Keywords: mediation, mediator, interlingual and intralingual mediation, intercultural mediation, translation, interpretation, social practice

1. The notion and practice of mediation

Mediation is the act of extracting meaning from visual or verbal texts in one language, code, dialect or idiom and relaying it in another, so as to facilitate communication. That is, mediation, which has also been defined at length elsewhere (Dendrinos 2007a,2007b[1] and 2006), entails providing information from a source text that an interested party has no access to, or explaining a message contained in a text (verbal or visual) to someone who does not understand it. As users of language(s) and informed about cultural and social practices, we are all potentially mediators. When we assume the role of mediator, it is so as to participate in a communicative event, acting as a go-between, an intermediary whose job is to help someone understand the message delivered. We intervene to help the flow of in­teraction and facilitate the exchange. The need usually arises when two or more parties interacting are experiencing a communication breakdown or when there is some type of communication gap between them. The mediator intercedes as a meaning negotiator, undertaking the task of reconciliation, settle­ment or compromise of meanings.

When we perform as mediators, we become meaning-making agents; that is, we create meaning for someone else, who is unable to understand what is going on, to comprehend a text, whether this is in a language s/he knows well or it is in a foreign language. We create and interpret meanings through speech or writing for our inter­locutor(s), with whom we may or may not share linguistic, cultural and/or social experiences.

From the above, it becomes obvious that we mediate when there is need to make accessible informa­tion that a friend, a colleague, a family member, etc. does not grasp; it originates from the need to have something clarified, to interpret or reinterpret a message, to sum up what a text says for one or more persons, for an audience, for a group of readers, etc.

Interpreting a word, phrase or a whole text as part of the act of mediation should not be confused with the job of the professional interpreter – at least not if one thinks along KPG lines. The KPG definition of mediation does not coincide with the definition provided in the Common European Framework of Languages (henceforth CEFR), where mediation is viewed as synonymous with translation and with interpretation. There is a rather sharp distinction between written mediation in the writing test paper of the KPG exam (Module 2) and professional translation of functional, technical or literary texts. Moreover, there is a very pronounced distinction between simultaneous interpreta­tion (at conferences, business meetings, etc.), consecutive inter­preta­tion (during speeches, guided tours) and other similar practices on the one hand, and KPG oral mediation on the other. The latter concurs with what the CEFR describes as informal interpre­tation in social and transactional situations for friends, family, clients – interpretation of signs, menus, notices, etc. However, mediation is not limited to interpretation. It may serve a series of other communicative purposes, such as reporting, explaining, directing, elaborating or providing gist, defining, instructing, and much more. 

Usually, when people in real life mediate – and this is true of KPG mediation too – they may resort to certain translation and interpretation techniques, but their job is not to produce a text or speech equivalent in meaning and similar in form as when translation and interpretation are at work. The very purpose of translation and interpretation is the production of configurations which are as close as possible to the original, i.e. to the source text. The task of translators and interpreters is to establish corresponding meanings between source and target text, with perfect respect for the source text – the what and the how it articulates its meanings.

Mediators, unlike translators and interpreters, have the prerogative of producing their own text; a text which may not be equivalent in terms of form, while it may be loosely connected in terms of the meanings articulated. Mediators bring into the end product their own ‘voice,’ often expressing their take on an issue. They select which meanings or messages to extract from a source text and then decide how to convey them. Their choice in real life is necessarily dependent on why the mediator is interfering, which means that the outcome of mediation (particularly where form is concerned) is task specific and addressor specific. The outcomes of translation and interpretation on the other hand are usually text specific, though the audience for whom the text is intended is always taken into account. Put differently, the translator’s and interpreter’s ‘loyalties’ lie with the source text, whereas mediators’ loyalties lie first and foremost with the interlocutor.

Finally, transla­tors and interpreters have no ‘right’ to change the discourse, genre or register of their text; it is to be the same as the source text. For example, say that someone undertakes the job of translating a theatrical play. What s/he must do is to produce a play in another language, not a summary or a review of that play! A mediation task on the other hand may involve just that. An advanced level KPG test paper could include a short one-act play in the candidates’ L1[2] and the mediation task could be to have candidates use L2 to speak on the meaning of the play (oral mediation task) or to have them write a review of that play (written mediation task).

Mediators have the right to change the discourse, genre or register of their text, and having the prerogative to do so is not an issue because it is an inherent component of their role as mediators. Imagine that you are a medical student and you visit the doctor with your father, who has been taken ill. Upon leaving the doctor’s office, your father asks you to tell him what the doctor’s diagnosis was because he didn’t understand a word of what she said – not because what she said was in a foreign language, but because she used medical discourse. So, your father puts you in the role of mediator, and this isn’t always easy not only because you have to think of how to turn medical discourse into plain language, but because you also have to interpret the communication breakdown between doctor and patient, which may be quite complex. No matter what, though, your job is to select those bits of the doctor’s message that you think are crucial and say them in a way that the patient (your father) can understand. If there is serious illness involved, you may have to modify or play down what the doctor said, or you may even consciously decide to conceal some information so as not to scare him.

The case discussed above is an instance of intralinguistic mediation. It does not involve relaying information from one language to another but in the same language. In other words the doctor spoke, say, Greek and you report the doctor’s diagnosis to your father also in Greek. Some teachers would say that this is an act of transforming direct to reported speech. But, obviously, this is more than a (grammatical) transformation exercise, because what the mediator has to do here is to select salient information provided in medical discourse used perhaps to account for the symptoms and conveyed in formal register (possibly to report, instruct, warn, etc.), and to relay all or part of this to the patient in simple, everyday, informal language.

Another example of intralinguistic real-life mediation is when, say, you tell your friends Joshua and Laura a joke in English. While you’re expecting both of them to laugh, Laura bursts out in loud laughter but Joshua just smiles and looks confused. Laura explains what’s funny about the joke, using English which is the common language between the three of you. The act she’s performing is intralinguistic mediation. She’s mediating in the same language, as you did when you went with your father to the doctor earlier. Though such tasks are not labelled as mediation, KPG does use intralinguistic mediation tasks, involving only the target language, i.e., the language tested. For example, the first activity of the C1 writing test paper asks the candidate to extract information from one text in English and to compose a script using the information extracted. In many instances the text to be produced is to be of a different genre and register than the source text, as for example where candidates are asked to read a webpage which provides information regarding ‘Education for sustainable development’ and are asked to write a review, recommending the webpage to website visitors (Appendix 1). 

Unlike intralinguistic, interlinguistic mediation involves two (or more) languages. A real life example is when you’re watching CNN and your mother who doesn’t understand English happens to see a scene which surprises her. Impressed by it, she asks you what’s going on. You’ve been listening to the news and you report to her in Greek the gist of what’s happened. It is this type of interlinguistic (oral) mediation task that we see in the KPG speaking test, with one signifi­cant difference: candidates are asked to mediate only from Greek to English and never the other way around. The same is true of written mediation, whereby candidates are asked to use information in a written text in Greek to compose one in English which may be of a similar or of a totally different text type, register and style, while the two texts may have totally different genre (that is, text type and communicative purpose). For example, a writing mediation task could originate from two Greek ads about houses that a teacher could easily locate on the internet. The mediation task that could be assigned is:

Read the two ads about houses for rent on the island of Kalymnos, where your friend Amelia wants to spend one month in the summer with her family. Write an email to her to inform her about what’s available and to recommend one of the two to her.

We can use the doctor-patient example provided above to explain further interlinguistic mediation – the only kind of mediation that the KPG actually labels as mediation. If the cause of the communication gap between doctor and patient was language – i.e., that the interlocutors spoke different languages (say, the doctor spoke English only and the patient Greek) – you, as mediator, would be called upon to relay salient points of the doctor’s message in English to the patient in Greek. Or vice-versa, if the language spoken by the doctor was Greek and the patient spoke English only. This latter type of communicative situation resembles the context set up for the KPG mediation tasks in the speaking test. For example, the November 2005 speaking test paper contained a magazine page with short numbered texts about fun activities for children. One of the related tasks was:

Help your friends who have two children (aged 10 and 12). They are in Greece for the summer. Give them advice about activities that their children would enjoy. Read texts 2 and 3 and explain why their children would enjoy these particular activities.

Obviously, mediation is both a spoken and written activity in our everyday life. Therefore, both spoken and written mediation are tested in KPG, where candidates produce an oral or a written text in L2 for the speaking and the writing test, respectively, based on one or several source texts, in L1. The Greek source texts from which candidates are asked to extract information are always written and often complemented with visuals (a graph, a map, a sketch, a photo).

What has just been explained in the above example concerns the linguistic mediator. However, increasingly important is the function of the cultural mediator[3] – a role that entails explaining the social meaning of specific cultural practices or traditions, filling in information gaps about social issues, customs and values, or accounting for the operation of social institutions, etc. We conventionally do these things for listeners or readers who do not share the same cultural experiences with us. In other words, we intervene not because our interlocutor lacks the linguistic resources but rather the cultural awareness required; s/he does not have insight into the cultural reality in question, or rather what we call ‘intercultural awareness’ – insight into one’s own and the foreign culture. The person in the role of cultural mediator does have it and is therefore able to explain things to someone who lacks it. Think, for example, of a situation where a group of Greek friends are talking politics; they are rather loud, they often interrupt one another and all talk at the same time. An English friend, watching, asks you why these people are fighting. Your friend’s question puts you in the position of cultural mediator so that you explain that they are not actually fighting; they’re just expressing their views in a passionate manner.

Actually, the KPG exams do assess target cultural awareness, but indirectly and not in isolation from language. For example, in the May 2007 English exam of B2 level, Activity 7 of the reading test (Module 1) is an acrostic quiz. Candidates are asked to read what people are saying in COLUMN B, to guess what their job is, and fill in the gaps in COLUMN A. To fill in item No. 69 below with the word ACTOR, the candidate must have quite a bit of cultural knowledge to pick up the cues and infer that the person speaking is an actor. 

Furthermore, in order to successfully complete the mediation tasks of the speaking and the writing tests of B level and C level exams, candidates are required to have developed not only (socio)linguistic awareness, but also (inter)cultural awareness. An example documenting this is the B2 level speaking test of the April 2005 exam in English, which contains a task where candidates are asked to look at various photos (Appendix 2) and explain what the purpose of each ceremony is, and what usually happens on such occasions. Also, candidates need to have developed cultural and intercultural awareness in order to respond to tasks such as the ones below, which are from a speaking test activity of the C1 level exam in English.

Tell us what you think people mean with the saying “Don’t put all your eggs in one basket,” and if you think that this is always true.

There is an English saying which goes: “Absence makes the heart grow fonder.” There’s also a Greek saying which is the exact opposite: “Μάτια που δε βλέπονται, γρήγορα λησμονιούνται”. Which one would you agree with and why?

Finally, comprehension tasks in the KPG exams often include cultural information aiming at a backwash effect for the development of intercultural awareness. For example, the May 2007 B1 level exam in English contains the activity below.

Fill in the gaps in items 46-50 with ONE word so that each rule makes sense.

2. What KPG mediation involves

What is labelled as mediation in the KPG exams involves verbal activity intended to bridge the gap between a source text in L1 and a target text in L2. The mediation tasks in the writing tests of the B1 and B2 level exams require that the candidate selectively extract informa­tion, ideas and specific meanings, and then produce a script which has the same thematic concerns but often is to be articulated in different discourse, genre, register and/or style. The B1 level Horoscope mediation task (Appendix 3) is a good example. In doing this task, though the thematic concern of the source and target texts is similar, successful completion necessitates production of different discourse, genre, register and style. That is, whereas the source text is a horoscope, the text that the candidate is asked to produce is an e-mail. Also, whereas the source text makes use of public, impersonal discourse and semi-formal language, the target text requires use of more private, personal discourse and informal language.

It becomes obvious that mediation is no easy job – neither linguistically nor cognitively. Mediators have to make complicated decisions about the information to be extracted from the source text, the content of the message to be delivered and the form of the text to be created, so that it is appropriate for the communicative event and useful for the other par­ticipant(s). On the other hand, mediation tasks usually demand degrees of literacy in both languages as well as various types of competences and skills. In other words, depending on what the task actually is, demands may be any one or more of the following (Table 1):

As already pointed out, the goal of mediation is to facilitate inter­action during a communicative event, to fill in a communication gap or resolve some sort of communi­cation breakdown.  The goal itself sounds uncomplicated, but the process is rather challenging, as Table 2 below shows.

All the steps that the mediation process entails are demanding, but step 7 is perhaps the most crucial of all, at least from a linguistic point of view.

3. The use of L1 and mediation in the KPG exams

It has been made clear that the writing and speaking test papers of the KPG exams in all languages test candidates’ oral and written mediation performance from the B1 level exam onwards. In the A level exams, mediation is not tested, though there is consistent use of the L1 in the reading and listening test papers. The use of L1 at this level functions as a facilitator to understanding the L2, as will be shown below.

3.1. The use of L1 in the A level exam

3.1.1. Reading comprehension

The texts that candidates have to read are always in L2, whereas the rubrics are consistently in both languages, as in the example in Appendix 4. The reading comprehension items which accompany the text (mostly objective-type items like multiple choice, multiple matching, True or False, etc.) are commonly in L2, with two exceptions: There are two activities whose items are articulated in L1. The function of L1 in this case is to help candidates demonstrate their understanding of content and the semantic/pragmatic meaning of parts of the text or of single utterances. The discussion that follows and the examples provided below illustrate these points.  

Knowing that foreign language readers understand much more than that which they are able to produce – partly because they lack the ‘vocabulary’ in the target language to express themselves – L1 is used to pose rather sophisticated questions (Appendix 4). The Step 1 task, originating from a text in English but with reading comprehension items in Greek, aims to test the reader’s understanding of the purpose and gist of the text. Such items would be too difficult for the A level reader to understand if they were posed in English rather than Greek. The second step aims to test language awareness, and L1 is used to pose questions about the pragmatic meaning of utterances in the text. Of course, this task also tests candidates’ ability to establish semantic equivalence between utterances in L1 and L2, which is cognitively quite a demanding job.

3.1.2.  Listening comprehension

Texts that KPG candidates listen to are always in the target language. The rubrics in Module 3, i.e. the listening comprehension test paper, are consistently in both the L2 and L1, like in Module 1. The listening comprehension items which accompany the text (mostly objective-type, such as multiple choice or True-False) are in L2. There is only one activity in the listening test where L1 is used for a similar purpose as in the reading comprehension test: to help candidates state what they have understood without having to use L2 at a level of competence they have not yet developed. For example, they listen to three recorded phone messages and they are asked to respond to True and False items with regard to the purpose of each phone message. The choices are in Greek.

4. Testing mediation in the B and C level exams

The CEFR does not provide a list of benchmarked illustrative descriptors for each level of language competence for mediation, as it does for other areas of language use. It is our long-term goal at the RCeL, on the basis of a large-scale research project which has already started, to provide a detailed account of mediation performance at each one of the levels where mediation is tested in the KPG exams. In the meantime, this paper explains, based on published KPG specifications and task description, what the expectations for mediation task completion in the KPG exams are

• in the two test papers that assess mediation, and

• at macro levels of proficiency where mediation is assessed, i.e. at B level (Autonomous user), which includes micro-levels B1 and B2, and at C level (Proficient User).

4.1.     Mediation in the writing test

Mediation tasks in the writing test originate from text(s) found in printed or website sources on issues that are relevant to the average Greek candidates’ cultural experiences and literacy.

Tasks are designed to encourage the use of the text as a source of information rather than as a meaningful entity in one language to be rendered as a whole into another language, as in the case of a translation task. Tasks in the B1, B2 and C1 level writing test increasingly require that candidates make reference to specific points raised in the Greek text as well as add additional information which stems from their world knowledge and experiences regarding the issue in question.

The assessment criteria for writing performance at each level, based on an L1 source text, are the same as the criteria for writing performance for tasks with cues or a source text in L2. To help the script rater assess candidates’ mediation performance, expectations for mediation task completion are articulated, though there is a need to standardize these expectations more for each micro-level, across languages.  

4.1.1.    B1 level writing task completion expectations

According to the KPG published specifications, candidates are expected to compose in the target language a script of about 100 words which:

• communicates the gist of a Greek text, or

• relays information extracted from the Greek text. 

The Greek texts in the B1 level test papers are very often from popular magazines and touch on topics such as health and diet, exercise and daily routines, travel and sports, work and school, human relations. In other words, they are on topics that do not require the use of specialized vocabulary. Writing mediation tasks stem from either several brief texts on one topic, appearing on a single page usually elaborately designed, or from a single text in sections, also richly illustrated. Extended narratives, news articles or reports are altogether avoided.

As writing mediation tasks (WMT) from different exam administrations show[5], the script to be produced is consistently of a genre that candidates are very familiar with – an e-mail message – to give advice, warn, inform, explain, describe, etc.:

WMT 01: The task of the May 2007 exam administration is based on several brief texts regarding myths and facts about nutrition from a Greek magazine. Candidates are asked to write an e-mail, giving their friend tips about healthy eating.

WMT 02: The task of the November 2007 is based on two horoscopes which are divided into sections about love and career, and which also contain a piece of advice. Candidates are asked to write an e-mail to a friend, Ursula, to warn her about spending too much money and to tell her that, based on what the horoscope says, her husband might get the job he’s been waiting for.

WMT 03: The task of the May 2009 administration is based on a single text about the Mediterranean diet, divided into sections and complemented with a visual, which labels the foods in English to help candidates with vocabulary. Candidates are asked to write an e-mail to their friend Scott, explaining what the Mediterranean diet is all about.

The cognitive demands in all of these tasks are related but different, and the linguistic requirements vary. In all three instances, candidates must select pertinent information from the source text and use it to convey a message in L2.

In the case of WMT 01, candidates have to use the information provided in L1 statement form to give their friend tips in English about healthy eating. These tips may be expressed in the form of suggestions (‘You should … ,’ ‘It’s a good idea to … ,’ etc.), in the form of commands (‘Do this … ,’ ‘Avoid that … ,’ etc.) or as factual statements, either in impersonal or personal forms (‘We should drink lots of water,’ or ‘People should not drink …,’ ‘We should not drink water or soft drinks with our food,’ or ‘People should not drink … ,’ etc.).

Relevant-to-the-task selection of information is required in WMT 02. But whereas in WMT 01 candidates have to read each short text very carefully and decide which tip they will use and how they will express it, in WMT 02 the required information is in specific parts of the text, each of which has a heading. So, candidates skim though and focus each time on the relevant part, interpret the message contained in accordance with what the task asks them to do, and use the information to a) warn in the one part of their script, and b) make a prediction in the other.

In the case of the third example, WMT 03, candidates are required to combine bits of information from the whole text in order to explain what the Mediterranean diet (described verbally and visually in the source text) is.

4.1.2. B2 level writing task completion expectations

According to published specifications, candidates are expected to compose a socially meaningful text in the target language – a script of 130-150 words which:

• conveys the main idea of a text in Greek, or

• makes a summary of the Greek text, or

• relays messages contained in the Greek text

Systematic task description at this level shows that the L1 texts in the B2 level test papers are on more sophisticated topics than those at B1 level, such as those from which tasks WMT 04-06 originated.[6] They are also from a wider variety of sources, such as promotion leaflets, newspapers, magazines, books and websites. Both the source text and the target text are of a wider variety of genres. That is, while at B1 level the source texts are frequently short, popular magazine texts on everyday topics and candidates are asked to write personal messages drawing information from the source texts, at B2 level there is a wider range of text types, such as a graph (November 2003), a tourist guide (June 2004), a webpage (November 2004), a newspaper article with a figures table (April 2005), and a website event announcement (November 2005). The text types to be produced are quite diverse also, as one can see in the examples provided below.

WMT 04: The task of the May 2006 administration is based on a Greek book announcement containing factual information about the novel (title and author, ISBN, cost, publishing house, etc.) with the story line of the novel articulated as a narrative. Candidates are asked to write a book announcement for the publisher’s English book catalogue.[7]

WMT 05: The task of the May 2008 administration is based on four movie briefs, of the type that one finds in the film section of a newspaper or magazine. Candidates are asked to write a text for the WHAT’S ON guide appearing in English, recommending two films for children and two for teenagers.

WMT 06: The task of the May 2009 administration is based on a webpage of the Greek Ornithological Society, promoting a volunteer project on the island of Syros. Candidates are asked to write an e-mail to their friend Martin, with whom they have decided to spend part of their summer vacation doing volunteer work. Using information from the website, they are to try and convince Martin that it’s a good idea for the two of them to take part in this project. They are not totally free to choose any bits of information; rather, it is suggested that they stress those aspects of the project which make it particularly ideal for them, i.e the location, the flexible dates, the cost, and the type of work they will be doing.  

4.1.3.    C1 level writing task completion expectations

According to the published specifications, C1 level candidates are expected to be able to use their knowledge and the communicative competencies they have developed as users of Greek and the target language, in order to act as mediators in the educational, professional or public sphere. More specifically, candidates are expected to compose a 200-word script in the target language in order to:

• convey the main idea or supporting details of a Greek text, or

• summarize a Greek text, or

• interpret in the target language the meaning or meanings of one or more messages in Greek.

Task analysis reveals that there are certain differences between the B2 and C1 level source texts used for mediation activity, such as length and sophistication of text. While a variety of genres are to be produced, as in B2 level mediation, C1 level production requires:

• an impersonal text articulating public discourse, or

• a text type which coincides with the source text, or

• more specialized vocabulary (motivated by the source text) and formal register (instigated by the task).

Actually, at C1 level, the mediation task obliges candidates to stick more closely to the source text and relay specific pieces of information rather than select those items they can write about in the target language. Below are some examples:

WMT 08: The task of the November 2006 administration is based on the website of the SOS Villages Greece, and candidates are asked to produce a report for an international organization which funds important social projects.[8] The purpose of the report is to promote the work being done in Greece and to stress its social usefulness so that they get the funding they need.

WMT 09: The task here (Nov 2007) is based on a newspaper article, translated from English into Greek, originally published in the Evening Chronicle. This article, which also contains two pie charts, presents the results of a survey on tourist services in Greece and specifically the percentage of tourists who believe that services in Greece are good, mediocre or bad and the percentages who believe that it is better, worse or the same as other EU countries. Candidates are asked to read the charts and the article and write a letter to the newspaper editor to a) express doubt that this is what people really think of Greece, b) point out that the article does not reveal how the survey was conducted and by whom, and c) present their own evaluation of tourist services in Greece.

WMT 10: The task of the November 2008 administration is based on the review (in Greek) of a book which originally had been written in Swedish and recently translated into English. Candidates are asked to use the information from the book review and write a book presentation for the catalogue of the publishing house they supposedly work for. 

The genres to be produced in all the examples above are obviously more demanding linguistically than those of B1 and B2 level: Twice candidates are asked to produce a report, and a third time a letter to the editor of a newspaper. But even when they are asked to produce a text of a similar type as that produced at B1 or B2 level, such as an e-mail, at C1 level the communicative purpose is quite different (Appendix 6).  

4.2. Mediation in the KPG Speaking Test

Mediation tasks in the speaking test require that information be extracted from the texts, which are chosen carefully to suit the average Greek candidate in terms of age and literacy. However, there is consideration given to the fact that there are both younger and older candidates taking part in the KPG exams[9]; as a result, texts are chosen and tasks are developed in a way that some are more conducive to the adult candidates’ cultural experiences and literacies, and others to those of younger candidates. Most importantly, however, the source texts are selected with a view to being a rich source of information which stimulates talk.

For those readers who are not familiar with the KPG examination battery, it should be mentioned that the speaking test (Module 4 of the exams in all languages and levels) involves the use of a Candidate Booklet and an Examiner Pack. The Candidate Booklet is an illustrated publication in full colour, and each titled page or page section, which contains texts on a single theme, is designed up to look like a page out of a magazine, a brochure, a website, etc. (Appendices 7a and 7b). For each text/theme, several tasks are developed and they are included in the Examiner Pack, which is not available to the candidates. The examiner may choose which task to assign to which candidate (for the B level test), or to which pair of candidates (for the C level test).

For the KPG speaking test, two examiners and two candidates are present in the exam room. One of the two examiners assumes the role of Interlocutor and assigns the tasks to candidates orally. For the B1 and B2 speaking tests, different tasks are assigned to each candidate, who is asked to address either only the examiner-interlocutor or everyone in the room. For the C1 level speaking test, the two candidates in the room are assigned the same task and are asked to interact with one another and exchange information from a Greek source text.

4.2.1. B1 and B2 level speaking task completion expectations

According to the published specifications, B level candidates are expected to use the target language to:

• relay selected information from L1 texts, or

• express the gist of L1 texts, or

• talk about an issue discussed in an L1 text.

Task description indicates that the source texts are on issues of everyday concern, such as health and diet, the environment and saving energy, travel and holiday, entertainment, home safety, work and education, public holidays and celebrations, means of transport.

The B level mediation test requires one-sided talk, which means that the source text must provide enough information/ideas to allow each candidate in the room to speak in the target language for about 2½ -3 minutes, performing the task assigned.

Analysis of B1 and B2 level speaking mediation tasks (SMT) reveals that each task, which is linked to a single page/theme/text, commonly points the candidate to a different part/section of the text. Each task has a different communicative purpose, a different addressee, and it often concerns a different person, while it may also set up a different situational context. Consider, for example, the two out of the four B1 level tasks for a page entitled Fruit in children’s diet from the speaking test of the November 2007 exam in English.

SMT 01:   Imagine I am your Belgian friend and my 14-year-old son never eats fruit. Read the text and give me some advice on what I should do to change his mind.

SMT 02:   Imagine I am your Swedish friend and my children do not like eating fresh fruit. Read the text and suggest ways to add fruit to their diet.

It is interesting to note that the person to be addressed in both tasks is the examiner in the role of a friend, and that the situational context is more or less the same. However, the language function to be performed in each case is somewhat different; that is, in SMT 01 the candidate is to give advice about what to do, and in SMT 02 to suggest ways of doing something differently. Also, each task concerns different parties and undertakings, which means that the attention of the candidate is directed to a different part of the text; in SMT 01 the candidate is to find information useful for getting a teenager to change his mind about eating fruit, while in SMT 02 the candidate is to find information useful for a parent interested in adding fruit to his/her child’s diet. 

Now consider two more B1 level tasks from a different page of the Candidate Booklet, entitled Sea and safety, from the same oral mediation test as above.

SMT 03:   Your Austrian friend and her family are going to spend their summer holidays on a Greek island. Read the text and tell her what she should be extra careful about when she takes her kids to the beach.

SMT 04:   You are the leader of an international camp for young children. Read the text and give advice to the young children on how to swim safely.

The person(s) whom the candidates are to address in both tasks SMT 03 and SMT 04 is not the examiner or other people in the exam room, but a third party they are to imagine that they are speaking to – their foreign friend, who is a parent. The same is true of SMT 04. Again, candidates are not asked to address the examiner but a third party, this time, a group of children. There are, of course, expectations that the candidates’ talk will be appropriate for this situational context, which is different from that in SMT 03.

Though the B1 level and the B2 level tasks have much in common, there are certain differences, which mainly have to do with the topic of the source text and the type of discourse to be produced. At B2 level, it is often semi-formal, impersonal, or requires the use of some specialized vocabulary. For example, see below the B2 level mediation tasks linked to a page entitled Recycling electrical goods, from the English exam of the same period as the B1 tasks above. It is not only the topic that calls for a more specialized vocabulary in the source text, but also each task that originates from this text. The discourse and register the candidate is expected to use when performing SMT 05 is quite different from that used for SMT 01-04 because the situational context requires the use of impersonal language, as in the case of SMT 06, where the candidate is asked to give advice to a friend not on a personal matter that has to do with human behaviour, relations, etc., but on acting in an environmentally friendly way.

SMT 05: Imagine you have been asked to present in English a new recycling programme for electrical goods. Using information from Text 1, tell us what points you will include in your presentation.

SMT 06:   Imagine your German friend Ingrid wants to get rid of her old computer. Using information from Text 2, give her some advice on how to recycle it.

The situational context is similar in the B2 level examples below, included here to explain the differences – even though they are subtle – with B1 level mediation. All four tasks are from the same test (November 2007 speaking test) as tasks 05 and 06, and require the production of some specialized vocabulary because of the topic. SMT 07 and 08 are associated with the Greek text on a page entitled Archery, and SMT 09 and 10 from a source text on a page entitled A successful job interview. In addition, the communicative act to be performed in each case is somewhat impersonal, requiring a semi-formal style of talk, i.e. to tell someone about the benefits of something (SMT 07), to tell others what points will be included in a talk about archery (SMT 08), to give someone advice on a successful job interview (SMT 09), and to tell others what advice would be offered to young people looking for a job (SMT 10).

SMT 07:   Imagine your Dutch friend Marcel wants to take up a new hobby. Read the text about archery and inform him about the benefits of the sport and the necessary equipment.

SMT 08:   Imagine you are responsible for the local sports centre. You’re going to give a talk in English about archery, a new sport to be offered at the centre. Using information from the text, tell us what points you will include in your talk.

SMT 09:   Imagine your Portuguese friend Paolo is very anxious because he has got an important job interview next week. Read tips 1-4 and give him some advice on how to be successful at his interview.

SMT 10:  Imagine you’re going to give a talk in English to young people who have just started looking for a job. Read tips 1-4 and tell us what pieces of advice for a successful job interview you will include in your talk.

4.2.2. C1 level speaking task completion expectations

The main difference between the B level and the C level oral mediation tasks is that the latter involve interaction and not merely one-sided talk. This means that candidates are required to initiate and sustain a conversation for about 10 minutes, and during that time to provide their interlocutor with information and converse with her/him in order to reach a decision, resolve a problem, arrive at a conclusion, etc., all of which demands negotiation of meanings, ideas and factual information.

The C1 level Candidate Booklet is, in fact, organized in a way that is suitable for this interactive mediation activity: the first half of the Booklet contains texts for Candidate A, and the second half texts for Candidate B. Texts A and B contain different chunks of information, but they are both on exactly the same issue and usually from the same source, as the examples in Appendices 7a and 7b, with texts giving rise to tasks SMT 11 and 12. Each candidate is instructed to look at her/his own text, on a different page, but the mediation task they are both assigned is one such as the following:  

SMT 11:  Imagine that you and your partner are planning a trip for the Christmas and New Year holidays. Exchange information from your texts and decide about the most interesting New Year’s celebration. This decision will also help you decide which country you might visit.

Alternatively, with another couple of candidates, the task originating from these texts is:

SMT 12:  Exchange information from your texts with your partner and together decide on the two most unusual customs to write about for the special Christmas issue of your school/local newspaper/magazine.

This C1 activity, with two different speaking mediation tasks, involves candidates in interaction and negotiation for which they must have the competences and skills presented earlier (Table 1), so they can go through processes also presented earlier (Table 2). The ultimate goal in each instance, where candidates must consider different circumstances, is for them to reach a common decision.

Similarly, tasks SMT 13 and SMT 14 below ask candidates to make a common decision. The texts upon which the tasks are based are on the issue of Anger control. Each candidate is instructed to look at her/his own page/text and to:

SMT 13:  Exchange information from your texts and together decide on the two most important things people should avoid doing when they are angry and on the two most effective ways to deal with anger.

Or, alternatively, with another set of candidates:

SMT 14:  Exchange information from your texts and together decide what pieces of advice you would give to a newlywed couple who have just had their first argument.

Likewise, in other C1 speaking mediation tasks, candidates are most commonly called upon to make a common decision, i.e.:

SMT 15:  Read brief book reviews and exchange information with your partner. Then together a) decide which two are the most likely to become best sellers, or b)which four books you should buy for your local/ school library.

SMT 16:   Read pop magazine article tips which might help you when shopping, and with your partner decide a) which two tips are the easiest to follow, or b) which two tips are mainly addressed to women and which to men.

SMT 17:  Exchange information from your texts, and with your partner decide about a) the two most suitable dogs for a family with children, or b) the two most suitable dogs for a woman living on her own.

5.     Mediation task difficulty

It is clear from the examples and earlier discussion that both lifeworld and test mediation tasks are challenging. In the KPG exams, mediation entails comprehension in L1 and production in L2 and, as many studies have argued,[10] language and code switching is demanding in any case, but it is even more so when it means encoding the message in the foreign language. Yet, language or code switching is not the only challenge that mediation poses. The preceding sections of this paper have shown that mediation tasks are also cognitively demanding. This makes it essential for activity and task developers to become increasingly aware of what is involved in each instance of mediation, so they can control the cognitive load and linguistic demands of each activity,[11] according to the level of language proficiency that is tested, as well as according to the age and literacy of the candidates to whom the exams are addressed.

The statement about consideration of the above variables (level, age and literacy) having been made, I must also add that the three are not necessarily dependent upon one another. There is no direct correlation, for example, between proficiency level and cognitive load, which means that there may be a more cognitively demanding mediation task for lower proficiency level candidates and vice-versa. However, cognitive load is contingent upon age and literacy, and this may be linked to higher level testing.  

There are also strong indications in the preceding sections that the demands and the linguistic requirements of each instance of mediation are both task specific and source text specific. This is why when a mediation activity at each level of KPG testing were presented earlier, the type of texts used as sources of information and the types of tasks set were discussed on the basis of test level. However, these are important considerations which need to be explicitly stated and further clarified – perhaps with examples.

Let us begin with an example about task specific difficulty and ask you, the reader, to consider two different tasks on the same topic, which is also a variable that has to do with activity complexity. The activity topic is Illegal immigration, and the two tasks with varying degree of difficulty are the following:

A.    Task 01 asks candidates to gather information about the issue from a variety of source texts and to present the pros and cons of the social inclusion of immigrants.

B.    Task 02 asks candidates to read an editorial about illegal immigration and say what they think the author’s opinion about the social integration of immigrants is.

Undoubtedly, the cognitive load of Task 01 is greater than that of Task 02. Both the cognitive load and the linguistic demands, on the other hand, are very much dependent on the source texts – how many there are for Task 1, what discourse they articulate (e.g. media or social science discourse), and how the source texts of Task 1 in particular are organized.

To illustrate further how mediation requirements are strongly related to the source text – i.e. they are source text specific – we should turn attention back to one of the mediation tasks discussed earlier, WMT 01. The language of the text on which this task is based is too difficult for B1 level candidates to translate. Therefore, if they want to be able to respond to the task at hand, they need to understand the information conveyed in L1, interpret it in relation to what is asked of them, and use their interpretation to give tips about a healthy diet to their friend in an e-mail message. Below are two of the Greek texts and sample responses to the task from candidates:

Candidates’ responses

1. You must drink a lot of water. It’s good for you and you don’t get headaches

2. Drink a lot of water every day not to have headaches and feel tired.

3. Did you know that our head is 80-85% water? Drink lots of water. You should not get dehydrated.

Candidates’ responses

1. You shouldn’t believe all that you hear. Some people say that it’s bad to drink water and stuff with our food. That’s not true.

2. It’s a lie that we should not drink liquids during and for a long time after we eat.

3. You should drink water or other things with your food. It helps you to digest better.

What we can see in the above responses is that candidates were able to function as competent mediators and, by using specific communication strategies, they were perfectly able to deal with the task at hand. In fact, this seems to be the case with mediation performance in the exams of all levels in English. Descriptive statistical analysis we have been carrying out at the RCeL shows that, despite EFL teachers’ worries about mediation tasks being ‘too difficult,’ there is no statistically significant deviation in the scores that KPG candidates’ are assigned for the two activities. What is even more interesting as we look at the results of our analysis is that sometimes candidates’ scores in the mediation task (Activity 2) is higher than in the writing task (Activity 1).

6.     Mediation performance

Mediation has a crucial role in today’s world of multicultural contexts that increasingly demand cultural and linguistic negotiation for successful participation in the communication process, ‘producing oral or written texts in which forms and words are manipulated to extend further understanding across cultures’ (Valero-Garcés 2009). This is the main reason that I have decided to concentrate on this important topic. In addition, I also feel the need to give the mediation component of the KPG exams the attention it deserves. Therefore, I am publishing on the subject (Dendrinos and Stathopoulou, 2010, Dendrinos 2006), working with postgraduate students (Stathopoulou 2008),[12] (Stathopoulou 2009, Voidakos 2007),[13] and with PhDs researching oral and written mediation (Stathopoulou, 2013a). Furthermore, I am directing a number of large-scale projects under way at the RCeL, which progressively provide a more accurate account of mediation performance and performance expectations. The work carried out in this fascinating area permits us to talk about mediation in an informed way and to identify the problem areas to be dealt with when teaching or coaching for mediation practices.

6.1 Written mediation performance

The major data bank we have developed at the RCeL, with its corpus of thousands of KPG candidate scripts classified according to the mark they have received by trained script raters, has opened the possibility to conduct systematic research on the written mediation performance of candidates.

Stathopoulou (2013a), having recently completed her PhD thesis on the topic gives us insights into what mediation entails and what types of written mediation strategies lead to the achievement of a given communicative purpose. Drawing data from the KPG Task Repository and the KPG English Corpus, she examined KPG test tasks and candidates’ scripts in order to create a levelled mediation task typology and an Inventory of Written Mediation Strategies, which may provide the basis for the creation of mediation-specific levelled descriptors (see also Stathopoulou 2013b). These will in turn make reliable assessment of the mediation performance.

One of the MA studies (Voidakos 2007), which set out to describe B2 and C1 level mediation performance by means of analyzing 283 scripts produced by KPG candidates, has provided insights which suggest that candidates’ have the following common problems[14] in  written mediation performance:

• They tend to include as much information as possible from the source text, rather than selecting only relevant-to-the-context pieces of information. This test-taking strategy does not favour candidates, not only because they are penalized when their scripts are longer than they are supposed to be, but also because they are more likely to end up with more instances of inaccurate or inappropriate English language use.

• They pay close attention to producing in L2 structures and forms which correspond to L1, as they believe that this is what is expected of them.

• They tend to pay little attention to the discourse, genre, register and style of the text to be produced, which may be of a different text type than the source text. Instead, they tend to produce the same text type and this results in inappropriate language use, since the language features of a text are dictated by its genre.

In the aforementioned study, Voidakos finds that there are indications that C1 level candidates perform slightly better in mediation tasks than B2 candidates. To qualify this claim, I should say that what she means is that C1 candidates seem to make more frequent and effective use of the necessary mediation test-taking strategies.

This claim raised a noteworthy question, for which there was no conclusive evidence in the study: Are more proficient L2 users better mediators? My guess at that point was that this is true only when there are other factors at play too, such as age (and therefore the cognitive capacity which develops with age), literacy level in both the source and the target language, and task-specific skills.

This assumption was an issue of concern in a later study, along with another hypothesis, born out of our ongoing examination of candidates’ mediation scripts. The hypothesis was that mediation performance is contingent not only on candidate, but also on task variables. In other words, mediation performance may be ‘better’ or ‘worse’ depending on who is doing it (her/his age, literacy, knowledge, skills, etc. in both languages) but also depending what s/he has to do.

Aspects of what I had come to suspect have now been investigated by Stathopoulou (2009).[15] Her study, exclusively on B2 level written mediation performance, was carried out with a view to supporting the claim that when mediating, the source text necessarily regulates the target text, and the visible traces may vary from weak to strong, depending on a series of factors. A total of 240 B2 level scripts were analyzed for the purposes of the study, which ultimately offers interesting results and conclusions about KPG mediation performance.

As in the case of the previous study, the scripts analyzed for this dissertation had also been randomly selected from the RCeL script corpus: half of the selected scripts were from the ‘fully satisfactory’ category, and the other half from the ‘moderately satisfactory.’ The two categories of scripts were compared, showing that fully satisfactory scripts were less regulated by the source text than the moderately satisfactory, and that the hybrid formations in fully satisfactory scripts are perfectly ‘acceptable’ in English; that is, as Stathopoulou puts it, they contain ‘fairly successful code meshing structures that create no problem of intelligibility to the reader.’ Other interesting findings in her study are the following:

• Hybrid formations on the borderline of being acceptable in English, in both fully or moderately satisfactory scripts, constitute violations on the level of pragmatics rather than the level of semantics or formal grammar.

• The scripts contain a number of strongly regulated constructions which are unsuccessful in relaying the message(s) of the source text.

• In moderately satisfactory scripts, one notes a tendency towards word-for-word translation of complete sentences, whereas in fully satisfactory scripts, this tendency is on the lexical rather than the sentence level. Moreover, moderately satisfactory scripts more frequently than fully satisfactory scripts transfer language elements from one language to the other without adjusting them so that they are appropriate for the linguistic and social context. Thus, they violate English word order restrictions, make inappropriate use of prepositions and other words, as well as language structures, such as modality.

• Fully satisfactory scripts are not as strongly regulated since the source text information is paraphrased, and when they do contain hybrid formations, these are considered acceptable in English; that is, successful code meshing language items.

• Finally, in moderately satisfactory scripts, information seems to have been selected on the basis of what information was easily transferable from one language to the other, rather than on the basis of what information was relevant to the communicative demands of the task. Stathopoulou observes that ‘any ideas that candidates were unable to relay, probably due to limited linguistic resources, were omitted.’

6.2 Speaking mediation performance

Greek privacy protection laws do not allow us to record the KPG speaking test on DVD; therefore, we have no access to digital information, on the basis of which we can study oral mediation performance. However, we do have valuable information from the feedback forms that our examiners complete after every exam administration, while we also have the detailed accounts of candidates’ performance from trained observers, whose job is to assess the speaking test tasks, the procedure, and the examiners. Analysis of this data is under way, and it will soon be published. Presently, I include below some of the remarks that have been recorded:

• Younger and linguistically less competent C1 candidates make serious attempts to translate the source text rather than to relay selected information.

• B1 and B2 level candidates who are obviously unprepared to take on the role of mediator seem to feel awkward and less confident about what it is they are to do.

• There are instances when B1 and B2 level candidates refrain from making any use of information in the source text. They simply speak on the theme of the L1 text. When their oral English is good, examiners are divided as to what to do since they have been given no thorough instructions on how to deal with this matter.

• In C1 level oral mediation, when candidates are required to interact and exchange information from their source texts in order to reach a common decision, they sometimes do so by using only a few points from the text, and then draw from their own experiences.

• Both candidates and teachers preparing them need to familiarize themselves further with the nature of mediation activities. KPG examiners also need more training on this new aspect of assessment.

7.     Training for mediation

If it is indeed true that mediation tasks are very difficult, the first question that comes to mind is whether it is fair – i.e. if it is ethical – to test it. In asking this question, it is only right to reveal that the inclusion of mediation in the test papers of the KPG exams has not been without reaction in the Greek language teaching scene – though, admittedly, those that have been positioned most strongly against it are L1 English speakers who have found jobs teaching English in Greece. These people, many of whom have not been trained to teach a foreign language or to teach at all, and who are nevertheless highly regarded simply because they are native speakers (NSs) of the language they are teaching, benefit from the exclusion of the learners’ mother tongue from the classroom, from the teaching materials, and so on.[16] There are also a number of Greek EFL teachers who react as strongly, unaware of what this reaction of theirs conceals. However,  the largest percentage of Greek EFL teachers are not negative; they are sceptical because candidates are not really prepared to perform as mediators since it is not part of the mainstream TEFL materials or practices. And, of course, it is only natural that teachers worry about an element in a suite of national exams that is not legitimated by the established international exams trade. If it were, it might more easily have been considered a valid test element and its inclusion would immediately seem more logical and justified. As things now stand, the legitimisation of mediation in the exams is through its endorsement via the CEFR.

The reason that mediation and, in general, the use of L1 is absent from TEFL textbooks and other materials produced by the international or multinational publishing industry is profit. If it were to localise textbooks and other materials, companies would not make as big a profit as they do now when they make one product for international use. And, if something is not in the textbook, it is not legitimate curricular knowledge – especially for those who ‘teach to the book’ –which is true of the largest percent of foreign language teachers in Greece and elsewhere (cf. Dendrinos 1992). Hence, the omission of mediation practice from foreign language textbooks is a basic factor for its prohibition from the foreign language programme. All this means, of course, that foreign language learners are not systematically trained for their role as mediators, though they frequently practice mediation in their daily lives.

This reality provokes us to amend the question above: Is it fair not to help foreign users learn to communicate in a way that is valuable in their daily lives, just because the dominant foreign language paradigm triggered by economic and symbolic profit does not promote it? The answer is No. It is not fair or ethical to refrain from training learners to exploit language in ways which will be of practical use to them.

But, how does one do that? Greek EFL teachers, who are increasingly turning attention to mediation, tell us that they don’t know what materials to use in order to ‘teach’ mediation – a problem which is easily solved since nowadays there are all sorts of authentic materials suitable for classroom mediation practice available on the internet. The more complicated question has to do with ways of teaching mediation. The most common EFL teacher question is: ‘What is it exactly that we should teach?’ ‘Do we teach,’ they wonder, ‘aiming at the development of all the knowledge and skills mediation seems to require?’

The response to this question is that this would not be possible; what is possible and quite feasible is the creation of conditions so that learners can progressively resort to the language awareness they have developed in both languages, to their socio-cultural knowledge and experiences, to their communication skills, and to their cognitive capacities in communicative contexts that require mediation.

Learners could be gradually trained to deal with both languages in ways that will help the mediation cause:[17]

• Start training at lower levels with the types of tasks that are used in the A level reading and listening comprehension tests of the KPG exams.
• Introduce comprehension in L2 and production in L1 rather than the other way around during first stages of mediation training.
• Progressively introduce practice on comparing articulations of socially purposeful meanings in the two languages so that learners become increasingly aware that similar meanings are verbalized with different forms and structures in the two languages.[18]

• Progressively, introduce practice whereby learners are asked:

• to express in their own words bits of information contained first in L2 and then in L1 texts

• to relay in their own words spoken and written messages 

• to express in one language the gist of a text in the other language

• to select information suitable to a context of situation from a text in one language and to relay it in the other

• to progressively conduct controlled, guided and free writing/ speaking practice in one language, using as cues texts in the other.

All the techniques above, it seems to me, would be easily employable by any foreign language teacher who, of course, is proficient in both the learners’ mother tongue and the target language. My experience also tells me that the great majority of educated-for-their-job teachers can also be quite resourceful and use the aforementioned teaching techniques plus many more if they are interested enough in the object of knowledge. And this brings us to the issue of teacher interest, or the interest of the curriculum designer or the planner of the foreign language programme that teachers follow. Reasons for teachers’ ambiguity or even reluctance to train for mediation have already been mentioned. What has not been openly stated is that there is a very definite backwash effect by the inclusion of mediation in the KPG, as it always happens with high stakes exams (cf. Tsangari 2006, Shohamy 2001), and interest in it is persistently growing in Greece. But, then one might rightly wonder: Does the end justify the means? Is it just to test this aspect of language use if it is unjust to the candidates for whom the exams are intended?

The question immediately above was at the heart of our investigation during the planning and the piloting phases of the KPG exams. Our early findings were not discouraging; quite the contrary, in fact. Though it became evident in our initial stages of research related to mediation that candidates who were in some way prepared or just coached for this aspect of the exam did better than those who had never before seen or practised doing a mediation task in the class context, still the differences of scores they received in other types of L2 production were not significant enough to warrant a great deal of apprehension.[19]

For those of us working for the KPG, the inclusion of mediation in these high stakes exams has raised several crucial questions which have instigated organized research, now being carried out at the RCeL. The ten most common questions that have been articulated during the five years that we have been developing the exams are the following, in random order:

1. What exactly does mediation involve?

2. When people mediate, what is it exactly that they do?

3. When the aim is assessing one’s ability to mediate, what exactly do you measure, and what criteria of assessment do you use?

4. What are the demands and expectations for mediation performance at different levels of proficiency?

5. What types of materials should be used to train learners and to coach candidates for mediation?

6. Which types of tasks should be used in a foreign language programme to serve which goals?

7. What types of tasks should be used in tests so as to be able to discern the different levels of proficiency?

8. What are the main problems that candidates encounter when mediating and how do these problems surface in their talk and scripts?

9. Which are the evaluation criteria for mediation performance at different levels of proficiency?

10. What can be done to prepare candidates to take the role of mediator successfully in a test situation?

The various questions immediately above, some of which have been addressed directly or indirectly in this paper, are all linked with concerns about the ‘teachability’ of mediation. Data from related research being carried out will yield helpful information and be presented in other papers. One of these papers, by Maria Stathopoulou, is included in this issue of Directions and presents findings from one of the larger scale projects investigating suitable test-taking strategies for the KPG exams. Stathopoulou (ibid) focuses on written mediation strategies, and her paper provides evidence that mediation test-taking strategies are indeed ‘teachable’ and that training for mediation performance leads to positive results.

In conclusion, it is important to stress that with training and coaching, foreign language learners and exam candidates are bound to have better results, as with all types of other language practices. With mediation, it is even more important to coach candidates so that they learn to use skills they have developed together with test-taking strategies in order to do something they have never before had a chance to do as part of their language training, since mediation is usually not ‘taught’ or tested in foreign language classes.

The reasons for the exclusion of mediation in mainstream foreign language programmes have been discussed in this paper. It has been maintained that, despite the launching of mediation through the CEFR, mediation training did not find fertile ground in the international business of language teaching and testing, mainly on account of the conventional monolingual ideology on which it is based and reproduced because of economic interests.[20] Therefore, foreign language teachers in Greece and elsewhere who have adopted the dominant paradigm of foreign language teaching and learning are still largely unfamiliar with the concept and techniques conducive to mediation. Hopefully, this paper will be a step forward to thinking about this significant topic, which is immersed in the cultural politics of teaching and testing.

References

Dendrinos, B. (1988). Using the L1 in the classroom. Views on EFL Teaching and Learning. Athens: The University of Athens Publications.

Dendrinos, B. (1992). The EFL Textbook and Ideology. Athens: N.C. Grivas Publications.

Dendrinos, B. (1995). Foreign language learning and the development of sociolonguistic skills to express translatable experiences. Actes du Collogue International Enseignement des Langues et Comprehension Europeene (pp. 55-64). Athens: University of Athens & Association pour le Rayonnement des Langues Europeenes.

Dendrinos, B. (1999). The Conflictual Subjectivity of the EFLPpractitioner. In A.F. Christidis (Ed.), 'Strong’ and 'Weak’ Languages in the European Union: Aspects of Hegemony. Vol. 2. Thessaloniki: Centre for the Greek Language, pp. 711-727.

Dendrinos, B. (2001). Linguoracism in European Foreign Language Education Discourse. In M. Reisigl & R. Wodak (Eds.), The Semiotics of Racism: Approaches in Critical Discourse Analysis (pp. 177-198). Vienna: Passagen Verlag.

Dendrinos, B. (2001). Plenary talk at the International Conference on Non-native speaking teachers in foreign language teaching, organised by the University of Lleida in Catalunya, Spain. Title of plenary address: The pedagogic discourse of ELT and the discursive construction of the NNS’ professional value.

Dendrinos, B. (2003). The Hegemony of English (in collaboration with Dοnaldο Macedo & Panagiota Gounari). Boulder, Colorado: Paradigm Publishers.

Dendrinos, B. (2004). Multilingual literacy in the EU: Alternative discourse in foreign language education programmes, In B. Dendrinos and B. Mitsikopoulou (Eds.), Policies of Linguistic Pluralism and the Teaching of Languages in Europe (pp. 60-70). Athens: Metehmio publications and University of Athens.

Dendrinos, B. (2006). Mediation in communication, language teaching and testing. Journal of Applied Linguistics, 22, 9-35.

Dendrinos, B. (2007a). Mediating between cultures and languages. Plenary talk at the TESOL Greece Annual Convention, Athens, March 2007.

Dendrinos, B. (2007b). Mediation practices and foreign language literacy. Plenary talk at the Greek Applied Linguistics Association Conference. Thessaloniki, December 2007.

Dendrinos, B. (2008). Los discursos que moldean la subjectividad periferica del docente de ingles como lengua extrangera. In Educación y Pedagogía, Vol. XX (pp. 65-76) May-August, 2008 (Journal of the University of Antioquia, Colombia.

Dendrinos, B. & Stathopoulou, M. (2010). Mediation activities: Cross-Language Communication Performance. ELT News, 249(12):  http://rcel.enl.uoa.gr/kpg/kpgcorner_may2010.htm

Heredia, R. R., & Altarriba, J. (2001). Bilingual language mixing: Why do bilinguals code-switch? Current Directions in Psychological Science, 10(5), 164-168.

Shohamy, E. (2001). The Power of Tests: A Critical Perspective of the Uses of Language Tests. London: Longman.

Stathopoulou, M. (2008). Performing written mediation: Strategies employed by KPG B1 level candidates. Unpublished paper submitted for credit in the Language Learning Theories course in the Applied Linguistics MA Programme, Faculty of English Studies, University of Athens.

Stathopoulou, M. (2009). Written mediation in the KPG exams: Source text regulation resulting in hybrid formations. Unpublished dissertation submitted for the MA degree in the Applied Linguistics Postgraduate Programme, Faculty of English Studies, University of Athens.

Stathopoulou, Maria. (2013a). Task dependent interlinguistic mediation performance as translanguaging practice: The use to KPG data for an empirically based study. (Doctoral thesis). Faculty of English Language and Literature, National and Kapodistrian University of Athens.

Stathopoulou, M. (2013b). Investigating Mediation as Translanguaging Practice in a Testing Context: Towards the Development of Levelled Mediation Descriptors. In the Proceedings of the International Conference Language Testing in Europe: Time for a New Framework? University of Antwerp, Belgium, May 2013.

Tay, M. W. J. (1989). Code-switching and code-mixing as a communicative strategy in multilingual discourse. World Englishes, 8(3), 407-417.

Tsangari, K. (2006). Investigating the washback effect of a high stakes EFL exam in the Greek context: Participants’ perceptions, material design and classroom applications. Unpublished dissertation submitted for the degree of Doctor of Philosophy, Lancaster University.

Valero-Garcés, C. (2009). Mediation as translation or translation as mediation? Widening the translator's role in a new multicultural society.  Retrieved from http://www.babelport.com/articles/32, 19 August 2009.

Voidakos, I. (2007). What mediators do: Analysing KPG candidates’ actual performance in written mediation tasks. Unpublished dissertation submitted for the MA degree in the Applied Linguistics Postgraduate Programme, Faculty of English Studies, University of Athens.

Wang, L. (2003). Switching to first language among writers with differing second-language proficiency. Journal of Second Language Writing, 12, 347–375.

Williams, S., & Hammarberg, B. (1998). Language switches in L3 Production: Implications for a polyglot speaking model. Applied Linguistics, 19(3), 295-333.

Woodall, B. R. (2002). Language switching: using the first language while writing in a second language. Journal of Second Language Writing, 11, 7-28.

Endnotes

Appendix 1
Appendix 2
Appendix 3
Appendix 4
Appendix 5
Appendix 6
Appendix 7