Acquisition of Causative Directed Manner of Motion by English-speaking Learners of Chinese
Item Type text; Electronic Dissertation
Authors Peng, Ke
Publisher The University of Arizona.
Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
Download date 27/09/2021 04:38:20
Link to Item http://hdl.handle.net/10150/202949
ACQUISITION OF CAUSATIVE DIRECTED MANNER OF MOTION BY ENGLISH- SPEAKING LEARNERS OF CHINESE
by
Ke Peng
______
A Dissertation Submitted to the Faculty of the
EAST ASIAN STUDIES
In Partial Fulfillment of the Requirements For the Degree of
DOCTOR OF PHILOSOPHY
In the Graduate College
THE UNIVERSITY OF ARIZONA
2011
2
THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE
As members of the Dissertation Committee, we certify that we have read the dissertation prepared by Ke Peng entitled Acquisition of Causative Directed Manner of Motion by English-speaking Learners of Chinese and recommend that it be accepted as fulfilling the dissertation requirement for the
Degree of Doctor of Philosophy
th ______Date: August 5 , 2011 Feng-hsi Liu
______Date: August 5th, 2011 Rudolph Troike
______Date: August 5th, 2011 Chia-lin Pao Tao
______Date:
______Date:
Final approval and acceptance of this dissertation is contingent upon the candidate’s submission of the final copies of the dissertation to the Graduate College.
I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.
______Date: August 5th, 2011 Dissertation Director: Feng-hsi Liu 3
STATEMENT BY AUTHOR
This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at the University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.
Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his or her judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.
SIGNED: Ke Peng
4
ACKNOWLEDGEMENTS
It would have been impossible for me to complete this thesis without the ongoing guidance and support of my committee. I owe many thanks to my adviser and chair of committee,
Feng-hsi Liu , who has mentored and guided me towards numerous intriguing questions related to
Chinese linguistics and Second Language Acquisition of Chinese. I have benefited tremendously from discussions with her and learning from her. Her broad knowledge and expertise of Chinese linguistics, and her open-mindedness as well as her full support for graduate students have made a long-term impact on my career development. Without her encouragement, direction and patience, I could not have finished my work. Professor Feng-hsi Liu read through different
versions of my draft with great patience and provided me with numerous valuable suggestions.
Unfortunately, time has not permitted me to incorporate all her suggestions into the draft, though
I hope to at a later date. Meanwhile, as the author of this manuscript, I acknowledge
responsibility for all the shortcomings the paper may contain.
I am also very grateful to the other two members of my committee. Both Rudolph Troike
and Chia-lin Pao Tao were very supportive and they always made themselves available
whenever I needed help. Professor Rudolph Troike was very resourceful: he not only provided
thorough feedback to my draft, but also referred me to an abundance of literature relevant to my
studies. I benefited quite a bit from the fascinating exchanges I had with both Professor Rudolph
Troike and Professor Chia-lin Pao Tao . I thank them for the positive energy and superb support
of my work. I also thank all the faculty and staff members I have worked with at University of
Arizona.
My experiment was conducted online with the assistance and participation of Jianfei
Chen and her students at Princeton University, Yi Xu and her students at University of Pittsburg, 5
Xia Zhang , Cindy Shih and their students at University of Arizona, Claire Wang and her students at Tucson Accelerated Learning Laboratory, and Jing Ma and his students at Tucson Unified
School District. I really appreciate the kind help I have received from all these teachers and student volunteers.
I want to extend my deep appreciation to my family and friends for their support. I thank my parents and sister for their unconditional love. They sacrificed themselves and provided me with the drive and means to achieve my dreams. They taught me how to practice true love in everyday life. I thank my dear friends, including Yang Yang , Lai Yao , Claire Wang , Tingda Li and many more, who accompanied and took good care of me after I lost my mom in May 2007.
Without these loving and caring friends, I would not have been able to make it through that difficult time of my life. I also want to express special appreciations to my friend and former colleague at University of Arizona, Lucas Wolf , who took the time to proofread my entire thesis and made a significant amount of constructive contribution to my writing. I am very grateful for having such a group of caring and supportive friends.
Additionally, it is truly a blessing that my beloved partner, Robert Allen , is such an angel and supports me in every way that he can. He has inspired and motivated me to further develop myself in many ways. I am also thankful to my mother-in-law, Marcia Allen , who has been supremely helpful all this time.
Finally, to my dear mother, Lanfang (‘Fragrance of Orchid’) Peng , I dedicate this dissertation. 6
TABLE OF CONTENTS LIST OF TABLES…………………………………………………………………………….…10 LIST OF FIGURES…………………………………………………………………………….. 12 LIST OF ABBREVIATIONS……………………………………………………………………13 ABSTRACT……………………………………………………………………………………...14
CHAPTER 1 INTRODUCTION……………………………………………………………………………….16 1.1 Cross-linguistic Variations………………………………………………………..……...16 1.2 L2 Learning Challenges…………………………………………………………..……...18 1.3 Overview of the Dissertation………………………………………………….…….…...22 CHAPTER 2 LINGUISTIC ANALYSIS………………………………………………………………………25 2.1 A Manner Language or A Path Language?………………………..…………..….……...26 2.2 Questions Unanswered in Talmy’s Typologies……………………..……………….…..29 2.3 Equipollently-framed Languages or A Case of Grammaticalization?...... 34 2.4 Questions Needed to be Answered………………………………...………………...... 38 2.5 Event Structure of Causative Directed Manner-of-motion………………………………41 2.6 Light Verb Syntax of Chinese……………………………………………………………48 2.6.1 Characteristics of Light Verbs………………………………………..………..…..48 2.6.2 Light Verbs in Chinese and English……………………………..………………...53 2.7 Licensing Causative Directed Motion with Light Verbs…………..…………………….56 2.7.1 Light Verb ba for CAUSE and zhe for ACCOMPANY………………….………..59 2.7.2 Light Verb dao for the GOAL……………………………………………………..65 2.7.3 Questions Re-addressed……………………………………………………………69 CHAPTER 3 STUDIES OF L2 ACQUISITION OF DIRECTED MOTION………………………………….73 3.1 Learnability Challenges …………………………..……………………………………..73
7
TABLE OF CONTENTS (CONTINUED) 3.2 Parameter Switch or Feature Assembly?……………………..……………….…………74 3.3 Previous SLA Studies of Directed Motion…..……………….……………..………...…76 CHAPTER 4 RESEARCH DESIGN AND METHODOLOGY……………………………………………….81 4.1 Introduction …………..…………………………..………………………………….…..81 4.2 Variables…. …………..…………………………..………….……………….…………84 4.3 Instrument Development……………………..……………….……………..………...…86 4.4 Tasks…. …………..…………………………..………….……………….……………..88 4.4.1 Sentence Token Types……………………………………………………………..89 4.4.2 One Example………………...……………………………………………………..92 4.5 Participants………..…………………………..………….….…………….……………..95 4.6 Reliability and Validity of the Experiment…..………….……………….………………98 4.6.1 Two Experiment Subsets…………………………………………………………..98 4.6.2 Overall Reliability of the Two Subsets…………………………………………….99 4.6.3 Descriptive Statistics of the Two Subsets………………………………………...100 4.6.4 Homogeneity Variances between Two Experiment Subsets……………………..102 4.6.5 Validity of the Instrument………………………………………………………...104 CHAPTER 5 RESULTS AND DISCUSSION………………………………………………………………..114 5.1 Introduction…………..…………………………..…………………………….…….…114 5.2 The Instrument and Participants………………………………………………………..115 5.2.1 Validity and Reliability of the Proficiency Test………………………………….116 5.2.2 Two Proficiency Levels: Correlation and Regression Analysis……………...... 116 5.3 The Results of the Picture Judgment Task………………………………………..…….118 5.3.1 Acquisition of TRANSITIVITY………………………………………………….118
8
TABLE OF CONTENTS (CONTINUED) 5.3.2 Acquisition of GOAL…………………………………………………..………...125 5.3.3 Acquisition of CAUSATIVITY……………………………………………….….130 5.3.3.1 Acquisition of CAUSATIVITY with Transitive Verbs..…………………...131 5.3.3.2 Acquisition of CAUSATIVITY with Intransitive Verbs..………………….136 5.3.3.3 Acquisition of the Light Verb ba…. ………………………………………..142 5.3.4 Acquisition of ACCOMPANY…………………………………………………...145 5.3.5 Acquisition of COMPLEX MANNER…………………………………………...149 5.3.5.1 Sentence Type 1 versus Type 7 with Intransitive Verbs..…………………..151 5.3.5.2 Sentence Type 1 versus Type 8 with Intransitive Verbs .……..……………154 5.3.5.3 Sentence Type 2 versus Type 7 with Intransitive Verbs .…………………..157 5.3.5.4 Sentence Type 2 versus Type 8 with Intransitive Verbs .…………………..161 5.3.5.5 Sentence Type 7 versus Type 8 with Intransitive Verbs .…………………..163 5.3.5.6 Summary………..……………………………………...... 166 5.4 Acquisition of Causative Directed Manner-of-motion Event………………..…………168 5.5 Summary……………………………………………………………………..…………173
CHAPTER 6 DISCUSSION AND CONCLUSION…………………….……………………………………174
APPENDIX A COMMON MANNER-OF-MOTION VERBS IN CHINESE...... ……………………….....…188 APPENDIX B VALIDITY OF EXPERIMENT INSTRUMENT (GOAL)…… .……………...……...... ……189 APPENDIX C VALIDITY OF EXPERIMENT INSTRUMENT (CAUSATIVITY)…...… .…………………194 APPENDIX D VALIDITY OF EXPERIMENT INSTRUMENT (ACCOMPANY).… .………………...……198 APPENDIX E VALIDITY OF EXPERIMENT INSTRUMENT (COMPLEX MANNER)… .………………202
9
TABLE OF CONTENTS (CONTINUED) REFERENCES………………………………………………………...……………………….205
10
LIST OF TABLES
Table 2.1 Differences between Main Verbs, Light Verbs and Auxiliaries………………………36
Table 2.2 Causatives of the Four Classes of Motion Verbs……………………………………...48
Table 4.1 Experiment Variables…………………………………………………………………69
Table 4.2 Sentence Token Types………………………………………………………………...74
Table 4.3 Subject Information (Age)…………………………………………………………….80
Table 4.4 Length of Instruction for L2 Participants……………………………………………..80
Table 4.5 Tested Verbs in the Experiment……………………………………………………….83
Table 4.6 Reliability of the Experiment……………………………………………………….....84
Table 4.7 Mean Ratings of the Sentence Types by Control Group A & B………………………85
Table 4.8 Levene’s Test for Equality of Variances between Control Groups…………………...87
Table 4.9 Descriptive Statistics of Sentence Types for TRANSITIVITY………………….……89
Table 4.10 Paired Samples t-test on TRANSITIVITY……………………………….………….89
Table 4.11 Complete List of Paired Samples t-tests on All Variables………………………….96
Table 5.1 Descriptive Statistics of Participants’ Judgment of TRANSITIVITY by Groups...…103
Table 5.2 Multiple Comparisons of the Groups in Their Judgment of Intransitive Verbs..…....104
Table 5.3 Summary of all Sub-group TRANSITIVITY Analyses……………………………..108
Table 5.4 Descriptive Statistics of Participants’ Judgment of GOAL by Groups…………...…109
Table 5.5 Multiple Comparisons of the Groups in Their Judgment on GOAL PPs…………....110
Table 5.6 Multiple Comparisons of the Groups in Their Judgment on Pre-verbal PPs……..…111
Table 5.7 Multiple Comparisons of the Groups in Their Judgment on Post-verbal PPs……….112
Table 5.8 Descriptive Statistics of Participants’ Judgment of CAUSATIVITY by Groups (with Transitive Verbs)….………..…………………………………………………………………..115
Table 5.9 Multiple Comparisons of the Groups in Their Judgment of CAUSATIVITY (with Intransitive Verbs)….………………………………..…………………………………..……..116
11
LIST OF TABLES (CONTINUED)
Table 5.10 Multiple Comparisons of the Groups in Their Judgment on Sentence Type 1 with Transitive Verbs for Causative Event with Ba Construction…………………………………...118
Table 5.11 Descriptive Statistics of Participants’ Judgment of CAUSATIVITY by Groups (with Intransitive Verbs)….………………………………………..………………………………....120
Table 5.12 Multiple Comparisons of the Groups in Their Judgment of CAUSATIVITY (with Intransitive Verbs)...………………………………..…………………………………..……....121
Table 5.13 Multiple Comparisons of the Groups in Their Judgment on Sentence Type 1 (with Transitive Verbs) for Causative Event with Ba Construction..…………………..…………….122
Table 5.14 Multiple Comparisons of the Groups in Their Judgment on Sentence Type 2 (with Intransitive Verbs) for Causative Event with Ba Construction..…………….…………………123
Table 5.15 Descriptive Statistics of Participants’ Judgment on Sentences with Ba construction …………………………………………………………………………………….127
Table 5.16 Descriptive Statistics of Participants’ Judgment of ACCOMPANY by Groups (with Transitive Verbs)….…………………………..……………………………………………...... 128
Table 5.17 Multiple Comparisons of the Groups in Their Judgment on Transitive Sentence Type 4 for Accompanied Actions with Zhe ……………..……………………………………………132
Table 5.18 Descriptive Statistics of Participants’ Judgment of COMPLEX MANNER by Groups……………..……………………………………………………………………………134
Table 5.19 Multiple Comparisons of Group Judgments on Sentence Type 1 (with Intransitive Verbs).………………………………..…………………………………………………………136
Table 5.20 Multiple Comparisons of Group Judgments on Sentence Type 7……………….…137
Table 5.21 Multiple Comparisons of Group Judgments on Sentence Type 8………...……..…140
Table 5.22 Multiple Comparisons of the Groups in Their Judgment on Sentence Type 2……………………………………………..………………………………..…………………143
Table 5.23 Multiple Comparisons of the Groups in Their Judgment on Sentence Type 7 and Sentence Type 8 with Intransitive Verbs…..………………………………………...…………146
Table 5.24 Acquisition of COMPLEX MANNER with Intransitive Verbs…..………………..152
Table 5.25 Summary of Results of Two-way Repeated Measures ANOVA…………………..156 12
LIST OF FIGURES
Figure 4.1 Example Picture………………………………………………………………………76 Figure 4.2 Example Picture………………………………………………………………………77 Figure 4.3 Result of Control Group A…...………………………………………………………85 Figure 4.4 Result of Control Group B…...………………………………………………………86 Figure 4.5 Effect of TRANSITIVITY ON Sentence Type 1…………………………………….90 Figure 4.6 Effect of TRANSITIVITY ON Sentence Type 2…………………………………….91 Figure 4.7 Effect of TRANSITIVITY ON Sentence Type 3…………………………………….92 Figure 4.8 Effect of TRANSITIVITY ON Sentence Type 4…………………………………….92 Figure 4.9 Effect of TRANSITIVITY ON Sentence Type 5…………………………………….92 Figure 4.10 Effect of TRANSITIVITY ON Sentence Type 6…………………………………...94 Figure 5.1 Judgment of TRANSITIVITY by Participant Groups……………………………...119 Figure 5.2 Judgment of GOAL by Participant Groups…………………….…………………...113 Figure 5.3 Judgment of CAUSATIVITY with Transitive Verbs by Participant Groups………119 Figure 5.4 Judgment of CAUSATIVITY with Intransitive Verbs by Participant Groups……..125 Figure 5.5 Judgment of Sentences with Ba Construction by Participant Groups……………....128 Figure 5.6 Judgment of ACCOMPANY with Transitive Verbs by Participant Groups………..133 Figure 5.7 Grammaticality Judgment of Sentence Type 1 vs Type 7 with Intransitive Verbs by Participant Groups……………………………………………………………….……………..138 Figure 5.8 Grammaticality Judgment of Sentence Type 1 vs Type 8 with Intransitive Verbs by Participant Groups……………………………………………………………… ……………..141 Figure 5.9 Grammaticality Judgment of Sentence Type 2 vs Type 7 with Intransitive Verbs by Participant Groups ……………………………………………………………………………..144 Figure 5.10 Grammaticality Judgment of Sentence Type 2 vs Type 8 with Intransitive Verbs by Participant Groups ………………………………………………………………..……………146 Figure 5.11 Grammaticality Judgment of Sentence Type 7 vs Type 8 with Intransitive Verbs by Participant Groups ………………………………………………………………..……………149 Figure 5.12 Grammaticality Judgment of Sentences with Intransitive Verbs by Participant Groups …………………………………………………………..……….……………….……151 13
LIST OF ABBREVIATIONS
ACC: Accusative Marker
BA: Ba Construction, ‘grasp/take’(Archaic), Object Marker
CFL: Chinese as a Foreign Language
CL: Classifier
DE: Adverbial Manner Marker
DAO: Preposition -dao
LE : Perfect Aspect Marker -le
LOC: Locative Marker
PST: Past Tense
SLA: Second Language Acquisition
TOP: Topic Marker
UG: Universal Grammar
YONG: ‘Use/Using’ , Co-verb, Instrumental marker
ZHE: Durative Aspect Marker -zhe 14
ABSTRACT
This study aims to explore how English-speaking learners of Chinese acquire directed manner of motion. The thesis consists of six chapters in an attempt to achieve two goals: firstly, to provide linguistic analysis to account for the Chinese-English differences in licensing causative directed manner of motion events; and secondly, to present a second language acquisition study informed mainly by a picture acceptability judgment task.
The causative directed manner of motion events in Chinese behave differently from their counterparts in English. While the causatives of transitive and intransitive use of manner of motion verbs share similar syntactic structures in English, for example, John walked Mary to her car. , John kicked the ball to the wall ., their counterparts in Chinese display distinctively different surface structures. As a result of my dissatisfaction with the initial analysis with Talmy’s (1975
& 1985) conflation typologies and Slobin’s (1996 & 2004) three-way proposal, I adopted another approach formulated within the Parameters and Principles Theory (Chomsky and Lasnik, 1993), namely the light verb analysis (Huang, 1997 & Lin, 2001) to explicate the Chinese-English distinctions in licensing complex motion events. Following Huang (1997 & 2006) and Lin’s
(2001) proposal, I argue that the argument structure of the causative directed motion is licensed via several light verbs at the syntactic level (S-syntax) in Chinese, and at the sub-syntactic level
(L-syntax) in English. The manner of motion verbs in English undergo a full conflation in the lexical computation and enter the syntax with full-fledged functions to formulate the phrase structure of a causative directed motion. However, the argument structure of causative directed motion in Mandarin Chinese is realized with the light verb ba for the causative sub-event, the light verb zhe for the accompanied action, and the light verb dao for the goal of the motion. 15
These cross-linguistic variations pose at least two questions regarding English-speaking learners of Chinese: (1) how do they acquire ways to describe causative directed motion in the target language; (2) if and/or when do they actually become aware of the parameter setting distinctions (L-syntax versus S-syntax). The results of the present acquisition study are informed by a picture grammaticality judgment and a translation task from a pilot study. A total number of seventy-three participants completed the study. Forty of them were native speakers of Chinese, serving as the control group. The other thirty-three were English-speaking learners of Chinese which were placed within two proficiency groups. Four acquisition patterns were identified based on a two-way repeated measures ANOVA.
16
CHAPTER1
INTRODUCTION
This dissertation is devoted to studying causatives of directed manner-of-motion exemplified in (1.1-1.3) from both cross-linguistic and acquisitional perspectives.
(1.1) a. John walked to the car. b. John walked Mary to her car.
(1.2) a. The ball rolled down the hill. b. John rolled the ball down the hill.
(1.3) a. * Mary pushed. b. Mary pushed the cart into the store. Directed manner-of-motion here refers to the composition of a manner-of-motion verb
(walk , roll , & push ) with a prepositional phrase indicating the goal of the activity (to her car, down the hill & into the store ). The causativization of directed manner-of-motion introduces an even more complicated semantic and syntactic structure by adding a causer to the event demonstrated in the (b) sentences above.
1.1. Cross-linguistic Variations
My interest in this topic first started with the unavailability of the equivalent causative form of (1.1b) in Chinese. It is noted that the same event described in the English sentence (1.1b) may be put in several ways in Chinese, like (1.4b) and (1.4c), but Chinese does not allow the causative form, shown in (1.4a). A close look at (1.4b) and (1.4c) shows that a lexicalization of the implied meaning “accompanying” is licensed in the syntactic representation, see pei in (1.4b) and he in (1.4c).
17
(1.4) a. *John 走 Mary 到她的汽车(那儿)。 John zou Mary dao ta de qiche (na’er) John walk Mary to her car (there) ‘John walked Mary to her car.’
b. John 陪(着)Mary 走到她的汽车(那儿)。 John pei(zhe) Mary zou dao tade qiche (na er). John accompany(-ing) Mary walk to her car (there) ‘John accompanying Mary walked to her car.’
c. John 和 Mary 走到她的汽车(那儿)。 John he Mary zou dao tade qiche (na er). John and Mary walk to her car (there) ‘John and Mary walked to her car.’
Unlike that of (1.1), the translation of (1.2a) and (1.2b) both exist in Chinese as shown in
(1.5a) and (1.5b). These examples illustrate that the unaccusative manner-of-motion verb roll in
Chinese can also participate in causatives of the directed motion with Ba construction.
(1.5) a. 那个球滚下了山。 Nage qiu gun xia le shan. That-CL ball roll down PERF hill. ‘That ball rolled down the hill. ‘
b. John 把那个球滚下了山。 John ba nage qiu gun xia le shan. John BA that-CL ball roll down LE hill. ‘John rolled the ball down the hill.’
c. ?John 滚那个球下了山。 John gun nage qiu xia le shan. John roll that-CL ball down PERF hill. ‘John rolled the ball down the hill.’
The transitive manner-of-motion verb push in Chinese can also function similarly as a causative of the directed motion with or without the use of Ba construction, (1.6b) and (1.6c) respectively.
18
(1.6) a. *Mary 推。 Mary tui. Mary push. ‘Mary pushed.’
b. Mary 推了那辆车到商店。 Mary tui le naliang che dao shangdian. Mary push PERF that-CL cart to store. ‘Mary pushed that cart to the store.’
c. Mary 把那辆车推到了商店。 Mary Ba that-CL cart push to PERF store. ‘Mary pushed that cart to the store.’
The surface distinctions in the syntactic representations of causatives of directed manner-
of-motion between English (1.1-1.3) and Chinese (1.4-1.6) is observed systematically and has
led to the first research question of this thesis: How to account for the differences between
Chinese and English in the representations of causatives of directed manner-of-motion ?
1.2. L2 Learning Challenges
Additionally, I am also interested in the acquisitional aspect of the issue. As a foreign learner of English myself, I have personally encountered some challenges in understanding the following directed motion events in the target language, and I am very uncertain of the odds that
I will catch myself using these structures in my own L2 output.
(1.7) The train rumbled through the tunnel. (1.8) She rustled out of the room. (1.9) The old man zigzagged towards the gate. (1.10) The hurricane whirled away the tree.
19
A preliminary examination seems to reveal that English has a rich inventory of motion verbs, for example, noise verbs, path-shape verbs, verbs describing manners, etc., and learners have to learn them one by one. What makes it even more confusing for Chinese learners like me is that when we resort to our L1 for reference, it turns out that these similar sentences in English are expressed in a totally different way in Mandarin Chinese, involving a variety of syntactic measures:
(1.11) 火车轰鸣着穿过了隧道。 Huoche Hongming-zhe chuan-guo le suidao. train rumble-ZHE pass-through LE tunnel. ‘The train rumbled through the tunnel.’
(1.12) 她跑出房间，衣服沙沙作响。 Ta pao chu fangjiang, yifu shasha-zuoxiang. She run exit room cloth rustle-sounded ‘She rustled out of the room.’
(1.13) 老人歪歪扭扭地走向大门。 Laoren waiwai niuniu de zou xiang damen. Old-man crookedly walk toward gate ‘The old man zigzagged towards the gate.’
(1.14) 飓风把树卷跑了。 Jufeng ba shu juan pao le. hurricane BA tree whirl away LE ‘The hurricane whirled the tree away.’
As noticed, in (1.11), a separate adverbial phrase marked by the continuative aspect marker zhe is used; in (1.12), one of the manner verbs shasha-zuoxiang ‘rustle-sounded’ is expressed as a separate clause; to describe the manner and path in (1.13), an adverbial phrase,
‘waiwai niuniu de’ is utilized; whereas in (1.14), ba construction is highly preferred. Simply by comparing the syntactic structures involved above, it seems that Mandarin Chinese speakers
20 have distinctively more complex ways of describing the same motion events than English speakers.
The cross-linguistic differences of motion verbs between Chinese and English have also caused challenges for English-speaking learners of Chinese. In order to delineate a picture of a girl hopping into the hospital in Chinese, English-speaking learners have come up with several versions. Consider the following examples in (1.15) – (1.18).
(1.15) 她单脚跳去了医院。 ta danjiao tiao qu le yiyuan. she single-foot jump go/to LE hospital ‘She jumped on one foot into the hospital.’
(1.16) 她单脚跳着去了医院。 ta danjiao tiao zhe qu le yiyuan. she single-foot jump ZHE go/to LE hospital ‘She went to the hospital, jumping on one foot.’
(1.17) 她用一只脚跳去了医院。 ta yong yi-zhi-jiao tiao qu le yiyuan. she use one-CL-foot jump go/to LE hospital ‘She used one foot (to) jump into the hospital.’
(1.18) ?她单脚去了医院。 ta danjiao qu le yiyuan. she single-foot go LE hospital ‘She went to the hospital, (on) single foot.’ Examples (1.15) – (1.18) display that there is no corresponding word for the meaning of
“hop” in Mandarin Chinese. The same semantic meaning can be expressed by combining some
words to form a compound phrase, for example Dan-jiao tiao ‘single-foot jump’ in (1.15).
These sentences are all acceptable in Mandarin Chinese, but each has varying emphasis.
Sentence (1.15) entails a telic meaning: the main VP is tiao ‘jump’ with a deictic element qu
‘go/to’ and dan-jiao ‘single-foot’ serves as the manner component of the motion verb tiao ‘jump’.
21
Qu can be a main verb, ‘go’, but in modern Chinese it can also function as a co-verb or preposition, ‘to’. (Li & Thompson, 1981). Sentence (1.16) has the main verb as qu ‘go/to’ with a
continuative adverbial phrase dan-jiao tiao zhe ‘single-foot jump ZHE’ as the manner
component. Sentence (1.17) has a serial verb construction with the first VP yong yi-zhi-jiao ‘use
one-CL-foot’ and the second VP tiao ‘jump’ with a deictic element qu ‘go/to’. The main verb for
(1.18) is qu ‘go/to’ with a NP dan-jiao ‘single-foot’ as the manner component.
In addition to (1.15) – (1.18), there might be other ways to express the same meaning of
English word ‘hop’ in Mandarin Chinese. But no matter what it is, the word ‘hop’ is represented
with complex semantic meanings and syntactic structures in Chinese by compositional measures.
If we draw the syntactic trees for these, we will find a vP on top of another VP for these
examples except (1.18) which has a NP dan-jiao ‘single-foot’ functioning as the manner
component.
These contrasting syntactic representations demonstrated in the acquisitional data, shown
in (1.11)–(1.18), bring about some acquisitional research questions. For instance, how do
English-speaking learners of Chinese and Chinese-speaking learners of English learn to portray a
complex motion event in the target language? These are broad intriguing research questions
deserving numerous studies. In this present study, the scope of the research is pragmatically
narrowed down to only one type of complex motion event – causative directed manner-of-
motion events, and only part of the learner population – English-speaking learners of Chinese.
Specifically, the second research question of this dissertation asks: How do English-speaking
learners of Chinese acquire causative directed manner-of-motion events ?
22
1.3. Overview of the Dissertation
The dissertation aims to first study the Chinese-English distinctions in licensing causatives of directed manner-of-motion structures and then investigate how these variations affect the second language acquisition of Chinese by English-speaking learners. The thesis consists of six chapters. Chapter 2 provides the major theoretical framework of this study. It starts with the most well-known of Talmy’s cross-linguistic typologies of lexicalization patterns for motion verbs and language acquisition studies adopting this model. A detailed review and critique of the application of this proposal is presented in an attempt to explain why I find this proposal cannot provide a unified coherent explanation for the cross-linguistic variances between
Chinese and English. Instead, a syntactic approach is adopted, using the Light Verb Syntax of
Chinese (Huang, 1997; Lin, 2001) to explicate the Chinese-English distinctions in licensing causatives of directed motions. According to Lin (2001), Chinese and English stand at the two extremities in the entire typological spectrum: arguments in English are licensed at the sub- syntactic level of grammar, or L-Syntax in Hale and Keyser’s definition (1991 & 1993), whereas arguments in Chinese must be spelled out and are hence licensed at the syntactic level, or S-
Syntax. This proposal is grounded on Principles and Parameters Theory (Chomsky, 1981 &
1995) and it advocates that Chinese and English differ in the parameter of lexicalization.
Chapter 3 extends the literature review from the formal linguistic analysis of the issue to
its application in Second Language Acquisition (SLA) studies. Whether Principles and
Parameters Theory has significant implications in SLA research is a highly controversial issue in
the field of SLA studies. Proponents are excited with its potential to more precisely characterize
the grammatical distinctions between a learner’s L1 and L2. SLA can be determined by a
learner’s degree of success in ‘resetting’ one or more parameters from the L1 value to those of
23
L2 (Haegeman, 1988). The use of parameters in acquisition studies may help discover clusters of language properties associated with one issue and the value setting of one particular parameter may lead to “a cascade of related effects” in enhancing the language learning experience.
However, other researchers hold opposite opinions; for instance, Lardiere (2008 & 2009) has critically reviewed the parameters resetting hypothesis and the rehabilitating contrastive analysis in second language acquisition. Lardiere (2009a) points out that the theoretical analyses with the application of parameters fail to fulfill the promise of the “deductive-consequences conjecture”.
She also suggests that the process of second language acquisition be best described with the re- selection and re-assembly of features into language-specific lexical items.
With the linguistic theories and SLA debates in mind, I have designed an experiment to investigate how English-speaking learners acquire the expression of causative directed motion in
Chinese. Chapter 4 explains the research design and methodology of this study, while Chapter 5 illustrates the results of the experiment. In addition to a pilot study, three groups of participants
(N=73) were recruited to participate in the formal experiment. Forty were native Chinese speakers serving as the control group and also to validate the reliability of the research instrument. The non-native speakers comprised seventeen intermediate learners and sixteen advanced learners of Chinese. This experiment examined the acceptability judgment of these three groups on five semantic variables: TRANSITIVITY, GOAL, CAUSATIVITY,
ACCOMPANY, and COMPLEX MANNER. The definition of these variables would be provided in Chapter IV, while the intricate results would be presented in great length in the
Chapter V. Four patterns were found to emerge in the acquisition of causative directed motion expression in Chinese.
24
The final chapter of this thesis addresses two issues: what do the data and results imply theoretically and pedagogically? Based on the results of our study, we find reasons to argue that language acquisition is unlikely to be a simplistic process of parameter re-setting, but rather it would be more likely to be a process of feature re-assembly. However, the lack of superior-level learner example in the sample refrain me from making any definite conclusion. I propose that explicit instruction of grammar is a necessity in teaching English-speaking learners how to describe causative directed-motions in Chinese, especially when negative evidence is lacking in the L2 input.
25
CHAPTER 2 LINGUISTIC ANALYSIS
This chapter surveys the literature on the linguistic analysis of causative directed motion in Chinese and English in an endeavor to answer the first research question of this study: how to account for the differences between Chinese and English? We will start with the best known typology of lexicalization patterns for motion events (Talmy, 1975, 1985) and explore how this proposal has informed, evolved and been applied in the analysis of cross-linguistic variation, especially with regard to Chinese and English. There are questions this particular strand of proposals fails to resolve, for example:
• Why does Chinese have a greater number of manner-adverbials, a feature that resembles Path
Languages like Japanese and Korean?
• Why cannot noise or emission words derive into manner-of-motion verbs in Chinese?
• Why does transitive 1 and intransitive use of motion verbs differ so distinctively in Chinese
but not in English?
Additionally, this proposal has few pragmatic implications for real classroom teaching from an
acquisition perspective. For instance, it is not very informative or helpful to explain the language
differences in a general sense and expect that learners will automatically learn how to break up a
1 Transitivity is a complicated issue and it is especially the case with verbs in Chinese which lacks a clear-cut distinction between the two notions. Some verbs have both transitive and intransitive uses, some verbs like walk and run that are commonly defined as unergative verbs in other languages have transitive surface structures in Chinese even though no transitivity is involved in the meaning per se . Transitivity in this thesis is defined from a lexical perspective, i.e. whether the effects of an action ‘passes/carries over’ or ‘transfers’ from an agent to a patient (Thompson, 1973; Hopper & Thompson, 1980). For example, zou shan-lu ‘walk (on) mountain road’ has a transitive surface structure, but it is not considered as a transitive verb in the sense that it does not ‘transfer’ or ‘carry over’ any effect of the walking action from the agent/the person to the patient/ the road. Instead, zou shan-lu is still treated as unergative in this study. For a more lengthy discussion of the nature and structure of transitive/intransitive verbs in Chinese, readers are referred to Huang (2006), Lin (2001), Tieu (2007) and Tsang (2003).
26 scene into several components and use separate action clauses to describe segment by segment what take places in this event.
As a result of my dissatisfaction with Talmy’s conflation typologies and Slobin’s three- way proposal especially from a second language acquisition standpoint of view, I adopted another approach formulated within the Parameters and Principles Theory, namely the Light
Verb Syntax of Chinese (Huang, 1995 & T. Lin, 2001). The following sections are structured following this line of reasoning: in Section 2.1, Talmy’s typologies of lexicalization patterns for motion events are briefly reviewed and applied to the analysis of Chinese-English distinctions, which leaves us some questions unanswered (Section 2.2). It turns out that defining Chinese as an equipollently-framed language (Section 2.3) is not informative and useful to distinguishing the differences between the two languages. For instance, when the inventory of lexical items is similar in two languages, i.e. the prevalence of manner-of-motion verbs, why does the transitivity nature of motion verbs affect the availability and productivity of causative directed motion? Furthermore, why are certain syntactic constructions highly preferred or even required in Chinese, but not in English, e.g. the ba construction, the durative aspect marker zhe , etc. It is
my belief that any successful theory of motion events must be able to identify cross-linguistic
variations and account for what gives rise to the differences in the argument structures for the
same motion event. With these redefined questions in mind (Section 2.4), I propose to
investigate the issue applying the light verb analysis to Chinese.
2.1. A Manner Language or Path Language?
The semantic and syntactic interaction of motion verbs has been a central focus of studies since Talmy (1975) and Jackendoff (1976). These constructions are of particular research value
27 in both cross-linguistic variation analysis and second language acquisition because they seem to reveal the causal connection between syntax and semantics (Folli & Harley, 2006) despite the debate on which is cause and which is effect. Amongst all the proposals explaining the availability and productivity of directed-motion constructions, the most well-known and well- studied explanation is Talmy’s works, especially his crosslinguistic typologies of lexicalization patterns (Tamly, 1975 & 1985) and event integration (Talmy, 1991 & 2000). His proposals are widely studied and have been applied to different language studies in the field of general linguistics (Italian & Korean: Zubizarreta, 2007; Hindi: Narasimhan, 2002; Chinese: Bai, 2007 &
Han, 2007, among others) and in the field of second language acquisition (Spanish & English:
Slobin, 1985 & Montrul, 2001; Japanese & English: Inagaki, 2001; Chinese- & Japanese- speaking learners of English: Yu, 1997, etc.).
Talmy (1975, 1985) is concerned with one major question: how semantic elements can profile the surface structure. His focus is on the semantics of motion verbs in this respect.
According to Talmy (1975, 1985), a motion event consists of both internal elements (Motion,
Figure, Ground, Path) and external elements (Manner/Cause). He defines each of the elements in the following way:
Motion : the fact that some entity changes its spatial location
Figure : the entity that moves
Ground : the entity with respect to which the figure moves
Path : the course followed by the figure with respect to the ground
Manner: the way in which the figure moves
Cause: the cause due to which the figure moves
28
Talmy also identifies three major ways in which different elements of motion are conflated in the meaning of a verb in language. The first type is one in which Manner or Cause is conflated with Motion in the meaning of a verb. Both English and Chinese languages are classified as having this conflation pattern as a dominant one (Talmy, 1985) with examples provided in (2.1) and (2.2). The second type is one in which the verb root expresses both Motion and Path. Romance languages, Spanish for example in (2.3), Japanese in (2.4) and Korean all belong to this type. The third type is the conflation of Motion and Figure as in (2.5).
(2.1) The bottle floated into the cave. [Talmy, 1985]
(2.2) 瓶子漂到了洞里。 [Chinese equivalent of Talmy’s example (2.1)]
Pingzi piao dao le dong li. Bottle float to LE cave inside ‘The bottle floated into the cave.’ (2.3) La botella entró a la cueva flotando. [Talmy, 1985] The bottle moved.in to the cave floating. ‘The bottle entered (moved-in) to the cave floating.’ (2.4) Taro wa kawa o aruite watat-ta. [Matsumoto, 2003] Taro TOP river ACC walk cross-PST ‘Tara walked across the river.” (2.5) It rained in through the bedroom window. [Talmy, 1985]
Since the third type – the conflation of Motion and Figure – is restricted to relatively
unfamiliar languages, most studies in the literature concentrate only on the previous two types.
The languages that have the dominant Manner + Motion pattern and the Path + Motion pattern are sometimes called manner languages and path languages, respectively (Wienold, 1995).
29
Manner Language (Satellite-framed lexicalization) MV (Manner/ Cause + Move) + Satellite (Path) Path Language (Verb-framed lexicalization) MV (Path + Move) (+ adjunct Manner/Cause expression) This strand of the proposal provides insightful explication of the cross-linguistic variations in some languages, for example, the inventory of lexical items in English exhibits enormous differences from Italian, with the prevalence of manner-of-motion verbs and availability of Path-denoting PPs in the former language and the salience of inherently directed motion and absence of Path-denoting PPs in the latter.
2.2. Questions Unanswered in Talmy’s Typologies Even though Chinese and English are both satellite-framed or manner languages (Talmy,
1985), i.e. the dominant pattern in these two languages is to encode the MANNER or CAUSE in the framing event and the PATH information is projected in the subordinate satellite component, there are fundamental distinctions between the two languages in their expression of manner-of- motion verbs. Firstly, they differ in the number and the formation of Manner-of-motion VPs.
Mandarin Chinese has a comparatively smaller number of Manner-of-motion verbs than English
(See Appendix 1). Levin (1993) has listed more than 100 Manner-of-motion verbs in English and this list is by no means exclusive. The run -class and roll -class of Manner-of-motion in English has about 100 tokens for each category. (2.6a) is only a partial list (cf. Levin’s run -class and roll -class (1993)) and it includes over 65 tokens.
In contrast, (2.6b) is Bai ’s (2007) list of Manner-of-motion verbs for the walk -class in
Mandarin. It is observable that many walk -class verbs in Mandarin are actually compound phrases combining the attributive descriptions/Manner-Adverbial with the general verb zou
30
‘walk’ or xing ‘move’. The Manner-Adverbials are usually marked with 地 de ‘DE’, which is the
Manner-adverbial marker in the language.
(2.6a) amble, bounce, bowl, canter, charge, clamber, crawl, creep, dart, dash, dodder, drift, flit, float, fly, gallop, gambol, glide, hasten, hike, hobble, hop, hurry, inch, jog, jump, leap, limp, lumber, march, mosey, pad, parade, plod, prance, prowl, race, roam, roll, run, rush, saunter, scramble, scurry, shamble, shuffle, skip, skulk, slide, slither, sneak, speed, stagger, streak, stride, stroll, strut, wager, sweep, swim, toddle, totter, tramp, trot, waddle, wade, walk, wander
(2.6b) 漫步 man-bu ‘amble’ or ‘stoll’, 弯着腰走 wan-zhe-yao zou ‘creep’, 一瘸一拐地走 yi- que-yi-guai DE zou ‘limp’, 艰难地走 jian-nan DE zou ‘plod’, 齐步走 qi-bu zou ‘march’, 缓慢而有规律地走 huan-man-er-you-gui-lu DE zou ‘pace’, 轻轻走 qing-qing zou ‘pad’, 大步走 da-bu zou ‘stride’, 趾高气扬地走 zhi-gao-qi-ang DE zou ‘strut’, 蹒跚而行 pan- shan-er-xing ‘stagger’, 僵直地走 jiang-zhi DE zou ‘stump’, 拖着脚步走 tuo-ZHE-jiao-bu zou ‘shuffle’, 悄悄走 qiao-qiao zou ‘sneak’, 踮着脚走 dian-ZHE-jiao zou ‘tiptoe’, 重踏步 走 zhong-ta-bu zou ‘tramp’, 缓慢而吃力地走 huan-man-er-chi-li DE zou ‘trudge’, 踉踉 跄跄地走 niang-niang-qiang-qiang DE zou ‘totter’, 闲逛 xian-guang ‘wander’, 摇摆地走 yao-bai DE zou ‘waddle’, 费力地走 fei-li DE zou ‘wade’
Furthermore, the fact that Mandarin has a smaller set of Manner-of-motion verbs in the walk -class and it makes finer manner distinctions in the Manner-adverbials reveals the similarities between Mandarin and Path-languages like Japanese and Korean. Wienold (1995) hypothesizes that “ if a language has a specific word class lexicalizing adverbs of a manner type, such a language will always be poor in motion verbs of manner ” (p.317). It seems that
Mandarin consists of a rich inventory of Manner-adverbials, which modifies the general class of motion verbs, but on the other side, Mandarin does not necessarily lack the variety of Manner-of- motion verbs. Take Pei Zhu’s (2007) enumeration for example, Modern Chinese has more than sixty frequently-used manner of motion verbs and this is not a complete inventory. What is more interesting is that Wienold’s hypothesis is supported in Japanese and Korean, both of which fall
31 into the category of Path-language, as they lexicalize manner primarily in adverbs and do not comprise a rich set of Manner-of-motion verbs (Matsumoto, 2003).
Another distinction between English and Mandarin concerns the availability of some types of Manner-of-motion. Once again, this is not addressed or accounted for in Talmy’s typologies of lexicalization. As example (2.11) and (2.12) in Chapter 1 demonstrate sound or emission words cannot directly turn into manner-of-motion verbs in Mandarin Chinese, but such words in English are versatile and can function as nouns, interjections and manner-of-motion verbs (Matsumoto, 2003). Instead, these words usually appear as manner-adverbials in Mandarin
Chinese (He, 2006), the feature of which mirrors the syntactic characteristics of Japanese and
Korean (Matsumoto, 2006). Ironically, both Japanese and Korean are Path Languages according to Talmy. Consider the examples in (2.7) – (2.9) in both English and Mandarin Chinese. The sound or onomatopoeic words become part of the manner-adverbials which are marked or can be marked by zhe (the durative aspect marker), or de (the manner adverbial marker).
(2.7a) The train rumbled through the tunnel.
(2.7b) 火车轰鸣着穿过了隧道。
Huo-che hong-ming zhe chuan guo le sui-dao. Train rumbling ZHE pass through LE tunnel ‘The train passes rumbling through the tunnel.’
(2.8a) She rustled out of the room.
(2.8b) 她跑出房间，衣服沙沙作响。
Ta pao chu fang-jian, yi-fu sha-sha zuo xiang. She run out room, clothes rustling as sound ‘She (clothes rustling) ran out of the room.’
32
(2.9a) The bullet whistled through her window.
(2.9b) 子弹从她的窗户里呼啸而过。
Zi-dan cong ta de chuan-hu li hu-xiao er guo. Bullet from she CL window LOC-inside whistling CL pass ‘The bullet whistling pass through her windown.’
Bai in 2007 also points out that English derives a set of verbs from NPs to depict the
manner-of-motion, which is usually not available in Mandarin. Her examples are given in
Chinese with the English annotation edited by me here in (2.10) and (2.11). Bai notices that
metaphorical lexical items are usually adopted instead, for instance, 像 xiang, 如 ru, 若 ruo ‘like’ or ‘seem’ in the manner-adverbials.
(2.10a) The train snailed up the steep grade.
(2.10b) 火车蜗牛般慢慢爬上陡坡。
Huo-che wo-niu ban man-man pa shang dou-po. Train snail like slowly climb up steep grade ‘The train snail-like climbed up the steep grade.’
(2.11a) Unemployment levels have rocketed to new height.
(2.11b) 失业率迅速上升到新的水平。
Shi-ye lu xun-su shang sheng dao xin de shui-ping. Unemployment rate rapidly up rise to new CL level ‘Unemployment rapidly rise up to the new level.’
33
Another point that Bai (2007) has made is that Mandarin employs compounding to derive complex manner-of-motion, but the inventory of such lexical items is limited and this type of compounding not very productive. Such examples include 鲸吞 jing tun (whale-eat) ‘engulf’, 蛇
行 she xing (snake move) ‘slithering’, 鼠窜 shu cuan (rat rush) ‘scurry (like rats)’ or ‘scamper’,
etc.
Similar to the compounding explanation, Matsumoto proposes a manner categorization
parameter in 2003 on the basis of Wienold’s hypothesis (1995) and the distributional differences
of onomatopoeic words in English versus Japanese and Korean. According to him, English
belongs to Manner-in-verb languages, while Japanese and Korean belong to Manner-in-adverb
languages, in which manner distinctions are primarily made by adverbials. This parameterization
concerns the “lexical categories” (p. 412) in which manner is coded, and it does not specify
“where in a sentence they are used”. Mastumoto emphasizes that this parameter does not aim to
set a clear-cut distinction between the two types (Matsumoto, 1995). Following Mastumoto’s
proposal, Mandarin has both features of Manner-in-verb and Manner-in-adverb languages.
In the above two sections, I have made an attempt to decipher the distinctions between
Chinese and English in licensing causative directed motion with Talmy’s lexicalization
dichotomy. However, this endeavor turns out to be unsuccessful and it has left us many
unanswered question. If both Chinese and English belong to Satellite-framed languages, why do
they differ so much in the number and the formation of Manner-of-motion VPs? What gives rise
to the availability, the formation and the distribution of manner-in-adverbs motion VPs in
Chinese in the first place? Does this give Chinese more of the features of Path languages? This
34 makes Talmy’s lexicalization dichotomy of motion events susceptible to questions and further verification.
2.3. Equipollently-framed Languages or A Case of Grammaticalization? As discovered from the previous analysis, Chinese exhibits features of both Satellite-
framed Lexicalization and Verb-framed Lexicalization which leaves us skeptical about this
simplistic proposal. As a matter of fact, Talmy’s classification has recently been challenged by
Slobin (1996, 2004), Tai (2003), and Chu (2004). Tai (2003) argues that Chinese is primarily a
verb-framed language, based on evidence that the path verb is the main verb, rather than the
manner verb.
Unlike Tai (2003), Chu argues in 2004 that both the satellite-framed pattern and verb-
framed pattern are available in Chinese for expressing the motion elements of Move, Path,
Manner , and Cause . But he is also aware that there are distributional differences between the two
patterns. On one hand, satellite-framed lexicalization profiles the Manner or Cause of motion in
Mandarin and is suitable for expressing realized motion. Under this frame of lexicalization,
either the Manner is significant within the speaker’s focus of attention, or the Cause is
highlighted in the causative of the motion. On the other hand, verb-framed lexicalization is also
frequently used in Mandarin. This frame of lexicalization does not specify the exact Manner and
Cause of the motion, and thus is the appropriate option for conveying motion when Manner and
Cause are not at issue in the conceptualization. The following two sentences are Chu’s (2004)
two counter examples against Talmy’s classification of Chinese (p.128-130).
35
(2.12) 三爷哪天下山？ sanye na tian xia shan? Sanye which day descend mountain? ‘When did Sanye go down the mountain?’
(2.13) 走，我们回家。 zou, women huijia. go we move-back home. ‘(Let’s go!) Let’s go back home.’ The main verb xia ‘descend’ in (2.12) and hui ‘move-back’ in (2.13) encoded both Move and Path of the motion. Chu (2004) indicates that words functioning as Path-verbs and Path- complements actually fall into a closed set (see the list in (2.13)). Therefore, a third type of lexicalization pattern is added to Talmy’s binary typology by some researchers. Slobin (2004) advocates a new category, equipollently-framed language like Chinese, for which the specifications of path and manner information are achieved by equivalent grammatical forms.
Slobin (1996, 2004) stresses such classifications of languages have an influence on rhetorical styles and habitual patterns of language use. He puts forward four hypotheses:
A. Speakers of satellite-framed languages consistently show both more frequency of
mention and greater lexical diversity with regard to the manner component of motion
events, as compared to speakers of verb-framed languages.
B. Speakers of satellite-framed languages also typically make more frequent and
elaborated ground description than speakers of verb-framed languages.
C. Speakers of satellite-framed languages tend to break up a scene into several
components and use separate action clauses to describe segment by segment what takes
places in this event, whereas speakers of verb-framed languages tend not to do so.
D. Finally, speakers of satellite-framed languages tend to devote more narrative attention
to the dynamics of movement, while their counterparts of verb-framed languages tend
36
to focus more on the static description of the physical setting in which the action takes
places.
To test Slobin’s hypotheses, Chen (2005) devotes his entire dissertation to investigating the development of motion event descriptions by native Chinese children from four age groups
(Age 3, 4, 5, and 9). Fifty-nine children participated in his experiment to describe the Chinese frog stories, which is based on Mercer Mayer’s classic picture book entitled: Frog, where are you ? The 29-paged picture book depicts a boy who loses his frog and romps through the woods to look for him. Along the way, he makes mischief and new friends. This picture book involves delightfully funny but dynamic actions and it is widely adopted to elicit production of motion events in both first and second language acquisition. The results of Chen’s (2005) study suggest that while Mandarin speakers have access to manner-of-motion verbs, path verbs, and neutral verbal compounds (Motion + Path) to describe motion events, they favor the use of manner-of- motion verbs, and particularly serial verb constructions involving a manner-of-motion verb plus one or more path complements.
His study confirms Slobin’s first hypothesis about satellite-framed languages, but at the same time he indicates a difference between Mandarin and English speakers (or Spanish, in this case) in that Mandarin Chinese encodes both the core information of path and the supporting information of manner of movement in two separate verbs of equal grammatical status with one as the small clause. Regarding Slobin’s second hypothesis about description of ground elements,
Chen (2005) finds that Chinese speakers behave more like speakers of verb-framed languages and do not emphasize ground information. For the third hypothesis on the tendency to mention more narrative segments of a complex motion event, Chinese speakers pattern with speakers of satellite-framed languages. They tend to break up a complex motion event into a large number of
37 components because they are habitually inclined to narrate several path components in a single clause (Chen, 2005; p. 71). Finally, Chinese speakers are like Spanish speakers in that they like to provide some description of the physical settings in which movement takes places.
In addition to the above dual-features of Mandarin Chinese observed in Chen’s study
(2005), Chu (2004) also finds that Mandarin has over 20 types of Path-verbs and these verbs can also serve as directional or resultative particles to appear in serial verbal constructions. Part of his list is included in (2.14):
(2.14)
出 chu ‘exit’/’out’ 上 shang ‘ascend’/ ‘up’ 到 dao ‘arrive’/’up-to’ 沉 chen ‘sink’ / ‘down’ 回 hui ‘return’/’back’ 倒 dao ‘topple’ / ‘over’ 起 qi ‘rise’ / ‘up’ 落 luo ‘fall’ / ‘down’ 下 xia ‘descend’ / ‘down’ 越 yue ‘pass’/ ‘past’ 进 jin ‘enter’/ ‘in’ 凑 cou ‘converge’ / ‘together’ 过 guo ‘pass’ / ‘past’ 没 mo ‘submerge’ / ‘down’ 开 kai ‘open’/ ‘away’ 围 wei ‘surround’
Despite the relatively small number of lexical items counted here, both directional and
resultative verb constructions are used at a very high frequency and salience level in Chinese.
The availability and frequent use of these path verbs makes Mandarin Chinese more similar to
Path-framed languages. Matsumoto (1997) has summarized the path verbs in Japanese, which
turns out to be thirty-three. Levin (1993) counts 20 path verbs existing in English, but a dozen of
them are compound VPs such as ‘go up’, ‘ go out’, ‘go across’, etc. Matsumoto (1997) concludes
that there are fewer path distinctions in any language than manner distinctions.
The evidence and data exhibited in Chu (2004) and Chen’s (2005) studies suggest that
Chinese is an equipollently-framed language and has both satellite-framed lexicalizations and
38 verb-framed lexicalizations in profiling motion events. According to them, this explains why
Mandarin has both features of Manner-in-verb and Manner-in-adverb languages and why
Mandarin differs from English in the projection of directed Manner-of-motion.
Is Chinese really an equipollently-framed language, or does this result simply from the fact that Mandarin displays dual typological features of classical Chinese and modern Chinese?
Peyraube (2006) has made a diachronic study of the directional complements, or path verbs in
Chinese, and found that Chinese underwent, some ten centuries ago, a typological shift from a verb-framed language to a satellite-framed language. Not surprisingly, the resultative, directional and deictic particles are all closely related to each other and have gone through a process of grammaticalization. Tai (2003) believes that just like the grammaticalization of resultative complements and directional complements in Chinese, path-verbs like shang ‘move-(up)-to’, xia
‘descend’, hui ‘move-back’, chu ‘move-out’, etc. have become grammaticalized to path- complements through frequent use by Mandarin native speakers.
2.4. Questions Needed to be Answered To define Chinese as a satellite-framed language, equipollently-framed language, or a language in the process of grammaticalization from a verb-framed language to a satellite-framed language may help answer some of the questions we had regarding the Chinese-English distinction in licensing manner-of-motion events in Section 2.2. However, it does not provide informative explanations for the following cross-linguistic differences in syntactic structures.
Firstly, when the inventory of lexical items for manner-of-motion verbs and goal- denoting PPs are both available and productive in these two languages, why there are such evident distinctions between the transitive and intransitive use of motion verbs in licensing causative directed manner of motion in Chinese but not in English? In the case of intransitive
39 motion verbs, why must the cause/ accompanied action and the manner of the motion be projected in two separate lexical units and connected with the durative aspect marker ZHE in
(2.7), otherwise, a completely different word must be used. As pointed out in the previous chapter, intransitive manner-of-motion verbs in Mandarin Chinese cannot license the causatives of directed motion in the same way as English. For example, it is grammatical to say sentence
(2.15a) in English, but the same expression in Mandarin Chinese (2.15b) is not acceptable unless we utilize another lexical item song ‘drop off’ or add to the main verb zou ’walk’ another verb pei ‘accompany’ marked with the durative aspect marker ZHE, as shown in (2.15c) and (2.15d) respectively. In (2.15d), even when ZHE is not overtly marked in spoken Chinese, the meaning is implied: most native speakers of Chinese share the intuition that the durative aspect marker
ZHE does not need to be spelled out in this sentence, but is inferred.
(2.15a) John walked Mary (back) home.
(2.15b) * 约翰走玛丽回家。
yuehan zou mali hui jia. John walk Mary back home. ’John walked Mary back home.’ (Intended meaning)
(2.15c) 约翰送玛丽回家。
yuehan song mali hui jia. John drop off Mary back home. ‘ John dropped off/walked Mary (at/back) home.’
(2.15d) 约翰陪(着)玛丽走回了家。
yuehan pei (zhe) mali zou hui le jia.
40
John accompany ZHE Mary walk back LE home. ‘John accompanying Mary walk back home.’
Another piece of evidence is related to the fact that transitive use of manner-of-motion verbs in Chinese are usually overtly marked with the BA construction. Sentence (2.16 a-c) illustrate this point. It is very rare amongst native Chinese speakers to use sentence (2.16b) to express the meaning of the English sentence (2.16a), even though (2.16b) is understandable and corresponds to the English sentence structure perfectly. Instead, native speakers usually use the
Ba construction in such situations i.
(2.16a) John kicked the ball to the wall.
(2.16b) ? 约翰踢球到墙边。
yuehan ti qiu dao qiang bian. John kick ball to wall side. ‘ John kicked the ball to (the side of) the wall.’ (Intended meaning)
(2.16c) 约翰把球踢到了墙边。
yuehan ba qiu ti dao le qiang bian. John BA ball kick to LE wall side. ‘John kicked the ball to (the side of) the wall.’
In short, the typologies of lexicalization patterns for motion verbs, either Talmy’s original dichotomy (1975 and 1985) or Slobin’s trichotomy (1996 and 2004), do not offer an explanation of why the transitivity use of a motion verb would affect the syntactic structure of such sentences.
Additionally, even when two languages share similar lexicalization pattern for motion events,
41 this particular strand of proposals does not explain why certain syntactic measures are highly preferred, if not required, in Chinese, e.g., the BA construction and the durative aspect marker
ZHE , while this type of syntactic requirements does not occur in English. Last but not least, this strand of proposals (Slobin, 1996 and 2004; Chen, 2005) predicts that the lexicalization pattern will determine the salience of the conflation pattern and the way of describing a motion event in one language, but it does not have informative implications for understanding what challenges second language learners may experience when distinctive differences are observed in the syntactic representations between their L1 and L2 in licensing complex motion events. These are the questions I intend to answer with another approach within the Principles and Parameters
Theory (Chomsky, 1975 & 1985). In the next two sections, I will first identify the lexical structure of causative directed manner-of-motion, and then analyze how it is projected in the syntax in both Chinese and English.
2.5. Event Structure of Causative Directed Manner-of-Motion
One central consensus in generative grammar research on the connection between syntax and the lexicon is that the realization of argument structure is based on event structure. In other words, syntax is a projection of the lexicon. An event can be examined from the internal structure of the situation, i.e. the aspectuality/(situation) aspects, or externally from the temporal perspective, i.e. tense. Vendler (1967) proposed a four-way classification of events into states
(e.g. know and think ), activities (e.g. walk and swim ), accomplishments (e.g. draw a circle ) and achievements (e.g. die and reach ) according to the aspectual properties of verbs. In this thesis, eventuality is used as the cover term for all these event types. Achievements and accomplishments are telic, inherently bounded and entail a change of state, while activities and
42 states are atelic, unbounded and embrace no terminal endpoint. The properties of these event types are decomposed into features below (cf. J. Lin, 2004; Chang, 2001; Anderson, 1990; Smith,
1991):
• Activities: [-telic, +durative, +dynamic] • States: [-telic, -durative, -dynamic] • Accomplishments: [+telic, +durative, +dynamic] • Achievements: [+telic, -durative, +dynamic]
It deserves noticing that these four Vendler’s event classifications may shift from one category to another depending on the situation context. For instance, the eventuality classification may change in accordance with such factors as the addition of an object, the way construction, fake reflexives, etc. (Rosen, 1999; Tenny, 1994; Chang, 2001). J. Lin (2004) emphasizes that this event classification does not apply to an individual verb, but rather to the entire verb phrase or even the entire utterance. To illustrate this point, if we can add a Goal-PP to the motion construction in (2.17) and a resultative to (2.18), a similar change happens to the event types: the activity turns into an accomplishment. Folli and Ramchand (2005) suggest that the resultatives and goal of motion are much alike in the sense that they add a ‘telos’ to the otherwise unbounded activity event, which consequently turns into a telic accomplishment. This explains why Directed Motion with Goal-PP shares many common properties with Resultative
Verb Constructions (RVCs).
(2.17) Goal-PP a. Activities John walked for an hour/ * in an hour.
b. Accomplishments John walked to school *for an hour/ in an hour. (2.18) Resultative
43
a. Activities John hammered the metal for an hour/ * in an hour.
b. Accomplishments John hammered the metal flat *for an hour/ in an hour. Vendler’s event types have laid the foundation for the lexical semantic models and theories of verbal argument structures. Dowty (1979) decomposes Vendler’s event types with primitive semantic elements DO, CAUSE, and BECOME exhibited below (Dowty, 1979; J. J.
Lin, 2004). In what later becomes a standard analysis adopted by later linguists, Dowty breaks causative sentences down into two sub-events: a causing sub-event in the first verbal phrase (V1) and a result sub-event in the second VP (V2). According to Dowty’s theory, the four event structures can be decomposed and represented as follows:
• Activities: DO (α1,[ πn (α1… αn)]) • States: πn (α1… αn) • Achievements: BECOME[ πn (α1… αn)] • Accomplishments: [[DO ( α1,[ πn (α1… αn)])] CAUSE[BECOME[ πn (α1… αn)])]
More recently, Rapport Hovav and Levin (1998 – hereafter RHL) also proposed a similar
theory of event templates to map arguments in the lexical semantic representation to syntactic
arguments with a decomposition of Vendler’s event classifications. It replaces DO in Dowty’s
model to ACT with either a MANNER or an INSTRUMENT and it further categorizes
accomplishments into two kinds.
• Activities: [ x ACT
44
RHL’s linking template is lexicalist in nature, i.e. verbs directly lexicalize or project complex event structures. RHL propose a process named Template Augmentation to explicate the semantically complex events. For example, the resultative form of a sweeping event (e.g. John swept the floor clean . [[John ACT
accomplishment by adding another sub-event, i.e. the floor becomes clean [BECOME [floor
templates. This important feature of RHL’s theory formulates the fundamental lexical approach
distinctive from syntactic approaches (Grimshaw, 1993; Hale & Keyser, 1991, 1993, 1999 &
2002; J. Lin, 2004; Huang , 1997;T. Lin, 2001).
Different from RHL’s lexical approach, Jimmy Lin (2004) posits that event structure in
Mandarin Chinese is composed purely syntactically from a particular set of primitives following
syntactic principles such as control and movement (p.163). He rejects the theory of linking
between the parallel representation architecture by pointing out that assuming linking rules
required in the licensing of argument structure itself needs to be independently motivated. His
major argument against aligning the argument structure with aspectually-defined event structures
is that Vendler’s classification of events does not apply to verbs in isolation. As a matter of fact,
Lin argues some factors such as the count/mass noun distinction of incremental themes in (2.19),
the real-world knowledge in (2.20), and the availability of a goal-PP in a motion event in (2.21),
can exert some effect on the aspectual interpretation of the sentence.
2.19. a. John ate the apple in an hour/ *for an hour. b. John ate pudding *in an hour/ for an hour. (J. Lin, 2004, p.22)
2.20. a. John ate pudding. (activity) b. John ate the apple. (accomplishment)
45
c. John ate the grape. (achievement) (J. Lin, 2004, p.22)
2.21. a. John swam. (activity) b. John swam across the pool. (accomplishment) (J. Lin, 2004, p.23) J. Lin (2004) incorporates the work of Dowty (1979), RHL (1998), Larson’s VP-shell analysis (1998), and Hale and Keyser’s (1993) syntactic approach into his proposal of argument structure in Chinese and English. J. Lin proposes that Mandarin Chinese only has two primitive verbal categories: activity and state. States are further classified into static events and inchoatives.
The decompositional structures of events are shown below. In J. Lin’s model, aspectual projections are put one layer above the syntactic representation of functional elements for the argument structure, which is not outlined in the diagrams of tree structures below.
VBE [- dynamic] = BE (states)
Vδ [+dynamic, + inchoative] = BECOME (change of state/ inchoative)
VDO [+dynamic, -inchoative] = DO (activities)
The V BE lexicalizes static situation and is only compatible with verbal roots denoting states, like John is tall in example (2.22).
2.22. John is tall. (J. Lin, 2004, p.37)
When a state event is embedded under the V δ, an inchoative event arises. The V δP is
assumed to be embedded under a TP. The specifier of V δ is the entity ( window ) that undergoes a
46
change to the state defined in the inner V BE P ( break ). English unaccusative verbs usually follow the same structure as break in (2.23).
2.23. The window broke. (J. Lin, 2004, p.37)
A V δP can be optionally incorporated into a V DO and that gives rise to the causative
sentence. The causative/inchoative alternation in English can be represented similarly as the
example of break exhibited in (2.24).
2.24. John broke the window. (J. Lin, 2004, p.38)
47
In J. Lin’s (2004) syntactic framework for event and argument structures in Mandarin
Chinese, since activity and state are the only two primitive verbal types, there are no simple mono-morphemic accomplishments or achievements in Chinese. The verbal meaning in Chinese is completely compositional and is “built up” adding one element to another. For example, achievements are syntactically derived from underlying stative roots by adding an inchoativizing element to states, whereas accomplishments are further derived from achievements by adding a causativizing element (J. Lin, 2004). J. Lin’s syntactic proposal for Chinese argument structure hinges upon two key elements: compositional semantics and overt syntactic process, i.e. head movement and control in a Minimalist Program framework. The composition of complex events in Mandarin is licensed and realized in transparent syntactic derivation.
In summary of the various frameworks on event structure, I tend to agree with J. Lin that
Chinese only has states and activities. It is evident that the causative directed manner-of-motion event in discussion here is undoubtedly an accomplishment in eventuality. Following compositional semantics principles, this complex motion event consists of several sub-events: the causative event, and the manner-of-motion event as activities, and the direction denoting the change of location, or the goal of the motion. The event structure of causative directed manner- of-motion is presented in (2.25).
2.25. Event Structure of Causative Directed Manner-of-motion: [x DO < MANNER >]CAUSE [BECOME [ Y < PLACE >]]] In spite of my general endorsement of J. Lin’s proposal (2004) for event structures in
Chinese, I disagree with J. Lin (2004) that the composition of such a complex motion event is realized in syntax directly. Unlike J. Lin (2004), who claims that the event structures are realized syntactically in Chinese, I adopt T. Lin (2001) and Huang’s (1997, 2006) proposal of light verb
48 analysis of Chinese, discussed in length and details in the next section. The composition and representation of the complex motion events is realized in S-Syntax via the function of light verbs. The observed cross-linguistic variations in the syntactic representations of causative directed motion between Chinese and English are therefore resulted from a parametric difference in this regard between the two languages. I propose that the Chinese-English variations in licensing causative directed manner-of-motion lie in the distinctive lexicalization of each sub- event, namely, the realization of DO< MANNER >, CAUSE, and GOAL or [BECOME [ Y <
PLACE >].
2.6. Light Verb Syntax of Chinese
2.6.1. Characteristics of Light Verbs
The term light verb refers to a particular class of verbal items that systematically
contributes certain properties of lexical meaning to a complex predicate and plays some syntactic
functions in the structure. It is a very interesting linguistic notion widely adopted in the current
generative framework. As cited in Butt and Geuder (2001, p.323), Jespersen (1965) first coined
the term light verb for English V+NP constructions like have a rest , give a shout, take a walk , etc.
The main verb have , give , or take in these constructions is semantically light as the nouns rest ,
shout , walk define the fundamental meaning of the VP instead, while the functional role of these
verbs is more prominent – to formulate a VP following the standard VP schema in English. In
general, the light verb in the complex predicates is semantically light but functionally heavy,
which affects the eventualities of the joint predication. It may add “boundedness”, “benefactive
49 readings”, “forcefulness”, “suddenness” “inception”, “volition” or other subtle modifications to the event (Butt, 2004; T. Lin, 2001).
Light verb constructions can be V+N or V+V structures. This study predominantly concentrates on the analysis of V+V light verb constructions in the manner-of-motion events.
There is tremendous literature on light verbs applied to V+V constructions in various languages, for example, Grimshaw and Mester’s (1988) study of the suru construction in Japanese, Rosen’s
(1989) analysis of Romance periphrastic causatives with make , Butt’s work on Urdu V+V complex predicates (Butt, 1995, 2004 & 2010), Hale and Keyser’s (henceforth HK; 1991, 1993) investigation of the derivation for denominal verbs like shelve, laugh , jump , etc.. The light verb or complex predicate structure is also well studied in subsequent Chinese linguistic works, including but not limited to, small-clause analysis of resultative verb constructions with ba and de (Huang, 1992), directional phrases (Scott, 1995; Butt & Scott, 2002), phrase structures (T. Lin,
2001), resultatives verb constructions (Huang, 2006).
There are different explanations about the nature of light verbs , which lead to distinctive approaches. Light verbs have been treated as a semantically empty predicate-licenser which serves as a functional element. Grimshaw and Mester (1988) find in their study that the inherent property of the noun which suru is compounded with determines the transitivity of the verbal expression. In their analysis, suru is semantically light and plays only a functional role. However,
Matsumoto’s later work (1996) proves that suru ‘do’ at the very least contributes a selectional restriction with respect to agentivity.
In HK’s (1991, 1993) and Chomsky’s recent work (1995, 1999), light verbs are also treated as empty place holders with barely minimum semantics in English verbal structures. The
50 light verb is the instantiation of little v, first introduced by Chomsky (1957) for auxiliaries and modals. In the Minimalist Program, v can be interpreted as a lexical, a functional, or a mixture- of-both category. In Chomsky’s recent works, the light verb serves as the head of a transitive predicate, which has phi-features and needs to be checked with the object in object-shifted languages such as Icelandic.
In another approach, light verbs are treated as one class of syntactic categories that are neither fully lexical nor fully functional, which is labeled as “semi-lexical heads” by Corver and
Van Riemsdijk (2001, p.9 & p.16). Semi-lexical categories behave quite distinctively from content categories/words or function categories/words. Content words are open classes of lexical items that have a relatively ‘specific or detailed’ semantic meaning, for example, objects (N), events (N), or locations/directions (P). Conversely, function words refer to the fixed roster of items that usually carry non-conceptual meaning, but fulfill certain ‘grammatical’ functions, such as tense/aspect marker, modality indicator, etc. (Corver & Van Riemsdijk, 2001, p.1)
Content words differ from function words in many aspects. First, the selection of a lexical category by a lexical head is top down, for instance, a placement verb selects certain types of objects but not others. On the other hand, the selection of a functional head is bottom up; for example, the gender form of a determiner is selected by the noun it determines. Second, lexical categories enter into theta marking, but functional categories do not. Additionally, the two differ in movement/displacement. Content words can license empty categories in their complement position (2.26c), but function words cannot (2.26b). Corver and Van Riemsdijk’s examples
(2001, p.2) are provided below.
2.26. a. I don’t believe [ CP that [ IP Mary hates soccer]]. b. *[ IP Mary hates soccer] i I don’t believe [ CP that t i].
51
c. [ CP That Mary hates soccer] i I don’t believe t i. Semi-lexical heads , like light verbs, display dual properties of lexical categories and functional categories. In complex predicates (full verb plus light verb), the light verbs can pattern with fully lexical items, but not with functional heads, so they are less “functional” than auxiliaries; on the other hand, light verbs may be lexically defective to various degrees, but they are not completely empty elements. The light verb participating in complex predicates contributes certain semantic meaning to the event description provided by the full verb.
Prepositions are another example of semi-lexical heads . According to Van Riesmsdijk’s works
cited in Corver and Van Riesmsdijk (2001), prepositions share both lexical and functional
characteristics – they do not really constitute an open class and are less “functional” than
determiners, and more “grammatical” than nouns, verbs and adjectives (p.4). This approach
partly explains the cross-linguistic differences in light verbs: in some languages, the lexical
properties of light verbs are more salient than the functional properties, and in other languages it
is the other way around.
Butt and Geuder (2001) adopt this approach to analyze the complex predicates in Urdu.
They argue that light verbs are not completely empty elements despite the lexical defectiveness.
Instead, a light verb conceptualizes certain lexical meaning in the VP that participates in
licensing the event structure. They emphasize that the light verb use and the full verbal function
of the same item should be a case of “lexical polysemy” instead of “grammaticalization” (p. 326).
Light verbs are syntactically verbs (of category V) and can only be licensed in
conjunction with a main verb, but at the same time, they also contribute additional semantic
features to the event description (Butt and Geuder, 2001, p. 358). The differences between main
52 verbs, light verbs, and auxiliaries are summarized in the following Table (Butt and Geuder, 2001, p. 365).
Table 2.1. Differences between Main Verbs, Light Verbs and Auxiliaries
Argument Structure Event Category Main Verb Full a-structure Event description V Light Verb Incomplete a-structure Modifies event V Auxiliary No a-structure Situates event AUX or I (e.g. Reichenbachian E,R,S)
2.6.2. Light Verbs in Chinese and English
Now that we are aware of the semi-lexical and semi-functional properties of light verbs, the next question is how light verbs are used in Chinese and English. According to Huang’s
(1997) hypothesis regarding lexical structures and syntactic projections in Chinese, light verbs in
Chinese provide evidence for a lexical decomposition approach to lexical semantics. The abstract light verb stays above an event-denoting predicate in Chinese and the event argument is the complement of the empty eventuality predicate DO, CAUSE, BECOME/OCCUR or BE/HOLD
(see examples in 2.27a-f). T. Lin (2001) extends Huang’s proposal to study the cross-linguistic variation between Chinese and English. He believes that Chinese and English are at the two ends of a continuum for the typology of light verbs – the arguments in Chinese are licensed by light verbs and merged to the phrase structures via their functions in S-Syntax; arguments in English are licensed at the sub-syntactic level of grammar, or L-Syntax in Hale and Keyser’s terms (1991,
1993). T.Lin (2001) believes that light verbs are predicates of aspects of eventualities.
Syntactically, they are verbs, with or without phonetic realizations; semantically, they are predicates of aspects that compose eventualities (T. Lin, 2001).
53
2.27. a. Activity: ku ‘cry’
b. Activity: kan-shu ‘read books’
c. Inchoative: pang ‘fat’
d. Stative: pang ‘fat’
e. Stative: xi-huan ‘like’
f. Causative: qi-si ‘to anger someone to death’
54
The light verb analysis is based on lexical decomposition of event structures. A number of theories have been proposed to decipher lexical decomposition, Jackendoff’s (1990) Lexical
Conceptual Structure (LCS), and Hale and Keyser (1990)’s Lexical Relational Structure (LRS), amongst others. Hale and Keyser (1990) argue that lexicalization/conflation (Talmy, 1985; Fodor and Lepore, 1999) is syntactic in nature. LRS suggests that the argument structure of a verb does not only comprise a list of argument positions in an order designated by the Thematic Hierarchy, i.e. the argument serving as the Agent will appear in the highest argument position of any sentence at D-Structure, and the Experiencer will appear in higher position than Goal, Theme, and Obliques will stay in the lowest position. But more importantly, the argument positions are actually placed in a “lexical syntactic structure” defined by the verb’s LCS. That means in a composite event, like the complex motion event in discussion here, while two or more eventualities (a causing, a manner of motion, and a change of location) are involved, the causing event is more prominent and asymmetrically c-commands the becoming event, which is in turn in a higher position than the change of location.
The example (2.28) illustrates HK’s (1990) proposal for causative predicate. The NP in the specific position of the higher VP is the agent of the causative event, while the theme is placed in the inner subject of a complex predicate, i.e. the specifier position of the lower VP in the structure. Following this line of thinking, the higher VP denotes the causing event, John
55 functions as the causer of this event, and the lower VP indicates change of state or location. Here
V ( put ) maps to the event e, and P ( on ) maps to a location or interrelation r. This leads to the lower VP in (2.28) reads as an event e implicates an interrelation r, e r. The NP in the lower
VP Spec, a book , is the subject that brings about the change of state, n> (e r). It is also the theme of the whole event. Therefore, the complete event structure for (2.28) is e1 n> (e2 r).
2.28. a. John put a book on the shelf. b.
Unergatives such as walk, run , etc. are known as “truly intransitive”, Example (2.29) demonstrates the derivation for the unergative verb, walk , where the head N of the NP complement incorporates to V. Similar derivation applies to the deadjectivals in (2.30), where A is conflated to the higher V.
56
2.29. unergative / denominal
2.30. unaccusative / deadjectival
In short, HK’s theory is to decompose and represent event structures in syntactic structures. Light verbs serve as the empty “VP shells” or building blocks for LRS to represent the thematic relations. Larson’s (1988) “VP shells” bear two major features: first, there can be more than one VP projection in the syntactic structure of a predicate and each VP projection is binary; second, the distribution of arguments in the syntactic structure is determined by the universal thematic hierarchy (Jackendoff, 1972; Grimshaw, 1990).
On the foundation of Hale and Keyser’s work (1991 , 1993), Huang (1997) and T. Lin
(2001) conduct a crosslinguistic analysis which shows that Chinese, Japanese and English form a continuum in terms of the lexicalization of the light verb structure: in English, the licensing of
57 the arguments applies at the L-Syntax; in Japanese, the licensing of the subject argument takes place in S-Syntax and other arguments in L-Syntax; in Chinese, the whole light verb structure is manifested and spelled out in S-Syntax. For instance, the derivation from the noun to the verb phone only happens in English lexicon but not in Chinese. The Chinese noun dianhua ‘phone’ cannot be denominalized directly into a verb. An overt light verb da ‘hit/ do…with hand’ must be encoded to the verbal structure.
Basically, the different realization of event structures between Chinese and English is a result of distinction in the Lexicalization Parameter (T. Lin, 2001, p.9):
…languages differ in the phrase structural height in a light verb structure at which lexicalization applies – in English, the whole light verb undergoes lexicalization, hence the verb in English contains rich internal eventuality information and a full-fledged set of arguments; …in Chinese, lexicalization trivially applies to the main verb, leaving the rest of the light verb structure intact, thus the whole light verb structure is sent to the syntactic representation. Chinese is thus identified as a Davidsonian language (Huang, 1997; T. Lin, 2001), since the constituting components of the phrase structures in Chinese display a one-to-one correspondence to eventuality predicates in the logical structure. All activity verbs (unergative and transitive) are complements to a predicate akin to DO; inchoative predicates are embedded under BECOME or OCCUR; statives are embedded under BE or HOLD; and causatives are embedded under CAUSE. The arguments in Chinese are licensed by light verbs and merged to the phrase structure via their respective functions.
2.7. Licensing Causative Directed Motion with Light Verbs
Now that the Lexicalization Parameter explains the Chinese-English distinctions in the representation of eventualities, referring back to our study, what is its implication in this study?
58
How is the event structure of causative directed manner-of-motion, illustrated in (2.25) and repeated below as in (2.31), realized in Chinese and English respectively? Will this parametric difference provide an answer to the Chinese-English distinctions in licensing causative directed manner-of-motion?
2.31. Event Structure of Causative Directed Manner-of-motion: [x DO < MANNER >]CAUSE [BECOME [ Y < PLACE >]]] Based on the Lexicalization Parameter, I propose that each sub-event of the causative
directed manner-of-motion is realized via one light verb in Chinese and overtly marked in PF;
however, the composition and licensing of these same sub-events are completed in LF of English
and is covertly marked in PF. As a result, the causativization of both transitive and unergative
manner-of-motion verbs are available in English syntax given that these verbs satisfy the lexical-
semantic restrictions, such as “accompanying” and “agentivity” interpretations (I will come back
to this in the next section), because these verbs go through a full conflation in the lexicon and
contain the internal eventuality information when entering syntax.
On the other hand, only the transitive manner-of-motion verbs in Chinese undergo
causativization and the causatives are projected in Chinese syntax with the use of light verb ba
(to be discussed in Section 2.7.1). The unergative manner-of-motion verbs, however, cannot
participate in causativization per se and all the lexical-semantic meanings involved are spelled out in PF. Therefore, a complex manner is formulated with unergative manner-of-motion verbs by the composition of an “accompanying” component with the “manner” of motion. The representation of the “accompanying action” emerges in the syntax with overt marking by the use of durative marker or light verb ZHE (to be discussed in Section 2.7.1). This answers one of
59 the earlier questions regarding the distinctive behaviors of transitive versus unergative manner- of-motion verbs in the two languages.
Additionally, the same parametric difference also explains the crosslinguistic variation in the realization of the goal ( [BECOME [ Y < PLACE >]]) between the two languages. While both the locative and directional meanings are incorporated in the LF of English prepositional phrases, the directional PPs in Chinese are syntactically complex and they are usually overtly marked with the directional preposition or light verb DAO and the localizer or place words (to be discussed in Section 2.7.2).
2.7.1. Light Verb ba for CAUSE and zhe for ACCOMPANY
In this thesis, some causatives 2 bear a direct causation, i.e. an event or situation A is said
to cause an event B if A is such that if A did not occur, B would not occur (Naoko, 2006). One
example of direct causation is John kicked the ball to the wall , because the event of John kicking
the ball causes the occurrence of the ball rolls or moves to the wall . Other causatives in the study may bear an indirect causation, for example, the equivalent translation of the English sentence
John walked Mary to the car into Chinese is realized by composition of a manner verb pei (zhe )
‘accompanying’ with an intransitive verb zou ‘walk’. Therefore, John pei(zhe) Mary zou dao/hui jia is an indirect causation. The definition of causative sentences in this thesis follows the interpretation of the English sentence. In other words, my interest in causatives in this study only concerns how the English causatives of directed motion are realized in Chinese. It is neither my
2 I should like to thank Dr. Feng-hsi Liu and Dr. Rudolph Troike for interesting discussions on the definition and availability of causatives in both Chinese and English. It is noted that some semantic features like agentivity and animacy affect the acceptability of causatives, for example, Walk an animal differs from walk a bicycle . Language variations have an impact on whether the causative in one language is causative in another. For example, to understand Marry carried the baby across the street as Mary caused the baby move across the street by carrying him may be acceptable in one language, but may sound tricky in another. Furthermore, individual judgment also influence whether the derivation a lexical causative is acceptable. For example, some linguists may find it makes perfect sense that “ kill ” is to “ cause to die ”, while others may disagree (Fodor, 1970).
60 intention nor the focus of the present study to provide a full account of causative constructions here. However, it is admitted that some semantic restrictions, language variations and individual judgment differences have significant impact on the derivation and interpretation of causatives.
Causative constructions and resultative constructions have been well studied in various frameworks, including Simpson’s (1983) work in the Lexical Functional Grammar, Levin and
Rapport’s (1988) proposal in the Lexical Conceptual Structure, Carrier and Randall (1992) in the
Government and Binding theory, and Kratzer (2004) in the Minimalist framework. The two constructions are correlated not only because of the frequent co-occurrences across languages, but also because the resultative construction contains a lexical causative predicate CAUSE and hence resultative constructions are causatives by nature (Sybesma, 1999; Cheng & Huang, 1995;
Tomioka, 2001).
Causative constructions in Chinese have also been much-discussed by numerous linguists in association with resultatives (Gu, 1992; Sybesma, 1992, 1999; Zou, 1994; etc.). For example,
Li (1990, 1991, 1995, 1998, 1999) has carried out multiple studies to analyze causatives and resultatives in Mandarin Chinese. He postulates V1 as the head of a resultative construction and approaches the issue from the perspective of theta role assignment (1990, 1998), the c-role assignment (1995), and the interaction of the two with the Direct Object Restriction (cf. Levin and Rappaport Hovav, 1995), the Thematic Hierarchy (cf. Grimshaw, 1990; Larson, 1988), and
Uniformity of Theta Assignment Hypothesis (cf. Baker, 1987). Following Li’s proposal of V1 as the head of resultatives, Cheng and Huang (1995) also adopt a syntactic approach and treat the argument structure of resultatives from the point of view of verb classes. They suggest that the aspectual property, instead of the transitivity property, of an argument structure determines the composition and properties of resultative compounds. Recently, Huang (2006) adopts a
61 parametric view to investigate why unergative objectless resultatives are available in Chinese, and why both unaccusative and unergative resultatives can be causativized in Chinese. He proposes that Chinese as a highly analytical language has the ability to merge an unergative verb as a manner modifier to an inchoative eventuality predicate BECOME, while English lacks this ability.
Given the similarities and correlations between causative constructions and resultative constructions, between resultatives and directional PPs, how can we explain the causativization with both unergative and unaccusative verbs in Chinese resultative constructions (Huang, 2006) on one hand, and the unavailability of causativization with unergative verbs in the directed motion? In other words, if resultatives are causative in nature and share a lot in common with goal PPs in terms of semantic and syntactic properties, what gives rise to the inconsistent availability of causatives for unergative and unaccusative verbs in Chinese resultatives and directed motions?
See below Huang’s example in (2.32a) and (2.32 b) for the causativization of an unergative verb in a resultative construction. I have provided additional two examples in (2.32 c) and (2.32 d) accordingly. (2.33) gives Folli and Harley’s (2007) English example of the causativization of an unergative verb in a directed motion event, and my complementary Chinese example in (2.34).
2.32. a. 他跳得满头大汗。 ta tiao-de man-shen-da-han. He dance-till whole-body-big sweat ‘He danced [himself] all sweaty.’
b. 一只探戈跳得他满头大汗。 yi-zhi tanggewu tiao-de ta man-shen-da-han one-CL tango dance-till he whole-body-big-sweat ‘A tango dance caused him to dance himself all sweaty.’
62
c. 一只探戈舞把他跳得满头大汗。 yi-zhi tanggewu ba ta tiao-de man-shen-da-han one-CL tango BA he dance-till whole-body-big-sweat ‘A tango dance caused him to dance himself all sweaty.’
d. ? A tango danced him all sweaty.
2.33. a. John waltzed (*Matilda). b. John waltzed Matilda into the bedroom.
2.34. a. John 跳了华尔兹(*Matilda) 。 John tiao le hua-er-zi (*Matilda) John dance PERF waltz (*Matilda) ‘John waltzed (*Matilda).’
b. John 陪着 Matilda 跳了华尔兹。 John pei zhe Matilda tiao le hua-er-zi John accompany ZHE Matilda dance PERF waltz ‘John waltzed (*Matilda).’
c. *John 把 Matilda 跳华尔兹到了卧室。 John ba Matilda tiao hua-er-zi dao le wo-shi li John BA Matilda dance waltz DAO PERF bedroom inside Intended meaning: ‘John waltzed Matilda into the bedroom.’
d. John 陪着 Matilda 跳华尔兹到卧室。 John pei zhe Matilda tiao hua-er-zi dao wo-shi li John accompany ZHE Matilda dance waltz DAO bedroom inside ‘John accompanying Matilda waltze to the bedroom.’
The examples (2.32) - (2.34) seem to be contradictory to each other. On the surface structure, examples in (2.32 (a)-(c)) show the availability of causative construction for the unergative verb tiao ‘ dance ’ with or without the use of ba construction in Chinese, and the unavailability of the same causativization in English (2.32 (d)). However, examples in (2.34) show that the opposite takes place in directed motion. (2.34) indicates that despite the availability in English (2.33), the same unergative verb tiao ‘ dance ’ cannot be causativized in the directed motion with or without the use of ba construction in Chinese. Instead, a complex manner is
63 composed (as in (2.34b) and (2.34 d)) with the durative marker zhe to indicate a continuous, stable, cotemporaneous status of the subject.
Before I move on to examine the semantic properties of the two sets of causatives, I would like to have a brief discussion of the two structures involved here: one is the ba construction and the other is the durative aspect marker zhe . The ba construction is the most studied structure in Chinese linguistics and it is a complicated structure with complex semantic, syntactic and functional properties. It comes under a number of labels, for example, it is a
‘disposal construction’ (Chao, 1968; Li & Thompson, 1981), an ‘executive construction’, and an
‘accusative construction’ (Sybesma,1992). Semantically, it involves a high degree of affectedness (Tenny, 1987) and transitivity. Transitivity is defined as “the carrying over of an activity from an agent to a patient” (Thompson, 1973). Aspectually, the ba NP denotes boundedness and specificity (Liu, 1997). Functionally, it relates to the topic of the discourse as suggested in F. Liu (2007) that the use of ba is more likely to occur when the ba NP denotes old information but is not highly topical, and when it carries new but relatively heavy information. In the present study, the ‘causative’ property of the ba construction is prominent. Sybesma (1992) holds that all ba -sentences are causatives and T. Lin (2001) also proposes that ba is the light verb for causatives as discussed earlier. Following these proposals, I suggest that ba functions as the light verb here by occupying the CAUSE in eventuality and licenses the causative sub-event in the directed motion.
The second particle, zhe, is defined as a progressive marker or a durative aspect marker
(Li and Thompson, 1981), indicating that the situation is enduring or continuing. It also signifies a state which is the result of an action (Smith, 1991; Sybesma, 1992): for instance, the resulting state of picking up a book is holding it, and that of putting on a jacket is wearing it. Normally,
64 zhe does not occur with resultatives (Klein et al, 2000), when it does co-occur with resultatives, it is only available to the stage statives (Yeh, 1993). T. Lin (2001) defines zhe as one of the commonly-used light verbs in Chinese, but he only acknowledges light verb zhe as a durative aspectual marker like other linguists mentioned above. In the present study, however, I find the aspectual interpretation of the light verb zhe is not salient. I propose that the light verb zhe does not necessarily need to entail aspectual information; instead, it can simply license the accompanied sub-event of a directed motion.
Now that we have a better understanding of the semantic, syntactic and functional properties associated with the light verb ba and the light verb zhe , I would like to go back to the analysis of the two sets of causatives displayed in examples (2.32-2.34). A further close examination suggests that the causer of the directed motion entails such semantic meanings as
[+agentive][+intentional][+accompany], while the causer of the resultative construction is [- agentive] [-intentional] [-accompany]. It turns out that the combinations of these semantic factors determine whether a causative can be formed in English. Consider the classifications of verbs proposed by Folli and Harley (2007) in attempt to explicate when a causative of directed motion can be licensed (See Table 2.2).
Table 2.2 Causatives of the Four Classes of Motion Verbs Status of Causer Argument
Verb Class -intentional, +intentional, +intentional, +accompany -accompany +accompany -Path, -Agent
(shudder ) * * * -Path, +Agent
(whistle ) * * +Path, +Agent
(walk ) * * +Path, -Agent
(roll )
65
As Table 2.2 (Folli & Harley, 2007) demonstrates, when a verb specifies neither an
Agent nor a Path, a causative cannot be formulated (* John shuddered the cart across the hallway ). When a verb specifies just an Agent, but not a Path, a causative may be formed with an intentional subject but definitely not with a non-intentional one. No cotemporaneous requirement is commanded ( Mary/ * The teakettle whistled the dog down the hill ). Verbs like walk require the existence of both an Agent and a Path, therefore, the causative is only available when the subject is intentional and there is an accompanied-action requirement (Mary walked the bike to the restaurant ). Roll -verbs implies a Path but not an Agent, and causatives are available with or without intentional subjects.
In the case of Chinese, unergative motion verbs simply cannot be directly causativized.
Chinese and English differ the following two aspects:
• The causative directed motion is subject to lexical-semantic requirements as demonstrated in
Table 2.2. When these semantic restrictions are met, the causatives of unergative verbs are
available in English, but they are not directly available in Chinese. Instead, a complex
manner is licensed with the durative marker zhe to denote the semantic meaning [accompany]
involved in the event.
• The resultatives are not subject to additional lexical-semantic requirements for the verb or the
causer. Both unergative and unaccusative verbs in Chinese can participate in causativization,
but only unaccusative verbs in English can be causativized in resultatives.
These differences provide another piece of evidence that English motion verbs can undergo full conflation and pick all the features at the end of the lexical computation and enter syntax with the corresponding structures (Huang, 2006). On the contrary, Chinese unergative manner-of-
66 motion verbs do not go through conflation as a lexical operation. They simply enter into syntactic computation with only a conceptual structure. The lexical-semantic restrictions that are required in the conceptual structure are licensed the argument structure with the appropriate light verbs.
2.7.2. Light Verb dao for the GOAL
The goal of a directed motion event indicates the endpoint of such motion and it is
usually signified by prepositional phrases. However, language variations and different strategies
are observed in how prepositional phrases are employed to denote located motion and directed
motion. For example, the choice of case of the complement of the PP in some languages (e.g.
Latin, Russian, German) determines whether it is located motion or directed motion, while in
some other languages (e.g. English), ambiguity exists in whether the PP indicates located or
directed motion. In Roman languages (e.g. Spanish), suggested by Talmy (1975), only located
motion is entailed by the combination of a manner of motion verb and a PP; therefore, directed
motion requires that the telic path verb inherently expresses the end point of motion with the
manner of the motion serving as an adjunct (Folli & Ramchand, 2005).
In Chinese, prepositions marking the located and directed motions have different
semantic requirements and syntactic distribution. Pre-verbal PPs usually indicate the general
location of an activity, while the post-verbal PPs denote the endpoint or result of an action.
Additionally, a localizer (Zhao, 1968) or locative particle (Li & Thompson, 1989) is usually
required to specify the orientation. This section only provides a brief account of the PPs involved
in the main research questions. For a comprehensive discussion of co-verbs or PPs in Chinese,
readers are referred to Tai (1975), Zhu (1981), Fan (1982), Wu (1996), Troike & Pan (1995),
and F. Liu (2009), among others.
67
The English and Chinese examples for located and directed motions are provided below in (2.35) and (2.36). Illustrated with the contrast between (2.35a) and (2.35b), the preposition is
‘obligatorily dynamic ’ (Folli & Ramchand, 2005) and provides an endpoint to the directed motion. Example (2.36a) and (2.36b), however, demonstrate that the preposition involved is ambiguous and it can be either dynamic or stative. The Chinese examples in (c) and (d) display the distinctive syntactic distribution in correspondence to the different semantic meanings of located and directed motion.
2.35. a. John ran to the store.
b. *John was to the store.
c. John 跑到了商店(里)。 John pao dao le shang-dian (li) John run DAO PERF store (inside) ‘John ran to the store.’
d. John 在/到商店里跑。 John zai/dao shang-dian li pao John DAO /ZAI store inside run ‘John is running in the store.’
2.36. a. The boat floated under the bridge.
b. The boat was under the bridge.
c. 船浮到了桥底下。 Chuan fu dao le qiao di-xia Boat float DAO PERF bridge underneath ‘The boat floated under the bridge.’
d. 船在桥底下浮着。 Chuan zai qiao di-xia fu zhe Boat ZAI bridge underneath float ZHE
In order to explicate the language variation in the availability and interpretation of Goal
PPs and resultatives, Folli and Ramchand (2005) propose that three event projections in (2.37)
68 are necessary to represent all the possible components of the event structure building processes of all languages.
2.37.
As represented in (2.37), the verb phrase contains three layers of projections and each projection is an instantiation of a possible sub-event. According to Folli and Ramchand (2005):
• vP introduces the causation event and licenses the external arguments;
• VP defines the nature of the process or change and licenses the object of such process or
change;
• RP provides the ‘telos’ or boundary of the event and licenses the object of result.
They employ this projection in the analysis of prepositional phrases involved in directed motion as an attempt to answer why the properties of preposition phrases are subeventally complex, containing both a direction/path (the process) and a final location/place (the result).
They follow Higginbotham (1995, 2000) and define these PPs as accomplishment prepositions with dual event projections entering the syntactic derivations. They label the ‘direction/path’ projection as P, and the ‘final location/ place’ as Rp. The complex structure of their example is shown in (2.38). Following this line of analysis, the accomplishment directional prepositions in
Chinese example is displayed in (2.39) below.
69
2.38.
2.39.
Not surprisingly, the differences between the two languages in licensing the accomplishment prepositions can be attributed to the parametric difference in lexicalization, i.e. the composition of the sub-event ‘direction/path’ and the sub-event ‘final location/place’ complete in the L-Syntax of English by incorporating the stative Rp ‘in/on’ with the dynamic P
‘to’. In Chinese, Rp and P share the same lexical item DAO, therefore the second DAO is dropped to avoid repetition, but the Spec of RpP, Shang-dian ‘store’, and the DP, li-mian ‘inside’ must enter the syntactic projection and spell out in PF.
2.7.3. Questions Re-addressed
70
In this section, I have adopted the parametric approach to interpret the Chinese-English differences in representing causative directed manner-of-motion events with the use of light verbs. The Lexicalization Parameter distinguishes Chinese from English in how the event structure is realized in the syntactic structure, which offers a more coherent and unified explanation why the formation, syntactic structures and distributions of directed manner-of- motion differ in these two languages. It advocates that argument structures of the causative directed motion are formulated at the sub-syntactic level in English while their counterparts are licensed at the syntactic level in Chinese.
One of the unanswered questions in Talmy (1975, 1985) and Slobine’s(1996, 2004) work on manner of motion concerns the frequent occurrence of the ba construction and the durative marker zhe in Mandarin Chinese. The analysis in this section reveals that causative directed
manner-of-motion events are accomplishments in eventuality. These complex accomplishment
events can be further decomposed into three elements: causing (causative), doing (activity), and
becoming (change of location). The English verbs undergo a full conflation and pick up all the
features at the end of the lexical computation, and then go into syntax with all arguments
licensed to formulate the phrase structure. However, the argument structure of causative directed
motion in Chinese must be realized in the syntactic level via the function of several light verbs.
To be specific, the causative event is licensed with the light verb ba, the activity is formulated
with a motion verb, and the achievement (change of location) is represented with the light verb
or coverb dao . The transitivity use of the motion verbs affects the argument structures in Chinese
because Chinese unergative verbs cannot participate in causativization in the case of directed
motion. Instead, a complex manner emerges in the syntactic computation and the light verb zhe is employed to indicate the accompanied action.
71
Employing the same lexical decomposition approach, I have also explained why the goal
PPs in Chinese behave differently from English. On one hand, only the post-verbal prepositional phrases can license the goal of the directed motion and denote the telicity of the accomplishment event. In this case, it is marked with light verb dao . Conversely, the pre-verbal prepositional phrases only indicate the general location where an event takes place. On the other hand, the decomposition of the prepositional phrases also accounts for why the locative particles stay behind the NP after the preposition in Chinese.
Following this line of reasoning, it also becomes clear why the two languages vary in the number and types of manner-of-motion verbs. Chinese resorts to compounding to denote various types of walking by composition of a manner-adverbial with a motion verb, which is not evidently productive in English. Generally speaking, when a motion verb does not designate manner in Chinese, it projects the manner information to the syntax with the use of adverbial phrases or adjuncts. It is a different story with English: most of the commonly used manner of motion verbs, such as run -class and walk -class, undergo the full conflation in English lexical computation and enter the syntactic representation with full-fledged meaning, structure and function. However, their counterparts in Chinese have to exploit the syntactic computation to return similar meaning. As a result, the parametric lexicalization difference gives rise to the differences in number and the type of manner-of-motion verbs that are available in these two languages.
i It is very rare to not use the ba constructions in the given situation, but sometimes children or foreign learners of Mandarin Chinese use this type of sentence pattern. For instance, Erbaugh (1985) find native-speaking children “overextended their highly transitive agent-action-pation prototype” (p.420). Children at the age of 2-3 tend to produce such agentive sentences without using the ba Construction. The following examples are elicited from Erbaugh’s analysis of the
72
children’s errors (p.420-422). The examples in (a) are children’s errors, while (b) illustrate their intended expressions.
1a. * 你上这个。 Ni shang zhe ge. You get on this. ‘(You) put on this.’ (intended meaning)
1b. 你把这个带上。 Ni ba zhe ge da shang. You Ba this CL put on ‘(You) put on this.’
2a. * 蜡笔跑掉了。 Labi pao diao le. Crayon run down LE. ‘The crayon rolled down.’ (intended meaning)
2b. 蜡笔滚下去了。 Labi gun xiaqu le. Crayon roll down LE ‘The crayon rolled down.’
3a. * 我坏了。 Wo huai le. I bad LE. ‘I broke it.’ (intended meaning)
3b. 我把这个弄坏了。 Wo ba zhe ge nong huai le. I BA this CL make broken LE ‘I broke it.’
73
CHAPTER 3
STUDIES ON L2 ACQUISITION OF DIRECTED MOTION
Before we move on to the actual SLA study, this chapter provides a review of SLA theories relevant to the present study and previous studies on L2 acquisition of directed motion.
Three questions are put forward. Firstly, based on the cross-linguistic variations covered in the previous chapter between Chinese and English in licensing causative directed manner-of-motion, what learning problems do we expect English-speaking learners of Chinese to experience?
Secondly, given that the linguistic analysis has revealed that the L1 and the L2 distinctions in licensing causative directed motion originate from the lexicalization parameter of argument structures in these two languages, can English-speaking learners of Chinese access UG to switch to reset the L2 parameter? Lastly, how can previous studies on acquisition of directed motion with goal PPs inform this study?
3.1. Learnability Challenges
The previous chapter was dedicated to investigating the semantic-syntactic interaction of directed motion constructions in both Chinese and English. The cross-linguistic variations between the L1 and the L2 usually cause difficulties in L2 acquisition (White, 2003). In order to fully acquire motion verbs in Chinese, L2 learners must first arrive at a representation for lexical items in the target language and map the representations from argument structure to syntax.
Given that the L1 and L2 realize argument structures parametrically differently, i.e. the arguments in English are licensed at the L-Syntax, while the arguments in Chinese are licensed at the S-Syntax, the acquisition of causative directed motion in Chinese by English-speaking learners is expected to be extremely difficult.
74
The Chinese-English variations in the lexical and syntactic correspondences observed in
Chapter 2 have posed some specific learning challenges for English-speaking learners of Chinese:
Will they acquire the distinction between transitive and intransitive motion verbs in Chinese?
Will they acquire the use of light verbs in representing directed motion in Chinese, specifically,
Ba for CAUSATIVITY, Zhe for ACCOMPANY, and Dao for GOAL? Will they acquire the understanding that the parameter setting determines the distinctive realizations of argument structures between the two languages?
Both overgeneralization and undergeneralization errors are expected in this case.
Overgeneralization errors in second language acquisition usually take place when a learner’s L1 allows for more syntactic representations of a particular argument structure than their L2. In this case, English-speaking learners of Chinese may overgeneralize sentences like John walked Mary to the house , Mary ran the dog to the park , to Chinese. At the same time, the light verbs are not spelled out in English in the same way as those in Chinese; therefore, undergeneralization errors may occur when English-speaking learners fail to fully grasp aspects of the L2 argument structure, i.e. the various light verbs.
3.2. Parameter Switch or Feature Assembly? Within the Minimalist Program framework and the Principle and Parameters approach, it is believed that there is a finite set of parameters that restrict the possible range of syntactic variation across languages (Chomsky, 1981 & 1995). Principles and Parameter theory has significant implications in SLA research for its potential to more precisely characterize the grammatical distinctions between a learner’s L1 and L2. Chomsky’s metaphor for language acquisition below is also adopted in a literal sense.
75
Acquisition of language is in part a process of setting the switches one way or another on the basis of the presented data, a process of fixing the values of the parameters (Chomsky, 1988, p.63) Hence, SLA is determined by a learner’s degree of success in ‘resetting’ one or more parameters from the L1 value to those of L2 (Haegeman, 1988; Radford, 1997). From the perspective of the proponents, the use of parameters in acquisition efforts may help learners discover clusters of language properties associated with one issue and the value setting of one particular parameter may lead to “a cascade of related effects” in enhancing the language learning experience.
However, Lardiere (2008, 2009) has critically reviewed the theory of parameter resetting and the rehabilitation of contrastive analysis in second language acquisition, which has led to a wide-spread debate in the field of second language acquisition research (Champaign, 2009;
Ottawa, 2009; Slabakova, 2009; White, 2009). Lardiere points out that theoretical analyses based on the application of parameters fail to fulfill the promise of the “deductive-consequences conjecture” (p.177). One major reason is that there are disparate types of parameters and there is little consensus on theoretical characterization of what sort of object can be counted (or not) as a possible parameter. One extreme example is Kayne (2005), who identifies all cross-linguistic syntactic differences as microparameters, whether they are associated with any cluster of related effects or not. He suggests that many parameters are too ‘coarsely characterized’ and need to be refined. Using Kayne’s contrastive example of degree adverb enough and assez in English and
French, however, Lardiere (2009a) argues that there is no clear feature that can both explain the
cross-linguistic distribution differences and also hold true for the other degree adverbs in English
(or French). Therefore, this unknown feature of one particular lexical item should not be referred
to as “parameter” (Lardiere, 2009a, p 177). She suggests that linguists need to be wary of risks
76 allowing parameters to proliferate and generating endless sets of micro-parameters that are no more general than the constructions of traditional grammar (Baker, 1996 and Lightfoot 1997 as cited in Lardiere 2009a & Travis, 2008).
Lardiere (2009a) explicitly distinguishes parameters from features, and distinguishes the process of parameter-setting from the selection and assembly of features into language-specific lexical items. For instance, the macro-parameter of the polysynthesis parameter distinguishes polysynthetic languages from isolating ones, but polysynthetic languages differ from one another in microparameters, that is, in certain features that carry the “idiosyncratic morpholexical properties” (Baker, 1996). Within the Generative Grammar framework, features are thus identified as the fundamental units of a language and can be bundled together into functional categories, such as C, T, or D. Features provide the basic information of the phonology, form, and semantics of a language, thus language variation are attributed to feature distinctions. Due to the fact that not all languages make use of all the features in universal grammar (UG), children’s exposure to their first language (L1) has triggered the parameter settings of some features, but disregarding, discarding, or forgetting the others. Features have become “the heart” of generative grammar in the past decades and according to Travis (2008, p.23), language acquisition studies done within this framework are actually turning into studies of “the acquisition of features”.
Second language acquisition (SLA) hence involves a process of re-assembling or reconfiguring features as learners bring “an already-full-assembled set of (L1) grammatical categories”
(Lardiere, 2009a, p. 175) to their SLA experience. Consequently, Lardiere (2008 & 2009 a) believes that a learner’s L2 proficiency is manifested in his/her degree of success in reconfiguring or remapping features from how these are represented in the L1 into that in the L2, which in some cases even involve new formal configurations on distinctive types of lexical items.
77
Cross-linguistic syntactic differences, or micro-parameters, are located in functional
categories, but Lardiere advocates that what matters most to learners are “feature matrices” and
how they are assembled for each language (Lardiere, 2009a) instead of simplistic judgment of
learners’ acquisition of broad functional categories. For instance, it is not clear what it means
that a learner has ‘acquired IP’, but it is more specific and useful to know if the learner has
mastered auxiliaries, agreement, tense marking, verb raising, or nominative case-marking.
Therefore, acquisition of functional categories in the target language is represented in assembly
of hierarchically-ordered bundles of feature matrices. For example, the functional category “T”
in English includes at least an EPP diacritic, the values [±past], uninterpretable/unvalued phi-
feature [ uPerson] and [uNumber], etc. In order to learn this functional category, a Chinese- speaking learner of English needs to learn the feature matrices, specifically, number/person agreement and the past-tense agreement with subject, which are not required in their L1. These features must be indexed to particular lexical items in English, namely the actual sound-meaning parings or morpholexical items mappings. The magnitude of this task demonstrates the general complexity of language acquisition as learners have to compose the feature bundles associated with each category feature by feature (Lardiere, 2009a) and their feature assembly will by no means be uniform across developmental stages of their language acquisition (Hegarty, 2005).
What makes SLA even more complicated in comparison with L1 acquisition is that learners bring to the process a “fully developed system of assembled lexical items and functional categories” (Hegarty, 2005) and they have to figure out how to realize the relevant features in the target language.
The core of Lardiere’s Feature Re-assembly Approach rests on the assumption that learners will need to acquire the morpholexical correspondences in the L2 through semantic or
78 grammatical comparisons to those in their L1 (Lardiere, 2008 & 2009). She is opposed to the
Representational Deficit Hypothesis or Interpretability Hypothesis (Hawkins, 2003; Tsimpli,
2003; Hawkins and Hattori, 2006, among others) and the Failed Functional Features Hypothesis
(Hawkins and Chan, 1997). These proposals predict that the acquisition of an L2 is constrained by the properties of the L1; in other words, the availability of features in the L2 is restricted even though they may have not been selected in the learner’s L1. Specifically, the presence or absence of certain features, e.g. interpretable or uninterpretable, in the L1 determines the acquirability in the target language. For instance, uninterpretable features cannot be acquired beyond a hypothesized critical period unless such features were selected in primary (or early, pre-critical period) language acquisition since they are hypothesized to disappear from the learner’s UG inventory and are no longer accessible. That is why the late L2 acquirer’s knowledge of some features is predicted to “permanently diverge” from that of native speakers (Hawkins and Hattori,
2006, p. 271).
One of the missions of the present study is to investigate and gather evidence from the data to understand whether parameter resetting or feature re-assembly occurs in advanced L2 learners of Chinese. Since the cross-linguistic variation between Chinese and English can be reduced to parameter setting distinctions in the lexicalization of argument structures, will we be able to find any evidence to support the parameter re-setting proposal from the L2 acquisition data?
3.3. Previous SLA Studies of Directed Motion The present study is informed by many SLA studies, including but not limited to those on conflation patterns in the interlanguage representation of argument structure by Inagaki (2001) and Montrul (2001). Ingaki (2001) carried on a study to investigate the L2 acquisition of motion
79 verbs with goal PPs in English and Japanese. Because the research design is a bi-directional investigation, it involves two studies: a study on Japanese learners’ acquisition of English
(Japanese N=42; English control group N=22), and a study on English learners’ acquisition of
Japanese (English N=21; Japanese N=43). The learners’ proficiency level in their respective target language is not comparable in that Japanese participants’ proficiency level in English was lower than the English participants’ proficiency level in Japanese.
A written grammaticality judgment task with pictures was used in both studies. Talmy’s ideas were applied in the research design; in each picture there was “figure” (an object that moves) and “ground” (an object with respect to which the figure moves). There was also an arrow in each picture, which indicated the direction and the endpoint of the motion depicted in the picture. A set of eight candidate sentences for grammaticality judgments were listed under each picture, with a five-point Likert scale ranging from -2 (completely unnatural) through 0
(not sure) to +2 (completely natural). A counter-balanced design was set up to cancel any possible variations caused in the order of the test taken. Each version of the test consisted of 11 target items: five manner-of-motion verbs (walk, run, swim, crawl and fly ) and six goal PPs (to, into, onto, under, over, behind ). There were four target sentence types in the English version and three types in the Japanese version.
3. 1. Four Sentence Types
Language English Japanese Type 1 MANNER V + PP PP + MANNER V Type 2 DIRECTED V + PP + - ING PP + TE + DIRECTED V Type 3 DIRECTED V + PP + BY-ING DIRECTED V + PP + (BY)-ING Type 4 TE + PP + DIRECTED V MANNER V AND DIRECTED V + PP
80
A two-way repeated measures ANOVA was conducted on the English-study data and then on the Japanese-study data. It was found from the English study that Japanese speakers accepted all sentence types except for Type 2 [DIRECTED V + PP + -ING]. English speakers accepted Type 1 [MANNER V + PP], but did not like the other sentence types, esp. Type 3
[DIRECTED V + PP + BY-ING]. Intermediate Japanese learners of English did not have difficulty recognizing the grammaticality of manner-of-motion verbs with goal PPs due to the availability of positive evidence. But they had not yet learned that what they recognized as
English equivalents of Japanese-type forms (Type 3 [DIRECTED V + PP + BY-ING] and Type
4 [MANNER V AND DIRECTED V + PP]) were marginal in English due to the lack of clear positive evidence.
Results of the Japanese study showed that English speakers accepted all sentence types including Type 1 [PP + MANNER V], favoring Type 2 [PP + TE + DIRECTED V]. Japanese speakers rejected Type 1 [PP + MANNER V] and accepted the others. Advanced English learners had difficulty recognizing that manner-of-motion verbs with goal PPs were ungrammatical in Japanese. They had not yet learned that Type 3 [TE + PP + DIRECTED V] was as natural as Type 2 [PP + TE + DIRECTED V] expressing a single motion event, due to their association of the former with [MANNER V AND DIRECTED V + PP] in English.
What this study demonstrates is that L1 transfer persists when an argument structure in
L2 constitutes a subset of its counterpart in the L1. L2 acquisition of argument structure will not be difficult when the L2 is a superset of the L1 due to the availability of positive evidence, and that it is difficult when the L2 is a subset of the L1 because L2 learners could not unlearn L1- based over-generalizations on the basis of positive evidence. This study is one of the first
81 systematic investigations of motion verbs with goal PPs in SLA. However, opposite results were found from Montrul’s study of transitivity alternations in Spanish (2001).
Montrul (2001) also carried out two experiments to examine the effects of L1 on L2 acquisition of argument structure with agentive verbs of directed motion (march, walk) and change-of-state verbs (break, melt) in Spanish and English. The agentive verbs of directed motion undergo a transitivity alternation in English when there is a prepositional phrase, for example, the captain marched the soldiers to the tents. But this transitivity alternation does not occur in Spanish, * El capitan marcho a los soldados hasta el campamento.
Two related experiments, one in English and the other in Spanish, were conducted to test whether learners could distinguish semantically and syntactically between the two classes of verbs that alternate in transitivity in English, i.e. agentive verbs of directed motion and change- of-state verbs. The English study included 19 native speakers as the control group, 17 Spanish- speaking learners, and 18 Turkish-speaking learners. The Spanish study consisted of 20 native speakers, 15 English-speaking learners and 19 Turkish-speaking learners.
All the participants first took a cloze test as an independent measure of proficiency. Then a vocabulary translation task was implemented to make sure they recognized the idiosyncratic meaning of the verbs tested in the experiment. The formal experiment was a picture judgment task. Learners were asked to judge each sentence on a seven-point Likert scale for its grammaticality or semantic appropriateness in the context provided by the picture.
The results showed that learners were able to distinguish between manner-of-motion verbs and change-of-state verbs in the natural-force subject and adjectival passive constructions.
Both Turkish and Spanish learners in the English study behaved similarly and undergeneralized
82 manner-of-motion verbs in the lexical causative construction in English, while Turkish and
English learners of Spanish displayed opposite patterns of response. The findings of both
Montrul’s study (2001) and Inagaki (2001) supported the Full Transfer/Full Access Hypothesis
(Schwartz and Sprouse, 1996), whereby the L1 representation is implicated in the interlanguage lexicon; lexical entries can be restructured on the basis of L2 input.
The next chapter will explain how the research designs of Inagaki (2001) and Montrul were incorporated in the present study. Unlike the parallel studies conducted in these experiments, the present study focuses on only one side of the acquisition direction, that is, how
English-speaking learners of Chinese acquire the expression of causative directed motion.
Chapter 4 discusses the research design and methodology of the experiment, and Chapter 5 presents the results of the study.
83
CHAPTER 4
RESEARCH DESIGN AND METHODOLOGY
4.1. Introduction
This chapter is devoted to the research design and methodologies of the second language acquisition of causative directed manner-of-motion by English-speaking learners of Chinese. A pilot study was conducted before the formal experiment was carried out. Two tasks, sentence translation and picture story-telling, were involved in the pilot study; however, these two tasks were eventually replaced with two other tasks in the main experiment. The major task for the experiment was acceptability judgment based on pictures, and the minor precedent task, picture description in one short sentence in English. I had to give up the original picture story-telling task used in the pilot study due to the fact that only a limited number of the verb tokens were directly related to the type of motion verbs I wanted to study in this project, although a variety of verb types were elicited from that task,. The data collected from the sentence translation task were not completely eliminated from the study, but were only briefly touched upon in the final discussion of Chapter 6. Therefore, the findings of this study were mainly informed by the acceptability judgment task. Chapter 4 and Chapter 5 will hence concentrate on the design and results of this task.
In order to obtain a clearer understanding of the acquisition pattern and sequence of components involved in the causative directed motion events and how the parameter setting of the syntactic structures interact with these segments, five variables were examined and three groups of participants were recruited for the study: a control group (native Chinese speakers), a group of intermediate learners, and a group of advanced learners. Before the experiment was 84
administered to the non-native speakers, its validity and reliability was initially validated by the data from the control group.
This chapter is organized in the following way: the five variables that are examined in this study are identified in Section 4.2; the construction and development of the experiment will be described in full detail in Section 4.3, after which the validity and reliability of the experiment instrument will be evaluated with the data collected from the control group (Section 4.4). I want to make sure that the experiment instrument was valid and reliable and as such both measured what I intended to test and elicited consistent results. Through the use of a quality experiment, then, I can build solid ground upon which to present my results in the next chapter.
4.2. Variables
As discussed in Chapter 2, causative directed manner-of-motion events consist of a number of sub-events. To master the proper way to describe such complex events demands an extremely high level of proficiency. In order to document the stages and processes learners have gone through, this experiment examined the acquisition of five semantic variables:
TRANSITIVITY, GOAL, CAUSATIVITY, ACCOMPANY, and COMPLEX MANNER (in sentences with intransitive verbs). These independent variables and the different levels of the variables are illustrated with examples in Table 4.1.
85
Table 4.1. Experiment Variables
Examples Variables Levels Transitive Verb Intransitive Verb (Used Transitively) 他搬椅子到教室。 *鸭妈妈游小鸭到河对岸。 Transitive tā b ān y ǐzi dào jiàoshì yā m āma yóu xi ǎoy ā dào hé duìàn he move chair to duck mom swim little duck to TRANSITIVITY classroom river opposite bank ‘He moved chairs to ‘Mother Duck swam the little Intransitive the classroom.’ duck across to the other bank of
the river.’ *鸭妈妈到河对岸游小鸭。 他到教室搬椅子。 yā m āma dào hé duìàn yóu xi ǎoy ā tā dào jiàoshì b ān y ǐzi duck mom to river opposite bank he to classroom move swim little duck Preverbal chair Intended meaning: ‘Mother Duck ‘He in the classroom on the other bank of the river moved chairs.’ swam the little duck.’ GOAL 他搬椅子到教室。 *鸭妈妈游小鸭到河对岸。 tā b ān y ǐzi dào jiàoshì yā m āma yóu xi ǎoy ā dào hé duìàn he move chair to duck mom swim little duck to Postverbal classroom river opposite bank ‘He moved chairs to Intended meaning: ‘Mother Duck the classroom.’ swam the little duck across to the other bank of the river.’ *鸭妈妈游小鸭到河对岸。 他搬椅子到教室。 yā m āma yóu xi ǎoy ā dào hé duìàn tā b ān y ǐzi dào jiàoshì duck mom swim little duck to Unmarked he move chair to river opposite bank (no Use of classroom Intended meaning: ‘Mother Duck ba construction) ‘He moved chairs to swam the little duck across to the the classroom.’ other bank of the river.’
CAUSATIVITY *鸭妈妈把小鸭游到河对岸。 他把椅子搬到教室。 yā m āma b ǎ xi ǎoy ā yóu dào hé tā b ǎ y ǐzi b ān dào duìàn jiàoshì duck mom Ba little duck swim to he Ba chair move to river opposite bank Marked with classroom Intended meaning: ‘Mother Duck ba construction ‘He moved chairs to swam the little duck across to the the classroom.’ other bank of the river.’ 86
*鸭妈妈游小鸭到河对岸。 他搬椅子到教室。 yā m āma yóu xi ǎoy ā dào hé duìàn tā b ān y ǐzi dào jiàoshì duck mom swim little duck to Unmarked he move chair to river opposite bank (no use of zhe ) classroom Intended meaning: ‘Mother Duck ‘He moved chairs to swam the little duck across to the the classroom.’ other bank of the river.’
ACCOMPANY *鸭妈妈游着小鸭到河对岸。 他搬着椅子到教室。 yā m āma yóu zhe xi ǎoy ā dào hé tā b ān zhe y ǐzi dào duìàn Marked jiàoshì duck mom swim zhe little duck to with zhe he move zhe chair to river opposite bank classroom Intended meaning: ‘Mother Duck ‘He went to the went to the other bank of the classroom, moving river, swimming the little duck.’ chairs.’ 鸭妈妈背小鸭游到河对岸。 yā mama b ēi xi ǎoy ā yóu dào hé duìàn Unmarked duck mom carry little duck swim Complex to river opposite bank Manner+Motion ‘Mother Duck carried the little duck (on her back) to swim across to the other bank of the river.’ COMPLEX Not Available MANNER 鸭妈妈背着小鸭游到河对岸。 yā mama b ēi zhe xi ǎoy ā yóu dào Marked hé duìàn Complex duck mom carry zhe little duck Manner+Motion swim to river opposite bank (with zhe ) ‘Mother Duck carrying the little duck (on her back) swam across to the other bank of the river.’
4.3. Instrument Development
The experiment was designed with the above variables in mind and each judgment question had at least six to eight options corresponding to these variables, so the basic content validity of the instrument was assured. In addition, three steps were taken to validate that the 87
observed differences amongst the subjects in the experiment were indeed their acquisition of the six semantic variables tested in the experiment,.
First, the frequencies of all the measured verbs were checked with HSK ( Hanyu Shuiping
Kaoshi ), the standardized Chinese Proficiency Test, and the Academia Sinica database. In this way, all the tested manner of motion verbs, transitive and intransitive, were assured to be of high frequency in the language input of modern Chinese. Second, both the pronunciation and English meaning citation of the tested vocabularies were provided next to the judgment task. Even if participants did not learn the tested verb at the time of experiment, they could check the provided vocabulary. This ensured that the experiment did not focus on participants’ recognition of these verbs, but how these semantic variables were projected in the interlanguage syntax. In addition, the experiment was first administered to a control group of 40 native speakers. Both the validity and reliability of the experiment were investigated, and subsequent revisions were made to experimental items based on the results. Section 4.6 will cover the results from the control group with a focus on the reliability and validity of the experiment.
A total number of sixteen sets of pictures and corresponding questions were created: half of these questions selected transitive manner–of–motion verbs as the main verbs; the other half took on intransitive manner–of–motion verbs. An additional eight sets of questions (with or without pictures) were included as the distracters and used as proficiency test. The experiment was conducted online using DatStat Illume Survey Management software, which allowed for randomization, HTML customization, advanced logic patterns and sophisticated validation. In the next three sections, the tasks and token types of the experiment will be addressed followed by an example afterwards to illustrate the construction of the experiment. 88
4.4. Tasks
In this experiment, all of the subjects took two tasks: a picture–description task in English, and a Chinese sentence acceptability/ naturalness judgment task with the same picture. The judgment task of the same picture took place immediately after the picture–description task. In the first task, subjects were asked to describe the picture concisely with one transitive sentence in
English. A guiding question was provided below the picture in the first task to help elicit language production relevant to the project. For example, if the picture showed a woman pulling a suitcase to her house, then the guiding question was: “what did the woman do to her suitcase?”.
After the picture–description task, the same picture was shown on the next page, and this time the relevant vocabularies (given in both Pinyin and English) and meanings were provided.
Subjects were asked to evaluate the acceptability/naturalness of six to eight Chinese sentences to describe the motion event depicted in the picture. Judgments were given on a five–point Likert scale ranging from –2 (very unnatural) through 0 (cannot decide) to +2 (very natural). Subjects took several examples for practice before participating in the formal experiment. The acceptability judgment task was made up of eight transitive manner-of-motion verbs (each with six sentence tokens to judge in Task 2), eight intransitive manner-of-motion verbs (each with eight sentence tokens to judge in Task 2), and eight distracters serving as proficiency test items.
The pictures were quite similar in nature, each depicting a complex motion event with one entity (a causer) doing something to another entity (causee) which resulted in a change of location (the ground , Talmy, 1985). Arrows were added to all the pictures in order to indicate the direction and the endpoint of the motion event. In other words, these pictures were selected and designed in such a way that descriptions of them encouraged, if not required, the use of 89
externally caused structure of directed manner–of–motion. The guiding question in the first task was used to induce language production of such structures. The reason for arranging the picture– description task before the judgment task was twofold: first, the picture–description task in L1 forced the subject to conceptualize what he/she observed in the picture – it had the same function as a translation task to control the relevancy of the elicited data, but it avoided the pitfalls of word–to–word translation from L1 to L2; second, the research design could be extended to a parallel acquisition study of English–speaking learners of Chinese and Chinese–speaking learners of English. This thesis, however, only focuses on the data collection, result analyses and reports of L2 acquisition of Chinese.
4.4.1. Sentence Token Types
To investigate the above listed independent variables, each with two or more levels, the
experiment focused on a total of 8 types of sentence tokens (See Table 4.2). Following the model
of these token types, six sentences were created for each transitive manner-of-motion verb and
eight sentences for each intransitive manner-of-motion verb.
90
Table 4.2. Sentence Token Types
Sentence Token Types Verb Types Examples 他搬椅子到教室。 tā b ān y ǐzi dào jiàoshì Transitive he move chair to classroom ‘He moved chairs to the classroom.’ Sentence Type 1 *鸭妈妈游小鸭到河对岸。 MANNER V + PP yā m āma yóu xi ǎoy ā dào hé duìàn Intransitive duck mom swim little duck to river opposite bank Intended meaning: ‘Mother Duck swam the little duck across to the other bank of the river.’ 他把搬椅子到教室。 tā b ǎ y ǐzi b ān dào jiàoshì Transitive he Ba chair move to classroom ‘He moved chairs to the classroom.’ Sentence Type 2 *鸭妈妈把小鸭游到河对岸。 ba...+ MANNER V +PP yā m āma b ǎ xi ǎoy ā yóu dào hé duìàn Intransitive duck mom Ba little duck swim to river opposite bank Intended meaning: ‘Mother Duck swam the little duck across to the other bank of the river.’ ?他到教室搬椅子。 tā dào jiàoshì b ān y ǐzi Transitive he to/in classroom move chair ‘He in the classroom moved chairs.’ Sentence Type 3 *鸭妈妈到河对岸游小鸭。 PP+ MANNER V yā m āma dào hé duìàn yóu xi ǎoy ā Intransitive duck mom to river opposite bank swim little duck Intended meaning: ‘Mother Duck on the other bank of the river swam the little duck.’ 他搬着椅子到教室。 tā b ānzhe y ǐzi dào jiàoshì Transitive he move zhe chair to classroom ‘He went to the classroom, moving chairs.’ Sentence Type 4 *鸭妈妈游着小鸭到河对岸。 MANNER V zhe +PP yā m āma yóu zhe xi ǎoy ā dào hé duìàn Intransitive duck mom swim zhe little duck to river opposite bank Intended meaning: ‘Mother Duck went to the other bank of the river, swimming the little duck.’ 他搬椅子搬到了教室。 Sentence Type 5 tā b ān y ǐzi b ān dào le jiàoshì Transitive Verb Copying he move chair move to classroom ‘He moved chairs to the classroom.’ 91
*鸭妈妈游小鸭游到河对岸。 yā m āma yóu xi ǎoy ā yóu dào hé duìàn Intransitive duck mom swim little duck swim to river opposite bank Intended meaning: ‘Mother Duck swam the little duck across to the other bank of the river.’ ?他到教室搬着椅子。 tā dào jiàoshì b ān zhe y ǐzi Transitive he to classroom move zhe chair ‘He went to the classroom, moving chairs.’ Sentence Type 6 *鸭妈妈到河对岸游小鸭。 PP+ MANNER V zhe yā m āma dào hé duìàn yóu xi ǎoy ā Intransitive duck mom to river opposite bank swim little duck Intended meaning: ‘Mother Duck on the other bank of the river swam the little duck.’ 鸭妈妈背小鸭游到河对岸。 yā mama b ēi xi ǎoy ā yóu dào hé duìàn Sentence Type 7 V + V + PP duck mom carry little duck swim to river opposite bank MANNER MOTION ‘Mother Duck carried the little duck (on her back) to swim across to the other bank of the river.’ Intransitive 鸭妈妈背着小鸭游到河对岸。 yā mama b ēi zhe xi ǎoy ā yóu dào hé duìàn Sentence Type 8 duck mom carry zhe little duck swim to river opposite MANNER Vzhe+MOTION V + PP bank ‘Mother Duck carrying the little duck (on her back) swam across to the other bank of the river.’
92
4.4.2. One Example
In this section, one example of the experiment will be demonstrated. Both simplified and traditional Chinese is provided for the experiment. Note that Task 2 is shown on a separate webpage after the participants finish Task 1. As shown in the picture (Figure 1), a man carries a sleeping baby in his arms and the red arrow points the direction of his action/motion towards the bed.
Figure 4.1 (a). Example Picture
Task 1: Try to describe the picture in a short transitive sentence in English. 尽量用一句
简单的(及物动词)英语描述此图/盡量用一句簡單的(及物動詞)英語描述此圖。Guiding
Question: What did he do to the baby? 他 对 孩 子 做 了 什 么 ? 他 對 孩 子 做 了 什 麼 ？
______
Task 1 asked participants to describe the picture in a transitive sentence, specifically, what he/the man did to the baby . This guiding question prompted participants to focus on the motion event with a result or destination implied. As expected, most participants came up with a sentence containing a manner of motion main verb and a prepositional phrase to indicate the goal of the motion. The most commonly produced English sentences for this picture were: 93
(4.1) He put/placed the (sleeping) baby in/on/to (the) bed. (4.2) He carried/ tried to carry the baby (over)to the bed. (4.3) He tucked the sleeping baby into bed.
Now that participants have cognitively and actively described the motion event, the subsequent Task 2 asked them to judge the naturalness or acceptability of six sentences in their
L2, Chinese. The provided vocabulary made certain that the semantic meaning of the words would not block their understanding and hence judgment of the sentences. This was a controlled task and it was fairly less powerful than a pure production task in terms of accurately demonstrating L2 learners’ level since the exact verb had already been determined for them. That being said, however, Task 2 did measure and manifest the knowledge and preference of the participants in using the provided sentences to delineate the motion event.
Figure 4.2(b). Task 2:
Vocabulary that may be used in the sentences:
抱: bào(hold...in the arm) 把: bǎ(Ba Construction) 他: tā(he) 小孩: xi ǎohái(little baby) 床上: chuáng(bed) shàng(top/up) 到: dào(to) 着/著: zhe(Zhe Construction)
Very Not Very Cannot Very Acceptable Sentence Type Unnatural Natural Decide Natural 還行 很不自然 不太自然 不知道 很好 他抱小孩到床上。
他把小孩抱到床上。
他到床上抱小孩。
他抱着小孩到床上。
他抱小孩抱到床上。
他到床上抱着小孩。
Compared with commonly operated production task in SLA studies, the two tasks for the current experiment were designed and presented in such an order/manner that the cognitive processing on L2 learners’ end was highly simplified and the attention could be focused on the 94
research questions and variables. In a pure production task, for instance, an oral description of the picture in L2, participants usually go through a rather complicated cognitive processing: they need to take a close look at the picture, understand the situation or sometimes put themselves in the context, interpret the key elements in the picture, and then recall what they have learned in their L2 that can help them verbalize this image. The challenges that many lower-level or intermediate-level L2 learners face in this type of task are that sometimes they get overwhelmed with the amount of information embedded in the picture or they get so stuck on one word or one grammar point that they do not have the cognitive resources to produce the simple sentence that researchers intend to elicit. Another challenge for adopting this natural task is that the same motion event can be interpreted and described from so many perspectives that it may or may not relate to the variables I want to test in the first place.
Since the main verb involved in this example was transitive in nature, only six sentence tokens were provided for participants to judge. In the case of intransitive verbs, a total number of eight sentence tokens were tested. The order in which these tokens appear was automatically randomized by the software program in order to reduce the internal variability. It was admitted that six or eight sentences for one judgment item might seem to be too much for the participants as it required a relatively longer time to complete, but an advantage was that these tokens looked quite similar and it forced participants to evaluate the tokens and to make judgment more carefully. On one hand, learners’ language intuition was crucial for this experiment; on the other hand, prudent appraisal and comprehensive comparison tended to reveal with confidence how much the L2 learners actually understood the contrasts and differences between these tokens.
The above example illustrated how I created the tasks for the experiment and what they looked like. 95
4.5. Participants
The subjects for the study consisted of three groups: the control group (40 Chinese native speakers) and two groups of English–speaking learners of L2 Chinese, 33 in total. The L2 learners had been taking or had taken Chinese for more than two semesters. This study excluded beginning learners of Chinese due to the complex nature of the phenomenon being investigated.
These 33 participants were divided into two proficiency groups (intermediate and advanced) based on two factors: the length of Chinese instruction they had received, and their performance in a proficiency test. Table 4.3 and Table 4.4 provide information on the subjects.
Table 4.3. Participant Information (Age)
Number of Age Range Group Subjects 12–18 18–25 26–35 36–45 Control 40 6(15%) 26(65%) 8(20%) Group Intermediate 17 10(58.82%) 7(41.18%) Advanced 16 7(43.75%) 6(37.5%) 3(18.75%)
Table 4.4. Length of Instruction for L2 Participants
Number of Length of Chinese Instruction Group Subjects <3 semesters 3–6 semesters >6 semester Intermediate 17 7(41.18%) 9(52.94%) 1(5.88%) Advanced 16 7(43.75%) 9(56.25%)
Forty native speakers of Mandarin Chinese were recruited from a popular online reading forum. They participated in this experiment as volunteers and were only rewarded with game points for that forum. All of these participants were educated native speakers of Chinese from 96
three age groups (see Table 4.3): 15% of them were college students or new college graduates between the ages of 18 to 25, 65% were college graduates between the ages of 26 and 35, 20% were between the ages of 36 and 45. These native speakers were included in this study for three major purposes. First, they were utilized to validate the internal reliability of our experiment instruments – if the experiment was broken down into two parts, would the two parts be equivalent in soliciting similar results? Do the native speakers show similar patterns in judging each type of sentence tokens? Second, the control group of native speakers was also used to check the validity of our experiment – do the instrument items test the variables we intended to measure? Third, they also serve as a criterion to evaluate the language proficiency of the non- native speakers. The results of the non-native speakers were compared to those of the control group to see if the differences were statistically significant.
Since the non-native speakers were recruited from five different institutions in the United
States, they were of different age and education levels (middle school, high school, college and graduate school) and the factor of the instructional quality was not controlled. It was therefore impossible to simply use the length of their Chinese learning as an indicator of their language proficiency. As a result, I made full use of filters for the experiment, which not only served as distracters, but also functioned as a proficiency test. The same filters/ distracters were administered to all subjects involved in the study. The validity and reliability of this measure will be discussed in Chapter 5.
Of the two groups of L2 learners, the intermediate group had 17 participants: seven of them took Chinese at least more than one and a half semesters, ten of them took the target language for more than three semesters but less than six semesters, one student took more than three years of Chinese in high school but he/she was included in this group due to his/her poor 97
performance in the proficiency test. The latter advanced group was made up of 16 participants:
43.75% of them took Chinese for more than three semesters, 56.25% of them for more than six semesters.
Before giving the experiment to non-native speakers, the experiment was administered to the control group (of 40 native speakers) so as to validate the internal consistency of the instruments. The validity and reliability of the experiments will be covered in Section 4.6.
4.6. Reliability and Validity of the Experiment
4.6.1. Two Experiment Subsets
The experiment comprised the same number items for each category: eight transitive
verbs, eight intransitive verbs, and eight distracters/proficiency test items. Some subjects
participated in the project as volunteers; other subjects were compensated with a ten–dollar gift
card. In either case, however, it was impractical to assign a long experiment to the subjects. Thus
it became especially crucial for me to be able to randomly assign a subject to a shorter test,
which nevertheless is reliable and achieves the desired goal. As a result, the experiment was
divided into two parts, Subset A and Subset B respectively, and was administered to the control
group. Whether or not a subset of the experiment generates reliable results will help me decide
what to do if some L2 participants only finish half or part of the experiment. Therefore, each
subset contained a total of 16 items (4 transitive verbs, 4 intransitive verbs, and 8 proficiency test
items). The acceptability judgment task included six sentence tokens for each transitive verb
(Type 1–6) and eight sentence types (Type 1–8) for each intransitive verb. Both subsets of the
experiment shared the exact same proficiency test items. The complete list of verbs used in the
sentence tokens are demonstrated in Table 4.5. 98
The items of the experiment were randomly assigned to two subsets and were administered to two groups of control groups (native speakers). Two additional advantages of dividing the experiment into two subsets were: first, it made it possible to have a short test for native speaker volunteers; second, it allowed a comparison between the two subparts.
Table 4.5. Tested Verbs in the Experiment
Experiment Sets Transitive Verbs Intransitive Verbs 滚 gǔn ‘roll’ 游 yóu ‘swim’ 追 zhu ī ‘chase’ 爬 pá ‘climb’ Subset A 挂 guà ‘hang’ 滑 huá ‘slide’ 推 tu ī ‘push’ 行军 xíngj ūn ‘march’ 扔 rēng ‘throw’ 飞 fēi ‘fly’ 抱 bào ‘hold’ 跑 pǎo ‘run’ Subset B 踢 tī ‘kick’ 跳‘tiào’ ‘jump’ 搬 bān ‘move’ 走 zǒu ‘walk’
A total number of 45 native speakers participated in the online research. I randomly selected the data of 40 participants and excluded the rest for analysis with the consideration of a statistically balanced design. After providing the language background information, each subject was given a chance to select a subset of the experiment to work on. The presentation order of subsets was randomized automatically: sometimes Subset A appeared as the first option, and other times Subset B. The report here only includes the results of 20 native speakers who took
Subset A (henceforth Control Group A) and 20 people who took Subset B (henceforth Control
Group B).
4.6.2. Overall Reliability of the Two Subsets 99
The reliability of the experiment was validated by analyzing the results of the control group (native speakers of Mandarin Chinese). Reliability refers to the internal consistency of the measurement. Cronbach's α (alpha) was used to inspect the internal consistency of the experiment tokens. Cronbach's α ranges from 0 to 1: when α =0, the true score is not measured
and all the items are error components; when α =1, all items measure only the true score, and there is no error component. By convention, a lenient cut–off of .60 is common in explanatory research. A Cronbach’s α of 0.7 is a rule–of–thumb as an acceptable level of agreement.
Table 4.6. Reliability of the Experiment
Cronbach's Cronbach's Alpha Based on Number of Items Alpha Standardized Items Subset A .854 .864 75 Subset B .839 .854 80
The analysis of the acceptability judgment task yielded a Cronbach’s α of .854 for
Experiment Subset A and .839 for Subset B (Table 6) respectively. This means that both subsets
of the experiment were very reliable measurement of the variables and they were able to elicit
consistent acceptability judgment from the 40 native speakers. The next section will concentrate
on the comparison of the two subsets.
4.6.3. Descriptive Statistics of the Two Subsets
A general analysis of the results from both control groups exhibits similar patterns. The
results of Group A are compared to those of Group B in Table 7. The results are also illustrated
in the following two graphs: Figure 4.2 and Figure 4.3. In general, for the causative transitive
verbs, Type 2 ba ...+ MANNER V +PP was much more preferred than any other form of expression;
and Type 1 ( MANNER V+PP) and Type 4 ( MANNER V zhe +PP) are comparatively equally preferred. 100
Native speakers in Group A preferred to use Type 4 by marking the ACCOMPANY form with zhe , whereas those in Group B exhibited an opposite pattern of preference. Neither group accepted the verb copying structures. Manner of motion verbs with preverbal prepositions with or without zhe , Type 3 and 6 PP+ MANNER V ( zhe) respectively, were regarded as unacceptable in
this context of directed motion.
Table 4.7. Mean Ratings of the Sentence Types by Control Group A & B
Group A Group B Average Sentence Token Types Transitive Intransitive Transitive Intransitive Transitive Intransitive
Type 1: MANNER V + PP 0.48 –1.11 0.75 –1.33 0.59 –1.21
Type 2: ba ...+ MANNER V +PP 1.62 –0.18 1.98 –1.39 1.81 –0.79
Type 3: PP+ MANNER V –1.18 –1.58 –1.33 –1.68 –1.28 –1.66
Type 4: MANNER V zhe +PP 0.79 –1.07 0.24 –1.58 0.51 –1.33 Type 5: Verb Copying 0.00 –1.21 –0.08 –1.29 –0.08 –1.25
Type 6: PP+ MANNER V zhe –1.42 –1.75 –1.53 –1.88 –1.49 –1.82
Type 7: MANNER V+ MOTION V + PP 0.63 1.73 1.18
Type 8: MANNER Vzhe +MOTION V +PP 1.24 1.67 1.47
For the causative intransitive verbs, none of the above mentioned sentence types were accepted, even though Group A speakers found Type 2 ba ...+ MANNER V +PP more acceptable than other types (Type 1 & Type 3–6). Instead, Type 7 and Type 8 MANNER V( zhe) + MOTION V + PP were rated as the most grammatical forms to express agentive directed manner–of–motion. 101
Figure 4.3. Results of Control Group A
2.00 Transitive 1.50 Intransitive 1.00
0.50
0.00 Type 1 Type 2 Type 3 Type 4 Type 5 Type 6 Type 7 Type 8 -0.50
-1.00
-1.50
-2.00
Figure 4.4. Results of Control Group B
2.00 Transitive 1.50 Intransitive 1.00
0.50
0.00 Type 1 Type 2 Type 3 Type 4 Type 5 Type 6 Type 7 Type 8 -0.50
-1.00
-1.50
-2.00
The general results from the two control groups confirmed that differences exist between the six variables and the two control groups showed similar patterns in their acceptability judgment of the sentence types. In the following sections, we will elaborate on how different these variables are with multiple T–tests. But before we get into the more detailed statistical 102
analyses of the variables, we need to make sure that the variance homogeneity exists between the two groups or two experiment subsets.
4.6.4. Homogeneity Variances between Two Experiment Subsets (Control Groups)
A Levene’s test for equality of variances was carried out to see if the variances of the two
control group samples are equal, which sets the precondition for the following Paired Samples
T–tests in Section 4.6.5.1. The results of the Levene’s test are presented in Table 8. Overall, the
two control group samples demonstrated equal variances in their judgment of the sentence tokens.
Table 4.8. Levene's Test for Equality of Variances between Control Groups/Experiment Subsets
Levene's Test for Levene's Test for Equality of Equality of Variances Sentence Token Types Variances Sentence Token Types F Sig. F Sig. Type 1 Type 1 3.833 0.058 0.53 0.472 (with Transitive Verbs) (with Intransitive Verbs) Type 2 * Type 2 27.36 0.000 3.42 0.072 (with Transitive Verbs) (with Intransitive Verbs) Type 3 Type 3 0.67 0.416 1.65 0.207 (with Transitive Verbs) (with Intransitive Verbs) Type 4 * Type 4 * 5.33 0.026 4.27 0.046 (with Transitive Verbs) (with Intransitive Verbs) Type 5 Type 5 0.02 0.903 0.002 0.968 (with Transitive Verbs) (with Intransitive Verbs) Type 6 Type 6 0.05 0.832 4.05 0.051 (with Transitive Verbs) (with Intransitive Verbs) Type 7 * 5.30 0.027 (with Intransitive Verbs) Type 8 3.97 0.054 (with Intransitive Verbs) * p< .05
However, four of the tests were found to be statistically significant: (1) Sentence Type 2 (with transitive verbs), F(38)= 27.36, p< .001; (2) Sentence Type 4 (with transitive verbs), F(38)= 5.33,
p= .026; (3) Sentence Type 4 (with intransitive verbs), F(38)= 4.27, p= .046; and (4) Sentence 103
Type 7 (with intransitive verbs), F(38)= 5.30, p= .027. In other words, the two control groups behaved differently in judging these four sentence types.
A closer examination of the data (Table 4.7, Figure 4.2, Figure 4.3 and raw data) in these four sentence types reveals that these two groups reached agreement on their general acceptability judgment of these sentences; for example, they all accepted Sentence Type 2 with
Transitive Verbs (ba ...+ MANNER V +PP) as the most natural way to describe the picture, but they disagreed in their rating scales –– Group B systematically rated higher than Group A speakers.
That is, speakers in Group B were much clearer with their judgment. When they accepted one token, they tended to assign +2 (Very Natural) to the given token; and when they disliked one sentence token, they were inclined to assign –2 (Very Unnatural) to the token. On the other hand, speakers in Group A acted more conservatively. Even when they accepted one token as the best or worst choice for the context, they tended to assign only ±1 (Natural/Unnatural) to that token.
This pattern of rating differences led to the variances that occurred in the other types of tokens as well.
To conclude, group preferences did exist, especially in Sentence Type 4 with Transitive
Verbs and Sentence Type 7 with Intransitive Verbs, but generally speaking, the equality of variances between the two control groups were satisfied. Since all the native–Chinese participants randomly selected the experiment subset that they themselves took, it is safe to conclude that the two subsets of the experiment met the requirement of homogeneity of variances, i.e. no matter which subset of the experiment, or even which experiment, a participant took, it was very likely that the experiment (subset) would accurately measure his/her judgment of the token types. In the rest of the analyses, no specification of experiment subset is necessary and the focus will be on the sentence token types instead. 104
4.6.5. Validity of the Instrument
Now that the Levene’s test for equality of variances have shown that the two experiment subsets are homogeneous in terms of variances, this section will be dedicated to analyses of distinctions within the six variables. I will cover the Paired Samples T–tests for each variable one by one. The results of the first variable, TRANSITIVITY, are presented here. I have used the same approach to evaluate the validity of the other four variables, GOAL, CAUSATIVITY,
ACCOMPANY, and COMPLEX MANNER, but the results of the four variables are included in
Appendix B-E instead. At the end of this section, a short summary of the validity for all five variables will be provided.
Table 4.9. Descriptive Statistics of Sentence Types for TRANSITIVITY
Std. Std. Error Sentence Type Mean N Deviation Mean
Sentence Type 1 with Transitive Verbs .594 40 .746 .118 Sentence Type 1 with Intransitive Verbs -1.213 40 .495 .078 Sentence Type 2 with Transitive Verbs 1.813 40 .324 .051 Sentence Type 2 with Intransitive Verbs -.788 40 .884 .140 Sentence Type 3 with Transitive Verbs -1.281 40 .698 .110 Sentence Type 3 with Intransitive Verbs -1.656 40 .399 .063 Sentence Type 4 with Transitive Verbs .51 40 .562 .089 Sentence Type 4 with Intransitive Verbs -1.331 40 .626 .099 Sentence Type 5 with Transitive Verbs -.075 40 .723 .114 Sentence Type 5with Intransitive Verbs -1.250 40 .555 .088 Sentence Type 6 with Transitive Verbs -1.488 40 .493 .078 Sentence Type 6 with Intransitive Verbs -1.819 40 .300 .0474
Six pairs of t–tests were run on the factor of TRANSITIVITY, comparing each sentence type between transitive verbs and intransitive verbs. Whichever type of sentence tokens was involved, the t–test showed statistically significant differences between the two types of verbs 105
(Table 4.9 & 4.10). Great details will be provided for the analysis of the first variable. Below we will go through the six pairs of t-tests one by one, and then examples for each pair of t-test are provided in (4.2) – (4.7) below with corresponding graphs for the mean values on the side.
Table 4.10. Paired Samples T–test on TRANSITIVITY
95% Confidence Interval of the Std. Std. Error Sig. Transitive –Intransitive Mean t df Deviation Mean Difference (2-tailed) Lower Upper
Sentence Type 1 1.806 .752 .119 1.566 2.047 15.189 39 .000 Sentence Type 2 2.600 1.050 .166 2.264 2.936 15.665 39 .000 Sentence Type 3 .375 .641 .101 .170 .580 3.703 39 .001 Sentence Type 4 1.838 .674 .107 1.622 2.053 17.254 39 .000 Sentence Type 5 1.175 .692 .109 .954 1.396 10.746 39 .000 Sentence Type 6 .331 .472 .0746 .180 .482 4.442 39 .000
A paired-samples t-test was conducted to compare the role of TRANSITIVITY in
Sentence Type 1 with Transitive Verbs and Intransitive Verbs (MANNER V + PP). There was a significant difference in native speakers’ judgment ratings for sentences with Transitive Verbs
(M= 0. 59, SD= 0.75) and Intransitive Verbs (M= -1.21, SD= 0.50); t(39) = 15.189, p < .001.
This suggests that for Sentence Type 1 ( V + PP), MANNER Type 1 (MANNER V + PP) whether the main verb is transitive or intransitive makes a 2.00 1.50 significant difference. This pair of comparison is 1.00 0.50 exemplified in (4.2a), and the intransitive verb, as in 0.00 -0.50 Transitive Intransitive (4.2b). On average, the mean acceptability of Sentence -1.00 Type 1 with Transitive Verbs like (4.2a) by the control -1.50 -2.00 groups was rated +0.59, whereas (4.2b) was rated – Figure 4. 5. Effect of TRANSITIVITY on Sentence Type 1 106
1.21 (See Table 4.9 & Figure 4.4). Given that +2 meant the sentence sounded very natural to these native speakers, +1 acceptable, 0 cannot decide, –1 not very natural, and –2 completely unnatural, the sentence with an intransitive verb (4.2b) was basically not correct in native
Chinese. On the other hand, the sentence with a transitive verb (4.2a) was understandable, but not quite acceptable.
4.2. Type 1 ( MANNER V + PP) a) 他搬椅子到教室. tā b ān y ǐzi dào jiàoshì he move chair to classroom ‘He moved chairs to the classroom.’
b) *鸭妈妈游小鸭到河对岸。 yā m āma yóu xi ǎoy ā dào hé duìàn duck mom swim little duck to river opposite bank ‘Mother Duck swam the little duck across to the other bank of the river.’ Comparing Sentence Type 2 of transitive verbs like 4.3(a) with intransitive verbs like
4.3(b), another paired-samples t-test was conducted to compare the role of TRANSITIVITY in
Sentence Type 2 with Transitive Verbs and Intransitive Verbs (ba ...+ MANNER V +PP). There was a
significant difference in native speakers’ judgment ratings for this sentence type with Transitive
Verbs (M= 1.81, SD= 0.32) and Intransitive Verbs (M= -0.79, SD= 0.88); t(39) = 15.665, p
< .001 (Table 4.9, Table 4.10 & Figure 4.5). This indicates that whether the main verb is transitive significantly impacts native speakers’ judgment of the same sentence type.
On average, the mean acceptability of Sentence Type 2 with Transitive Verbs by the control groups was rated +1.81, whereas the ones with Intransitive Verbs was –0.79. In licensing the causatives of directed manner–of–motion, Type 2 ( ba ...+ MANNER V +PP) was regarded as the most natural way for transitive verbs and the use of ba construction made the sentences with
intransitive verbs (–0.79) sound not as bad as that of Type 4.1(b), which was rated as –1.21, even
though it was considered unnatural. The reason why 4.3(b) sounded better than 4.2(b) was that 107
even though the intransitive verb yóu ‘ swim ’ in Chinese could not participate in the causative expressions, the use of ba construction did signal the semantic meaning of CAUSATIVE.
4.3. Type 2 (ba...+ V +PP) MANNER Figure 4.6. Effect of TRANSITIVITY on Sentence Type 2 a) 他把搬椅子到教室。 tā b ǎ y ǐzi b ān dào jiàoshì Type 2 (ba ...+ MANNER V +PP) he Ba chair move to classroom 2.00 ‘He moved chairs to the classroom.’ 1.50 1.00 0.50 b) *鸭妈妈把小鸭游到河对岸。 0.00 yā m āma b ǎ xi ǎoy ā yóu dào hé duìàn -0.50 Transitive Intransitive duck mom Ba little duck swim to river -1.00 opposite bank -1.50 Intended meaning: ‘Mother Duck swam the -2.00 little duck across to the other bank of the river.’
Preverbal prepositional phrases (Type 4.3), whether the verb was transitive (4.4a) or intransitive (4.4b), were unanimously rejected in the context of directed manner-of-motion: –
1.28 and –1.66. But the sentences with intransitive Figure 4.7. Effect of TRANSITIVITY on Sentence Type 3 verbs sounded much worse than those with transitive Type 3 (PP+ MANNER V) verbs, and the difference was statistically significant in 2.00 1.50 the third paired samples t-test: Sentence Type 3 with 1.00 Transitive Verbs (M= -1.28, SD= 0.70) and Intransitive 0.50 0.00 Verbs (M= -1.66, SD= 0.40); t(39) = 3.70, p < .001. -0.50 Transitive Intransitive -1.00 (Table 4.9, Table 4.10 & Figure 4.6). The reason for the -1.50 -2.00 significant difference was self-evident. 4.4(b) was not only inappropriate for the context, but also grammatically incorrect, so it was rejected; on the 108
other hand, 4.4(a) was grammatically correct, although it was not appropriate for the context, so it sounded a little better.
4.4. Type 3 (PP+ MANNER V) a) ?他到教室搬椅子。 tā b ǎ yǐzi b ān dào jiàoshì he to/in classroom move chair ‘He in the classroom moved chairs.’
b) *鸭妈妈到河对岸游小鸭。 yā m āma b ǎ xi ǎoy ā yóu dào hé duìàn duck mom to river opposite bank swim little duck Intended meaning: ‘Mother Duck on the other bank of the river swam the little duck.’
Sentences 4.5(a) and 4.5(b) highlight the meaning of ACCOMPANY or Accompanied
Action with the use of zhe. Again, the TRANSITIVITY effect differentiates the two: the
acceptability of sentence (4.5a) is 0.51, while that of (4.5b) is -1.33. Another paired-samples t-
test was conducted to compare TRANSITIVITY in Sentence Type 4 with Transitive Verbs and
Intransitive Verbs (MANNER V zhe +PP). Again, we Figure 4.8. Effect of TRANSITIVITY on Sentence Type 4 found another significant difference in judgment ratings Type 4 (MANNER V zhe +PP) for Sentence Type 4 with Transitive Verbs (M= 0.51, SD= 2.00 1.50 0.56) and those with Intransitive Verbs (M= -1.33, SD= 1.00 0.50 0.63); t(39) = 17.254, p< .001. (Table 4.9, Table 4.10 & 0.00 Figure 4.7). The results imply that the TRANSITIVITY -0.50 Transitive Intransitive -1.00 nature of the main verb determines the acceptability of -1.50 Sentence Type 4. Sentences with transitive verbs like (4.5a) -2.00
are barely acceptable, but (4.5b) is mostly unacceptable in a native speaker’s L1 grammar.
4.5. Type 4 ( MANNER V zhe +PP) a) 他搬着椅子到教室。 tā b ānzhe y ǐzi dào jiàoshì 109
he move zhe chair to classroom ‘He went to the classroom, moving chairs.’
b) *鸭妈妈游着小鸭到河对岸。 yā m āma yóu zhe xi ǎoy ā dào hé duìàn duck mom swim zhe little duck to river opposite bank Intended meaning: ‘Mother Duck went to the other bank of the river, swimming the little duck.’
The use of Verb Copying in Sentence 4.6(a) and 4.6(b) was not the focus of the current study, but I included the results here to ensure the comparison between sentences with transitive verbs and intransitive verbs was complete. The sentences with the Verb Copying or
Reduplication were controversial. An investigation of the raw data showed that 45% of the native speakers rated the sentences with transitive verbs as very unnatural or not very natural
(rating <0), an equal number of the native speakers, 45% as well, regarded this sentence as acceptable or very natural (rating >0), and the rest 10% found it hard to decide (rating =0). These contradictory interpretations brought the mean rating to down to – 0.08. On the contrary, the judgment on the intransitive counterpart was much more consistent. It was unanimously rejected, with a mean rating = -1.25. The ambiguous interpretations of the Sentence Type 5 with
Transitive Verbs as in 6(a) and the consistent rejection of the Type 5 (Verb Copying) Intransitive Verbs as in 4.6(b) led to another statistical 2.00 1.50 significant difference. A paired-samples t-test was 1.00 0.50 implemented to compare TRANSITIVITY in Sentence Type 0.00 -0.50 Transitive Intransitive 5 with Transitive Verbs and Intransitive Verbs (MANNER V zhe -1.00 +PP). Again, we found a significant difference in judgment -1.50 -2.00 ratings for Transitive Verbs(M= -0.08, SD= 0.72) Figure 4. 9. Effect of TRANSITIVITY on Sentence Type 5 and Intransitive Verbs (M= -1.25, SD= 0.56); t(39) 110
= 10.746, p< .001. (Table 4.9, Table 4.10 & Figure 4.8).
4.6. Type 5 (Verb Copying) a) 他搬椅子搬到了教室。 tā b ān y ǐzi b ān dào le jiàoshì he move chair move to classroom ‘He moved chairs to the classroom.’
b) *鸭妈妈游小鸭游到河对岸。 yā m āma yóu xi ǎoy ā yóu dào hé duìàn duck mom swim little duck swim to river opposite bank Intended meaning: ‘Mother Duck swam the little duck across to the other bank of the river.’
Unsurprisingly, the effect of TRANSITIVITY on Sentence Token Type 6, exemplified in
4.7(a) and 4.7(b), was also statistically significant in the Paired Samples t-test: Sentence Type 6 with Transitive Verbs (M= -1.49, SD= 0.49) and Intransitive Verbs (M= -1.82, SD= 0.30); t(39)
= 4.442, p< .001. (Table 4.9, Table 4.10 & Figure 4.9). The native speakers as a whole declined the sentences with both transitive and intransitive verbs, but the latter was regarded to be much more awkward than the former. Figure 4.10. Effect of TRANSITIVITY on Sentence Type 6
4.7. Type 6 (PP+ MANNER V zhe )
a) ?他到教室搬着椅子。 Type 6 (PP+ MANNER V tā dào jiàoshì b ān zhe y ǐzi zhe ) he to classroom move zhe chair 2.00 ‘He went to the classroom, moving chairs.’ 1.00 b) *鸭妈妈到河对岸游小鸭。 yā m āma dào hé duìàn yóu xi ǎoy ā 0.00 duck mom to river opposite bank swim little duck -1.00 Intended meaning: ‘ Mother Duck on the other bank of the river swam the little duck.’ -2.00
In this section, we have presented the six Paired Samples T-tests to examine the effect of
TRANSITIVITY on different Sentence Token Types. All these comparisons turned out to be 111
statistically significant. This confirms that there are constant differences between transitive and intransitive manner-of-motion verbs regardless of the syntactic environments these sentences occur in. Transitive manner-of-motion verbs behave in a distinctly different fashion from their intransitive counterparts with respect to ba, Verb Copying, or position of PPs, and the use of zhe .
Using the same methodology, I analyzed and verified the validity of other four variables.
The results of those variables are included in Appendix B-E. Readers who are interested to learn more about it are referred to the appendices. To summarize, I have taken several measures to examine the validity of the experiment instrument, specifically the picture judgment task, in measuring the five variables: TRANSITIVITY, GOAL, CAUSATIVITY, ACCOMPANY, and
COMPLEX MANNER. The general results from the paired samples t-tests or ANOVA are re- organized and re-presented again below in Table 4.19.
The results show that there is a statistically significant effect of TRANSITIVITY on native speakers’ judgment on the six sentence types (Type 1 –6). This indicates that whether the main verb is transitive determines the acceptability of the sentences. There is also a significant effect of GOAL on their judgment. The judgment results support the linguistic distinctions between pre-verbal and post-verbal prepositional phrases in the manner-of-motion VP. Only the latter can be used to indicate the GOAL of the motion.
Even though both Sentence Type 1( MANNER V + PP) and Type 2 ( ba ...+ MANNER V +PP) are grammatical in Chinese, native speakers give high preference to the second marked with Ba
Construction. This implies that Chinese native speakers tend to use the light verb ba to express
CAUSATIVITY.
112
Table 4.11. Complete List of Paired Samples t–tests or ANOVA on all variables
Sig. (2– Variables Paired Samples T–tests/ ANOVA t/ F df tailed) Sentence Type 1 15.189 39 0.000 * (Transitive vs Intransitive Verbs) Sentence Type 2 15.665 39 0.000 * (Transitive vs Intransitive Verbs) Sentence Type 3 3.703 39 0.000 * (Transitive vs Intransitive Verbs) TRANSITIVITY Sentence Type 4 17.254 39 0.000 * (Transitive vs Intransitive Verbs) Sentence Type 5 10.746 39 0.000 * (Transitive vs Intransitive Verbs) Sentence Type 6 4.442 39 0.000 * (Transitive vs Intransitive Verbs) Sentence Type 1 vs Type 3 12.154 39 0.000 * (with Transitive Verb) Sentence Type 1 vs Type 3 5.458 39 0.000 * (with Intransitive Verb) GOAL Sentence Type 4 vs Type 6 * (with Transitive Verb) 21.000 39 0.000 Sentence Type( 4 vs Type 6 5.420 39 0.000 * (with Intransitive Verb) Sentence Type 1 vs Type 2 –11.418 39 0.000 * (with Transitive Verb) CAUSATIVITY Sentence Type 1 vs Type 2 –3.887 39 0.000 * (with Intransitive Verb) Sentence Type 1 vs Type 4 0.551 39 0.585 (with Transitive Verb) Sentence Type 1 vs Type 4 1.289 39 0.205 (with Intransitive Verb) Sentence Type 3 vs Type 6 ACCOMPANY 1.941 39 0.059 (with Transitive Verb) Sentence Type 3 vs Type 6 2.924 39 0.006 * (with Intransitive Verb) Sentence Type 7 vs Type 8 –2.932 39 0.006 * (with Intransitive Verb) Sentence Type 1 vs Type 7 –16.243 39 0.000 * COMPLEX (with Intransitive Verb) MANNER Sentence Type 4 vs Type 8 –21.053 39 0.000 * (with Intransitive Verb)
*. The difference is significant at the 0.05 level. 113
In sentences with a transitive main verb, ACCOMPANY tends to be obviated, so whether the manner-of-motion verb is marked with zhe to accentuate the accompanied action does not make a big difference in the acceptability of the sentence. If an unmarked sentence with a transitive verb is appropriate and grammatical, the use of zhe will not significantly increase the
degree of acceptance; if an unmarked sentence with a transitive verb is inappropriate and
ungrammatical, the use of zhe will not save the sentence. When the main verb is intransitive, the factor of ACCOMPANY then becomes relevant. The use of zhe to highlight the accompanied action is actually preferred in the causativization of intransitive manner-of-motion verbs.
Specifically, of the two sentence types: Simple Sentence Type 4 ( MANNER Vzhe +PP) and
Complex Sentence Type 8 ( MANNER Vzhe + MOTIONV + PP), the latter is much preferred. In those ungrammatical sentences with intransitive manner-of-motion verbs, the use of zhe actually
significantly increases the rejection rate of such token types.
The test on COMPLEX MANNER explores the acceptability of causative directed
manner-of-motion English-like, in comparison with expressing such events with a decomposed
MANNER and MOTION. The significant differences obtained to some extent support the
validity of the judgment task. 114
CHAPTER 5
RESULTS AND DISCUSSIONS
5.1 Introduction
This chapter reports and analyzes L2 learners’ performance in the picture judgment task, which will help us understand how these learners acquire ways to describe causative directed manner-of-motion in the target language, Mandarin Chinese. The results will also reveal whether foreign learners of Chinese go through the process of “parameter resetting” or “feature assembly”. I begin this chapter by validating the reliability of the instrument and the group classification of the non-native speakers, who were divided into two groups based on their proficiency levels, as determined by two factors: their performance in a short proficiency test and the length of Chinese language instruction they had received by the time of the experiment
(Section 5.2). Then the focus of the chapter is to find out how L2 learners acquire knowledge of complex motion events with respect to five features (Section 5.3). Six questions will be raised and addressed.
1. How do English-speaking learners of Chinese acquire the expression of TRANSITIVITY?
2. How do English-speaking learners of Chinese acquire the expression of GOAL?
3. How do English-speaking learners of Chinese acquire the expression of CAUSATIVITY?
4. How do English-speaking learners of Chinese acquire the expression of ACCOMPANY?
5. How do English-speaking learners of Chinese acquire ways to license the notion of
COMPLEX MANNER when the motion verb is intransitive?
6. Is there a sequence in their acquisition pattern of the above variables? 115
After answering the first five questions with detailed statistical analyses, an overview of the results will be re-examined in Section 5.4 before the last question is considered.
5.2 The Instrument and Participants
In addition to the 40 native speakers, whose data were used to analyze the validity and
reliability of the experiment items covered in Chapter 4, a total number of 62 learners of Chinese
participated in the online experiment, but only 33 completed at least half of the experiment,
either Survey A or Survey B. 31 (94%) of the non-native speakers indicated that they were still
taking Chinese classes, while the rest confirmed that they had studied Chinese before, but were
not taking classes at the time of their participation in the experiment. 13 (40%) of them finished
only Survey A, 5 (15%) completed Survey B, and 15 (45%) completed the entire survey. Since
both Survey A and Survey B are comparably reliable, the data of these 33 learners are included
in the study.
In order to understand how learners of Chinese acquire the five elements of complex
motion events and see whether there is an acquisition sequence, it is essential to differentiate
between the language proficiency levels of the learners. In this study, learners’ language
proficiency was determined by two factors: the length of Chinese instruction these participants
have received and their’ performance in a short proficiency test. Before we get into details of the
data analysis, it is important to first validate the proficiency test. Only when the test is valid and
reliable can we use this measurement instrument to assess learners’ judgment and language
performance. 116
5.2.1 Validity and Reliability of the Proficiency Test
Instead of administering an independent test, the fillers in the experiment were utilized to evaluate learners’ language proficiency. This was an attempt to make the entire experiment a more natural and coherent process. The proficiency test consisted of eight multiple–choice questions. These questions were selected or edited from a placement test, which was originally developed by the Chinese instructors in the department of East Asian Studies at the University of
Arizona. The departmental placement test has been administered to hundreds of students over the past decades and it has proved to be a valid instrument for formative assessment purposes.
In addition to constructional validity, the reliability of the proficiency test was also checked. As in the test on the control grouop, a cut-off of 0.7 is an acceptable level of agreement. The proficiency test yielded an α value of 0.701. Given the small number of items involved, this test was accepted as a reliable measurement instrument.
5.2.2 Two Proficiency Levels: Correlation and Regression Analyses
Now that the validity and reliability of the proficiency test have been validated, the length of instruction was also factored into the determination of the proficiency level of the participant.
Most of the students had at least 3 semesters of Chinese language instruction at the time of the experiment. The number of semester was utilized to measure the length of language instruction.
Taking into consideration both learners’ performance in the proficiency test and the length of language instruction, the non-native speakers were divided into two groups: intermediate and advanced. To validate the grouping decision, three sets of correlation and regression analyses were conducted. 117
First, the length of instruction and learners’ test scores in the proficiency test were strongly correlated, r(29) = .401, p < .05. One outlier case was deleted from the sample based on the scatterplot of the results. Regression analysis was employed to test if the length of instruction significantly predicted learners’ grades in the proficiency test in this case. The results of the regression indicated that the predictor explained only a small portion (16.1%) of the variance,
2 R = . 161, F 1,29 =5.554, p< .05, but the length of instruction significantly predicted the learners’
performance grades in the proficiency test, β = .401, t(29) =3.642, p < .05. To put it in a simple way, the longer students received Chinese language instruction, the more likely will they perform better in the proficiency test.
Second, the length of instruction and the grouping of the learners were also significantly correlated, r(29) = .560, p = .001. The regression was a poor fit, describing only 31.4% of the
2 variance (R adj = 29.1%), but the overall relationship was statistically significant F 1,30 =13.716, p
= .001. The length of instruction also significantly predicted which group a learner would be
assigned to in this experiment, β = .560, t(30)= 3.800, p = .001. In other words, the longer students received Chinese language instruction, the more likely would they be placed in a higher proficiency level group.
Lastly, the test scores also significantly correlated with the grouping of the learners, r(29)
= .781, p < .001. The regression was a good fit and the learner’s test scores explained a
2 significant portion (61.1%) of variance in the proficiency level (R adj = 59.7%), F 1,30 =45.469, p
< .001. Learners’ test scores also significantly predicted their proficiency level, β = .781, t(30)=
2.502, p < .001. That is to say, the higher the score, the more likely it was that they would also be
placed in a higher proficiency level group. 118
Taking all these correlation and regression results collectively, these three factors, namely, the length of language instruction, the test score, and the final determination of learner’s proficiency level were strongly correlated with each other. It can be reasonably argued that the proficiency level of the participants was determined by the length of language instruction and that their test score in this study was an accurate indicator of their actual Chinese language proficiency.
5.3 Results of the Picture Judgment Task
This section will be devoted to the questions raised in Section 5.1, together with statistical analysis. I will address Chinese learners’ interlanguage development and the acquisition issues associated with the five variables in causative directed motion events, i.e.,
TRANSITIVITY, GOAL, CAUSATIVITY, ACCOMPANY, COMPLEX MANNER. The following sub-sections are organized in this way. The first five sub-sections (5.3.1 – 5.3.5) answer the first 5 questions. Section 5.3.6 is a summary of the results. The last question will be discussed in Section 5.4.
5.3.1 Acquisition of TRANSITIVITY
I will begin with the first research question, which concerns learners’ acquisition of
TRANSITIVITY. I analyzed the data using SPSS statistical package. A two-way repeated- measures analysis of variance (ANOVA) was conducted in order to evaluate judgments of subjects in each group. The independent variable, or between-subject factor, was the groups, and the dependent variable, or within-subject factor, was participants’ judgment ratings of the sentences with transitive and intransitive manner-of-motion verbs. A significance level of .05 was chosen since the population sample for the study was relatively small and there were an uneven number of participants in each group. 119
TABLE 5.1 Descriptive Statistics of Participants’ Judgment of TRANSITIVITY by Groups
Std. 95% Confidence Interval Proficiency TRANSITIVITY Mean Error Lower Bound Upper Bound Transitive 0.11 0.08 -0.04 0.26 Intermediate Intransitive -0.08 0.12 -0.31 0.15 Transitive 0.2 0.08 0.04 0.35 Advanced Intransitive -0.6 0.12 -0.84 -0.36
Native Transitive 0.01 0.05 -0.09 0.11 Speakers Intransitive -1.37 0.08 -1.52 -1.22
The mean scores on the judgment of transitive and intransitive verbs for each of the experimental groups were included in Table 5.1. The results showed that there was a significant main effect of verb types: F(1,70)=212.50, p < .001, which meant that there were significant
differences between the means of judgments on sentences with transitive and intransitive verbs.
The main effect of groups was also statistically significant: F(2,70)=28.96, p < .001, which indicated that the mean judgment scores amongst the three groups were very different from each other. There was a significant interaction of the verb types and groups: F(2,70)=48.42, p < .001.
The post hoc comparisons showed that there were significant differences between the native
speakers (M= -0.68, 95% CI [-0.79, -0.57]) and the non-native speakers at both the intermediate
level (M= 0.02, 95% CI [-0.15, 0.18]), p < .001 and the advanced level (M= -0.20, 95% CI [-
0.37, -0.03]), but no significant differences were identified within the two groups of the non- native speakers.
In addition, a significant interaction was observed when the participants’ judgment towards the two sentences in one group significantly varies from that of the other group(s). No 120
interaction was observed if the judgment of sentences made by one group is the same as the judgment made by the other groups. In this case, the significant interaction of the participant groups and verb types suggested different judgments towards sentences with transitive and intransitive verbs were held amongst the three groups. However, this significant interaction did not reveal which group differed from another, nor how much they varied. To address these questions and understand what contributed to the significant interaction, I conducted a series of follow-up analyses and post hoc comparisons.
TABLE 5.2 Multiple Comparisons of the Groups in Their Judgment of Intransitive Verbs
(I) Mean Std. 95% Confidence Interval (J) Proficiency Sig. Proficiency Diff (I-J) Error Lower Bound Upper Bound Advanced 0.52 * 0.16 0.007 0.13 0.91 Intermediate Native Speakers 1.29 * 0.14 0.000 0.96 1.62 Intermediate -0.52 0.16 0.007 -0.91 -0.13 Advanced Native Speakers 0.77 * 0.14 0.000 0.43 1.1
Native Intermediate -1.29 0.14 0.000 -1.62 -0.96 Speakers Advanced -0.77 0.14 0.000 -1.1 -0.43
Because of the significant interaction, tests of the simple effect of verb types or uses
(TRANSITIVITY) were conducted for each level of the participant groups. No significant effect of TRANSITIVITY was found for intermediate learners: F(1,16)=0.04, p = .845, which implied that intermediate learners were not able to differentiate transitive verbs from intransitive verbs in the complex motion events. However, there was a significant effect of TRANSITIVITY for both advanced learners (F(1,15) = 24.92, p < .001) and native speakers (F(1,39) = 719.66, p < .001). 121
That suggested that advanced learners, like the native speakers, were able to identify the distinction between transitive and intransitive manner-of-motion verbs.
Due to the significant interaction, additional tests of the simple effect of participant groups were also conducted for each type of verbs. There was no significant effect of participant groups for the transitive verbs: F(2,72)=2.23, p = .116, which indicated that the three groups
made similar judgments about transitive verbs. In contrast, the simple effect of participant groups
for the intransitive verbs turned out to be significant: F(2,72)=48.53, p < .001, suggesting that
the three groups made quite variant judgments about intransitive verbs. The post hoc
comparisons shown in Table 5.2 indicated that the judgment ratings of the native speakers on
sentences with intransitive verbs (M= -1.37, 95% CI [-1.52, -1.22]) were significantly lower than
those of the advanced learners (M= -0.6, 95% CI [-0.84, -0.36]), p < .001 and intermediate
learners (M= -0.08, 95% CI [-0.31, 0.15]), p < .001, the judgment ratings of the advanced
learners were also significantly lower than those of intermediate learners. In other words, in spite
of the apparent distinction between native and non-native speakers in their judgment of sentences
with intransitive verbs, advanced learners were able to make significantly more native-like
judgments than intermediate learners with respect to sentences with intransitive verbs.
Figure 5.1 provides a graphic representation of the participants’ judgment ratings of
sentences with transitive and intransitive verbs. In short, these three groups behaved distinctively
in their judgment of sentences with transitive and intransitive verbs that describe causative
directed manner of motion. The significant interaction of verbs (transitive vs intransitive) and
participant groups and the post hoc comparisons proved that native speakers held significantly
more varied judgments on TRANSITIVITY distinctions than the non-native speakers. The 122
advanced learners were not significantly better than the intermediate learners in their overall knowledge of TRANSITIVITY distinctions.
FIGURE 5.1 Judgment of TRANSITIVITY by Participant Groups
Further investigation revealed that intermediate learners did not have knowledge of the syntactic differences between transitive and intransitive manner-of-motion verbs in Chinese at all, whereas advanced learners were able to identify the TRANSITIVITY distinction. The follow-up analysis showed that if we just focused on the participants’ judgment of sentences with transitive verbs, the differences amongst the three groups were not very significant. The major group distinctions lay in their acceptability judgment of the sentences with intransitive verbs: native speakers differed significantly from non-native speakers, and advanced learners differed significantly from the intermediate learners. Even though the results presented in this section were based upon the mean average of all sentences with transitive verbs in comparison to all 123
sentences with intransitive verbs, the same acquisition patterns were found in almost every pair of analysis as illustrated in Table 5.3.
To answer the first research question, the results proved that TRANSITIVITY was difficult but definitely acquirable for English-speaking learners of Chinese. In this study, advanced learners were able to identify the differences between transitive and intransitive manner-of-motion verbs. The major learning hurdle was not so much in learning transitive directed manner-of-motion verbs, but more in learning the intransitive manner-of-motion verbs in Chinese.
124
TABLE 5.3 Summary of all Sub-group TRANSITIVITY Analyses Simple effect of Main Interaction Simple effect of Groups Main Factor Effect Factor Effect of Post-hoc of Groups Compariso M A NSs Transitive Post-hoc Intransitive Post-hoc Factor n TRANSITIVITY 1 Sig. Not sig. Sig. / Sig. Sig. Sig. Not sig. / Sig. NSs vs NNSs MANNER V + PP TRANSITIVITY 2 Not NSs vs M Sig. Sig. Sig. / Sig. Sig. Sig. Sig. NSs vs NNSs Ba+ MANNER V +PP sig. A vs M
TRANSITIVITY 3 Not Not NSs vs NSs vs M Sig. Sig. / Sig. Sig. Sig. Sig. PP + MANNER V sig. sig. NNSs A vs M
TRANSITIVITY 4 Not Not Sig. Not sig. / Sig. Sig. Sig. NSs vs M Sig. NSs vs NNSs MANNER Vzhe +PP sig. sig.
TRANSITIVITY 5 Not NSs vs M Sig. Sig. Sig. NSs vs M Sig. Sig. Sig. / Sig. Verb Copying sig. A vs M
TRANSITIVITY 6 Not Not NSs vs Sig. Sig. Sig. NSs vs M Sig. Sig. Sig. NSs vs NNSs PP+ MANNER Vzhe sig. sig NNSs
Abbreviations: NSs: Native Speakers; NNSs: Non-Native Speakers; M: Intermediate Learners; A: Advanced Learners 125
5.3.2 Acquisition of GOAL
Another two-way repeated-measures analysis of variance (ANOVA) was conducted to evaluate how the three groups interpret the pre- versus post-verbal prepositions in licensing the
GOAL of the motion events. As explained in earlier chapters, the GOAL information is realized through post-verbal prepositions in Chinese. The pre-verbal PPs have a different function, i.e. to indicate the general location of an event. The independent variable, or between-subject factor here, was the participant groups, and the dependent variable, or within-subject factor, was their judgment ratings of the sentences with pre-verbal and post-verbal PPs. The mean scores of the participants’ judgment of pre-verbal PPs and post-verbal PPs for each of the experimental groups are included in Table 5.4.
TABLE 5.4 Descriptive Statistics of Participants’ Judgment of GOAL by Groups
Std. 95% Confidence Interval Proficiency GOAL Mean Error Lower Bound Upper Bound Pre-verbal PPs -0.29 0.17 -0.64 0.06 Intermediate Post-verbal PPs 0.15 0.16 -0.16 0.47 Pre-verbal PPs -0.79 0.18 -1.15 -0.44 Advanced Post-verbal PPs -0.28 0.16 -0.6 0.05
Native Pre-verbal PPs -1.48 0.11 -1.71 -1.25 Speakers Post-verbal PPs -0.33 0.1 -0.53 -0.12
The results showed that there was a significant main effect of GOAL PPs: F(1,70)=
30.43, p < .001, which suggested that there were significant differences between the
participants’ mean acceptance with pre-verbal PPs and post-verbal PPs. The main effect of
groups was also statistically significant: F(2,70)= 19.28, p < .001, which meant the three groups 126
demonstrated large differences in their mean judgment scores of the PPs. The interaction of the prepositional phrases and groups was also statistically significant: F(2,70)= 4.12, p = .020. This significant interaction of the prepositional phrases and groups indicated that the three groups held different judgments towards the sentences with pre- and post-verbal PPs.
TABLE 5.5 Multiple Comparisons of the Groups in Their Judgment on GOAL PPs Mean (I) Std. 95% Confidence Interval (J) Proficiency Diff (I- Sig. Proficiency Error J) Lower Bound Upper Bound Advanced 0.47 * 0.16 .015 0.08 0.86 Intermediate Native Speakers 0.84 * 0.14 .000 0.51 1.16 Intermediate -0.47 * 0.16 .015 -0.86 -0.08 Advanced Native Speakers 0.37 * 0.14 .027 0.04 0.70 * Native Intermediate -0.84 0.14 .000 -1.16 -0.51 Speakers Advanced -0.37 * 0.14 .027 -0.70 -0.04 Based on observed means. The error term is Mean Square(Error) = .222. *. The mean difference is significant at the .05 level.
The post hoc comparisons of the participants’ judgment on the prepositional phrases (Table 5.5) indicated that the judgment ratings of the native speakers on pre- vs post-verbal PPs for licensing
GOAL of the motion events (M= -0.90, 95% CI [-1.05, -0.76]) were significantly lower than those of the advanced learners (M= -0.54, 95% CI [-0.77, -0.30]), p < .05 and intermediate learners (M= -0.07, 95% CI [-0.30, 0.16]), p < .001, the judgment ratings of the advanced
learners were also significantly lower than those of intermediate learners, p < .05. This meant that there were distinctions between native and non-native speakers in their judgment of the two prepositional phrases for representing the GOAL of the motion events. Advanced learners again 127
showed significantly more native-like judgments than intermediate learners with respect to the two PPs.
To further explicate what caused the significant interaction of the groups and GOAL PPs, tests of the simple effect of the GOAL PPs were conducted for each level of the participant groups. No significant effect of the two PPs was found for intermediate learners: F(1,16)=4.21, p = .057, which implied that intermediate learners were not able to differentiate pre-verbal prepositional phrases from post-verbal prepositional phrases in representing the GOAL of the complex motion events. There was a significant effect of the two PPs for both advanced learners
(F(1,15) = 24.92, p < .001) and native speakers (F(1,39) = 719.66, p < .001). That suggested that the advanced learners made similar judgments as native speakers did and were able to distinguish the use of pre-verbal PPs from that of post-verbal PPs in directed motion events.
TABLE 5.6 Multiple Comparisons of the Groups in Their Judgment on Pre-verbal PPs
(I) Mean Std. 95% Confidence Interval (J) Proficiency Sig. Proficiency Diff (I-J) Error Lower Bound Upper Bound Advanced 0.51 0.25 .114 -0.09 1.10 Intermediate Native Speakers 1.19 * 0.21 .000 0.70 1.69 Intermediate -0.51 0.25 .114 -1.10 0.09 Advanced Native Speakers 0.69 * 0.21 .005 0.18 1.19 * Native Intermediate -1.19 0.21 .000 -1.69 -0.70 Speakers Advanced -0.69 * 0.21 .005 -1.19 -0.18 *. The mean difference is significant at the .05 level.
Additionally, tests of the simple effect of the participant groups were also performed for both pre-verbal PPs and post-verbal PPs. Both simple effects were found to be statistically 128
significant: for the pre-verbal PPs, F(2,72)= 17.83, p < .001, and also for the post-verbal PPs,
F(2,72)= 3.36, p < .05, implying that all three groups made significantly varied judgments on sentences with pre-verbal PPs and those with post-verbal PPs. The post hoc comparisons revealed that in the case of pre-verbal PPs (Table 5.6), the native speakers (M= -1.48, 95% CI [-
1.71, -1.25]) made distinctive judgments from non-native speakers at both an advanced proficiency level (M= -0.79, 95% CI [-1.15, -0.44]), p = .005, and at an intermediate level (M= -
0.29, 95% CI [-0.64, 0.06]), p < .001. No significant differences were identified between the intermediate and advanced learners, p = .114. This meant that even the advanced learners did not fully acquire the use of pre-verbal PPs.
TABLE 5.7 Multiple Comparisons of the Groups in Their Judgment on Post-verbal PPs
(I) Mean Std. 95% Confidence Interval (J) Proficiency Sig. Proficiency Diff (I-J) Error Lower Bound Upper Bound Advanced 0.43 0.23 .149 -0.12 0.98 Intermediate Native Speakers 0.48 * 0.19 .035 0.03 0.94 Intermediate -0.43 0.23 .149 -0.98 0.12 Advanced Native Speakers 0.05 0.19 .963 -0.41 0.52 * Native Intermediate -0.48 0.19 .035 -0.94 -0.03 Speakers Advanced -0.05 0.19 .963 -0.52 0.41 *. The mean difference is significant at the .05 level.
In the post hoc multiple-comparisons of groups on their judgment of post-verbal PPs (see
Table 5.7), a different pattern emerged. There were no obvious differences between the advanced learners (M= -0.28, 95% CI [-0.60, 0.05]) and the native speakers (M= -0.33, 95% CI [-0.53, -
0.12]), p = .963, or between the advanced and intermediate learners (M= 0.15, 95% CI [-0.16, 129
0.47]), p = .149, in their acceptability of the post-verbal PPs as the GOAL in the motion events.
However, there was a significant difference between native speakers and the intermediate learners, p < .05. The results suggested that advanced learners did not have as much knowledge of the post-verbal PPs as native speakers, but they were significantly more knowledgeable than the intermediate learners.
FIGURE 5.2 Judgment of GOAL by Participant Groups
Graphic presentation in Figure 5.2 demonstrated the participants’ distinctive acceptability
judgment of the pre- versus post-verbal preposition phrases that signal the GOAL of a directed
motion. In addition to the significant differences between native speakers and non-native
speakers, the advanced learners made significantly more native-like judgments than intermediate
learners in distinguishing the two PPs. In other words, advanced learners did not achieve the
same intuitive level as native speakers regarding the different semantic meanings and functions 130
of pre- versus post-verbal PPs in the context of directed manner-of-motion events, but they were definitely more knowledgeable than the intermediate learners.
Even though the advanced learners were already aware of the differences between the two PPs, when we focused on their performance in judging each PP, it turned out that the advanced learners were only able to acquire a better intuition about post-verbal PPs. While they were probably more knowledgeable about pre-verbal PPs than their counterparts in the intermediate proficiency level, this difference was not statistically significant. The intermediate learners were not able to recognize the differences between the pre- and post-verbal PPs in licensing the GOAL for a directed manner-of-motion event. However, they did acquire more knowledge about post-verbal PPs. The results also indicated that the major learning hurdle they faced was mainly concerned with the use of the pre-verbal PPs.
5.3.3 Acquisition of CAUSATIVITY
In this section, we will use the same two-way repeated-measures ANOVA to understand how the non-native learners acquire the lexicalization of CAUSATIVITY in directed motion events in comparison to the native speakers in the control group. As explained in Chapter 2,
Chinese differs from English in that it prefers to use a Ba Construction to license a causative event. We will first compare the three groups’ interpretation of Sentence Type 1 ( MANNER V + PP)
and Sentence Type 2 ( ba ...+ MANNER V +PP) with transitive verbs, and then we will compare that with intransitive verbs. The independent variable was the three participant groups as in the previous two sections, and the dependent variable, or within-subject factor, was their acceptability judgment of Sentence Type 1 and Sentence Type 2. In the former, Lexical
CAUSATIVITY is allowed in English for sentences with both transitive and intransitive verbs; 131
in the latter, syntactic CAUSATIVITY with ba Construction is only appropriate for transitive verbs in Chinese.
5.3.3.1 Acquisition of CAUSATIVITY with Transitive Verbs
This section will be dedicated to analyzing the effect of CAUSATIVITY with transitive
verbs. When the main motion verb is transitive in nature, both Sentence Type 1 ( MANNER V + PP) and Sentence Type 2 ( ba ...+ MANNER V +PP) are acceptable in Chinese. However, as indicated by
the mean scores in Table 5.8, the latter sentence type with a ba construction is highly preferred.
On the other hand, the mean acceptability by native speakers is 0.59, which implies that it
considered barely acceptable. A similar pattern is found in advanced learners’ data.
TABLE 5.8 Descriptive Statistics of Participants’ Judgment of CAUSATIVITY by Groups (with Transitive Verbs) 95% Confidence Interval Std. Proficiency CAUSATIVITY Mean Error Lower Upper Bound Bound Sentence Type 1 (no Ba ) 0.37 0.20 -0.03 0.76 Intermediate Sentence Type 2 (with Ba ) 0.77 0.12 0.53 1.00 Sentence Type 1 (no Ba ) 0.13 0.20 -0.27 0.54 Advanced Sentence Type 2 (with Ba ) 1.61 0.12 1.37 1.85
Native Sentence Type 1 (no Ba ) 0.59 0.13 0.34 0.85 Speakers Sentence Type 2 (with Ba ) 1.81 0.08 1.66 1.96
The results from the two-way repeated measures ANOVA indicated that there was a significant main effect of CAUSATIVITY: F(1,70)= 86.82, p < .001, which meant that there were significant differences between the participants’ mean acceptance of Sentence Type 1
(MANNER V + PP) and Sentence Type 2 ( ba ...+ MANNER V +PP) when the main verbs was transitive. 132
The main effect of groups was also statistically significant: F(2,70)= 9.86, p < .001, which attested that the three groups demonstrated different interpretations of the two sentences for
CAUSATIVITY. The interaction of the two causative sentences and the three participant groups was also statistically significant: F(2,70)= 7.45, p < .001. This significant interaction suggested that the three groups gave different acceptability judgments towards the sentences representing
CAUSATIVITY, with or without ba.
TABLE 5.9 Multiple Comparisons of the Groups in Their Judgment on CAUSATIVITY (with Intransitive Verbs)
(I) Mean Std. 95% Confidence Interval (J) Proficiency Sig. Proficiency Diff (I-J) Error Lower Bound Upper Bound Advanced -0.30 0.18 0.204 -0.73 0.12 Intermediate Native Speakers -0.64 * 0.15 0.000 -0.99 -0.28 Intermediate 0.30 0.18 0.204 -0.12 0.73 Advanced Native Speakers -0.33 0.15 0.077 -0.69 0.03 * Native Intermediate 0.64 0.15 0.000 0.28 0.99 Speakers Advanced 0.33 0.15 0.077 -0.03 0.69 Based on observed means. The error term is Mean Square(Error) = .258. *. The mean difference is significant at the .05 level.
The post hoc comparisons of the participants’ judgment on the two sentence types in licensing CAUSATIVITY (Table 5.9) indicated that the native speakers differ significantly (M=
1.20, 95% CI [1.04, 1.36]) from the intermediate learners (M= 0.57, 95% CI [0.32, 0.81]), p <
.001. No significant differences were found between the native speakers and the advanced
learners, or between the two non-native speaker groups. It was evident that the advanced learners 133
behaved more like the native speakers, but the results only provided us limited information as to how well the two non-native speaker groups actually understood the two sentences types.
To further explicate what contributed to the significant interaction of the groups and
CAUSATIVITY, tests of the simple effect of the two sentence types were conducted for each level of the participant groups. No significant effect of CAUSATIVITY was found for intermediate learners: F(1,16)= 2.28, p = .151, which implied that intermediate learners did not provide very different judgments toward Sentence Type 1 and Sentence Type 2 with transitive verbs. In other words, whether or not Ba was used in licensing the CAUSATIVITY of the complex motion events did not affect their acceptability judgment of the sentences. There was a significant effect of CAUSATIVITY for the advanced learners: F(1,15) = 32.50, p < .001, the
same for the native speakers (F(1,39) = 130.37, p < .001. The advanced learners made similar
judgments to the native speakers and preferred to use Sentence Type 2 with Ba construction to
indicate the causative nature of the complex motion event.
Additionally, tests of the simple effect of the participant groups were also performed for
both Sentence Type 1 ( MANNER V + PP) and Sentence Type 2 ( ba ...+ MANNER V +PP) with transitive
verbs. The simple effect of the former was not statistically significant: F(2,72)= 1.92, p = .155,
but the simple effect of the latter was significant: F(2,72)= 28.99, p < .001. This result indicated
that the three groups were similar in their judgment of the Sentence Type 1 ( MANNER V + PP), but
they made significantly different judgments towards sentences with Ba construction. 134
TABLE 5.10 Multiple Comparisons of the Groups in Their Judgment on Sentence Type 1 with Transitive Verbs for Causative Event with Ba Construction
(I) Mean Diff Std. 95% Confidence Interval (J) Proficiency Sig. Proficiency (I-J) Error Lower Bound Upper Bound Advanced -0.85* 0.17 0.000 -1.24 -0.45 Intermediate Native Speakers -1.05* 0.14 0.000 -1.38 -0.72 Intermediate 0.85* 0.17 0.000 0.45 1.24 Advanced Native Speakers -0.20 0.14 0.328 -0.54 0.14
Native Intermediate 1.05* 0.14 0.000 0.72 1.38 Speakers Advanced 0.20 0.14 0.328 -0.14 0.54 *. The mean difference is significant at the .05 level.
The post hoc comparisons revealed that in terms of representing a causative motion event
(Table 5.10), the native speakers (M= 1.81, 95% CI [1.71, 1.92]) were significantly more fond of
Sentence Type 2 with Ba Construction than the intermediate learners (M= 0.76, 95% CI [0.36,
1.17]), p < .001. The same significant preference was also demonstrated by the advanced learners
(M= 1.61, 95% CI [1.42, 1.80]) compared to the intermediate learners. It was very clear that the advanced learners had already acquired and preferred to use Ba construction with transitive verbs in describing a causative motion event, whereas the intermediate learners had only started to gain a sense of its importance(M= 0.76, 95% CI [0.36, 1.17]).
Figure 5.3 illustrated the average acceptability judgment of the three participant groups with respect to the two sentence types with transitive verbs and their realization of the
CAUSATIVITY of the complex motion event. In summary, the significant interaction of 135
FIGURE 5.3 Judgment of CAUSATIVITY with Transitive Verbs by Participant Groups
CAUSATIVITY and the participant groups and the post hoc comparisons indicated that the advanced learners behaved more like the native speakers. In contrast, the intermediate learners behaved significantly differently from the native speakers. Further exploration of the simple effects suggested that intermediate learners failed to make significantly different judgments when rating Sentence Type 1 ( MANNER V + PP) and Sentence Type 2 ( ba ...+ MANNER V +PP), and they displayed similar acceptance to both the sentence types. Advanced learners, on the other hand, were clearly aware of the differences between the two sentence types in licensing
CAUSATIVITY. The advanced learners, like native speakers, showed much preference to the sentence with ba construction for the causative motion event. In other words, advanced learners have already acquired the use of ba construction for CAUSATIVITY when the main verb was transitive. In the next section, we will look at their judgment towards similar sentence types, but 136
with intransitive verbs. This will inform us as to how exactly the learners interpret the ba construction with intransitive verbs.
5.3.3.2 Acquisition of CAUSATIVITY with Intransitive Verbs
Will the transitivity nature of the verb change learners’ judgments about the two sentence types? In other words, will the learners develop different acceptability judgments regarding different ways of expressing CAUSATIVITY when the main verb is intransitive? This section will focus on the effect of CAUSATIVITY with intransitive verbs. As discussed in Chapter 2, unlike English, intransitive motion verbs in Chinese cannot participate in syntactic causativization. Other strategies are used instead; for example, the use of zhe to mark
accompanied action, the use of causative verbs shi or rang . Since causative expressions are
available with both transitive and intransitive verbs in English, this cross-linguistic difference is
expected to interfere with the learners’ acquisition process.
TABLE 5.11 Descriptive Statistics of Participants’ Judgment of CAUSATIVITY by Groups (with Intransitive Verbs) 95% Confidence Interval Std. Proficiency CAUSATIVITY Mean Error Lower Upper Bound Bound Sentence Type 1 (no Ba ) -0.16 0.16 -0.48 0.15 Intermediate Sentence Type 2 (with Ba ) 0.56 0.23 0.10 1.02 Sentence Type 1 (no Ba ) -0.65 0.16 -0.97 -0.32 Advanced Sentence Type 2 (with Ba ) 0.63 0.24 0.15 1.10
Native Sentence Type 1 (no Ba ) -1.21 0.10 -1.42 -1.01 Speakers Sentence Type 2 (with Ba ) -0.79 0.15 -1.09 -0.49
137
Table 5.11 provides the mean scores of their acceptability judgments, from which we can tell that native speakers rejected both sentences with intransitive verbs, but the use of ba construction made Sentence Type 2 sound less awkward than Sentence Type 1. Both English- speaking learner groups showed a tendency of accepting the causativized sentences (with intransitive verbs) with the use of ba construction .
The two-way repeated measures ANOVA indicated that there was a significant main effect of CAUSATIVITY with intransitive verbs: F(1,70)= 56.25, p < .001, implying that there were significant differences between the participants’ mean acceptance of Sentence Type 1
(MANNER V + PP) and Sentence Type 2 ( ba ...+ MANNER V +PP) with intransitive verbs. The main
effect of groups was also statistically significant: F(2,70)= 22.99, p < .001, which meant the
three groups demonstrated different interpretations of the two sentences with intransitive verbs
for CAUSATIVITY. The interaction of the two causativized sentences with intransitive verbs
and the three participant groups was also statistically significant: F(2,70)= 5.81, p = .005.
TABLE 5.12 Multiple Comparisons of the Groups in Their Judgment on CAUSATIVITY (with Intransitive Verbs)
(I) Mean Std. 95% Confidence Interval (J) Proficiency Sig. Proficiency Diff (I-J) Error Lower Bound Upper Bound Advanced 0.21 0.24 0.661 -0.37 0.79 Intermediate Native Speakers 1.20 0.20 0.000 0.72 1.68 Intermediate -0.21 0.24 0.661 -0.79 0.37 Advanced Native Speakers 0.99 0.21 0.000 0.50 1.48 -1.20 0.20 0.000 -1.68 -0.72 Native Intermediate Speakers Advanced -0.99 0.21 0.000 -1.48 -0.50 Based on observed means. The error term is Mean Square(Error) = .481. *. The mean difference is significant at the .05 level. 138
The post hoc comparisons of the participants’ judgment on the two sentences with
intransitive verbs types in licensing CAUSATIVITY (Table 5.12) indicated that the native
speakers differed significantly (M= -1.00, 95% CI [-1.22, -0.78]) from the English-speaking
learners at the intermediate level (M= 0.20, 95% CI [-0.14, 0.53]), p < .001 and at the advanced
level (M= -0.01, 95% CI [-0.36, 0.33]), p < .001. No significant differences were found between
the two non-native speaker groups.
TABLE 5.13 Multiple Comparisons of the Groups in Their Judgment on Sentence Type 1 (with Transitive Verbs) for Causative Event without Ba Construction
(I) Mean Std. 95% Confidence Interval (J) Proficiency Sig. Proficiency Diff (I-J) Error Lower Bound Upper Bound Advanced 0.49 0.23 .089 -0.06 1.03 Intermediate Native Speakers 1.05 * 0.19 .000 0.60 1.50 Intermediate -0.49 0.23 .089 -1.03 0.06 Advanced Native Speakers 0.56 * 0.19 .013 0.10 1.03 * Native Intermediate -1.05 0.19 .000 -1.50 -0.60 Speakers Advanced -0.56 * 0.19 .013 -1.03 -0.10 *. The mean difference is significant at the .05 level.
To further understand what contributed to the significant interaction of the groups and
CAUSATIVITY in sentences with intransitive verbs, tests of the simple effect of the two sentence types were conducted for each of the participant groups. Significant effect of
CAUSATIVITY was found for intermediate learners: F(1,16)= 8.75, p = .009, which implied that intermediate learners assigned different ratings toward Sentence Type 1 and Sentence Type
2 with intransitive verbs. In other words, the intermediate learners were able to identify the 139
differences between using and not using the Ba construction in licensing CAUSATIVITY of complex motion events. The same significant effect was also observed for the advanced learners:
F(1,15) = 25.99, p < .001, and for the native speakers (F(1,39) = 15.11, p < .001.
TABLE 5.14 Multiple Comparisons of the Groups in Their Judgment on Sentence Type 2 (with Intransitive Verbs) for Causative Event with Ba Construction
(I) Mean Diff Std. 95% Confidence Interval (J) Proficiency Sig. Proficiency (I-J) Error Lower Bound Upper Bound Advanced -0.07 0.33 .978 -0.85 0.72 Intermediate Native Speakers 1.35 * 0.27 .000 0.69 2.00 Intermediate 0.07 0.33 .978 -0.72 0.85 Advanced Native Speakers 1.41 * 0.28 .000 0.74 2.08 * Native Intermediate -1.35 0.27 .000 -2.00 -0.69 Speakers Advanced -1.41 * 0.28 .000 -2.08 -0.74 *. The mean difference is significant at the .05 level.
Additionally, tests of the simple effect of the participant groups were also performed for both Sentence Type 1 ( MANNER V + PP) and Sentence Type 2 ( ba ...+ MANNER V +PP) where the main verbs were intransitive. The simple effect of groups was statistically significant for
Sentence Type 1, F(2,72)= 16.42, p < .001. This implied that the three groups differed from each
other in their judgment of Sentence Type 1. The post hoc comparisons (see Table 5.13) revealed
that the native speakers (M= -1.21, 95% CI [-1.42, -1.01]) were significantly more likely to reject
the transitivization (or transitive use) of an intransitive motion verb than both the intermediate
learners (M= -0.16, 95% CI [-0.48, 0.15]), p < .001, or the advanced learners (M= -0.65, 95% CI
[-0.97, -0.32]), p < .05. No significant differences were identified between the intermediate and 140
advanced learners in their acceptability judgment of Sentence Type 1, p =. 089. In short, the native speakers rejected this sentence type because to use an intransitive motion verb in a transitive syntactic structure was not grammatical in Chinese. Neither of the two learner groups had fully acquired this knowledge.
The simple effect of groups also turned out to be statistically significant for Sentence
Type 2 with the ba construction, F(2,72)= 19.28, p < .001. This implied that the three groups
differed from each other in their judgment of Sentence Type 2 (with intransitive verbs). To find
out how much each mean varied from another, post hoc comparisons were conducted. Table
5.14 revealed that even though the use of the ba construction to express causation was not judged
as bad as the transitive use of an intransitive verb in Sentence Type 1, the native speakers (M= -
0.99, 95% CI [-1.48, -0.50]) still rejected Sentence Type 2 more often than both the intermediate
learners (M= 0.56, 95% CI [0.10, 1.02]), p < .001, and the advanced learners (M= 0.63, 95% CI
[0.15, 1.10]), p < .001. No significant differences were identified between the intermediate and advanced learners in their acceptability judgment of the Sentence Type 2, p =. 978.
Figure 5.4 illustrated the average acceptability judgment of the three participant groups with respect to the two sentence types with intransitive verbs that express CAUSATIVITY.
Overall, the post hoc comparisons indicated that the significant interaction of CAUSATIVITY and the participant groups mostly resulted from the distinction between the native speaker group and the two non-native speaker groups. Further exploration of the simple effects suggested that intermediate learners, like the advanced learners, started to realize the differences between
Sentence Type 1 ( MANNER V + PP) and Sentence Type 2 ( ba ...+ MANNER V +PP) where the main verb was intransitive, but both of them failed to understand that intransitive motion verbs in
Chinese cannot participate in transitivization or the 141
FIGURE 5.4 Judgment of CAUSATIVITY with Intransitive Verbs by Participant Groups
causativization process. They were aware of the differences between the two sentence types in licensing CAUSATIVITY, but they did not realize that intransitive motion verbs behaved distinctively from those in English, and they overgeneralized the pattern from L1 to the target language. In order to make sure that these results are an accurate description of what was happening in the learners’ interlanguage development, in the next section, we will utilize their judgment on Sentence Type 2 (with transitive versus intransitive verbs) with the ba construction
for triangulation purposes.
5.3.3.3 Acquisition of the Light Verb Ba
Sentence Type 2 contains the Ba construction ( ba ...+ MANNER V +PP), which is a unique syntactic structure in Mandarin Chinese. It can be used to express CAUSATIVITY of complex 142
motion events. To learn when and how to use ba construction properly has been a challenge for most English-speaking learners of Chinese. This is not only because there is no equivalent in their L1 input, but also because of the complexity involved in the structure itself.
In the previous two sections, we have obtained conflicting results on whether the intermediate learners realized the differences in the use of ba construction for CAUSATIVITY.
In Section 5.3.1, the intermediate learners did not assign significantly different scores to the sentence with a ba construction for CAUSATIVITY when the verb was transitive; however, in
Section 5.3.2, when the verb was intransitive, they rated the sentence with a ba construction as
being significantly more favored than the one without the ba construction. In an attempt to
understand whether the intermediate learners were really able to identify the use of ba construction in realization of CAUSATIVITY, two tests were implemented: first, we examined how the intermediate learners of Chinese interpreted the use of ba construction with both transitive and intransitive verbs; second, we collapsed the data together, taking and averaging all students’ ratings for Sentence Type 1 (without ba construction), and then compared this mean with that of Sentence Type 2 (with ba construction).
The first measure was originally one sub-group analysis of TRANSITIVITY covered in
Section 5.3.2, but the focus here will concentrate heavily on the performance of intermediate learners. The means of all three groups are provided in Table 5.15 and graphically presented in
Figure 5.5. Apparently, the similar mean values demonstrate that the acceptance levels of
Sentence Type 2 by the intermediate learners were very close.
143
TABLE 5.15 Descriptive Statistics of Participants’ Judgment on Sentences with Ba construction 95% Confidence Interval Std. Proficiency CAUSATIVITY Mean Upper Error Lower Bound Bound Sentence Type 2 0.77 0.12 0.53 1.00 (with Transitive Verb) Intermediate Sentence Type 2 0.56 0.23 0.10 1.02 (with Intransitive Verb) Sentence Type 2 1.61 0.12 1.37 1.85 (with Transitive Verb) Advanced Sentence Type 2 0.63 0.24 0.15 1.10 (with Intransitive Verb) Sentence Type 2 1.81 0.08 1.66 1.96 Native (with Transitive Verb) Speakers Sentence Type 2 -0.79 0.15 -1.09 -0.49 (with Intransitive Verb)
To understand the statistical significance of the seemingly similar ratings, the data were once again analyzed by using a two-way repeated-measure ANOVA, with the three participant groups as a between-subjects factor and their acceptance ratings of the sentences (Sentence Type
2 with Transitive Verbs or Intransitive Verbs) as a within-subjects factor. The main effect of
Sentence Type 2 (F(1,70)= 99.20, p < .001) and the three Participant Groups (F(2,70)= 6.69, p =
.002) were both significant. The interaction of Sentence Type 2 and the Participant Groups was
also significant: F(2,70) = 39.67, p < .001. This significant interaction indicated that the
participants in these three groups showed different preferences for the two sentences. The post-
hoc comparisons of the groups did not reveal anything particularly interesting. In other words,
the rating of one group was not significantly different from another in the post-hoc comparisons.
144
FIGURE 5.5 Judgment of Sentences with Ba Construction by Participant Groups
2.00 1.80 1.60 1.40 1.20 1.00 0.80 0.60 0.40 0.20 0.00 -0.20 Intermediate Advanced Learners Native Speakers -0.40 Learners -0.60 Sentence Type 2 (Transitive Verb)
Because of the significant interaction, tests of the simple effect of Sentence Type 2 were conducted for each participant group. The simple effect of Sentence Type 2 was significant for both the advanced learners (F(1,15) = 10.54, p = .005) and the native speakers (F(1,39) = 245.39,
p < . 001). However, there was no significant effect of Sentence Type 2 for the intermediate
learners (F(1,16) = 2.71, p = .119). In other words, statistically the intermediate learners assigned
similar acceptability ratings to Sentence Type 2 both with transitive and intransitive verbs. The
rest of this analysis will not be discussed in full detail here. Consult Table 5.3 for more
information.
The second measure was based on the mean of sentences with or without ba construction .
We averaged all students’ ratings for Sentence Type 1 (without ba construction), and compared 145
them with those of Sentence Type 2 (with ba construction). The simple effect of
TRANSITIVITY for intermediate learners was not significant, F(1,16) = 4.28, p = .055, which meant that the intermediate learners held similar judgment for sentences with or without ba construction . Taking all the results (Section 5.3.4.1 – Section 5.3.4.3) into consideration, the
results suggested that intermediate learners were not able to identify the CAUSATIVITY
differences between Sentence Type 1 (without ba construction) and Sentence Type 2 (with ba
construction).
5.3.4 Acquisition of ACCOMPANY
In this section, we will use the same two-way repeated-measures ANOVA to understand how the non-native learners acquire the use of zhe , which lexicalizes ACCOMPANY in the
complex motion events. In order to avoid confusion with respect to learners’ acquisition of zhe
and their acquisition of intransitive verbs, we will concentrate on the three groups’ interpretation
of merely Sentence Type 1 ( MANNER V + PP) and Sentence Type 4 ( MANNER V + zhe +PP) with
transitive verbs as the main verbs. The use of zhe in sentences with intransitive verbs will be touched upon in the next section instead. Generally speaking, it is optional to use zhe to license
ACCOMPANY in Chinese. For instance, both sentences in (5.1) and (5.2) are grammatical and acceptable to native speakers, but they have slightly different emphasis.
5.1. 他搬椅子到教室。 Tā b ān y ǐzi dào jiàoshì he move chair DAO classroom ‘He moved the chairs to the classroom.’
5.2. 他搬着椅子到教室。 Tā b ān zhe y ǐzi dào jiàoshì he move ZHE chair DAO classroom ‘He moving chairs to the classroom.’
146
How learners of Chinese interpret the subtle differences between the two sentences is the focus of this section. The independent variable, as that in the previous two sections, was the three participant groups, and the dependent variable, or within-subject factor, was their acceptability judgment of Sentence Type 1 and Sentence Type 4. From the mean ratings of the native speakers, it is obvious that both Sentence Type 1 (MANNER V…+ PP) and Sentence Type 4
(MANNER V zhe +PP) are acceptable in Chinese. As indicated by their mean scores in Table 5.16,
the acceptability of these two sentence types was quite similar: M (Type 1) =0.59 and M (Type 4) =0.51.
TABLE 5.16 Descriptive Statistics of Participants’ Judgment of ACCOMPANY by Groups (with Transitive Verb) 95% Confidence Interval Std. Proficiency ACCOMPANY Mean Error Lower Upper Bound Bound Sentence Type 1 (no zhe ) 0.37 0.20 -0.03 0.76 Intermediate Sentence Type 4 (with zhe) -0.05 0.18 -0.40 0.30 Sentence Type 1 (no zhe ) 0.13 0.20 -0.27 0.54 Advanced Sentence Type 4 (with zhe) 0.05 0.18 -0.31 0.42
Native Sentence Type 1 (no zhe ) 0.59 0.13 0.34 0.85 Speakers Sentence Type 4 (with zhe) 0.51 0.11 0.28 0.73
The results from the two-way repeated measures ANOVA indicated that the main effect of ACCOMPANY was not significant: F(1,70)= 2.89, p = .093, which meant that no obvious differences existed between the participants’ mean acceptance of Sentence Type 1 ( MANNER V +
PP) and Sentence Type 4 ( MANNER V zhe +PP). However, the main effect of groups was statistically significant: F(2,70)= 4.16, p < .020, which implied that the average judgment ratings
of the three groups for ACCOMPANY were distinctive from each other. The interaction of the
two sentence types and the three participant groups was not significant statistically: F(2,70)= 147
0.90, p = .41. This suggested that the mean judgment ratings for the two sentence types were quite similar across the three groups.
To further understand the learners’ interpretation of the two sentence types marked with
or without zhe , tests of the simple effect of the two sentence types were conducted for each level
of the participant groups. No significant effect of ACCOMPANY was found for advanced
learners, F(1,15) = 0.21, p = .655, or the native speakers, F(1,39) = 0.30, p = .585, which implied
that both groups assigned similar judgment toward Sentence Type 1 and Sentence Type 4. In
other words, the use or non-use of zhe in licensing the ACCOMPANY of the complex motion
events did not affect the acceptability of the sentences. There was a significant effect of
ACCOMPANY for the intermediate learners: F(1,16) = 4.69, p < .050. The intermediate learners
had a tendency to accept Sentence Type 1 but reject Sentence Type 4, which revealed that they
did not understand or recognize sentences marked with zhe .
Additionally, tests of the simple effect of the participant groups were also performed for
both Sentence Type 1 ( MANNER V + PP) and Sentence Type 4 ( MANNER V zhe +PP). The simple effect of the former was not statistically significant: F(2,72)= 1.92, p = .155, but the simple effect of the latter was significant: F(2,72)= 4.50, p < .050. This result indicated that the three groups gave similar judgment to the Sentence Type 1 ( MANNER V + PP), but their judgment on sentences with zhe was quite distinctive. The post hoc comparisons revealed in Table 5.17 demonstrated that the only significant differences amongst the three groups in their acceptability judgment of
Sentence Type 4 was between the native speakers (M= 0.51, 95% CI [0.33, 0.69]) and the intermediate speakers (M= - 0.05, 95% CI [-0.51, 0.41]), p < .050. No significant differences
were found between the native speakers and the advanced learners (M= 0.05, 95% CI [-0.42, 148
0.53]), p= .096, or between the advanced learners and the intermediate learners, p= .907, in the acceptability judgment of Sentence Type 4.
TABLE 5.17 Multiple Comparisons of the Groups in Their Judgment on Sentence Type 4 for ACCOMPANIED Actions with Zhe
(I) Mean Std. 95% Confidence Interval (J) Proficiency Sig. Proficiency Diff (I-J) Error Lower Bound Upper Bound Advanced -0.11 0.25 .907 -0.71 0.50 Intermediate Native Speakers -0.56 * 0.21 .026 * -1.06 -0.06 Intermediate 0.11 0.25 .907 -0.50 0.71 Advanced Native Speakers -0.45 0.21 .096 -0.97 0.06 * * Native Intermediate 0.56 0.21 .026 0.06 1.06 Speakers Advanced 0.45 0.21 .096 -0.06 0.97 *. The mean difference is significant at the .05 level.
Taking these results collectively, we can deduce that the advanced learners had mastered the use of zhe to express ACCOMPANY of a complex motion event. Like native speakers, they were aware that whether or not zhe was used had little effect on meaning or grammaticality of
Sentence Type 1 and Type 4. In contrast, the counterparts in the intermediate proficiency level tended to accept Sentence Type 1, but reject Sentence Type 4, and their acceptability judgment on these two sentence types was significantly different. This group of learners had not acquired the use of zhe yet. These results are more clearly represented in Figure 5.6. 149
FIGURE 5.6 Judgment of ACCOMPANY with Transitive Verbs by Participant Groups
5.3.5 Acquisition of COMPLEX MANNER
This section will be dedicated to the learners’ acquisition of complex motion events when
the main verb is intransitive. As discussed before, intransitive manner-of-motion verbs in
Chinese differ from their counterparts in English in that they cannot participate in transitives or
causatives directly. Instead, they occur with a transitive verb, which expresses the MANNER.
Together they induce indirect causation, i.e. the combination of a transitive manner verb and an
intransitive motion verb achieves the effect of a resultative state of the object (of the transitive
manner verb). Here, I will refer to the verb incorporation as the COMPLEX MANNER.
Therefore, Sentence Type 1 ( MANNER V + PP) with intransitive verb is not acceptable as it would be in English. Instead, Sentence Type 7 ( MANNER V + MOTION V + PP) and Sentence Type 8 150
(MANNER Vzhe + MOTION V + PP) must be used. Of the two, the latter was preferred by native speakers (See Table 5.18 for more information).
TABLE 5.18 Descriptive Statistics of Participants’ Judgment of COMPLEX MANNER by Groups 95% Confidence Interval Std. Proficiency COMPLEX MANNER Mean Error Lower Upper Bound Bound Sentence Type 1 -0.16 0.16 -0.48 0.15 (MANNER V + PP) Sentence Type 2 0.56 0.86 0.21 0.12 (ba ...+ V + PP) Intermediate MANNER Sentence Type 7 -0.07 0.18 -0.44 0.30 (MANNER V + MOTION V + PP) Sentence Type 8 0.00 0.17 -0.33 0.33 (MANNER Vzhe + MOTION V + PP) Sentence Type 1 -0.65 0.16 -0.97 -0.32 (MANNER V + PP) Sentence Type 2 0.63 1.16 0.29 0.01 (ba ...+ V + PP) Advanced MANNER Sentence Type 7 0.76 0.19 0.38 1.14 (MANNER V + MOTION V + PP) Sentence Type 8 0.69 0.17 0.35 1.03 (MANNER Vzhe + MOTION V + PP) Sentence Type 1 -1.21 0.10 -1.42 -1.01 (MANNER V + PP) Sentence Type 2 -0.79 0.88 0.14 -1.07 Native (ba ...+ MANNER V + PP) Speakers Sentence Type 7 1.18 0.12 0.94 1.42 (MANNER V + MOTION V + PP) Sentence Type 8 1.47 0.11 1.25 1.68 (MANNER Vzhe + MOTION V + PP)
The data analysis was conducted in five dimensions: first, I would like to find out if learners understand that between Sentence Type 1 and Sentence Type 7, the latter is a way to express indirect causation, with the combination of a transitive verb expressing manner and an 151
intransitive motion verb (discussed in Section 5.3.5.1). Second, I am also interested in whether they have acquired the same preference between Sentence Type 8 versus Sentence Type 1
(discussed in Section 5.3.5.2). Third, how did learners understand the differences between
Sentence Type 2 and Sentence Type 7? Did they recognize that the use of Ba construction cannot save the sentence from ungrammaticality (in Section 5.3.5.3). Fourth, how did the learners interpret the differences between Sentence Type 2 and Sentence Type 8 (covered in Section
5.3.5.4). Lastly, I would like to know if learners noticed the differences between Sentence Type
7 and Sentence Type 8 (discussed in Section 5.3.5.5). At the end of this section, all the data will be pulled together. We can then evaluate how learners acquired the causative directed manner- of-motion with an intransitive verb (summarized in Section 5.3.5.6).
5.3.5.1 Sentence Type 1 versus Type 7 with Intransitive Verbs
A two-way repeated-measures ANOVA was carried out to evaluate how the three groups understood or acquired the causative directed manner-of-motion when the main verb was intransitive. The between-subject factor was the participant groups, while the within-subject factor was their judgment ratings of the Sentence Type 1 and Type 7 with intransitive verbs.
There was a significant main effect of the sentence types here: F(1,70)= 98.71, p < .001,
which meant that there were significant differences between the participants’ mean acceptance of
Sentence Type 1 and Type 7. The main effect of groups was not statistically significant: F(2,70)=
0.51, p = .605, which meant no significant differences were found amongst the three groups in
the average scores of the two sentence types. The interaction of the two sentence types and
groups was statistically significant: F(2,70)= 30.66, p p < .001. This significant interaction of the two sentence types and the groups implied that the three groups held different judgments towards 152
the two sentence types. The post hoc comparisons found no significant variances between each of the two groups.
Tests of the simple effect of the two sentence types were conducted for each level of the participant groups. No significant effect was found for intermediate learners: F(1,16)=0.15, p =
.706, which showed that intermediate learners did not assign significantly different ratings for these two sentence types. In other words, they were unable to differentiate the ungrammatical
Sentence Type 1 from the grammatical Sentence Type 7. There was a significant effect for both advanced learners (F(1,15) = 20.37, p < .001) and native speakers (F(1,39) = 263.86, p < .001).
These results suggested that advanced learners made similar judgments as native speakers and were able to distinguish between ways of expressing causation, with intransitive verbs. In other words, the advanced learners demonstrated significant understanding of the fact that the direct way, using ba , is ungrammatical, while the indirect way, using a manner-denoting transitive verb, is grammatical.
TABLE 5.19 Multiple Comparisons of the Groups in Their Judgment on Sentence Type 1 (with Intransitive Verbs) (I) Mean Std. 95% Confidence Interval (J) Proficiency Sig. Proficiency Diff (I-J) Error Lower Bound Upper Bound Advanced 0.49 0.23 .089 -0.06 1.03 Intermediate Native Speakers 1.05 * 0.19 .000 0.60 1.50 Intermediate -0.49 0.23 .089 -1.03 0.06 Advanced Native Speakers 0.56 * 0.19 .013 0.10 1.03 * Native Intermediate -1.05 0.19 .000 -1.50 -0.60 Speakers Advanced -0.56 * 0.19 .013 -1.03 -0.10 *. The mean difference is significant at the .05 level.
153
Additionally, tests of the simple effect of the participant groups were also performed for both Sentence Type 1 and Sentence Type7. Both simple effects were found statistically significant: for the Sentence Type 1, F(2,72)= 16.42, p < .001, and for the Sentence Type 7,
F(2,72)= 16.20, p < .001, implying that all three groups made significantly different judgment on these two sentence types. The post hoc comparisons revealed that in the case of Sentence Type 1
(originally listed in Table 5.13 and repeated here as Table 5.19), there were significant differences between the native speakers and the non-native speakers. It was evident that non- native speakers did not quite master the fact that unlike their L1 English, the Sentence Type 1 was ungrammatical in Chinese.
TABLE 5.20 Multiple Comparisons of the Groups in Their Judgment on Sentence Type 7
(I) Mean Diff Std. 95% Confidence Interval (J) Proficiency Sig. Proficiency (I-J) Error Lower Bound Upper Bound Advanced -0.82* 0.26 0.008 -1.46 -0.19 Intermediate Native Speakers -1.25* 0.22 0.000 -1.78 -0.72 Intermediate 0.82* 0.26 0.008 0.19 1.46 Advanced Native Speakers -0.43 0.22 0.145 -0.97 0.11
Native Intermediate 1.25* 0.22 0.000 0.72 1.78 Speakers Advanced 0.43 0.22 0.145 -0.11 0.97 *. The mean difference is significant at the .05 level.
In the post hoc multiple-comparison of groups on their judgment of the Sentence Type 7 with the COMPLEX MANNER (shown in Table 5.20), a different pattern was observed. There was an obvious difference between the native speakers (M= 1.18, 95% CI [0.94, 1.42]) and the intermediate learners (M= -0.07, 95% CI [-0.44, 0.30]), p < .001. Additionally, there was also 154
significant difference between the intermediate and advanced learners (M= 0.76, 95% CI [0.38,
1.14]), p = .008, in their acceptability of the two sentence types. However, no significant difference existed between the native speakers and the advanced learners, p > .05. The results demonstrated that the intermediate learners had little knowledge about the grammaticality of the two sentence types. However, the advanced learners were able to identify the ungrammatical
Sentence Type 1 from the grammatical Sentence Type7 when the verb was intransitive. In other words, the advanced learners have acquired the knowledge of expressing causation with intransitive verbs.
FIGURE 5.7 Grammaticality Judgment of Sentence Type 1 vs Type 7 with Intransitive Verbs by Participant Groups
1.40 Sentence Type 1 1.20 1.00 Sentence Type 7 0.80 0.60 0.40 0.20 0.00 -0.20 Intermediate Learners Advanced Learners Native Speakers -0.40 -0.60 -0.80 -1.00 -1.20 -1.40
155
The graphic presentation in Figure 5.7 demonstrates the participants’ distinctive acceptability judgment of Sentence Type 1 and Sentence Type 7. Even though both groups of non-native speakers had difficulty in identifying the ungrammaticality of Sentence Type 1 in general, the advanced learners were well aware of the differences between Sentence Type 1 and
Sentence Type 7. They significantly preferred the Sentence Type 7. The mean ratings of the intermediate learners indicated that they did not have explicit knowledge about the two sentence types.
5.3.5.2 Sentence Type 1 versus Type 8 with Intransitive Verbs
The results of the two-way repeated-measures ANOVA showed that there was a
significant main effect of the sentence types here: F(1,70)= 137.00, p < .001, which meant that
the mean acceptance of Sentence Type 1 and Type 8 was significantly different from each other.
The main effect of groups was not statistically significant: F(2,70)= 1.16, p = .319, which meant
no significant differences existed amongst the three groups in their average judgment scores of
the two sentence types. The interaction of the two sentence types and groups was statistically
significant: F(2,70)= 45.86, p < .001. This implied that the three groups held different judgments towards the two sentence types. I failed to find any significant differences between each of the two groups in the post hoc comparisons.
Tests of the simple effect of the two sentence types were conducted for each level of the participant groups. No significant effect was found for intermediate learners: F(1,16)=0.462, p =
.506, which demonstrated that intermediate learners did not have a preference for any one of these two sentence types. They assigned similar ratings to the ungrammatical Sentence Type 1 and the grammatical Sentence Type 8. There was a significant effect for both advanced learners 156
(F(1,15) = 19.31, p < .001) and native speakers (F(1,39) = 487.59, p < .001). Advanced learners and native speakers were well aware of differences in the grammaticality of the two sentences.
Additionally, tests of the simple effect of the participant groups were also performed for both Sentence Type 1 and Sentence Type 8. Both simple effects were found to be statistically significant: for the Sentence Type 1, F(2,72)= 16.42, p < .001, and for the Sentence Type 8,
F(2,72)= 28.80, p < .001, implying that all three groups made significantly different judgments on these two sentence types. The post hoc comparisons of Sentence Type 1 were previously discussed in Table 5.19).
TABLE 5.21 Multiple Comparisons of the Groups in Their Judgment on Sentence Type 8
(I) Mean Std. 95% Confidence Interval (J) Proficiency Sig. Proficiency Diff (I-J) Error Lower Bound Upper Bound Advanced -0.69* 0.24 0.015 -1.26 -0.11 Intermediate Native Speakers -1.46* 0.20 0.000 -1.94 -0.99 Intermediate 0.69* 0.24 0.015 0.11 1.26 Advanced Native Speakers -0.78* 0.20 0.001 -1.26 -0.29
Native Intermediate 1.46* 0.20 0.000 0.99 1.94 Speakers Advanced 0.78* 0.20 0.001 0.29 1.26 *. The mean difference is significant at the .05 level.
In the post hoc multiple-comparison of groups on their judgment of the Sentence Type 8
(Table 5.21), there were obvious differences between the native speakers (M= 1.47, 95% CI
[1.25, 1.68]) and non-native speakers: the intermediate learners (M= 0.002, 95% CI [-0.33,
0.33]), p < .001, and the advanced learners (M= 0.69, 95% CI [0.35, 1.03]), p = .001, in their acceptability of the two sentence types. At the same time, the mean acceptability ratings also 157
differentiated the advanced learners from the intermediate learners, p = .015. The results showed that the three groups were at three different stages in their judgment of Sentence Type 8: native speakers liked it more than the advanced learners, who in turn also like it more than the intermediate learners. This pattern is clearly depicted in Figure 5.8.
FIGURE 5.8 Grammaticality Judgment of Sentence Type 1 vs Type 8 with Intransitive Verbs by Participant Groups
1.80 1.60 Sentence Type 1 1.40 1.20 Sentence Type 8 1.00 0.80 0.60 0.40 0.20 0.00 -0.20 Intermediate Learners Advanced Learners Native Speakers -0.40 -0.60 -0.80 -1.00 -1.20 -1.40
5.3.5.3 Sentence Type 2 versus Type 7 with Intransitive Verbs
In this section, I will present and discuss the results based on the comparisons between
Sentence Type 2 ( ba ...+ MANNER V +PP) and Sentence Type 7 ( MANNER V + MOTION V + PP) with
intransitive verbs. The former employs the ba construction. The results reported in Section 5.3.4 showed that non-native speakers took the ba construction as a way to express causation, even with intransitive verbs. Actually, Sentence Type 7 and Type 8 are the way to go with intransitive 158
verbs. In this section, the center of discussion is on Sentence Type 7, and the next section will be dedicated to Sentence Type 8. A comparison between Sentence Type 2 and Sentence Type 7 will reveal how well the non-native speakers understood the two patterns.
The main effect of the sentence types was significant: F(1,70)= 7.10, p = .010, which suggested that Sentence Type 2 and Type 7 were rated significantly differently from each other.
The main effect of the groups was statistically significant: F(2,70)= 6.84, p =.002, which meant that the three groups varied in their average judgment scores of the two sentence types. The interaction of the two sentence types and groups also turned out to be statistically significant:
F(2,70)= 22.57, p < .001. This implied that the degree of differences in their judgments on the two sentences was uneven across the groups.
Tests of the simple effect of the two sentence types were conducted for each level of the participant groups. As a result of the grammaticality distinction between the two sentence types, the effect of the two sentence types for the native speakers proved to be significant: F(1,39) =
77.01, p < .001. There was no significant effect for intermediate learners: (F(1,16) = 4.46, p >
.05, or for the advanced learners: F(1,15) = 0.093, p = .764. In other words, the judgment ratings by both groups of the non-native speakers obviated the distinctions between Sentence Type 2 and
Type7. 159
TABLE 5.22 Multiple Comparisons of the Groups in Their Judgment on Sentence Type 2
(I) Mean Std. 95% Confidence Interval (J) Proficiency Sig. Proficiency Diff (I-J) Error Lower Bound Upper Bound Advanced -0.07 0.33 0.978 -0.85 0.72 Intermediate Native Speakers 1.35* 0.27 0.000 0.69 2.00 Intermediate 0.07 0.33 0.978 -0.72 0.85 Advanced Native Speakers 1.41* 0.28 0.000 0.74 2.08
Native Intermediate -1.35* 0.27 0.000 -2.00 -0.69 Speakers Native Speakers -1.41* 0.28 0.000 -2.08 -0.74 *. The mean difference is significant at the .05 level.
Some of the results were covered in the previous sections, I will incorporate them in the discussion here. Tests of the simple effect of the participant groups were exerted for both
Sentence Type 2 and Sentence Type 7. Both simple effects were found to be statistically significant: for the Sentence Type 2, F(2,72)= 19.28, p < .001, and for the Sentence Type 7,
F(2,72)= 16.20, p < .001, implying that the three participant groups made significantly different judgments on these two sentence types respectively. See Table 5.22 for results of the post hoc comparisons on Sentence Type 2. Significant differences were found between native speakers
(M= -0.79, 95% CI [-1.07, -0.50]) and non-native speakers (Intermediate Learners: M= 0.56,
95% CI [0.12, 1.00], and Advanced Learners: M= 0.63, 95% CI [0.01, 1.24]). The post hoc multiple comparisons of the groups on their judgment of the Sentence Type 7 (exemplified in
Table 5.20) indicated that significant differences existed between advanced (M= 0.76, 95% CI 160
[0.34, 1.17]) and intermediate learners (M= -0.07, 95% CI [-0.46, 0.33]), as well as between native speakers (M= 1.18, 95% CI [0.94, 1.42]) and intermediate learners.
Comparing the results provided in Table 5.22 and Table 5.20, it became very obvious that advanced learners tended to accept Sentence Type 7, but they still mistakenly overestimated the use of ba construction in Sentence Type 2. The intermediate learners, however, regarded
Sentence Type 2 as a grammatical way to express a causative directed motion with an intransitive verb. They tended to reject Sentence Type 7 probably because it was quite different from their L1. Figure 5.9 clearly demonstrates the pattern.
FIGURE 5.9 Grammaticality Judgment of Sentence Type 2 and Type 7 with Intransitive Verbs by Participant Groups
1.40 Sentence Type 2 1.20 1.00 Sentence Type 7 0.80 0.60 0.40 0.20 0.00 -0.20 Intermediate Learners Advanced Learners Native Speakers -0.40 -0.60 -0.80
-1.00 161
5.3.5.4 Sentence Type 2 versus Type 8 with Intransitive Verbs
Continuing the same line of logic, this section is devoted to the comparisons between
Sentence Type 2 ( ba ...+ MANNER V +PP) and Sentence Type 8 ( MANNER V + MOTION V zhe + PP). The former employs the ba construction, while the latter differs from Type 7 by the presence of the marker zhe . The main effect of the two sentence types (Type 2 versus Type 8) was significant:
F(1,70)= 12.60, p = .001, which testified that Sentence Type 2 and Type 8 were rated as significantly different from each other. The main effect of groups was not significant: F(2,70)=
2.76, p > .050, which meant that the average judgment scores of the two sentence types did not differ significantly across the three groups. The interaction of the two sentence types and groups was statistically significant: F(2,70)= 34.84, p < .001. In other words, the degree of differences in the acceptability judgments on the two sentences was uneven across the groups.
FIGURE 5.10 Grammaticality Judgment of Type 2 and Type 8 with Intransitive Verbs by Participant Groups
1.60 Sentence Type 2 1.40 1.20 Sentence Type 8 1.00 0.80 0.60 0.40 0.20 0.00 -0.20 Intermediate Learners Advanced Learners Native Speakers -0.40 -0.60 -0.80 -1.00
162
Tests of the simple effect of the two sentence types were conducted for each level of the participant groups. Similar to the distinction between Sentence Type 2 and Type 7, discussed in the previous section, the effect of the two sentence types here was also significant for the native speakers: F(1,39) = 150.26, p < .001. Surprisingly, the effect of the two sentence types was
significant for the intermediate learners: F(1,16) = 4.976, p < .05, but not for the advanced
learners: F(1,15) = 0.200, p = .891. A close investigation (consult Figure 5.10) revealed that not
only were the intermediate learners unable to distinguish the grammatical Sentence Type 8 from
the ungrammatical Type 2, on the contrary, they took the ungrammatical pattern as grammatical.
Tests of the simple effect of the participant groups were performed for both Sentence
Type 2 and Sentence Type 8. Since the result of the former was presented in the previous
section, only the results of the latter are covered here. The simple effect of the participant groups
was statistically significant for the Sentence Type 8, F(2,72)= 28.80, p < .001, which suggested
that the three participant groups made significantly different judgments. Table 5.21 provides
results of the post hoc comparisons on Sentence Type 8. The three participant groups were
distributed in three tiers: native speakers (M= 1.47, 95% CI [1.28, 1.65]) made significantly
different judgment from the Intermediate Learners: M= 0.003, 95% CI [-0.38, 0.38], p < .001,
who were also significantly different from the Advanced Learners: M= 0.69, 95% CI [0.22,
1.16], p = .015. The advanced learners were also significantly different from the native speakers,
p = .001.
163
5.3.5.5 Sentence Type 7 versus Type 8 with Intransitive Verbs
This section focuses on the results of the comparison between Sentence Type 7 and Type
8. The main effect of the sentence types was not significant: F(1,70)= 1.52, p = .221, which
suggested that the participants did not consider the two types to be very different. The main
effect of groups was statistically significant: F(2,70)= 25.87, p <.001, which meant that the three
groups varied in their average judgment scores of the two sentence types. The interaction of the
two sentence types and groups, however, was not statistically significant: F(2,70)= 2.20, p =
.119. This implied that the degree of differences in their judgments was not significant. The post
hoc comparisons of the three groups (Table 5.23) on the average judgments of the two sentence
types showed that the native speakers were significantly different from the non-native speakers,
while the advanced learners were significantly different from the intermediate learners. The level
of acceptability was distributed among three tiers.
TABLE 5.23 Multiple Comparisons of the Groups in Their Judgment on Sentence Type 7 and Sentence Type 8 with Intransitive Verbs
(I) Mean Diff Std. 95% Confidence Interval (J) Proficiency Sig. Proficiency (I-J) Error Lower Bound Upper Bound Advanced -0.76* 0.23 0.004 -1.30 -0.21 Intermediate Native Speakers -1.36* 0.19 0.000 -1.81 -0.90 Intermediate 0.76* 0.23 0.004 0.21 1.30 Advanced Native Speakers -0.60* 0.19 0.008 -1.07 -0.14
Native Intermediate 1.36* 0.19 0.000 0.90 1.81 Speakers Native Speakers 0.60* 0.19 0.008 0.14 1.07 *. The mean difference is significant at the .05 level.
164
Tests of the simple effect of the two sentence types were conducted for each level of the participant groups. There was a significant effect for native speakers (F(1,39) = 8.58, p = .006, who gave higher preference for Sentence Type 8. However, no significant effect was found for the intermediate learners: F(1,16)=0.184, p = . 674 or the advanced learners: F(1,15) = 0.339, p
= .569, which indicated that both groups of non-native speakers did not recognize the differences between Sentence Type 7 and Type 8.
FIGURE 5.11 Grammaticality Judgment of Sentence Type 7 vs Type 8 with Intransitive Verbs by Participant Groups
1.60 Sentence Type 7 1.40 Sentence Type 8 1.20 1.00 0.80 0.60 0.40 0.20 0.00
Intermediate Learners Advanced Learners Native Speakers
In summary, in terms of participants’ acceptability judgments on Sentence Type 7, there were no significant differences between native speakers and the advanced learners, but both advanced learners and native speakers made significantly different judgments from the intermediate learners. In the case of judgment on Sentence Type 8, besides significant differences between the advanced learners and the intermediate learners, there were also 165
significant differences between the advanced learners and the native speakers. The graphic representation in Figure 5.9 gave us a clear illustration. Comparing the participants’ judgments on these two sentences, it is fair to conclude that the advanced learners have acquired the structure of Sentence Type 7, but they still had difficulty using zhe in such sentences. The intermediate learners, on the other hand, had little knowledge of these two sentence types.
5.3.5.6 Summary
In this section, we have investigated how learners of Chinese acquire the COMPLEX
MANNER with five comparisons involving four sentence patterns: Sentence Type 1 ( MANNER V +
PP), Sentence Type 2 ( ba ...+ MANNER V +PP), Sentence Type 7 ( MANNER V + MOTION V + PP) and
Sentence Type 8 ( MANNER Vzhe + MOTION V + PP). A summary of statistical results is provided in
Table 5.24. The first two types were ungrammatical in Chinese, whereas the latter two are ways
to express indirect causation of directed motion with intransitive verbs. Of the two
ungrammatical sentence types, native speakers rated Sentence Type 2 as being unacceptable. In
other words, the use of ba construction in Sentence Type 2 did not make the sentence
grammatical to native speakers, but it somehow increased the acceptability of the sentence. Of
the two grammatical sentence types, native speakers gave preference to Sentence Type 8 with the
overt marking of zhe . The differences between Sentence Type 7 and 8 were subtle to native
speakers, but statistically significant. On the whole, the distinctions between grammatical and
ungrammatical sentence types were statistically supported by the results in native speakers’
performance.
As the graph in Figure 5.11 shows, the same pattern can be seen in intermediate and
advanced learners. That is, they both recognized that Sentence Type 1 was not grammatical in
Chinese (although it was in English), but they both accepted Sentence Type 2 as grammatical, 166
assuming that the use of the ba construction could be a strategy to express a causative motion event. The two groups differ in that the advanced learners, but not the intermediate learners, were able to recognize the two grammatical sentence types, Type 7 and Type 8. They assigned statistically higher ratings for these two sentence types in comparison with the ungrammatical
Sentence Type 1. The intermediate learners, however, had nearly no knowledge of the two types,
Sentence Type 7 and 8. The first piece of evidence pointing to this conclusion comes from the fact that the mean acceptability ratings of these two sentence types were about 0, which was defined as “do not know” in the judgment scale. The second piece of evidence is the fact that there were no significant differences between their ratings for the ungrammatical Sentence Type
1 and the two grammatical sentence types. The available data suggests that they treated these two sentence types much in the same way as they did the ungrammatical Sentence Type 1.
FIGURE 5.12 Grammaticality Judgment of Sentences with Intransitive Verbs by Participant Groups
167
Table 5.24 Acquisition of COMPLEX MANNER with Intransitive Verbs Summary of Results Based on Two-way Repeated Measures ANOVA Simple effect of Main Interaction Simple effect of Groups Main Factor Effect Factor Effect of Post-hoc of Factor Factor Groups Compariso M A NSs Post-hoc Post-hoc Factor Level 1 Level 2 n Sentence Not NSs vs M sig Not sig sig / sig sig sig NSs vs NNSs sig Type 1 vs 7 sig A vs M
Sentence Not NSs vs NNSs sig Not sig sig / sig sig sig NSs vs NNSs sig Type 1 vs 8 sig A vs M
Sentence Not Not NSs vs M sig sig sig / sig sig NSs vs NNSs sig Type 2 vs 7 sig sig A vs M
Sentence Not NSs vs NNSs sig Not sig sig / sig sig sig NSs vs NNSs sig Type 2 vs 8 sig A vs M NSs vs Sentence Not Not Not NSs vs M NSs vs NNSs Not sig sig NNSs A vs sig sig sig Type 7 vs 8 sig sig sig A vs M A vs M M Abbreviations: NSs: Native Speakers; NNSs: Non-Native Speakers; M: Intermediate Learners; A: Advanced Learners; Tran: Transitive; Intran: Intransitive; Pre-v PP: pre -verbal prepositional phrases; Post -v PP: post -verbal prepositional phrases; 1 -4: Sentence Type 1 -4 168
The advanced learners outperformed the intermediate learners in being able to differentiate the grammatical sentence types from the ungrammatical ones. However, they failed to recognize the subtle differences between Type 7 and Type 8. Their judgment showed an opposite pattern to that of the native speakers, who gave preferences to Sentence Type 8, with zhe . This contrast suggests that the presence of zhe makes Sentence Type 8 more complicated in
structure and presents additional challenges to the learners.
5.4 Acquisition of Causative Directed Manner-of-motion Event
So far in this chapter we have presented the data collection, analysis and results of the
acceptability judgment task in an attempt to explore the second language acquisition of complex
motion events in Chinese. The structures of complex motion events, or technically, the causative
directed manner-of-motion events, were decomposed into five core variables, namely
TRANSITIVITY, GOAL, CAUSATIVITY, ACCOMPANY, and COMPLEX MANNER. The
general inquiry has been thus broken down into six sub-questions (p.111). The first five
questions were answered in Section 5.3. In this section, I will summarize the results first, and
then consider the last question. That is, can we identify an order or sequence in which learners of
Chinese acquire these core components of complex motion events?
The first question addressed English-speaking learners’ acquisition of TRANSITIVITY
in Chinese. The results showed clearly that advanced learners were able to make a distinction
between transitive and intransitive verbs in licensing causative directed motion; however, their
intermediate counterparts were not aware of such differences. On one hand, no significant
differences were found amongst the three groups for sentences with transitive verbs; on the other 169
hand, significant group differences existed regarding their acceptability judgments of sentences with intransitive verbs. This contrast revealed that intransitive manner-of-motion verbs were challenging for the learners. Further investigation revealed that even though both advanced and intermediate learners had problems with intransitive manner-of-motion verbs, the advanced learners made significantly more native-like judgments than the intermediate learners.
The second research question involved the acquisition of GOAL. Statistical results suggested that only the advanced learners were able to recognize the differences between pre- verbal PPs and post-verbal PPs with respect to expressing the GOAL in motion events. That being said, however, they faced the same challenges as intermediate learners; both need to learn how pre-verbal PPs are used. The third research question concerned learners’ acquisition of the notion of CAUSATIVITY. It was noted that advanced learners had mastered the use of ba construction when the manner-of-motion verb was transitive in nature. They were also aware that intransitive verbs in Chinese could not participate in transitivization, but they assumed that the use of ba construction would improve the grammaticality of sentences within intransitive verbs.
The statistical evidence pointed out that the intermediate learners did not acquire the use of ba construction whether the main verb was transitive or intransitive. I am prone to believe that some of the intermediate learners may have had some implicit knowledge of using the ba construction to describe causative events, but the majority of the intermediate group had not yet fully mastered the use of the ba construction.
The fourth research question concerned Chinese learners’ acquisition of the notion
ACCOMPANY in sentences with transitive verb. It can be reasonably argued that advanced learners have mastered the use of zhe for accompanied action, but intermediate learners have not.
The acquisition of ACCOMPANY in sentences with intransitive verbs was related to the fifth 170
research question regarding the acquisition of COMPLEX MANNER. It turned out that the advanced learners, but not the intermediate learners, were able to make the correct judgment distinguishing the grammatical sentences from the ungrammatical sentences in realization of
COMPLEX MANNER. But they were not completely in line with native speakers who tended to use overt marking in such situations. In other words, the marker zhe made complex motion events with intransitive verbs even more complicated and provided an additional level of difficulty to the already challenging learning task.
Taking the results of all five variables collectively, it is worth noting that English- speaking learners of Chinese acquire these components in a sequence. Overall, to describe causative directed-motion events, transitive verbs are learned before intransitive verbs (Pattern
1). This pattern is explicitly demonstrated in the acquisition of every semantic component. Thus, in the acquisition of CAUSATIVITY, they acquire the use of the ba construction with transitive verbs before the intransitive verbs, and in the case of GOAL, they acquire the distinctions between pre- and post-verbal PPs with transitive verbs prior to intransitive verbs.
• Pattern 1: Transitive Manner-of-motion Verbs < Intransitive Manner-of-motion Verbs 1 Secondly, when the main motion verb is transitive, the learners acquire the notion of
CAUSATIVITY first, followed by ACCOMPANY, and then GOAL (Pattern 2). In other words,
the English-speaking learners first learn that the use of the ba construction is a strategy of expressing CAUSATIVITY in a directed motion event; they next learn the use of zhe to indicate
1 “A < B” means that A is easier or learned earlier than B. Conversely, “A > B” indicates that A is more difficult or learned later than B. 171
Table 5.25. Summary of Results of Two-way Repeated Measures ANOVA Main Main Interaction Simple effect of Factor Simple effect of Groups Factor Effect of Effect of Post-hoc Factor Factor M A NSs Post-hoc Post-hoc Factor Groups Comparison Level 1 Level 2 Not (Tran) (Intran) NSs vs NNSs 1. TRANSITIVITY sig sig sig NSs vs NNSs sig sig / sig Not Sig Sig A vs M 2.GOAL PPs NSs vs NNSs Not (Pre-v PP) (Post-v PP) sig sig sig sig sig NSs vs NNSs NSs vs M (Transitive Verbs) A vs M sig sig Sig 2.GOAL PPs Not NSs vs NNSs Not Not (Pre-v PP) NSs vs NNSs (Post-v PP) sig sig sig NSs vs NNSs (Intransitive Verbs) sig A vs M sig sig sig A vs M Sig 3.CAUSATIVITY Not (Tran1) (Tran2) NSs vs M sig sig sig NSs vs M sig sig / (Transitive Verbs) sig Not Sig Sig A vs M 3.CAUSATIVITY (Intran1) (Intran2) sig sig sig NSs vs NNSs sig sig sig NSs vs NNSs NSs vs NNSs (Intransitive Verbs) sig Sig 3.Sentence Type 2 Not (Tran2) NSs vs M (Intran2) Transitive Verbs vs sig sig sig / sig sig NSs vs NNSs sig Sig A vs M Sig Intransitive Verbs
4. ACCOMPANY Not Not Not (Tran1) (Tran4) Not sig sig / Sig / NSs vs M (Transitive Verbs) sig sig sig Not Sig sig
4. ACCOMPANY Not Not Not Not (Intran1) (Intran4) Not sig sig NSs vs NNSs NSs vs NNSs NSs vs NNSs (Intransitive Verbs) sig sig sig sig Sig sig
5.COMPLEX See Table 5.24 MANNER Abbreviations: NSs: Native Speakers; NNSs: Non-Native Speakers; M: Intermediate Learners; A: Advanced Learners; Tran: Transitive; Intran: Intransitive; Pre-v PP: pre- verbal prepositional phrases; Post -v PP: post -verbal prepositional phrases; 1 -4: Sentence Type 1 -4. 172
accompanied action. Finally, they learn that post-verbal PPs, but not pre-verbal PPs, are used to express the GOAL of directed motion. When the main verb is intransitive, all the semantic and syntactic elements are so entangled together that English-speaking learners are not able to acquire all of these features simultaneously. Instead, their acquisition goes through several stages and the acquisition of each feature enforces will influence their performance concerning the next variable. Initially, they learn that the causative directed –motion events of intransitive verbs in
Chinese are expressed differently from their L1. Two strategies are involved, lexicalizing the
CAUSE of the event, relying on another verb to express ACCOMPANY. However, this recognition per se does not assist them in achieving native-like proficiency until they learn other aspects of grammar, including differences between pre- and post-verbal PPs, restriction of ba construction to transitive verbs, and the use of zhe in the context of ACCOMPANY. This is given as Pattern 3 below.
• Pattern 2 (transitive verbs): CAUSATIVITY < ACCOMPANY < GOAL
• Pattern 3 (intransitive verbs): COMPLEX MANNER < GOAL < CAUSATIVITY < ACCOMPANY The fourth pattern reflected in the data is that the English-speaking learners of Chinese acquire Post-verbal Prepositional Phrases before they acquire the Pre-verbal Prepositional
Phrases. The former is syntactically similar as that in their L1, while the latter appears saliently very early in the L2 input.
• Pattern 4: Post-verbal PPs < Pre-verbal PPs
173
5.5 Summary
In this chapter, I have analyzed L2 learners’ performance in a reliable picture judgment
task with great details as an attempt to understand how English-speaking learners acquire ways
to describe causative directed motion in Mandarin Chinese. Based on the linguistic analysis
provided in Chapter 2, the complex motion under investigation here is decomposed into three
sub-events and five core semantic factors. The complicated L2 acquisition of the complex
motion is hence observed in five tasks: acquisition of TRANSITIVITY, GOAL,
CAUSATIVITY, ACCOMPANY, and COMPLEX MANNER respectively. The statistical
results are extracted from a number of two-way repeated-measures ANOVA on each factor first.
Then the data are pulled together to inform us the acquisition sequence of these factors. As a
result, four acquisition patterns are identified and presented at the end of the chapter. In the next
chapter, I will address the theoretical and pedagogical implications of these results in the context
of literature re-contemplation and through the lens of some translation data.
174
CHAPTER 6
DISCUSSION AND CONCLUSION
So far I have presented the results of the L2 acquisition study mostly based on the
statistical analysis. This ending chapter completes the thesis by extending the results from the
current L2 study to a more general discussion of theoretical and pedagogical implications. It
blends the following three parts holistically to shed light for future research on this topic:
• firstly, to draw some primitive conclusions on how English-speaking learners of Chinese
acquire the ability to describe causative directed manner-of-motion in the target language;
• secondly, to discuss how these results relate to the current second language acquisition (SLA)
theory, and what pedagogical implications this study provide for teaching Chinese as a
foreign language (CFL);
• and finally, to provide a brief summary of the contributions and limitations of the study for
future research.
In Chapter 2, I have first provided a thorough linguistic analysis of the cross-linguistic variations between Chinese and English in licensing causative directed manner-of-motion events.
Initially, the most well-known conflation typologies (Talmy, 1975 & 1985; Slobin, 1996 & 2004) are applied to extricate the distinctions between the two languages. These unsuccessful attempts did not address some of the structural distinctions between the two languages. One of the significant contributions of my study lies in the fact that I have initiated a lexical decomposition approach and light verb analysis to examine the semantic-syntactic interplay of the complex motion event. Following the Lexicalization Parameter proposal raised by Huang (1997 & 2006) and Lin (2001) within the framework of Principles and Parameters Theory, the causative directed
175 manner-of-motion is thus decomposed into three sub-events licensed with different light verbs depending on the transitivity property of the main verb. This provides me with the means to explore how learners acquire the expressions of each sub-event with the use of light verbs.
In Chapter 3, I have touched upon the application of Principles and Parameters Theory to
SLA. There have been ongoing discussions on whether second language acquisition falls into a parameter resetting paradigm or a feature assembly/reassembly configuration (Haegeman, 1988;
Radford, 1997; Lardiere, 2008 & 2009; Champaign, 2009; Ottawa, 2009; Slabakova, 2009;
White, 2009). On one end, in a series of keynote articles, Lardiere (2008, 2009) cogently proclaims that parameter resetting is insufficient to describe cross-linguistic variations and it is also inadequate in explaining individual learner variations. Therefore, she proposes that feature assembly/reassembly in the L2 be employed to understand the difficulty of L2 acquisition. On the other end, other applied linguists (Slabakova, 2009; White, 2009; among others) defend the linguistic theory underlying the generative approach to SLA. They regard Lardiere’s proposal as a form of constructionist approach and find her proposal runs the risk of losing the deductive and explanatory power of the language learning mechanism. Slabakova (2009) refutes that parameters are actually universal constraints on feature assembly and re-assembly because while the magnitude of the learning task may vary, language variation is not limitless and therefore learner behavior can be predicted.
As a matter of fact, SLA specialists, who conduct contrastive analysis in SLA, have always been interested in making predictions about learner behaviors based on the L1-L2 variations. For example, White (2003) points out that differences in how languages realize arguments morphologically and syntactically are expected to cause difficulties in L2 acquisition.
When L1 forms a superset of L2 and permits more ways of realizing a particular argument
176 structure in the syntactic structure than L2, overgeneralization from L1 to L2 is predicted.
Conversely, undergeneralization is predicted when L1 forms a subset of L2 and does not provide as many ways of licensing a particular argument structure in syntax which are nevertheless exemplified in L2 input.
To discuss the theoretical implications of the present study, there are two sets of questions I will address. Firstly, do the results of the present acquisition study support parameter resetting paradigm or feature assembly/reassembly configuration? What evidences do we have or lack to buttress this claim? Secondly, what overgeneralization and undergeneralization patterns are predicted based on the contrastive analysis of complex motions in Chinese and English? Are these predictions verified in the results?
To answer the first sets of questions, I will start with the identification of the parameter and features investigated in the study. The linguistic analysis in Chapter 2 reveals that the event structure of causative directed manner-of-motion contains three sub-events and the Chinese-
English variations in licensing these complex motions are originated from the Lexicalization
Parameter difference between the two languages. Based on the Lexicalization Parameter, I propose that each sub-event is realized via one light verb (ba for CAUSE, dao for GOAL, and zhe for ACCOMPANY) and overtly projected to syntax in Chinese. In contrast, English verbs go through a full conflation in the lexical computation before entering the syntax, so the composition and licensing of these same sub-events are completed in LF (or L-syntax) of English.
In further analysis and design of the acquisition study, I have decomposed complex motion events into five semantic factors, or namely, semantic features: TRANSITIVITY,
CAUSATIVITY, GOAL, ACCOMPANY and COMPLEX MANNER. Note that linguists have different definition of features. Features can be primitive phonological and semantic elements
177 that make up the lexical inventory of every language. Feature can also be bundled together onto functional categories in different language-specific ways (Lardiere, 2009). When you analyze each of these features in great details, they may even be more like parameters. I will leave this discussion for future researchers.
Now that the parameter, i.e. Lexicalization Parameter, and the semantic features, i.e. the five factors, are identified for the present research, I will try to answer the above-mentioned follow-up questions regarding whether the results of the present study provide any direct evidences for either side of two SLA proposals, and whether predictions of overgeneralization and undergeneralization are valid in the actual results. Because the Lexicalization Parameter does not involve superset and subset distinctions, the discussion of over- and under- generalization issue requires me to first zoom in and focus on one feature at a time. Later this chapter, I will zoom out and collapse the results together in order to answer the more general questions and obtain a better understanding of the macro picture.
I will recapitulate and extract the main findings drawn from the acceptability judgment data in combination with the results collected from the translation task in the pilot study. The first feature being studied here is TRANSITIVITY, or [±transitive]. It is noted that this feature is selected and configured differently in Chinese and English. As discussed in Chapter 1 and
Chapter 2, both transitive and intransitive/unergative motion verbs in English can participate in causativization if the semantic restrictions are met (Folli and Harley, 2007). However, only transitive motion verbs can be causativized in Chinese. Unergative motion verbs in Chinese cannot participate in the equivalent causativization per se. The significant distinctions between transitive and intransitive motion verbs are confirmed in the judgment data of native speakers
(Chapter 4, Section 4.6.5.1). To describe the language variations in another way, the causative
178 directed motion has no selectional restriction for the feature [±transitive] in English, but it selects a different value, i.e. [+transitive], in Chinese. Overgeneralization is thus predicted in English- speaking learners’ re-assembly of this feature, i.e. L2 learners will accept the causativization of intransitive verbs as it is acceptable in their L1.
Now we will look at the results of TRANSITIVITY. As the results in Chapter 5 demonstrate, English-speaking learners acquire the transitive use of manner of motion verbs before the intransitive use in expressing causative directed-motion events in Mandarin Chinese
(Pattern 1). Advanced learners have acquired this feature at the time of experiment, while intermediate learners have not. The major learning challenge is with intransitive verbs, or the [- transitive] setting of this feature. This is because no significant group differences were found in the judgment of transitive verbs, but the intransitive verbs in such complex events significantly differentiated the language proficiency of these participants. Advanced learners made significantly more-native like judgment on sentences with intransitive verbs. These contrasting results thus indicate that the major challenge for students to acquire the [±transitive] feature exists more with [-transitive] than the other value. The overgeneralization is confirmed by the results. Both intermediate and advanced learners found the causative structure with an intransitive verb acceptable.
Next, I will recap the second and third acquisition patterns observed in the L2 acquisition study. Then I will take the same approach as above to analyze how L2 learners acquire each of the features involved. The second acquisition pattern of the results points out that when the main motion verb is transitive, English-speaking learners acquire the feature of CAUSATIVITY before that of ACCOMPANY, and lastly GOAL. In other words, English-speaking learners of
Chinese acquire the use of light verb ba first, followed by the use of light verb zhe , and lastly
179 light verb dao with understanding of the distinction between pre- and post-verbal goal PPs. This sequence may be a result of several factors: (1) the input salience of ba construction in the target language, (2) imperfective aspectual markers are typically acquired after perfective aspectual markers (Ke, 2005), and (3) the distinctive semantic and functional features associated with the different distribution of prepositional phrases are extremely difficult for English-speaking learners of Chinese. The same type of difficulty is also reflected in the translation task, where the prepositions are misused or even avoided. Examples in (6.1b) and (6.1c) show that learners do not know where to put the preposition in the sentence, so they either avoid using it or they use it between the serial VPs.
6.1 a. I walked to the store to buy some clothes.
b. *我走(路)商店(去)买衣服。 Wo zou(lu) shang-dian (qu) mai yi-fu I walk (road) store (go/to) buy clothes
c. * 我走路商店到买衣服。 Wo zou-lu shang-dian dao mai yi-fu I walk-road store DAO buy clothes
On the other hand, when the main verb is intransitive, the acquisition sequence displays a completely different pattern. Advanced learners demonstrate that they are able to identify the grammatical way of describing a causative directed motion. The ba construction adds some complication to the judgment task, and the participants seem to posit that the causative marker ba
provides an equally appropriate way to license the causative directed motion. This
misinterpretation explains the underlying reasoning behind the ungrammatical translation
displayed below in (6.2). Here learners have exploited several cognitive strategies and resorted to
180 using various verbs in Chinese to signify the causative event: in (6.2 b-d), they have attempted to utilize the light verb ba ; in (6.2 e) they have tried the instrument light verb yong ‘use’; in (6.2 f- g), they have also tested two other verbs rang and shi , both of which entail the literary meaning of ‘cause’ and ‘give rise to’.
6.2 a. Mary jumped the horse over the fence.
b. * Mary 把马跳(上/过)籬笆。 Mary ba ma tiao (shang/ guo) li-ba. Mary BA horse jump (above/ over) fence
c. Mary 把马跳外部是籬笆。 Mary ba ma tiao wai-bu shi li-ba Mary BA horse jump outside is fence
d. Mary 把马籬笆跳過。 Mary ba ma li-ba tiao-guo Mary BA horse fence jump-over
e. Mary 用马起籬笆。 Mary yong ma qi li-ba Mary use horse rise fence
f. Mary 让马跳(上/过)围栏去。 Mary rang ma tiao (shang/guo) wei-lan qu Mary let/cause horse jump (above/ over) fence go
g. Mary 使马跳籬笆。 Mary shi ma tiao li-ba Mary cause horse jump fence Now, let us take a close look at the features involved in the acquisition Pattern 2 and
Pattern 3. As concluded in Chapter 2, the lexical items and syntactic representations of these
variables are assembled quite differently between Chinese and English, each selecting different
co-occurring features. For example, CAUSATIVITY has no restriction on the feature value of
[±transitive] in English, but it goes into a binary situation in Chinese. On one side, when it
181 selects [+transitive], light verb ba is highly preferred or even required. This feature selection is verified in native speaker’s judgment of CAUSATIVITY in Chapter 4. On the other side, when it selects [-transitive], Chinese uses the composition strategies to incorporate manner and motion, which is referred to be COMPLEX MANNER in the study.
Additionally, if Folli and Harley’s (2007) account of the availability of causatives for directed motion is proper, the feature ACCOMPANY selects [+Agentive] [+Intentional] in all languages, including English and Chinese, but only Chinese bears the syntactic conditions that light verb zhe is projected in syntax. In the case of GOAL, the [+ post-verbal] feature is realized through certain lexical computations in English, e.g. the composition of dynamic PPs like to with stative PPs such as in , on , and at . This happens covertly in English syntax, but it is an overt computation schema in Chinese as the value of this feature [±post-verbal] affects the semantic interpretations and determines the grammaticality of the sentence in a given context. In short, all these features come with co-occurring features and different conditioning environments in
Chinese. Since Chinese (L2) requires the above-listed additional feature restrictions for
CAUSATIVITY, ACCOMPANY, GOAL, and COMPLEX MANNER that are not present in
English (L1), undergeneralization of light verb ba , zhe , dao and COMPLEX MANNER is thus predicted in the English-speaking learners’ acquisition of these Chinese features.
The acquisition of CAUSATIVITY discussed in Chapter 5 show that advanced learners have acquired the [+transitive] restriction for CAUSATIVITY, but they experience the same learning difficulty as their intermediate peers with respect to the indirect causativization for intransitive verbs.
182
The results of COMPLEX MANNER show that advanced learners can recognize the grammatical expressions to describe the causative motion event with intransitive verbs, but they fail to reject the ungrammatical expressions when ba is used with intransitive verbs. Instead of the predicted undergeneralization problems, the examples in (6.2 a-g) actually prove that L2 learners have made overgeneralization errors. These results illustrate that advanced English- speaking learners of Chinese are able to interpret and detect most of the features associated with
CAUSATIVITY, but they also have great difficulty in assembling just the right combination of features into the right lexical and functional items. They also have difficulty determining the appropriate conditioning environments to express the features, exemplified in their overgeneralization of ba to causativize intransitive verbs in this case.
The acquisition results of ACCOMPANY presented in Chapter 5 suggest that advanced
English-speaking learners have mastered the major configurations with this feature, but they have not completely acquired the nuances. The acceptability judgments of the native speakers
(Chapter 4, Section 4.6.5.4) indicate that the light verb zhe is preferred when COMPLEX
MANNER is used to describe the indirect causation with intransitive verbs. In this sense, L2 learners do make undergeneralization errors as predicted with the use of the light verb zhe. The similar acquisition pattern also exists in their acquisition of the feature GOAL. Advanced learners have acquired partial knowledge of the word-order restrictions [± post-verbal] on the representation of GOAL, but confusion arises with the [- post-verbal] position (Pattern 4).
These inadequate acquisitions of features are intertwined together and make the learning task of causative directed motion even more difficult; however, I am not surprised with these results. I suspect that only learners at superior proficiency level can have full mastery of the
Lexicalization Parameter and the different features associated with it. Each of these features
183 poses some challenges to the L2 learning task, and the fact that the interpretation and representation of these features may even further partitioned (i.e. has dependent features) has made it even more challenging (Cowper, 2005). Lardiere (2008, 2009) asserts that the greater difficulty for the second language acquisition lies in reconfiguration and remapping the right combination of features from the L1 to the L2, especially in cases where such features
(interpretable or uninterpretable, over or cover) do exist in the L1 but are configured differently.
Now that we have gone through the examination of the features bundled together in the representation of causative directed motion, I would like to pull all the results together and try to answer some broad questions as follows:
1. Is the Lexicalization Parameter observed in the L2 acquisition data?
2. If (1) is positive, have the settings of the value of this parameter deterministically produced a cascade of related effects to make the task of L2 learning correspondingly easier for English- speaking learners of Chinese?
3. If (1) is negative, do the results suggest that L2 learners associate the existing L1 (English) features with different lexical or syntactic items in the L2 (Chinese) and acquire L2 language- specific configurations of features?
Following the parameter resetting paradigm, if L2 learners have learned to reset the
Lexicalization Parameter in their L2, they will gain these features free or in a “cascade of effects”
fashion. On the other hand, in accordance with the feature assembly/reassembly proposal, L2
learners have to “reconfigure or remap features from the way these are represented in the L 1
into new formal configurations” on quite different types of lexical items and syntactic structures
184 in the L2 (Ladiere, 2009, p.175). What kind of evidences can we draw from the results and bring to light here?
Earlier in (6.2 b-g), L2 learners are found to make overgeneralization to describe causatives with intransitive verbs employing light verb ba and similar verbs with causative meanings. What motivates them to do so in the first place? The fact that learners find it a necessity to describe the causative event with the use of these lexical items, suggests that learners have taken cognitively and semantically conscious actions to decompose the complex motion events into sub-events: in this case, the causative sub-event. The examples in (6.2) seem to give evidence for the parameter re-setting proposal: advanced learners may be aware that the event structures are licensed at S-syntax via the function of light verbs in Chinese, so they have attempted to spell them out here. However, I do not have other production data to further validate this claim.
If we consider from another angle, all the result taken collectively, no participant in the non-native speakers groups has demonstrated that they have acquired the knowledge of all the features investigated here. There are two possibilities to explain this phenomenon. First, L2 learners at superior proficiency level are needed to see if they have acquired the Lexicalization
Parameter. Second, the parameter resetting proposal is not valid as no cascade of related effects is observed even though some L2 learners have acquired the Lexicalization Parameter.
Based on the data we have, what we learn from the complex acquisition patterns of all these features is that L2 learners have partial knowledge of these features, for example, the
[+transitive] restriction for CAUSATIVITY, the [+post-verbal] position requirement for GOAL with transitive verbs. The incomplete acquisition of each feature and the acquisition sequence
185 reflect that the uneven development of their interlanguage. Therefore, these intertwined but sequenced patterns may be used as evidence in support of the feature assembly/re-assembly proposal. At this point of time, I do not have sufficient evidence to answer the above three questions with confidence and the data do not prove one way or another.
The present study also sheds light on pedagogical issues in CFL classrooms. One of the interesting finding of this study suggests that the complication of the light verb ba in sentences with intransitive verbs makes the judgment significantly more confusing to the learners.
Different from the previous L2 acquisition studies on ba construction (Du, 2004) or causative constructions (Hui, 2010) in Chinese, only part of this study deals with the acquisition of the light verb ba as a causative marker in a complex event. However, the result is very informative.
Acquisition Pattern 3 in Chapter 5 indicates that in the case of sentences with an intransitive verb,
CAUSATIVITY marked with the ba construction is actually acquired after the goal PPs, which is the opposite to the acquisition sequence in sentences with transitive verbs. This finding has pedagogical implication in the sense that ba construction as a unique syntactic structure in
Chinese makes undergeneralization an easy pitfall for learners; however, after learners acquire its use in transitive sentences, they tend to make overgeneralization errors in intransitive sentences. This U-shape behavior in the acquisition of the ba construction indicates that when negative evidence is not sufficient in L2 input, explicit language instruction and exercise need to take place. Teachers of Chinese not only need to show the positive evidence by teaching students when the ba construction can be utilized to license causative event, but also need to provide negative evidence by demonstrating to students ungrammatical sentences with overgeneralized use of the structure.
186
Additionally, the results of the SLA study have also helped us to identify the extreme difficulty English-speaking learners experience with pre- and post-verbal prepositional phrases.
Based on Pattern 4, it is learned that English-speaking learners of Chinese acquire Post-verbal
Prepositional Phrases before they acquire the Pre-verbal Prepositional Phrases to license the
GOAL of a directed motion event. The typical Chinese textbooks used in the United States seldom points out explicitly the differences between the two prepositional phrases. Students usually learn the pre-verbal PPs at some point before they come across the post-verbal PPs at a later point. The results of the study suggest that post-verbal PPs are easier for English-speaking learners to acquire. Instructors teaching Chinese as a foreign language in an English-speaking setting should consider teaching post-verbal PPs before pre-verbal PPs. After both types of prepositional phrases are taught, instructors still need to provide further explicit explanation with regard to the semantic meaning differences caused by the different syntactic distribution. In other words, I advocate that students need explicit instruction regarding when to use pre-verbal
PPs, and when to use the other.
Inevitably, the present study suffers from some limitations. The lack of successful production task in the experiment design and un-impeccable learner samples without superior proficiency level have restricted me from supporting either viewpoint with full confidence. The grammaticality judgment task helped elicit perception data, while the translation task in the pilot study elicited some informative productive data, but it was not collected from the exact same population sample. In other words, some participants took part in both the pilot study and the formal experiment; others only participated in part of the research. To further improve the research design in future studies, I would suggest a more comprehensive production task conducted on the same population who take the acceptability judgment task. The correlation
187 between the two will provide a more informative understanding of not only how learners rate or judge the acceptability of sentences, but more importantly, how they use these structures in authentic communication tasks.
To conclude, even though the present study provide no direct evidence for students’ full
awareness of Parametric Lexicalization Patterns or gradual reconfiguration with feature
assembly/ re-assembly, this study has undoubtedly enhanced our understanding of the specific
interlanguage development and acquisition processes English-speaking learners encounter in
acquiring different ways to describe causative directed motion in the target language of Chinese.
It is my belief that employing explicit and well-thought out instruction for students when
negative evidence is missing from their L2 input, (for example, when not to use the ba
construction and where not to use the pre-verbal prepositional phrases) will greatly benefit them
in their goal of acquiring a native grasp of Chinese as a spoken and written language.
188
APPENDIX A COMMON MANNER-OF-MOTION VERBS IN CHINESE
Transitive:
摆 bai ‘put’ 、搬 ban ‘move’ 、赶 gan ‘goad’ 、卷 juan ‘roll’ 、包 bao ‘wrap’ 、拉 la ‘pull’ 、落 la ‘leave…behind’ 、挪 nuo ‘move’ 、扑 pu ‘throw oneself on’ 、扔 reng ‘throw away’ 、推 tui ‘push’ 、托 tuo ‘support with (hand)’ 、拖 tuo ‘pull’ 、掀 xian ‘lift’ 、运 yun ‘transport or carry on vehicle’ 、追 zhui ‘chase’ 、钻 zuan ‘dig-into’ 、提 ti ‘lift’ 、泡 pao ‘soak’ 、抛 pao ‘throw (upward)’ 、刨 pao ‘shave’ or ‘excavate’ 、挂 gua ‘hang’ 、骑 qi ‘ride’ 、开 kai ‘drive’ 、按 an ‘press’ 、涂 tu ‘paint’ 、摸 mo ‘touch’ 、抹 mo ‘wipe’ 、抢 qiang ‘grab’ 、盛 cheng ‘fill’ or ‘scoop’ 、吸 xi ‘suck (in/up)’ 、排 pai ‘put in order’ 、压 ya ‘press’ 、押 ya ‘escort’ 、解 jie ‘release’ 、接 jie ‘catch’ 、吞 tun ‘swollow’ 、吐 tu ‘throw up/out’ 、敛 lian ‘collect’ 、翻 fan ‘turn over’ 、蹬 deng ‘tread’ 、踩 cai ‘step on’ 、踏 ta ‘tramp’ 、勒 lei ‘tie’ 、切 qie ‘cut’ 、割 ge ‘cut’ 、喷 pen ‘spray’ or ‘gush’ 、伸 shen ‘strech’ 、摇 yao ‘shake’ 、宰 zai ‘butcher’ 、砍 kan ‘hack’ or ‘slash’ 、跺 duo ‘stamp; 、剁 duo ‘chop’ 、甩 shuai ‘throw’ 、驱散 qu-san ‘disperse’ 、装载 zhuang-zai ‘load’ 、推动 tui-dong ‘push (to move)’ 、运输 yun-shu ‘transport’ 、 移(动) yi(dong) ‘move’
Intransitive:
奔 ben ‘run’ 、蹦 ben ‘jump’ 、跳 tiao ‘jump’ 、跑 pao ‘run’ 、冲 chong ‘rush’ 、闯 chuang ‘barge in/into’ 、飞 fei ‘fly’ 、滚 gun ‘roll’ 、流 liu ‘flow’ 、落 luo ‘fall’ 、爬 pa ‘climb’ 、漂 piao ‘float (on the liquid’ 、飘 piao ‘float (in the air) 、绕 rao ‘circle’ 、逃 tao ‘fleed’ 、游(泳) you(yong) ‘swim’ 、走 zou ‘walk’ 、悬 xuan ‘hang’ 、涌 yong ‘surge’ 、烫 tang ‘burn’ 、冒 mao ‘emit’ or ‘come out’ 、升 sheng ‘raise/rise’ 、摔 shuai ‘fall down’ 、散步 san-bu ‘amble’ 、漫步 man-bu ‘stroll’ 、摇摆 yao-bai ‘sway’ 、弥漫 mi-man ‘permeate’ 、延绵 yan-mian ‘strech’
189
APPENDIX B VALIDITY OF EXPERIMENT INSTRUMENT (GOAL)
Following Chapter 4, Section 4.6.5, this is part of the experiment validation. The validity of the second semantic factor GOAL is investigated here. Regardless of the transitivity of the main verb, we want to find out from the judgment of these 40 native speakers if they interpret and use those sentences with a pre-verbal prepositional phrase dao (Type 1 and Type 4) differently from those sentences with a post-verbal prepositional phrase dao in a function
indicating the goal of a motion event. Four Paired Samples T-tests were implemented: Sentence
Token Type 1 versus Type 4 for both transitive and intransitive verbs ( MANNER V + PP versus PP+
MANNER V), and Sentence Token Type 4 versus Type 6 ( MANNER V zhe +PP versus PP+ MANNER V
zhe ). The former contrasts the simple post-verbal PP with pre-verbal PP, while the latter involves
an additional element of zhe .
The results exhibited in Figure 1demonstrate that to express the GOAL of a motion event,
post-verbal prepositional phrase with dao was the only construction that native speakers accepted.
In other words, to license the GOAL of a directed manner-of-motion event, pre-verbal PP was
rejected and post-verbal PP was required. The distinctions between the pre- and post-verbal
prepositional phrases were found out to be statistically significant (See Table 1 & 2).
First, a paired-samples t-test was implemented to compare the licensing of GOAL in
Sentence Type 1 ( MANNER V + PP ) and Sentence Type 3 (PP+ MANNER V) when the verbs were 190
Figure 1. Pre- vs Post-veral PPs for GOAL
2
1.5
1
0.5 Post-verbal PP Pre-verbal PP 0 Post -verbal PP + Zhe Transitive Intransitive -0.5 Pre-verba PP +Zhe
-1
-1.5
-2
transitive. There was a significant difference in native speakers’ overall judgment ratings for
Sentence Type 1 (M= 0.59, SD= 0.75) and Sentence Type 3 (M= -1.28, SD= 0.70); t(39) =
12.154, p< .001. (Table 1 & Table 2). When the main verb was intransitive, the same pair of t- test also turned out to be significant: Sentence Type 1 (M= -1.21, SD= 0.50) and Sentence Type
3 (M= -1.66, SD= 0.40); t(39) = 5.458, p< .001.
Table 1. Descriptive Statistics of Sentence Types for GOAL Std. Std. Error Sentence Type Mean N Deviation Mean Sentence Type 1 (with Transitive Verbs) .594 40 .746 .118 Sentence Type 3 (with Transitive Verbs) -1.281 40 .698 .110 Sentence Type 4 (with Transitive Verbs) .51 40 .562 .089 Sentence Type 6 (with Transitive Verbs) -1.488 40 .493 .078 Sentence Type 1 (with Intransitive Verbs) -1.213 40 .495 .078 Sentence Type 3 (with Intransitive Verbs) -1.656 40 .399 .063 Sentence Type 4 (with Intransitive Verbs) -1.331 40 .626 .099 Sentence Type 6 (with Intransitive Verbs) -1.819 40 .300 .0474
191
Next, another paired-samples t-test was implemented to compare the licensing of GOAL in Sentence Type 4 ( MANNER V zhe +PP) and Sentence Type 6 (PP PP+ MANNER V zhe ) where the
verb was transitive. Again, there was a significant difference in the ratings for Sentence Type 4
(M= 0.51, SD= 0.56) and Sentence Type 6 with transitive verbs (M= -1.49, SD= 0.49); t(39) =
21.000, p< .001. In the case of sentences with intransitive verbs, the same pair of t-tests was also
found to be significant: Sentence Type 4 (M= -1.33, SD= 0.63) and Sentence Type 6 (M= -1.82,
SD= 0.30); t(39) = 5.420, p< .001. (Consult Table 1 & Table 2 for more information).
Table 2. Paired Samples T–test on GOAL
95% Std. Confidence Sig. Std. T-test Mean Error Interval of the t df (2- Deviation Mean Difference tailed) Lower Upper
Type 1 – Type 3 1.875 .976 .154 1.563 2.187 12.154 39 .000 (Transitive Verb)
Type 4 – Type 6 1.994 .600 .095 1.802 2.186 21.000 39 .000 (Transitive Verb)
Type 1 – Type 3 .444 .514 .081 .279 .608 5.458 39 .000 (Intransitive Verb) Type 4 – Type 6 .488 .569 .090 .306 .669 5.420 39 .000 (Intransitive Verb)
Consider the examples in (4.8) – (4.9):
4.2. Type 1 ( MANNER V + PP ) vs Type 3 ( PP+ MANNER V) a) 大象滚南瓜到农场。 dàxiàng g ǔn nánguâ dào nóngch ǎng elephant roll pumpkin to barn ‘(The) elephant rolled (the) pumpkin to (the barn).’
b) *大象到农场滚南瓜。 Dàxiàng dào nóngch ǎng g ǔn nánguâ elephant to barn roll pumpkin 192
‘(The) elephant went to (the) barn to roll pumpkin.’
c) *她走老太太到家。 tâ z ǒu l ǎotàitai dào jiâ she walk old-granny to home Intended meaning: ‘She walked (the) old granny home.’
d) *她到家走老太太。 tâ dào jiâ zǒu l ǎotàitai she to home walk old-granny Intended meaning: ‘She walked (the) old granny home.’
4.3.Type 4 ( MANNER V zhe +PP ) vs Type 6 ( PP+ MANNER V zhe ) a) 大象滚着南瓜到农场。 dàxiàng g ǔn zhe nánguâ dào nóngch ǎng elephant roll ZHE pumpkin to barn ‘(The) elephant rolling (the) pumpkin (went) to (the barn).’
b) *大象到农场滚着南瓜。 Dàxiàng dào nóngch ǎng g ǔn zhe nánguâ elephant to barn roll ZHE pumpkin ‘(The) elephant went to (the) barn rolling pumpkin.’
c) *她走着老太太到家。 tâ z ǒu zhe lǎotàitai dào jiâ she walk ZHE old-granny to home Intended meaning: ‘She (was) walking (the) old granny home.’
d) *她到家走着老太太。 tâ dào jiâ zǒu zhe lǎotàitai she to home walk ZHE old-granny Intended meaning: ‘She (was) walking (the) old granny home.’ Sentence (a) and (b) of (4.8) and (4.9) contain a transitive manner-of-motion verb as the
main verb (gǔn roll ) with a post-verbal PP and pre-verbal PP respectively, while Sentence (c) and (d) comprise an intransitive manner-of-motion verb, zǒu walk . Of the four sentences, (a) –
(d), only Sentence (a) – the transitive verb with a post-verbal prepositional phrase – was accepted.
Even though both (c) and (d) of the sentences with an intransitive verb were rejected, the one with a pre-verbal prepositional phrase as the GOAL of the motion, Sentence (d), sounded much worse than the one with a Post-verbal PP, Sentence (c) (See Figure 1). In this sense, the 193 collective judgment of the Mandarin native speakers holds that Post-verbal PP is required to denote the GOAL of the directed motion.
Example (4.8) and (4.9) differ in the use of durative marker zhe , which emphasizes the accompanied action rolling and walking . For instance, (4.8a) involves a simple use of Post- verbal PP to signify the direction of the motion, while (4.9a) marks the accompanied action with a zhe . A typical reading of Sentence (4.9a) without looking at the corresponding picture is that
The elephant went to the barn with the pumpkin rolling along with him. The mean scores exhibited in Figure 1 display that Post-verbal PP is still required to designate the GOAL in sentence where the ACCOMPANY is marked with zhe .
194
APPENDIX C VALIDITY OF EXPERIMENT INSTRUMENT (CAUSATIVITY)
Following Chapter 4, Section 4.6.5, and Appendix B, the third variable I want to examine
concerns CAUSATIVITY in the directed manner-of-motion. In English, the causative meaning
is already embraced in the main motion verb; for example, move entails CAUSE to move, walk
in the directed motion entails CAUSE to walk. In Chinese, CAUSE is often licensed by Ba
Construction (Sybesma, 1999). As discussed in Chapter 2 (Section 2.7.1), this distinguishes
English and Chinese; lexicalization of causation in English happens at lexical level, while in
Chinese it occurs at the syntactic level. To license the causative motion, Ba construction is used.
Syntactically, Sentence Token Type 1 ( MANNER V + PP ) is grammatical, so is Type 2
(ba ...+ MANNER V +PP). T he intuitions and preferences of the 40 Chinese native speakers provided us
insightful information on this lexicalization process. It turned out that the use of Ba construction
made a significant impact on the judgment of the two types of sentences. Figure 1. Use of Ba Construction for CAUSATIVITY
2
1.5
1
0.5 Transitive 0 Intransitive Type 1: MANNERV + PP Type 2: -0.5 ba...+MANNERV +PP -1
-1.5
-2
As exhibited in Figure 1, in the case of the sentences with transitive verbs, if Type 1
(MANNER V + PP) in (4.10a), was considered barely acceptable, Type 2 ( ba ...+ MANNER V +PP) 195 illustrated in (4.10b) was almost unanimously judged as perfect. Undoubtedly, Type 2 was the most preferable way to describe the directed transitive manner-of-motion event. The reason why the mean rating of Type 2 was a little below +2.0 (Very Natural) was due to individual rating differences. Some speakers were conservative raters, and they only assigned +1.0, not +2.0, for the most natural sentence. For the intransitive verbs, where a causative was unavailable in
Chinese when the verbs were intransitive, the use of the Ba Construction ( Sentence Type 2, e.g. ba ...+ MANNER V +PP) made the sentence (4.11b) sound better than one without Ba . Sentence
Type 1 with intransitive verb ( MANNER V + PP) is exemplified in (4.11a).
4.2.Transitive: Type 1 ( MANNER V + PP) vs Type 2 ( ba ...+ MANNER V +PP) a) 大象滚南瓜到农场。 dàxiàng g ǔn nánguâ dào nóngch ǎng elephant roll pumpkin to barn ‘(The) elephant rolled (the) pumpkin to (the barn).’
b) 大象把南瓜滚到农场 dàxiàng bǎ nánguâ g ǔn dào nóngch ǎng elephant BA pumpkin roll to barn ‘(The) elephant (took the pumpkin and) rolled (the) pumpkin to (the barn).’
4.3.Intransitive: Type 1 ( MANNER V + PP) vs Type 2 ( ba ...+ MANNER V +PP) a) *她走老太太到家。 tā zǒu l ǎot àitai d ào ji ā she walk old-granny to home Intended meaning: ‘She walked (the) old granny home.’
b) *她把老太太走到家。 tā bǎ l ǎot àitai z ǒu d ào ji ā she BA old-granny walk to home Intended meaning: ‘She walked (the) old granny home.’ A paired-samples t-test was implemented to compare the lexicalization of
CAUSATIVITY in Sentence Type 1 ( MANNER V + PP) and Sentence Type 2 (ba ...+ MANNER V +PP)
with transitive verbs. There was a significant difference in native speakers’ judgment ratings for
the two: Sentence Type 1 (M= 0.59, SD= 0.75) and Sentence Type 2 (M= 1.81, SD= 0.32); t(39) 196
= –11.418, p< .001. There was a statistically significant effect of CAUSATIVITY on the intransitive counterparts: Sentence Type 1 (M= -1.213, SD= 0.50) and Sentence Type 2 (M= -
0.79, SD= 0.88); t(39)= –3.887, p< .001 (Table 4.13 & Table 4.14) .
The results of these t-tests suggest that use of the Ba Construction in licensing Causatives
has a significant effect on the judged acceptability of the sentence: for sentences with both
transitive and intransitive manner-of-motion main verbs, when Ba Construction was adopted to
lexicalize the CAUSATIVITY, the acceptability was statistically much higher than otherwise.
197
Table 4.13. Descriptive Statistics of Sentence Types for CAUSATIVITY
Std. Std. Error Sentence Type Mean N Deviation Mean Sentence Type 1 (Transitive Verb) .594 40 .746 .118 Sentence Type 2 (Transitive Verb) 1.813 40 .324 .0512 Sentence Type 1 (Intransitive Verb) -1.213 40 .495 .0783 Sentence Type 2 (Intransitive Verb) -.788 40 .884 .140
Table 4.14. Paired Samples T–test on CAUSATIVITY
95% Confidence Std. Sig. Std. Interval of the Mean Error t df (2- Deviation Difference Mean tailed) Lower Upper
Sentence Type 1 – Type 2 -1.219 .675 .107 -1.435 -1.003 -11.418 39 .000 (Transitive Verb)
Sentence Type 1 – Type 2 -.425 .692 .109 -.646 -.204 -3.887 39 .000 (Intransitive Verb)
198
APPENDIX D VALIDITY OF EXPERIMENT INSTRUMENT (ACCOMPANY)
This section follows Chapter 4, Section 4.6.5, Appendix B and Appendix C. I examine the validity of the fourth variable, ACCOMPANY, in the experiment. According to Heidi and
Folli (2007), whether the Causer accompanies the directed motion event determines whether causative is available with some verbs. Since Chinese is an S-syntax language and everything must be spelled out, we are going to test if the spelled out or marked form of ACCOMPANY influences the acceptability of the sentences.
Five pairs of paired-samples t-tests were conducted (Table 1 & 2). First, a paired-samples t-test was conducted to compare the projection of ACCOMPANY in unmarked Sentence Type 1
(MANNER V + PP) in Example (4.12a) and marked Sentence Type 4 ( MANNER V zhe +PP) in (4.12b) where the main verbs were transitive. There was not a significant difference in native speakers’ judgment ratings for the two: unmarked Sentence Type 1 (M= 0. 59, SD= 0.75) and marked
Sentence Type 4 (M= 0. 51, SD= 0.56); t(39) = 0.55, p = .585. Neither did we find a significant difference in cases with intransitive verbs between unmarked Sentence Type 1 (M= -1.21, SD=
0.50), exemplified in (4.13a), and marked Sentence Type 4 (M= -1.33, SD= 0.63) in 17(b); t(39)
= 1.29, p = . 205. The results suggest that the use of zhe did not make a big difference in how
native speakers interpreted the sentences. Both (4.12a) and (4.12b) with the transitive main verb
were regarded as barely acceptable; neither (4.13a) or (4.13b) sounded natural.
Table 1. Descriptive Statistics of Sentence Types for ACCOMPANY
Std. Std. Error Sentence Type Mean N Deviation Mean Sentence Type 1 (Transitive Verb) .594 40 .746 .118 Sentence Type 4 (Transitive Verb) .51 40 .562 .089 Sentence Type 3 (Transitive Verb) -1.281 40 .698 .110 Sentence Type 6 (Transitive Verb) -1.488 40 .493 .0780 199
Sentence Type 1 (Intransitive Verb) -1.213 40 .495 .0783 Sentence Type 4 (Intransitive Verb) -1.331 40 .626 .0990 Sentence Type 3 (Intransitive Verb) -1.656 40 .399 .063 Sentence Type 6 (Intransitive Verb) -1.819 40 .300 .047 Sentence Type 7 (Intransitive Verb) 1.18 40 .751 .119 Sentence Type 8 (Intransitive Verb) 1.47 40 .569 .090
Table 2. Paired Samples T–test on ACCOMPANY
95% Confidence Interval of the Std. Std. Difference Sig. Mean Error t df Dev. (2-tailed) Mean Lo Upper wer
Type 1 – Type 4 (Transitive Verb) .088 1.004 .159 - .409 .551 39 .585 Type 3 – Type 6 (Transitive Verb) .206 .672 .106 .23- .421 1.941 39 .059 Type 1 – Type 4 (Intransitive Verb) .118 .583 .0922 .00- .305 1.289 39 .205 Type 3 – Type 6 (Intransitive Verb) .163 .352 .056 .05.06 .275 2.924 39 .006 Type 7 – Type 8 (Intransitive Verb) -.283 .611 .097 01- -.088 - 39 .006 .47 2.932
4.2.Transitive Verb: Type 1 ( MANNER V + PP) vs Type 4 ( MANNER V zhe +PP) a) 大象滚南瓜到农场。 dàxiàng g ǔn nánguâ dào nóngch ǎng elephant roll pumpkin to barn ‘(The) elephant rolled (the) pumpkin to (the barn).’
b) 大象滚着南瓜到农场 dàxiàng g ǔn zhe nánguâ g ǔn dào nóngch ǎng elephant roll ZHE pumpkin roll to barn ‘(The) elephant rolling (the) pumpkin (went) to (the barn).’
4.3.Intransitive: Type 1 ( MANNER V + PP) vs Type 4 ( MANNER V zhe +PP) a) *她走老太太到家。 tā zǒu l ǎot àitai d ào ji ā font she walk old-granny to home Intended meaning: ‘She walked (the) old granny home.’
b) *她走着老太太走到家。 tā zǒu zhe l ǎot àitai z ǒu d ào ji ā font she walk ZHE old-granny walk to home Intended meaning: ‘She walked (the) old granny home.’ 200
Another two pairs of paired-samples t-tests were conducted to compare the projection of
ACCOMPANY in unmarked Sentence Type 3 (PP+ MANNER V) in Example (4.14a) and marked
Sentence Type 6 (PP+ MANNER V zhe ) in (4.14b) with the main verb as transitive verb. Both Type
3 and Type 6 were appropriate to describe the directed motion events, as pre-verbal prepositional phrases were only used to denote stative events. Once again, there was not a significant difference in native speakers’ judgment ratings for the two sentence types with transitive verbs: unmarked Sentence Type 3 (M= -1.28, SD= 0.70) and marked Type 6 (M= -1.49, SD= 0.49); t(39) = 1.94, p = .059.
However, when the verb is intransitive, note the difference between unmarked Sentence
Type 3 (M= -1.66, SD= 0.40) exemplified in (15a) and marked Sentence Type 6 (M= -1.82, SD=
0.30) in (15b); t(39) = 2.924, p = .006. This implies that even though both sentence types were
not acceptable, the use of zhe made sentences like (15b) significantly more awkward.
4.4.Transitive Verb: Type 3 (PP+ MANNER V) vs Type 6 (PP+ MANNER V zhe ) a) *大象到农场滚南瓜 dàxiàng dào nóngch ǎng gǔn nánguâ elephant to barn roll pumpkin ‘(The) elephant (went) to (the) barn to roll(the) pumpkin.’
b) *大象到农场滚着南瓜 dàxiàng dào nóngch ǎng g ǔn zhe nánguâ elephant to barn roll ZHE pumpkin ‘(The) elephant (went) to (the) barn rolling (the) pumpkin rolled.’
4.5.Intransitive Verb: Type 3 (PP+ MANNER V) vs Type 6 (PP+ MANNER V zhe ) a) *她到家走老太太。 tā dào ji ā zǒu l ǎot àitai font she to home walk old-granny Intended meaning: ‘She (went) home to walk (the) old granny.’
b) *她到家走着老太太。 tā dào ji ā zǒu zhe l ǎot àitai z ǒu font she to home walk ZHE old-granny walk 201
Intended meaning: ‘She (went) home walking (the) old granny home.’ We ran the last paired-samples t-test to see if the markedness of the ACCOMPANY element makes a difference in the judgment of the unmarked Complex Sentence Type 7
(MANNER V + MOTION V + PP) as in (4.16a) versus marked Complex Sentence Type 8 ( MANNER Vzhe
+ MOTION V + PP) in (4.16b). The comparison turned out to be statistically significant: the former
(M= 1.18, SD= 0.75) and the latter (M= 1.47, SD= 0.57); t(39) = -2.932, p = .006. Specifically, of the two complex manner-of-motion sentences with an intransitive motion verb and additional manner verb, Chinese native speakers accept both sentences, but significantly tend to lexicalize the accompanied action with the durative marker zhe .
4.6. Intransitive Verb: Type 7( MANNER V+ MOTION V+ PP) vs Type 8( MANNER Vzhe +MOTION V + PP) a) 她陪老太太走到家。 tā péi l ǎot àitai z ǒu d ào ji ā font she accompany old-granny walk to home ‘She accompanying (the) old granny home walked home.’
b) 她陪着老太太走到家。 tā péi zhe l ǎot àitai z ǒu d ào ji ā font she accompany ZHE old-granny walk to home ‘She accompanying (the) old granny home walked home.’
202
APPENDIX E VALIDITY OF EXPERIMENT INSTRUMENT (COMPLEX MANNER)
This section follows Appendix B-D. It is the last part of the experiment validation
(Chapter 4, Section 4.6.5). Here I analyze the validity of the fifth variable, COMPLEX
MANNER, of the experiment. Of all the sentence types discussed so far, only two tokens shown
in (16a) and (16 b), are grammatical when the verb is intransitive. Unlike English which allows
intransitive verbs to participate in causative expressions, such as Sentence Type 1 (MANNER V +
PP), Chinese lexicalizes causatives of the intransitive manner-of-motion verbs in a complex way
(MANNER Vzhe +MOTION V + PP). This distinction is based on Chinese grammar, and we want to
find out if native speakers also perceive and function in the same way, which forms one of the
major goals of this section.
Another goal of this section is to further confirm that the conclusions we reached in the
last section are supported. In the last section (Appendix D), there were a total of five pairs of t-
test comparisons. Two opposite patterns stood out: First, regardless of the transitivity nature of
the main verb, no significant differences existed between the unmarked Sentence Type 1
(MANNER V + PP) and marked Sentence Type 4 ( MANNER V zhe +PP) and native speakers preferred to use the unmarked Sentence Type 1 without the marker zhe to symbolize the accompanied action. Second, a significant difference was observed between the unmarked Complex Sentence
Type 7 ( MANNER V + MOTION V + PP) and the marked Complex Sentence Type 8 ( MANNER Vzhe +
MOTION V + PP), and the same group preferred the marked form with zhe .
To further investigate the issue revealed from these contradictory statistics, in this section, the judgment results from the control group of native speakers will be analyzed with a focus on the contrast between grammatical and ungrammatical use of intransitive manner-of-motions in four sentence types: unmarked Sentence Type 1 (MANNER V + PP) and unmarked Complex 203
Sentence Type 7 ( MANNER V + MOTION V + PP) first, and then marked Sentence Type 4 ( MANNER V
zhe +PP) and marked Complex Sentence Type 8 ( MANNER Vzhe + MOTION V + PP). In other words,
we want to understand the contrast between sentences like (4.13a) and (4.16a), and that between
(4.13b) and (4.16b). Native speakers’ judgment about these sentences can also provide us some
clues on how to explain later results from non-native speakers. With these two pairs of tests, we
expect to find out when English-speaking learners of Chinese start to realize the differences
between their L1 and L2.
A paired-samples t-test (illustrated in Table 1 & 2) was conducted to compare native
speakers’ judgment of COMPLEX MANNER in Simple Sentence Type 1 ( MANNER V + PP) and
Complex Sentence Type 7 ( MANNER V + MOTION V + PP) where the main verbs are intransitive. As
expected, there was a significant difference between the two: Simple Sentence Type 1 (M= -1.21,
SD= 0.50) and Complex Sentence Type 7 (M= 1.18, SD= 0.75); t(39) = -16.243, p < .001.
Table 1. Descriptive Statistics of Sentence Types for COMPLEX MANNER
Std. Std. Error Sentence Type Mean N Deviation Mean Sentence Type 1 (Intransitive Verb) -1.213 40 .495 .078 Sentence Type 7 (Intransitive Verb) 1.18 40 .751 .119 Sentence Type 4 (Intransitive Verb) -1.331 40 .626 .099 Sentence Type 8 (Intransitive Verb) 1.47 40 .569 .090
Table 2. Paired Samples t–test on COMPLEX MANNER
Std. 95% Confidence Std. Sig. Mean Error Interval of the t df Dev. (2-tailed) Mean Difference Low Upper Type 1 – Type 7 (Intransitive Verb) -2.396 .933 .148 er- -2.097 -16.243 39 .000 2.69 Type 4 – Type 8 (Intransitive Verb) -2.798 .841 .133 - -2.529 -21.053 39 .000 3.06
204
Another paired-samples t-test (Table 1 & 2) was also conducted to compare their judgment on COMPLEX MANNER marked with zhe in Simple Sentence Type 4 ( MANNER Vzhe
+PP) and Complex Sentence Type 8 ( MANNER Vzhe + MOTION V + PP). Another significant difference was observed between the two: Simple Sentence Type 4 (M= -1.33, SD= 0.63) and
Complex Sentence Type 8 (M= 1.47, SD= 0.57); t(39) = -21.053, p < .001. These two significant differences demonstrate that the judgment task is valid. It captures the distinction between grammatical and ungrammatical sentences.
205
REFERENCE Andersen, Roger. W. 1990. Papiamentu tense-aspect, with special attention to discourse. In Pidgin and creole tense-mood-aspect systems , ed. by J.V. Singler. pp. 59-96. Amsterdam, The Netherlands: John Benjamins. Ayoun, Dalila. 1998. Revisiting traditional assumptions in principles-and-parameters theory. Arizona Working Papers in SLAT , 61-79. Bai, Guo-Fang. (2007). A comparative study of verb lexicalization pattern and degree in English and Chinese. Journal of Pingdingshan Institute . 100-128. Baker, Mark. 1987. Incorporation . Chicago: Chicago University Press. Beck, Sigrid. & William Snyder. 2001. Complex predicates and goal PP’s: Evidence for a semantic parameter. In Proceedings of the 25 th Annual Boston University Conference on Language Development, ed. by A.H.-J. Do, L. Dominguez, and A. Johansen, BUCLD 25, Vol 1, pp.114-122. Somerville, MA: Cascadilla Press. Bender, Emily. 2000. The syntax of Mandarin Ba : Reconsidering the verbal analysis. Journal of East Asian Linguistics , 9, 105-145. Borer, Hagit. 2005. Structuring sense, Volume II: The normal course of events . Oxford: Oxford University Press. Butt, Miriam. 1995. The structure of complex predicates . Stanford: CSLI Publications. Butt, Miriam. 2004. The light verb jungle. In Papers from the GSAS/ Dudley house workshop on light verbs , ed. by G. Aygen, C. Bowern. & C. Quinn, 1-50. Cambridge, Harvard Working papers in Linguistics. Butt, Miriam. & Wilhem Geuder. 2001. On the (semi)lexical status of light verbs. In Semi-lexical Categories: The function of content words and the content of function words , ed. by N. Corver, & H.v. Riemsdijk, 323 – 370. New York: Mouton de Gruyter. Carrier, Jill. & Janet H. Randall. 1992. The argument structure and syntactic structure of resultatives. Linguistic Inquiry 23, 173-234. Chang, Jung-Hsing. 2003. Event structure and argument linking in Chinese. Language and linguistics , 4(2), 317-351. Chao, Yuen Ren. 1968. A grammar of spoken Chinese . Berkeley: University of California Press. Chen, Liang. 2005. The Acquisition and Use of Motion Event Expressions in Chinese . Doctoral dissertation, University of Louisiana at Lafayette. Cheng, Lisa L-S. & C-T. James Huang. 1995. On the argument structure of resultative compounds. In In honor of William Wang: Interdisciplinary studies on language and language change, ed. By Matthew Chen and Ovid Tzeng, 187-221. Taipei: Pyramid Press. Cheng, Lisa L-S. & ,C-T. James Huang., Y.-H. Audrey Li, and C. –C. Jane Tang. 1997.
206
Causative compounds across Chinese dialects: a study of Cantonese, Mandarin and Taiwanese. 《中国境内语言及语言学》4, 199-224. Chomsky, Noam. 1986. Knowledge of language . New York: Praeger. Chomsky, Noam. 1981. Lectures on government and binding . Dordrecht: Foris. Chomsky, Noam. 1995. The minimalist program . Cambridge, Mass.: MIT Press. Chomsky, Noam. 1998. Minimalist inquiries: the framework. MIT working papers in linguistics . Cambridge, MA: MITWPL, 15, p.1-56. Chomsky, Noam. 2005. Three factors in language design. Linguistics Inquiry , 36, p. 1-22. Chu, Cheng-Zhi. 2004. Event Conceptualization and Grammatical Realization: The Case of Motion in Mandarin Chinese . Doctoral dissertation, The University of Hawaii. Corver, Norbert. & Henk. C.Van Riemsdijk. 2001. Semi-lexical Categories: The function of content words and the content of function words . New York: Mouton de Gruyter. Cowper, Elizabeth. 2005. The geometry of interpretable features: Infl in English and Spaish. Language , 81, 10-46. Dowty, David. 1979. Word meaning and Montague grammar. Dordrecht, The Netherlands: D. Reidel Publishing Company. Dowty, David. 1991. Thematic proto-roles and argument selection . Language 67, 547-619. Du, Hang. 2004. The acquisition of the Chinese ba-construction by adult second language learners . Doctoral dissertation, University of Arizona. Erbaugh, Mary. 1985. The acquisition of Mandarin. In The crosslinguistic study of language acquisition, ed. by Slobin, D.I. Vol. 3. (pp.373- 456). Mahwah, New Jersey: Lawrence Erbaum Associates. Fillmore, Charles. 1968. The case for case. In Bach, E. & Harms, R. (Eds.) Universals in linguistic theory . (pp. 1-88.). Holt, Rinehart and Winston. Fillmore, Charles. et al. 1997-2002. FrameNet. Retrieved in May, 2008 http://www.framenet.icsi.berkeley.edu/ . Folli, Raffaella. & Gillian Ramchand. 2005. Prepositions and results in Italian and English: An analysis from event decomposition. In Perspectives on aspect, ed. by H.J. Verkuyl, H. de Swart and A. van Hout, 81-105. Netherlands: Springer. Folli, Raffaella. & Heidi Harley. 2006. On the licensing of causatives of directed motion: Waltzing Matilda all over. Studia Linguistica, 60 (2), 121-155. Goldberg, Adele. 1995. Constructions: A Construction Grammar Approach to Argument Structure. Chicago: University of Chicago Press. Grimshaw, Jane. 1990. Argument Structure . Massachusetts: MIT Press. Grimshaw, Jane. 1993. Semantic structure and semantic content. Unpublished manuscript, Rutgers University, New Jersey. Gu, Yang. 1992. The syntax of resultative and causative compounds in Chinese. Doctoral
207
dissertation, Cornell University. Hale, Ken. and Samuel Keyser. J. 1993. On argument structure and the lexical expression of syntactic relations. In The view from building 20: Essays in linguistics in honor of Sylvain ed, by Bromberger, Cambridge, K. Hale and S. J. Keyser, Massachusetts: MIT Press. Hale, Ken. and Samuel Keyser. J. 1998. The basic elements of argument structure. In Papers from the UPenn/ MIT rountable on argument structure and aspect: MIT working papers in linguistics 32, ed. by H. Harley, 73-118, Cambridge, Massachusetts: MITWPL. Hale, Ken. and Samuel Keyser. J. 1999a. Bound features, merge, and transitivity alternations. In L. Pylkkanen, A. V. Hout (eds.), Papers from the UPenn/ MIT rountable on lexicon: MIT working papers in linguistics 35, (pp. 73-118), Cambridge, Massachusetts: MITWPL. Hale, Ken. and Samuel Keyser. J. 1999b. A response to Fodor and Lepore, ‘impossible words?’. Linguistic Inquiry , 30, 453-466. Hale, Ken. and Samuel Keyser. J. 2002. Prolegomenon to a theory of argument structure. Cambridge, Massachusetts: MIT Press. Han, Da-Wei. 2007. Lexicalization Patterns of “Path”. Journal of Northeast Normal University , 227 (3), 155-159. Hegarty, Michael. 2005. A feature-based syntax of functional categories: the structure, acquisition, and specific impairment of functional systems . New York: Mouton de Gruyter. Hay, Jennifer., Kennedy, Christopher. & Levin, Beth. 1999. Scalar structure underlies telicity in ‘degree achievement’. In Proceedings of SALT IX . ed. by Mathews, T. & Strolovitch, D. 127-144. Ithaca: CLC Publications. He, Hong-Feng. 2006. Shift of adverbials of manner into attributes in Modern Chinese. Journal of Jianghan University . 25(6), 66-70. Hsiao, Hui.-Chen Sabrina. 2009. Motion event descriptions and manner-of-motion verbs in Mandarin . Doctoral dissertation, State University of New York at Buffalo. Hopper, Paul.J. & Thompson, Sandra. A. 1980. Transitivity in grammar and discourse. Language , 56(2), 251-299. Huang, C.-T. James. 1992. Complex predicates in control. In Control and grammar, ed. By James Higginbotham et al, 109-147. Dordrecht: Kluwer Academic Publishers. Huang, C.-T. James. 1994. More on Chinese word order and parameter theory. In Syntactic theory and first language acquisition: cross-linguistic perspectives, ed.by B. Lust, M. Suner, & J. Whitman., 15-35. Lawrence Erlbaum Associates, Publishers. Huang, C.-T. James. 1997. On lexical structure and syntactic projection. Chinese Languages
208
and Linguistics 3:45-89. Huang, C.-T. James. 2006. Resultatives and Unaccusatives: a Parametric View. Bulletin of Chinese Linguistic Society of Japan . 253. 1-43. Huang, C.-T. James. 2007. Thematic Structures and Syntactic Projections of Chinese Verbs. Linguistic Sciences . 汉语动词的题元结构与其句法表现. 《语言科学》 Huang, C.-T. James., Y.-H. Audrey Li, and Yafei Li. 2009. The Syntax of Chinese. Cambridge University Press. Hui, Chang. 2011. A study of the acquisition of Chinese causative constructions by native English speakers. Chinese Teaching in the World . Inagaki, Shunji. 2001. Motion verbs with goal PPs in the L1 acquisition of English and Japanese. Studies in Second Language Acquisition , 23, 153-170. Iwasta, Seizi. 2002. Does MANNER count or not?: Manner-of-motion verbs revisisted. Linguistics , 40(1), 61-110. Jackendoff, Ray. 1972. Semantic interpretation in Generative Grammar. Cambridge, MA: MIT Press. Jackendoff, Ray. 1983. Semantics and cognition . Cambridge, MA: MIT Press. Jackendoff, Ray. 1990. Semantic structures . Cambridge, Mass: MIT Press. Johnson, Mark. 1987. The body in the mind: The bodily basis of meaning, imagination, and reason . Chicago: University of Chicago Press. Juffs, Alan. 1996. Semantics-syntax correspondences in second language acquisition. Second Language Research, 12(2) . 177-221. Ke, Chuan-ren. 2005. Patterns of acquisition of Chinese linguistic features by CFL learners. Journal of Chinese Language Teachers Association , Vol 40, 1-24. Kayne, Richard.S. 2008. Some notes on comparative syntax, with special reference to English and French. In Cinque, G. and Kayne, R.S. (eds.), The Oxford handbook of comparative syntax . New York: Oxford University Press, p. 3-69. Klein, Wolfgang., Ping Li., and Henriette Hendriks. 2000. Natural Language & Linguistic Theory , 18: 723-770. Kratzer, Angelika. 2004. Building resultatives . Ms. Amherst, Mass. Krifka, Manfred. 1998. The origins of telicity. Rothstein, S. (ed.) Events and grammar. 197-236, Dordrecht: Kluwer. Lakoff, George. 1987. Women, fire and dangerous things . Chicago: University of Chicago Press. Lakoff, George. & Mark Johnson. 1999. Philosophy in the flesh . New York: Basic Books. Langacker, Ronald. 1987. Foundations of cognitive grammar , Vol.I. Stanford, CA: Stanford University Press. Langacker, Ronald. 1991. Foundations of cognitive grammar . Vol. II. Stanford, CA: Stanford
209
University Press. Larson, Richard. 1988. On the double object construction. Linguistic Inquiry , 19, 335 -391. Lardiere, Donna. 2008. Feature assembly in second language acquisition. . In J.M. Liceras, H. Zobl, & H.Goodluck (Eds.), The role of formal features in second language acquisition (pp. 106-140). New York: Lawrence Erlbaum Associates. Lardiere, Donna. 2009a. Some thoughts on the contrastive analysis of features in second language acquisition. Second Language Research,25(2), 173-227. Lardiere, Donna. 2009b. Further thoughts on parameters and features in second language acquisition: a reply to peer comments on Lardiere’s ‘Some thoughts on on the contrastive analysis of features in second language acquisition’ in SLR 25(2). Second Language Research,25(3), 409-422. Levin, Beth. 1993. English verb classes and alternations. Chicago: University of Chicago Press. Levin, Beth. 2000. Aspect, lexical semantic representation, and argument expression . Proceedings of the 26 th meeting of the Berkeley Linguistic Society. Levin, Beth. and Malka Rappaport Hovav. 1995. Unaccusativity: At the syntax-lexical semantics interface . Cambridge, MA: MIT Press. Levin, Beth. and Malka Rappaport Hovav. 1998. Building verb meanings. In The projection of arguments: Lexical and conceptual factors , ed. by Miriam Butt and Wilhelm Geuder, 97-134. Stanford: CSLI Publications. Levin, Beth. and Malka Rappaport Hovav. 1999. Two structures for compositionally derived events. In Proceedings of SALT 9, ed. by T.Mathews & D. Strolovitch, 199-223. Ithaca: CLC Publications. Li, Charles and Sandra Thompson. 1981. Mandarin Chinese: A Functional Reference Grammar. Berkeley: University of California Press. Li, Yafei. 1990. On V-V Compounds in Chinese. Natural Language and Linguistic Theory 8. 177-207. Li, Yafei. 1991. The syntactic representation of causativization. Ms., Cornell University. Li, Yafei. 1995. The thematic hierarchy and causativity. Natural Language and Linguistic Theory 13. 255-282. Li, Yafei. 1998. Chinese resultative constructions and uniformity of theta assignment hypothesis. In New approaches to Chinese word formation: morphology, phonology and the lexicon in modern and ancient Chinese . ed. by J. L. Packard, 285-310. Berlin: Mouton de Gruyter. Li, Yafei. 1999. Cross-componential causativity. Natural Language and Linguistic Theory 17. 445-497. Li, Yafei. 2005. X0: A theory of the morphology-syntax interface . The MIT Press.
210
Liceras, Juana.M., Zobl, Helmut. & Goodluck, Helen. 2007. (eds.). The role of formal features in second language acquisition. New York: Lawrence Erlbaum Associates. Lin, Tsong-Hong. Jonah. 2001. Light verb syntax and the theory of phrase structure . Doctoral dissertation, University of California Irvine. Lin, Jimmy. 2004. Event structure and the encoding of arguments: The syntax of the Mandarin and English verb phrase . Doctoral dissertation, MIT. Liu, Feng-hsi. 1997. An aspectual analysis of ba . Journal of East Asian Linguistics , 6(1),51-99. Liu, Feng-hsi. 1998. A Clitic analysis of locative particles in Chinese. Journal of Chinese Linguistics , Vol 26, 48-70. Liu, Feng-hsi. 2007. Word order variation and ba sentences in Chinese. Studies in language , 31(3), 649-682. Liu, Feng-hsi. 2009. Aspect and the postverbal zai phrase in Mandarin Chinese. In Janet Zhiqun Xing (ed.). Studies of Chinese Linguistics: Functional Approaches . (pp. 103-129). Hong Kong: Hong Kong University Press. Matsumoto, Yo. 2003. Typologies of lexicalization patterns and event integration: Clarifications and reformulations. In Shuji Chiba et al. (eds.) Empirical and theoretical investigation into language: A Festschrift for Masaru Kajita , (pp. 403-418). Tokyo: Kaitakusha. Matsumoto, Yo. 1997.Linguistic expressions of spatial motion and their extentions. In S. Tanaka & Y. Mastumoto, Kuukan to Idoo no Hyoogen , (pp.126-229), Kenkyusha. Montrul, Silvina. 1999. Causative errors with unaccusative verbs in L2 Spanish. Second Language Research , 15(2), 191-219. Montrul, Silvina. 2001. First-language-constrained variability in the second-language acquisition of argument-structure-changing morphology with causative verbs. Second Language Research , 17(2), 144-194. Montrul, Silvina. 2001. Agentive verbs of manner of motion in Spanish and English as second languages. Studies of Second Language Acquisition , 23, 171-206. Montrul, Silvina. 2009. Putting parameters in their proper place. Second Language Research 25(2), 291-311. Narasimhan, Bhuvana. 2003. Motion events and the lexicon: a case study of Hindi. Lingua, 113, 123-160. Narasimhan, Bhuvana. 2008. Unaccusative or unergatie? Verbs of manner of motion. Retrieve in November, 2008 at http://alphalinguistica.sns.it/QLL/QLL96/BNVDTCMV.Unaccusative.pdf . Norman, Jerry.1988. Chinese . Cambridge: Cambridge University Press.
211
Oshita, Hiroyuki. 2001. The unaccusative trap in second language acquisition. SSLA 23, 279-304. Palmer, Martha. 1995. Verb Semantics for English-Chinese Translation. Machine Translation . 10. 59-92. Peyraube, Alain. 2006. Motion events in Chinese: A diachronic study of directional complements. In Hickmann, Maya and Stéphane Robert (eds.), Space in Languages: Linguistic Systems and Cognitive Categories , pp. 121–135. Pinker, Steven. 1989. Learnability and Cognition: The Acquisition of Argument Structure. Cambridge, MA: MIT Press. Schwartz, Bonnie. & Eubank, Lynn. 1996. What is the “L2 initial state”?. Second Language Research, 12, 1-5. Schwartz, Bonnie. & Sprouse, Rex. 1996. L2 cognitive states and the full transfer/ full access model. Second Language Research, 12, 40-72. Shen, Yang. & Sybesma, Rint. 2005. Small clause results and the internal structure of the Chinese resultative small clause . Ms., Beijing University and Leiden University. Simpson, Jane. 1983. Resultatives. In Papers in lexical-functional grammar , ed. by L. Levin et al, 143-157. Bloomington, Indiana: Indiana University Linguistics Club. Slabokova, Roumyana. 2009. Features or parameters: which one makes second language acquisition easier, and more interesting to study? Second Language Research,25, 313-324. Slobin, Dan. I. The Universal, the Typological, and the Particular in Acquisition. In Slobin, D.I. (Ed.) The crosslinguistic study of language acquisition: Vol. 5. Expanding the contexts . (pp.1- 40). Mahwah, New Jersey: Lawrence Erbaum Associates. Slobin, Dan. I. 1996. From “thought and language” to “thinking for speaking”. In Gumperz, J., John, J. & Levinson, S. C. (Eds.), Rethinking linguistic relativity. (pp.70-96). Cambridge, MA: Cambridge University Press. Slobin, Dan. I. 2003. The many ways to search for a frog: Linguistic typology and the expression of motion events. In S. Stromqvist & L. Verhoeven (Eds.) Relating events in narratives: Typological and contextual perspectives . Mahwah, NJ: Lawrence Erlbaum Associates. Slobin, Dan. I. 2004. The many ways to search for a frog: Linguistic typology and the expression of motion events. In S. Stromqvist & L. Verhoeven (Eds.), Relating events in narrative: Typological and contextual perspectives (pp. 219-257). Mahwah, NJ: Lawrence Erlbaum Associates. Smith, Carlota. S. (1991). The parameter of aspect . Dordrecht, The Netherlands: Kluwer Academic Publishers. Sybesma, Rint. 1992. Causatives and accomplishments: the case of Chinese ba . Doctoral
212
dissertation, Leiden University. Sybesma, Rint. 1999. The Mandarin VP . Dordrecht, The Netherlands: Kluwer Academic Publishers. Tai, J. H.-Y. 1975. On the two functions of place adverbials in Mandarin Chinese. Journal of chinese linguistics 3, 154-79. Tai, J. H.-Y. 1993. Iconicity: motivation in Chinese grammar. Eid Mushira & Gregory Iverson (eds.), Principles and Prediction: The Analysis of Natural Languages. Amsterdam: John Benjamins Publishing Company, pp. 153-174. Tai, J. H.-Y. 1993. Conceptual structures of Chinese spatial expressions. Chicago Linguistic Society, 29(2), 347-362. Tai, J. H.-Y. 2003. Cognitive relativism: Resultative construction in Chinese. Language and Linguistics , 4(2), 301-316. Talmy, Leonard. 1975. Semantics and syntax of motion. In J.P. Kimball (ed.), Syntax and Semantics , Vol. 4, (pp. 181-238). New York NY: Academic Press. Talmy, Leonard. 1983. How language structures space. In H.L. Pick & L. P. Axredolo. (Eds.), Spatial orientation: Theory, research and application . New York: Plenum Press, 225-282. Talmy, Leonard. 1985. Lexicalization patterns: Semantic structure in lexical forms. In T. Shopen, (ed.), Language typology and syntactic description (3):Grammatical categories and the lexicon. Cambridge: Cambridge University Press. 36-149. Talmy, Leonard. 1991. Path to realization: A typology of even conflation. Proceedings of the Seventeenth Annual Meeting of the Berkeley Linguistics Society, 480-519. Berkeley, CA: Berkeley Linguistics Society. Talmy, Leonard. 2000. Toward a cognitive semantics , vol. 1 & 2. Cambridge, MA: The MIT Press. Tang, Sze-Wing. 1997. The parametric approach to the resultative construction in Chinese and English. UCI Working Papers in Linguistics 3, 203-226. Tenny, Carol. 1987. Grammaticalizing aspect and affectedness . MIT. PhD Dissertation. Thompson, Sandra. 1973. Transitivity and the ba construction in Mandarin Chinese. Journal of Chinese Linguistics 1:208-221. Tieu, Lyn Shan. 2007. Transitivity requirements in Chinese: Putting the generic object in context. Proceedings of the 2007 annual conference of the Canadian Linguistic Association . Tomioka, Naoko. 2006. Resultative constructions: Cross-linguistic variation and the syntax-semantics interface. Doctoral dissertation, McGill University. Travis, Lisa. 2000. Event structure in syntax. In C. Tenny and J. Pustejovsky, (eds.), Events as grammatical objects: the converging perspectives of lexical semantics and syntax,
213
145-185. Stanford: CSLI. Travis, Lisa. 2008. The role of features in syntactic theory and language variation. In J.M. Liceras, H. Zobl, & H.Goodluck (Eds.), The role of formal features in second language acquisition (pp. 22-47). New York: Lawrence Erlbaum Associates. Troike, Rudolph C. & Pan, Junlin. (1995). In Defense of P': The Status of Chinese Locative Particles. In Camacho, Jose. & Choueiri, Lina. (eds.) Proceedings of the 6 th North American Conference on Chinese Linguistics . Vol 1. Tsang, Chi Chung Aaron. 2003. Transitivity in Cantonese . Doctoral dissertation, University of Hawaii. Vainikka, Anne. & Young-Scholten, Martha. 1996a. Gradual development of L2 phrase structure. Second Language Research, 12, 7-39. Vainikka, Anne. & Young-Scholten, Martha. 1996b. The early stages in adult L2 syntax: additional evidence from Romance speakers. Second Language Research, 12, 140-76. Vendler, Zeno. 1967. Linguistics in philosophy . Ithaca, New York: Cornell University Press. Wang, Lingling. 2001. A study of resultative constructions in Mandarin Chinese . Doctoral dissertation, The Hong Kong Polytechnic University. Wang, W.B. & Zhou, C.B. 2004. A comparison of "LOOK" type verbs in English and Chinese: Lexical meanings and lexicalization. Foreign Language Teaching and Research , 36(6). White, Lydia. 2003. Second language acquisition and universal grammar . Cambridge University Press. White, Lydia. 2009. Some questions about feature re-assembly. Second Language Research , 25 (2), 343-348. Wienold, Gotz. 1995. Lexical and conceptual structures in expressions for movement and space: with reference to Japanese, Korean, Thai, and Indonesian as compared to English and German. In U. Egli et al. (eds.) Lexical knowledge in the organization of language , (pp. 301-340). John Benjamins. Wu, Xiao-Qi. 1996. ‘ Zai ’, ‘ Dao’ , and ‘ Gei’ Constructions – A Study of Chinese Word Order. Dissertation: The Ohio State University. Yan, Li. 2004. The syntax of the ba -constructions. Kansas working papers in linguistics , Vol. 27, 92-109. Yeh, Meng. 1993. The stative situation and the imperfective zhe in Mandarin. Journal of the Chinese Language Teachers Association , 28, 69-98. Yuan, Bo.-Ping. 2001. The status of thematic verbs in the second language acquisition of Chinese: against inevitability of thematic-verb raising in second language acquisition. Second Language Research, 17, 248-272. Yuan, Bo.-Ping. 1999. Acquiring the unaccusative/ unergative distinction in a second
214
language: evidence from English-speaking learners of L2 Chinese. Linguistics , 37-2, 275-296. Yuan, Bo.-Ping. 2010. Domain-wide or variable-dependent vulnerability of the semantics-syntax interface in L2 acquisition? Evidence from wh-words used as existential polarity words in L2 Chinese grammars. Second Language Research, 26, 219-260. Yu, Li-Ming. 1997. The role of cross-linguistic lexical similarity in the use of motion verbs in English by Chinese and Japanese learners . Doctoral dissertation, University of Toronto. Zou, Ke. 1994. Resultative V-V Compounds in Chinese. MIT working papers in linguistics 22, 271-290. Zhu, Pei. 2007. A general introduction of motion verbs in Modern Chinese. Journal of Mudanjiang Normal University . 137(1). 59-61. Zhu, De-xi. 1981. ‘Writing on the blackboard’ and relevant sentences. Language teaching and research , Vol 1, 4-18. 在黑板上写字及相关句式《. 语言教学与研究》. 第 1 期， 4-18 页. Zubizarreta, Maria Luisa. A compositional analysis of manner-of-motion verbs in Italian. Retrieved in October, 2008 from http://www-rcf.usc.edu/~zubizarr/MLZ.Papers%20&%20Chapters/LSRL36.final.pdf Zubizarreta, Maria Luisa. & Eunjeong Oh. 2007. On the syntactic composition of manner and motion. The MIT Press: Cambridge.