The chapter discusses the results of the present study in relation to the three sets of research questions.

CHAPTER FIVE DISCUSSION

The chapter discusses the results of the present study in relation to the three sets of research questions.

Causative Verb Constructions in Senior High School Textbooks

The content analysis has revealed several inadequacies in the textbook series: (1) simultaneous presentation of different causative verb constructions, (2) isolated, uncontextualized example sentences, (3) overemphasis on controlled practice of form, and (4) mismatches between grammar activities and text input. The four inadequacies will be addressed in detail in the following sections.

Simultaneous Presentation of Causative Verb Constructions

Textbook A and B show the tendency to present syntactically similar or identical causative constructions simultaneously. Such presentation, however, may mislead senior high school students to form misconceptions about causative verb constructions in English. For instance, the permissive causative let is semantically different from the true causatives make and have. A juxtaposition of the two distinct types of causative verbs in Textbook B (i.e., have / have / let + O + V) seems to suggest that the let-causative are semantically similar to the make- and have-causatives. Without given semantic explanations, senior high students are likely to misuse the let-causative in contexts which require true causatives. Similarly, the simultaneous presentation of the make- and have-causative in Textbook A and B may create a false impression that the two causatives are synonymous.

Like their active counterparts, the embedded-passive causative constructions

are often presented simultaneously. As noted in Chapter Two, the e-passive have- and get-causative structures can be used in overlapping contexts (e.g., we had/get the whole house renovated). Co-occurrence of the two structures in the grammar activities

1

For convenience, embedded-passive will be referred to as e-passive.

seems to be more reasonable. However, the e-passive make-causative structure differs from the above two structures in meaning and use. Simultaneous presentation of the three structures may result in learners’ unawareness of the limited range of past participles that the make-causative can take.

To sum up, misleading presentation of a structure in the textbooks may result in what James (1998) termed “materials-induced errors”, a subtype of “induced-errors”

(p. 178). Since every causative structure exhibits its own meaning and use, it is confusing to present the sentence patterns of different causative structures together with no semantic explanations provided. It is therefore suggested that presentation of causative structures in the textbook books should be designed on the basis of not only syntactic but also semantic considerations.

Isolated, Uncontextualized Example Sentences

All the textbook series incorporate only isolated, uncontextualized example sentences in their grammar activities. However, such sentences are inadequate for causative verb instruction. For one thing, grammar should be viewed as a set of

“discourse-affected choices” (Hughes & McCarthy, 1998, p. 280). The use of an appropriate causative verb structure is dependent on contextual clues such as how the causer achieves the intended result and what the relationship between the causer and the causee is. Uncontextualized example sentences fail to reflect natural language use of causative structures. Just one example should suffice to illustrate.

(1) Mrs. Smith makes Linda separate the garbage.

(From Volume 3, Lesson 6 of Textbook A)

As shown in (1), this discrete sentence does not contain any information that

requires the use of the make-causative. Without sufficient contextual clues, more than

one causative can fit into this sentence despite the fact that different causatives result

in different interpretations of the sentence. For example, the have-causative can replace the make-causative in this sentence to encode a situation where Mrs. Smith assigns Linda the task of separating the garbage without expending much effort. The existence of more than one possible causative in (1) suggests that isolated example sentences do not supply adequate context for the use of the target causative structure.

As Celce-Murcia (1991) aptly pointed out, “when learned as a decontextulized sentence-level system, grammar is not very useful to learners as they listen, read, speak, and write in their second or foreign language” (p. 466). Therefore, more contextualized examples should be incorporated into textbook grammar activities to raise learners’ awareness of when and how to use the target causative structures.

Overemphasis on Controlled Practice of Form

Textbook A and B place a great emphasis on controlled practice of form. Even though Textbook B provides some context-embedded grammar exercises, the context is almost superfluous to the practice of the target form(s) (cf. Fortune, 1998)

. That is, such exercises can be accomplished without much reference to semantic and contextual considerations.

The problem with the reliance on controlled practice of form can be discussed from two aspects. From the perspective of course design, causative verb instruction should preferably come with an increased focus on meanings and uses, as opposed to a constant manipulation of forms

. Since several causative constructions (e.g., make / have / let + O + V) have been introduced in junior high school textbooks, instruction of such structures in senior high textbooks should therefore put more emphasis on their meaning contrasts. For example, in Textbook B, the make- and have-causative

2

This inadequacy is also noted in Fortune’s (1998) survey review of grammar practice books.

3

Skierso (1991), in his guide to textbook evaluation, suggests that exercises and activities should be

graded “to provide a progression from manipulation to communication” (p. 437). Senior High School

English Curriculum Standards promulgated in 1995 also indicates that when a grammatical structure is

recycled, it should be presented and practiced with a different focus and difficulty level.

constructions are first presented in volume 2 and then recycled in volume 6. Instead of the manipulation of forms, the grammar exercises in the later volume should provide students with opportunities to choose between the two target causatives on the basis of contextual clues.

The inadequacy of overemphasis on form can also be explained from the perspective of grammar teaching. As has been reiterated in Larsen-Freeman (1991, 2001, 2002), grammar consists of three interrelated dimensions—form, meaning and use. On the basis of this view on grammar, grammar instruction is aimed at the facilitation of learners’ use of linguistic forms in an accurate, meaningful and appropriate manner. We should strike a balance between the pedagogical attention given to the form, meaning and use of causative verb constructions.

Overall, the grammar activities in the senior high textbooks under examination are deficient in the provision of opportunities for learners to make a choice about which of the target causatives best suits the context. This deficiency may result in senior high students’ low awareness of the appropriate uses of the target causatives.

Fortunately, newer editions of Textbook A and Textbook C incorporate some exercises which require students to choose between the make- and let-causatives on the basis of contextual clues, as illustrated in (2) and (3)

.

(2) Fill in each blank with “let” or “make” to complete the dialogue. Change the form of the verb if necessary.

Mom: Kacey, how was school today?

Kacey: Not bad. The coach made me dribble the ball for a whole hour.

Mom: You must be tired now.

Kacey: Not at all. Let me tell you something great, Mom. The coach will let me play in the game against Dallas High School next week….

(From Volume 1, Lesson 4 of the newer edition of Textbook A)

4

The answers to the questions in (2) and (3) are provided by the teacher’s manuals.

(3) Based on the dialogue, write a sentence by using either “let or “make”

Tommy: Mom, can I go outside to play?

Mother: Sure, Tommy. Go ahead.

Æ Tommy’s mother lets Tommy go outside to play.

(From Volume 1, Lesson 5 of the newer edition of Textbook C)

However, a close examination into the exercise in (2) reveals an inappropriate use of the make-causative, as shown in (4).

(4) Kacey: Not bad. The coach made me dribble the ball for a whole hour.

The make-causative in (4) implies Kacey’s unwillingness to dribble the ball for a whole hour and is thus incompatible with his previous utterance “Not bad.” Therefore, although some improvements have been made in exercise design, continuing efforts are needed to ensure each grammar exercise is properly contextualized.

Mismatches between Grammar Activities and Text Input

The linkage between units and exercises is one of the factors that need to be taken into consideration when we evaluate ELT textbooks and materials (Sheldon, 1988). However, the textbook analysis has identified several mismatches between grammar activities and text input with regard to causative verb constructions.

First of all, all the textbook series fail to provide sufficient grammar activities for senior high students to practice using the make + O + Adj. construction, despite its high frequency in the textbook texts. The make + O + V construction, in comparison, receive more attention from the three series. This inadequacy may explain Liu &

Shaw’s (2001) observation of Taiwanese university students’ redundant use of verbal complements with the make-causative in contexts where native speakers prefer the use of adjectival ones.

Second, some causative verb construction are selected as the target grammatical

points but are absent from the textbook texts. As noted by Fotos (2001), “input

provides essential positive evidence, the language data that allows acquisition to occur” (p. 271). Without sufficient exposure to the language data, language learners may not be able to associate the form with its meaning and use.

Finally, some inconsistencies are identified between the meanings and uses of the causative constructions occurring in the texts and those practiced in the grammar exercises. These gaps may reflect the tendency to focus more on form than on meaning and use in grammar instruction. In order to provide language learners with consistent example and practice sentences, it is necessary to design grammar activities with reference to semantic and contextual considerations.

In conclusion, the previous sections have discussed four major inadequacies in the textbook series and stressed the importance of providing sufficient input and communicative grammar activities. As suggested by Doughty & Williams (1998),

“salience in the input” and “communicative function or meaningfulness in the output”

are two of the various explanations for the acquisition of linguistic forms (as cited in Haley & Rentz, 2002, p. 5). Therefore, textbooks should enhance the salience of target grammatical structures in input by the use of techniques such as highlighting the target structures in a given passage or choosing texts where the target structures are frequent (Larsen-Freeman, 2001, p. 257). Moreover, textbooks should provide opportunities for learners to practice producing the target structures to convey meaning appropriately.

Instruction of Causative Verb Constructions in Teacher’s Manuals

One of the dimensions to evaluate a teacher’s manual is to examine if it provides

sufficient linguistic background information for its users (Daoud & Celce-Murcia,

1979; Skierso, 1991). However, the three sets of teacher’s manuals are deficient in the

following aspect. While considerable efforts are devoted to the formal explanations of

causative constructions, surprisingly few efforts are made to provide their functional

explications

. The following subsections aim to discuss several causative structures in the teacher’s manuals.

The Active Make-Causative Construction

The embedded-active make-causative structure encodes two distinctive senses:

the coercive and non-coercive sense (Baron, 1974; De Cock & Granger, 2004). Both senses should be fully explained in the teacher’s manuals and yet the coercive sense seems to have been overlooked. Although Textbook A incorporates many example and practice sentences of the coercive sense, its accompanying teacher’s manuals fail to provide adequate information on this sense. The make-causative is often translated into the Chinese causative shi. This Chinese causative, however, cannot be used to encode situations where both the causer and causee are agentive (Teng, 1989) and thus is unable to capture the coercive sense of the make-causative. Other Chinese expressions such as bi po ‘逼迫’ or qiang po ‘強迫’

would be more appropriate to describe the sense of coercion involved in the causation.

The Active Have-Causative Construction

The e-active have-causative construction encodes effortless causation where “the causer does not have to exert any special pressure to achieve the desired effect”

(Wierzbicka, 1998, p. 121). The causee is willing to perform the task assigned by the causer due to a payment in return or a relation of authority involved (Goldsmith, 1984;

Wierzbicka, 1998; Stefanowitsch, 2001). The typical uses of the have-causative are reflected in almost all the have-causative expressions in the three sets of textbook texts and grammar activities, as exemplified in (5), (6) and (7).

(5) After the play, Mrs. Poe had the students discuss it.

(From Volume 6, Lesson 9 of Textbook A)

5

The terms “formal explanation” and “function explication” were from Paulston and Bruder (1976).

6

The two Chinese expressions imply the sense of force.

(6) He had his secretary call Mr. Williams this morning.

(From Volume 2, Lesson 6 of Textbook B)

(7) The first clocks were so big and heavy that some rich people had servants carry their clocks.

(From Volume 1, Lesson 4 of Textbook C)

All the above examples imply a hierarchical relation between the causer and causee.

The examples in (6) and (7) further suggest regular payments in return.

Although these features are frequently associated with the have-causative expressions in the textbook series, they are not well addressed in their accompanying teacher’s manuals. Only similar but incomplete Chinese translations are provided for the teacher users. To raise teachers’ awareness of the semantic and pragmatic features of the have-causative, the teacher’s manuals should supply more detailed explanations rather than short and simple Chinese translations.

The Active Get-Causative Construction

The e-active get-causative construction encodes the effortful causation where much effort or difficulty is involved in achieving the intended result (Gilquin, 2003;

Martin, 1981; Stefanowitsch, 2001). This effortful sense is often not well addressed in the teacher’s manuals. The get-causative tends to be translated into the Chinese causative shi and presented as a synonym with the have-causative. Such presentation is likely to result in misconception about the get-causative. For one thing, the Chinese causative shi does not imply the effortful sense and thus can not explain the meaning of the get-causative. For another, the effortful sense of the get-causative contradicts with the effortless sense of the have-causative. It is therefore misleading to present the two causatives as if they were synonymous.

While the effortful sense is not represented in the teacher’s manuals, it is implied

in nearly all the get-causative expressions extracted from the textbook texts and the

grammar activities. The expression it took him a long while in (8) explicitly suggests

more time and effort involved in achieving the intended result.

(8) It took him a long while to get his father to understand why he chose English literature as his major.

(From Volume 6, Lesson 9 of Textbook A)

Moreover, the above example implies the use of verbal persuasion in overcoming the resistance on the causee’s part. The get-causative in (8) is therefore similar in meaning to the expressions talk into and persuade. Such sentences in the textbooks give English teachers great opportunities to teach the semantic features of the get-causative.

To help teachers fully explain such sentences, the teacher’s manuals should offer sufficient information on the meaning and use of the target structures.

Active Let-Causative Construction

The let-causative differs from the make-, have- and get-causatives in that it depicts permissive causation rather than true causation (Comrie, 1989; Shibatani, 1973). In order to facilitate appropriate use of the let-causative, the teachers’ manuals should provide information on the semantic distinction between the let-causative and the other three causatives, especially when the let-causative is simultaneously presented with true causatives.

The Embedded-Passive Causative Constructions

Instruction in the e-passive make-, have- and get-causative constructions often places a predominant focus on the conditions under which the three causatives are followed by past participles. Little attention is given to their respective meanings and uses. This often results in learners’ inability to use these constructions appropriately.

To address this inadequacy, the teacher’s manuals should provide functional

explanations of these structures: (1) make + O + p.p. is used to encode successful

transmission of a signal to others, (2) have + O + p.p. is predominantly used to

encode the service frame, and (3) get + O + p.p. is also used to encode the service frame but is preferred over have + O + p.p. when some difficulty is involved.

In conclusion, the majority of senior high teachers in Taiwan are nonnative speakers of English. Although they are familiar with the formal features of causative constructions, they might not be aware of their semantic and pragmatic differences. It is therefore essential to help EFL teachers in Taiwan to fully explain the target causative constructions by providing sufficient functional explications.

Effectiveness of the Two Types of Causative Verb Instruction

This section explores possible explanations for the two groups’ test performance before and after the instructional treatments and discusses the two groups’ attitudes toward the instructional treatments in relation to their test performance. The discussions in this section are expected to show whether the form-based causative instruction is effective in promoting the learning of the form and meaning/use of the target causative instructions and whether the meaning-based one is an effective alternative.

Explanations for the Two Groups’ Test Performance

The following subsections will provide possible explanations for the two groups’

test performance before and after the instructional treatments.

Before the Instructional Treatments

The make-causative. All the participants performed well on the form and the non-coercive sense of the make + O + V structure before the instructional treatments;

however, they showed great difficulties with the coercive sense (see Table 32 & 34).

This discrepancy may be teaching-induced. As noted earlier, little instructional

attention has been given to the coercive sense of the make-causative. This observation

was further confirmed by questionnaire responses. When asked about their previous

understanding of the make-causative construction, only 12% of the participants

provided the Chinese expression qiang po. This shows that most of the participants had not received sufficient instruction in the coercive sense from their previous teachers before the present study.

Although the participants performed well on the test items for the non-coercive sense, it is still worth discussing the incorrect responses. The let-causative was found to account for the majority of the incorrect answers, representing the most popular option the participants mistakenly resorted to (e.g., let me realize) (see Appendix J &

K). This is largely due to their inadequate knowledge of the let-causative in English.

When asked about their previous understanding of the let-causative, about 60% of the respondents provided the Chinese causative rang. However, this causative is broader in meaning and use than the English causative let, encoding both true and permissive causation (Weng, 2005). Negative L1 transfer occurs when a similar but incomplete L1-L2 correspondence can be found (Helms-Park, 2001, p. 80).

While the participants performed well on the form of the make + O + V structure, they performed worse on that of the make + O + p.p. structure. This difference can be explained by several factors. First, not all the participants had received instruction in the make + O + p.p. structure (see Table 5). Second, the make + O + p.p. structure is far less frequent than the make + O + V structure. As pointed out by Baron (1972), the make-causative is infrequently used with past participles (cf. Gilquin & Lecoutre, 2004). Due to its limited occurrence, the participants might not have had sufficient exposure to this structure and thus were not familiar with its form. Another possible factor is concerned with insufficient understanding of the item in (9).

(9) George got trapped in the bushes and tried to make himself noticed (notice).

The participants may have falsely perceived the pronoun himself as the agent of the

action notice and thus mistakenly filled in the blank with the bare infinitive notice.

The first two factors discussed above may also account for the participants’ poor performance on the meanings and uses of the make + O + p.p. structure. Furthermore, their difficulties may result from negative L1 transfer. Almost 35% of all the participants misused the let-causative with the past participles understood and heard (see Appendix J & K). This further suggests that the participants were influenced by the Chinese causative rang but unaware of its meaning and use differences from the English causative let.

The have-causative. A considerable proportion of the participants showed awareness of the use of bare infinitives with the have-causative before the present study (see Table 32). This finding is not surprising since 87% of the control group and 95% of the experimental group participants had previous learning experience with the have + O + V structure (see Table 5). However, both groups showed difficulties with the semantic and pragmatic aspects of this structure (see Table 34 & Table 37). Only around 40% of the participants used the have-causative to describe the effortless causation where “one person has a right to make requests of or assign tasks to another person” (Ackles, 2003, p. 114). This shows their low awareness of the meanings and uses of the have-causative.

The participants’ inadequate knowledge of the have-causative structure may be teaching-induced. The teacher’s manuals under discussion place primary emphasis on the formal aspect of the have-causative structure but pay little attention to its semantic and pragmatic features. Without given proper instruction, the participants were unable to differentiate the have-causative from the make-, get-, and let-causatives and thus misused the other three in the contexts which required the have-causative.

Compared with that of the have + O + V structure, the form of the have + O +

p.p. structure presented more problems for the participants (see Table 32). This may

be explained by the fact that fewer participants had received instruction in the have +

O + p.p. structure before this study (see Table 5). Contrary to our expectation, however, both groups performed better on the functions of the have + O + p.p.

structure than those of the have + O + V structure (see Table 34). This surprising result is mainly due to their better performance on the test item targeted on the common expression had/ got her hair cut short (see Table 38). Since the participants performed poorly on less common expressions (e.g., have her tonsils removed), they were still unfamiliar with the use of the have + O + p.p. structure to encode the service frame. This finding is not unexpected since the functions of the have + O + p.p. structure are often not well addressed in the instructional materials for EFL learners in Taiwan.

The get-causative. Both groups performed worse on the test items designed for the get-causative constructions (see Table 32 & 34). Their poor test performance may be attributable to two major factors. The first factor is concerned with the lack of previous learning experience. Quite a few participants had not received instruction in the get-causative constructions, especially the embedded-passive one (see Table 5).

The second factor pertains to inadequate instruction in the get-causative constructions.

While much attention has been given to the complement types the get-causative takes, little effort has been made to provide learners with information on the effortful sense implied by this causative. This may explain the participants’ difficulty in using the get-causative to encode the effortful causation (see Table 39 & 40).

The last point worth discussing here concerns the form of the get+ O + to V

structure. Although 70% of all the participants had previous learning experience with

this form (see Table 5), more than half of them misused bare infinitives with the

get-causative. This surprising result may be due to the fact that the get-causative is the

only target causative that takes to-infinitives. If this formal feature is introduced in an

arbitrary manner and practiced in a mechanical way, it is harder for learners to

retrieve the rule ‘the get-causative is followed by to-infinitives’ later.

The let-causative. Both groups were able to provide answers to the test items for the let-causative without much difficulty. This may be a consequence of three major factors. First, the let-causative is introduced to EFL learners in Taiwan in junior-high schools. All the senior-high participants had previous learning experience with this causative. Second, the let-causative is common in conversation and fiction (Biber et al, 1999, p. 269). The participants may have often encountered the let-causative in their textbook reading and conversation texts and thus were familiar with its form and meaning. The last factor is concerned with how the let-causative is usually taught.

Instruction in the let-causative often focuses on its form. Its meaning is often addressed by the similar but incomplete Chinese translation rang. While this translation may mislead EFL learners in Taiwan to use the let-causative beyond its legitimate boundary, it does help them to learn the permissive sense of the let-causative. Therefore, when provided with clear contextual clues, the participants did not show much difficulty using the let-causative to express the permissive sense, as exemplified in (10).

(10) Diane thinks television is a waste of time, so she doesn’t (2a) let her children (2b) X watch TV.

However, the participants did perform poorly on two particular let-causative expressions, as shown in (11) and (12).

(11) …But in other cultures people insist that a mother ignore her baby when it is crying. They think if you do not (31a) let a baby (31b) X cry, it will become spoiled…

(12) Barbara: Oh, I also need to cancel classes for the next week. Please ask the

work-study student to inform my students of the canceled classes. I’ll (25a)

let you (25a) X take care of the rest of the details.

The participants’ poor performance on (11) may be due to their failure to fully understand the item and their unfamiliarity with the use of the let-causative with the caused event that is already happening. As a result, 50% of the control group and 39%

of the experimental group resorted to the expression they were more familiar with, i.e.

make a baby cry (see Appendix J & K). Similarly, the participants’ poor performance on (12) may also result from their insufficient understanding of the test item and their inadequate knowledge of the let-causative.

In comparison to the other three causatives, the let-causative seems to present less difficulty for the participants. However, the participants’ misuses of the let-causative and difficulties with certain let-causative expressions indicate that their knowledge of this causative was still inadequate before the present study. This inadequacy may also be a consequence of inadequate teaching.

In sum, the foregoing sections have offered three major explanations for the participants’ poor performance on the pretest: (1) lack of previous learning experience, (2) insufficient input, and particularly (3) inadequate teaching.

After the Instructional Treatments

The make-causative. Distinct differences were identified between the two

groups in their posttest performance on the meanings and uses of the make + O + V

structure. While the coercive sense remained difficult for the control group, it became

a lot easier for the experimental group (see Table 42). This finding confirmed the

hypothesis that the participants’ poor performance on the coercive sense was largely

due to inadequate instruction. After given the meaning-based instruction, the

experimental group was able to make significant improvement in the use of the

make-causative to encode the coercive sense (see Table 31). This can be best

illustrated by their performance on the parallel pre- and posttest items in (13) and

(14).

(13) Johnny hates all kinds of vegetables. Last evening, his father shouted at him and (4) made him eat some spinach. He was really upset. (From the pretest) (14) Peggy hates doing homework very much. Yesterday evening, her parents

yelled at her and (5) made her do her math homework assignments. She felt extremely upset. (From the posttest)

While only 13% of the experimental group provided the intended causative make to the pretest item in (13), up to 92% did so to the posttest item in (14).

However, despite their raised awareness of the coercive sense, the experimental group performed poorly on (15).

(15) Sara: The answer is 5.34.

Mrs. Smith: No. You can try again to get it right, but only if you want to.

Summary: Mrs. Smith didn’t (19) make Sara try again.

This may be due to the linguistic complexity of this item. As pointed out by some experimental group interviewees, they had difficulty understanding the expression but only if you want to and thus failed to use the make-causative. Apart from their poor performance on (15), the experimental group showed a better understanding of the coercive sense after the meaning-based instruction.

On the other hand, the experimental group failed to make significant progress in the use of the make-causative to encode the non-coercive sense and to describe

“successful transmission of a spoken message” (Lee, 1996, p. 407). This group even performed less well than the control group on (16), (17) and (18).

(16) The things my friend said yesterday (6) made me wonder if she is in some kind of trouble.

(17) Warm colors such as orange and yellow can (8) make your bedroom seem much smaller than it actually is.

(18) Newborn babies can quickly learn to (11) make themselves heard through

coos, babbles, laughs and cries.

The unexpected findings are largely due to time constraints. As mentioned in Chapter 3, each instructional treatment only lasted for 70 minutes. It was therefore difficult for the experimental group to absorb all the information on the meanings and uses of the target causatives. Overgeneralization occurred when some participants from this group were aware of the core meanings of certain causatives but unaware of their boundaries. Around 20% of the participants misused the have-causative for the make-causative in (16) and (17). Some of them pointed out in the interview that they perceived the effortless sense and thus chose to use the have-causative. The same factor may also account for the unexpected low accuracy rate of (18).

The last point worth discussing here is concerned with the two groups’ posttest performance on the form of the make + O + p.p. structure. The control group was found to score significantly higher than the experimental group on the posttest items for this form (see Table 28). This finding suggests that the form-based instruction was more effective than the meaning-based one in promoting the learning of the form of the make + O + p.p. structure.

However, the control group did not perform on the test items for this form as well as expected, especially the one in (19). Only 47% of the control group and 34%

of the experimental group provided the intended answer known.

(19) Joe always tries to make himself known (know) to those around him as much as possible.

This can be partially explained by inadequate understanding of this item. The

participants may have misinterpreted the pronoun himself as the agent and thus

misused the bare infinitive know. As observed by the researcher, the participants were

more familiar with the use of past participles with inanimate objects than with

reflexive forms of person pronouns. Since the make + O + p.p. structure often

involves reflexive forms of person pronouns, its form was more difficult for both groups to learn. If the make + O + p.p. structure is selected as a target grammar point in the textbook, more input should be provided for learners to notice its form and meaning.

The have-causative. The two groups differed significantly in their performance on the posttest items for the meanings and uses of the have + O + V structure. While almost 80% of the experimental group participants used the have-causative to encode the effortless causation, only 40% of the control group participants did (see Table 42).

This finding confirmed the inadequacy of the form-based instruction and the effectiveness of the meaning-based one.

Yet the experimental group seemed to show particular difficulties with the items in (20) and (21).

(20) Ms. Allen: Fernando, you didn’t hand in your homework assignment today.

I’d like you to give it to me tomorrow.

Fernando: Sure, Mr. Allen. I’ll do that.

Summary: Ms. Allen (18) had Fernando hand in his assignment.

(21) (From an article about how charities try to get people to donate money)

…The (radio) station also (32) have volunteers answer phone calls from those who would like to make a donation…

Up to 42% of the experimental group participants used the make-causative for the

intended causative have in (20) (see Appendix M). This can be attributed to their

perception of the teacher-students relationship. Several participants indicated that they

regarded the ‘hand in homework assignment’ context as a representation of the

coercive causation and therefore used the make-causative. This phenomenon reflects

the fact that “the choice of a structure is dependent both on the intended meaning and

on how the speaker construes the situation” (Larsen-Freeman, 2002, p.116). Although

(20) does not imply any use of coercive means on the teacher’s part, the participants

still used the make-causative to describe what they perceived as a coercive situation.

Therefore, the experimental group participants were able to use the make-causative to express their intended meaning after the meaning-based instruction.

In addition to (20), the experimental group did not perform well on (21). It is noteworthy that (21) is the last item on the posttest. The participants may not have had enough time to finish reading the context of the test item. Consequently, up to 42% of the participants were not able to provide the causative which best suits the context.

The two groups also showed significant differences in their performance on the test items designed for the have + O + p.p. structure. The control group was found to significantly outperform the experimental group on the form of this structure (see Table 28) and to make significant in this form (see Table 29). These findings suggest the effectiveness of the form-based instruction in helping senior high students learn the form of the have + O + p.p. structure. On the other hand, the experimental group was found to significantly surpass the control group in the use of this structure to encode the service frame (see Table 30). This significant difference shows the effectiveness of the meaning-based instruction in promoting the learning of the functions of this structure.

The get-causative. The experimental group participants showed significantly better performance on the test items for the effortful sense of the get + O + to V structure than its control group counterparts after the meaning-based instruction (see Table 30). Moreover, this group made significant improvement in this sense, with its average accuracy rate increasing from 19% to 68% (see Table 34 & 42). These findings suggest that it is effective to provide senior high students with semantic explanations of the get + O + V structure.

However, the experimental group failed to perform on the posttest item in (22) as

well as expected.

(22) Ms. Chen: David, place your top front teeth against your bottom lip and say

“f-, fruit.”

David: “f- fruit.”

Mrs. O.: Good! That’s it!

Summary: Ms. Chen finally (21a) get David to pronounce the word “fruit.”

Almost one-third of the participants used the have-causative instead. Some of them indicated in the interview that they regarded this context as a situation where the teacher simply assigned her student the task of pronouncing the word “fruit” and thus used the have-causative to describe what they perceived as the effortless causation.

The interview responses suggest that the experimental group would have performed much better on this item if they had been provided with more contextual clues. In spite of their poor performance on (22), the experimental group participants showed enhanced understanding of the meanings and uses of the get + O + to V structure.

On the other hand, the experimental group showed more difficulties with the effortful sense of the get + O + p.p. structure. One explanation for this is that most of the experimental group participants (63%) did not have previous learning experience with this structure (see Table 5). It was therefore more challenging for them to pay attention to both form and meaning of the structure during the time-constrained instructional period. Another explanation is that when the get- and have-causatives take past participles, their differences become more subtle. Therefore, about 28% of the experimental group participants misused the have-causative in the items for the e-passive get-causative expressions. This finding shows that the effortful sense of the get + O + p.p. may be too subtle to introduce to senior high school students.

The last point worth noting here is concerned with the form of the get + O + to V

structure. This form remained difficult for both groups, with only 62% of the control

group and 54% of the experimental group being able to use to-infinitives with the

get-causative. To facilitate the learning of this form, it may be helpful to introduce to

learners the iconic relation between the effortful sense of the get-causative and its selection of to-infinitives as the complement verbs. As suggested by Larsen-Freeman (2000), language teachers should focus on teaching “reasons, not rules” (as cited in Larsen-Freeman, 2001, p. 265).

The let-causative. Both groups performed well on the pretest items designed for the form of the let-causative structure. However, the control group showed particular difficulty with the posttest item in (23).

(23) We let our cat go out (go out) at night if it wants to.

More than one-fourth of the control group participants mistakenly filled in the blank with the past participle gone out. There are two possible explanations for this regressive behavior. For one thing, the control group may not have been aware of the fact that the let-causative does not take past participles. For another, they may have misinterpreted the object our cat as the patient and thus misused the past participle gone out. The experimental group, in comparison, showed higher awareness of the incompatibility of the let-causative with past participles and the agentive role of the object out cat after given the meaning-based instruction. Only one experimental group participant filled in the blank with the past participle gone out. This shows the necessity to provide semantic explanations even when the instructional focus is on form.

The meaning-based instructional treatment also enabled the experimental group to outperform the control group on all the items for the meanings and uses of the let-causative. However, this group showed particular difficulty with (24).

(24) Some believe if you always pick up and try to soothe a crying baby you are

in danger of spoiling them, while others think if you (3) let a baby cry you

are neglecting them.

Their poor performance on (24) shows that the participants were still less familiar with the use of let-causative with the caused even that is already happening. In other words, they were less aware of the broad meaning of the let-causative, i.e. ‘not prevent something from happening’ (Dixon, 1991, Wierzbicka, 2002). To improve learners’ knowledge in this regard, grammar exercises in the textbooks should incorporate more let-causative expressions of this kind.

In conclusion, the foregoing discussions have supported the hypothesis that inadequate teaching was largely responsible for the two groups’ poor performance on the pretest items for the meanings and uses of the target causative structures. Once given the meaning-based instruction, the experimental group was able to make significant improvement in this regard. On the other hand, the control group continued to perform poorly on the test items for meaning and use after the form-based instruction. This contrast confirms the negative consequences of the inadequacies in the textbooks and teacher’s manuals, i.e. simultaneous presentation of different causative verb constructions with no explicit explanations provided to address their differences in meaning and use, overemphasis on controlled practice of form and the use of incomplete Chinese translations to explain the semantics of the target causatives. Moreover, this contrast provides empirical evidence for the effectiveness of the meaning-based instruction in helping senior high school students to learn the semantic and pragmatic features of the target causatives.

The foregoing discussions have also offered implications for the teaching and

learning of the formal features of causative verb constructions. The control group was

found to significantly outperform the experimental group on the forms of the make +

O + p.p. and have + O + p.p. structures after the form-based instruction. Notice that

the two structures are usually introduced to Taiwanese EFL learners later than the

embedded-active causative structures make + O + V, have + O + V, and let + O + V.

This finding thus suggests that mechanical drills are more effective in familiarizing learners with newly-introduced forms.

This suggestion is further supported by the finding that the control group made significant gains on the forms presented and practiced primarily in senior high schools, i.e. have + O + p.p., get + O + p.p. and get + O + to V (see Table 29). In other words, students do not benefit much from mechanical practice of the forms they are already familiar with. Since the make/ have/ let + O + V structures are introduced to Taiwanese EFL learners early in junior high schools, instruction of these structures in senior high schools should focus more on their meanings and uses. As suggested by Yalden’s proportional approach (1987), an increasing emphasis should be placed on communicative functions and discourse skills as students progress toward the advanced level. More instructional efforts should be made to equip learners with the ability to use the target causative structures not only accurately but also meaningfully and appropriately (cf. Larsen-Freeman, 2001, p. 256).

The Two Groups’ Attitudes toward the Instructional Treatments

This section aims to compare the two groups’ attitudes toward the effectiveness of the two types of causative verb instruction with their test performance to determine whether there is any connection between the two.

Overall, the two groups’ questionnaire responses are highly associated with their test performance. More than one-fourth of the control group participants provided neutral responses to the helpfulness of the form-based instruction. One participant even chose the option “Not helpful.” The questionnaire responses are closely related to the test result that the control group did not perform on the posttest items for form as well as expected. The questionnaire responses also suggest that mechanical drills of form seemed tedious to the control group and thus led to decreased positive responses.

It is noteworthy that although the experimental group participants were not given

much instruction in the target forms, they provided more positive responses to this question than their control group counterparts. These findings show the necessity to include more meaningful and communicative grammar activities into textbooks.

The experimental group, in comparison, provided a high percentage of positive responses (92%) when asked about the helpfulness of the meaning-based instruction.

This survey result corresponds well to the experimental groups’ significant gains on the posttest for meaning and use (see Table 31).

This group was also asked to choose causative verbs whose meanings became clearer to them after the instruction. Of the four target causatives, the most frequently chosen one was the get-causative (82%), followed by the have-causative (61%), the make-causative (53%) and the let-causative (37%). The survey result reflects the extent to which the experimental group participants made improvement in the target causatives in terms of meaning and use. They made the greatest progress in the effortful sense of the get + O + to V structure, the effortless sense of the have + O + V structure and the coercive sense of the make + O + V structure (see Table 31). On the other hand, they made the least progress in the permissive sense of the let + O+ V structure (see Table 31). This may explain why the let-causative was the least frequently chosen one by the experimental group. However, it is noteworthy that 26%

of the participants chose all the four causatives. This indicates that more than one fourth of the experimental group participants regarded the meaning-based instruction as an effective method for improving their understanding of the meanings and uses of all the target causatives.

In conclusion, both questionnaire and test results suggest the effectiveness of the

meaning-based instruction in promoting the learning of the meaning and use of the

target causative constructions.