Une approche ergonomique de la situation d'AMO (Apprentissage Médié par l'Ordinateur) intégrera nécessairement une réflexion sur les objectifs et contenus linguistiques de la formation (axe enseignant ↔ langue), sur l'organisation de cette interaction linguistique dans l'ensemble de la formation (axe enseignant ↔ apprenant), enfin sur le type d'interaction souhaitée entre la langue et l'apprenant et ses implications sur la relation entre la langue objet et la langue-cible (axe langue-apprenant). (Bertin 2002: 77)
- 4 Cf. Cazade 1999a & 1999b, 2000; Toma 1997, 1999a & b, 2003; Rézeau 2002; Bertin 2002.
18In the same way and although their work is not situated in the ergonomic theoretical framework, a number of researchers who are at the same time teachers in French language centres4 also call for CALL or multimedia designs which take into account the potential users’ characteristics and not just various potentialities of machines. For this purpose, it is necessary to examine the learners' interactions not just with an instrument (a computer or a textbook), but with the whole learning system devised by the teachers. Some French researchers will speak of the users’ interactions with a “dispositif”. In English, the closest concept to “dispositif” seems to be the notion of environment, yet the two concepts do not actually overlap as the word “dispositif” has a functional, pragmatic side which is not present in the more general concept of environment.
19A “dispositif”is the association of a material and a symbolical entity, just like an instrument, but it is larger in its scope. The learning environment, in this study, a SAR, included all the actors that interacted while working in it, teachers, students and the administrative staff. It comprised a technical apparatus, plus a pedagogical apparatus, plus an administrative one. It had didactic but also cultural and sociological features which became the subject of the following investigations carried out in a socio-constructivist – not just a cognitive – perspective (Wolton 1999; Raby 2003).
20There are two ways of looking at the usages of modern instruments in the language class. The first one belongs to the designer's sphere, it consists in conceiving all the different possible utilisations that you can make of an instrument, while the second one, anthropological, belongs to the users' sphere and consists in looking at what the actual users are doing with these instruments (Flichy 1995). The two approaches are necessary. First, because no efficient instrument could be designed that would not take into account the target-user from its conception; secondly, because if one wants to observe users, one needs to confront observations to model, and this model is partly provided by what the designers expect the end-users to do. To account for the way in which an individual appropriates or fails to appropriate a learning artefact, Rabardel and Vérillon have developed a theory of the instrument very much inspired by the activity theory of both Vytgosky (1978) and Leontiev (1981).
21The term instrumental genesis refers to all the mental processes related to the specific strategies or rules implemented by learners as they are getting used to a new learning environment or a new instrument.
22If cognition evolves, as genetic epistemology has shown, through interaction with the environment, then it can be expected, in the course of its genesis, to have accommodated to particular specific functional and structural features which characterise artefacts. Does this have an effect on cognitive development, on knowledge construction and processing, on the nature itself of the knowledge generated? If so, through what macro and micro-genetic processes can this effect be thought to be actuated? (Rabardel & Vérillon 1995 : 77)
23Instrumental genesis implies that subjects are first users who, by repeating their uses, stabilise their good (or bad) strategies turning them into rules of usage. These strategies or rules are made visible through specific behaviours, looking, typing, pressing a button, speaking aloud… which can only be indirectly interpreted as behaviours of identification, completion, regulation or control of their task. These behaviours have been defined by Baillé and Raby (1997) and Raby (2003) as instrumental behaviours.
24In this process, operators or teachers and learners are constantly re-defining their task, what they should do and in which conditions. Conditions include the choice of available media, the time constraints and the modes of help and control. This process goes on all along the working process depending on how successful operators are with their work. Leplat (1997) calls it redefinition of the task. In the end, there might be a big discrepancy between the expected or prescribed task designed by the teacher and the task redefined by the student. Yet, this discrepancy does not necessarily reveal a flaw in the didactic system; it may simply attest to the subjects’ independence and autonomy.
25These two studies will serve to illustrate the ergonomic research procedure. Sometimes, ergonomists cannot rely on previous studies to build up hypotheses on which to base their research. In such cases, they start with observing the work situation and after a period of observation, which may be long and detailed, they possibly end up with some hypotheses. In this study, the pre-observations of the SAR had yielded a rather blurred image of the learning environment with a great variety of behaviours and, at the same time, what looked like repetitive patterns. Therefore, researchers have decided to use a construct which would generate a macroscopic image of the SAR seen as a work environment.
26The SAR of this language centre was seen as a working space in which a variety of behaviours were displayed. These behaviours followed didactic/institutional constraints or the students’ own goals and strategies.
27The 116 students belonged to the same university and were characterised through four main variables, their field of study (e.g., eco for economics, pol for political science, soc for social sciences, law for law); the language they studied (e.g., ger for German, spa for Spanish…), their year of study y1, y2, y3, and their sex, (m for men and w for women).
Table 1. Subjects
|
Academic speciality
|
Year of study
|
Language studied when observed
|
Sex
|
|
Law :law
|
Year 1 : y1
|
English : eng
|
Men : m
|
|
Economics : eco
|
Year 2 : y2
|
German : ger
|
Women : w
|
|
Politics : pol
|
Year 3 : y3
|
Italian : ita
|
|
|
Social Sciences /soc
|
|
Spanish : Spa
|
|
28Each student was observed during fifty minutes and his/her moves, actions and spontaneous verbalisations in the course of action were noted down on a grid. The time devoted to each was also taken down. Data will be represented on the tables by codes (see table 1 and table 2) that will later appear in the maps issued from the statistical model we have been using (see figures 1 & 2).
29Some students chose one medium and, then, stuck to it, others changed their medium during the work session, moving, for instance, from the printed medium to the computer. These data were ranged under the label: media selection
30 Other behaviours were related to the way in which learners controlled their tasks, altering sub-goals or procedures and resorting to different kinds of helps. Such behaviours were ranged under the label: regulating behaviours.
31Finally, the time variable was also very relevant insofar as it influenced the management of the task. The amount of time spent by a learner in front of a medium was ranged under the label: work duration.
32The Multiple Analysis Correspondence model demands that data should be coded in a dichotomic/exclusive way (see figure1 & 2). Example: if the video was used, it will be noted video; if it was not used, it will be noted nvideo and the same will apply to all data. If a learner used a dictionary, it will be noted dic; if he did not, it will be noted ndic.
Selection of the medium by the students
33There were four media: the computer (compu/ncompu), the video (video/nvideo), printed didactic documents (dd/ndd), documents with a teaching/learning purpose), newspapers and periodicals (pap/npap).
Regulating behaviours
34They are noted in the maps which describes these behaviours as inq if learners inquired before starting to work, and ninq if they got down to work immediately. In the same way, they might or not use a dictionary (dic/ndic) take notes (note/nnote) they were proceeding with the task in order to discover and memorise some vocabulary and expression. Sometimes, they had a break; sometimes, they never stopped working (pause/npause). They either talked or not with their peers (com/ncom). Finally, some students asked for help either about the content of their task (helpc/nhelpc) or when they had a technical problem (helpt/nhelpt).
Work duration
35Work duration is calculated by subtracting the total amount of interruptions (pause) from the total time spent in front of the medium. After a period of pre-observations, the work duration was split into four classes: wd1 = less than 5 minutes; wd2 is comprised between 5 and 10 minutes, wd3 between 10 and 15 minutes and finally wd4 =16 minutes and more.
Table 2. Instrumental behaviours
|
Media selection
|
Regulating behaviours
|
Work duration in mns
|
|
Computer
|
compu/ncompu
|
Preliminary inquiry
|
inq/ninq
|
0<X<5
|
wd1
|
|
Video
|
Vall/nvideo
|
Makes pauses
|
pause/npause
|
6<X<10
|
wd2
|
|
Didactic printed documents
|
dd/ndd
|
Takes notes
|
note/nnote
|
X>10>15
|
wd3
|
|
Papers
|
pap/npap
|
Communicates with peers
|
com/ncom
|
X=>16
|
wd4
|
|
Seeks for content help
|
helpc/nhelpc
|
|
|
Seeks for material help
|
helpm/nhelpm
|
|
|
Uses a dictionary
|
dic/ndic
|
|
36The model used to describe the SAR as a working space is the Multiple Correspondence Analysis (MCA) (Benzecri 1973, 1980, 1992).
37For the 116 subjects, the researchers noted down the 26 behavioural occurrences that defined as many characters in the space. Characterising this space consists in looking for proximities between these characters. The values taken by the discrete variables are called categories. Knowing that inertia measures the discriminating capacity between categories, the method consists in building a first axis whose inertia is maximum. Then, iteratively, the method creates a succession of orthogonal, two by two axes, with the same mode of a maximum inertia. At each step, the axes can be visualised through a correspondence map. A correspondence map displays two of the dimensions which emerge from principal components analysis of the distances between points (see pp13-14).
38To interpret the maps, one proceeds in two stages, a first stage consists in elaborating an interpretation of the overall structure yielded by the axes; a second stage consists in studying the contribution of each category to the determination of each axis. The contribution of each factor is measured by a percentage of inertia.
Figure 1. Axis one and two with projection of the supplementary variables pertaining to the subjects
y2, nhelpt and npause overlap so we have set their labelling aside
Figure 2. Axis one and three with projection of the supplementary variables pertaining to the subjects
law, eng and work duration overlap so we have set their labelling aside
39In figure 1, axis 1 (horizontal) represents 19.9 % of the total inertia. It is the main axis. From the point of view of media selection, it is the axis of didactic documents and papers (ddand pap) with 30% of the inertia, whereas from the point of view of instrumental behaviours, it can be defined as the axis of peer-communication (com), note-taking (note) and dictionary-use (dic). Axis 2 (vertical) represents 12.8% of the total inertia. It is predominantly the axis of instrumental behaviours: pause (12%), help-seeking on content (helpc =12%), as well as technical problems (helpt = 25%) and preliminary inquiry (inq = 0.08%).
40The weak contribution of the electronic media to the determination of the first two axes was undoubtedly the most surprising finding of this research. It went against the expectations of the university ruling staff when they decided to create a highly equipped language centre and of the language centre's teachers who had devised a SAR highly based on the expected attraction and use of computers and TV sets. But the two maps also yield several interesting contrasts.
41From the point of view of media use, on axis two one major contrast was found, with the printed media (dd and pap) on one side and the use of the computer and the video (compu and video), on the other side.
42One can locate a proximity between electronic media use (compu and video) and the absence of peer communication (ncompu), the absence of note-taking (nnote) and a limited use of dictionaries (ndic). On the contrary, the printed media (dd and pap) seemed to be associated with communication (com), note-taking (note), ask for help about content (helpc) and the use of dictionaries (dic).
Contrasts between the electronic media and the printed media in work durations
43What was more surprising, though, was a clear contrast between the two electronic media, the video (video) and the computer (compu). The students using the video did not seem to inquire, did not ask for help, did not take notes (nnote) and did not make pauses (npause) or use dictionaries (ndic). They also had the longest work durations (wd4=15 minutes or more). The longest work durations (wd1 and wd4) seemed to be linked to the electronic media (compu and video), the shortest (wd1 and wd2) to the printed media (pap and dd). Students who used the computer did not display a lot of instrumental behaviours but asked for technical help (helpt) and had several breaks (pauses). These pauses indicated a sequential organisation of work, which clearly differentiated computer-use from all other media-use.
44In map one, the students that carried out some preliminary inquiry (inq) were those that did not ask for help, either technical or didactic, whereas those who did not inquire (ninq) were those who later sought help (helpt and helpc). In map two, axis 3, one can notice a proximity between (compu) and (inq), which means that students who chose to work on the computer looked for information before setting down to work and then went on with their work with no further help-seeking. The opposition between preliminary research (inq) and help-seeking (helpt and helpc) is confirmed.
- 5 Let us recall that GER or ITA do not mean German or Italian students, but students who are studyin (...)
45In map one, the projection of supplementary variables did not allow strong inferences to be made, yet some conclusions may be tentatively elaborated in the light of the local context. Students who were studying economics seemed to resort to the electronic media and to show few instrumental behaviours. There was proximity between students that were studying politics (pol) and German (ger). They evinced the same characteristics as those in their third year (y3), used predominantly newspapers or periodicals (pap) and sought help on content (helpc). On the other side of axis one, we found students that were studying social sciences (soc) and law (law), predominantly women. They seemed to work on didactic documents (dd) and not on electronic media (nvideo and ncompu). It followed that they did not ask for technical help (nhelpt) and did not need to have many breaks (npause). On each side of axis two, an opposition appeared between Spanish students (spa) who took notes (note), communicated (com) and did not use electronic media, and Italian students (ita) who used the electronic media (compu and video) and did not take notes (nnote) or communicate (ncom). As for the sex variable, one could notice that men and women were on opposite sides of the axes but their differentiation was really weak.
46The first exploratory research had ignored the effect of the pedagogical context. In other words, no distinction had been made between students working in guided autonomy, that is to say those whose work was piloted and assessed by a tutor-teacher, and students whose work was not piloted and assessed. In the second investigation, researchers sought to know if the fact that students were guided made any difference in the way they behaved in the SAR. The pedagogical context was apprehended through two modes, guided autonomy students will be noted as (GA) and non-guided students will be noted as (NG). Incidentally, the MCA method had shown that the computer medium was clearly associated with the longest work duration (wd4 =16 minutes or more); therefore, it was decided to compare work durations with the computer to work durations with any other medium, in the two modes: GA and NG.
47Four independent and equivalent groups were created taking into account the medium with which they had chosen to start working. Among GA students, ten had chosen the computer and ten other students had chosen another medium; the same applied to non-guided students, ten of them had chosen the computer while ten other students had chosen any other medium.
48Each student was observed three times running while they were working on their own. In the SAR, there were some teacher-tutors or student-tutors, but they did not address students upon their arrival, they only responded to students who sought help. Observers had an observation grid in which they took down the number and the nature of media shifts from one session to another and within a single session.
49We used the Analysis of Variance (ANOVA) statistical hypothesis testing procedure first developed by Fisher (1953) to analyse data generated by experimental designs.
50The ANOVA makes it possible to know the influence of two independent variables (here the didactic contract and the medium) on a dependent variable (here the work duration). It is also able to point out possible interactions of these treatment effects on the single response variable (here the amount of time expressed in minutes). Here, the pedagogical variable (GA versus NG) is taken as the source of variation of the work duration in front of each medium (video or computer).
Table 3. Work durations measured in minutes
|
Computer
|
Others
|
Means
|
|
GA
|
68.3
|
81.5
|
74.9
|
|
NG
|
94.12
|
57.4
|
75.8
|
|
Means
|
81.2
|
69.5
|
|
Table 4. Main effects and interaction
|
df
|
Mean square
|
df
|
Mean square
|
|
|
|
Effect
|
Effect
|
Error
|
Error
|
F
|
Level p
|
|
Medium computer
|
1
|
1392,40002
|
36
|
684,905579
|
2,032981
|
0,162533
|
|
Pedagogical:
guided autonomy
|
1
|
8,10000038
|
36
|
684,905579
|
0,011826
|
0,914005
|
|
Interaction
|
1
|
6250
|
36
|
684,905579
|
9,13
|
0,004618
|
51One notices a real difference in work durations in front of the different media. The longest work duration in NG applies to the computer, whereas in AG the longest work duration applies to the other media. The ANOVA analysis points to a strong global interaction: F=9.13; p≤.004 for 3df.
52Both studies accurately illustrate the concept of an exploratory research, that is to say a preliminary stage which aims at testing models and at formulating hypotheses for further research which can be more focused or more structured.
53Results of the first study can only be discussed in the light of the local context of this particular language centre where each language teacher brought his/her own conceptions and teaching strategies which in turn influenced the students' own working strategies. For instance, German students and Italian students were clearly opposed. The former worked on newspapers and did not choose the computer, while the latter were always associated with computer work. This is undoubtedly due to a specific trait of this particular language centre where there is only one Italian teacher and only one German teacher. In a further research the two teachers were interviewed about their work (Borges 2001). The Italian teacher had a long experience of CALL teaching, had received a specific training in computer science and believed very much in student autonomy. On the contrary, the German teacher was a newly trained teacher, had received little training in computer-assisted teaching and devoted her classes to the study of printed documents. She sent her students to the SAR mainly to learn more about Germany through the study of newspaper articles. She admitted to being rather suspicious of new educational technologies.
54As far as the procedure is concerned, results have shown that Multiple Correspondence Analysis may be an interesting approach, not in terms of external validity, for results should be replicated in the same conditions and this is extremely difficult owing to the specificity of each CALL environment, but in terms of its internal validity, for the model seems to yield a relevant picture of a particular learning space and raises acute questions for the researcher.
55First, it is interesting to recall that Raby and Baillé (1997) had already found the same effect of the didactic guidance on rhythmical behaviours, students who were guided selected a greater variety of media and changed more often during their working sessions. In the present study, the chronometric and the rhythmical experiment also tend to show that the students who are guided build up a wider repertoire of behaviours and procedures. This is an illustration of an old tenet still current in scientific psychology which considers that temporal and rhythmical variations during the realisation of the task reveal variations in the procedural treatments and variations in the orientation of action (Sperandio 1980).
Instrumentation and instrumentalisation
56There are some results that no teacher or observer would have been able to predict. Such findings actually modify the vision we have formed of those learning environments and are not so easy to interpret. Among them, there is the weak part played by the electronic media compared to that played by the printed ones. Some would probably suggest that students are reluctant to use the computer or the video because there is always a risk that they might break down; others will argue that it is more comfortable to stick to an instrument which you are accustomed to and will put forward a piagetian interpretation in terms of schemes (Piaget 1967). Another interpretation can be proposed which is inspired by the very nature of the instruments themselves, by their functioning constraints. Printed documents are more easily transportable and one can quickly and easily move forward and backward in a book or a newspaper without fearing to lose some information. On the contrary, electronic equipments, essentially the video but also the computer, are not so easily manipulated. The recurrent opposition between instrumental behaviours linked to the printed media and the electronic media (preliminary inquiries, help-seeking, pauses, work durations) could be interpreted in the same way. It is necessary to insist on this question because, too often, CALL designers tend to focus too much on the pedagogical assets of their inventions ignoring difficulties of instrumentalisation. Being experts in both fields (computer design and language learning) they do not always anticipate the problems encountered by the end-users. As Raby and Baillé (1993) emphasised, the functioning problems (instrumentalisation) often check the process of instrumentation (using the instrument for a specific learning task). This is even truer when students are working on their own, without the mediation of a tutor who is able to help them to overcome these difficulties. The capacity to regulate one's task often linked to reflexive thinking is crucial to autonomous forms of work.
- 6 Also, cf., in the same issue of ASp about learning situations, Percebois 2002, 141-154.
[…] ce qui importe n'est plus la quantité de connaissances sur la langue qu'a un apprenant, mais sa capacité de choisir parmi ses connaissances, d'utiliser ses stratégies cognitives et des procédures linguistiques afin d'effectuer des tâches diverses dans des situations de communication. (Walski 2002: 169) 6
Relevance of the construct of instrumental genesis
57The construct of instrumental genesis has been put forward to describe, analyse and interpret students' diverse behaviours. Although the didactic apparatus and the technical apparatus were the same for all students, behaviours were different. But a difference doesn't entail confusion, some patterns actually emerged in the Multiple Correspondence Maps since specific behaviours (pauses, communications, work durations etc.) were attached to specific media. This validates an anthropological construct of the instrument.
58In the case of this study, the reasons for the diversity of behaviours lay in the psychological diversity of the students, their knowledge of the SAR, their work conceptions, their knowledge of the foreign language, etc. In the same way, although it was impossible in the context of these observations to extract the mental processes of redefinition of the task, results also suggested that the process at work actually implied a process of task redefinition.
59Therefore, results show the relevance of the descriptive model of a learning environment seen as a pattern of behaviours. Of course, behaviours only account for a small part of the system, but by combining these results with those derived from the research centred on teachers and on tutors, one can get an idea of how the actors of a self-access room manage or not to appropriate their learning environment.
60Ergonomists do not limit themselves to behavioural data since what is important, for them, in fine, is to apprehend the mentalprocesses which lie behind these behaviours and which follow from the actors’ knowledge and representations. The second stage of the program consisted of a qualitative study focused on six students working on a comprehension task. Using the cross-checking method devised by Raby (1996), three main kinds of data were cross-checked, the behaviours of students while realising the task, their conceptions obtained thanks to questionnaires, interviews or journals and, finally, traces of their activity, the notes they jotted down on their rough copies or their working sheets. Results showed that three cognitive models were emerging, epistemic, procedural and combined (Raby & Baillé 1997), and that students under-used the resources at their disposal and rarely asked for help (Raby 2003). Finally, the discourse analysis of their journals led the authors to qualify their first results. Indeed, contrary to what the behavioural patterns had suggested, the journals revealed that their greatest difficulties were not with the electronic equipment, or the innovating didactic system, but with the foreign language itself, an epistemic difficulty. At the same time, the journals showed that there was some confusion in the minds of students about their students-instructors (in French, their moniteurs) and their teacher-instructors. They did not seem to know very well what they should expect from these two kinds of tutors in guided-autonomy. For this reason, a further survey followed, focused on the teaching side of the language centre.
61The research was carried out by Collombet who showed how differently tutor-students saw their tasks and pointed out that these tutor-students suffered from a lack of training for their new functions (Collombet 1998). Later, Borges and Raby (2000) and Borges (2001) studied the behaviours and conceptions developed by the teachers of the language centre. They compared the instrumental and didactic behaviours displayed by teachers when they were giving a traditional language practise class to the way they behaved when they were acting as instructors in the SAR. They found a lot of differences between teachers either in language practise classes or guided autonomy classes. They also found very different ways of appropriating the innovating working place, the SAR. One teacher, for instance, did not change at all and behaved exactly in the same way as when teaching in a language practise class, while another dramatically changed his actions, turning from leader to tutor; others seemed to hesitate between the two roles.
62At each stage of the research, results were presented to the managing staff of the language centre and some adjustments and improvements were introduced in the wake of these discussions. Teachers wrote out some didactic leaflets intended for students who were not assisted by a tutor-teacher; they decided to spend less money on buying course-wares and more money on making didactic documents; they gave their student-tutors more pedagogical training; and, finally, they altogether changed the signal system in the SAR. In particular they made the student-tutor more visible by asking them to wear a little flag of their native country. These changes support the idea that empirical investigations in language centres actually make sense. However, it is not always very easy to work hand in hand with practitioners for a variety of reasons and it is often made easier when one researcher in the research team is at the same time a language teacher because he/she often, then, serves as a go-between.
63The authors of this paper hope to have been able to suggest that ergonomics may serve as a unifying theoretical and methodological framework for CALL research seen as a cross-discipline investigation, essentially because ergonomics makes it possible to integrate the three pillars on which CALL research is often based: cognitive and social psychology, second language acquisition and technology. Another asset of educational ergonomics is the variety of procedures it offers: qualitative and quantitative investigations, observations and experimentations, macroscopic and micro-genetic approaches. It is clear that educational cognitive ergonomists support those who fight the widespread and harmful idea according to which there would be an opposition between quantitative and qualitative studies. The question is not "quantitative" versus "qualitative" research, but rather to choose the right method corresponding to a specific problematics and to use the right scientific tools in order to avoid ad hoc conclusions. Moreover, researchers should refrain from generalising their results when they have not been using a method which entitles them to do so.
- 7 Cf. Chapelle 1989, 1997, 2001; Salaberry 1999, Hegelheimer 2000.
- 8 Cf. the work that we are undertaking with Carol Chapelle and Lynn Chue (Chapelle et al. 2003). Our (...)
64Of course, the danger involved when applying ergonomics is that it often borrows from other disciplines such as semiotics, psycholinguistics, pragmatics or anthropology and there actually is a danger of syncretism (Chapelle & al. 2003). But, once other disciplines, with their own models, knowledge and results, have found a unifying theoretical and methodological framework, such as a theory of the instrument or of a learning environment, for instance, the potential dangers of syncretism are avoidable. In the end, to go back to the debate launched by Chapelle and others in LLT7, the work we are carrying out as a team does not exclude other highly relevant and totally different approaches in their scopes, methods and in their theoretical foundations as well. L.L.T, Alsic or Asp offer numerous examples of the wide range of approaches in CALL research. We all agree that the role of research is not to provide technophiles or technophobes with ideological arguments about the pros and cons of CALL but to enable CALL actors to better understand their learning environments and transform them accordingly (Baillé & Raby 1999; Baillé 2001). Therefore, it follows that researchers should be wary of comprehensive, holistic methods which do not lend themselves to any validation. This does not mean taking up a determinist, positivist view of scientific inquiry, just offering one’s discourse to other researchers’ testing and questioning.8