The intermediary role of reformulation in learning French as a foreign language: The case of students at the University of Ghana

Résumé . Le rôle intermédiaire de la reformulation dans l’apprentissage du français langue étrangère : le cas des étudiants de l’université du Ghana. Cet article vise une réflexion sur le rôle intermédiaire de la reformulation dans l’apprentissage d’une langue étrangère par des étudiants de français à l’Université du Ghana. Notre objectif est de montrer que, après avoir écouté un extrait de texte, les apprenants sont capables de reformuler les différentes séquences du texte avec des degrés différents de difficulté. Nous essayons de savoir ce qui est reformulé, comment c’est reformulé et le lien avec l’apprentissage implicite ou explicite d’une langue. La démarche d’analyse de cette étude exploratoire consiste en une comparaison entre les séquences produites par les étudiants et les séquences du texte source. Nous examinons les séquences à partir de critères sémantiques et syntaxiques pour savoir quel type de reformulation est produit et si les différences perçues dépendent des niveaux des apprenants. Notre observation est que les apprenants reformulent beaucoup sans qu'il n'y ait de répétition. Cependant, il y a une différence dans la complexité de la reformulation selon les niveaux. Abstract. The objective of this article is to provide an analytical perspective on the intermediary role of reformulation in foreign language learning by students of the French language at the University of Ghana. Our aim is to demonstrate that after having listened to an extract of a text, the learners are able to reformulate the different sections of the text with varying degrees of difficulty. We hope to find out what is reformulated and how it is reformulated and the link between implicit and explicit language learning. The method of analysis of this exploratory study consists in comparing the sequences produced by the students to the sequences of the source text. We examine the sequences based on semantic and syntactic criteria to know what kind of reformulation is produced and if the perceived differences are based on the level of the learners. Our observation is that the learners produce a lot of reformulation and there is no case of direct repetition. However, there are differences in the complexity of the reformulation that is based on the level of the student.


Introduction
Learning a foreign language involves a complex combination of both acquisition and learning. Paradis (2009) distinguishes between these two terms with acquisition referring to the implicit items and processes and learning referring to explicit or conscious items and processes. Whereas for native speakers, learning seems to be implicit, research has shown that foreign language learners need some explicit processing in learning the new language. (Ellis, 2008). Moreover, learners of a foreign language are often confronted with the challenge of how much of a language they have imbibed with this normally being measured through their performance. In expressing the difference between L1 and L2 acquisition by stating that: The acquisition of L1 grammar is implicit and is extracted from experience of usage rather than from explicit rules-simple exposure to normal linguistic input suffices and no explicit instruction is needed. Adult acquisition of second language (L2) is a different matter in that what can be acquired implicitly from communicative contexts is typically quite limited in comparison to native speaker norms, and adult attainment of L2 accuracy usually requires additional resources of explicit learning. The various roles of consciousness in second language acquisition (SLA) include: the learner noticing negative evidence; their attending to language form, their perception focused by social scaffolding or explicit instruction; their voluntary use of pedagogical grammatical descriptions and analogical reasoning; their reflective induction of metalinguistic insights about language; and their consciously guided practice which results, eventually, in unconscious, automatized skill. (Ellis 2008:1) This study focuses on language knowledge and languages use by learners of French as a foreign language in the University of Ghana. We seek to determine the interface between their linguistic and extra linguistic knowledge in French (their L2) and even in English (their L1) and how this translates into their oral production of the French language. This paper has a two-fold objective. First of all, we seek to observe the types of reformulation outlined by Martinot (1994Martinot ( , 2000Martinot ( , 2015 that are expressed in the productions of students through the following means: i) The substitution of lexical items, ii) reconstruction from direct speech to reported speech, and iii) the manipulation of tenses. Secondly, we hope to explore how the types of reformulation observed reveals the knowledge and use of the French language by students of French as a foreign language in the University of Ghana.
premise of Harris (1988Harris ( , 2007 that language is used to explain language. By extension, it can be affirmed that it is possible to have semantically similar utterances even if they are structurally different since you will be trying to explain what a statement means usually by using 'your own words'. Reformulation was first described in Gülich (1983) in considering the analysis of conversation. For the author, reformulation makes reference to the various linguistic actions of rephrasing and paraphrasing which characterize verbal interaction. Reformulation has also been studied by Clark (1985), Kail (1994), andHickman (1997). Their research was based on a psycholinguistic viewpoint with the aim of testing learners' comprehension in relation to their production. These reformulatory actions have been identified and classed by specialists in the field of the teaching of the mother tongue. Indeed, Martinot (2000) describes reformulation as being a common and spontaneous activity of any speaker who is producing language. It is not necessarily with the aim of giving explanations and is not limited to conversations. The activity of reformulation is a linguistic enterprise which can be undertaken either by a group of people who progressively adjust their words to produce a particular meaning, or it can be practiced by an individualsuch as an author searching for the right words to use to convey a certain effect.
When used in the domain of the teaching of foreign languages, this term usually refers to the corrective action of the teacher as he tries to reformulate the production of the foreign language learner into more acceptable renditions.

Definition of reformulation
For the purposes of this article, we will use the definition of Martinot (2015:3) in which she refers to and maintains her definition from 1994. This is defined as : « …tout processus de reprise d'un énoncé antérieur qui maintient, dans l'énoncé reformulé, une partie invariante à laquelle s'articule le reste de l'énoncé, partie variante par rapport à l'énoncé source… ».
This definition can be translated as follows : « … any process of restating a previous statement which maintains, in the reformulated statement, an invariable part to which the rest of the statement which could be different from the source statement is attached ... » Two characteristics of reformulation that are spelled out in the definition by Martinot are that firstly, the reformulated extract must have its origins in a previous statement. Secondly, there must be similarity of content and/or form for it to be called a reformulation of the source statement. According to Martinot (2015), this definition makes room for paraphrastic reformulations, non paraphrastic reformulations, and repetitive reformulations. It also enables one to situate the linguistic level of the invariable part of the statementwhether it is lexical, syntactic or semanticand also the linguistic level of the modified part that is introduced into the reformulated statement.
For example, a reformulated statement can have the same lexical items and the same construction but the meaning will not be the same: Secondly, a reformulated statement can have the same lexical items and the same meaning but have a different construction. These are transformations.
3) SS: The teacher beat the student 4) RS: The student was beaten by the teacher Thirdly, a reformulated statement can have the same lexical items and the same meaning but have a different structures. This is the case for restructured statements. 5) SS: John loaded the books into the car. 6) RS: John loaded the car with books.
Finally, we have the scenario where a reformulated sentence can have the same meaning but have different lexical items and a different construction. These are considered semantic paraphrases.
7) The woman feeds her baby. 8) The mother gives food to her new-born child.
On the whole, a learner receives input (oral or written) which he reformulates and produces as output (oral or written). The process between inputting the information and bringing it out as an output reveals the capacity of the learner to carry out the various types of reformulation cited above and to produce more complex or simplified sentences from the source statement that they heard or read.
Even though the existing work on Reformulation by Martinot (1994Martinot ( , 1996Martinot ( , 2000Martinot ( , 2013Martinot ( , 2015 has often focused on first language acquisition, we observe that the main thrust and the ideas can also be applied to foreign language learning since the learners also have to deal with prior information and produce language based on this prior information. There is therefore a processing of input that leads to an output. Moreover, reformulation is important in foreign language acquisition because it demonstrates the capacity of learners to carry out syntactic transformations and to simplify or make their utterances more complex. Our research reveals very scanty literature in this domain with the only relevant one we found being Gerolimich (forthcoming) who carried out a study on reformulation into and from French by Italian native speakers. The aim of her study was to investigate if reformulation can serve as a means of understanding the process of acquiring a foreign language. In her article, the author makes a distinction between mother tongue acquisition and foreign language learning in that usually the learner of a foreign language usually has acquired the linguistic system of the mother tongue and has also reached a relatively advanced stage of cognitive developmentespecially in the case of adolescents and adult learners. Due to this distinction, someone who is learning a foreign language is likely to employ various means that a child acquiring the mother tongue is not likely to employ. Some of these include: i.
Depending on their extra-linguistic knowledge. ii.
Depending on prior linguistic knowledge of the mother tongue or another foreign language learnt before the current target language. This usually leads to linguistic transfers (both positive and negative). iii.
Their implicit and explicit knowledge of the target language. iv.
The context and the co-text.
These factors come into play in one way or the other and are demonstrated in the oral and written productions of the learners.

Data Collection
Data was gathered qualitatively from students at the Department of French at the University of Ghana. We used ten volunteers from the third year Bachelor's degree students and ten volunteers from the final year Bachelor's degree students. We must state that there is a two year gap between the students in the third year and the students in the final year since the final year students go for a one year language immersion program in France or Benin to improve their French before coming back to take their fourth year courses. We will therefore rate the third year students as A2/B1 in French in the European Common Framework for languages and the final year students at B2 on the average with the very good students being rated as C1. They listened to a story in French which was read to them twice 3 . The students were then asked to retell the story as best they could immediately after hearing it the second time. The production of the students in retelling the story served as the data for this research.

Choice of story
The extract from the story La princesse et la grenouille, the French version of the Princess and the Frog, fairy tale in a 2012 Disney collection entitled 365 histoires pour le soir : princesses, (page 67) was selected because it is considered to be written in a simple, clear style and therefore accessible to all the students irrespective of their level. The vocabulary was also considered to be vocabulary they should be familiar with save for a few expressions. Finally, this story was selected because it is a storyline that learners are likely to be familiar with and therefore their comprehension will be facilitated. We considered the approach of reformulating from a source text relevant because this made it possible for the researcher to control the initial input which can be more easily compared to the output of the students.

Experimental protocol
20 volunteers were sought from among the third year and final year students of the French Department of the University of Ghana -10 students from each class. The students were aged between 19 and 25 years across levels. Students in the third year usually have A2+ competence in French and students in the final year usually have B2 competence. The final year students spend a year after the third year in a francophone countryeither Benin or Franceand so there is actually a two-year gap between them and the third-year students. Each class of 10 students was collectively read an extract from La Princesse et la Grenouille (in French) and each class was read the story twice. After the reading, 5 students in each class were asked to retell the story orally and 5 were asked to retell the story in writing. The oral responses were recorded with their permission and transcribed. The transcribed responses and the written responses of the other groups were then fit into the sequences of the story to find out what the students had formulated and what they had not, as well as how they had formulated the sequences that were formulated. Due to some technical challenges, we could not access two of the oral recordings for those in final year and so on the whole we had 18 productions instead of the expected twenty.
This paper does not present every single aspect of reformulation but we present the reformulations which show evidence of the changes in lexical items, change from direct to reported speech, and manipulation of tenses. We will discuss how these reformulations show evidence of implicit and explicit learning.

Oral reconstruction versus written reconstruction
The following table presents a summary of the difference in the retelling of the story based on the mode of reformulation i.e. whether oral or written. An analysis of Table 1 shows that, on the whole, the story was reformulated more by those who retold it in writing, than by those who retold the story orally. With the exception of the 12 th sequence which had one more sequence orally reformulated than the written reformulations and sequences 4 and 11 which were not reformulated at all either orally or in writing, every other sequence has more written reformulations than oral reformulations. Could this be because at this level of acquisition, students are still more comfortable writing in French than speaking it? Our general observation of students of the French language in Ghana is that, they do better at written production than at oral production. This has been attributed to many factors including the emphasis on writing the language rather than speaking it right from the high school. However, another reason could be that the students who retold the story in writing had more time to write than those who had to retell the story orally. This may have led to higher recall. It can also be observed that certain sequences are formulated with greater frequency than others (cf. sequences 1,3,6,7,10,12). If these sequences are put together, they present a summary of the text and the main highlights of the story the storyline. The other sequences seem to be explanations or extensions, providing additional information which may explain why they are not reformulated as much.  have a slightly greater number of reconstitutions than the third year students. This could be because on the whole, they have a better level of competence since they have also benefitted from a year abroad program, which has given them the opportunity to immerse themselves in the language. We also observe that even though there is not much difference when it comes to the quantity of reformulations, there is a difference in the quality of reformulations. That is to say, there is a difference in the complexity of the reformulations including the using of relatives, transformation of words from one category to the other etc. with the final year students displaying higher performance in this regard. In this paper, we will not dwell so much on the differences as the focus is to show that foreign language learners in general are able to input information and output it in their own way.

Syntactical analysis of reformulation by students
Results demonstrate that most students reformulated at least seven out of the twelve sequences of the story. It is also observed that none of the students completely repeated any of the sequences. However, there was evidence of semantic paraphrases and situational paraphrases. We will now present the syntactic aspect of our observations focusing on the three phenomena mentioned earlier in this paper: i) The substitution of lexical items, ii) Reconstruction from direct speech to reported speech, and iii) The manipulation of tenses.
In the following section, we present some examples of the three phenomena focusing on the particular sequences where they are most prevalent. For each example, we present the source statement (SS) i.e. the sequence observed as well the reformulated statements (RS) that are linked to it. We also indicate the level and the age of the participant who produced the said reformulation. The individual students are identified using the alphabet (a to r). As much as possible, the utterances are reproduced as they are though there are few modifications such as placing of accents to help clarify meaning.

The substitution of lexical items
It was observed that several of the participants in the study replaced certain words and expressions in the source text with other lexical items. These included replacing adjectives and verbs with some synonyms or even using phrases to try to explain when they do not have the right word. Some examples are as follows:
Written and oral reformulation : We observe that in all the cases observed, the adjective terrorisée is mostly replaced by the past participles effrayée and terrifiée which are other forms of expressing fear. This is likely to be because they are more exposed to these two expressions être effrayé and être terrifié as being the means of being fearful. The expression avoir peur is also used once during the oral reformulation. It is the simplest form for expressing the notion of being afraid. être effrayé expresses a lesser degree of fear than être terrifié, which is closer in intensity to the adjectif used in the source statementterrorisée. Interestingly, the final year students used the expressions of lesser degree (être effrayé and avoir peur) than the third year students (être terrifié).
We also observed that one final year student used the expression …elle a essayé d'échapper la scène. We can infer from this that the student understood that Tiana did not find meeting the frog a pleasant experience and therefore she tried to escape from the scene. Certainly, one of the reactions to fear is to try to escape from the source of fear. We can therefore consider this a situational paraphrase. We observe in the examples above that the students substitute the verb assommer with frapper, écraser, étouffer and tuer. A possible explanation is that even though assommer is not a verb they are familiar with, they understood it's usage in context used verbs they were more familiar with which had a similar meaning. Even though some of the verbs are closer in meaning than to assommer than others, they all carry the meaning of causing harm or death.
Secondly, the verbe empoigner is replaced with prendre, tirer, and enlever. In this case, the generic verb is prendre and the other verbs are used with varying degrees of correctness but they all carry elements of the generic verb and thus can be said to be acceptable in this context. The verb transformer is rendered changer by 4 out of the 6 students who reformulated this sequence. We also observe that one fourth year students maintains the concept transform but nominalizes it to transformation. Another final year student substitutes devenir (become) for transformer (transform). This is also completely understandable since devenir implies a change of state or a transformation of some sort.

Reconstruction from direct speech to reported speech
It was observed that some of the participants reformulated the sequences by moving from direct speech to reported speech. This also demonstrates their ability to process information and render the information considering discursive factors. Consider the following examples: Sequence 5, SS : « Je ne voulais pas vous effrayer », s'excusa la grenouille en la rejoignant dans la chambre. Elle sauta sur un meuble afin de voir Tiana de plus près.

Written reformulation
25) RS : 3rd year, A, 22yrs: La grenouille l'approcha dans la chambre, saute sur une table et dît à la princesse de ne pas s'inquiéter parce qu'elle n'a pas l'intention de l'effrayer. 26) 4th year, F, 22yrs : La grenouille a dit qu'il n'avait pas l'intention à l'effrayer et il s'est mis à s'introduire. There was no oral occurrence of Sequence 5 indicating reformulation from direct speech to reported speech. However, we observe that even though the tenses were sometimes mixed up, student A rightly reformulated from direct speech to reported speech by using the expression dire de…. Student F also uses the expression dire que which is also used to introduce a reported speech. With sequence 7, we see a lot of variety in changing from direct speech to reported speech with expressions such as : il a dit que…, Il se décrit comme, il a expliqué que, et Il s'est présenté comme. These expressions all announce reported speech and the type of verbs used indicate giving information as is noted in the source statement. We observe that two out of the four written reformulations are produced by 3rd year students and the other two by 4 th year students. However, all the four oral reformulations are produced by 4 th year students. interesting scenario when talking about reformulating from direct speech to reported speech. We observe that while the generic verb dire is used in the expression dire que as in the case of sequence 7 above, at least two of the 4 th year students display an understanding of speech acts as they use demander à to show that it was not just a matter of giving information as in sequence 7 but that this time the prince was making a request. They thus indirectly specify the nature of the reported speech by using the expression demander à instead of the generic expression dire que. Sequence 12 is reformulated both orally and in written form in most cases with the direct to reported speech containing the verb refuser in various forms. However, there are some instances where the student reformulates using other semantic paraphrases. For example, student Q (oral reformulation) uses …elle a dit non (she said no) to indicate refusal. We also have student A (written reformulation) who uses dire que + negation of the verb embrasser which implies a refusal to kiss the frog.

The manipulation of tenses
The students also demonstrated an ability to manipulate tenses and easily substituted the simple past tense (passé simple) with the present tense which can serve as a past in narrative constructions, and the passé composé. The two main sequences in which this was demonstrated are the sequences 8 and 12. In sequence 8, we observe that the first verb empoigner is replaced by another verb in all cases. Moreover, it is changed from the simple past tense to the passé composé in three out of four cases. It is only student H who maintains the simple past tense when he uses the verb tirer. The second verb tenter (tenta) seems more challenging for the students and they avoid reformulating it and use words and expressions showing intention rather than attempt e.g. avoir l'intention de and pour. The simple future tense in Sequence 12 (refusa) is reformulated in several ways including the 'passé compose' (a …refusé), a combination of the 'passé composé' and the near future tense ( a dit … qu'elle ne va pas…), and the imperfect tense (disait). Interestingly, the notion of emphasis rendered by catégoriquement in the source statement is reformulated by complètement, jamais and absolument.

Strategies used by the students in reformulating
The study also takes a look at some of the misrepresentations of the story during the reformulation process. Some of the errors represented lexical gaps while others demonstrated a lack of understanding of the source text. We confirm that these learners made use of strategies such as meaning negotiation, avoidance and paraphrasing to make up for their deficiencies in performance as indicated by Westhoff, 1991, andLittlemore, 2001. We also observe that the students fall on their L1 (English) or even some elements of their L2 (French) to fill these gaps in order to make their communication complete this making use of prior knowledge to sometimes infer unknown elements. The earlier section of our paper has already dealt 14 SHS Web of Conferences 38, 00007 (2017) DOI: 10.105110. /shsconf/20173800007 COULS 2016 with paraphrasing and so we will not treat that again but we will briefly look at the other two strategies meaning negotiation and avoidance as evidenced in our data.

Meaning negotiation
This is mostly demonstrated in the use of some lexical items to replace one they are not familiar with. In our case, some of them used another word or produced the word as they thought they heard it. An example from our data is the reformulation of the first sequence. In this example of sequence 1, we observe that the students do not seem to recognize or know the word balcon. Some reformulate by using close-sounding words (baobab and bardon) while others seem to make inferences of where a frog can be found (banc and chambre). In both cases, we can argue that the students are using the strategy of meaning negotiation.

Avoidance
In our text, we see a clear case of avoidance where sequences 4 and 11 were not reformulated at all by any of the students. These sequences do not seem to be sequences the students will be used to or will be familiar with. We also have situations where one portion of a sequence is reformulated but then another portion is left out. For example, an observation of the refomrulations of Sequence 3 -Tiana rentra dans la chambre, terrorisée. Elle n'arrivait pas à y croirereveals that the students reformulated the first part of the sequence but no one reformulated the second part. It was almost as if that part did not exist in the text.

Conclusion
Ellis (2005) indicates that implicit language learning can be said to occur with fluent comprehension and production. On the other hand, explicit language learning is expressed in the conscious efforts made by a learner to negotiate meaning and construct communication. The activity undertaken by the students can be said to comprise both implicit language learning and explicit language learning as some parts of the extract were easily understood and reproduced by the students. However, there were also cases of meaning negotiation and the construction of communication as the students struggled with some expressions, some vocabulary, and some syntactic structures. Their use of the language through reformulation demonstrates linguistic competence as proposed by Canale and Swain (1980) as they display their knowledge of the language code including grammatical rules, vocabulary, and to some extent spelling and pronunciation depending on whether they reformulated orally or through a written production. Indeed, the fact that they did not merely repeat what they heard but relied on their own repertoire to reconstitute the extract of the story they heard shows that they have individual linguistic baggage that they can fall on to express meaning. In all certainty, the students demonstrated an ability to negotiate meaning as they made an effort, sometimes wrongly, to reformulate what they had heard. However, the success cases show the students using semantic paraphrases to express the meaning of the sequences that they had heard and thus displaying implicit language knowledge.