Exploring the effects of input-based treatment and test on the development of learners’ pragmatic proficiency

Available online at www.sciencedirect.com

Journal of Pragmatics 41 (2009) 1029–1046 www.elsevier.com/locate/pragma

Exploring the effects of input-based treatment and test on the development of learners’ pragmatic proficiency Masahiro Takimoto * Tezukayama University, 1-15-30-1517, Imafukunishi, Joto-ku, Osaka 5360004, Japan Received 22 March 2006; received in revised form 14 November 2008; accepted 1 December 2008

Abstract The present study evaluates the relative effectiveness of three types of input-based instruction: comprehension-based instruction (proactive explicit information + structured input task), structured input instruction (structured input task), and consciousnessraising instruction (consciousness-raising task). The present study also investigates how the effectiveness of these different types of input-based instruction varies according to the method of assessment of learners’ pragmatic proficiency. Treatment group performance was compared to that of a control group on pre-, post-, and follow-up tests, which took the form of a planned discourse completion test, a planned role-play test, an unplanned listening judgment test, and a planned acceptability judgment test. The results of the data analysis indicate that the three treatment groups performed significantly better than the control group, but that the comprehension-based instruction group did not maintain the positive effects of the treatment between the post-test and followup test in the listening test. The results of the data analysis also reveal a significant main effect for Test and the three types of inputbased instruction varies according to the method of testing. # 2008 Elsevier B.V. All rights reserved. Keywords: Input-based task; Structured input task; Consciousness-raising task; Proactive explicit information; Pragmatic proficiency

1. Introduction Prior research into cross-cultural pragmatics between Japanese and native speakers of English has shown that such speech acts as requests, refusals, and apologies display cross-cultural variations (Beebe et al., 1990). It is therefore necessary to provide learners with opportunities that facilitate the development of many aspects of second language (L2) pragmatics. A key issue is to what extent it is possible to arrange learning opportunities in such a way that learners can develop pragmatic proficiency. In order to test claims about the role of awareness and attention in L2 learning, a number of interventional studies in pragmatics teaching have been conducted. Most of these studies have demonstrated that explicit instruction is more effective than implicit instruction in teaching pragmatic knowledge (e.g., House, 1996; House and Kasper, 1981; Takahashi, 2001; Tateyama et al., 1997). This finding reflects the tentative generalization supported by Norris and Ortega’s (2001) quantitative meta-analysis of 49 instructional studies (with a predominant emphasis on morphosyntax), which indicates the superiority of explicit instruction over implicit instruction (e.g., DeKeyser, 1995; Robinson, 1996). In addition, some of the interventional studies focusing on pragmatic features have indicated that pragmatic features are effectively learned when taught explicitly together with * Tel.: +81 6 6936 3182. E-mail address: [email protected]. 0378-2166/$ – see front matter # 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.pragma.2008.12.001

1030

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

some sort of input enhancement activities (e.g., House, 1996; Rose and Ng, 2001; Takahashi, 2001; Tateyama et al., 1997). The present study looks into how input enhancement activities can be best combined with explicit teaching technique and furthermore it examines how different input enhancement activities vary according to the method of testing in L2 pragmatics. 2. Input-based instruction in L2 learning VanPatten and Cadierno (1993) argued that instruction should focus on altering how learners perceive and process input rather than changing how learners produce language output. That is, instruction that changes the way input is perceived and processed by learners is more likely to become intake. Ellis (1997) proposed that it is the manipulation of input rather than output that is more likely to result in the integration of intake into learners’ implicit/declarative knowledge. Swain (1985), however, arguing that there are roles for output in second language acquisition, formulated the Output Hypothesis as a complement to Krashen’s (1985) Input Hypothesis. Ellis (1997) emphasized the importance of communicative language use and practice to automatize already learned knowledge. In addition, VanPatten (2003) suggested that output may help learners develop fluency. Thus, output is indeed also an important factor in second language acquisition but for different reasons. Output involves retrieving language data from the system, while input is itself potential data for intake into the system. In other words, input is used as the means by which acquisition takes place, whereas output is used to stimulate acquisition processes (Swain, 1998). In the present study, input is defined as the target pragmatic features participants are exposed to in tasks, whereas output is the target pragmatic features participants produce in speech and writing. The present study focuses on the role of input in developing learners’ pragmatic proficiency. Ellis (2003) has explained that two types of input-based approaches, structured input task and consciousness-raising task, can be best used in teaching languages. These two types of inputbased tasks rely on input enhancement, which is defined by Sharwood Smith (1993) as any pedagogical intervention that is used to make specific target features of the input more salient as an effort to draw learners’ attention to these features. 3. Previous research 3.1. Differences in task types: structured input vs. consciousness raising Processing instruction is a term coined by VanPatten (1991). According to VanPatten, processing instruction entails three basic features: 1. Learners are given information about a linguistic structure or form. 2. Learners are informed about a particular input processing strategy that may negatively affect their picking up of the form/structure during comprehension. 3. Learners are pushed to process the form/structure during activities with structured input—input that has been manipulated in particular ways so that learners become dependent on forms and structure to get meaning and/or to privilege the form/structure in the input so that learners have a better chance of attending to it (i.e., learners are pulled away from their natural processing tendencies toward more optimal tendencies) (1991: 764–765) Ellis (2003) has emphasized the third feature, i.e., structured input tasks, in which input ‘‘is purposefully ‘prepared’ and ‘manipulated’ to highlight particular grammatical features based on hypothetical input processing strategies believed to be employed by learners’’ (VanPatten, 1993: 438). VanPatten (1996: 67–70) argued further that structured input activities should observe the following six guidelines: ‘‘(a) teach only one thing at a time, (b) keep meaning in focus, (c) learners must do something with the input, (d) use both oral and written input, (e) move from sentences to connected discourse, and (f) keep the psycholinguistic processing mechanisms in mind.’’ Ellis (1997) argued that structured input texts need to be contrived in such a way that the target forms are frequent, the meanings of the target forms are clear, and so that comprehending the target forms is essential for comprehending the whole text. According to Ellis, there are two ways of structuring input: input flooding (e.g., Trahey and White, 1993) and interpretation tasks (e.g., VanPatten and Cadierno, 1993). Ellis (1997: 155–159) explained some general principles for the design of interpretation tasks:

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

1031

1. An interpretation activity consists of a stimulus to which learners must make some kind of response. 2. The stimulus can take the form of spoken or written input. 3. The response can take various forms (e.g., true/false, check a box, select the correct picture, draw a diagram, perform an action) but in each case the response will be either completely non-verbal or minimally verbal. 4. The activities in the task can be helpfully sequenced to require first attention to meaning, then noticing the form and function of the grammatical structure and finally error identification. 5. Interpretation tasks should require learners to make a personal response (i.e., relate the input to their own lives) as well as a referential response. Employing the interpretation tasks proposed by Ellis (1997), the present study defines such tasks as activities designed to invite learners to engage in intentional learning by consciously noticing how a target pragmatic expression is used in input specially contrived to contain numerous exemplars of the structure. A number of more recent studies have contrasted structured input- and output-based instruction in grammar teaching (e.g., Allen, 2000; Collentine, 1998; DeKeyser and Sokalski, 1996; Hazzard, 1999; Kim, 2001; Nagata, 1998; Salaberry, 1997; Tanaka, 1996, 2001; Toth, 1997). The results of all the studies comparing structured input with output-based instruction in grammar teaching have indicated that the structured input group performs as well on comprehension tasks as the output-based group. In addition, a number of studies (e.g., Allen, 2000; Collentine, 1998; DeKeyser and Sokalski, 1996; Toth, 1997) have reported equal gains by both the structured input and output-based instruction groups on comprehension tests and production tests, while some studies have found that the structured input tasks resulted in statistically significant greater gains either on comprehension tests or on production tests than output-based instruction (e.g., Hazzard, 1999; Tanaka, 1996). It is not possible, however, to determine from these studies whether explicit explanation before the structured input activities is beneficial to the processing instruction. To address this, VanPatten and Oikkenon (1996) conducted a study that examined the comparative effects of three types of instruction: (1) proactive explicit instruction with structured input tasks; (2) explicit instruction only and (3) structured input tasks. The participants were high school students in the U.S. and the target structure was Spanish clitic pronouns. The data were collected through interpretation and sentence-level production tasks the day after the instructional treatments were completed. The results showed that the participants in the structured input instruction and proactive explicit instruction with structured input tasks groups performed equally well and better than those in the explicit instruction group on the comprehension test. VanPatten and Oikennon (1996: 508) concluded that ‘‘significant improvement on the interpretation tests is due to the presence of structured input activities.’’ Since the study conducted by VanPatten and Oikennon, similar results have been obtained by Benati (2003), Farley (2003), and Wong (2003). These studies have shown that structured input tasks by themselves were very effective to improve learners’ proficiency level. This harmonizes with Wong and VanPatten’s (2003) argument that structured input task alone is sufficient to cause improvement in learner performance and knowledge. Consciousness-raising tasks, on the other hand have been noted by Ellis (2003) to differ from structured input tasks. Structured input tasks are designed to lead to implicit learning, whereas consciousness-raising tasks are intended to cater to explicit learning. That is, the structured input tasks cause learners to attend to specific target features in the input, identify and understand the meanings they convey, and facilitate the process of comparison between their existing representation of these features and the actually observed features in the input, while the consciousnessraising tasks are designed to develop explicit knowledge about some target features, which later helps learners to identify and understand the meanings they convey, and facilitate the process of comparison between their existing representation of these features and the actually observed features in the input. Ellis (2003: 163) has identified the main characteristics of consciousness-raising tasks: 1. There is an attempt to isolate a specific linguistic feature for focused attention. 2. The learners are provided with data that illustrate the target feature and they may also be provided with an explicit rule describing or explaining the feature. 3. The learners are expected to utilize intellectual effort to understand the target feature. 4. Learners may be optionally required to verbalize a rule describing the grammatical structure. Furthermore, Ellis (2003: 163) has argued that ‘‘a consciousness-raising task consists of: (1) data containing exemplars of the target feature, and (2) instructions requiring the learners to operate on the data in some way.’’

1032

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

Consciousness-raising activities are designed in accordance with the data options, i.e., ‘‘(a) authentic vs. contrived, (b) oral vs. written, (c) well-formed vs. deviant, and (d) gap vs. non-gap,’’ and operations, i.e., ‘‘(a) identification (i.e., the learners are invited to identify incidences of a specific feature in the data by, for example, underlining it), (b) judgement (i.e., the learners are invited to judge the correctness or appropriateness of features in the data), and (c) rule provision (i.e., in the case of inductive tasks learners may or may not be asked to give a rule to account for the phenomena they have investigated)’’ (Ellis, 1997: 161–162). There are some studies that have investigated the effectiveness of consciousness-raising tasks in developing explicit knowledge of the L2. Fotos and Ellis (1991) compared the effects of direct consciousness-raising instruction (teacherfronted grammar explanations) with the effects of indirect consciousness-raising instruction (consciousness-raising tasks only) on two groups of Japanese EFL college students. The target grammatical feature was dative alternation and the data were collected through grammaticality judgment tests. The results indicated that both groups made significant gains on the grammaticality judgment test with the more durable gains for the direct method. In a follow-up study, Fotos (1994) compared the effects of direct consciousness-raising instruction with the effects of indirect consciousness-raising instruction on Japanese EFL learners. The target grammatical features were adverb placement, dative alteration, and relative clauses. The results showed that there was no statistically significant difference between the two groups. Based on the results of two studies, Fotos and Ellis (1991) and Fotos (1994) concluded that the consciousness-raising tasks may be regarded as useful activities because the tasks promote noticing and proficiency gains. 3.2. Interventional studies of L2 pragmatics As mentioned in the previous sub-section, most of the interventional studies in pragmatics teaching have indicated a general trend in support of explicit instruction. Among them, some studies (e.g., House, 1996; Rose and Ng, 2001; Takahashi, 2001; Tateyama et al., 1997) have employed input-based approaches in teaching pragmatics. Tateyama et al. (1997), for example, investigated how beginning learners of Japanese as a foreign language developed Japanese pragmatic proficiency under two instructional treatments. The pragmatic features to be learned were three functions of the routine formula sumimasen, i.e., as an attention getter, apology, and thanking expression. In the explicit group, students first discussed the different functions of sumimasen. The instructor then provided further examples and explanations with a handout which illustrated and explained the differences in use of the routine formulae. Finally, they were shown short video clips of examples of the pragmatic routines under study. The implicit group did not engage in the explicit metapragmatic activities, but they were shown the same video clips as the explicit groups. Tateyama et al. (1997) found that the explicit group had an advantage over the implicit group. In another study in support of explicit instruction with intermediate learners, Takahashi (2001) proposed a broad view of input enhancement and set up four input enhancement conditions: explicit instruction, form-comparison, form-search, and meaning-focused conditions. These four conditions differed from each other in terms of the degree of input enhancement, with the explicit instruction showing the highest degree of input enhancement and the meaningfocused condition the least. The result demonstrated that the explicit group learned all of the request strategies more successfully than the other three groups. Another study with high-intermediate learners by Rose and Ng (2001) investigated the effectiveness of explicit and implicit approaches in teaching compliments and compliment responses. Both experimental groups followed the same procedure with one exception. The implicit group was not provided with any teacher-fronted metapragmatic information about the target structures, but was exposed to film segments and additional examples of the structures, and provided with questions to guide their discovery about the target features. After six lessons (30 min each), the results of three questionnaires (self-assessment questionnaires, discourse completion questionnaires, and metapragmatic questionnaires) showed that although both explicit and implicit instruction led to gains in pragmalinguistic proficiency, only explicit instruction effectively developed sociopragmatic proficiency. Rose (2005) commented that the finding for similar improvement of both explicit and implicit instruction in pragmalinguistic proficiency is due to the advanced proficiency level of participants or the low level of difficulty of the pragmalinguistic target features. In a similarly designed study with high-intermediate to advanced level proficiency EFL learners, House (1996) looked into how German university students improved on initiating and responding to speech acts and conversational routines in two versions of the same communication course. In the explicit version of the course, students received teacher-fronted explicit metapragmatic information about the sociopragmatic conditions governing the use of routines and their pragmatic functions. The implicit version did not provide explicit metapragmatic information about the

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

1033

target features. Students in both versions listened to tapes of their own language behavior and samples of tape-recorded conversation at various stages of the course indicated that both groups improved over the instruction period (14 weeks), but the explicit groups were superior to the implicit groups. The above studies suggest that learning of the target pragmatic features is possible when they are taught explicitly with some sort of input enhancement and that it would be interesting to see how input-based tasks such as structured input task and consciousness-raising task can be best combined with explicit teaching techniques. 4. The present study 4.1. Research questions Although previous research supports the idea that explicit instruction with some sort of input enhancement activities is effective, it does not suggest how input-based tasks can be best combined with explicit teaching techniques (e.g., providing proactive explicit information). Moreover, no studies have looked into how different input-based instruction varies according to the method of testing in L2 pragmatics. To address these gaps, this study will investigate the following research questions: 1. What are the relative effects of comprehension-based instruction, consciousness-raising instruction, and structured input instruction on the development of Japanese learners’ pragmatic proficiency? 2. To what extent do the effects of the different input instruction vary according to the method of testing? 4.2. Participants Prospective participants were sought through an employment advertisement in a weekly magazine, From A, as well as on the internet, From A navi, in Japan, both of which target students. After checking applicants’ scores on the Test of English for International Communication (TOEIC), only intermediate English proficiency level (TOEIC scores 500– 700) learners were selected for inclusion in the study. Sixty participants were randomly assigned to one of four groups consisting of three treatment groups and a control group (n = 15). The three treatment groups included a comprehension-based instruction group (CB) (n = 15), a consciousness-raising instruction group (CR) (n = 15), and a structured input instruction group (SI) (n = 15). The participants had studied English from 5 to 22 years, had Japanese as their first language, and were in the 18–40 year age range. 4.3. Target structures According to studies about learners’ requestive strategies conducted by Faerch and Kasper (1989), Hill (1997), and Takahashi (1998, 2001), non-native speakers of English typically lack the L2 pragmalinguistic knowledge to enable them to mitigate English by means of lexical/phrasal downgraders (subjectivizer and downtoner) and syntactic downgraders (aspect and tense). Thus, this study focuses on teaching lexical/phrasal downgraders and syntactic downgraders in English requestive forms. Lexical/phrasal downgraders soften the imposition force of a request by modifying the head act internally, whereas syntactic downgraders modify the head act internally by mitigating the level of imposition of a request through syntactic choices. A list of internal modifiers (adapted from Blum-Kulka et al., 1989: 273–286) is shown below. Internal modifier

Example

Internal syntactic downgraders Aspect (durative aspect marker) Tense

I am wondering if you could lend me a book. I wanted to ask you to come here.

Internal lexical and phrasal downgraders Subjectivizer Downtoner

I wonder if you could come here. I’m afraid you are going to have to move your desk. Could you possibly lend me your textbook?

1034

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

4.4. Instructional treatments Each teaching session for the four groups lasted 40 min. All directions were given in Japanese during the instruction. The sessions were conducted twice weekly by the same instructor for 2 weeks at a major English conversation school in Osaka, Japan. The three instructional treatments were matched for target structure and all four groups were matched for the amount of instructional time. The first class for all the treatment was on lexical/phrasal downgraders of English requests, the second on syntactic downgraders, the third on a review of the first class, and the fourth class on a review of the second class. The comprehension-based instruction treatment consisted of two components: (a) teacher-fronted explanation of the target downgraders, and (b) structured input tasks. In each lesson, the CB group was given handouts with a brief summary and examples of the target structures in English. The instructor took approximately 10 min to read the summary and examples aloud in English and then explained the summary and examples in Japanese. Then, the group engaged in the structured input tasks, which consisted of referential and affective activities (see Appendix). The group was given handouts with three referential activities and three affective activities. In the referential activities, participants read each situation and dialogue and selected the more appropriate request form out of the two choices offered for each underlined part. Then, they listened to an oral recording of the dialogue and underlined the actual request. In the affective activities, participants read each dialogue in the handouts and then listened to an oral recording. Secondly, they were asked to rate the level of appropriateness of each underlined request on a five-point Likert scale. The consciousness-raising intstruction treatment consisted of three components: (a) form-comparison; (b) answering analysis questions and (c) listing ways in which the requests in two dialogues differed. In each lesson, participants formed small groups of two to four and were given handouts which contained six English dialogues. Participants read and subsequently listened to each situation and dialogue. Secondly, they were asked to copy and compare the underlined request forms in the two dialogues. Then, they were asked to first find, and then discuss and comment (in Japanese) on the differences between the two requests. The groups also answered three analysis questions regarding the relationship between the characters in the dialogues, the difficulty of the requests, and the appropriateness of the requests in both dialogues. Finally, they were asked to list ways in which the requests differed between dialogues. After answering the analysis questions and making the lists in their groups, each group leader presented the answer to the analysis questions and the list to the whole class in English. The structured input instruction treatment consisted of only one component: participants engaged in the same structured input tasks as the CB group did. Lessons for the control group were designed to help participants do well on the TOEIC. Participants in this group engaged in reading comprehension exercises and were not exposed to the target structures at all. 4.5. Testing instruments and procedures This study used a pre-, post-, and follow-up test. The pre-test was administered 2–3 days prior to the instructional treatment, the post-test 8–9 days after the treatments and the follow-up test in the fourth week following instruction. Each test consisted of two input-based tests, such as the listening test (LT) and acceptability judgment test (AJT), and two output-based tests, such as the discourse completion test (DCT) and role-play test (RP). All situations in the four testing instruments had one speech act (requests) with three sociolinguistic variables (Power, degree of Imposition, and Distance) examined. The present study will focus attention on situations with a high level of Imposition (HI) combined with a lower level of requestor Power and a greater relational Distance. Situations with low level of Imposition (LI) were added as distractors in order to increase the reliability of each instrument. The DCT, RP, and AJT consist of 20 situations (10 HI items and 10 LI items), while the LT consists of 12 situations (six HI items and six LI items). The situations with a high level of imposition validated by Hill (1997), Hudson et al. (1995) and Takahashi (1998, 2001) were modified and used incorporating also a range of interlocutors related to two contexts; student life outside of school and student life at school. Three versions (A, B, and C) of the four tests (the DCT, RP, LT, and AJT) were developed and counterbalanced for order of presentation of the same situations across the pre-tests, post-tests, and follow-up tests. Three versions were used so that any test learning effect would be minimized. The pre-tests, the posttests, and the follow-up tests were administered in the following order; the DCT, RP, LT, and AJT. The two input-based

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

1035

tests were administered last because of concern that they might provide participants with models that could be used in the production tests. Although participants were instructed to complete the four tests within 2 h, only the LT had a timing constraint of 2 s per item for responses. The DCT required participants to read in English short descriptions of 20 situations and asked them to write in English what they would say in each situation. Although there was no time limit for the DCT, most participants took 30–45 min. Each response was rated by two native speakers of English according to the appropriateness of the request forms on a five-point Likert scale. The test contained 20 items, the maximum score being 100. (A sample item from the DCT is provided in Appendix). The RP consisted of short descriptions of 20 situations written in English with a Japanese supplement and required participants to play particular roles with the instructor. The instructor, a non-native speaker of English, was chosen as an interlocutor because Ellis (2003) has suggested that a testing situation where the examinees interact with other nonnative speakers rather than with native speaker examiners is more likely to elicit the examinees’ best performances. Prior to the role-play participants were given role cards describing the situations and their roles, and were required to start the role-play by asking for something from their interlocutors. Participants on average took 2–3 min preparing for each role-play, which was tape-recorded and their individual performance was rated for appropriateness of request forms on a five-point Likert scale by two native speakers of English. The test contained 20 items, the maximum score being 100. (Sample item provided in Appendix). The LT consisted of 12 situations in English. It required participants to listen to a dialogue between a Japanese university student and a native speaker of English in 12 different situations and then to score the appropriateness of the Japanese university student’s request forms on a five-point Likert scale. The test involved an audio-recording of the dialogue and had a timing constraint of 2 s per item for responses. When a participant rated a request appropriately in accordance with English native speakers’ baseline data, five points were given. The test contained 12 items, the maximum score being 60. (Sample item in Appendix). The AJT was a computer-delivered test. Participants were required to read written descriptions in English of 20 situations with a Japanese supplement. Participants were then presented with a series of isolated requests and instructed to score the first request on an 11-point Likert scale and then to score subsequent responses proportionally higher or lower in accordance with the degree of perceived acceptability. When a participant rated a request appropriately in accordance with English native speakers’ baseline data, five points were given. The test contained 20 items, the maximum score being 100. One sample item is shown below. 5. Reliability and validity Interrater reliability was estimated by examining the extent to which the raters’ scores correlated with each other. The correlation coefficients for the DCT and RP were .995 and .994, respectively, which were statistically significant ( p < .001). Cronbach alpha reliability estimates for each measure were calculated (see Table 2). The internal consistency estimates for all the tests are fairly high with the exception of the LT, which was low in reliability because of eight problematic items. By deleting the 8 problematic items out of the 20 items, a higher level of reliability was achieved. However, the reliability estimate for the LT is still somewhat lower than the others and this might be related to the

Table 1 Treatment features of each group. Group

Treatment

Teacher-fronted explanation

Comprehension-based instruction (CB) (n = 15)

Teacher-fronted explanation (10 min) + referential activities (18 min) + affective activities (12 min)

Yes

Consciousness-raising instruction (CR) (n = 15)

Form-comparison (10 min) + answering analysis question (10 min) + listing ways in which the requests in two dialogues differed (10 min) + discussion (10 min)

No

Structured input instruction (SI) (n = 15)

Referential activities (24 min) + affective activities (16 min)

No

Control group (n = 15)

TOEIC reading comprehension exercises (40 min)

No

1036

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

Table 2 Cronbach alpha reliability for each test.

DCT RP LT AJT

Version A

Version B

Version C

Total

.942 .935 .791 .868

.919 .950 .714 .858

.911 .915 .801 .881

.917 .926 .853 .893

Table 3 Factor analysis for DCT, RP, LT, and AJT. Variable

Factor 1

Factor 2

Factor 3

h2

DCT RP LT AJT Proportion of variance %

.95 .95 .26 .16 47.27

.21 .21 .95 .16 25.43

.15 .13 .17 .98 25.45

.96 .96 1.00 1.00 98.15

narrower rating scales in this test. That is, the LT used a five-point Likert scale, while the AJT used a 11-point Likert scale. According to Hatch and Lazaraton (1991), a broader range in scale encourages more precision in respondents’ judgments. Construct validity was examined on the basis of a factor analysis shown in Table 2. The factor analysis produced a four-factor solution but examination of the eigenvalues greater than 1 and scree plots showed that a three-factor solution was appropriate for the data and varimax rotations provided fairly clear patterns of loadings. According to Table 3, the DCT and RP tests loaded highest on Factor 1, the LT loaded highest on Factor 2, and the AJT loaded highest on Factor 3. One possibility is that Factor 1, Factor 2, and Factor 3 seem to be related to the test method. That is, the DCT and RP tests use productive test types (i.e., they require subjects to produce written or oral language), while the AJT and LT tests are judgment test types (i.e., they require subjects to select answers and they are not required to produce any language). The LT is a listening-based judgment test and the AJT is a computer-based judgment test. Thus, Factor 1 might be considered a productive test method factor, Factor 2 a listening-based judgment test method factor, and Factor 3 a computer-based judgment test method factor. 6. Results The results of MANOVA for the HI items reveal a significant effect for Instruction (F = 72.57, p = .000), whereas the results of MANOVA for the LI items indicate no significant effects for Instruction (F = 1.43, p = .181). Therefore, further exploration of the instructional effect for the HI items is warranted, while there is no warrant for any further examination for the LI items. The following section examines the results for each testing instrument more closely. In the data analysis a Bonferroni adjustment was employed in order to maintain an approximate experiment-wise .05 alpha level. In other words, the overall alpha level was set at .05, but with four group comparisons (the DCT, RP, LT, and AJT) for one item type (HI item). Therefore, .05 was divided by the number of comparisons (four), resulting in a p value of .0125 for individual statistical decisions. 6.1. Results from discourse completion test The results of a two-way ANOVA with repeated-measures reveal a significant main effect for Instruction (the CB, CR, and SI), F(3, 56) = 99.92, p = .000, a significant main effect for Time (the pre-test, post-test, and follow-up test), F(3, 56) = 583.57, p = .000, and a significant interaction effect between Instruction and Time, F(9, 56) = 48.64, p = .000. Results displayed in Fig. 1 show that although there are no statistically significant differences among the four groups on the pre-test scores, F(3, 56) = 1.06, p = .373, the three treatment groups made gains from the pre-tests to the post-tests and the follow-up tests, F(3, 56) = 127.12, p = .000; F(3, 56) = 178.42, p = .000, and positive effects for the three treatments were maintained, F(2, 42) = .58, p = .566. Therefore, the effects of Instruction were greater on the post-test and the follow-up test than on the pre-test. Furthermore the interaction shows the relative superiority of the

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

1037

Fig. 1. Interaction plot for DCT. Note. CB = comprehension-based instruction; CR = consciousness-raising instruction; SI = structured input instruction.

three treatment groups over the control group with no crossovers between the three treatment groups and the control group after the treatments. Post hoc Scheffe´ tests conducted on the post-test and follow-up test scores for the main effect for treatment show the following contrasts: The CB, CR, and SI groups perform significantly better than the control group; there are no significant differences among the three treatment groups. 6.2. Results from role-play test The results of a two-way repeated-measures ANOVA show a significant main effect for Instruction, F(3, 56) = 83.93, p = .000, a significant main effect for Time, F(3, 56) = 502.61, p = .000, and a significant interaction effect between Instruction and Time, F(9, 56) = 33.88, p = .000. Results displayed in Fig. 2 indicate that although there are no statistically significant differences among the four groups on the pre-test scores, F(3, 56) = .54, p = .654, the three treatment groups made gains from the pre-tests to the post-tests and the follow-up tests, F(3, 56) = 161.05, p = .000; F(3, 56) = 133.90, p = .000, and positive effects for the three treatments were kept, F(2, 42) = 161.05, p = .564. Accordingly, the effects of Instruction were greater on the post-test and the follow-up test than on the pre-test, and the interaction shows the relative superiority of the three treatment groups over the control group with no crossovers between the three treatment groups and the control group after the treatments. Post hoc Scheffe´ tests performed on the post-test and follow-up test scores indicate the following

Fig. 2. Interaction plot for RP. Note. CB = comprehension-based instruction; CR = consciousness-raising instruction; SI = structured input instruction.

1038

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

Fig. 3. Interaction plot for LT. Note. CB = comprehension-based instruction; CR = consciousness-raising instruction; SI = structured input instruction.

contrasts: The CB, CR, and SI groups perform significantly better than the control group; there are no statistically significant differences among the three treatment groups. 6.3. Results from listening test The results of a two-way repeated-measures ANOVA show a significant main effect for Instruction, F(3, 56) = 27.48, p = .000, a significant main effect for Time, F(3, 56) = 81.27, p = .000 and a significant interaction effect between Instruction and Time, F(9, 56) = 7.97, p = .000. The results displayed in Fig. 3 reveal that although there are no statistically significant differences among the four groups on the pre-test scores, F(3, 56) = 1.91, p = .139, the three treatment groups made gains from the pre-tests to the post-tests and the follow-up tests, F(3, 56) = 33.98, p = .000; F(3, 56) = 36.18, p = .000. However, Fig. 3 shows that unlike the CR and SI groups, the CB group did not maintain the positive effects between the post-test and follow-up test, F(2, 42) = 8.36, p = .001. A separate one-way ANOVA performed on the follow-up test scores shows a significant difference, F(2, 42) = 13.10, p = .000, in how the three treatment groups performed on this test. Although only the CB group did not retain the positive effects between the post-test and follow-up test, the effects of Instruction were greater on the post-test and the follow-up test than on the pre-test. Furthermore the interaction shows the relative superiority of the three treatment groups over the control group with no crossovers between the three treatment groups and the control group after the treatments. Post hoc Scheffe´ tests performed on the post-test and follow-up test scores indicate the following contrasts: the CB, CR, and SI groups perform significantly better than the control group on the post-test and follow-up test; there are no statistically significant differences among the three treatment groups on the post-test; there are no statistically significant differences between the CR and SI groups on the follow-up test; however, the CR and SI groups are significantly better than the CB group on the follow-up test. 6.4. Results from acceptability judgment test The results of a two-way repeated-measures ANOVA indicate a significant main effect for Instruction, F(3, 56) = 7.32, p = .000, a significant main effect for Time, F(3, 56) = 43.07, p = .000, and no significant interaction effect between Instruction and Time, F(9, 56) = 3.21, p = .000. The results displayed in Fig. 4 show that although there are no statistically significant differences among the four groups on the pre-test scores, F(3, 56) = .33, p = .801, the three treatment groups made gains from the pre-tests to the post-tests and the follow-up tests, F(3, 56) = 16.56, p = .000; F(3, 56) = 14.69, p = .000, and positive effects for the three treatments were kept, F(2, 42) = 3.06, p = .058. Therefore, the effects of Instruction were greater on the post-test and the follow-up test than on the pre-test and the three treatment groups performed significantly better than the control group. Post hoc Scheffe´ tests on the post-test and follow-up test scores for the main effect from treatment reveal the

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

1039

Fig. 4. Interaction plot for AJT. Note. CB = comprehension-based instruction; CR = consciousness-raising instruction; SI = structured input instruction.

following contrasts: the CB, CR, and SI groups perform significantly better than the control group; there are no statistically significant differences among the three treatment groups. 6.5. Test effect results from comprehension-based instruction group A one-way within-subject ANOVAs of the four tests was performed. The result revealed a significant main effect for Test, F(3, 11) = 29.82, p = .000. A post hoc test was also conducted and the results indicated the following contrasts ( p < .05): DCT; RP; AJT > LT The participants performed significantly better on the DCT, RP, and AJT than on the LT. 6.6. Test effect results from consciousness-raising instruction group A one-way within-subject ANOVAs of the four tests showed a significant main effect for Test, F(3, 11) = 4.70, p = .006. In addition, the result of a post hoc test indicated the following contrasts ( p < .05): DCT; RP; AJT > LT As in the CB, the results showed that the participants performed significantly better on the DCT, RP, and AJT than on the LT. 6.7. Test effect results from structured input instruction group A one-way within-subject ANOVAs of the four tests revealed a significant main effect for Test, F(3, 11) = 20.48, p = .000, and a post hoc test indicated the following contrasts ( p < .05): DCT; RP > AJT > LT The participants performed significantly better on the DCT and RP than on the LT and AJT and the AJT scores were significantly higher than the LT scores. 7. Discussion The three different types of input-based instruction proved to be effective. In addition, the results for the two types of input-based task indicate that they are equally effective and show that the development of L2 pragmatic proficiency can be influenced by manipulating input.

1040

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

There are two possible reasons for the effectiveness of the structured input task and the consciousness-raising task. One is the possibility that the two different treatments directed the participants’ attention to form in the input that they received. Although the two treatments involved different conditions, they may have made the target structures equally salient. The participants in the SI engaged in referential and affective activities where they needed to pay attention to the target linguistic forms and meanings they conveyed. In the referential activities, the participants chose the more appropriate request form from the two offered for each highlighted part, and in the affective activities they rated the level of appropriateness of each bold-faced underlined request in accordance with their beliefs or opinions. On the other hand, the CR participants were asked to pay attention to the highlighted requests in two dialogues, copying them and comparing the request in one dialogue with that of the other dialogue. Then, they were asked to discuss the differences in form and meaning between the two requests. Therefore, both treatments are likely to have led to the participants to pay conscious attention to the linguistic forms. The second possible reason for the effectiveness of the two types of input-based tasks is related to the deeper processing that arises where meaning is involved. The relationship between level of processing and long-term memory retention was discussed by Craik and Lockhart (1972), who claimed that the quality of a memory trace depends on the level or depth of perceptual and mental processing, where meaning plays a very important role. In other words, when the participants focus more on the meaning of the target feature, they are inclined to process it at a deeper level, which leads to greater retention. As in prior studies that looked into the effects of structured input tasks, the structured input tasks in the present study were also designed to focus heavily on meaning. On the other hand, as in prior studies that investigated the effects of the consciousness-raising tasks, the participants in the present study also had the opportunity to discuss the target feature with each other. During such activities the meaning of the target structure was emphasized, allowing for processing at a deep level by the participants. Research question 1 asked about the relative effects of comprehension-based instruction, consciousness-raising instruction, and structured input instruction on the development of the pragmatic proficiency of Japanese L1 learners of English. The results of this study indicate that three types of instruction seem to have similar effects on the development of Japanese learners’ pragmatic proficiency as measured by the DCT, RP, and AJT. However, regarding the LT, the CB, CR, and SI groups resulted in significant improvement in test scores, but, while the CR and SI groups maintained the improvement in the follow-up test, the CB group did not and the difference between the three types of instruction in the follow-up test was significant. Why did the CB group perform as well as the CR and SI groups on the post-test of the LT but not as well on the follow-up test of the LT when no difference was evident on any of the other tests? Any answer to this question is necessarily speculative as no information relating to the psycholinguistic processing involved in either the treatments or the test is available. What distinguishes the LT from the other tests is that it calls for on-line processing. Such tests are demanding on working memory, as participants have to process and respond to the stimuli rapidly. The CB, CR, and SI treatments can be assumed to have provided participants with explicit knowledge but they differed in how this was achieved: In the case of the CB treatment, the participants were simply given explicit information, and in the case of the CR and SI treatments, they had to discover the rules for themselves and also had the opportunity to talk about the data. It is possible, then, that the CR and SI participants attended to the forms and meanings of the target structures more deeply than the CB participants, that is, the CR and SI treatments involved greater depth of processing, resulting in knowledge that was more firmly embedded and thus more easily accessed. Now, in all the immediate post-tests, this difference was not apparent because the explicit knowledge was fresh in the participants’ memories. However, in the follow-up test of the LT, they were less successful in accessing this explicit knowledge. The CR and SI participants were still able to cope with the demands of this test as the explicit knowledge they had learned was firmly entrenched. The CB participants were not able to cope with the test to the same extent because the demands of the test were taxing on their working memories, making it difficult for them to access their more weakly established explicit knowledge. Research question 2 asked about to what extent three types of input-based instruction varies according to the method of testing. The results revealed a significant main effect for Test and the participants for the CB and CR performed significantly better on the DCT, RP, and AJT than on the LT, whereas the participants for the SI performed significantly better on the DCT and RP than on the LT and AJT while the AJT scores were significantly higher than the LT scores. The LT scores were the lowest among the four tests for three treatment groups and there are two possible explanations for this result. The first is associated with the complexity of judgment-type tests. The LT is a judgment test, which required the participants to listen to a dialogue between a Japanese university student and a native speaker of English and then to score the appropriateness of the Japanese university student’s request forms on a five-point Likert scale. A problem with this

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

1041

test is that some participants chose a neutral option (like the three on the five-point Likert scale). Birdsong (1989) noted that learners with low proficiency lack confidence in their judgment with the result that their choice hovers around neutrality. Another possible explanation is related to time pressure. Because this test required participants to judge the appropriateness of the stimuli under time pressure (2 s), it is likely to have required participants to access their proceduralized knowledge of the target structures. In other words, in the LT, the participants experienced difficulty in accessing the knowledge they had acquired from the treatments because this knowledge was not yet fully proceduralized. The results of the present study do not support the theoretical claims of skill-learning theory, but they lend support to the theoretical claims of information-processing theory. Skill-learning theory asserts that a particular skill is learned as a result of practice. According to Anderson’s (1983) ATC theory, language learning is like any other kind of skill learning in that it involves a progression from declarative knowledge to procedural knowledge through extensive practice in using the L2. While declarative knowledge is not automatic, procedural knowledge is. In skill-learning theory, practice through input-based instruction only develops the participants’ ability to comprehend the target features and does not lead to development of the participants’ ability to produce the target structures. However, in the present study, practice through input-based instruction in the CB, CR, and SI developed the participants’ ability to comprehend and produce the target structures in the DCT, RP, LT, and AJT. This is in line with information-processing theory, which claims that input-based instruction will serve to develop the participants’ ability to comprehend and produce the target features drawing on the same underlying knowledge source. Robinson’s review of informationprocessing theories (1995) explains that general cognitive mechanisms process information in the input to achieve a mental representation of the target structures. This knowledge is then accessed through other cognitive mechanisms and enables the participants to comprehend and produce the target structures. In the present study, the three inputbased instruction groups performed significantly better than the control group where both comprehension and production were concerned, which means that the input-based instruction developed the same underlying knowledge source that is called on in both comprehension and production. 8. Conclusion The present study examined the relative effects of input-based tasks and also the method of testing in L2 pragmatics. This study can be considered to be of practical use especially in a foreign language context where learning English pragmatics rather than English grammar has become one of the most significant area of focus, and where exposure to English and class time available for teaching English is limited. The present study indicates that input-based tasks can work effectively when they provide an emphasis on forms and meanings in teaching pragmatics in an EFL context. The pedagogical implication then for teachers is that they should be aware that effective learning occurs when the tasks used provide learners with the opportunity for processing both the forms and meanings of the target features. It is possible, then, that structured input tasks and consciousness-raising tasks can be used together and can complement each other. Such tasks may prove of great value in improving learners’ pragmatic proficiency. The present study suggests that there is a limitation that future research needs to consider. The assessment instruments in this study consisted of four different kinds of tests: a planned written-production test, a planned oralproduction test, an unplanned oral-judgment test, and a planned written-judgment test. For the unplanned judgment test, a listening test was used as it required automatic processing because of the time constraint provided for the task. For the planned judgment test, a reading test without time constraint was used as the participants were likely to resort to controlled processing because they had plenty of time for the task. By the same token, for the planned writtenproduction test and planned oral-production test, a writing test and a role-play test without any time constraint were used. Accordingly, for the assessment of controlled processing skills, judgment and production modes of testing were adopted, while for the assessment of automatic processing skills, only the judgment mode was used. In view of the fact that judgment skills are examined through listening and reading tests, whereas production skills were measured by means of a speaking and a writing test in the present study, an on-line speaking test should ideally have been added in order to measure automatic processing skills. Since the time constraint provided for the production tests plays a significant role in distinguishing spontaneous language processing from controlled language processing, the on-line speaking test would have served as the counterpart of the on-line listening test. Finally the author hopes that the results of the present study will provide researchers and teachers with clearer guidelines for effective ways of teaching pragmatics in an EFL context.

1042

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

Acknowledgement I am very much grateful to Professor Rod Ellis for his skillful guidance and great support. I would also like to thank the anonymous reviewers for their invaluable feedback and suggestions. Appendix A. Examples of instructional materials used A.1. Structured input task Referential activity: Read the following situation and the dialogue and choose the more appropriate request form out of two offered for each underlined part and indicate your choice by circling ‘(a)’ or ‘(b)’. Then, listen to an oral recording of the dialogue and indicate whether the actual request used in the dialogue is ‘(a)’ or ‘(b)’. Situation: Yuka is about to start her car when she notices that her car battery has gone flat. She needs to go to school now and she does not have any other means but to ask her landlord, Mr. Brown, whom she has never spoken to before, to give her a ride to school. Her landlord is extremely busy, but she decides to ask her landlord to drive her to school. Brown: Yuka: Brown: Yuka:

Brown:

Hello. Hi, you are Mr. Brown, aren’t you? That’s right. I’m a tenant next door. My car battery has just gone flat and I can’t start my car. I really need to get to school. 1. (a) I was just wondering if I could by any chance get a lift; (b) I am just wondering if I could by any chance get a lift. Well, actually, I am really busy helping other tenants moving into this apartment. So, I can’t really help you.

Affective activity: Read the following situation and the dialogue and answer the following questions. Situation: John is living in an apartment. He is extremely busy working on his assignment, but he needs to send a big parcel to England today. His landlady, Mrs. Taylor, whom he has never spoken to before, is extremely busy, but he decides to ask his landlady to send the big parcel. John sees the landlady. John: Taylor: John: Taylor: John: Taylor: John: Taylor:

Hi, you are Mrs. Taylor, aren’t you? That’s right. Hello. My name is John. Oh, you are the tenant. Yes. I live next door. How is it going? Pretty good, thank you. I’m very busy working on my assignment. 1. I wondered if I could possibly ask you a favor. What’s the favor?

Indicate the appropriateness level of four underlined requests from your point of view on a scale as below. 1. very unsatisfactory 1—2—3—4—5 completely appropriate A.2. Consciousness-raising task Read the dialogue A and the dialogue B and answer the following questions. Dialogue A. Situation: Yuka is about to start her car when she notices that her car battery has gone flat. She needs to go to school now and she does not have any other means but to ask her neighbor, James, whom she knows well, to give her a ride to school. She sees her neighbor go out by car and she decides to ask her neighbor to drive her to school.

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

James: Yuka: James:

1043

Hi, Yuka. Hi, Jim. My car battery has just gone flat and I can’t start my car. I really need to get to school. (a) Could I have a lift ? Sure. Get in.

Dialogue B. Situation: Yuka is about to start her car when she notices that her car battery has gone flat. She needs to go to school now and she does not have any other means but to ask her landlord, Mr. Brown, whom she has never spoken to before, to give her a ride to school. Her landlord is extremely busy, but she decides to ask her landlord to drive her to school. Brown: Yuka: Brown: Yuka: Brown:

Hello. Hi, you are Mr. Brown, aren’t you? That’s right. I am a tenant next door. My car battery has just gone flat and I can’t start my car. I really need to get to school. (a)’ I wonder if I could get a lift. Well, actually, I am really busy helping other tenants moving into this apartment. So, I can’t really help you.

1. Write up the requests Yuka makes in the dialogue A and the dialogue B in the table below and also comment on the differences between the two.

2. Indicate the level of relationship between Brown and Yuka in the dialogue A and the dialogue B? the dialogue A: very distant 1—2—3—4—5 very close the dialogue B: very distant 1—2—3—4—5 very close 3. In consideration of respective relationship between Brown and Yuka in the dialogue A and the dialogue B, do you think that Yuka is asking for something difficult or simple in the dialogue A and the dialogue B? the dialogue A: very simple 1—2—3—4—5 very difficult the dialogue B: very simple 1—2—3—4—5 very difficult 4. Look at the requests you found in #1 and answer the following questions. (1) Do you think Yuka is being polite in the dialogue A? Indicate the appropriateness level of the requests by Yuka in the dialogue A. very unsatisfactory 1—2—3—4—5 completely appropriate (2) Do you think Yuka is being polite in the dialogue B? Indicate the appropriateness level of the request by Yuka in the dialogue B. very unsatisfactory 1—2—3—4—5 completely appropriate 5. How does Yuka try to be polite making requests? Appendix B. Examples of test items B.1. Discourse completion test You are writing a difficult paper for Professor Hill. You need some help with the paper but Professor Hill is away for a month. A friend of yours has suggested you go and see Professor Watson. Although you do not know Professor Watson and Professor Watson is extremely busy, you have decided to ask Professor Watson to look

1044

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

through your long paper before you hand it in the next day. What would you ask Professor Watson? (based on Takahashi, 1998, 2001) Note. imposition = +; power = ; distance = +; + = more; = less;  = equal

B.2. Role-play test You must write an essay on a topic and turn it in by tomorrow. You have just found a very long interesting article on the topic, but it is written in Chinese. You can’t read the article in Chinese and you need to translate it into Japanese. You have just heard that there is a student in another department who is bilingual in Chinese and Japanese. Although you have never spoken to the student before, you decide to ask the student to translate the article into Japanese. (based on Hill, 1997) Note. imposition = +; power = ; distance = +; + = more; = less;  = equal Now: You see the student. B.3. Listening test Taro is working in a restaurant. The owner has asked Taro to get each customer to complete a very long questionnaire about the quality of the food and the service in the restaurant. Taro has given the questionnaire to a customer. Taro notices that the customer has not filled it in but is about to leave in a hurry. Taro needs to have the questionnaire filled by the customer. (based on Hudson et al., 1992, 1995) Note. imposition = +; power = ; distance = +; + = more; = less;  = equal Taro: Excuse me. I can see you are in a hurry. But please fill in this questionnaire. Brown: Oh, look. I’m sorry I really haven’t got the time. not appropriate at all 1—2—3—4—5 completely appropriate B.4. Acceptability judgment test Professor King at your university is a famous psychologist. You are now reading one of Professor King’s books and finding it very complicated. You would like to ask Professor King some questions about the book. Professor King does not know you and Professor King is extremely busy. However, you decide to go and ask Professor King to spare you some time for some questions. What would you ask Professor King? (based on Takahashi, 1998, 2001) Note. imposition = +; power = ; distance = +; + = more; = less;  = equal a: I want to ask you some questions. not appropriate at all 0—1—2—3—4—5—6—7—8—9—10 completely appropriate b: I was wondering if it would be possible for me to ask you some questions. not appropriate at all 0—1—2—3—4—5—6—7—8—9—10 completely appropriate c: Could I possibly ask you some questions ? not appropriate at all 0—1—2—3—4—5—6—7—8—9—10 completely appropriate References Allen, Linda, 2000. Form-meaning connections and the French causative: an experiment in processing instruction. Studies in Second Language Acquisition 22 (1), 69–84. Anderson, John, 1983. The Architecture of Cognition. Harvard University Press, Cambridge, MA. Beebe, Leslie, Takahashi, Tomoko, Uliss-Weltz, Robin, 1990. Pragmatic transfer in ESL refusals. In: Scarcella, Robin, Anderson, Elaine, Krashen, Stephen (Eds.), Developing Communicative Competence in a Second Language. Newbury House, New York, pp. 55–73.

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

1045

Benati, Alessandro, 2003. The effects of structured input activities and explicit information on the acquisition of Italian tense. In: VanPatten, Bill (Ed.), Processing Instruction. Erlbaum, Hillsdale, NJ. Blum-Kulka, Shoshana, House, Juliane, Kasper, Gabriele, 1989. Cross-cultural Pragmatics: Requests and Apologies. Ablex, Norwood, NJ. Birdsong, David, 1989. Metalinguistic Performance and Interlinguistic Competence. Springer-Verlag, Berlin. Collentine, Joseph, 1998. Processing instruction and the subjunctive. Hispania 81, 576–587. Craik, Fergus, Lockhart, Robert, 1972. Levels of processing: a framework for memory research. Journal of Verbal Learning and Verbal Behavior 11, 671–684. DeKeyser, Robert, 1995. Learning second language grammar rules: an experiment with a miniature linguistic system. Studies in Second Language Acquisition 46 (4), 613–642. DeKeyser, Robert, Sokalski, Karl, 1996. The differential role of comprehension and production practice. Language Learning 46 (4), 613–642. Ellis, Rod, 1997. SLA Research and Language Teaching. Oxford University Press, Oxford. Ellis, Rod, 2003. Task-based Language Learning and Teaching. Oxford University Press, Oxford. Faerch, Claus, Kasper, Gabriele, 1989. Internal and external modification in interlanguage request realization. In: Blum-Kulka, Shoshana, House, Juliane, Kasper, Gabriele (Eds.), Cross-cultural Pragmatics: Requests and Apologies. Albex, Norwood, NJ, pp. 221–247. Farley, Andrew, 2003. Processing instruction and the Spanish subjunctive: is explicit information needed? In: VanPatten, Bill (Ed.), Processing Instruction. Erlbaum, Mahwah, NJ. Fotos, Sandra, 1994. Integrating grammar instruction and communicative language use through grammar consciousness-raising tasks. TESOL Quarterly 28 (2), 323–351. Fotos, Sandra, Ellis, Rod, 1991. Communicating about grammar: a task-based approach. TESOL Quarterly 25 (4), 605–628. Hatch, Evelyn, Lazaraton, Anne, 1991. The Research Manual: Design and Statistics for Applied Linguistics. Heinle & Heinle, Boston, MA. Hazzard, Michael, 1999. The effects of instructional methodology on the comprehension of problematical grammar structures in Spanish. Unpublished doctoral dissertation. Temple University, Philadelphia, PA. Hill, Thomas, 1997. The development of pragmatic competence in an EFL context. Unpublished doctoral dissertation. Temple University, Japan. House, Juliane, 1996. Developing pragmatic fluency in English as a foreign language: routines and metapragmatic awareness. Studies in Second Language Acquisition 18 (2), 225–252. House, Juliane, Kasper, Gabriele, 1981. Zur rolle der kognition in kommunikationskursen [The role of cognition in communication courses.] Die Neueren Sprachen 80, 42–55. Hudson, Thom, Detmer, Emily, Brown, James, 1992. A framework for testing cross-cultural pragmatics (Technical Report #2). University of Hawai‘i Press, Honolulu, HI. Hudson, Thom, Detmer, Emily, Brown, James, 1995. Developing prototypic measures of cross-cultural pragmatics (Technical Report #7). University of Hawai‘i Press, Honolulu, HI. International Communication Incorporated, 1996. TOEIC Friends, 2, 37–41. Kim, Sook, 2001. Structured input and production practice in foreign/second language learning. Unpublished doctoral dissertation. Temple University, Japan, Tokyo. Krashen, Stephen, 1985. The Input Hypothesis: Issues and Implications. Longman, London. Nagata, Noriko, 1998. Input vs. output practice in educational software for second language acquisition. Language Learning & Technology 1 (2), 23–40. Norris, John, Ortega, Lourdes, 2001. Does type of instruction make a difference? Substantive findings from a meta-analytic review. In: Ellis, Rod (Ed.), Form-focused Instruction and Second Language Learning, Language Learning, 51 (Suppl. 1), 157–213. Robinson, Peter, 1995. Attention, memory, and the ‘noticing’ hypothesis. Language Learning 45 (2), 283–331. Robinson, Peter, 1996. Learning simple and complex second language rules under implicit, incidental, rule-search, and instructed conditions. Studies in Second Language Acquisition 18 (1), 27–67. Rose, Kenneth, 2005. On the effects of instruction in second language pragmatics. System 33 (3), 417–435. Rose, Kenneth, Ng, Connie, 2001. Inductive and deductive teaching of compliments and compliment responses. In: Rose, Kenneth, Kasper, Gabriele (Eds.), Pragmatics in Language Teaching. Cambridge University Press, New York, pp. 145–170. Salaberry, Rafael, 1997. The role of input and output practice in second language acquisition. Canadian Modern Language Review 53, 422– 451. Sharwood Smith, Michael, 1993. Input enhancement in instructed SLA: theoretical bases. Studies in Second Language Acquisition 15 (2), 165– 179. Swain, Merrill, 1985. Communicative competence: some roles of comprehensible input and comprehensible output in its development. In: Gass, Susan, Madden, Carolyn (Eds.), Input in Second Language Acquisition. Newbury House, Rowley, MA, pp. 235–253. Swain, Merrill, 1998. Focus on form through conscious reflection. In: Doughty, Catherine, Williams, Jessica (Eds.), Focus on Form in Classroom Second Language Acquisition. Cambridge University Press, Cambridge, pp. 64–81. Takahashi, Satomi, 1998. Quantifying requestive imposition: Validation and selection of situation for L2 pragmatic research. Studies in Languages and Cultures, vol. 9. Institute of Language and Cultures, Kyushu University, pp. 135–159. Takahashi, Satomi, 2001. The role of input enhancement in developing interlanguage pragmatic competence. In: Rose, Kenneth, Kasper, Gabriele (Eds.), Pragmatics in Language Teaching. Cambridge University Press, New York, pp. 171–199. Tanaka, Takeo, 2001. Comprehension and production practice in grammar instruction: does their combined use facilitate second language acquisition? JALT Journal 23 (1), 6–30. Tanaka, Yoshihiro, 1996. The comprehension and acquisition of relative clauses by Japanese High School students through formal instruction. Unpublished doctoral dissertation. Temple University Japan, Tokyo.

1046

M. Takimoto / Journal of Pragmatics 41 (2009) 1029–1046

Tateyama, Yumiko, Kasper, Gabriele, Mui, Lara, Tay, Hui-Mian, Thananart, Ong-on, 1997. Explicit and implicit teaching of pragmatic routines. In: Bouton, Lawrence (Ed.), Pragmatics and Language Learning, Monograph Series, vol. 8. Division of English as an International Language, University of Illinois, Urbana-Champaign, IL, pp. 163–178. Toth, Paul, 1997. Linguistic and pedagogical perspectives on acquiring second language morpho-syntax: a look at Spanish se. Unpublished doctoral dissertation, University of Pittsburgh, Pittsburgh, PA. Trahey, Martha, White, Lydia, 1993. Positive evidence and preemption in the second language classroom. Studies in Second Language Acquisition 15, 181–204. VanPatten, Bill, 1991. Grammar instruction and input processing. Paper presented at the special colloquium on the Role of Grammar Instruction in Communicative Language Teaching, Concordia University and McGill University, Montreal, July, 1991. VanPatten, Bill, 1993. Grammar teaching for the acquisition-rich classroom. Foreign Language Annals 26 (4), 435–450. VanPatten, Bill, 1996. Input Processing and Grammar Instruction in Second Language Acquisition. Ablex, Norwood, NJ. VanPatten, Bill, 2003. From Input to Output: A Teacher’s Guide to Second Language Acquisition. McGraw-Hill, New York. VanPatten, Bill, Cadierno, Teresa, 1993. Explicit instruction and input processing. Studies in Second Language Acquisition 15, 225–243. VanPatten, Bill, Oikkenon, Soile, 1996. Explanation versus structured input in processing instruction. Studies in Second Language Acquisition 18 (4), 495–510. Wong, Wynne, 2003. Processing instruction in French: the roles of explicit information and structured input. In: VanPatten, Bill (Ed.), Processing Instruction. Erlbaum, Mahwah, NJ, pp. 187–205. Wong, Wynne, VanPatten, Bill, 2003. The evidence is IN: Drills are OUT. Foreign Language Annals 36 (3), 403–423.