Developing pragmatic fluency in an EFL context (Julia Barón & María Luz Celaya 2010)
Developing pragmatic fluency in an EFL context in an EFL context
Júlia Barón and María Luz Celaya
University of Barcelona
Studies on pragmatic development, especially on the development of pragmatic fluency, are still scarce in the area of Interlanguage Pragmatics. The present study analyses whether EFL learners (N = 144), from Primary to University levels, who have not been instructed in pragmatics nevertheless show development in pragmatic fluency. A wide variety of measures were used to analyse the learners’ production in open role-play. The results in the present study show that pragmatic fluency indeed develops as proficiency increases (the learners develop their use of gambits and routines, they are capable of changing topics by themselves and they produce appropriate time responses) but also that, in contrast, the development in the use of patterns stops at Grade 11 and there is no development in the opening and the closing phases. These results are discussed in the light of cognitive models of second language acquisition.
1. Background
1.1 Interlanguage pragmatics (ILP) and development
Researchers in Interlanguage pragmatics (ILP) investigate the learners’ knowledge, perception, comprehension, and acquisition of the pragmatics of the second language (L2) on the one hand, and on the other, the use and production of the L2 pragmatics, both in second and in foreign language learning contexts (Alcón, 2008). There is much research on ILP which has focused on comparing the use and production of L2 pragmatics to native speakers’ production, but, as several authors have noted, studies on development are still scarce (Kasper & Schmidt, 1996; Rose, 2000; Kasper & Rose, 2002; Barron, 2003; Achiba, 2003; Häkasson & Norrby, 2005). The present study was conducted in response to the increasing interest over the last decade in tracing the developmental patterns of pragmatic acquisition to examine how L2 learners acquire and develop pragmatic competence.
In order to explain pragmatic development in both second language (SL) and foreign language (FL) contexts, insights have been gained from theories from second language acquisition (SLA) research and other disciplines. Cognitive theories like Schmidt’s Noticing Hypothesis (1993) and Bialystok’s Two-Dimensional Model (1993, 1994) aim to explain learners’ pragmatic development by means of the different cognitive processes they go through. Schmidt’s hypothesis is based on conscious and unconscious learning. According to this theory, L2 learners ‘notice’, conscious or unconsciously, a specific linguistic or pragmatic feature of the L2 which is kept in the learners’ memory. Once such feature is analyzed and recognized by the learner, the process of understanding begins. As regards Bialystok’s model, firstly, learners go through a process of analysis in which they acquire the cultural and social representations of the language. Secondly, they go through the process of control where they select attention on a specific representation with a specific aim or purpose. According to Bialystok, the main difference between children and adults is that children need to go through the process of analysis even in their L1. In the case of adults, since they have already gone through the aforementioned processes in their L1, they only need to develop the control over the L2 representations in order not to fail in a communicative situation.
Sociocultural Theory (SCT) has attempted to present pragmatic development not only as a result of cognitive processes, but as a result of the social and cultural rules that learners acquire by means of interaction, which in turn enable them to acquire the grammatical rules of the target language (TL). Third, Conversation Analysis (CA) sees the pragmatic exchange from an interactional perspective, where the learning process is not only explained from a cognitive perspective, but rather as a social phenomenon which is acquired and used interactively by the speakers of a language (Firth & Wagner, 2007).
One of the earliest studies on the development of L2 pragmatic competence compared the acquisition of pragmatics o the development of the L2 grammar by Wes, a Japanese learner of English as a foreign language (EFL) who stayed in Hawaii for three years (Schmidt, 1983). The results of this study showed that the learner’s pragmatic performance indeed developed, but not his morphology, syntax or his grammar to any significant extent. In contrast, a study by Schmidt and Frota (1986), which analyzed Schmidt’s 5-month stay in Brazil with a focus on the speech act of answering questions affirmatively, showed that even if some grammatical aspects (such as tense/aspect choice or mood) were more difficult to acquire, the learner’s grammatical abilities were better than in his conversational production.
The development of pragmatic competence and grammar has also been analyzed from the perspective of pragmatic and grammatical awareness. Bardovi-Harlig and Dörney (1998) studied the development of pragmatics and grammar in a group of English-as-a-second-language (ESL) learners and a group of EFL learners, each comprising several subgroups of English proficiency. The results showed that ESL learners recognized more pragmatic errors than grammatic errors, and considered them to be more severe. On the contrary, the EFL learners recognized more grammatical errors than pragmatic errors, and considered the former to be more serious. All in all, the results should be interpreted as an indication of development from grammatical awareness to pragmatic awareness as proficiency increases. A study which supported the findings of Bardovi-Harlig and Dörney’s study is Schauer (2006), who studied pragmatic awareness in German ESL learners in a stay-abroad context (SA) during an academic year. The learners’ pragmatic and linguistic awareness was elicited through tasks with grammatically correct but inappropriate utterances and vice versa (the same tasks as in Bardovi-Harlig & Därney, 1998). The results showed that both grammatical and pragmatic awareness increased after the SA period.
As can be seen in these studies, a parallel development can be traced between the acquisition of grammar and pragmatics. The difference between these studies and the studies by Schmidt (1983) and Schmidt and Frota (1986) reviewed above is that there were probably other factors influencing the acquisition of pragmatic competence. In the case of the Wes study, the participant was accepted in the new community; which helped him develop his pragmatic abilities, but not so much his grammatical competence. In the Schmidt and Frota (1986) study, Schmidt was receiving instruction in Portuguese, but the opportunities for him to interact with the new culture were more limited, a fact which might have led to the lack of pragmatic competence. In the case of Bardovi-Harlig and Dörney’s study and also in Schauer’s research, the main difference can be found in the participants, who were studying in a formal context even if they were studying abroad. Grammar and pragmatics are then seen as developing together, even considering the differences between the EFL group which are provided with less input and opportunities for practice and contact with the L2 culture than the ESL and the SA participants.
Another approach in the field has been the development of sociopragmatic ability on the one hand, and the acquisition of pragmalinguistics on the other, through the study of the concepts of appropriateness of speech acts and appropriateness of form of such speech acts. In a cross-sectional analysis by Scarcella (1979), politeness strategies were looked at from beginners to advanced ESL learners by comparing their pragmatic use of politeness to that of the native speakers (NSs). The results showed that politeness forms appeared before the learners had acquired the rules to be able to use such forms in an appropriate way, thus leading to the conclusion that the learners had acquired the pragmalinguistic forms but not the sociopragmatic norms of the L2. Belz and Kinginger (2002) found a similar pattern in the acquisition of sociopragmatics and pragmalinguistics in a study where the aim was to investigate the development of the use of the second person pronouns (du/Sie) in American learners of German in the USA and German learners of English in Germany. By means of a computer-mediated task, each group had to write different texts to the other group; the results showed development toward the appropriate use of the pronouns.
However, opposite results have been found in Bardovi-Harlig and Hartford (1993), who investigated the development of suggestions and rejections by adult learners of English in a formal context The learners showed development toward native-like use in the choice of appropriate speech acts but they failed to choose the appropriate forms of such speech acts; such results, contrary to those in Scarcella, seem to favor sociopragmatics over pragmalinguistics.
The development of requests has been the focus in a large number of studies, especially the development of requests) specifically, the movement from direct requests toward the use of indirect requests, which would be the same development in English first language (L1) acquisition (Ellis, 1992; Trosborg, 1995; Hill, 1997; Rose, 2000; Achiba, 2003; Félix-Brasdefer, 2007). Ellis (1992) followed the development of two ESL beginners in London, aged 10 and 11, in a classroom setting during a period of 15 and 21 months, respectively. The results supported the movement from a direct to an indirect use of requests, although the two learners failed to develop most of the different request types as well as the linguistic means to perform the type of requests that they had acquired. Ellis argued that the sociolinguistic instruction that the learners had received in class was not realistic, and that this might have been an explanation for their failure in sociolinguistic competence.
The same pattern of development of requests was found by Trosborg (1995), who studied the use of requests together with apologies and complaints in EFL Danish learners at three different levels, and by Rose (2000), who looked at the development of requests, apologies, and compliments in Japanese children at three different levels. However, no development could be traced in the use of apologies and complaints in Trosborg’s study and in Roses’s study, apology and compliment strategies did not appear until the learners reached the highest level of proficiency.
It should be noted that the study by Takahashi and Dufon (1989) showed the opposite results in the development from direct to indirect requests, probably due to L1 influence; their female Japanese learners of English, separated into two groups of low and high proficiency, tended to move from indirect strategies to more direct strategies as their proficiency increased, which, as has been claimed, is the opposite pattern of development that occurs both in native speakers of English at the early stages of acquisition and in L2 acquisition.
The study by Achiba (2003) presented a longitudinal case study of her daughter Yao, a 7-year-old Japanese girl, who spent 17 months in Australia. Yao went through four developmental phases, from the use of formulaic utterances, such as routines and patterns, to the production of more elaborate patterns in the second phase, the appearance of indirect strategies and the use of modals in the third, and finally, the production of more complex structures together with new strategies, such as stating preparatory and query preparatory and mitigation. Similar stages of development were found in Félix-Brasdefer (2007), who followed Kasper and Rose’s (2002) proposal of five stages of development in request realization, and in Schauer (2006, 2007), who found that both groups of learners developed toward an increased use of external modification (internal modification was first acquired) in her German participants, in the SA, and the at-home (AH) groups. The difference between the two groups was that the SA group had a broader repertoire than the AH.
Few studies have focused on the development of pragmatic fluency as measured via a wide variety of fluency measures and groups of learners at different ages and levels of proficiency. We address this issue in the present study. The next section revises both the concept of pragmatic fluency and the most relevant measures used in previous studies.
1.2 Defining and measuring pragmatic fluency
According to House (1996: 228), pragmatic fluency can be defined as “the combination of appropriateness of utterances and smooth continuity of ongoing talk”. Being a pragmatically fluent speaker would imply that “the NNSs’ talk must meet the expectations of the NSs of the foreign language and it must represent acceptable language behavior” (House, 1996: 229). In other words, pragmatic fluency could be seen as being pragmatically appropriate in the TL on the one hand, and being a fluent speaker in the TL on the other.
House (1996) followed the development of pragmatic fluency in two groups of adult learners: one group was instructed in conversational routines, and the other group was enrolled in a communicative classroom. The development of pragmatic fluency was measured through the use of gambits, discourse strategies, speech acts, turn-taking, speech rate, pauses, and repairs. Gambits, the linguistic features frequently used in spoken interaction, and also called discourse markers, were classified by Edmonson and House (1981) and by House (1996) into the following types: the uptaker, which provides the information that the message has been received and accepted; the clarifier, which is the current speaker’s utterance; the appealer, which is what the speaker expects the interlocutor to do; and the starter, which anticipates that the speaker is going to say something. Expressions such as ok, well, now, you know, right, yeah, among others, are then considered gambits. As far as discourse strategies are concerned) House’s (1996) main aim was to investigate learners’ knowledge of the different interaction structures that are typically used in order to accomplish a specific conversational goal. As regards speech acts, House (1996) focused on how the learners opened and closed speech acts by means of using routine formulas. Two other measures of pragmatic fluency in her study were how topics were introduced and changed in the speech act, and the appropriate use of turn-taking in relation to the topic discussed in the task. As far as fluency measures are concerned, House (1996) examined the speech rate of the learners’ production as well as the frequency of filled and unfilled pauses. It should be mentioned, though, that not much information is provided about the specific measures used to assess fluency in the study. The results overall showed that the two groups improved in their use of routines. In contrast, as for gambits and strategies, the explicit group was superior to the implicit group. However, no differences were found in speech act realization.
Taguchi (2007a, 2007b) used fluency measures in relation to proficiency and pragmatic competence. Taguchi (2007a) traced the development of pragmatic comprehension ability in a group of Japanese learners of English in indirect refusals and opinions. The participants were assessed by means of a listening task to test their pragmatic comprehension, while speed was analyzed by the average time that the learners took to answer each of the items they were requested correctly. The results showed that pragmatic knowledge did not always seem to match the learners’ processing capacity Taguchi (2007b) studied task difficulty in L2 oral output in requests and refusals in different power relation situations. Appropriateness, planning time, and speech rate were analyzed in the data. The results showed that, in the situations where there was a higher power relation, speech rate was generally slower and the learners at a low proficiency level needed more planning time.
CA methodology can also shed some light on this type of study in the way that CA considers response time in preference organization. The use of delays in some responses can be seen as initiating a dispreferred turn. For example, if a request is made, the preferred response would be an acceptance, and such response should be produced immediately. However, if there is a delay in such response, a refusal might be expected. Nevertheless, a learner may want to produce an acceptance, but due to his or her difficulties in the L2, a delay could appear.
Conversational routines have also been considered as a way to promote fluency and pragmatic competence (Kasper, 1995; House, 1996, 2003; Kanagy, 1999). Conversational routines are linked to specific social situations and carry a strong pragmatic force. Such expressions have been classified into two: routines, which are wholly memorized structures such as ‘how are you’; and patterns, which are only partially memorized structures with open slots such as ‘Can you X’ (Hakuta 1974). As mentioned above, given the scarce amount of studies on pragmatic fluency, together with the need to have more developmental studies in JLP, the research question of the present study is the following:
Does pragmatic fluency develop in EFL contexts, from beginner to more advanced learners, where learners have not been instructed in the pragmatics of the FL?
2. The current study
2.1 Participants
The participants for the present study belong to the BAF (Barcelona Age Factor) Project, which is one of the projects that GRAL (Grup de Recerca en Adquisició de Llengües) carried out (see Muñoz, 2006). The whole sample comprised learners from thirty schools in Barcelona (Spain) where English was taught as a FL. The participants were Catalan-Spanish bilinguals, and as regards the social status of the participants’ families, a mixed population was included: low-middle class, middle class, and professionals. The BAF Project was triggered by a change of an Education Law implemented in Spain which affected both primary and secondary schools. One of the objectives of the new Education Law was to advance the age of learning English from 11 years old (Grade 6) to 8 years old (Grade 3). Researchers of the BAF Project collected data from students who had begun learning English at the age of 11 (Late Starters = LS) and learners who had been first exposed to English at school at the age of 8 (Early Starters = ES), with the new Education law. The aim was to investigate whether the Age of Onset (AU) affected the acquisition of EFL in a formal context. Because of a different aim in the present study, only ES learners, together with university learners (see below) are analyzed here.
As can be seen in Table 1, data were collected at four different moments (Times 1, 2, 3, and 4). At Time 1, the participants had been exposed to English for 200 hours and they were 10 years old (Grade 5). At Time 2, the learners had received 416 hours of English instruction, they were 12 years old, and at Grade 7. At Time 3, data were collected after 726 hours of instruction, when the learners were 16 (Grade 11). Time 4 took place after 800 hours of instruction at Grade 12 and when they were 17 years old.
GRAL has recently begun a new project with data from undergraduates who had been exposed to English at school and who have continued studying English at university The data collection also took place in Barcelona and the participants’ social status was kept similar to the BAF Project. Three universities participated in this project, although for the present study only one of them was included (Communication Studies), since the learners studied EFL whereas those excluded were undertaking a degree in English Studies. This study, then, consists of a cross-sectional sample of data, from beginners (Grade 5) to more advanced learners (University level) with a total of 144 learners (see below).
2.2 Instrument
The participants of the BAF Project carried out a battery of tasks both written and oral: a background questionnaire, a dictation, a doze test, a listening comprehension test, a grammar test, a phonetic imitation and discrimination task, three oral tasks (interview, story-telling, and a role-play), and a written composition (see Muñoz, 2006). The university participants fulfilled the background questionnaire, a vocabulary test, a phonetic discrimination task, the written composition, and the three oral tasks. The present study uses the background questionnaire in order to obtain personal information from each participant and focuses on the open role-play to assess pragmatic development, since open role-plays allow learners to participate in quite natural situations in which pragmatic use is required.
Two different background questionnaires were used: one for the participants at primary and secondary education, and a different one for the university participants, since it was adapted to adults, and con lamed information on their past experience with English. The background questionnaires were extremely important in order to find out whether the participants were taking extra English classes outside the school or university and whether they had been abroad to learn English, since only participants who had been exposed to English at school and at university were included. However, since the aim is to look at interaction between two participants, in those cases where a participant with no extra English courses or a SA carried out the task with a non-valid participant (those learners with extra exposure), the latter was also included in order to observe the whole interaction. Non-valid participants were carefully analyzed in order to see if they behave differently from the valid participants.
The role-play carried out by school learners consisted of a task fulfilled by two students in which one played the parent, and the other the child. The son/daughter had to ask for permission from the mother/father to host a birthday party at home. They were asked to discuss the time the party would start and finish, the number of friends that they would invite, the food and drinks they could buy, and any other topics that they would find it necessary to discuss. Finally, after negotiating these issues, the mother or father had to decide whether or not he/she would give permission to the son/daughter to hold the party at home. In the case of the role-play for university students, they were not asked to play the roles of parent and child for obvious reasons of age differences; they had to pretend that they were roommates and one of them wanted to celebrate his/her birthday party at home and had to ask the other roommate for permission. The topics they had to discuss were exactly the same as in the case of the school learners.
2.3 Measures
To assess pragmatic fluency, some of the measures that were used in the study by House (1996) are used in the present study, such as how the learners open and close the speech act, how topics are introduced and changed by learners, how turn-taking is developed throughout the role-play, and the use of gambits, routines, and patterns. The time that the learners take to respond to requests is also considered, as well as the time that the ‘requester’ may take if an offer is made in return.
Opening and closing the speech act
In a speech act carried out by NSs or by pragmatically proficient non-native speakers (NNSs), one of the appropriate rituals of a speech act is to provide an opening and a closing phase. According to House (1996) and Edmonson and House (1981), the opening phase can be divided into two parts: the greeting, where the speaker meets the interlocutor (e.g. Hello and the How-are-you), and where the speaker shows an interest in his/her interlocutor. As regards the closing phase, House also divides it into two categories: the first one is the extractor, where the speaker anticipates that he/she is leaving the conversation (e.g. I’m sorry but I should move on), and the second one is the leave-take, in which the speaker leaves the conversation (e.g. Goodbye).
Due to the nature of the data in the present study, the different sub-categories of the opening and closing phase of the speech act have been simplified. This decision was made because most of the participants tended to go straight to the task and omitted the opening and closing phases of the speech act. However, there are cases in which the learners open and close the speech act in a simpler way to that in House’s classification. The following examples, taken from the data in the present study, illustrate how this measure has been coded and analyzed:
(1) Grade 7
Opening phase
*R1: hello dad! [Opening_Phase]
*R2: hello. [Opening_Phase]
Closing phase
*R1: ok thank you by. [Closing_Phase]
*R2: bye. [Closing_Phase]
Topic introduction and topic change
The ability to introduce and change topics in a conversation is used as a discourse strategy and also considered as a sign of pragmatic fluency (House, 1996). As was seen in the description of the role-play used in this study, the participants were already given instructions on the topics that they could discuss. The analysis of the topic introduction and topic change has been undertaken from the perspective of how the participants are able to introduce and change topics by themselves. Since the investigator plays an important role, especially at the first stages of acquisition and he/she usually helps the participants even with topic changes. In the present study, the main focus is on topic change, since topic introduction is already provided before the task begins. The aim of analyzing the use of topic changes is to observe the ability of the speakers to change from one topic to another without the help of the researcher, as in Example 2:
(2) University level
*R1: yeah we can buy pizzas and [/] and potatoes.
*R2: but the other people have to put money.
*R1: and until what time? [Topic Change]
*R2: ten o'clock?
Turn-taking
As considered in House’s study, the role of appropriate uptaking in a conversation is also a sign of pragmatic fluency. In the present research, turn-taking is one of the measures to analyze the development of pragmatic fluency. The number of turns of each participant is taken as a unit of analysis in order to check to what extent each learner participates in the task. It should be taken into account, though, that in most cases, there are participants whose number of turns is high, but their interventions are only one- or two-word turns. In such cases, turns are considered as “minimal” in length (Swain, 1988). The example provided below illustrates these two types of turns:
(3) Grade 11
*R1: hmm@p uh@p can I do a party next Wednesday?
*R2: yes. [Minimal turn]
*R1: ok I want #1 explain to you the the my plan.
*R2: mmhm@i. [Minimal turn]
*R1: I want that invite hmm # fifteen fifteen people and I don’t know! want to # to dancing and put music and and make some sandwiches and buy drinks.
*R2: ok. [Minimal turn]
*R1: dancing and # and eat things?
*R2: yes. [Minimal turn]
Apart from the participants’ turns, the investigator’s turns are also considered, since the researcher’s role seems to be important in this study.
Use of gambits
As mentioned in Section 1.2, expressions such as well, ok, right, now, yeah, oh were considered as gambits by House (1996) and Edmonson and House (1981). The context in which such expressions are produced by learners may provide information about pragmatic fluency; children might use these expressions as chunks and sometimes not appropriately, whereas it should be considered whether adult and more proficient learners use them more frequently than children and in an appropriate way.
Below are some examples of the use of gambits that can be found in the data of the present study (Example 4):
(4) University students
*R2: I’m ok [Gambit] and you?
*R1: yeah I’m ok [Gambit].
*R1: listen [Gambit] hmm@p this weekend it’s my birthday so I thought [/] I thought maybe I could make a little party at home and I just wanted to ask you if you were agree with me, if it’s fine for you that I... (the conversation goes on)
*R1: listen [Gambit] it’s my birthday party &=laugh, okay [Gambit]?
*R1: I just wanna make something with my friends and not play to the
with the ...
*R2: I mean [Gambit] I don’t mind &=laugh if you want to drink.
Routines and patterns
The aim of analyzing the use of routines and patterns is to investigate (i) whether such conversational routines are present in the current data as a sign of pragmatic fluency, and (ii) Lo check whether there is a development from a low frequency in the use of routines and patterns at the first stages toward an increase of such use in more advanced learners. Examples 5 and 6 below illustrate the use of routines and patterns:
(5) Grade 5
*R1: uh@p it’s a [Routine] party.
*R2: yes.
*R1: it’s a [Routine] about a party.
*R2: hmm@p.
*R1: it’s a [Routine] party.
*R2: yes.
*R1: com es diu begudes? (how do you say drinks?)
*INV: drinks.
*R1: it’s a [Routine] drinks.
(6) University level
*R1: how are you [Routine]?
*R2: fine [/] fine thank you [Routine].
*R2: and you [Routine]?
*R1: fine [Routine].
Response time
In House (1996), speech rate, frequency, and length of pauses produced by the learners were some of the units to measure fluency. In the present study, it was decided that such measures were not reliable in the current data, since the bad quality of some of the recordings made it difficult to measure the silent pauses and speech rate in the task. Therefore, we decided to use a measure which could possibly relate pragmatics and fluency to pragmatic appropriateness. What was used to investigate the development of pragmatic fluency, therefore, was the time that the participants took when responding to a request or to an offer or suggestion. This should be related to preference organization, where in preferred turns, for example in an acceptance to a request or a pre-request, a participant would probably respond immediately after the request. If instead of responding immediately, the learner delayed (over 0.4 seconds), then this response can be considered as not pragmatically appropriate (Schegloff, 2007). However, as Wong (2000) pointed out, it should be taken into account that the time learners take may be due to communicative problems, or it could even be argued that at that moment the participant could be distracted by some external factors. Examples 7 and 8 show how the response time has been analyzed in the present study data:
(7) Grade 11
*R1: can I make a [/] a party in home?
*R2: (0.8) ok.
(8) University level
*R1: I just thought that today is my birthday and +/.
*R2: hey happy birthday! We could make a party!
*R1: (0.3) I would like to make a party.
3. Results
Since the learners participating in this study were playing two roles (child or roommate asking for permission on the one hand, and parent or roommate giving permission on the other), the results are reported separately for each role (R1 = role of son/daughter/roommate, R2 = role of father/mother/roommate) because the participants may be behaving in different ways depending on the role.
First of all, the aim was to see if there were differences between grades from a developmental perspective, so a Pearson’s Chi-square and a Bonferroni correction were used; the level of significance was set at p < 0.1.
As regards the opening and closing phases of the role-play, the data analysis shows that the learners do not commonly open and close the task. Some learners only open the task but do not close it, or vice versa. The statistical analysis shows that no statistically significant differences are found in the groups as regards the opening phase neither for Rl nor for R2 (p = 0.45 1 and p = 0.188, respectively). As far as the closing phase is concerned, no statistically significant differences are found in the different groups and in the two roles (p = 0.281 for R1 and p = 0.281 for R2).
The use of gambits tends to appear as proficiency increases. At the early stages, the learners may use gambits, but this may be so because they are repeating what the researcher has said or because they cannot follow the conversation, as can be seen in the following Example 9:
(9) Grade 7 (R2_27)
*R1: you buy the bananas?
*R2: ok.
*R1: you go supermarket?
*R2: ok.
However, the data analysis shows that, as has been mentioned above, more proficient learners use gambits in a way which promotes their pragmatic fluency, as Example 10 shows:
(10) University level (R2_72)
*R1: and do you want to help me with the party?
*R2: yeah okay of course we can go together to buy things that would be okay.
Statistical tests present strong statistically significant differences in groups (p = 0.000) for both roles. The results show that there are also strong significant differences in the use of routines for both roles (p = 0.000). Therefore, the statistical analysis supports what has been analyzed in the data, that there are differences between the groups as regards the use of routines, which tend to be frequently used by University learners (11):
(11) University level (R1_77) (R2_77)
*R1: hi how are you?
*R2: fine thanks and you?
Regarding patterns, the data analysis has also found differences between the groups, which means that learners behave differently depending on the grade they are in. Statistically significant differences can then be found for both R1 (p = 0.00 1) and R2 (p = 0.025) (see Example 12):
(12) Grade 11 (R2_54)
*R2: do you want something especial for the party?
The results also show strong statistically significant differences in the participants’ ability to change topics also for both roles (p = 0.000). This finding suggests that as learners become more proficient. they have the ability to change topic by themselves — without the help of the investigator, who seems to be present, especially at very early stages, helping the learner carry out the task (the role of the investigator will be further discussed later on in the paper). As far as the response time is concerned, the data analysis seems to show that there is a development from Grade 5 to University level. This is also supported by the statistical analysis, which shows that once again there are statistically significant differences in both roles (p = 0.000).
The second aim was to see where the differences lay between groups; Custom Tables and Comparison of column proportions were used. Custom Tables report the results with the total and percentage of the use of each variable in relation to each grade group. This way, it is possible to trace where there is an increase or decrease of each variable. With a similar aim, Comparison of column proportions marks those grade groups in which a variable presents significant differences. When using Custom Tables and the Comparison of column proportions, the test which processes the data is the Binomial test, which is the most common test used to examine differences between different sample proportions with small samples of data. The statistical tests used mainly report whether the learners have produced the pragmatic feature analyzed or not, against the total of participants (N = 18 R1, and N = 18 R2 for each grade).
As can be seen in Figure 1 and Figure 2 below, gambits are less often used by Grade 5 and 7 participants, both by R1 (22.2% and 11.1%) and R2 (16.7% both grades), than by Grade 11 participants and undergraduates, also by R1 (61.1% and 88.9%) and R2 (72.2% and 94.4%). The results also show a development toward an increased use of routines in the highest grades, but there is a clear increase at University level, with a 100% of use in the participants playing R1, and a 94.4% of use in the participants playing R2.
The most common type of routines used at University level is already memorized structures which promote fluency, as Example 13 shows:
Figure 1. Results gambits, routines and patterns (Role 1)
Figure 2. Results gambits, routines and patterns (Role 2)
(13) University level (R2_65) (R1_65)
*R2: how are you?
*R1: I’m fine thanks.
As far as the use of patterns is concerned, not many learners at Grades 5 and 7 make use of patterns. Grade 11 students and undergraduates use more patterns than students in Grades 5 and 7; however, this variable shows that patterns are not the most widely used as compared to gambits and routines. For example, both in Grade 11 and University level, 44.4% of the participants use patterns; therefore, both Grade 11 and University learners behave in a similar way as regards the use of patterns.
As regards the response time, there is a development toward an appropriate response time in the later stages, in R1 (100%) and in R2 (88.9%), as is shown ¡n Figure 3 below.
Figure 3. Results response time
Another aspect worth analyzing is the fact that the participants’ ability to change the conversational topic during the task seems to increase as they become more proficient However, this development can be better traced in those participants playing R2, where in Grade 5 only 16.7% of the learners change topic by themselves without the help of the interviewer versus 94.4% of the undergraduates, who change topic by themselves.
In relation to this issue, turn-taking also provides information on how the learners are able to carry out the task by themselves. Each of the individual participants’ interventions and the investigators’ interventions are considered separately. The aim is to see whether there would be differences between the participants’ turns and the investigators’ turns. A One-way ANOVA was used to search for interactional effects between the independent variables. As a post-hoc test, the Tukey honestly significant difference (HSD) was used in order to see where differences appeared. However, no statistically significant differences were found between the participants’ turns. It seems that all the learners behave in a similar way as regards turn-taking. The main difference can be found in the investigators’ turns, where — as can be seen in Figure 4 – the investigator is present at the very early stages, while at higher proficiency levels, the investigator’s presence is scarce. This is related as well to the learners’ ability to change topics by themselves and to their ability to carry the weight of the conversation on their own.
Figure 4. Results investigator turns
Regarding the minimal turns, there are no differences between Grades 5 and 7 on the one hand, and between Grade 11 students and undergraduates on the other. However, differences can be found between the two first grades and the two last grades. Differences between Grade 5 and University learners are only marginally significant (p = 0.160) when participants are playing R1. There are also marginal statistically significant differences (p = 0.148) between Grades 5 and 7 in R2. It should be argued that participants in Grades 5 and 7 usually perform minimal turns by means of using answers such as yes/no because of their lack of understanding or because they did not know what to say in order to participate in the task more actively. The most proficient participants used minimal turns by means of expressions such as ok, yeah, or right as a sign of consent and agreement, which shows an understanding of the task.
4. Discussion
The results in the present study show that there is development as regards pragmatic fluency. To begin with, the learners develop their use of gambits and routines as their proficiency increases. In the earliest stages, there is a scarce use of gambits and routines if compared to Grades 11 and 12 and undergraduates; besides, whenever gambits and routines are produced in early stages, their use is no the same as their use by more advanced learners. Younger learners tend to use gambits and routines either from the input received in class or else because they repeat expressions that the researcher has said — such as the use of ‘okay’ or ‘right’. Such features seldom appear in a pragmatically appropriate way, due to the lack of linguistic means. For example, if the researcher had said ‘okay’, one of the learners might use the same expression to answer any of the questions asked by the other participant. It is argued here that this use of gambits and other routines in similar situations does not imply pragmatic fluency. Instead, this behavior is more related to the use of formulaic language in early stages; as Kanagy (1999) pointed out, such routines may be helpful in children in order to acquire the L2 grammar and also the L2 pragmatics, especially in immersion contexts. Further analysis of the use of gambits and routines reveals a similar behavior in Grades 5 and 7 as opposed to Grades 11, 12, and undergraduates. At the highest grades and university level, learners tend to use gambits and routines in a more pragmatically appropriate way, and especially at University level, such expressions help to promote the learners’ fluency (Kasper, 1995).
The analysis of the use of patterns, even if also showing a clear development as proficiency increases, does not provide much information as regards the development of pragmatic fluency in the present study, since at Grade 11 the development stops.
As regards topic change, the present study did not only aim to investigate whether the participants are capable of changing topics by means of different discourse strategies and discourse lubricants (House, 1996), but also, before looking at how participants change topics, to see if they could change topic by themselves. This was due to the fact that the data samples of the present study also consist of children who, in most cases, are not capable of carrying out the task by themselves. Consequently, the role of the investigator has also been considered. In the earliest stages the investigator is always present and, as proficiency increases, the presence of the investigator decreases. This also means that younger participants do not change topic by themselves, but with the help of the investigatory and when they change topics, they do not use discourse markers (gambits and routines) to do so. In contrast, more proficient learners change topics by themselves by means of such discourse markers, as was found in House (1996); the participants playing R2 seem to carry the weight of the conversation more than the participants playing R1.
Contrary to what has been stated thus far, the participants in the present study do not show any development in the opening and the closing phases, since on the whole they do not open and close the task. What can be seen in the highest stages is that, if they do open the task, they use pragmatically appropriate routines that make them pragmatically fluent in that phase. This result may be due to the fact that they do not find it necessary to open and close the task.
As regards the response time, in a different task and with a different purpose to Taguchi (2007), who measured the response time to pragmatically appropriate responses, it could be seen whether the participants produced appropriate responses as their proficiency increased, along with a development from childhood to adulthood. More advanced learners do not only produce delays when a dispreferred act, such as a refusal, is coming up, but also make use of discourse markers such as ‘well’, which shows that the participants hesitate before refusing the request or pre-request.
However, even if advanced learners seem to be competent in this aspect, the results for the young learners might be misleading. The fact that they produce delays, even in preferred and dispreferred responses, may be caused by their lack of pragmatic competence or necessary linguistic means, so they have to think before they make a response. This finding might support Wong (2000), who claims that delays in sequence organization should not always be attributed to pragmatic inappropriateness or low proficiency, but also to other factors such as nervousness and distractors which might be influential as well.
5. Conclusion
This study has analyzed the development of pragmatic fluency in four groups of EFL learners from less to more advanced levels of proficiency. One of the main claims of this study has been the scarcity of studies focusing on pragmatic fluency, and also the need to study ILP from a developmental perspective, as many authors have claimed. Previous studies have focused on either adults or children who were learning English in the target community (as Ellis, 1992; Achiba, 2003) or on children in an EFL context (Rose, 2000). However, the focus on pragmatic fluency in the present study, the group of adult learners included, and the wide variety of measures of pragmatic fluency specially adapted from previous research for the present study are features which might provide a new perspective to the analysis of the development of pragmatic fluency.
The results discussed above definitely point to the existence of development in pragmatic fluency from beginner to more advanced learners in an EFL context with no instruction on pragmatics. This can be clearly seen in the case of gambits and routines, topic change, and response time.
At the early stages, any instance of gambits and routines appears as a result of repetition from the input received and does not consequently imply the acquisition of pragmatic fluency. On the contrary, at Grades 11 and 12 and undergraduates, the use of gambits and routines is pragmatically appropriate. As regards topic change, it is only at the highest grades that learners are able to change topics by themselves, especially those learners playing R2. Finally, the development of pragmatic fluency as measured through response time is also directly related to an increase in proficiency.
However, as has been seen, not all the measures used in the analysis provide such clear-cut information on the development of pragmatic fluency. The use of patterns seems to get stuck at Grade 11 and, thus, the data obtained in the present study allows no further analysis. Similarly, no development can be analyzed in the opening and closing phases, since learners do not usually open or close the task, as has been discussed above.
Since the learners in the present study belong to a formal EFL context, the most plausible explanation to the findings in the development of pragmatic fluency has to be sought in cognitive explanations to second language acquisition, as described at the beginning of the paper, and not in relation to socio-cultural models. Interestingly, even if our learners had not been instructed in pragmatics, they were able to become more pragmatically fluent, as reflected in most of the measures. Bialystok’s Two-Dimensional Model, which claims that adult learners might take advantage of their L1 pragmatic competence, appears as a good theoretical framework for the present study. Therefore, the learners’ development in pragmatic fluency could be possibly explained by means of Bialystok’s Two-Dimensional Model, which as pointed out at the beginning, adult learners only need to control the representations of the L2, since they have already acquired and controlled the representations of their L1. This could possibly explain why the adult learners are the most pragmatic fluent participants in this study. It could be argued then, that adult learners might take advantage over children, since they are already pragmatically competent in their L1. In the case of children, they have to go through both the process of analysis and the process of control in their L1 and in their L2.
Future research is needed to further develop the concept of pragmatic fluency in order to include new measures which could possibly link pragmatic fluency to pragmatic appropriateness in the TL. Further research could also provide relevant suggestions for educational policies.
Acknowledgments
The authors acknowledge the Ministry of Science and Technology for support through grant HUM2007-64302 to the GRAL research group. We also thank the anonymous reviewer for the comments. We would like to dedicate this paper to the memory of Dr W. Edmonson.