Analysis of an Educational Software for language learning: insights from the Theory of Structural Cognitive Modifi ability and Human-Computer Interaction

The aim of the study is to analyze a software for teaching English as a foreign language reporting (i) the interaction between the software and the learner; (ii) the cognitive/mental operations required to perform the tasks in software and (iii) the pedagogical strategies implemented by the software. Human-Computer Interaction (HCI) aspects of the software were also analyzed so as to evaluate its degree of interactiveness and usability (Ergolist, 2011). Results of the study suggest that the software is content-oriented and the ergonomic analysis revealed that the didactic resources applied by the software meet most usability criteria, requiring few modifi cations. Key-words: educational software; theory of structural cognitive modifi ability; human-computer interaction.


Introduction
Several ideas have been advocated in favor of the use of technological devices in class based on the fact that computers permeate learners' daily activities and may serve as tools to motivate those students who do not usually feel motivated by a traditional class design (Burns; Griffi n; Snow, 1999 cited in Meskill, 2005).
According to Valente (1993), information technology should be seen as a way to open new opportunities for learners and teachers, in a move from more traditional to more modern pedagogical practices centered on the learning, rather than on the teaching process. Though, as pointed out by Reis (2008), a change in the pedagogical paradigm does not relate only to implementing new ways to use computers in class but also needs to be meaningful for all involved in the learning process: learners, teachers, and schools.
What seems to be more relevant regarding the use of new technology in education, apparently, is not the decision of using it or not, but whether these pieces of technology are indeed promoting learning. This concern is also present in the fi eld of English language teaching (hereafter ELT). As argued by Chapelle (1996;2007), the great amount of computer-mediated tasks to which English as a foreign language (hereafter EFL) learners are exposed to and their innovative and in-teracting nature may impact the acquisition of the foreign language once access to new forms of input is constant in computer mediated (hereafter CM) interaction, thus providing extended time, frequency and possibly depth of exposure to the target language.
Another important aspect to be taken into consideration regarding the use of computer-mediated activities in education is related to the mediating characteristics of such activities. From a socio-cultural perspective (Feuerstein and Feuerstein, 1994;Feuerstein, 1997), learning develops thorough mediated experiences as they enable learners to transcend their actual knowledge stage so as to perform tasks which lead to the development of autonomous reasoning and behavior. In this sense, the mediator plays a major role in selecting and modifying the stimulus learners will receive. When addressing the learner-computer interaction, educational software may mediate between stimuli and learners as they are designed to offer users a way to apply the knowledge they have acquired to solve problems and perform different activities. To do that (in the case of educational software designed to teach English), the software designer needs to take into account the knowledge learners already have about the language, their learning necessities, and most importantly, their reasoning patterns in order to enable mediation and therefore cognitive modifi ability, as proposed by Feuerstein (1997) and Feuerstein and Feuerstein ( 1994).
Whether the software is perceived as a relevant tool for learners seems to depend on its degree of interactiveness with the user and the level of complexity. Put differently, aspects of human-computer interactions appear to impact learners' perception of the educational software as well as their performance on the tasks proposed.
With that in mind, the main objective of the present study was to analyze the validity of an educational software designed to teach English as a foreign language to beginners. Taking into account the theory that sees learning as the modifi cation of cognitive structures through mediated learning experiences (Feuerstein and Feuerstein, 1994;Feuerstein, 1997), four secondary objectives were pursued in the present study, namely: (i) to assess aspects of the interaction between the software and the learner; (ii) to investigate the cognitive/mental operations learners need to undergo in order to perform the tasks required by the software; (iii) to investigate the pedagogical strategies implemented by the software and (iv) to evaluate Human-Computer Interaction (HCI) aspects of the software so as to determine its degree of interactiveness and usability. In what follows, a brief review of the main theories underlying this investigation is presented.

Theory of Structural Cognitive Modifi ability
The Theory of Structural Cognitive Modifi ability (hereafter SCM), as proposed by Feuerstein (1997), sees the human mind as being capable of modifying itself. This capability is related to the generation of mental schemes and operations that can be applied to different learning contexts when necessary. According to Feuerstein (Op. cit. in Noguez, 2002), the SCM results from a combination of biological and sociocultural factors and refers to a mutable rather than a static state. Therefore, it develops through changes in the cognitive structures of the learner by means of mediation.
To Feuerstein (1997) and Feuerstein and Feuerstein (1994), mediated learning experiences (hereafter MLE) are essential to human cognitive development since they foster the modifi cation of mental structures. In this sense, the mediator is seen as the potentializer of the learning process, interfering between the stimulus and the learner, leading him, according to his learning needs, to the restructuring of his reasoning patterns.

Mediated Learning Experiences Functions
As proposed by Feuerstein and Feuerstein (1994), mediated learning experiences have two main objectives: 1. To transmit socio-cultural values to learners, these values are intrinsically related to the context and require another individual to be shared. In this case, the mediation generates new necessities for learners, who become more willing to transcend their limits in search for new knowledge.
2. To interfere in the stimulus, adapting it to learners' needs. In this case, the mediator selects the stimulus, manipulates its intensity and frequency in order to emphasize what he/she believes is more appropriate to foster learners' modifi ability cognitive potential. Feuerstein (1997) and Feuerstein and Feuerstein (1994) argue that not all kinds of mediation can be considered an MLE. The three mediating criteria that characterize an MLE are: intentionality/reciprocity, transcendence and meaning.
• Intentionality/Reciprocity: to Feuerstein (1997), every teaching act is intentional, since a specifi c result is expected through a pedagogical intervention. Thus, the mediator (teacher) needs to clarify his/her intentions regarding the instructional contents and learners' expected behavior. With that, it is expected that learners assume a pro-active attitude, collaborating and engaging themselves into the learning process.
• Transcendence: the objective of mediation, according to Feuerstein (1997), is not only the solution of immediate problems. It is also its function to allow the mediated learner to establish connections between what he is learning and what was already internalized, projecting new ways of facing and understanding future demands. The transcendence also permits the mediated leaner to build a net of interconnect knowledge and abstract connections that will, eventually, lead to the creation of generalizations to be applied in new learning situations.
• Meaning: as proposed by Feuerstein (1997), the mediation of meaning is important to foster the propensity of the cognitive modifi ability of the mediated learner, as it makes clear the relevance of instructional contents and the activities proposed by the mediator. In order for learning to occur, in this case, mediator and mediated must interact so as to change values and life experiences, constructing meaning from each other and getting emotionally engaged in the learning process.

Ergonomic aspects of educational softwares: humancomputer interaction
Human-Computer Interaction (hereafter HCI) is seen as part of a broader fi eld of knowledge known as Software Ergonomics. Its objec-tive is to facilitate the understanding of the system, taking into account users' adaptability to it (Barbosa;Silva, 2010) The HCI also aims at understanding how users interpret softwares, assuming that every system can be modifi ed and be constantly improved through feedbacks, making it possible for the software designer to make adjustments according to the mental model of a particular kind of user, creating high-quality interfaces.
HCI aspects are of great relevance due to the fact that human natural language is very different from the complex codes that constitute machine language. In this context, the interface plays a major role once it allows interaction between humans and the machine. Users are responsible for approving and accepting the system as a valid working/ learning tool (Moran, 1981).
To Cybis, Betiol and Faust (2007), there are several techniques to evaluate the ergonomic aspects of a software. One of them is the heuristic assessment, which applies usability patterns to evaluate the software interface at any time of the project, either during its execution (designing of the software) or when it is already in the market.
To assess the quality of the interface of the software selected for this study, a set of usability criteria proposed by Ergolist (2011) was analyzed from a heuristic perspective.

Methodology
The present study was conducted to analyze the validity of an educational software -Interchange Arcade by Cambridge University Press -, designed to teach English as a foreign language to beginners.
The choice of the educational software was based on the fact that it is part of a well-known English series, used by several English as a foreign Language (hereafter EFL) learners around the world. This software provides an on-line companion of grammar, vocabulary and listening exercises to aid learners in the review and practice of content already studied in class. The software is in fact a complement for the Student Book, since it presents the same contents in the same order of the printed material usually used by the teacher. In order to investigate the validity of the educational software as a tool to promote EFL learning based on the tenets of the Theory of Structural Cognitive Modifi ability, a set of criteria adapted from Gomes (2001) was analyzed. This set was chosen so as to refl ect the main principles of the theory and its implications to learning in general and EFL learning in particular, as can be seen in Table 1.

Mediation of Intentionality/ reciprocity
This criterion takes into account whether the software clearly presents its educational intentions to the student by establishing a communication channel with them and expects reciprocity through participation in the proposed tasks.

Mediation of Meaning
It analyses whether the software approximates students and activities, by proposing meaningful and relevant tasks.

Mediation of Transcendence
It analyses whether the software helps students understand the learning process by making them generalize to new contexts.

Knowledge and content
This criterion analyses whether the knowledge is structured in a logical and coherent way. It also takes into consideration whether the software organizes the content according to its particularities, allowing the student to develop specifi c kinds of reasoning.

Form of presentation
It analyses the kinds of languages used by the software such as: numbers, symbols, schemes, visual, audio and audio-visual language. It also analyzes whether the software uses different kinds of languages to reach the greatest number of students. Cognitive functions/ Mental operations It analyses whether the software requires students to undergo specifi c mental operations such as analysis and synthesis. It verifi es whether the software tackles only the fi nal product of the tasks or the underlying cognitive processes.

Level of abstraction
It analyses whether the software utilizes concrete and abstract information interchangeably, according to the students' needs. Level of complexity This criterion analyzes the complexity level of the tasks regarding the number of stimuli or elements involved. For example, a task that requires the student to solve 2+2 is relatively less complex than a task in which he needs to solve X= (2+2+2+2).3/4. Level of effi ciency It analyses whether the software allows students to achieve a reasonable level of effi ciency, trying to establish a balance between diffi cult and easy tasks.

Pedagogical strategies implemented by the software
Observation This criterion analyzes whether the software interacts with the students leading to a metacognition of the observed data, through a series of specifi c and general interventions. Analysis It verifi es whether the software motivates the student to analyze the elements of the tasks, leading them to fi nd relevant information through the mental operation of analysis. Conceptualization It analyzes whether the software helps the student build concepts and name the elements of the tasks. Solution planning It analyzes whether the software helps the student to identify specifi c and general strategies to solve the tasks as well as provides information that helps students build planning strategies. Comparison It analyzes whether the software provides information about the importance of the mental operation of comparison.

Argumentation and logical reasoning
This criterion analyzes whether the software provides terms of the process of argumentation. It also verifi es if the software helps students understand their own steps, motivating them to verbalize their reasoning (meta-cognitive process). Generalization It analyzes whether the software provides concepts that help students transcend the task and assess the mental process involved in learning.
Because of the subjectivity of this type of analysis, three raters were asked to rate the set of criteria based on a scale designed to evaluate how well each criterion is implemented by the software. All raters were professors at a private graduate school. None of them hold a degree in English or is an EFL student. Table 2 displays an excerpt of the data collection instrument and the scale used.  Raters received instructions and a list with the description of each criterion to be analyzed in their L1 -Brazilian Portuguese. So as to assess the ergonomic features of the educational software, its degree of interactiveness and usability, a set of criteria proposed by Ergolist (2011) was evaluated by a fourth rater who is a professor at a Federal Institution and the only expert in Computing among raters. This rater was not asked to assess the validity of the educational software as a pedagogical tool (see the set of criteria in Appendix 1).

The software as a pedagogical tool
The analysis carried out in order to investigate the pedagogical validity of the educational software Interchange Arcade aimed at assessing (i) the interaction between the software and the learner; (ii) the cognitive/mental operations learners need to undergo in order to perform the tasks required by the software and (iii) the pedagogical strategies implemented by the software. Table 3 displays the results taking into account each rater and their assessment of each criterion related to the aforementioned objectives. As it can be seen in Table 3, raters have different opinions about the criteria evaluated. In what follows, an analysis of their evaluation is presented.
=> Mediation of intentionality/reciprocity: the software displays the objective of each task at the bottom of the screen and makes it clear what kind of answer is expected from learners. However, the objective is not about learning, in general, it is related to the content being studied. Therefore, the software intention seems more contentthan process-oriented.
=> Mediation of meaning: all tasks are meaningful if we consider that they provide an opportunity for students to practice and reinforce grammar aspects of the language learned in class. On the other hand, because most of the tasks are concerned with grammar only, they might not be meaningful for those students interested in developing a more communicative (lexical) competence.
=> Mediation of transcendence: because the tasks are presented from the less to the more complex ones, students can generalize about how they were able to solve the previous task when they are to perform the next. => Knowledge and content: the software organizes and displays the content in a logical sequence, from micro to macro structures. In the case of the Simple Past, the fi rst task asks learners to use some verbs in the past to fi ll in the blanks, whereas the next tasks require them not only to use knowledge regarding the correct form of the verb, 30.1 but also a greater understanding of the content, since statements and paragraphs with simple past verbs are introduced.
=> Form of presentation: the software makes use of different types of languages, such as verbal, visual and audio-visual. A reason for the discrepancy among raters' evaluation must be related to the lack of schemes. The software does not usually present schemes to summarize key points of the instructional content. Maybe the use of schemes from which learners could abstract grammar aspects required by the task could serve to reinforce the rules, show learners why their response was not correct and enable them to create generalizations to be applied in future contexts. Schemes could be displayed after the feedback given in the end of each task.
=> Cognitive functions/mental operations: raters' assessment of this criterion may be due to the fact that the software seems more concerned with the fi nal product of the tasks than with the underlying cognitive processes of learning. The lack of more detailed feedbacks after the completion of tasks, for instance, might be an indication that the tasks were designed more specifi cally as a way to operationalize the content studied in class, without deeper concern about developing reasoning/learning strategies. => Level of abstraction: the level of abstraction of the software was rated as good because it makes use of abstract and concrete information. Although grammar exercises usually focus on the abstraction of rules, there are tasks that display information through visual and audiovisual resources.
=> Level of complexity: raters evaluated this criterion very differently. Taking into account the content is presented from micro to macro structures, it is possible to say that the level of complexity seems to increase from task to task. On the other hand, analyzing the instructions given to learners one could argue that it may be somehow diffi cult for beginners to understand as they are complete statements in English.
=> Level of effi ciency: raters tended to have a more positive view towards this criterion. This might be due to the fact that there seems to be a good balance between easy and diffi cult (less/more complex) tasks, providing learners with the opportunity to reach a good efficiency level. => Observation: it seems that raters had a different understanding of what this criterion means. The software does interact with learners in some tasks through feedback messages such as 'you can do better next time', good work'. However, more general instructions about learning how to learn or about the importance of observing certain aspects of the task and their elements seem not to be the focus of the software, since it appears to be more content-than process-oriented.
=> Analysis: raters seem to demonstrate some agreement concerning this criterion. Although tasks do not explicitly require learners to analyze elements to come up with an answer, learners do need to apply this mental operation so as to solve the tasks adequately.
=> Conceptualization: there seems to be a greater disagreement among raters regarding this criterion. Because the software is more content-oriented and focused on performance, there seems to be little concern for the building of concepts that allow the transfer of knowledge to different learning situations. => Solution planning: this criterion is somehow related to the building of general learning concepts. The lack of such concepts might undermine learners' opportunities to improve planning strategies that can be used in other language-use contexts.
=> Comparison: this criterion seems to have caused considerable doubt to raters, since two out of three rated it Excellent and Good, and the third replied not having understood it at all. Indeed, the instructions to the activities presented by the software do not make clear whether learners should compare different elements to perform the tasks. However, if learners are strategic enough to generalize on what they have previously learned, it is probable that they will eventually make use of this mental operation (comparison) each time a new tasks with a higher level of complexity is proposed.
=> Argumentation and logical reasoning: raters' assessments diverge on this criterion. All the tasks proposed by the software do require learners to make use of logical reasoning if they are to be performed correctly. However, there is no explicit instruction to learners aiming at motivating them to think about the steps they took to solve the tasks. Little concern seems to be given to the development of meta-cognition in the way tasks are proposed by the software.
=> Generalization: raters' evaluation of this criterion represents very different points of view. This might be related to the auto-didactic characteristic of the software. That is, due to a more content-oriented approach, it seems that the software was designed to be self explanatory. To some tasks, when learners perform incorrectly, the software provides the correct response as a feedback, but does not give emphasis on the grammar rule itself. If learners are not able to deduce why they were wrong, it is likely that they will have diffi culty to create generalizations to these rules. Even if they are able to do that, it does not guarantee generalizations will be correctly applied in new tasks. However, generalizing leads learners to form and test hypotheses about language in use, which seems to make it a very important strategy.
Summing up, data analysis revealed that the software Interchange Arcade seems to be more content-than process-oriented, as little concern is given to the development of learners' metacognitive awareness. It seems that the software also lacks more detailed feedback messages regarding specifi c aspects of language that learners did not get right when performing the tasks. Another drawback of the analyzed software is its little emphasis on the building up of learners' reasoning and learning patterns. It is essential for learners to know what kind of mental operations (as well as how to apply them) may be used in other learning situations, extrapolating from that specifi c language rule he/ she is learning at the moment to a broader context in which a particular cognitive behavior is necessary to foster cognitive modifi ability and thus, learning.

Ergonomic assessment
Interchange Arcade is entirely available on the Internet (Interchange, 2012).The ergonomic criteria proposed by Ergolist (2011) and analyzed in the present study were: promptness, grouping by location, grouping by format, feedback, legibility, concision, minimal actions, informational density, explicit actions, user control, fl exibility, user experience, protection against errors, error message, error correction, consistency, meaning and compatibility. In what follows, a qualitative analysis of each individual criterion is presented: => Promptness: all the necessary information for the user to interact with the software is available, facilitating the use of the software and directing students' attention to learning.
=> Grouping by location: the software follows a logical sequence regarding the distribution of the objects on screen, which are easily identifi ed by students, contributing to keep learners attention on the tasks.
=> Grouping by format: there is a visible difference among commands, options, data and instructions presented by the software, either regarding the kind of letter or the colors and audio signals.
=> Feedback: the software provides learners with information regarding its actions. That is, it displays a specifi c symbol on the screen each time a piece of information is being loaded. This kind of resource helps to direct students' focus of attention to the learning object.
=> Legibility: the icons used by the software are readable, allowing for a fast interpretation of their meanings, mainly because they are followed by written statements.
=> Concision: the software makes a good distribution of icons, pictures, drawings and other pieces of information in relation to the space on the screen. => Minimal actions: there are actions required by the software that would not be necessary if this criterion were accomplished. As for instance, when the learner needs to fi ll in the blank gaps in a sentence, the cursor does not appear on the gap that should be fi lled out. In this case, to decrease the number of learners' actions, the cursor should appear where the text is to be placed, speeding the execution of the tasks and freeing students' focus of attention from irrelevant aspects. Besides, learners should be able to move through the software by using not only the mouse, but also specifi c shortcuts in the keyboard. For an advanced computer user, this might facilitate navigation and save performance time.
30.1 2014 => Informational density: the amount of information (data density) displayed by the software is in accordance with the windows' size or spaces reserved to this end, helping learners to make sense of the content presented.
=> Explicit actions: the software waits until learners give specifi c commands to perform corrections. That is, an explicit action is necessary from the learner to prompt a particular software response, which is a way of providing learners with the control over the completion of the tasks.
=> User control: The software does not require any pre-requisite in order to perform the tasks. This means that students are allowed to repeat the exercises/tasks as many times as necessary, in any sequence they prefer, giving them control over their own performance regarding the tasks separately. However, learners seem not to have a good control over the great scenario, that is, the software does not provide any kind of record containing scores and other relevant information for students to access in order to have a view of the whole learning process, becoming aware of which contents they are not mastering or what aspects of the tasks they need to improve.
=> Flexibility: this criterion is not accomplished by the software, since it does not allow learners to choose specifi c kinds of screen configuration, such as particular features for beginners and advanced users, different menu layouts, different background colors, among others.
=> Users experience: in order to fully meet this criterion, the software should allow learners to use the keyboard to replace some functions which are only made by using the mouse, such as the selection and execution of some actions needed to perform the tasks. The software does not take into account the fact that students may be either experienced or novel computer users.
=> Error protection: the way the software was designed contributes to avoid usability errors, since learners are advised when the data input is not correct and must be done again. There is also no danger concerning data loss as they are not recorded for further actions.
=> Error message: in general, messages to users are clear, impartial and polite, meeting this criterion well. 30.1 2014 => Error correction: the software offers students the possibility of error correction by giving them the chance to repeat the exercise as many times as needed.
=> Consistency: there is a good consistency among the different windows displayed by the software. That is, the icons and symbols used in one window always express the same meanings when presented in other windows. The same happens with their color, format, and size of letter used.
=> Meaning: in most cases, the statements that precede tasks are objective and self-explanatory, which, for some learners, may facilitate the understanding of what needs to be done.
=> Compatibility: there is a certain level of compatibility between learners' expectations and needs regarding the software, once it provides tasks to different levels of knowledge, allowing students to choose what exercises they want to perform. If expectations are not met in the fi rst time, they can decide to go over the same exercise so as to solve their doubts.
In general, data analysis revealed that the software Interchange Arcade meets most of the criteria evaluated, being a relevant tool to be used by learners of English as a foreign language.

Final considerations: limitations and further research
In general, data analysis revealed the educational software presents a content-oriented approach to EFL learning. Apparently, there is little concern with the underlying mental processes involved in the performance of the tasks, once there is lack of more detailed feedbacks that could lead to the development of learners' meta-cognitive and strategic behavior and thus to cognitive modifi ability.
It might be that, if the software provided more detailed feedbacks, including the reasons for errors and details about the grammar rule learners should master to solve the task, they would have more chances to repeat the task and get a positive result, instead of going through tentative and error trials. Though learners could search for the rules in the class book, it would still be interesting to have the chance to choose whether to do it or not.
Regarding the ergonomics analysis, the resources used by Interchange Arcade meet most usability criteria, requiring few modifi cations. The most important interface design features that need re-elaboration, as revealed by data analysis, concerned the following criteria: minimal actions, fl exibility, users' experience and user control.
Data analysis also showed it might be important for learners to have the possibility to check their progress through the performance of the tasks in order to have a picture of what they need to improve. A report presenting some statistics regarding learners' performance at the end of each unit would allow for a more accurate analysis, by the learner of which aspects of the instructional content might not have been internalized, or might need more practice or even a better understanding of the language rules 1 . By analyzing which language aspects need improvement, learners may develop a more pro-active behavior towards language learning as they will probably be able to build a set of meta-cognitive strategies to be applied in different learning situations, thus fostering cognitive modifi ability.
The present study suffered from some limitations. First, regarding the relevance of the educational software as a pedagogical tool to promote EFL learning, data analysis revealed a reasonable disagreement among raters for most criteria evaluated. It is believed that certain level of diffi culty to evaluate some criteria may be due to the fact that raters were nor teachers nor students of English. As a result, in order to overcome the aforementioned drawback, further research should try to implement the following actions: Re-elaborate the assessment criteria so as to make them clearer and more specifi c in order to minimize doubts from raters; Have English teachers as raters; Check the internal consistency of the data collection instrument by applying appropriate statistical tests; Check for inter-rater reliability.
Second, the ergonomic analysis was conducted by one rater only, since it was the only computing expert among raters. Future research should have other raters assessing the ergonomic aspects of the software in order to gather different points of view about the same pedagogical tool. Moreover, it would be interesting to have learners assess some features of the educational software so as to obtain a different perspective of its degree of usability and interactiveness.
In sum, although the educational software Interchange Arcade seems to be an appealing technological resource to be used in the teaching/learning of English as a foreign language, more research is needed to fully evaluate its contribution to foster structural cognitive modifi ability and thus, learning. Verifi ca a compatibilidade do sistema com as expectativas e necessidades do usuário em sua tarefa.
The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked