2.1 Issues in Pragmatics

CHAPTER 2

LITERATURE REVIEW

The literature review of this chapter is based on three theoretical frameworks.

The first one discusses issues in pragmatics, in which Speech Act Theory, Politeness Model, and the speech act of complaining will be introduced. The second one addresses issues in interlanguage pragmatics (ILP), which deals with pragmatic competence and related ILP studies. The last one concerns pragmatics in language teaching, which discusses the pedagogical interventions and instructional effects on learners’ interlanguage pragmatics development.

2.1 Issues in Pragmatics

According to Leech (1983), general pragmatics includes two major parts. The first part is sociopragmatics, which focuses on the sociological part of pragmatics.

The sociopragmatic studies are generally culture-specific in that the Cooperative Principle and the Politeness Principle operate variably in different cultures of language communities. The other part of pragmatics is pragmalinguistics, which deals with the more linguistic part of pragmatics (Leech, 1983).The pragmalinguistic studies are generally language-specific in that they consider the particular resources of a given language for conveying particular illocutions (Leech, 1983). Of all the issues concerning language usage, speech act theory has aroused widespread interests because the acquisition of speech acts is considered a prerequisite for the acquisition of language in general (Levinson, 1983).

Therefore, in this section, three important issues relevant to this study will be

discussed: the speech act theory, the politeness principles, and the speech act of

complaining. The first two issues are the major concerns in pragmatics, and the last

one particularly addresses the speech act under investigation in this study.

2.1.1 Speech Act Theory

The speech act theory began with the work by J. Austin (1962), who argues that language is used to perform actions instead of determining truth values. What is important for an utterance is to distinguish between sense, which refers to the meaning of the speaker's words, and force, which refers to the speaker's communicative intention. He further distinguished three types of forces: locutionary force, which refers to the actual sounds and words uttered; illocutionary force, which refers to the intention underlying the words; and perlocutionary force, which refers to the effect of the utterance on the hearer.

Searle (1969) drew on Austin's work and had two major contributions to the speech act theory. The first contribution has to do with formalization of speech acts, in which classification and felicity conditions of speech acts were proposed.

The second contribution relates to indirectness, which will be addressed in detail in this section because it is highly relevant to the present study.

Indirect speech acts occur when there is a mismatch problem between the expressed and the implied meaning (Thomas, 1995). Searle (1975: 60) stated that in indirect speech acts

...the speaker communicates to the hearer more than he actually says by way of relying on their mutually shared background information, both linguistic and nonlinguistic, together with the general powers of rationality and inference on the part of the hearer.

Searle further distinguished two kinds of indirectness in speech acts:

1 Searle (1975a) identified five major categories of communicative acts: representatives, directives, commissives, expressives and declarations. Representatives refer to the acts used to describe states of affairs (e.g. stating, asserting). Directives refer to the acts used to ask the hearer to do something (e.g.

requesting, ordering). Commissives refer to the acts used to commit the speaker to do something (e.g.

promising, guaranteeing). Expressives refer to the acts used to express the emotional feelings of the speaker (e.g. apologizing, thanking). Declarations refer to the acts used to change the status of state of affairs (e.g. appointing, naming).

nonconventional and conventional. Nonconventional indirectness is open ended in propositional content, linguistic form and pragmatic force (Blum-Kulka, 1989). Searle described ten processing steps for a hearer to arrive at the primary illocutionary force from the literal illocution. In Step 1, the hearer draws facts about the conversation. In Step 2, the hearer assumes that the speaker observes principles of conversational cooperation and that his/her remarks are therefore relevant to the conversation. In Step 3, the hearer assumes that the speaker produces a relevant response. In Step 4, the hearer recognizes that the speaker’s literal meaning of the utterance is not a relevant response inferred from Steps 1 and 3. In Step5, the hearer assumes that the speaker means more than he/she says. In Step 6, the hearer further judges the speaker’s utterances from factual background information, or, knowledge of the world. In Step 7, the hearer makes inferences from Step 6. In Step 8, the hearer draws on the preparatory condition to determine the propositional content the speaker conveys. In Step 9, the hearer infers from Step 1 (facts about conversation), Step 7 and Step 8 (theory of speech acts). In Step 10, the hearer arrives at the speaker’s primary illocutionary point. In general, the more indirect an utterance is, the more cognitive processing efforts are needed (Sperber & Wilson, 1995).

The other type of indirectness is conventional indirectness. According to

Blum-Kulka (1989), conventions of usage are represented in conventions of means

and conventions of forms. Conventions of means refer to the kinds of sentences that

are typically used. On the other hand, conventions of forms refer to the wording used

in the utterances. For example, Searle (1975: 65-7) identified six means to perform

conventionally indirect directives, which include the address to the H's ability to

perform A (e.g. “Can you reach the salt”), S's wish that H will do A (e.g. “I would like

you to go now”), H's doing A (e.g. “Will you quit making that awful racket”), H's

desire or willingness to do A (e.g. “Would you be willing to write a letter”), reasons

for doing A (e.g. “You should leave immediately”) and embedding of the previous elements into one another (e.g. “Might I ask you to take off your hat”).

Blum-Kulka (1989) further argues that the two types of indirectness differ in the inherent pragmatic ambiguities in each, which can be realized in terms of sentence meaning, utterance meaning and speaker meaning. Sentence meaning is similar to the interpretation assigned to linguistic structures only. Utterance meaning refers to the pairing of linguistic expressions with specific contexts. Speaker meaning refers to what the speaker intends to convey to make the hearer recognize his/her intentions.

She claims that the pragmatic ambiguity of non-conventional indirectness lies at the level of speaker meaning, while the pragmatic ambiguity of conventional indirectness exists at the level of utterance meaning.

To conclude, Austin’s pioneering works proposed the notion of speech acts.

Based on Austin’s work, Searle was devoted to the formalization of speech acts. His other contribution, indirectness, is closely related to politeness, which will be discussed in the next section.

2.1.2 Politeness Model

Fraser (1990) identified four current perspectives on politeness: the social-norm view, the conversational-maxim view, the face-saving view and the conversational-contract view. Of the four perspectives to a construct of politeness, the face-saving model advocated by Brown & Levinson (1987) has been best known.

Brown & Levinson started with the assumption that all competent interactants have

two kinds of properties: face and rationality. Drawing on Goffman's (1967) notion of

face, they claim that people cooperate in communication to maintain each other's face

(Brown & Levinson, 1987). In this model, there are two types of face wants. The first

type is negative face wants, which refer to "the basic claim to territories, personal

preserves, rights to non-distraction" (Brrown & Levinson, 1987: 61). The second type

is positive face wants, which refer to "the positive consistent self-image or personality claimed by interactants" (Brown & Levinson, 1987: 61). According to Brown &

Levinson, certain speech acts are face-threatening in that they may threaten the speaker's or hearer's negative or positive face. The other property of interactants prescribed in the model is rational capacities. Combining these two properties together, a competent adult, when considering a face-threatening act (FTA), must be able to assess the payoffs, which are usually influenced by three sociological variables:

the social distance between speaker and hearer, the relative power between the speaker and hearer, and the absolute ranking of impositions in a particular culture.

He/she may choose from the five superstrategies—Do the FTA without any redress, Do the FTA with positive politeness redress, Do the FTA with negative politeness redress, Do the FTA off record and Do not do the FTA.

The first three are on-record strategies. The first strategy is bald-on-record, which corresponds roughly to the maxims described in Grice's (1975) Cooperative Principle.

The second and third ones are on-record strategies with redressive actions, in which the speaker attempts to give face to the hearer. Positive politeness redress is approach-based in that the speaker intends to minimize the face threat by saying what the hearer wants. Negative politeness redress, on the other hand, is avoidance-based in that the speaker intends to maintain the hearer's basic wants of self determination.

The fourth strategy relates to hints, with which the hearer may not commit himself to particular intentions. The four strategies can be measured on a directness scale, with bald-on-record strategy the most direct, off-record strategy the most indirect, and

2 In Logic and Conversation, Grice (1975) proposed four conversational maxims, which include Maxim of Quantity, Maxim of Quality, Maxim of Relation and Maxim of Manner. Maxim of Quality refers to the amount of information a speaker should provide. The information should not be more or less than required. Maxim of Quality claims that a speaker should tell the truth, and do not say what he/she believes to be false. Maxim of Relation refers to the relevance of information a speaker should provide to the course of conversation. Maxim of Manner states that a speaker should be concise and orderly when speaking and avoid obscurity and ambiguity of expressions.

redressive strategies in between. The last strategy relates to opting out, in which the speaker avoids performing the FTA. The payoffs for this strategy choice are that no offense is made to the hearer. However, the speaker also fails to achieve desired outcomes in an interaction.

Grundy (2000: 158) provides good examples of how these strategies can be used in performing the speech act of requesting. For example, if someone parks the car right in front of your house, you could say:

(1) Bald on record: “Don't park your car outside our house anymore!”

(2) Positive politeness redress: “Bill, my old mate, I know you want me to admire your new car from my front room, but how about moving it across the road and giving yourself the pleasure?”

(3) Negative politeness redress: “I'm sorry to ask, but could you possibly park your car in front of your own house in future?”

(4) Off-record: “Is your car all right outside our house?”

(5) Do not do anything.

There are two important issues in the politeness model. One concerns the relationship between politeness and indirectness, and the other deals with universality and culture-specificity. In this model, Brown & Levinson claim that politeness parallels indirectness. In other words, they believe that the more indirect the strategy is, the more polite the speaker is. However, this claim has later been criticized by other researchers. Blum-Kulka (1987) argues that indirectness does not necessarily parallel politeness, and her argument has been supported by an experiment conducted in the Cross-cultural Speech Act Realization Project (CCSARP), where the Hebrew speakers rated non-conventionally indirect requests as the most impolite forms.

According to Blum-Kulka (1987), politeness can be observed from two

perspectives: the need for pragmatic clarity and the need to avoid coerciveness.

Take the speech act of requesting as an example. Blum-Kulka (1987) distinguished three types of requests found in the CCSARP: direct requests, conventionally indirect requests and non-conventionally indirect requests. In these three kinds of requests, conventionally indirect requests have been found to be the most polite forms and most frequently used across languages. This is because conventionally indirect requests strike a balance between pragmatic clarity and face threat. It is pragmatically clear because the hearer can recognize the requests through conventionalized linguistic forms without too much cognitive burden. For example, when hearing the sentence

"Can you open the door", the hearer would first appeal to the requestive interpretation unless there is a mismatch between the interpretation and the context. It is less face-threatening because it leaves the hearer an option to reject the request. Direct and non-conventional requests, on the other hand, are comparatively less polite in that the former lack face concern and that the latter lack pragmatic clarity. Because of the imbalance between pragmatic clarity and coerciveness, these two kinds of requests are considered to be less polite.

The other issue in the politeness model pertains to universality and culture-specificity. Although Brown & Levinson (1987: 244) acknowledge the cross-cultural variations in the general level and compositions of weightiness of an FTA, they principally claim for the universality of the politeness model, which amounts to the following three points:

(i) The universality of face, desirable as two kinds of wants.

(ii) The potential universality of rational action devoted to satisfying others' face wants.

(iii) The universality of the mutual knowledge between interactants of (i) and (ii).

However, there is a challenging line of research which argues that Brown &

Levinson's model is only applicable to the western cultures and not appropriate to the

non-western cultures (Gu, 1990; Mao, 1994; Matsumoto, 1988; Pan, 1995; Shi, 1994).

The most frequently criticized aspect is the notion of face, which has been seen differently in the western and non-western worlds. First, Brown & Levinson's notion of face differs from Goffman's face. Goffman (1967: 5) defines face as "the positive social value a person effectively claims for himself by the line others assume he has taken during a particular contact". In other words, Goffman regards face not as an internalized property, but as a public property endorsed by others' judgments in the flow of events. By contrast, Brown & Levinson (1987: 61) define face as "the public self-image that every member wants to claim for himself". They characterized face as an internalized property, which deviates from Goffman's theory considerably. Such a self-image description of face can be applied to the western cultures, but can cause problems in the non-western world (Mao, 1994). The problem of the western and non-western differences arises from the fact that the former values individual rights and the latter individual's relative position in a society. Matsumoto (1988) points out that in Japan, where the interdependence on each other is strong, the juniors must show deference to the seniors by acknowledging dependence, while the seniors must show the responsibility to take care of the juniors. Furthermore, Pan (1995) indicates that in Chinese society, deference is shown in only one direction: from lower level to higher level in the power hierarchy. Yu (2003) also notes that since there is a lack of recognition of relative power in American culture, certain speech acts such as requests and commands are considered very face-threatening. As a result, Americans tend to use more face redressive strategies. From these studies, it is obvious that the notion of face in western and non-western worlds is quite different.

Bearing the different notions of face in mind, it is questionable whether positive

and negative politeness described in Brown & Levinson's model is appropriate in a

non-western context. Matsumoto (1988) argues that Brown & Levinson's notion of

negative face wants sounds odd to the Japanese people. She states that the linguistic politeness system in Japanese can reflect this oddity. For example, in the formulaic expressions to show deference to the others, Japanese may say: "I ask you to please take care of me", which appears to be different from negative politeness through a Westerner's eyes. In Chinese, there are two words which carry the meaning of "face":

mianzi (面子) and lian (臉). Hu makes a functional distinction between these two terms. Mianzi refers to "a reputation achieved through getting on in life, through success and ostentation" (Hu, 1944: 45). Lian, on the other hand, refers to "the respect of the group for a man with a good moral reputation" (Hu, 1944: 45). The loss of mianzi occurs when a person’s social performance fails to reach the expectations of others. The loss of lian, on the other hand, occurs when a person transgresses the boundary of moral codes of the society. In other words, loss of mianzi goes with failure and can be regained through compensatory actions, while loss of lian goes with transgression and may not be regained (Huang, 1987). Mao (1994) criticized the associations of mianzi with negative face and of lian with positive face in that Chinese mianzi is obtained through one's dependence on society's recognition. (Mao, 1994), rather than through one's claim in freedom from imposition (Brown & Levinson, 1987). Moreover, Chinese lian seems to be similar to positive face in that they both recognize a person's desire to be liked. However, this resemblance is limited because the moral tone of lian is not reflected in the notion of positive face (Mao, 1994).

Overall, then, cross-cultural data on speech acts have shown that speech act

performances are influenced by interlocutors’ face concern. Brown & Levinson (1987)

point out that some speech acts such as request, suggestion and advice threaten the

addressee’s negative face wants, while others such as complaint, disapproval, and

disagreement threaten the addressee’s positive face wants. Politeness is therefore

reflected in indirect speech acts to some extent to reduce face threats. However, as

mentioned above, although the notion of face is universal, the face value appears to be different in western and non-western cultures.

2.1.3 The Act of Complaining

Having reviewed the speech act theory and politeness model, the focus of this section shifts to the speech act of complaining examined in the present study. In this section, the concept of complaining and studies relating to this speech act will be discussed below.

2.1.3.1 The Concept of Complaining

There are two distinct lines of research on the act of complaining. The first line of research investigates the expression of dissatisfaction to an interlocutor who is not present. Boxer (1991, 1993) termed this type of complaint as indirect complaint (IC), which is taken from D’ Amico-Reisner’s (1985) work with instances of direct complaint or disapproval. (D). The second line of research investigates how the speaker makes complaints to the hearer whom is believed to be responsible for the offense. Since these two types of complaints differ greatly in intent and outcome (Boxer, 1991), they have generally been examined separately. The present study focuses on the complaint addressed to the hearer, as discussed in the following.

Searle (1975) set out a series of felicity conditions for a given speech act to be performed: propositional act, preparatory condition, sincerity condition and essential condition. According to Searle, the conditions prescribed for the act of complaining to take place include:

Propositional act: Speaker (S) expresses annoyance of disapproval for a past act (A) done by hearer (H).

Preparatory condition: S believes that A makes him/her inconvenient.

Sincerity condition: S is unhappy with A.

Essential condition:

1. Counts as an expression of annoyance or disapproval for A.

2. Counts as a request for repair of A or a request to cease for doing it.

House & Kasper (1981a), in their discussion of politeness markers in English and German, state that in the act of complaining, the addressee Y must have done an action P which the speaker regards as bad for him/her. They argue that expressives which have the features of "post-event/anti-X" can be referred to as complaints. In English, these verbs include “criticize”, “accuse” and “reproach”.

Edmonson (1981: 145) claims that "in making a complaint, a speaker potentially disputes, challenges, or bluntly denies the social competence of the complainee". Under this circumstance, a complaint flouts the hearer's supportive maxim. However, since the hearer has already flouted this maxim in a socially offensive event, the ground for the complaint is justified. Edmonson (1981: 278) further points out that from a conversational perspective, the ground which constitutes a complaint can form the basis for an apology for the interlocutor:

(i) APOLOGIZE: S did P, P bad for H (ii) COMPLAIN: H did P, P bad for S

Olshtain & Weinbach (1987, 1993) conducted two studies on the speech act of complaining. Their first study (Olshtain & Weinbach, 1987) investigated the expression of complaining among native speakers of Hebrew and learners of Hebrew.

Their second study (Olshtain & Weinbach, 1993) was designed to explore the

variations among two English-speaking groups (British and American) and one

Hebrew-speaking group in terms of complaint strategies. Olshtain & Weinbach set out

the preconditions in which a complaint may take place. First, the speaker expects a

favorable event to occur, but the hearer performs a "socially unacceptable act" (SUA),

whose consequences are perceived by the speaker as offensive for him/herself, or

even for the general public. Second, the speaker regards hearer as responsible for the

SUA. Finally, the speaker decides to express his/her displeasure or frustration verbally to convey the illocutionary force of censure.

Trosborg (1995: 311-2) argues that complaining is “an illocutionary act in which the speaker (the complainer) expresses his/her disapproval, negative feelings, etc. towards the state of affairs described in the proposition (the complainable) and for which he/she holds the hearer (the complainee) responsible, either directly or indirectly”. Laforest (2002: 1597), in the examination of complaining in family life, disagrees with Brown & Levinson's (1987) distinction between acts of disapproval, complaint, criticism and accusation and claims that the act of complaining includes acts of disapproval, accusation, warning and threatening.

In short, a complaint is an illocutionary act in which the speaker expresses negative feelings towards the hearer. The speaker does so because he/she thinks that the hearer should be responsible for a past socially unacceptable event. In the next section, the characteristics of complaints will be discussed.

2.1.3.2 Characteristics of Complaining

In general, the act of complaining has three characteristics. First, it is a face-threatening act. Second, it can be addressed at different directness levels. Finally, it has no obligatory pair part. The three characteristics will be discussed in the following.

As mentioned in section 2.1.2, Brown & Levinson's (1987) politeness model argues that all competent members of a community have two kinds of basic wants:

negative face want and positive face want. Negative face want involves the want of every competent adult member of not being imposed upon, while positive face want involves the want of every adult of being liked by others (Brown & Levinson, 1987).

The act of complaining is a face-threatening act because the speaker holds a negative

attitude towards the offense which is thought to be the hearer’s responsibility. It

threatens the hearer’s positive face because the speaker passes the moral judgment and does not care about the hearer’s desires of being admired or appreciated. In addition, Brown & Levinson note that a complaint may intrinsically threaten the hearer's negative face, which mostly occurs when a complaint is accompanied by a request for compensation.

Since complaining is a face-threatening act, the complainer is facing the danger of destroying the relationship with the complainee. A rational complainer must evaluate the sociological variables in the first place to make the best choice for complaining. According to Brown & Levinson (1987), all competent interlocutors are rational beings, who compute the weightiness of an FTA based on three criteria: social distance (D), relative power (P) and ranking of imposition (R). Social distance refers to the degree of intimacy between interlocutors, which can be identified on four dimensions: strangers, acquaintances, friends and relatives (Olshatin & Weinbach, 1987). Relative power relates to the social status between interlocutors, which can be identified on three dimensions: the speaker is higher in status than the hearer, the hearer is higher in status than the speaker, and both the speaker and hearer are of equal status. Finally, ranking of imposition refers to the degree to which an FTA is considered to interfere with an agent's wants of self-determination or of approval, which can be measured on a continuum ranging from the most explicit to the least explicit mutual commitment between interlocutors (Olshtain & Weinbach, 1987).

The evaluation of sociological variables is helpful to the complainer because

he/she is faced with a variety of options. The first decision the complainer has to

make is whether to issue a complaint or not. He or she may benefit from being

considered as tactful or polite for the choice of opting out. However, the complainer

may also run the risk of not being able to air the frustration or disappointment. Second,

if the complainer decides to issue the complaint, there are several means to achieve

this goal. The decision he/she has to make in the first place is to perform the act either on record or off record. For the on-record path, there are three choices: complaining without any redress, complaining with positive redress and complaining with negative redress. In bald-on-record complaining, the complainer expresses the censure unmitigatedly and explicitly mentions the socially unacceptable act (SUA) and the hearer as the violator. Complaining oriented to the hearer’s positive face reveals the speaker's concern about the hearer, and it usually occurs in in-group contexts. In addition, complaining oriented to the hearer’s negative face is generally recognizable as a conventional request for repair. On the other hand, off-record decisions refer to giving some hints which are related to the offense without explicitly mentioning either the offense or the hearer (Olshtain & Weinbach, 1993). On a directness scale, bald-on-record option is the most direct, and off-record option is the most indirect.

On-record options with redressive actions are in between.

Gender may influence the evaluation of sociological variables. For example, Lin (2005) collected 720 expressions of complaint in Mandarin Chinese. She found that females are more likely to impose obligation on the hearer. They feel more easily disturbed and tend to make complaints when facing a complaint situation. As for the strategies used, males made stronger complaints such as threat more frequently than females. Such gender difference is in agreement with the general belief that females are more polite conversationalists than males.

The last characteristic of a complaint is that it has no typical corresponding second part of an adjacency pair (Edmonson, 1981; Laforest, 2002). This is caused by the fact that the perlocutionary act of a complaint is negotiable (Edmonson, 1981).

The act of complaining is by nature not explicit because many assertions intending to

carry the illocutionary function of complaining may be regarded as simple comments,

while many observations which do not intend to carry such illocutionary function may

be taken as complaints (Laforest, 2002).

The inherent variations of the second pair part of a complaint have led a few researchers to investigate the responses to a complaint, but the research to date has been scarce (Laforest, 2002). Trosborg (1987) argues that in the process of issuing a response, the complainee is encountering a paradox of his/her own social competence and the denial of the social standing of the complainer. Frescura (1993) identified seven semantic formulas of complaint responses, which were classified into two major categories: hearer-supportive and self-supportive. Hearer-supportive formulas include apology, responsibility, repair and forbearance, while self-supportive formulas include denial, explanation and appeal. Laforest (2002) has shown that complaint responses can be classified into four major categories, including acceptance of the complaint, partial acceptance of the complaint, rejection of the complaint and disregarding of the complaint. She also notes that the combination of various complaint response strategies is possible within one single conversational turn, and that a complaint can be followed by any kind of response. From the above studies, it is clear that complaint has no typical corresponding second part because it may be followed by any complaint response strategies, which are generally classified in accordance with the degree of responsibilities the complainee takes.

In summary, section 2.1 provides theoretical basis on which interlanguage

pragmatics is built. Three issues are discussed: the speech act theory, the politeness

model and the act of complaining, which are interwoven in an intricate manner. Most

of the complaint studies to date address interlingual comparisons: English and

German (House & Kasper, 1981a), English and Danish (Trosborg, 1987, 1995),

English and Hebrew (Olshtain & Weinbach, 1987, 1993) and English and Italian

(Frescura, 1993), but it seems that the comparison between English and Chinese has

not yet been fully discussed. Therefore, the present study chooses to investigate

American and Chinese complaints to widen the scope of cross-cultural comparisons.

2.2 Issues in Interlanguage Pragmatics

The notion of pragmatic competence originated from the pioneering work on communicative competence by Hymes (1972), which describes an individual's ability to convey and interpret messages appropriately in communicative contexts within a specific speech community. Following Hymes’ notion of communicative competence, Canale & Swain (1980) and Canale (1983) and captured the concept of pragmatic competence. There are four components in their model—grammatical competence, discourse competence, sociolinguistic competence and strategic competence. The competence which has a close relationship with what is later identified as pragmatic competence is sociolinguistic competence, which refers to an understanding of the social context where language is appropriately used (Savignon, 1983). The first researcher who formally used the term pragmatic competence in the language competence model was Bachman (1990), who distinguished between organizational competence, which deals with forms of language, and pragmatic competence, which deals with function of language.

The increasing attention to pragmatic competence in SLA research has consequently resulted in ILP development, which is defined as "the study of nonnative speakers' use and acquisition of linguistic action patterns in a second language" (Kasper & Blum-Kulka, 1993: 3). Being one of the divisions in SLA research, ILP contrasts itself with other interlanguage studies in morphology, phonology, syntax and semantics because it deals with language learners’

development in pragmatic areas such as speech acts, implicatures and conversation

analysis (Kasper & Blum-Kulka, 1993). However, unlike other interlanguage studies

which have a longer history, the research on interlanguage pragmatics, which has

started since the early 1980s, has received relatively less attention. The two issues

discussed in this section include studies on speech act interlanguage and pragmatic transfer. These issues are chosen because they are frequently discussed in ILP research. The details of each issue will be addressed respectively as follows.

2.2.1 Research on Speech Act Interlanguage

In interlanguge pragmatics, the most important and fruitful exploration has been the study of speech acts. The ILP research has concentrated on a narrow scope of well defined speech acts, e.g. apology, request, compliment, thanking, refusal, and so forth (Flowerdew, 1990). Moreover, most of the ILP studies have focused on non-native speakers' pragmatic competence in comprehension and production, rather than development (Kasper & Blum-Kulka, 1993; Kasper & Schmidt, 1996), which are different from other interlanguage studies that analyze language learners' acquisitional patterns over time. Kasper & Schmidt (1996) explained that it is because ILP comes from empirical cross-cultural pragmatics, rather than second language acquisition research. Therefore, the major issues investigated in ILP research are basically the same as those studied in cross-cultural pragmatics.

It has been demonstrated by many ILP studies that nonnative speakers generally produce speech act behaviors different from those of native speakers in terms of choice of speech act, semantic content, semantic formulas, sociolinguistic forms and utterance length (Bardovi-Harlig, 2001; Cohen, 1996; Olshtain & Weinbach, 1993). In the following, each of them will be discussed.

2.2.1.1 Choice of Speech Act

The first difference between native and nonnative speakers lies in the fact that they may implement different speech acts in the same situation (Cohen & Olshtain, 1993). Coulmas (1981) points out that apologies seem to be more appropriate to Japanese in situations where gratitudes are expected in western languages.

Bardovi-Harlig & Hartford (1993) investigated native and non-native performances in

authentic academic sessions and found that nonnative students used more rejections than suggestions in the earlier sessions of talks. American students, by contrast, used more suggestions than rejections in the talks with their advisors. Murphy & Neu (1995) investigated how Americans and Korean learners of English reacted to an unfairly graded paper. The results show that Americans tended to make more complaints. The linguistic devices used by Americans in issuing a complaint included pronoun “we”, questions to ask for advice, depersonalization of the problem, mitigators and acceptance of partial responsibility. However, Korean learners tended to place the blame on the professor and make more criticisms, which included the use of second person plus modal “should”, personalization of the problem, and refusal to accept responsibility for the problem.

2.2.1.2 Semantic Content

Semantic content refers to the amount of information given by the speaker. In general, the information provided by nonnative speaker appears to be less specific than that by native speakers. Beebe et al. (1990) investigated the performance of refusals by Japanese and American subjects in the request, invitation, offer and suggestion situations. In their study, although most of the refusals in all the situations contained excuses, Americans seemed to produce specific excuses (Beebe, et al, 1990), while Japanese seemed to produce vague excuses by the American norms.

Takahashi & Beebe (1993), in the investigation of Japanese and American corrections,

also revealed that the Japanese learners' correction displayed formulaic patterns but

provided little semantic content. The Americans, on the other hand, displayed more

specific, personalized information. In a longitudinal study of the acquisition of

pragmatic competence, Bardovi-Harlig & Hartford (1993) investigated the differences

between native and nonnative performances of the speech acts of suggestion and

rejection during academic talks with their advisors. The study revealed that the

content of rejections was a major problem for the nonnative graduate students because they frequently provided non-credible explanations (Bardovi-Harlig & Hartford, 1993). Therefore, the success rate of nonnative graduates’ rejections was much lower than that of native speakers of English.

2.2.1.3 Semantic Formulas

Many ILP studies have indicated that nonnative speakers may employ different semantic formulas within the same speech act. Blum-Kulka (1982) notes that learners may not choose the strategies most frequently used by native speakers of the target language. She investigated the speech act of requesting produced by Hebrew learners of English and native speakers of English. The study revealed that learners tended to use less direct strategies under certain circumstances. Olshtain & Cohen (1983), in a review of the apology studies, showed that Hebrew learners of English deviated from native speakers of English because they did not express an apology or offer of repair as much as native speakers in a given situation. Fukushima (1990) examined how Japanese learners of English performed the speech acts of offers and requests and found that Japanese learners used direct strategies more often than Americans.

Olshtain & Weinbach (1993) investigated the complaint strategies employed by a Hebrew speaking group and an English speaking group. Their study showed that native Hebrew speakers preferred more severe strategies in making a complaint, but the American and British learners of Hebrew preferred less severe strategies.

Takahashi & Beebe (1993) also pointed out that Americans tended to employ positive remarks before they made the speech act of correction. However, Japanese learners of English used comparatively less positive remarks than native speakers of English. Lin

& Chen (2006) found that Chinese learners of English used imperatives and want

statement more frequently than Americans.

2.2.1.4 Sociolinguistic Forms

The native and non-native speakers of the target language may also differ in sociolinguistic forms. Olshtain & Cohen (1983) found that Hebrew learners of English did not intensify their regrets in apology situations as much as native speakers did. Eisenstein & Bodman (1986) found extensive syntactic and lexical problems in the ESL learners’ productions of gratitudes. These problems included different use of intensifiers, tense, word order, mangled idioms, prepositions and choice of words.

Trosborg (1987) examined the apology strategies produced by native speakers of English, native speakers of Danish and Danish learners of English. It shows in the study that learners were not as polite because native speakers of English produced more modality markers such as downtoners, understaters, hedges, subjectivizers, intensifiers, commitment upgraders and cajolers. In a series of studies investigating the performances of native and nonnative graduate students in the academic advising sessions, Bardovi-Harlig & Hartford (1990, 1993, 1996) found that although nonnative speakers learned to use the same speech acts as native speakers did, they did not change in the use of forms. American graduate students used mitigators in nearly all suggestion performance, and aggravators were not observed in their productions. By contrast, nonnative graduate students tended to use fewer mitigators and aggravators were identified where native speakers would not use. Yu (1999a), in an investigation of request strategies performed by native speakers of English and Chinese learners of English, also found that the learners used fewer downgraders than Americans.

2.2.1.5 Utterance Length

The issue of utterance length between native and nonnative speakers of a

target language was first raised by Blum-Kulka (1986), who states that nonnative

speakers used too many words to express a certain communicative intent. Such

verbosity may cause pragmatic failure and is termed as waffling, which refers to the too much use of linguistic forms to fill a discourse slot to achieve a particular communicative goal (Edmonson & House, 1991). Some studies are in support of the waffling phenomenon for nonnative speakers. For example, House (1989) reported that German learners of English used more supportive moves than British English speakers in all the five requestive scenarios in the CCSARP (the Cross-cultural Speech Act Realization Project). Bergman & Kasper (1993) reported that Thai learners of English did waffle in apology situations. Olshtain & Weinbach (1993), in the investigation of interlangauge features of complaints, also found that learners at the advanced and intermediate levels tended to use more words than native speakers of English. However, there are studies which showed contrastive findings. For instance, Eisenstein & Bodman (1993) reported that nonnative role-plays in gratitude situations were fifty percent shorter than those of native speakers. The overall performances of nonnative speakers were found not long enough to adequately express gratitudes. The contradictory research findings can be explained by Edmonson & House’s (1991) claim that the waffling phenomenon is only evidenced in written DCTs, rather than in face-to-face interactions elicited by open-ended role plays. The explanation they provided is that in the DCTs, learners have more time to adjust their communicative plan, which reflects their lack of confidence and strategic uncertainty.

In short, it can be seen from the above review that learners often deviate from

native speakers of a target language in terms of choice of speech act, semantic content,

semantic formulas, sociolinguistic forms and utterance length. According to Corder

(1967), learners’ deviations or errors can be attributed to either L1 transfer or

proficiency. Therefore, pragmatic transfer and its relationship with proficiency is

discussed in the next section.

2.2.2 Pragmatic Transfer

One of the most important reasons why native and nonnative speakers differ in the speech act performance is the result of pragmatic transfer. Odlin (1989: 27) defines generic transfer as "the influence resulting from similarities and differences between the target language and another language that has been previously (and perhaps imperfectly) acquired". In the domain of ILP, pragmatic transfer has always been the transfer of speech act knowledge and is defined as "the influence exerted by learners’ pragmatic knowledge of languages and cultures other than L2 on their comprehension, production and learning of L2 pragmatic information” (Kasper, 1992:

206-7). Transfer may have positive and negative outcomes. Positive transfer occurs when there is a form-function correspondence between L1 and L2 (Kasper & Rose, 2001). For example, the past modal verbs “could” and “would” in English have equivalent formal, functional and distributional counterparts in German (i.e.

“koenntest” and “wuerdest”) (Kasper & Rose, 2001).

While positive transfer results in successful communication, negative transfer leads to nonnative use (or avoidance) of speech acts and frequently results in unsuccessful communication (Bardovi-Harlig, 2001). Kasper (1992) admits that positive transfer has been given less attention than negative transfer in ILP research.

Basically, there are two kinds of negative transfer: pragmalinguistic and

sociolinguistic transfers. Thomas (1983) states that pragmalingustic failures are

caused by inappropriate transfer of speech act strategies from L1 to L2. On the other

hand, sociopragmatic failures are caused by the influence of equivalent L1 context on

L2 learners' social perceptions in the understanding and production of L2 linguistic

actions. For example, the degree of politeness to invest in a face-threatening act

depends on how the interlocutor assesses context-external factors, which includes

social power and social distance, as well as context-internal factors, which includes

degree of imposition. A certain situation may be regarded as severe and require more politeness in a particular culture than in others (Kasper, 1992).

The first researcher who systematized positive or negative transfer of semantic formula, strategy and linguistic form was Gabriele Kasper. She criticized the rough estimations of similarities and differences between L1 and L2 in earlier ILP studies. It is claimed in Kasper & Dahl (1991) that the canonical design of ILP research should include IL, L1 and L2 data, and statistically significant differences between the three groups on a particular feature should be calculated to determine pragmatic transfer.

Therefore, according to Kasper (1992), positive transfer occurs when there is lack of statistically significant differences between L1, L2 and IL. By contrast, negative transfer occurs when there are statistically significant differences between IL-L2 and L1-L2, as well as lack of statistically significant differences between IL and L1.

In addition, Kasper (1992) notes two nonstructural factors which may influence pragmatic transfer. The first one concerns learning context. Takahashi & Beebe’s (1987) study on refusals by Japanese EFL and ESL learners revealed that L1 influence was found more frequent in EFL learners’ performances rather than ESL learners’

performances. Yu (1999b) investigated the productions of compliment and

compliment responses by Chinese learners of English in Taiwan (EFL) and in the

States (ESL). It was found that the EFL group displayed speech act behaviors more

like those of native Chinese speakers, while the ESL group displayed speech act

behaviors more like those of native American English speakers. Lee (2000) studied

the complaint-apology speech act sequences between Korean learners of English in

Korea (EFL) and in the States (ESL). The investigation shows that both ESL and EFL

learners displayed transfer effects from their L1. Although Lee did not specifically

address the question whether EFL or ESL learners were affected by L1 to a greater

extent, the analysis shows that the amount of interaction in English has the greatest

effect on learners’ speech act performance in comparison with other factors such as language proficiency and length of stay in the target culture. Matsumura (2001) also points out the importance of exposure to the target language. In the study, Japanese learners’ perceptions of offering advice to people of equal or lower status changed dramatically after they came and studied in Vancouver, while the learners living in the home culture did not show such progress.

The other factor which may influence L1 interference relates to L2 proficiency (Kasper, 1992). The research shows that L2 proficiency exerts an effect on L1 transfer, but whether the correlation is positive or negative has been under debate. Some research indicates that L1 interference is more likely to occur in cases where the learners are at higher proficiency. Takahashi & Beebe (1987) hypothesized that the more advanced learners were assumed to conduct negative L1 transfer. Although their study did not support this assumption, other studies have been found to confirm this hypothesis. For example, Koike (1995) studied first-year, second-year and advanced English-speaking learners of Spanish in comprehending the speech act of suggestion.

The results showed that it was difficult for first- and second-year students to comprehend the utterances even though they were formulated in the same way in both languages. However, advanced learners appeared to be more likely to transfer L1 knowledge if they found the lexical or syntactical items matched those of English.

Other research shows that less proficient learners tended to conduct negative transfer from L1. Olshtain & Cohen (1989) claimed that transfer of learners’ L2 apology strategies and modifying mechanism was a result of insufficient L2 proficiency.

Maeshiba et al. (1996) examined the apology strategies employed by Japanese

learners of English at intermediate and advanced levels. Their study also found that

Japanese learners of lower proficiency were more likely to transfer L1 apology

strategies than learners of higher proficiency.

On the other hand, some research does not identify a clear relationship between L2 proficiency and L1 transfer. Trosborg (1987) compared the apology strategies produced by three groups of Danish learners of English at intermediate, lower advanced and higher advanced levels with those produced by native speakers of English and Danish. This study confirmed that negative pragmatic transfer did exist in all the learners’ role play enactments, but did not identify whether such transfer correlated positively or negatively with L2 proficiency. Takahashi (1996) examined transferability of five Japanese request strategies to English request situations.

Subjects in this study were classified into two groups: Japanese EFL learners at lower and higher proficiency levels. The study revealed that both groups transferred their L1 request strategies and conventions, but there was no definite tendency regarding positive or negative correlation between L2 proficiency and L1 transfer. Among the five request strategies in L1, only one strategy was found to be more transferable for lower proficiency group than for higher proficiency group.

In summary, section 2.2 reviews ILP studies and focuses on the differences

between native and nonnative speakers in speech act productions. The research has

shown that nonnative speakers generally differ from native speakers in terms of

choice of speech act, semantic content, semantic formulas, sociolinguistic forms and

utterance length. Such differences can be attributed to negative pragmatic transfer

from L1 to L2. Research has also shown that EFL learners tend to conduct pragmatic

transfer more frequently than ESL learners. However, it has been controversial as to

the relationship between pragmatic transfer and proficiency. Therefore, one of the

purposes of the present study is to investigate the relationship between pragmatic

transfer and proficiency. Furthermore, since the studies have demonstrated that

learners deviate from the natives in the performances of a given speech acts, it is

reasonable to think if instruction can be facilitative. In the next section, the discussion

will shift to pragmatics in language teaching.

2.3 Pragmatics in Language Teaching

In this section, two major issues will be discussed. The first one deals with the rationale of instruction in pragmatics, and the second one deals with instructional approaches in teaching pragmatics.

2.3.1 Rationale for Instruction in Pragmatics

In SLA research, there has been a consensus that the learner, no matter in a tutored or an untutored learning environment, must pass through a certain stage regardless of their age and L1 background (Larsen-Freeman & Long, 1991). Since the routes of learning for all the language learners are identical, researchers such as Krashen (1982) argues that formal instruction should be abandoned and that learners can benefit most if they are allowed to construct their interlanguage system along a natural route. However, there has been another challenging line of research which claims that although formal instruction cannot change the order of acquisition, it has two major advantages. First, it has a beneficial effect on the process of language development, rate of acquisition and ultimate level of attainment (Pavesi, 1984; Pica, 1983; Sharewood Smith, 1981; Tarone, 1983). Second, through practice, routinization and consciousness-raising in the classroom (Sharewood Smith, 1981), formal instruction can transform from learning to acquisition (Stevick, 1980), from explicit to implicit knowledge (Bialystock, 1982), and from controlled processing to automatic processing (McLaughlin, 1978).

The effect of instruction in SLA research has been demonstrated by a large

body of morphosyntactic studies, but there have been relatively fewer experiments in

ILP research which compare pragmatic competence acquired in a natural setting and

in a classroom setting. The most representative study which has demonstrated that

instruction shortens the time in the acquisition of pragmatics is Bouton (1999). He

found that implicatures can be acquired in a natural setting, but the acquisition requires 33-months to 4-7 years. Therefore, he conducted an experiment to investigate the development of non-native speakers’ skill at interpreting implicatures in American English through explicit pedagogical interventions. He developed handouts with six different types of implicatures, which could be broadly categorized as idiosyncractic and formulaic implicatures. Idiosyncractic implicatures were easier but formulaic implicatures were more difficult for newcomers to the United States. The results showed that after six hours of formal instruction, the learners’ interpretations of formulaic implicatures reached the same level as those who have resided in the United States for seventeen months to seven years. However, idioscyctactic implicatues, which were easier at the beginning, proved to be resistant to explicit instruction.

Bouton’s study is significant because it has evidenced that teaching pragmatic knowledge may not alter the route of acquisition, but may accelerate the process of acquisition. Moreover, the other influence which explicit instruction has on the learners, as Bouton put it, is that learners became interested in using implicatures as indirect communication.

The other question concerning whether teaching pragmatics is necessary is

discussed by Kasper (1997). Since some pragmatic knowledge such as the

Cooperative Principles (Grice, 1975) and Politeness Principles (Brown & Levinson,

1987) are universal, adult learners seem to be able to get a considerable amount of

pragmatic knowledge for free. The universality of pragmatic knowledge may lead

learners to conduct positive transfer successfully if there are parallels between L1 and

L2 in terms of forms and functions (Kasper, 1997; Kasper & Rose, 2001). However,

as Kasper (1997) pointed out, learners do not always make good use of what has

already been available to them. In addition, the appropriate use of a given strategy

may not be identical across languages. As reviewed in section 2.2.2, ILP research has

repeatedly shown cultural variability in that differences exist in the performances of a given speech act between native and non-native speakers of the target language.

These studies have confirmed Gumperz’s argument that second language learners may have a good command of grammar and lexicon of the target language, but may

“contextualize their talk by relying on the rhetorical strategies of their first language”

(Gumperz, 1996: 383). In other words, negative pragmatic transfer from L1 may occur if pragmatics is not one of the components in a language curriculum.

Having recognized the role of instruction in pragmatics, it is necessary to understand the three hypotheses underlying the intervention studies in SLA research.

These three hypotheses include Schmidt's noticing hypothesis, Swain's output hypothesis and Long's interaction hypothesis (Kasper, 2001), each of which will be discussed in detail as follows.

2.3.1.1 Schmidt’s Noticing Hypothesis

Schmidt's noticing hypothesis acknowledges that both conscious and unconscious processes are important to second language learning (Schmidt, 1990).

According to Schmidt (1995), all learning, whether intentional or incidental, requires attention. Input can become intake only when the learner notices it. According to Schmidt (1993), implicit learning, which refers to nonconscious generalizations from examples, seems to be the most effective for the learning of fuzzy patterns, while explicit learning, which refers to conscious problem solving, seems to be the most effective for the learning of rules. To learn pragmatics in a second language, Schmidt (1993) argues that it is important for the learner to consciously pay attention to the input and to analyze relevant features such as linguistic forms, functional meanings and contextual clues. Therefore, explicit teacher-fronted approach is helpful because these features residing in the input are sometimes not salient enough for the learner.

The consciousness-raising approach in classroom teaching can help accelerate the

speed of learning and reorganize what is contained in the input.

2.3.1.2 Swain’s Output Hypothesis

Swain (1985), in contrast with Krashen’s (1982) input hypothesis, highlights the value of output in second language acquisition. She formulated the output hypothesis, which proposes that language acquisition may take place through producing the language and argues that output can promote both fluency and accuracy in second language acquisition (Swain, 1993). Intuitively, fluency can be enhanced through constant production of the target language. Accuracy, on the other hand, can be reflected upon the three functions of output: the noticing/triggering function, the hypothesis-testing function, and the metalinguistic function (Swain, 1993). The noticing/triggering function can force the learner to shift from semantic processing to syntactic processing. In comprehension, the learner may pay full attention to the meanings of the messages and ignore the linguistic details such as definite/indefinite articles and singular/plural forms. However, as he/she produces the target language, the learner would start to notice the gaps between what has already been recognized and what has not been identified in the linguistic knowledge (Swain, 1993).

In addition, noticing can trigger cognitive processes, which generally produce new linguistic knowledge for the learner through explicit or implicit feedback from the interlocutor (Swain, 1996). The hypothesis testing function views output as the learners’ hypothesis testing (Swain, 1996). To put it another way, output provides the opportunity to try out the learner's hypotheses during the negotiation of meaning.

Also, the feedback which the learner receives provides information about the extent of

comprehensibility and well-formedness of the utterances (Swain, 1993).The last

function of output is that it provides conscious and metalinguistic reflection. This

reflective process is generally not demanded in the learner's communicative tasks

where attention is largely paid to meaning rather than form. Swain thus suggests that

one classroom activity which can lead the learners to the reflective process is to ask them to negotiate the form through the use of the target language.

2.3.1.3 Long’s Interaction Hypothesis

Long’s (1996) interaction hypothesis suggests that negotiation for meaning between the native speaker and non-native speaker provokes interactional adjustments, which include linguistic forms, conversational structures and message content. It facilitates learners’ acquisition because the process productively associates input, internal learner capacities, and output (Long, 1996). The exposure to linguistic input is important because it consists of positive and negative evidence with which a competent interlocutor can provide the learner. Positive evidence offers native models of grammaticality and acceptability, while negative evidence provides direct or indirect, explicit or implicit grounds for ungrammaticality produced by the learner.

In an NS-NNS interaction, the native speaker frequently simplifies the language through omission, expansion and rearrangement to make the input more comprehensible. However, Long proposes that comprehensible input alone is not enough, particularly when an adult learner intends to achieve native-like proficiency.

There are two reasons for his argument. First, the learner may not attend to a particular linguistic form even if it exists in the input. Selective attention is therefore necessary to extract items from an array of stimulus (Long, 1996). Second, linguistic support in the form of comprehensible input may sometimes prohibit learning because the learner tends to understand the message content without noticing the syntactic structures or lexical items encoded in the language. Therefore, Long agrees with Swain’s (1985) output hypothesis that a learner should be given the opportunity to produce the language.

In short, this section reviews the rationales for pedagogical interventions in

pragmatics. Although adult learners possess pragmatic universals, they may not be

able to employ them when learning a second/foreign language. The cultural variability even complicates their learning. Schmidt’s noticing hypothesis, Swain’s output hypothesis and Long’s interaction hypothesis provide theoretical grounds for instructions in SLA, which can also be applied to the teaching of pragmatics.

Bouton’s study in implicatures has provided empirical data demonstrating that instruction can speed up the process of acquisition. Since pedagogical intervention has its place in enhancing learners’ pragmatic competence, the next concern is to search for effective instructional approaches, which leads us to the discussion of the following section.

2.3.2 Approaches to Teaching Pragmatics

The research on instructional approaches in pragmatics has started to take shape since the early 1990s (Bou-Franch & Garces-Conejos, 2003). However, Rose’s (2005) comprehensive review shows that the relevant reports are not very many in comparison with those on how instruction influences other aspects of second language acquisition (Kasper, 2001). The teaching targets of these studies include discourse markers (House & Kasper, 1981; Soler, 2005; Yoshimi, 2001), pragmatic routines (Tateyama et al., 1997; Wildner-Bassett, 1984, 1986, 1994), implicatures (Bouton, 1994; Kubota, 1995), pragmatic fluency (House, 1996); sociostylistic variation (Lyster, 1994), and speech acts (Barron, 2005; Bardovi-Harlig & Griffin, 2005;

Liddicoat & Crozet, 2001; Martinez-Flor & Fukuya, 2005; Olshtain & Cohen, 1990;

Rose & Ng, 2001; Koike & Pearson, 2005; Soler, 2005; Takahashi, 2001, 2005).

These studies attempt to address two major research questions. First, do

learners who receive treatments outperform those who do not? Second, if treatments

improve learners’ pragmatic competence, which teaching approach is the most

effective? For the first question, the studies have demonstrated that the learners who

receive treatments perform better than those who do not. For example, Rose & Ng

(2001) experimented with first-year university students in Hong Kong. They organized the subjects into three groups, which included two treatment groups and one control group. The former received instruction in compliment and compliment responses, while the latter did not. The results showed that there was a sharper increase in the use of compliment strategies by the two treatment groups than by the control group. Yoshimi (2001) experimented with third-year learners of Japanese in Hawaii. A story telling task was administered to the control group and treatment group to examine the effects of instruction in interactional discourse markers. The results demonstrated that the treatment group outperformed the control group in overall frequency and accuracy of interactional discourse markers in the posttest.

Marinez-Flor & Fukuya (2005) experimented with Spanish learners of English and found that those who received instruction showed improvements in English suggestions, while those who did not receive any instruction revealed no such progress in their e-mail and phone tests. Koike & Pearson (2005) worked with English learners of third-semester Spanish. To examine the effectiveness of teaching pragmatic information, a unit on Spanish suggestions was constructed. Results indicated that the groups that experienced instruction and feedback performed significantly better than the control group in multiple choice items. Soler (2005) investigated the efficacy of instruction at the pragmatic level and found that those who were exposed to instruction on English requests outperformed those who were not.

To answer the second question, it is important to know that there are basically

two kinds of teaching approaches in ILP research: explicit and implicit. The major

difference is that explicit teaching provides metapragmatic information, while implicit

teaching does not. In Rose & Ng’s (2001) study, the subjects were divided into two

groups: the deductive group and the inductive groups. Both groups received

instruction in compliment and compliment responses, but in a different manner. The deductive group was taught explicitly the metapragmatic information prior to practice activities, while the inductive group was expected to generalize relevant pragmatic information through practice activities. The results showed that the deductive group outperformed the inductive group in terms of compliment response strategies.

Takahashi (2001) taught Japanese college students request strategies by classifying them into groups of explicit teaching, form-comparison, form-search, and meaning focus, each of which is distinguished by different degrees of input enhancement.

The results indicated that the explicit teaching group, who received the most input enhancement, performed the best compared with the other three groups. Tateyama (2001) used both explicit and implicit instruction to teach two groups of beginning learners of Japanese about attention getters, gratitudes, and apologies. The results showed that there were no significant differences in the role plays and multiple choice tasks between the two groups. Nevertheless, the subjects receiving explicit instruction were more successful in correctly choosing the items which required higher degree of formality. Koike & Pearson (2005) found that the subjects who experienced explicit instruction and explicit feedback performed significantly better than those who experienced implicit instruction and implicit feedback. Soler (2005) also found that after instruction, the explicit group showed an advantage over implicit group in the ability to use request strategies in English.

Since a general consensus has reached regarding the importance of explicit instruction in teaching pragmatics, some studies used only this type of instruction to compare learners’ performances of the target feature before and after treatment.

Liddicoat & Crozet (2001) provided explicit instruction on French interactional norms

for a group of university students of French studying in Australia. Role-plays

conducted immediately after the instruction showed that the French interactional

norms for the question “Did you have a good weekend” were incorporated into learners’ talk in which they were engaged. In Yoshimi’s (2001) study, pre- and posttests were administered to the experimental group who received explicit instruction in the use of interactional discourse markers in Japanese. The analysis indicated that the experimental group revealed dramatic gains in their use of interactional discourse markers in the post extended tellings.

A closer look at the intervention studies shows that the major steps of explicit teaching of pragmatic rules follow the traditional “Three P’s methodology”

(Presentation-Practice-Production) (McCarthy, 1998). At the presentation stage, the native speaker model (the NS model) is important because it offers examples of language in use (Olshtain & Cohen, 1990). The model can be short extracts derived from authentic materials such as films (Rose & Ng, 2001), telephone conversations (Olshtain & Cohen, 1991), and television programs (Tateyama, 2001). Or it can simply be interactions between two native speakers of the target language (usually the instructor and another native speaker) (Yoshimi, 2001). In addition to the NS model, the teacher would provide learners with explanatory handouts which identify metapragmatic details necessary for the acquisition of the target form (Bouton, 1994;

Rose & Ng, 2001; Takahashi, 2001; Tateyama, 2001; Yoshimi, 2001).

At the practice stage, learners are given tasks which can reinforce the pragmatic knowledge they have acquired. At this stage, both pragmalinguisitc and sociopragmatic information of the given tasks are encouraged to be clarified by the learners. Therefore, learners may have the opportunities to ask the instructor about the vocabulary and structures they intend to employ to complete the tasks (Yoshimi, 2001).

Finally, at the production stage, some researchers used written DCTs to elicit

learners’ performance (Rose & Ng, 2001; Takahashi, 2001; Tateyama, 2001), while