Serious games, conjugaison de jeux d'apprentissage et de jeux de la connaissance

La notion de jeu est l'une des premières composantes de la construction de modèles dans le cadre de la théorie des situations didactiques (TSD):

"Modéliser une situation d'enseignement consiste à produire un jeu spécifique du savoir visé, entre différents sous-systèmes : le système éducatif, le système élève, le milieu, etc." Mais, écrit Brousseau (1986/ in 1998 p.80), "Il ne s'agit pas de décrire précisément ces sous-systèmes autrement que par les relations qu'ils entretiennent dans le jeu."
- Au regard de la connaissance : "le jeu doit être tel que la connaissance apparaisse sous la forme choisie, comme la solution, ou le moyen d'établir la stratégie optimale [...]" (ibid. p.80)
- Au regard de l'activité d'enseignement :"le jeu doit permettre de représenter toutes les situations observées dans les classes (sinon les déroulements particuliers) même les moins satisfaisantes dès lors qu'elles parviennent à faire apprendre à des élèves une forme de savoir visé. Il doit pouvoir engendrer toutes les variantes, même les plus dégénérées. Elles seront obtenues par le choix des valeurs de certaines variables caractéristiques de ce jeu." (ibid. p.81)

Ainsi le jeu, source de motivations, peut par ses règles, ses représentations et ses stratégies,  accompagner l'apprenant vers la connaissance enjeu de l'apprentissage.

Le diaporama ci-dessous a servi de support à un exposé introductif à une discussion lors d'un séminaire de l'équipe MeTAH en juin 2010 sur le thème des jeux sérieux. Il met en relation la problématique du jeu au sens de la TSD et la problématique des jeux sérieux.

Design heuristics for authentic simulation-based learning games

Just published in IEEE Transactions on Learning Technologies (TLT) : "Design heuristics for authentic simulation-based learning games", a paper based on the PhD research of Celso Gonçalves supervised by Muriel Ney who created and led the project Laboratorium of epidemiology, to which I was associated.

Here is the abstract :

"Simulation games are games for learning based on a reference in the real world. We propose a model for authenticity in this context as a result of a compromise among learning, playing and realism. In the health game used to apply this model, students interact with characters in the game through phone messages, mail messages, SMS and video. Perceived authenticity is measured after the game in 196 phone interviews
that yield quantitative and qualitative results. We show evidence of relationships between attributes of the game environment and perceived authenticity. This yields a list of parameters that can be adjusted to favour authenticity. We also study three situations of interaction and show when and why they are perceived as authentic, or not. These results lead to recommendations for the design of simulation games that can be perceived as authentic."

[Get the paper]

#ocTEL MOOC (week 3 A33) Learning forward, designing backward

The third activity for this week 3 on Designing active learning is to design an activity and to review a learning activity. I didn't design one specifically for this MOOC, but I am happy to share one which I designed for a Doctoral school a few years ago, it was about the design of learning game, starting by inviting students to play a game...

About learning games and the design of learning spaces

The idea is simple: invite students to play a game first alone against the teacher who manages to sometimes loose, sometime win. This the time to acquire the rules. Then the students play against each other, first alone, then in team with a spokesperson who will play the strategy of the team. There are two levels of debriefing, the first one specific to the game as such, the second to understand the structure and the function of the game as a learning situation. Eventually, students are invited to analyse a simulation game in epidemiology. The sequence closes with a more theoretical analysis of the role of games in learning.

The lesson learned from this exercise is that while learning goes forward from action to articulated knowledge, the design of a learning situation must go backward from the targeted learning outcome back to the optimal situation to engage learner in the process. This situation could be a game but not necessarily, it must essentially be a situation which allows learners to mobilise what they know, whatever it is, in order to make the first step towards the target. The sequence of situation is a journey allowing the construction of the required mental constructs, then language, then means to evaluate and give ground to the piece of knowledge which has emerged.  This is a quick summary, but the essential is there.

It is with this in mind that I reviewed two activities proposed by (@James Kerr), History of Educational Technology-A Collaborative Timeline Project, and (@ElizabethECharl), Webquest – a hunting we will go. In both cases, the difficulty is to figure out precisely what will be the learning outcome and how the situations are appropriate for this objective. Kerr activity is interesting as such, it could stimulate conversations on the history of educational technology and beyond on the role of technology in education. It is an open situation which could give ground to several different learning objective. Elizabeth activity is more focussed on information search on the net. It is a starter, and actually presented as such, which fruitfulness will depend on the follow up either by new situations or by the teacher -- here a librarian. As a learner, I am now in standby in both cases...

#ocTEL MOOC (week 3 A32) While playing, one cant' help learning

The focus on this week second task is on game-based learning as the best example of good case of active learning. The reasons given are brief and clear: "It encapsulates many principles of active learning, such as engagement in an authentic context, learning by mistake-making and reflection, experiential learning, collaborative learning and learning by problem-solving". Apart from the word "authentic" that I would discuss, I agree with the list. But is it enough? In my opinion: "no", because the issue is not that some learning occurs but to be able to tell what learning occurs and, even better, that an intended learning objective has been reached. For this, it is not enough to engage the learners in an active play.

Let's take the case of the proposed games, of which I tried two: the adventure game Lost in the City and the strategy game Westward.  After 15 minutes of play (recommended), I stopped, I stepped back and I tried to respond to the question: "What do you think you could learn playing this game?" The only response I could offer is that we could learn how to play these games and that it may take some time. Then what we could learn once being reasonably familiar with the game is not obvious, although there could be a general statements (I prefer to leave the floor to a knowledgeable other): "The game “Lost in the City” is interesting as an exercise in following directions and solving puzzle" (@James Kerr), "Westward [...]  felt as though it wrapped entertainment around learning very well, and could present learning in an engaging way" (@James Kerr). Yes, but which learning? James Kerr refers to "The Oregon trail" as a similar game. If I got it well, it is both a role-play and a simulation game of a period in the history of the US (as a matter of fact, following a link from the wikipedia page of "The Oregon trail" one reaches "Westward!" and learn that it is an online adaptation of it  - but may be not to confuse with Westward - without an exclamation point).

So, before being lost (or loosing my reader, if one happens to reach this line), I must tell what I learned today from activity 3.2. The first thing is that I learned a bit how to play these games which I didn't know before; and indeed, while playing, I cant' help learning. The second thing is that one cannot say clearly and precisely what can be learned when playing a game; almost every learning is possible from learning how to play, learning some attitude, some skills and serendipitously some content or know-how which could have a meaning and a utility outside the universe of the game. The analysis is almost impossible.

Hence, the reasonable approach is to question the game from the perspective of the learning outcome one targets. I will come back to this point with the week 3 activity 3.3.

They play. So far, so good. But, what do they learn?

Retrieved from a post of Nicolas Balacheff on the SOA scientific portal on February the 26th, 2010
 
Martin Oliver and Caroline Pelletier contribution to the edited book "Digital Generations" is quite interesting and stimulating, taking up the challenge of contributing to our effort to understand "what, if anything, people are learning by playing games" (p.69) Their contribution is based on activity theory, referring primarily to Vygotsky, built on the system formed by the Tool as a mediator between the Subject and the Object (the latter meaning the intention of the subject) and its contemporary extension by Engeström and others which takes into account the social determination of both the Subject and (his/her) Object(ive). The authors make the relevant remark that taking a systemic perspective means that properties which may be identified cannot be ascribed to the Subject as an isolated part of that system. What raises a theoretical and methodological difficulty when the problématique is to understand learning (or Subject semantic/meaning attached to a behaviour). Meeting this difficulty with the cK¢ model [*], I solved it (if I may say so) by considering what could be seen as the projection of the system model onto one of its components, the learner (or onto the Tool). In the case of cK¢ it leads me to propose the (P, R, L, Σ) quadruplet to model the learner conception (what could be mirrored by a quadruplet of the same kind to model the Tool). So, it is clear that I am interested in the method of analysis which the authors propose in order to operationalize the theory.

Then, looking precisely at the proposed methodology, I see a few issues which may be interesting to discuss: contraction, action/operation and in the end the reference to learning and the related question "what is learned?"

Contradiction is a difficult concept to manipulate from a methodological point of view. As Piaget analysed it, contradiction exists if there is a witness of its existence and it can be noticed only if there is an explicit awareness of an objective or an expectation. So there may be a contradiction from the point of view of the observer and not from the point of view of the Subject. How to decide on that? Which observed behaviours can inform the observer? These are difficult questions but critical ones when learning is at stake (as pointed by the authors). So we cannot diagnose a contradiction if there is not an evidence that it is the case for the Subject and hence if we cannot state what is the Object from the Subject point of view. This points a new question: is the Object what the designers or the researchers or the observers claim to be? This question which is important to model the game-playing activity is indeed critical from a learning perspective (it is directly related to identifying learning outcomes). The authors identified in the discussion section, in relation to the interpretation and classification of observed behaviours, the "such claim are difficult to justify without assuming (rather than knowing) the intention of the player" (p.83). My own position is that this is a central issue for learning and that our research effort must start from that point : an explicit hypothesis on the learner intention.

The delicate distinction between action and operation could be better addressed if it was contextualised by such a claim about the intention of t he Subject or the Object of the activity. The authors express their expectations of a progress if a finer grained reading of the actions or the behaviours (eg eye tracking) was possible. My claim is that it may be of no help if the observer cannot relate it to an intention or an objective. Actually it is the identification of the Object in the system and/or the intended learning outcomes (at least as research hypotheses) which will determined the reasonable level of granularity we have to reach.

Eventually, in my opinion, the question "what is learned?" cannot be answered without responding to the question of the objective, intention, aim of the game and the situation which contextualises it. If we do not start from that point, we will progress as blind researchers and in the end respond "they learn how to play" (p.86), which may be a disappointing and quite unhelpful answer. We may agree that this applies also to the problem of understanding the Subject intention, then the nature of the Object and in the end the whole question of learning in a game environment. This issue may be peripheral from a strict game-play perspective, where whatever is learned the motivation and the interest in the game is the thing which counts, but it is critical from an educational point of view.

Note: - Piaget et al. (1974) Recherches sur la contradiction Paris: Presses univ. de France, 2 volumes. - (P, R, L, Σ) stands for Problem, Operators (in French "règles"), Representation (semiotic system), Control structure

A note after the reading of: Oliver M., Pelletier C. (2006) Activity theory and learning from digital games: developing an analytical methodology. In: Burckingham D., Willett R. (eds)  Digital generations (pp. 67-92). Mahwah, NJ: Lawrence Erlbaum.