, 1997).

I prefer to use the term “learning environment” to mean: a complex set of nested structures which provide the physical setting for the work of a community of learners. This physical setting can include all sorts of learning tools, spaces and resources – what we conventionally think of as hardware and software but also other knowledge objects produced through interactions between members of the learning community. I also take “physical” to include both the “material” and the “virtual,” while recognising that each interpenetrates the other.

On this view, a learning environment may be constituted from things as small as a mouse and as large as a campus; as small as a button, as large as the Web. For any one specific situated activity, not everything in the world can usefully be deemed part of the situating environment. Those things which readily “come to hand” are more significant; as may be those things which intrude on perception. But the notion of “nesting” is also important, not least because proximal and distal affordances (and constraints) can interfere with each other (Luckin, 2010; Resnick, 2010). Affect plays into this conception of learning environment in two main ways: first, any device/machine which is affect sensitive is a component of the learning environment; second, the feelings of people engaged in learning activity will be shaped, in part, by characteristics of the learning environment.

Architecture

I take architecture to be a set of practices which focus on understanding and arranging relationships between form and function. The etymological roots refer to a chief (ὰrci) builder (Τεκτονική). Conventionally, architecture is the art and science of designing buildings and other structures. By extension, it is now also used to refer to the masterminding of structures in complex computer systems (e.g. hardware and software architecture; systems architecture) including information systems (information architecture). Some inherent attractiveness of the term means that it is also being applied much more broadly (e.g. organisational architecture; policy architecture) and loosely (e.g. “experience architecture”). In my view, the last of these takes the term too far since it muddles experience (a relation between a person and a phenomenon) with the phenomenon itself. An architect might legitimately be said to design and/or create a set of things which are available to experience, but they cannot create the experiences themselves. In the same way, it is a mistake to think of a “learning architect.” One may design and/or create an environment conducive to learning, but one cannot create learning for others. Architecture, as a practice, is both analytic and creative. It seeks to understand relations between form and function but also to create new forms aligned with desired functions. Through analysis, it seeks to understand how a building (for example) may be decomposed into constituent parts and how the relations between these parts (the structure) relates to function – the way people use that building. (“Use” here needs to be interpreted broadly, so that it includes how people understand the building, how they feel about it, what they are readily able to do in it, etc.)

Affordance

The cluster of ideas around the term “affordance” provides some useful ways of understanding relations between form and function, between built space and its use, between a learning environment and learning activity. As is now well-documented, the term originates with the ecological psychologist James Gibson (e.g. Gibson, 1977) and was introduced into the human–computer interaction and product design communities – with a twist – by Don Norman (e.g. Norman, 1990; see also Norman, 1999; Turner, 2005). It penetrated into writing about the psychology of learning (e.g. Greeno, 1994) and is used both rigorously and loosely in the literature of educational technology (Conole & Dyke, 2004; Oliver, 2005). Recent writing critical of the concept has tended to dwell on slippage in usages, which have muddied thinking about whether affordances are objective features of the physical world, or ways in which the environment’s opportunities are perceived. Design theorists have introduced the idea of “conventional affordances,” to capture such things as the suggestions of use embodied in the appearance of items on a computer desktop. (“Desktop” being an example.) I do not want to dwell on these important arguments and reservations, but observe that it is useful to have a term whose connotations include the provision of thoughtful guidance, balancing structure with user autonomy, principled scaffolding for activity, etc. (Maier, Fadel, & Battisto, 2009). This allows us to tread a more credible and productive middle course, from both analytic and designerly perspectives, between the extremes of determinism and laissez faire.

Conviviality

I have appropriated this term from the writings of Ivan Illich (notably, Illich, 1980). Like Illich, I don’t use it to mean “tipsy jolliness” (though that affective state is pleasant enough from time to time). Rather, it is a way of denoting human, social and working relationships in which the interests of autonomously creative but politically interrelated individuals are served, rather than the interests of managers or capital. For Illich, conviviality is in stark contrast to highly industrialised modes of production, and their associated values and relationships.

A convivial society should be designed to allow all its members the most autonomous action by means of tools least controlled by others. People feel joy, as opposed to mere pleasure, to the extent that their activities are creative; while the growth of tools beyond a certain point increases regimentation, dependence, exploitation, and impotence.

Illich is careful to explain that he uses the term “tool” very broadly, to include

…simple hardware such as drills, pots, syringes, brooms, building elements, or motors, and not just large machines like cars or power stations [but also] productive institutions such as factories that produce tangible commodities like corn flakes or electric current, and productive systems for intangible commodities such as those which produce “education,” “health,” “knowledge,” or “decisions.” I use this term because it allows me to subsume into one category all rationally designed devices, be they artifacts or rules, codes or operators… School curricula or marriage laws are no less purposely shaped social devices than road networks.

Moreover, tools are not just means for accomplishing practical goals, they are intrinsic to social relationships:

An individual relates himself in action to his society through the use of tools that he actively masters, or by which he is passively acted upon. To the degree that he masters his tools, he can invest the world with his meaning; to the degree that he is mastered by his tools, the shape of the tool determines his own self-image. Convivial tools are those which give each person who uses them the greatest opportunity to enrich the environment with the fruits of his or her vision. Industrial tools deny this possibility to those who use them and they allow their designers to determine the meaning and expectations of others. Most tools today cannot be used in a convivial fashion. (Illich, 1980)

Combining these ideas, we may sketch a conceptual frame in which a learning environment provides tools that afford certain kinds of valued human activity. Understanding the nature and evolution of the architecture of such learning environments is scientifically valuable – there is scope for interesting and worthwhile investigative work. It also has practical value in helping us see how to engage with others in designing new, or helping improve existing, learning environments. Notions of improvement inevitably invoke questions of value – what do we mean by “better”? Acknowledging the importance of human feelings in the design space means that our notions of “better” have to move on from mere instructional efficiency.

Illustrating the Argument

It is time to use some analysis of concrete examples to illuminate these abstract ideas about relationships between tools, learning, conviviality and feelings. There is only space in this chapter to choose two such examples, and to provide a first-pass analysis, but I trust this will be sufficient to clarify the basic point.

Diabetes Daily

Diabetes daily (DD) is a large and complex Web site dedicated to the needs of those affected by diabetes, including people who have diabetes as well as their friends, family, carers, etc. Among many other tools and resources, it hosts a set of online discussion fora. The individual fora are clustered under a number of themes, such as “peer groups” and “daily living” (the former allows peer groups, such as teens with diabetes, or parents with diabetic children, to talk together). One can read the posts without being a registered member of the site. As of 12 August 2010, the DD community had over 34,000 members and over 400,000 posts.

The following posts come from the “complications: eyes” forum. The selected thread begins on 22 June 2010. It is headed: blurry vision and freaking out. (In the text below, T2 refers to Type 2 diabetes, BG(L) is blood glucose level, Endo is endocrinologist.)

Plinstedt: Hello- I am T2 and thinking I could keep my BG under control w/ diet, exercise and oral meds, I quit taking my insulin for 3 months. Had my a1c on 6/1 and it went from 6.1 to10! Needless to say, I am back on humalog and levemir. A few days after starting the insulin again, my vision became very blurry. Saw my ophthalmologist this morning and he said he saw no diabetic eye problems. He said my vision might return to normal or might not. Endo says wait at least a month before purchasing glasses. I used to use 1.25 reading glasses but now need 2.5 glasses. Please give me some encouragement that I will have my vision will return to normal.

This post elicited seven replies in the next 24 h, of which the following are a sample:

Nan-OH: We can’t tell you that it will happen the way you want. Many people will have improvement in their vision when their bgl comes back under control. It takes a bit of time. Listen to your Endo, and keep as much control of your bgl as you can. Diabetes and its control is an individual thing. Eating less carbs works for all of us, but what foods work for one person does not work for another. Don’t get too upset, yet. Relax and see what happens in the next few months.

NewdestinyX: Our bodies LOVE consistency. When we choose to make a major change like coming off the insulin -- or coming off another med.. and that change affects our Bg level balances enough to go from 6 to 10 in a short period of time -- the body reacts. Now -- you’re back on the wagon and things SHOULD readjust. My doc also finds now diabetic eye complications but yet I have this ‘shifting sand’ with my eyes too. Remember too that for many of us we’re in mid-life or older and eye changes, even quick ones are commonplace even amongst NON-diabetics.. So hang in there and look at the long haul. I take EVERY weirdness in my 48 year old body as if it were a potential complication of my Diabetes and keep tightening my control. I’m not a Dr Bernstein - all numbers under 100 - guru/zealot - but I still like to keep tighter control than the ADA standards -- hedging against risk.

CalgaryDiabetic: I think it will although 3 months of super high BG is a long time. Remember many so called type 2 are insulin deficient and no amount of exercise, low carb or metformin will do any good.

Premonitions of blindness are understandably emotion-charged and I think two points are worth making here. First, the other contributors to this thread are framing their responses in ways that intertwine emotional support with expertise. The emotional support has a tough strand to it (“we can’t tell you it will happen the way you want…”; “don’t get too upset, yet”; “hang in there and look at the long haul”). Moreover, many of the contributions mobilise specialist knowledge, expressed in specialist language. This is clearly a place in which “expert patients” gather and converse. Fox, Ward, and O’Rourke (2005) use this term to denote “those who can manage their own illnesses and conditions by developing knowledge relevant to maintaining health and countering illness”.

Then consider the environment(s) in which this building of shared understanding is taking place. Whether one wants to conceive of it as a single environment, or a set of intersecting environments – one per participant – the discourse within the forum only makes sense through shared reference to a multiplicity of tools, test results, drugs, symptoms and people outside the forum. The forum alone would not be sufficient to constitute an effective learning environment. It has to be understood as part of something more extensive: part of a Web or network of tools that collectively enable that which we are glimpsing. Emotions are influenced by this distributed environment, but also emotions influence how the environment is perceived – among other things, they influence the focus of attention (Most, Scholl, Clifford, & Simons, 2005), and thereby shape the functioning of affordances.

Lastly (for want of space), it is worth noting that the basic form of the forum has been around for some while now – threaded discussion has been in use for several decades. Although connection speeds and interface design have improved, with some consequent benefits for ease and efficiency of use, the DD fora could function perfectly well with the technology of the 1980s. How would new affect-sensitive technology make a difference? What would be the most appropriate level at which to model affect? Is it most sensible to try to detect the changing affective states of each participant in the forum, or even of each reader? Does it make sense to model the affective charge between individuals (Nan-OH and Plinstedt, or NewdestinX and Plinstedt)? Is there likely to be value in modelling the affective state of a system or community? (We speak loosely of the “mood of the meeting” or the “morale of the workforce”. If such constructs help with coming to understand complex situations, and deciding how to act, then presumably they should be “in scope” for R&D in affective computing).

Inception

I now want to shift to what may appear to be a trivial illustration – insofar as the focus of the sense-making activity I am about to describe may seem far-removed from the gritty realities of managing chronic disease. But this example also has some impressive characteristics, and I would like any accounts that we create of the relationships between learning, technology and affect to be able to deal with such examples.

In July 2010, Warner Brothers released the science fiction movie Inception. The film received critical acclaim and grossed over half a billion dollars (US) in its first month. A premise of the movie is that dreams and reality are interwoven, which makes it hard to summarise the plot. Suffice to say that it involves dreams within dreams within dreams (nested to four or perhaps five levels), characters interacting within and across levels, including characters who (separately) have the dream and “architect” the landscape within which a dream plays out. Understanding which “level” contains the current action, and whether the “top level” is reality or actually another, encompassing, dream, are key intellectual and perceptual challenges for the audience.

Within hours of its release, movie commentators had written pieces on the Web – including on movie blog sites that allow readers to provide feedback by adding comments. One example is the site RopesofSilicon, hosted by Brad Brevet.

Brad Brevet (2010) provided his own interpretation of the movie in a 1700 word blog posting, illustrated with stills from the movie, on sixteenth July (the first day on general release). He invited readers to comment on his analysis, especially of the key question of whether the final scenes are reality or another (level of) dream.

By 10 August 2010, there were 425 postings in response. The comment tool in this particular blog allows people to post a response to the original piece by Brevet or to add responses to other people’s responses. Thus the comments take the form of a shallow threaded discussion.

Almost all of the comments are contributions to a shared project of sense-making. (A few comments talk about how much the respondent enjoyed the film, but do not add to the interpretive work). Many comments refer to very brief and apparently inconsequential passages in the movie – such as whether the lead character was or was not wearing a wedding ring at a such and such a point in the action. Observations of this kind are deployed in arguments that seek to establish whether – at that time – we were witnessing a dream or reality and, if a dream, at what level the dream was nested.

An important element of the architecture here is that most respondents were making contributions based on (recent) memories of the film. (Some may have already acquired bootleg copies of the movie, but in many cases, respondents are referring to a complex event in which they participated some hours previously.) Crucially, there is very limited access to the key source (the movie) and so the responses have to use text to establish the context for any given piece of observation/evidence, relying on the accuracy of the readers’ memories of the film in so doing.

That said, what’s striking about the unfolding discourse is the extent of careful reasoning that draws on both (a) observation of details in the movie and (b) an emerging shared theory about how things might be, in the world of Inception. Take the following sequence as an example:

Mason (July 17th, 2010-11:47 am) I’m not going to go too far into the movie right now, because I’d like to see it at least once more before I come up with my official interpretation. However, I just have a little thought that I’d like to propose and I don’t know if it’s been brought up yet because I didn’t exactly want to read through all 97 comments.

It seems to me that throughout the whole movie, Christopher Nolan is not only trying to make us question what is reality and what is just a dream in context of the film, but also, it is to raise questions about our actual lives. Could it be possible that our real lives take place within our dreams, and this “reality” is actually just an escape? This seems to be the thought that Mal contemplates, resulting in her suicide. Nolan may be trying to get us to see the same thing as Mal. Therefore, the final frame becomes a bit more understandable. Whether or not the ending is a dream or reality is practically irrelevant to me. I do think it is a dream, yes. But the fact that the top wobbles in the last frame shows to me the parallel between the real world and the dream world. It shows that perhaps our perception of reality is actually quite shaky.

Idk (July 17th, 2010-1:50 pm) How come when Fischer woke up, he didn’t recognize that the people around him were all a dream? I think he was in on it. Or that Miles did really plan a double-inception.

After all, Dom did say that Miles was the one who taught him to manipulate minds. Miles also taught architecture. Maybe HE was the true master of inception? I do agree with the idea that Ellen Page’s character was specially chosen, somehow. And that Mol was right in killing herself to go BACK to reality. I also believe there is some connection between Fischer’s relationship with his father, and Miles and Dom’s relationship. Ultimately, I believe that Mol and Miles were trying to get him back home, and to release everything and wake up from his dream. THAT is what Mol’s totem was able to fall at the end. But again, I am not sure with my conclusion. There are MANY scenes left in the film that are needed to be pieced together. DOES ANYBODY CARE TO HELP ME OUT?

Joe (July 17th, 2010-4:47 pm) To idk: I am glad you have brought this up. After my post last night I began thinking about the moment when Fischer must have woken up and seen all of the people sitting on the plane around him. Surely he would have immediately recognized them and known they had performed an inception on him. Thus it could not have been real. Saito, Ariadne, and the others are not real. Ariadne’s name is what convinces me that she is not real (in Greek mythology Ariadne helps Theseus escape the labyrinth).

I agree that Mal was right and she does return to reality, while Dom remains behind. I think the “fourth-level” (actually fifth-level) confrontation with Mal is the moment when he finally comes to terms with his guilt. This is the real inception, planted by Ariadne, who has been constructed by Miles. Dom then returns to the first-level dream world which we see at the end of the movie.

The point is that Dom no longer has his guilt to interfere with him. He will ultimately realize that he is still in a dream, and finally be able to return to reality. Anyway, I have not had a chance to see it for a second time yet, I am interested to see how all of these theories will hold up after a second viewing.

Setting aside, for a moment, the fictive referent of this discussion, one cannot fail to be impressed by the careful writing, intricate reasoning and use of evidence. (Many teachers would be delighted to get this quality of work in their students’ online discussions.) The text is also imbued with motivation – the contributors are driven by a passion to make sense of the film’s puzzles, and they expertly invoke motivational explanations for the characters’ actions, and the intentions of the director. A sceptic may argue that this passion for collaborative sense-making is misdirected – why waste intellectual energy on a fictional puzzle? But I think we should learn from this example that the right combination of deep mystery, discussion tools and urgency can prove very fruitful.

The question of time plays differently in our two examples. In a few short weeks, the Inception discussions will be of historical interest. Somebody happening across the fora next year may be intrigued by the discussion, but it will be too late to contribute. The writers and readers will have moved on. Conversely, the DD fora gain value over time – every day sees an addition of new experience, insights and explanations that can be searched by subsequent visitors. Computing affect would be tricky here. It would, at least, need a theory of currency and decay, allowing reasonable inferences to be drawn about changes over time in relevance and affective charge. For example, what is the relationship, if any, between the emotions frozen into the fora and the feelings of the contributors today? If we imagine an emotionally intelligent agent sitting over the DD fora, what personal theories of affect would it need in order to compute whether to link newcomers with new issues (like Plinstedt) to people who had felt deeply about these issues 2 years ago?

Another contrast between the two fora stems from their respective relationships to face-to-face (F2F) discussions. It is easy to see the Inception forum as a close analogue of the kinds of discussion that a group of friends might have as they come out of the movie – and for several hours afterwards. Of course, it has different affordances. The forum has the virtue of crystallising ideas, evidence and inferences – they are frozen in time and can be read and re-read. It is much harder to remember and revisit ideas (etc.) that are articulated in a rapid F2F group conversation, where words disappear into the ether. In contrast, there really is no satisfactory F2F equivalent of the DD fora. Of course, there are self-help groups which meet F2F, but they cannot have the numbers, breadth of experience or record of past discussions that are core to DD.

Discussion

There are thousands of fora like DD. There are even several fora that have run in parallel with the Brevet-stimulated interpretations of Inception. This kind of colla­borative technology-mediated sense-making and knowledge sharing – often involving the joint creation of complex explanations – is not rare. Within these activities and spaces we can see expressions of what Illich would label “conviviality” – people are appropriating simple online tools and using them in joint creative enterprises which, one way or another, matter in their lives. (It is worth mentioning that Illich wrote speculatively about the value of such “learning webs” in Deschooling Society – 20 years in advance of the technology that now supports them).

How then should we be scoping the field of learning, technology and affect? I suggest that we need to consider at least the following:

  1. (a)

    We are unlikely to make a success of modelling feelings that matter to people if we ignore, or assume away, large parts of the environment in which they are thinking, learning, deciding, conversing, etc. A classic user: computer dyad may feel manageable, for the purposes of advancing the technology of affective computing. Improvements in the effectiveness of educational tools and tutors, brought about by better handling of affect, are also of practical value – when judged within the frames of current educational goals and practices. However, from time to time we should also pause and reflect on a properly ambitious scope for the field, and also ask ourselves whose framing of what is educationally valuable are we reinforcing? In my view, a systemic, holistic, ecological or architectural conception of learning environments is likely to be necessary to scientific progress.

  2. (b)

    R&D in affective technology is set in a dialectic between predicting technological trends and creating new technology. (We help shape the world, but not in conditions of our own making. Making best guesses about the pace and direction of technological change is necessary if we are to time our own research such that it intercepts with people’s needs and practices.) There is value in taking a moral position – it is partly in our hands to decide whether our tools are tools for conviviality or machines that perpetuate serfdom.