中文條件句之第一語言習得

全文

(1)國立臺灣師範大學英語學系碩士論文 Master Thesis Graduate Institute of English National Taiwan Normal University. 中文條件句之第一語言習得 First Language Acquisition of Mandarin Chinese Conditionals. 指導教授：陳純音博士 Advisor: Dr. Chun-yin Doris Chen 研究生：呂建慧 Student: Rose Chien-hui Lu 中華民國一百零一年七月 July, 2012 .

(2) 摘要. 本研究旨在探討以中文為母語的兒童對條件句之第一語言習得。主要研究五種中文條件句在標記性、語境、題型及年齡是否會影響兒童對此結構的習得。為了避免潛在的實驗效應，本研究採用兩種實驗題型，解釋測驗與模仿測驗。研究對象包含實驗組：依年齡三到七歲分為五組，每組 18 人，共 90 人；及對照組： 18 位中文為母語的成人。實驗結果顯示，在條件句的習得中，標記性、語境、題型及年齡皆對兒童認知及表達造成影響。五種條件句可分為無標記性及標記性。對兒童而言，無標記性的句首條件副詞條件句最容易習得，而最讓兒童感到困難的則是標記性的句末條件副詞條件句。此現象同時印證了無標記條件句較標記條件句較早習得之假設。在語境方面，受試者在假設條件句表現地較反事實條件句佳，呼應文獻中兒童的認知發展模式為假設先於反事實。而題型效應則顯示受試者在解釋題型中明顯表現地比模仿測驗好，說明此結構的理解優先於表達。此外，模仿測驗的其他句型分析中，五種不同於標的句型大多來自於受試者較不熟悉的標記性條件句。而受試者最常使用插入策略，亦即插入無標記性的句首條件副詞。在年齡效應方面，六歲是兒童習得中文條件句的轉變期，同時也於此時達到成熟階段。. 關鍵字：母語習得、條件句、標記性、題型效應、年齡效應 i .

(3) ABSTRACT In first language acquisition, conditionals are regarded as one of the most complicated syntactic constructions, enjoying extensive discussion in the literature. However, few researchers have conducted an empirical study to investigate the competence and performance of Chinese children’s acquisition of conditional sentences. Therefore, the present study aims to probe into Chinese children’s development by investigating the markednesss issues, scenario differences, task effects, production analysis and age effects on the five types of conditionals. A comprehension task (i.e., interpretation task) and a production task (i.e., imitation task) were assigned to 90 Chinese children (aged 3-7), divided into five age groups, and a control group of 18 Chinese speaking adults. The overall results indicated that the markedness, scenarios, tasks and age were determinant factors in the acquisition of Mandarin Chinese conditionals. The five types exhibited different degrees of difficulty. Type 2 (i.e., clause-initial adverbial conditionals) was found the easiest, and Type 1 (i.e., double adverbial conditionals) and Type 4 (i.e., non-wh-word conditionals) were relatively easier than Type 5 (wh-word conditionals) and Type 3 (i.e., clause-final adverbial conditionals). The results also showed that children performed better on the unmarked types (Types 1, 2 and 4) while the marked types (Types 3 and 5) were found challenging. Concerning the scenarios, it was found that the hypothetical conditionals were acquired prior to the counterfactual conditionals, a result in accord with the literature. As for the task effects, the subjects performed the comprehension task significantly better than the production task. With regard to the production data, it was found that most of the children’s production resulted from Type 3 and Type 5. In addition, Insertion was commonly used by most subjects and they tended to add an unmarked clause-initial ii .

(4) conditional adverbial ruguo ‘if’ to produce typical conditional sentences. Finally, it was found that age six was a cutting point where our children performed significantly better than preschoolers and reached the adult grammar.. Keywords: L1 acquisition, conditionals, markedness, task effects, age effects. iii .

(5) . ACKNOWLEDGEMENTS The thesis could not have been achieved without many people’s tender guidance and generous assistance. It is with pleasure that I acknowledge their support. First and foremost, I would like to express my sincere gratitude to my thoughtful thesis advisor, Dr. Chun-Yin Doris Chen, who led me to the field of first language acquisition. Since I was in her class, Dr. Chen’s humorous expressions and excellent analysis of linguistic research sparked my interest in L1 acquisition. When I was depressed during the thesis writing, she never lost her constant belief in me. Her countless patience, opportune encouragement and unwavering optimism always comforted my diffident spirits and helped me continue working energetically throughout the bittersweet process of accomplishing this thesis. She truly is my mentor but words cannot express my deepest gratitude. I am also grateful to my considerate committee members, Dr. Jen-Yi Li and Dr. Rueih-Ling Sharon Fahn. Their scrutiny of my thesis proposal accounted for the appropriate design of the experiments and the unbiased results. Their gentle reminder and insightful suggestions enlightened the direction of the revision and improved the quality of the thesis. Their kind advice really plays a significant role of the study. Sincere appreciation is also given to the teachers who taught me during the graduate studies in NTNU, Dr. Miao-Hsia Tammy Chang, Dr. Miao-Ling Hsieh, Dr. Su-Kai Hsieh, Dr. Ing Cherry Li, Dr. Chen-Jer Charles Lin, Dr. Hsieh-O Lin, Dr. Kwock-Ping John Tse, and Dr. Jing-Lan Joy Wu by alphabetic order. I am also appreciative to Dr. Mei-Jin Lillian Huang when I was an undergraduate. All of these teachers’ intellectual instruction and precise guidance introduced me to the field of linguistics and inspired me to finish the thesis. Furthermore, a great indebtedness is owed to the passionate participation of my iv .

(6) subjects recruited from Taoyuan Green Kindergarten, Taoyuan Little Angle Kindergarten and Taoyuan Da-Lun Elementary School. Special thanks also go to my warmhearted friends Tai-Yuan Tony Huang, Xing-Lun Liu and Wei-De Wu for their kindly assistance of recording the designed stories in the experiments. My heartfelt appreciation also extends to the colleagues of Taoyuan Da-Lun Junior High School, who rendered their invaluable time to participate in the tasks and encouraged me to complete the thesis. Special thanks go to the colleagues in the Office of Student Affairs because they shared my distress when I felt dispirited. My warm thanks are also given to my loving friends Chien-Fu Chiou and Hui-Ting Xu because they greatly smoothed the difficulty during the period. The thesis would not have been completed but for their cordial participation and enthusiastic assistance. Moreover, thanks go to all my classmates in NTNU: Justin Chang, Marco Chang, Wallace Chang, Ray Chen, Adam Chou, Linda Huang, Patricia Huang, Rebecca Li, Terry Li, Ansel Lin, Mars Lin, Martha Lin, Mina Lo, Debby Pan, Jenny Tung and Joseph Yu by alphabetic order. My graduate school life would have been humdrum without their company. Special thanks go to Ansel Lin, who promised me to code the raw data without hesitation even though he was on vacation and Joseph Yu, who kindly encouraged me with his marked terms and participated in my presentation. Last but most importantly, I would like to thank my beloved parents and brothers for their gentle understanding and unconditional love. They were always tolerant of my emotional swings and provided me with attentive care even though it was midnight when I painstakingly reanalyzed the data. When I needed them, they were always there to support my attempt of the linguistic study and helped me overcome the tough problems. I realize that their endless love build the base of the thesis and I cannot repay them for what they have devoted to me in my whole life. To my family, I dedicate the thesis. v .

(7) TABLE OF CONTENTS CHINESE ABSTRACT………………………………………………………………. i ENGLISH ABSTRACT……………………………………………………………….ii ACKNOWLEDGEMENTS…………………………………………………………..iv TABLE OF CONTENTS……………………………………………………………..vi LIST OF TABLES.……………………………………………………….…………...ix LIST OF FIGURES……………………………………………………………………x LIST OF ABBREVIATIONS…………………………………………………………xi Chapter One Introduction…………………………………………………………...1 1.1 Motivation and Purpose………...……………………………………………1 1.2 Research Questions…………………………………………………..5 1.3 Significance of the Study.…………………………………………………….6 1.4 Organization of the Thesis….………………………………………………...6 Chapter Two Literature Review and Classification of Chinese Conditionals...…….7 2.1 Markedness Theory……..……………………………………………………7 2.2 Previous Theoretical Studies of Chinese Conditionals………………………9 2.2.1 Li and Thompson (1981)……………………………………………...9 2.2.2 Wu (1994)……………………………………………………………12 2.2.3 Cheng and Huang (1996)…………………………………………... 15 2.2.4 Chierchia (2000)……………………………………………………..19 2.2.5 Su (2005)……….……………………………………………………21 2.2.6 Summary…………………………………………………………….22 2.3 Previous Empirical Studies on L1 Acquisition of Conditionals…………….23 2.3.1 Reilly (1986)………………………………………………………...23 2.3.2 Harris, German and Mills (1996)……………………………………26 2.3.3 Crutchley (2004)…………………………………………………….29 2.3.4 Perner, Sprung and Steinkogler (2004)……………………………...32 2.3.5 Summary…………………………………………………………….35 2.4 Classification of Conditionals in Mandarin Chinese………………………..35 2.4.1 Double Adverbial Conditionals………………..…………………….36 2.4.2 Clause-initial Adverbial Conditionals………………..……………...38 2.4.3 Clause-final Adverbial Conditionals…………………...……………39 2.4.4 Non-wh-word Conditionals………………………...………………..41 2.4.5 Wh-word Conditionals…………………………………………...….42 2.4.6 Summary…………………………………………………………….43 vi .

(8) 2.5 Summary of Chapter Two…………………………………………………..43 Chapter Three Research Design…………………………………………………...45 3.1 Subjects……………………………………………………………………..45 3.2 Methods and Materials……………………………………………………...46 3.2.1 The Interpretation Task………………………………………………47 3.2.2 The Imitation Task..………………………………………………….50 3.3 Procedures…………………………………………………………………..53 3.3.1 Pilot Study…………...………………………………………………53 3.3.2 Formal Study………………………………………………………...54 3.3.3 Scoring and Statistical Analysis……………………………………..56 3.4 Summary of Chapter Three…………………………………………………58 Chapter Four Results and Discussion……………………………………………...59 4.1 Markedness Effects…………………………………………………………59 4.1.1 A Comparison of the Five Types of Chinese Conditionals……………59 4.1.2 A Comparison of the Unmarked vs. Marked Conditionals…………...62 4.1.3 General Discussion……………………………………………………64 4.2 Hypothetical vs. Counterfactual Conditionals………………………………69 4.2.1 Comparison of the Hypotheticals and Counterfactuals……………….69 4.2.2 A Second Look at the Hypothetical and Counterfactual Conditionals 71 4.2.3 General Discussion……………………………………………………73 4.3 Task Effects…………………………………………………………………76 4.3.1 The Results of the Interpretation vs. Imitation Tasks………………...76 4.3.2 General Discussion……………………………………………………79 4.4 Production Analysis………………...………………………………………82 4.4.1 Patterns Different from the Target Sentences in the Imitation Task….82 4.4.2 General Discussion……………………...…………………………….86 4.5 Age Effects…………………………………………………………………90 4. 6 Summary of Chapter Four………………………………………………….90 Chapter Five Conclusion…………………………………………………………..92 5.1 Summary of the Major Findings……………………………………………92 5.2 Limitations of the Present Study and Suggestions for Future Research…....94 Bibliography………………………………………………………………………....96 Appendix A: The Scenarios Used in the Interpretation Task…………………......102 vii .

(9) Appendix B: The Story Used in the Imitation Task……...…………………………115 Appendix C: Results of the Pilot Study……………………………………………129 Appendix D: Consent Form……...…………………………………………………130. viii .

(10) LIST OF TABLES Table 1-1 Schachter’s Hierarchy of Conditionals……………………………………..2 Table 2-1 Types of Chinese If-words in Wu (1994)………………………………….13 Table 2-2 Five Types of Chinese Conditionals in Wu’s Study……………………….15 Table 2-3 The Acquisition of Temporals and Conditionals in Reilly (1986)………...24 Table 2-4 Mean Scores of Correct Replies as a Function of Age and Question……..27 Table 2-5 Mean Scores of References per Child to the Rejected Option…………….28 Table 2-6 Mean Scores of References per Child to Alternatives…………………….29 Table 2-7 Most Commonly-occurring Conditional Types in Crutchley (2004)……...30 Table 2-8 Frequency and Percentage of Non-adult like Responses………………….31 Table 2-9 The Taxonomy of Chinese Conditionals in the Present Study…………….43 Table 3-1 A Summary of the Subjects………………………………………………..45 Table 3-2 Test Conditional Types for Both Tasks…………………………………….48 Table 3-3 An Example of the Test Scenario in the Interpretation Task…...………….50 Table 3-4 An Example of the Test Scenario in the Imitation Task……………..…….52 Table 3-5 The Coding System Used in the Present Study……………………………57 Table 4-1: Production Types in the Imitation Task…………………………………..84. ix .

(11) LIST OF FIGURES Figure 2-1 Percentages of Correct Answers in Experiment 1………………………..33 Figure 2-2 Percentages of Correct Answers in Experiment 2………………………..34 Figure 3-1 Characters Used in the Interpretation Task…………………..…………...49 Figure 3-2 Characters Used in the Imitation Task…………………..………………..52 Figure 4-1 A Comparison among the Five Types of Conditionals…………………...60 Figure 4-2 A Comparison of the Unmarked and Marked Conditionals……………...63 Figure 4-3 A Comparison of the Hypotheticals and Counterfactuals………………..69 Figure 4-4 A Revised Comparison between Hypothetical and Counterfactual……...72 Figure 4-5 Subjects’ Responses to the Two Tasks on the Unmarked vs. Marked Conditionals……………………………………………………………...77 Figure 4-6 Subjects’ Responses to the Two Tasks on the Two Scenarios…………...78 Figure 4-7 A Comparison of the Two Tasks…………………………………………79. x .

(12) LIST OF ABBREVIATIONS ADV CF CI CL COP IM IP IT L1 NEG NP PT PAS S V VP. adverb clause-initial conditional adverbial clause-final conditional adverbial classifier cooperative marker the imitation task inflection phrase the interpretation task the first language negator noun phrase particle passive marker subject verb verb phrase . xi .

(13) Chapter One Introduction. 1.1 Motivation and Purpose In logic, conditionals are considered as a relation between two propositions, the protasis (p), i.e., the antecedent clause and the apodosis (q) i.e., the consequent clause (Braine 1978). In the logical relation, the conditional situation is established if p and q are both true, or p is false and q is true, or p is false and q is false but the possibility of p being true while q is false is excluded (Braine and O’Brien 1991, Comrie 1986). However, the majority of conditionals lack a clear-cut boundary because the consequent clause may have illocutionary force, such as a request or a question. In other words, the belief triggered by the antecedent clause places insufficient constraints on the possible antecedent models (Jonson-Laid 1983 and1986). Furthermore, conditionals show a strong causal link between two clauses in which the protasis (the antecedent clause) is considered as cause and the apodosis (the consequent clause) as effect (Comrie 1986, Reilly 1986). The causal relation suggests that “the content of the protasis must be interpretable as a cause of the content of the apodosis (Comrie 1986:80).” Nevertheless, causal relations generally not only represent the literal meaning but also the speaker’s motivation to make the proposition, as shown in (1): (1) If you want to know, ten isn’t a prime number.. (Comrie 1986:81). There is no direct causal relation between the two propositions since the addressee’s asking does not cause ten not to be a prime number. It is the addressee’s desire of asking whether ten is a prime number (speech act) makes the propositions causal (Comrie 1986). 1 .

(14) In terms of pragmatics, some researchers state that children’s implicit conditionals may be segmented across turns (De Castro Campos 1981, Bowerman 1986). Such a serial reflects the sequence of question-confirmation-implicative assertion (Jespersen 1940) and children gradually learn conditionals because of internalized dialogues. With regard to the conditionals themselves, Grice (1967) state that conditionals are correspondent to material implications from which some implicatures are derived due to conversational conventions while Stalnaker (1968) note that they cannot be material implications. Then, philosophers have evaded the issue and claim conditionals are a matter of context (Comrie 1986). Conditional constructions typically reveal different degrees of hypotheticality of the truth of the propositions which rely on backshifting of tense (Comrie 1982 and 1986). In this case, greater hypotheticality utilizes nonfuture time reference to show lower probability while lower hypotheticality is with future time reference (Comrie 1986). Schachter (1971) classifies conditionals into two types in terms of relative hypotheticality: simple and imaginative conditionals, as illustrated in Table 1-1. Between these two types, the forms in imaginative conditionals which represent different likelihood of possibility in the real world are more complicated in tense. Table 1-1 Schachter’s Hierarchy of Conditionals 1. Simple conditionals. Examples. Present. If I touch my eye, it hurts.. Past. If the cat is in the kitchen, he’s eating the meat.. Generic. If the tortoise has a runny nose, he sleeps in the house.. Predictive. If Kate sees the ice cream, she will want some.. 2.Imaginative conditionals. Examples. Hypothetical. If he ate all those doughnuts, he would be ill.. Counterfactual. If I were a boy, I would have curls. If you had been awake, you would have heard the coyote. 2. .

(15) As mentioned above, conditionals represent a diverse degree of hypotheticality. In the literature, Athanasiadou and Dirven (1997) claim that imaginative conditionals (ICs) should be prototypical conditionals. Among imaginative conditionals (ICs), the different degree of markedness relies on the commitment to the realization of the situation. If the statement of an IC is potentially real, it is an unmarked case. On the contrary, it is a marked case if the situation is unreal (Athanasiadou and Dirven 1997). Since Eckman (1983) argues that the degree of difficulty in language accord with the degree of markedness, a counterfactual conditional will be more difficult for children to acquire and they tend to make some errors in producing the structure. In addition, many findings (Riggs et al. 1998, Riggs and Peterson 2000, Robinson and Beck 2000) have shown that 3-year-old children had little difficulty in reasoning hypothetically1 while they had problems with counterfactual conditionals. In the field of language acquisition, conditionals, given the nature of their cognitive and logical complexity between two clauses, have always been the focus and significant issues in language acquisition since they can reveal children’s mental integrity of propositions, as in (2) and (3): (2) If you go out without the umbrella, you’ll get wet. (Comrie 1986:78) (3) Ruguo tuo henjiou bu qilai, (Yeh 2000:370) if put off long while not get up ni hai keyi zai ba wo waqilai you still can again make I wake up ‘If I put it off for a long while and didn’t get up, you can try to wake me up again.’. Studies on English conditionals have demonstrated that children’s first typical conditional utterance is at the second half of their third year (Clancy, Jacobsen and                                                         1. . Their studies focused on future hypotheticals, in which events may occur in the future while the present study will focus on hypotheticals in which all events appear in the designed story may not occur in the real life. 3.

(16) Silva 1976, McCabe et al. 1983, Reilly 1982). However, the critical age of acquiring conditionals is at age 4 (Harris, German and Mills 1996, Reilly 1982& 1986). Children’s production ability may not be fully acquired at age 8 (Reilly 1986) and some production forms may be refined from the existing system even in the teenage years (Crutchley 2004, Reilly 1982 and 1986). In addition to the discussion about the age effect of conditional acquisition, many scholars have investigated linguistic issues in children’s late acquisition of conditionals. Some claim (Bloom et al. 1980, Reilly 1982) that cognitive complexity may influence children’s comprehension while others (De Castro Campos 1981, McCabe et al. 1983) argue that pragmatics, such as speech acts and discourse are significant for late emergence of conditionals. Since English conditionals consist of diverse inflection markings about speech and reference time, syntactic complexity is highly likely to influence children’s acquisition of conditionals (Rutherford 1982). Moreover, conditionals, as known, are not only complicated in syntax, but also pragmatically complex (Crutchley 2004). Granted that conditionals are backshifting of tense crosslinguistically, Mandarin, a serializing language, does not yield similar tendency (Li and Thompson 1981). English relies on inflections to target the speech time and reference time while Mandarin Chinese mostly employs various conditional adverbials (i.e., if-words) in different places of an antecedent clause to produce conditionals. According to Wang (1996) and Yeh (2000), Mandarin Chinese conditionals have more clause-initial conditional adverbials, which occupy almost 50% of the occurrences in the corpus. Furthermore, jiu ‘then’ is often employed in the consequent clause of Mandarin Chinese conditionals (Li and Thompson 1981, Wu 1994). Therefore, conditional adverbials in Chinese conditionals are assumed a decisive factor affecting children’s conditional acquisition in the present study. 4 .

(17) In addition, much research on Chinese conditionals presents the characteristics of conditional types in terms of semantics and syntax (Cheng and Huang 1996, Chierchia 2000, Wu 1994) and demonstrates that Mandarin Chinese conditionals rely on context to differentiate time frames (Li and Thompson 1981). Accordingly, Mandarin Chinese conditionals are widely discussed in the linguistic field while few empirical research on L1 acquisition of Mandarin Chinese conditionals has been carried out. The present study aims to probe into children’s acquisition of Chinese conditionals. Hence, preschoolers and elementary school students will be asked to complete a comprehension task and a production task to investigate their conditional development. In addition, conditional adverbials in Mandarin Chinese will be assumed a significant factor affecting children’s conditional development. Given the assumption that the more frequent forms in occurrences correspond to accessibility in language acquisition in the Markedness Theory (Eckman 1977 and 1986), the conditional type with clause-initial conditional adverbials are assumed the easiest for our subjects to acquire in the present study.. 1.2 Research Questions Motivated by the aforementioned discussion, the present study aims to address the following questions: 1). Are conditional adverbials significant factors affecting children’s acquisition of Chinese conditionals? If so, do various types of conditionals yield a varying degree of difficulty for Chinese children?. 2) Do hypothetical and counterfactual scenarios influence the difficulty of L1 acquisition of Chinese conditionals? 3) Do Chinese children perform equally well on different tasks in L1 acquisition of Mandarin Chinese conditionals? 5 .

(18) 4) What do the children’s utterances different from the target sentences in the production task imply in L1 acquisition of Mandarin Chinese conditionals ? 5) Is age a factor affecting the acquisition of Chinese conditionals?. 1.3 Significance of the Study Conditional constructions have drawing linguists’ interests in the past decades. In the literature, English-speaking children’s conditional reasoning has been animatedly investigated and analyzed with theoretical and empirical approaches (Athanasiadou and Dirven 1997, Beck et al. 2006, Comrie 1986, Riggs et al. 1998, Riggs and Peterson 2000, Robinson and Beck 2000). Even though the theoretical issue concerning Mandarin Chinese conditionals has been highly discussed for years, there has been paucity of acquisition studies on this construction. Thus, the research questions addressed in the present study will shed light on children’s developmental patterns in comprehension and production of Mandarin Chinese conditionals. With the discussion of differences in properties and scenarios, the present study will bridge the gap between theoretical and experimental arguments. Furthermore, it is hoped that the present study will attract more research on the acquisition of Mandarin Chinese conditionals.. 1.4 Organization of the Thesis The present study is organized as follows. Chapter Two reviews theoretical issues in conditionals, classification of the structure as well as empirical studies of L1 acquisition of conditionals. Chapter Three introduces the research design and Chapter Four reports the results and presents the analysis of conditional types of Mandarin Chinese. Finally, Chapter Five summarizes the major findings of the present study. 6 .

(19) Chapter Two Literature Review and Classification of Mandarin Chinese Conditionals. In this chapter, the arguments for the Markedness Theory, the previous theoretical studies of Mandarin Chinese conditionals, L1 acquisition of English conditionals and the classification of Mandarin Chinese conditionals will be discussed. Section 2.1 discusses the Markedness Theory and its related issue in language acquisition. Section 2.2 explores five theoretical studies on Mandarin Chinese conditionals. Section 2.3 reviews some empirical studies on L1 acquisition of conditionals. The classification of Mandarin Chinese conditionals is elucidated in Section 2.4. Finally, Section 2.5 is the summary of this chapter.. 2.1 Markedness Theory Markedness Theory (or the Theory of Markedness) is the study of how languages differ from each other and to map some linguistic phenomena into a specific language. It investigates universal grammar and also one particular linguistic phenomenon. The term “markedness” is that “a phenomenon A in some languages is more marked than B if the presence of A in a language implies the presence of B; but the presence of B does not imply the presence of A” (Eckman 1977:320). Based on the logical definition, the occurrence of a marked form can predict the occurrence of an unmarked form. Markedness Theory is highly discussed in syntax and phonology, however, it is understood as a map and device of language acquisition research (Williams1981). In language acquisition, an unmarked case, which is conceived as children’s initial comprehension or hypothesis of a language, emerges earlier than a marked one (Williams 1981). A marked case, however, acquires later in children’s speech. In 7 .

(20) addition, Eckman argues that the degree of difficulty in language is in accord with the degree of markedness. In other words, a marked case is more difficult for children to acquire and some errors are likely to be made when children produce marked forms. With regard to the presumption, Rutherford (1982) claims that markedness is equal to psycholinguistic complexity and sentential complexity. Taking an English conditional for instance (If you had seen my younger sister, you would have known that she was pregnant.), the lexical forms of a counterfactual conditional is marked since it consists of relatively complicated inflectional markings and it takes much more effort for children to anchor a time frame like the event time and the speech time. Other scholars also explore markedness of conditionals. Athanasiadou and Dirven (1997) examined the prototypicality of English conditionals. They suggested imaginative conditionals (ICs), including hypothetical and counterfactual conditionals, should be prototypical. Among imaginative conditionals (ICs), the different degree of markedness relies on the commitment to the realization of the situation. In other words, the decisive factor is whether readers or hearers disengage themselves from a real situation but consider potential realities. If the statement of an IC is potentially real, it is unmarked. On the contrary, it is a marked case if the situation is unreal (i.e., contrary to the present or past). Even though the concept of markedness is widely discussed, the criteria to discover an unmarked case deserves further investigation. According to Cairns (1983), two criteria, overt marking and frequency counts, are problematic in the Markedness Theory. However, his argument stems from the phonological and morphological studies of Markedness Theory. The researcher would like to investigate whether the observation needs to be reconsidered in the present study.. 8 .

(21) 2.2 Pervious Theoretical Studies of Mandarin Chinese Conditionals This section is concerned with the theoretical studies of Mandarin Chinese conditionals. In the literature, Li and Thompson (1981) discuss Mandarin Chinese conditionals in discourse, and following Li and Thompson (1981), Wu (1994), further examin Chinese conditionals from the syntactic and semantic perspectives. However, Cheng and Huang (1996) probe into some particular conditional structures of Mandarin Chinese: bare conditionals, ruguo conditionals and dou conditionals in syntax. Then, Chierchia (2000) argues against Cheng and Huang’s (1996) observation by proposing a semantic constraint. Finally, Su (2005) discusses the relationships of Chinese conditionals in terms of discourse. These studies will be reviewed in the following subsections.. 2.2.1 Li and Thompson (1981) According to Li and Thompson (1981), Mandarin Chinese conditionals are one type of forward linking sentences, which consist of two parts, an antecedent (if…) and a consequent (then clause) clause. The antecedent clause introduces the condition of a sentence first, and the consequent clause, which represents a true proposition, makes the meaning of the sentence complete. Li and Thompson argue that Chinese conditionals can be made in certain ways. First, they can be made with linking words (i.e., clause-initial conditional adverbials) in the antecedent clause, such as dehua 'if,' as can be seen in (1):. ( 1) Ta you qian dehua jiu bu hui xiang wo jie qian. he have money if then NEG will to me borrow money ‘If he has money, he won’t borrow money from me.’ Second, some linking words can be used in Chinese conditionals like ruguo, jiaru, 9 .

(22) jiashi and yaoshi ‘if,’ as in (2): ( 2 ) Yahoshi jintian fan jia jiu hao le. if today set free holiday then good PT ‘If we are on holiday today, that would be good.’ Third, no overt linking element, with jiu ‘then’ in the consequent clause, may be used in conditionals, as shown in (3): ( 3) Wo shuo keyi jiu keyi. I say OK then OK ‘If I say it’s OK, then it’s OK.’ Nevertheless, this type of usage shows that the proposition of the antecedent clause does not depend on the consequent clause, indicating that the antecedent clause can independently appear in other contexts. Thus, speakers may interpret a conditional sentence as a temporal sentence like ‘Once I say it’s OK, then it’s OK.’ Interrogatives, lastly, can be employed in no-linking conditionals such as shei ‘who’ and sheme ‘what,’ as shown below:. (4) Shei. zui. le. jiu. fa. shi. kuai. qian.. who drunk PT then fine ten dollar money ‘Whoever is drunk, h/she is fined by ten dollars.’ Furthermore, Li and Thompson examine meanings of conditionals from discourse and semantic perspectives. Three basic scenarios of Mandarin Chinese conditionals are classified: (1) Reality, (2) Hypothetical, and (3) Counterfactual conditionals1. Reality conditionals indicate the relation between two propositions in the real world, as in (5):.                                                         1. . Hypothetical and counterfactual conditionals belong to Imaginative Conditionals (Li and Thompson 1981). 10.

(23) (5) Jiaru taiyang chulai if the sun comes out women jiu keyi qu haibin. we then can go the beach ‘If the sun comes out, we can go to the beach.’. (Li and Thompson 1981:448). The consequent clause, as in (5), is an expectation that may be realized in the real world about the antecedent clause. Furthermore, it can represent a suggestion, a comment and an assertion. The second type, hypothetical conditionals, shows a possibly true situation in an imaginative world where speakers imagine a context (the antecedent clause) and make a comment, an assertion or even as a question based on the proposition (the result clause). Sentence (6) is a hypothetical conditional, as shown below:. (6) Jiaru. ni. gei. qiche zhuang dao. (Li and Thompson 1981:451). if you PAS car hit down women zenme ban? we how manage ‘If you were hit by a car, how would we manage?’ Finally, counterfactual conditionals, as the name displayed, represent a counterfactual situation in which the event of the sentence may be hypothetically true i.e., contrary to present or past. For example, the listener in (7) did not take the speaker’s advice, so h/she suffered. (7) Jiaru ni ting wode hua, (Li and Thompson 1981:449) if you listen my words ni jiu buhui shou ku le. you then NEG get sufferings PT ‘If you had listened to me, you wouldn’t have suffered.’ Though three kinds of scenarios of Mandarin Chinese conditionals are introduced, it 11 .

(24) is not necessary to differentiate the time frame when we hear a conditional. Li and Thompson suggest that context and speakers’ judgments can influence the interpretation of sentences. When the speaker utters a conditional, hearers may make a reality interpretation because h/she wants or plans to make the propositions possibly true. Probably, hearers may regard (7) as a hypothetical conditional because both the speaker and the hearer can imagine a situation that the hearer does not listen to the speaker’s words. In this case, a conditional sentence can be considered as a reality, hypothetical or counterfactual conditional. Therefore, context, especially the context that the speaker and the hearer co-construct and identify, is crucial when we assess Chinese conditionals. Overall, Li and Thompson (1981) discuss elements of Mandarin Chinese conditionals. Furthermore, they describe that forms themselves may not represent different interpretations of conditionals. Nevertheless, discourse and semantic factors are scrutinized in Chinese conditionals. Context, which is significant, provides different intents of message of the speaker and the hearer in speech. To sum up, Li and Thompson’s (1981) statement of Mandarin Chinese conditionals is descriptive and explicit. Nevertheless, they only describe the literal markings but do not mention the syntactic or semantic differences of the four types of conditionals they classify. To them, Chinese conditionals, hence, are considered of equal importance without hierarchical structures being compared.. 2.2.2 Wu (1994) Different from Li and Thompson emphasizing the importance of context, Wu (1994) examines Chinese conditionals from the semantic and syntactic perspectives. She argues that there are five explicit grammatical markers in Chinese conditionals. Among these markers, clause-initial conditional adverbials are the most typical in 12 .

(25) expressing conditionals such as ruguo, yaoshi, ruoshi, and jiashi ‘if’. In addition to listing these clause-initial conditional adverbials, Wu further classifies these words into three categories according to the likelihood of supposition, as shown in Table 2-1: Table 2-1 Types of Chinese If-words in Wu (1994) Category Chinese. Gloss in English. 1. ruguo yaoshi. ‘in case’ ‘if, in case’. 2. jiaru jiashi jiashe wanyi yaobushi. ‘if, supposing’ ‘if, supposing’ ‘given that’, ‘provided that’ ‘in case……should…….’ ‘were it not the case…’, ‘had it not been the case…’. 3. jishi jiushi jiusuan zhi(yao) chufei. ‘even if’ ‘even if’ ‘even if’ ‘only if’ ‘unless’. The clause-initial conditional adverbials of the first category are neutral in the degree of possibility. In other words, they show no other intended meanings. Those of the second category reveal the least likelihood of fulfillment and the negative meanings. The if-words of the third category contain “various meanings of ‘if’ and different degrees of likelihood (Wu 1994:148).” However, ruguo and jiaru, which are often used as typical and unmarked conditional adverbials in Chinese, belong to different categories in Wu’s argument. Such a categorization may encourage more sophisticated observation and the researcher supposes that these two words may belong to the same category. Wu also examines a clause-final conditional adverbial dehua ‘the supposition that, the case that,’ as previously discussed in Lü’s (1942) and Chao (1968). Although dehua can be omitted in Chinese conditionals, Wu argues that dehua is an 13 .

(26) independent conjunction, indicating an equal degree of importance compared to other Chinese if-words. In addition, Wu claims that jiu ‘then, would’ in the consequent clause of Chinese conditionals helps to indicate a hypothetical meaning. Jiu is not equivalent to English then or would since it cannot be deleted in Chinese conditionals but then or would can be dropped in English. Furthermore, jiu itself shows a thorough conditional meaning without the presence of any if-words; hence, it is not simply a particle. It occupies a significant position in Chinese conditionals. Following Lü (1942) and Chan (1968), Wu also examines negators in Mandarin Chinese conditionals. It is argued that negators display conditional or even counterfactual meanings without any if-words as in (8): (8) Ni bu lai wo bu qu. you NEG come I NEG go ‘If you don’t come, I don’t go.’. (Chao 1968:116). In addition to presenting detectable conditional adverbials of Mandarin Chinese conditionals, Wu reviews Lü and Li and Thompson’s findings that no conjunctions are included in Chinese conditionals. However, covert markings may suggest some linguistic clues in conditional messages. In covert marking of conditionals, there is a noticeable pause, a subordinate marking, between the antecedent and the consequent clause. Since pausing is a suprasegmental marking in sentences, Wu suggests that it should be treated as a Chinese grammatical element in conditionals. Interestingly, this kind of sentence can be considered as a temporal expression since the meaning of the sentence mainly depends on context. Based on the previous discussion, Wu classifies Mandarin Chinese conditionals into the following five types (Wu 1994:155). 14 .

(27) Table 2-2 Five Types of Chinese Conditionals in Wu’s Study Conditional clauses. Consequent clauses. 1. The typical conditional ruguo/ yaoshi ‘if’. jiu ‘then’. 2.Clause-final particle, with optional ‘if’ Ruguo/yaoshi 'if' ....... dehua. jiu ‘then’. 3. Adverbial conjunction in the consequent clause, without ‘if’ …….. jiu ‘then’. 4. No linking element, with a pause between the two clauses …….,. (pause)……... 5. Negative in one or both clauses Bu-shi/ mei-you ‘not; no’. Bu-shi/ mei-you ‘not; no’. To sum up, Wu reviews some previous studies and states that jiu ‘then’ serves as a syntactic status of Chinese conditionals. Moreover, she argues that negators and suprasegmental markings are used in Mandarin Chinese conditionals. Thus, she proposes more types of conditionals than the previous researchers.. 2.2.3 Cheng and Huang (1996) Cheng and Huang (1996) discuss Mandarin Chinese conditionals with indefinite wh-words that have the semantic meanings of donkey sentences. They argue that Chinese conditionals with donkey anaphora can be classified into two paradigms. One type is called bare conditionals and the other is further divided into two subtypes, ruguo- and dou- conditionals. Bare conditionals are conditionals without a clause-initial conditional adverbial such as ruguo ‘if’ in an antecedent clause or a quantifier like dou ‘all’ in a consequent clause, as shown in (9):. 15 .

(28) (9) Shei xian lai shei xian chi. who first come who first eat ‘If X comes first, X eats first.’. (Cheng and Huang 1996:127). In their opinion, the wh-words in the antecedent and the consequent clauses must be identical, as in (9). Conditionals will be ill-formed if the wh-word in the consequent clause is substituted by an overt pronoun like ta ‘he,’ covert pronoun [e] or a definite NP like nageren ‘that person’. Based on the Discourse Representation Theory (DRT), Cheng and Huang suggest that wh-words are indefinite NPs without inherent quantification force. Therefore, an external licenser, called a necessity operator with the force of universal quantification, licenses and binds two distinct wh-words simultaneously. This kind of quantification, therefore, is called ‘unselectively’ binding. In addition, wh-words, treated not as quantifiers but variables, are independent in nature and neither of them is syntactically anaphoric to another. As mentioned before, wh-words in Chinese bare conditionals must appear in pairs. Cheng and Huang, then, adopt Kratzer’s Prohibition Against Vacuous Quantification to explain the tempting occurrence of wh-words. (10) Prohibition Against Vacuous Quantification (Kratzer 1989:155) For every quantifier Q, there must be a variable x such that Q binds an occurrence of x in both its restrictive clause2 and its nuclear scope3. The argument indicates the necessity of two identical wh-words in bare conditionals. If we just use one wh-word in either the antecedent or consequent clauses, (10) will be violated. To avoid overt pronouns in the consequent clause, they state that the overt                                                        . 2 3. . A restrictive clause is an antecedent clause in conditionals. A nuclear scope is a consequent clause in conditionals. 16.

(29) pronouns can be considered as: (1) bound variables or (2) E-type pronouns4 (Evans 1980). If the overt pronoun in the consequent clause is a bound variable, the accessibility condition will be violated since the wh-word in the antecedent clause does not c-command it and none of its element is accessible. As for the status of E-type pronouns, it will also lead to vacuous quantification since E-type pronouns are not variables. Considering the constraint on empty pronouns and definite NPs in the consequent clause, Cheng and Huang apply the Parallelism Constraint on Operator Binding (PCOB), as in (11) to account for their unavailability. (11) The Parallelism Constraint on Operator Binding (PCOB) (Safir 1984:607) If O is an operator and x is a variable bound by O, then for any y, y a variable of O, x and y are [α lexical]. The [α lexical] stands for the identical occurrence of variables. In this case, the variables in bare conditionals must be either all lexical [+lexical]5 or all empty [-lexical]. The second type conditionals with donkey anaphora, ruguo- and dou-conditionals, nevertheless, display the complementary scenario compared with bare conditionals. Briefly speaking, if there is a wh-word in the antecedent clause, the identical wh-word cannot be used. Nonetheless, a pronoun, an empty category, or a definite NP are welcomed in the consequent clause. Under these circumstances, the grammatical properties are discussed for the complementary distribution between the two main conditional types, as shown in (12) and (13):.                                                         4. 5. . E-type pronouns, as Evans states (1980:340) refer to “the objects that verify the antecedent quantifier-containing clause,” and “the truth of the clause containing them requires that all the relevant objects satisfy the predicate…” In addition, E-type pronouns cannot take negative quantifiers like no one or nobody as their antecedents. ‘[+lexical]’ refers to the same word usage in either the antecedent or consequent clauses. 17.

(30) (12) Ruguo ni if you. kando shei see who. (Cheng and Huang 1996:142). qin jiao ta lai jian wo. please tell him/her come see me ‘If you see someone, please as him/her to come see me.’ (13) Ni jiao shei jinlai wo dou jian ta. (Cheng and Huang 1996:145) you ask who enter I all see him/her ‘Whoever you ask to come in, I’ll see him/her.’ To begin with, Cheng and Huang expound on the ungrammaticality of the wh-word in a consequent clause of ruguo and dou conditionals. In ruguo conditionals, ruguo ‘if’ licenses the wh-word in an antecedent clause. The wh-word, hence, is considered as an existential quantifier in the scope of the antecedent clause. It represents that a complete tripartite conditional structure is manifested in the antecedent clause through Quantifier Raising. On account of the property, there is no need for wh-word to occur in the consequent clause. If it does appear in the consequent clause, the binding and the licensing problems will result in the ungrammaticality of these conditionals. Dou-conditionals, however, seem to be interpreted in different ways. The antecedent clause of a dou conditional is claimed to be an embedded question, inherently having the force of existential quantification. Given the existential nature of the clause, the wh-word in it is treated as an existential quantifier just like that in a ruguo conditional. Similar to ruguo conditionals, the grammatical movement in the antecedent clause of dou conditionals definitely denies the availability of another wh-word in the consequent clause. Second, Cheng and Huang illustrate why anaphoric elements can appear in the consequent clause of ruguo and dou conditionals. They propose that the anaphoric forms can be regarded as E-type pronouns. In aforementioned discussion, one feature of E-type pronouns is that they cannot occur in a negative-quantification sentence. With the nature, the ungrammatical example shown in the following clearly verifies 18 .

(31) that the anaphoric form in the conditional is an E-type pronoun. (14 ) *Ruguo meiyo shei if not-have who. ma ni scold you. (Cheng and Huang 1996:144). ni jiu jiao ta lai jian wo. you then ask him/her come see me ‘*If no one scolds you, then you ask him/her to come see me.’ To sum up, Cheng and Huang only discuss the donkey conditionals, which have not been mentioned in the previous studies. Because of the relatively peculiar forms of these conditional types, they employ two constraints on the interpretation of these conditionals. The bare conditionals are accounted for by unselective binding while ruguo and dou conditionals are analyzed as a construction with E-type pronouns. Wh-words, in this case, show different status in these two types of conditionals. Cheng and Huang’s approach skillfully solved the unspecified status of wh-words. However, whether or not the occurrence of wh-words in conditionals may cause the acquisition difficulty for children is not addressed in their study.. 2.2.4 Chierchia (2000) Chierchia, following Cheng and Huang (1996), analyze two types of donkey Chinese conditionals as well. He points out problems with Cheng and Huang’s arguments and probes into the property of wh-words based on Dynamic Semantics. According to Cheng and Huang, wh-words, with identical forms in two clauses of bare conditionals, have to occur in pairs. The wh-words, according to Chierchia, are not independent at all. He assumes that a novelty condition has to be found when there is a wh-word in the antecedent clause. However, the wh-word in a consequent clause suggests a familiarity condition since it is semantically bound by the previous wh-word. The specificity of wh-words in bare conditionals makes the status 19 .

(32) unpredictive and context dependent. Following the view of Dynamic Semantics, Chierchia first examines the nature of Chinese wh-words. Examining the occurrence of wh-words in different types of Chinese conditionals, he claims that wh-words should be indefinite pronouns since they are in accordance with the following properties, as in (15): (15) Properties of indefinite pronouns (Chierchia 2000:33) a. Existential interpretation in episodic contexts. b. Can be targeted for disclosure by Q-adverbs (quantification adverbs). c. No c-command binding d. No (or only loose) anaphoric links among indefinite pronouns. Then, he argues that the licensing of wh-words in bare conditionals is the disclosure operator (the index i), which associates with the quantificational adverb (ADVi), as shown in (16): (16). IP CP. ADVi. (Chierchia 2000:36) IP. wo xian da sheii sheii xian jinlai who first enters I first hit who ‘If X enters first, I hit X first.’ In Dynamic Semantics, an adverb of quantification can be either [+ affective] or [affective]. Since a [+ affective] quantificational adverb has the force to license wh-words, based on the diagram, it must operate in both clauses. Thus, the second wh-word cannot be substituted by another anaphoric element. In addition, the disclosure operator can wipe out the existential force of wh-words in the sentence. Thus, the truth condition of (16) will be: For every x, if x enters first, I first hit x. Such an explanation implies that wh-words are variables and we may also infer that 20 .

(33) the unselective binding proposed by Cheng and Huang still works. Chierchia manipulates a different theoretical perspective to specify wh-words in Chinese conditionals. Basically, his claim mainly maintains the spirit of Cheng and Huang’s proposal of bare conditionals. However, it seems that Chierchia does not provide us with a thorough discussion about the violation of the novelty condition in the consequent clause of bare conditionals. Thus, a semantic account for the occurrence of the wh-word in the consequent clause will be needed if we want to find out the truth of the property.. 2.2.5 Su (2005) Su investigates Mandarin Chinese conditionals with a discourse analysis method and discusses the relations of conditionals in natural context and their possible interpretations. In her study, ruguo ‘if’ (28.4%) and ruguo…dehua ‘if’ (14.81%) occupy the largest percentage among various linguist forms of Chinese conditionals. Furthermore, Su presents three types of conditional relationships: content-based, epistemic and speech act conditionals. Content conditionals occupy the largest portion (60.49%), followed by epistemic conditionals (29.63%) and the speech act conditionals (9.88%). Aside from these three basic conditional relationships, there are still possible further chains (i.e., the chain type) in natural context, such as content-epistemic (41.98%), content-epistemic-speech act (6.17%), content-speech act (2.47%), epistemic-speech act (9.88%). Su addresses that discourse-pragmatic concern servers as a driving force of the speaker’s various conditional productions. She suggests the antecedent clause, which ties what will be said and what has come before, shows intersubjectivity because it demonstrates the speaker’s stance (positive, negative or neutral) and the involvement 21 .

(34) of the hearer. In addition, the consequent clause presents the speaker’s undesirability of the real world. In other words, she considers that the undesirability of the consequent clauses is decisive for the intersubjective nature in terms of hypothetical use associated with the antecedent clause. She argues that linguistic forms are simply cues to activate our knowledge about the conditional concept while utterance meanings are products emerging from the interaction between what the context gives to the speaker and what is evoked by the form. In addition, Su claims that the interpretation of conditional should be identifiable because the conditional constructions create coherent mental space (Fauconnier 1985) and serve communicative needs. Lastly, her statement shows that language “serves as a powerful means of prompting dynamic online constructions of meaning that go far beyond anything explicitly provided by the lexical and grammatical forms.” (Fauconnier 2003:251).. 2.2.6 Summary The studies on Chinese conditionals reviewed in the section can be grouped into three frameworks. First, Li and Thompson (1981) and Wu (1994) only discuss fundamental markings in conditionals. To them, some conditional adverbials appear in the front or back of an antecedent clause while covert markings may be employed in conditional utterance. Second, scholars like Cheng and Huang (1996) and Chierchia (2000) examine peculiar types of conditionals. Wh-words in bare conditionals cannot be replaced by other segments such as pronouns and definite NPs while those in ruguo conditionals can since they play different syntactic roles. Finally, Su (2005) presents three kinds of conditional relationships and claims that conditionals establish upon the interlocutor’s beliefs. In addition, Li and Thompson classify Chinese conditionals into three time 22 .

(35) frames in terms of discourse: reality, hypothetical and counterfactual. However, the boundary among these time frames has not been clearly addressed since context may convey more messages than words themselves, which may intensify but not necessarily employ in hypothetical or counterfactual reasoning.. 2.3 Pervious Empirical Studies on L1 Acquisition of Conditionals This section mainly reviews some studies on L1 acquisition of English conditionals since there are no empirical studies on Chinese conditionals to be explored. It is believed that English-speaking children’s development of conditional sentences still can shed light on the present study. Four empirical studies are discussed in the following subsections: Reilly (1986), Harris, German and Mills (1996), Crutchley (2004) and Perner, Sprung and Steinkogler (2004).. 2.3.1 Reilly (1986) In most studies, conditionals are highly related to some complex constructions such as temporal and causal adverbials. Reilly (1986) also probed into interactive factors. affecting. language. acquisition. of. conditionals,. focusing. on. the. interrelationship between temporals and conditionals. According to Reilly, temporals and conditionals are highly similar since they both describe simultaneous or sequential events in speech. In addition, these constructions can denote causal relation. However, temporal and conditionals are different in that temporals refer to certainty or speakers’ expectations toward the event while conditionals express speakers’ uncertainty and supposition of the speech. These intriguing characteristics of temporals and conditionals may be interrelated to children’s acquisition of temporals and conditionals. Thus, Reilly hypothesized that children might acquire these two constructions concurrently and use them before they 23 .

(36) master these constructions. In her experiment, Reilly collected both naturalistic and experimental data. The naturalistic data were taken from five children in naturalistic play situations or diary speech, ranging in age from 12 months to 3 years 6 months. The experimental data were collected from 28 children, ranging in age from 2½ to 9 years and they were divided into seven age groups, each of which consisted of four children in each group. The experiment consisted of five tasks: the first three were designed to test children’s acquisition of conditionals while the last two were employed to investigate temporals. Reilly argued that there were seven stages, with some overlaps, in acquiring temporal and conditionals, as shown in Table 2-3 (Reilly 1986:323): Table 2-3 The Acquisition of Temporals and Conditionals in Reilly (1986) Stage. Age. Functions. Forms. I. 2½. Sequence. Juxtaposed propositions (unmarked). II. 2-3. Predictive(future sequence). when. III. 2-3. Predictive/Simultaneous (proto-generic). when. IV. 2-3. Predictive/Simultaneous/ Present. when and/ or if (individual variation). Co-temporal past punctual Past fantasy. when. Predictive/ Present / generic. when and if. Co-temporal past punctual Past fantasy. when. Hypothetical production. if. V. 3. VI. 4. Hypothetical (comprehension). if. VII. 5-8. Predictive/ Generic. when and if. Past punctual Past stative or habitual (reality). when. Hypothetical (production and comprehension). if. 24 .

(37) The children were forward to acquire unmarked forms (juxtaposition) in temporals. Then, when was used to express predictives and also simultaneous events. At Stage IV (aged from 2;6-3;2 in her study), if was used in conditionals. However, there were individual variations among children. Some children used when and if in protogeneric contexts and predictives while others preferred to use them in specific situations. In addition, when was found to have other functions. First, it was used to describe events in the past, where one clause referred to a punctual event. Second, when was used to introduce fantasy when the children had no world knowledge in comprehending a past stative sentence. At Stage V, the hypothetical use of if appeared. However, Reilly discovered that children had difficulty in responding to what if6 questions in the experiment. She argued that the children might be incapable of understanding the complicated scenario or misinterpret adults’ intended speech act. For the subjects, it was easier to produce or manipulate the construction than to clarify the specific linguistic components in speech. At Stage VI, what if questions were acquired. Counterfactuals were also acquired at the stage (age four). During the last period, when was for past statives and past habituals. The children fully understood how to use when and if in appropriate situations. To sum up, the children first acquired when referring to the real world first, gradually moved to the hypothetical and counterfactual uses of if and past habituals for when. In Reilly’s study, the chaotic situation at Stage V reflected that children readjusted their linguistic knowledge in acquiring language. The children actively used their temporal knowledge to understand and produce conditionals. In addition, Reilly demonstrated that the children started to use conditionals around 2 ½ years old,                                                         6. The sample question is as follows: Adult: What if you eat three ice cream cones? Katie: You don’t have three hands. 25. .

(38) with a limitation on the hypothetically immediate context and they fully acquired the conditional construction at age 8. As for the acquisition of counterfactual conditionals, 4-year-olds succeeded in the comprehension task and they could produce counterfactual responses in certain situations. Reilly’s findings corresponded to the previous results that age 4 is a crucial stage for acquiring conditionals. Although Reilly’s findings shed light on the acquisition of English conditionals, her study still demonstrated some limitations. First, her study consisted of a small subject pool, which may influence the analysis of children’s general language development. Therefore, a quantitative study is welcomed for future research. Second, her experiment did not include a control situation. Thus, it may be difficult to infer her findings to other conditional studies.. 2.3.2 Harris, German and Mills (1996) Harris, German and Mills (1996) investigated children’s counterfactual thinking in causal scenarios. In their study, children aged 3 to 5 years participated in three experiments which highly emphasized the relation between causal and counterfactual thinking. In the first experiment, 26 children, divided into two groups, were observed in a play with a causal situation and were told to answer two control questions7 and one test question8, which was contrary to the fact, after watching the play. Table 2-4 shows the correct responses of Experiment 1 (Harris, German and Mills 1996:239):.                                                        . 7 8. . The two control questions were Now and Before questions: (1) Now: Is the floor dirty now? (2) Before: Was the floor dirty before? The test question was: What if Carol had taken her shoes off – would the floor be dirty? 26.

(39) Table 2-4 Mean Scores of Correct Replies as a Function of Age and Question Question Group. Control Now. Before. Test. Mean SD. Mean. SD. Mean. SD. Younger. 4. 0. 3.69. 0.6. 3. 1.1. Older. 4. 0. 4. 0. 3.46. 0.7. The result showed that children almost responded correctly to the control questions. In the test questions (what-if questions), they made few errors or even no errors, indicating that the children, even though they were three years old, could demonstrate counterfactual thinking. Furthermore, the accurate responses also manifested that children’s causal thinking preceded their responses and helped them respond to counterfactual situations. In Experiment 2, 3-year-old children’s counterfactual thinking were further explored. Similarly, they were divided into two groups and were given control and test questions to answer. However, they were asked two versions of counterfactual scenarios, in one of which A would not have caused B, while in the other of which A would also have caused B. After they heard these situations, two test questions (preserve and change questions9) were designed in the experiment to evaluate how they differentiated these situations. The results showed that they successfully differentiated the two scenarios, and they performed better on before than now questions. In addition, more correct responses were found to preserve than change questions. Even though they showed preferences for certain questions, they still performed equally well on the two scenarios, indicating that they employed the causal scenario to anticipate contrary-to-fact situations. In other words, causal conditions helped them think                                                         9. . A sample preserve question is like this: If Teddy hadn’t painted the floor with his brush, would the floor be clean now? Change question is as follows: If Teddy had painted the floor with his fingers instead , would the floor be clean now? 27.