〈、、4 T三3 C二 CD
「u) 8 Q)= c\d
m~ "
τ， co
E 由 6 0'一 日立的
令J
EEEQ) τ三守“
豆 U主 5
S 2u
C三Q) 這二出 i三3
f/Y /y
，.-<三/
. \ \ \ /
﹒~-~-:三于\/
NES LEPL
Groups
AEL
Figure 3
Differences Between Boundaries1 and 2 for Contexts 1 and 2
DISCUSSION AND PEDAGOGICAL IMPLICATIONS
(a)
InVestIgate to
耳vere
The pu中oses of the present study
whether Taiwanese English learners reliably use prosodic cues to disambiguate syntactically ambiguous sentences when reading them aloud, and (b) whether English proficiency level leads to different prosodic behavior patterns. The main results of this reading aloud task at sentence level can be summarized as follows: (1) both native speakers and advanced English learners reliably produced prosodic boundaries to convey the intended meaning of an utterance when informed the ambiguous points in advance; furthermore,the prosodic boundaries made by advanced learners were more distinguishable than those by native speakers, (2) limited English learners did not
73
其語教學 'E物品'sfi '1主acfiil穆巴tLearning
34.3 (Fall201 0)
create prosodic phrasing effects in such a reading aloud condition, and (3) English proficiency level did lead to different prosodic behavior patterns.
Reliable Use of Prosodic Disambiguation by Informed Native Speakers
Prior studies have consistently indicated that informed native speakers are able to properly express the intended meaning of an ambiguous utterance by altering intonational phrasing. Based on the prior research, Snedeker and Trueswell (2003) claimed that speakers of a language share certain implicit knowledge about the relationship between prosody and syntax? They argued, however, that speakers do not necessarily use such implicit knowledge to produce prosodic boundaries to disambiguate syntactically ambiguous sentences. On the contrary,the production of prosodic disambiguation relies mainly on the fact that speakers know the ambiguity situation and think it is needed. For example, if speakers are aware of the alternative meanings of an ambiguous utterance, they might think of prosodic means as needed. Or if no other means other than prosodic means are available for syntactic disambiguation, speakers might consciously produce prosodic boundaries. In this sense, use of prosodic disambiguation in spoken language may occur under certain specific cIrcumstances.
To replicate the findings that informed native speakers reliably use prosodic cues to disambiguate ambiguous utterances,we provided
3 According to Ellis(2008),implicit knowledge about a language takes place as a result of children acquiring the language by engaging with their caretakers in natural meaningful communication
Yang: A Preliminary Study of Prosodic Disambiguation in L2 Production our native speakers with the possible meanings of an ambiguous utterance and enough time to prepare for the reading aloud task before starting formal recordings. Like Lehiste (1973) and Allbritton et al.
(1996), we found that our native speakers showed prosodic phrasing effects on syntactic disambiguation. Speci日cally， since the subjects were informed of the ambiguities and had enough time to monitor the reading aloud task at sentence level, they realized that they must effectively lengthen a critical boundary so as to reflect the syntactic structure of an ambiguous utterance. These results provide evidence that our native speakers not only had some implicit knowledge about the relationship between prosody and syntax but they were also capable of using such implicit knowledge to disambiguate ambiguous utterances through prosodic grouping.
In summary, under the conditions where the ambiguities were first pointed out in advance and when no other information was available for syntactic disambiguation, the native speakers in our study reliably conveyed the intended interpretation of an ambiguous utterance through prosodic cues or boundaries.
Developmental Stage of L2 Prosody
Previous studies have shown that L2 proficiency level is a factor in the prosody of L2 production and perception. Inother words, there is a developmental stage of L2 prosody. Such a developmental stage is further justified by the present study. More specifically, like native speakers, advanced learners could alter prosodic boundaries to express the appropriate meaning for an ambiguous utterance. On the other hand, limited learners behaved differently from native speakers in that they were unable to produce prosodic boundaries.
75
-9t語教學!£句“正心acfii句“ ζearning
34.3 (Fall 2010)
First of all,when advanced learners wanted to read aloud the meaning of which the adjective modified only the noun immediately following 泣， they lengthened Boundary 2. By contrast, when they read aloud the interpretation of which the adjective modified both the two nouns after it, they lengthened Boundary 1. This suggests that advanced learners, like native speakers, owned some knowledge about the relationship between prosody and syntax in English.
Furthermore,they definitely made effective use of the knowledge to produce prosodic cues to syntactic disambiguation.
It is interesting to find, however, that advanced learners used more distinguished pause durations to produce prosodic boundaries than native speakers. For Interpretation Context 1, the size of boundary effect 切 2
=
.977) for advanced learners was bigger thanthat 切 2
=
.678) for native speakers. Also,for Interpretation Context 2,the effect size (η2
=
.951) for advanced learners was bigger than that(η2
=
.778) for native speakers. The findings reveal that advanced learners' knowledge about the relationship between prosody and syntax in English was explicit,rather than implicit.As mentioned previously, our advanced learners received Chinese language schooling up until their university years. They were unlikely to acquire knowledge about the relationship between prosody and syntax in English by engaging with their caretakers in natural meaningful communication when they were young. In other words, unlike native speakers, these advanced learners learned the knowledge explicitly rather than implicitly. In general, explicit learning involves attention and draws on the conscious monitoring of the behavior to be acquired or retrieved (Destrebecqz & Cleermans, 2001; Fischer, Drosopoulos, Tsen, & Born, 2006; Reber, 1989;
Yang: A Preliminary Study of Prosodic Disambiguation in L2 Production Willingham, 2001). Thus, it is very possible that explicit learning made advanced learners overproduce prosodic boundaries when they struggled to express the appropriate meaning of an ambiguous utterance. The possibility provides a reason why advanced learners used more distinguished prosodic boundaries to resolve the ambiguities than native speakers when reading them aloud.
Next, limited learne時， unlike native speakers, did not show prosodic disambiguation when reading aloud the ambiguities. There are three possibilities to explicate the learners' prosodic behaviors.
The first possibility is that they had not yet succeeded in learning concepts of the syntactic structure of an utterance or sentence.
Therefore, these learners did not know how to capitalize on prosodic cues to syntactic disambiguation. The second is that even if they had learned concepts of the syntactic structure of an utterance, they did not have knowledge about the relationship between prosody and syntax in English. Thus, the fact that prosodic features can help clarify the syntactic structure of an utterance was a completely strange concept to them. The last possibility is that even if they had such declarative knowledge, their declarative knowledge could not successfully be converted into procedural knowledge. They,therefore, failed to produce prosodic boundaries. Regardless of which possibility led to the prosodic behaviors of limited learners, limited English proficiency was doubtless a cause for their failure.
From the perspective of a developmental stage in second language acquisition, it was predicted that advanced learners can behave in a way more similar to native speakers than limited learners could. The present study supported this prediction,which in tum further implied the presence of a developmental stage of L2
77
~語教學 P.ng{is反侵acfiingef,Learning 34.3 (Fall 2010)
prosodic disambiguation production.
Explicit Instruction of Syntax and Prosody
The failure to use prosodic boundaries for an ambiguous utterance impedes successful communication. Therefore,it is of prime importance that English learners have the abilities of perceiving and producing L2 English suprasegmental features. The findings obtained in the present study provide pedagogical implications of English education in Taiwan. First,it is necessary to enhance explicit teaching of what syntactic structures or constituents are, such as verb phrases, noun phrases, and prepositional phrases. This is an initial step to cultivate knowledge about the relationship between prosody and syntax. Selkirk (1984) indicated that syntax is the mediation of sound and meaning.
Moreover, it is important for English teachers in Taiwan to explicitly execute pronunciation training in terms of knowledge, perception, and production of suprasegmental features such as duration, stress, pitch, and intonation contour. The best way of pronunciation training is to expose learners to native English-speaking environments. However, it may not always be feasible to create such environments in English classrooms. As for this problem, Meng et al. (2009) and Zhang et al. (2010) suggested that computer-aided language learning (CALL) platform may be a good alternative to native English-speaking environments. This is because CALL is not restricted to any occasions and places. In brief, explicit instruction is essential to English learners developing English suprasegmental features, and further effectively perceiving and producing prosodic cues.
Yang: A Preliminary Study of Prosodic Disambiguation in L2 Production