This is a more considered response to Clark Quinn's paper, in which he argues for the use of computer games in education on the grounds that they engage learners and hence motivate them. So the issues are: what are the relationships between fun, engagement, and motivations for learning. This note only considers theoretical issues, which as noted at the end may have limited application to practice!
What is the relationship of enjoyment and learning? Some blend of novelty with familiarity seems required. Both extremes are too much to enjoy. (These relate to Malone's elements of challenge and curiosity, mentioned by Clark.)
Terry Mayes (a friend working at another university here in Glasgow) has a specific theory about enjoyment and learning. He thinks that accretion type learning, where you are adding new facts to existing schema, is experienced as enjoyable and effortless--which is why people read magazines for fun and why football fans effortlessly remember incredible amounts of facts in their area of enthusiasm. However, he argues that restructuring, i.e., the construction or reconstruction of new schema, is both effortful and aversive--which is why people do not read textbooks for fun and combine a liking for novelties that don't change their way of thinking with a strong preference for data that support their views. (Bigotry is fun for the bigot, it just doesn't fit the world very well, and I am most likely to indulge in it in areas like art appreciation and football that are far from my professional interests, and where I don't want to spend the effort to acquire more complex views.) Terry's view seems to imply that games might be effective for follow-up work adding detail in a topic but couldn't work on the most important bit of education--forming new concepts.
However, I don't think Terry is altogether right in this, plausible as his argument sounds to me at first. My experience of what television I find engaging depends on how alert versus how tired I am. When tired, I can't take in even "accretion" type facts, and will doze off over a wild life program (say) that at other times has my full attention and from which I retain at least some facts. An old Clint Eastwood film, however, can manage to keep me awake despite the fact (or because of it: How does anticipation work?) that I actually already know everything that happens. This is engagement at its purest in one way. However if I am wide awake then I want the opposite. I find thrillers too shallow to be interesting (engaging), but can be enthralled by an educational program. Apparently I am not alone in this. The British Open University broadcasts TV programs as part of its teaching, using a public broadcast channel at pretty unsocial hours. However they have discovered that a large majority of their viewers are not their intended audience of registered Open University students. Now these programs are NOT designed to be games, or to be fun in any conventional sense, yet they attract a considerable voluntary audience. This doesn't fit Terry's theory; and it doesn't really fit Clark's either. I think it nearer the truth to say that serious learning (restructuring concepts as opposed to picking up a few more facts that fit familiar patterns) does require considerable effort, but that whether this is attractive or aversive depends on how energetic we feel: in this it is no different from any sport such as hill walking. And that effortful learning (like hill walking) does attract people some of the time without any specially added fun and games to draw them in.
This also suggests that fun is at best only one of several quite different possible motivations for learning--so it is unnecessary to make learning fun, even if it may sometimes be useful. One of the responses to Clark's paper [Thomas, 23 Jan 97] argued against fun and in favor of showing learners how the material is useful to them, which rang a bell with me to some extent. But this seems to me an equally one-sided partial truth as considering only fun in games as a motivation for learning. A third motivation is just sheer, idiosyncratic "intrinsic" interest, as expressed in Papert's passage on "gear wheels" in the preface to his Mindstorms book. A more common example than gear wheels is astronomy, which fascinates a significant minority of people, but not because it applies in their lives or will ever have the slightest benefit or application to them. We probably also need this to explain the motives of learners on courses on literature (few of whom will get a job related to literature), and those who sign up to do degrees (as quite a number do at the Open University) after the end of their working lives.
I think this all interacts with the points Lloyd [Rieber, 24 Jan 97] made about games. Thus for instance, some adult learners who have paid money for a course may be averse to the label "game" while actually a pattern of interaction essentially like games may still be very effective for them. Conversely, exercises not consciously designed as games may come to seem like games when they have become sufficiently easy for a given learner. You have only to consider puzzle books, or how physicists are likely to leap at a calculation someone else begins or mentions: it is often a pleasure to exercise a skill once practice has reduced the perceived effort it demands. So I agree with Lloyd that, on the one hand, it is hard to define where a game begins and ends and, on the other, that there is a very pervasive pattern here whether we call it "game" or not.
In referring to Brenda Laurel's work on the notion of engagement, constructed on the experience of what engages theater audiences, Clark brings up a point whose truth I feel is very hard for us to accept in the field of learning and teaching. The most intense theatrical experiences, in which audiences are maximally engaged, and often described as "spellbound," do not depend upon realism in the sensory presentation, and they do not depend upon participation by the audience. Both of these go against our intuitions.
First, the minor point about sensory realism. Theater may use a lot of scenery and costume, but it need not and success is not correlated with that. Even if you consider mass audience trends in our society, sensory intensity and realism are not the most valued thing. The highest intensity technologies before virtual reality using surrounding screen and sound are still only seen in one or two special theaters. Ordinary cinema is much more immersive than TV, but TV attracts overwhelmingly bigger audiences. Realism and sensory immersion can be fun, but do not dominate other factors in contributing to fun. It is even more questionable, as Ian Hart [27 Jan 97] argued, whether they are associated with learning. University libraries are deliberately low-stimulation places (silence, soft lighting, etc.), and even so many prefer to work elsewhere as they find libraries too distracting for serious learning work. So neither realism nor intense sensory stimulation are necessary or sufficient for fun, engagement, or learning.
The second, more important point is about the importance of interaction, "doing," or being active as opposed to passive. At its best, theater is intensely engaging, yet the audience is "passive" on all external behavioral measures. Similarly my own most intense learning experiences are of this kind. Fun, engagement, and learning can all take place without external interaction, activity, or "doing." This is hard to take to heart. At least since Dewey, the best theorists have argued persuasively for a view that we often summarize as "learning by doing." Furthermore I know, as no doubt we all do, that if I fill a lecture session with my own monologue, then students struggle to remain attentive; while if I introduce some interactive activity, then the great surge in the level of their engagement is immediately apparent and I believe, that in those cases where measures of learning have been taken, more learning often occurs. So both experience and the best theorists seem to say that action is important to learning, yet reflection turns up crushing counter-examples.
The resolution is that the action (or "construction") that matters is mental, not physical. Engagement is essentially mental. The actions that are essential for learning are internal mental ones--the right kind of processing. However this means that there is no support whatsoever here for thinking that interaction, in the sense of visible learner actions, are likely to promote learning. If we interpret "learning by doing" literally in terms of physical action then it is false; while if we interpret it correctly in terms of mental action then it loses most of its value as a heuristic for designing teaching, and certainly its application as an argument for educational games. What is important for learning about participatory activities in class, as the Laurillard model tells us, is when they require students to "re-express" the material, which in turn requires mental processing of a different kind than mere reception. This shift to an alternate mental task on the same material is what is crucial.
The argument that Clark actually made, of course, was more subtle and is not undermined by this: he argued that games may engage learners mentally, and it is this engagement that can promote learning. There is a trap here which we may fall into, although Clark does not, about what is meant by "engagement." If I see someone wrapped up in what he calls a "twitch" game, where the player is highly interactive and processing motor reactions at full stretch, it is natural to call them "engaged." However this will act to prevent conceptual learning, not promote it. Clark explicitly remarks on this, but it remains as a problem for whether an analysis in terms of engagement is helpful for designers of educational games.
Fun is one, but only one, possible motivation for learning.
Learning has a variable relationship to enjoyment. Different types of learning require different degrees of effort, and our inclination for effort depends on how energetic we feel as well as how large other motivations for learning are.
Externally observable activity is neither necessary nor sufficient for either engagement or learning. Thus some of the best learning experiences occur with no external "doing" whatsoever, and conversely, lots of activity and/or lots of external stimulation may not promote learning.
Most of the points I have raised were already acknowledged piecemeal in Clark's paper, yet he argues for a model designers are converging on that ignores these disclaimers. That means that the model is, I believe, wholly wrong as a theory of what is necessary and sufficient for learning. However it may still be a useful heuristic model of the blunderbuss kind for instructional design. If you use all the elements (demonstrate applications, scaffolding, reflection, etc.), then, even though none are necessary, most learners in a given set will be favorably affected by one or more of the elements, so learning outcomes will be quite good. It would be possible in principle to investigate this by a study of a set of learners that took measures for each individual after each element, thus showing which elements mattered for which individuals.