以中文為母語的兒童頻率副詞詮釋之實證研究

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. 以中文為母語的兒童頻率副詞詮釋之實證研究 An Empirical Study of Children’s Interpretations of Frequency Adverbs in Mandarin. 指導教授：陳純音博士 Advisor: Dr. Chun-yin Doris Chen 研究生：鄭羽辰 Student: Yu-chen Eliza Cheng. 中華民國一百零六年六月 June, 2017.

(2) 摘要本研究旨在探討以中文為母語之兒童頻率副詞詮釋的發展。本研究主要包含兩階段理解測驗：第一階段為辨認與排序測驗，可得知兒童對於頻率副詞的理解差異；第二階段為接受度測驗，可觀察兒童對於頻率副詞修飾動詞事件種類的接受程度是否有別，語境效應亦融入此階段測驗中。兩階段測驗皆探討年齡效應。研究對象包含實驗組的八十位兒童，依其就讀年級分為四組：幼兒園大班、小學二年級、四年級、六年級，每組各二十位，以及對照組的成人二十位。研究結果顯示，頻率副詞之習得隨著年齡增長，理解能力逐漸與成人趨向一致。典型頻率副詞的語意較非典型頻率副詞模糊，因此兒童較難理解典型頻率副詞，六年級兒童亦尚未具備成人母語者理解典型頻率副詞的能力，但四年級兒童已可正確理解非典型頻率副詞。關於頻率副詞用於修飾的動詞事件方面，除了大班兒童外，其他組兒童皆對於頻率副詞修飾有界動詞事件的接受度遠高於頻率副詞修飾無界動詞。而且，四年級、六年級兒童及成人組對於頻率副詞修飾有界動詞事件的次類型有相似的接受度。由於四年級兒童已具備理解非典型頻率副詞意思的能力，亦能依頻率副詞能修飾的事件區分出有界與無界動詞事件，對於有界動詞的各項次分類，亦皆能接受其被頻率副詞修飾，因此，四年級是習得中文頻率副詞的關鍵期。最後，語境效應在本研究呈現負面效果，受試者在提供語境的試題中，對句子的接受度普遍降低。. 關鍵詞：頻率副詞、第一語言習得、理解、中文. i.

(3) ABSTRACT The present empirical study aims at investigating children’s interpretations of Chinese frequency adverbs. Two phases were designed to elicit the child subjects’ interpretations. In Phase 1, the children’s interpretations of frequency adverbs were obtained from the identification and the ordering tasks. Similarly, in Phase 2, the children’s acceptability of various kinds of event types was investigated by the acceptability task. A comparison between a context-free task and a context-provided task was also explored in this phase. Age differences were examined in the both phases. A total of eighty children studying in an elementary school in New Taipei City were recruited and further divided into four experimental groups according to their grades (i.e., the kindergartners, Grades 2, 4 and 6). Moreover, twenty adults were recruited as the control group. The results indicated an age-related pattern in the acquisition of Chinese frequency adverbs. It was found that Grade 4 was able to interpret atypical frequency adverbs. However, even Grade 6 had not yet acquired adult-like interpretations of typical frequency adverbs. Therefore, typical frequency adverbs were more challenging for the subjects to interpret than atypical ones because typical frequency adverbs were vaguer in meaning than atypical ones. As for the events modified by frequency adverbs, it was found that telic events were more acceptable than atelic ones. All the groups, except for the kindergartners, accepted frequency adverbs when they were used to modify telic ii.

(4) events in a higher degree than those used to modify atelic events. Moreover, the subtypes of telic events displayed a similar degree of acceptability when they were modified by frequency adverbs for the older children (Grades 4 and 6) and the control group but not for the younger children (the kindergartners and Grade 2). Because Grade 4 was able to interpret atypical frequency adverbs, differentiate telic from atelic events and accept frequency adverbs when they were used to modify the subtypes of telic events in a similar degree, it was concluded that Grade 4 was a critical period for the acquisition of Chinese frequency adverbs. Finally, it was found that contextual clues showed a negative effect on the interpretations of frequency adverbs. The subjects generally gave lower acceptability ratings to the context-provided task than the contextfee task.. Keywords: frequency adverbs, first language acquisition, interpretations, Mandarin Chinese. iii.

(5) ACKNOWLEDGEMENTS I was fortunate to receive valuable assistance from many people during the period of doing the research. Each of their assistance was indispensable to the completion of this thesis. It is my pleasure to express my gratitude for their kind support. First and foremost, I would like to express my deepest appreciation to my thesis advisor, Dr. Chun-yin Doris Chen, for her professional and patient guidance. Whenever we met to discuss about the research, she always provided me with insightful comments. Similarly, when I encountered difficulties, she always offered me constructive suggestions and guided me to overcome the difficulties. Without her great support and considerable encouragement, the thesis would not have been completed. I would also like to show my genuine appreciation to my committee members, Dr. Jyun-Gwang Fred Chen and Dr. Rueih-Lirng Sharon Fahn. Their insightful remarks and valuable suggestions enlightened me on the improvement of my research and helped me enhance the quality of my thesis. I deeply appreciate their passionate participation and thoughtful comments. Sincere gratitude is also given to teachers, Ms. Ling-Xin Isabel Chang and Ms. Yu-Mei May Zou, who kindly helped me recruit subjects in their schools. The subject recruitment would not have been successful without their suitable and careful arrangement.. iv.

(6) Additionally, I would like to give special thanks to the teachers who taught me in the linguistic program: Dr. Gerardo Fernández-Salgueiro, Dr. His-Yao Su, Dr. HsiaoHung Iris Wu, Dr. Hui-Shan Nissa Lin, Dr. Jen-I Li, Dr. Jen Ting, Dr. Li-Hsin Ning, Dr. Miao-Hsia Tammy Chang, and Dr. Miao-Ling Hsieh, by alphabetical order. Their professional instruction introduced me to the world of linguistics and their generous support encouraged me to complete my thesis. I am also grateful to my dear classmates and partners in the graduate program: Amy Chen, Andrew Syue, Eileen Lin, Francis Chen, Howard Su, Johnny Hu, Kabby Chen, Lily Hsiao, Louisa Zheng, Mark Tu, Nick Hu, Shawn Lin, Stephanie Ye and Vivien Cheng, listed in alphabetical order. I was fortunate to have their warm company. Special thanks go to Amy Chen, Andrew Syue, Eileen Lin and Louisa Zheng, who promised without hesitation to help me conduct the tasks of my experiment even when they had much work to do. Last but not least, I would like to thank my beloved family for their constant support and considerable encouragement. They always supported every decision I made and encouraged me to fulfill my dreams. Their absolute faith in me boosted my confidence during my thesis writing and their unconditional love supported me whenever I met difficulties. Thus, I would like to dedicate this thesis to my dearest family. v.

(7) TABLE OF CONTENTS CHINESE ABSTRACT .................................................................................................. i ENGLISH ABSTRACT ................................................................................................. ii ACKNOWLEDGEMENTS .......................................................................................... iv TABLE OF CONTENTS .............................................................................................. vi LIST OF TABLES ......................................................................................................viii LIST OF FIGURES ....................................................................................................... x Chapter One Introduction .............................................................................................. 1 1.1 Motivation ........................................................................................................ 1 1.2 Theoretical Background ................................................................................... 2 1.3 Research Questions .......................................................................................... 6 1.4 Significance of the Study ................................................................................. 7 1.5 Organization of the Thesis ............................................................................... 7 Chapter Two Literature Review ..................................................................................... 9 2.1 Theoretical Studies of Frequency Adverbs ...................................................... 9 2.1.1 Wu and Ding (2005)............................................................................ 10 2.1.2 Zou (2006) .......................................................................................... 15 2.1.3 Piao (2011) .......................................................................................... 21 2.1.4 Summary ............................................................................................. 25 2.2 Empirical Studies of Vague Linguistic Terms ................................................ 26 2.2.1 Kuczaj (1975)...................................................................................... 27 2.2.2 Beyth-Marom (1982) .......................................................................... 30 2.2.3 Hoffner, Cantor and Badzinski (1990) ................................................ 33 2.2.4 Huang and Huang (2011) .................................................................... 37 2.2.5 Summary ............................................................................................. 40 2.3 Summary of Chapter Two .............................................................................. 43 Chapter Three Research Design................................................................................... 45 3.1 Subjects .......................................................................................................... 45 3.2 Methods and Materials................................................................................... 46 3.2.1 Types of Frequency Adverbs............................................................... 46 3.2.2 Types of Events ................................................................................... 47 3.2.3 Four Tasks ........................................................................................... 52 3.2.3.1 Phase 1 ……………………………………………………….53 3.2.3.1 Phase 2 ……………………………………………………….56 vi.

(8) 3.3 Procedures ...................................................................................................... 61 3.3.1 Formal Study....................................................................................... 61 3.3.2 Data Analysis ...................................................................................... 62 3.4 Summary of Chapter Three ............................................................................ 63 Chapter Four Results and Discussion .......................................................................... 64 4.1 Typical and Atypical Frequency Adverbs ...................................................... 64 4.1.1 Overall Findings.................................................................................. 65 4.1.2 A Second Look at Typical Frequency Adverbs ................................... 66 4.1.3 General Discussion ............................................................................. 70 4.2 Telic and Atelic Events .................................................................................. 73 4.2.1 Overall Findings.................................................................................. 73 4.2.2 General Discussion ............................................................................. 78 4.3 Three Subtypes of Telic Events...................................................................... 79 4.3.1 Overall Findings.................................................................................. 80 4.3.2 General Discussion ............................................................................. 85 4.4 Contextual Clues ............................................................................................ 86 4.4.1 Overall Findings.................................................................................. 87 4.4.2 General Discussion ............................................................................. 91 4.5 Age Differences.............................................................................................. 93 4.6 Summary of Chapter Four ............................................................................. 97 Chapter Five Conclusion ............................................................................................. 98 5.1 Summary of the Major Findings .................................................................... 98 5.2 Limitations of the Present Study and Suggestions for Future Research ...... 100 Bibliography .............................................................................................................. 101 Appendix A: The Test Items Used in Phase 1 ............................................................ 106 Appendix B: The Worksheet Used in Phase 1 ........................................................... 114 Appendix C: The Test Items Used in the Context-free Task in Phase 2 .................... 115 Appendix D: The Test Items Used in the Context-provided Task in Phase 2 ............ 118 Appendix E: The Worksheet Used in Phase 2 ........................................................... 128 Appendix F: Consent Form ........................................................................................ 129. vii.

(9) LIST OF TABLES Table 1-1 Vendler’s Classification of English Verbs ........................................................ 3 Table 2-1 The Five Levels and Their Examples of Frequency Adverbs ........................ 11 Table 2-2 The Youngest Age of Children Producing Each Frequency Adverb .............. 38 Table 2-3 Major Findings and Limitations of These Previous Empirical Studies ......... 41 Table 3-1 A Revised Classification of Frequency Adverbs ............................................ 47 Table 3-2 Five Types of Events ...................................................................................... 51 Table 3-3 The Distribution of the Test Items Used in the Two Tasks in Phase 1 ........... 54 Table 3-4 An Example of the Identification Task in Phase 1 ......................................... 55 Table 3-5 Examples of the Ordering Task in Phase 1 .................................................... 56 Table 3-6 The Three-point Scale in Phase 2 ................................................................... 57 Table 3-7 The Distribution of the Questions in Phase 2................................................. 58 Table 3-8 An Example Question in the Context-free Task ............................................. 59 Table 3-9 An Example Question in the Context-provided Task ..................................... 60 Table 4-1 Subjects’ Overall Performance on the Interpretations of Typical and Atypical Frequency adverbs .......................................................................... 65 Table 4-2 A Second Revised Classification of Frequency Adverbs ............................... 67 Table 4-3 Revised Subjects’ Overall Performance on the Interpretations of Typical and Atypical Frequency adverbs .................................................................... 68 Table 4-4 The p-values for the within-group Differences between the Interpretations of Typical and Atypical Frequency Adverbs .................................................. 69 Table 4-5 The p-values for the between-group Differences in Typical and Atypical Frequency Adverbs ........................................................................................ 69 Table 4-6 Subjects’ Overall Acceptability of Frequency Adverbs Modifying Telic and Atelic Events ........................................................................................... 74 Table 4-7 The p-values for the within-group Differences between the Acceptability of Frequency Adverbs Modifying Telic and Atelic Events ............................ 75 Table 4-8 The p-values for the between-group Differences in Acceptability of Frequency Adverbs Modifying Telic and Atelic Events ................................ 77 Table 4-9 Subjects’ Overall Acceptability Ratings for Frequency Adverbs Modifying the Three Subtypes of Telic Events................................................................ 80 Table 4-10 The p-values for the within-group Differences in the Acceptability of Frequency Adverbs Modifying the Three Subtypes of Telic Events ........... 82 Table 4-11 The Interaction Effects in the Acceptability of Frequency Adverbs Modifying the Three Subtypes of Telic Events ........................................... 83 Table 4-12 The p-values for the between-group Differences in the Acceptability of Frequency Adverbs Modifying the Three Subtypes of Telic Events ........... 84 viii.

(10) Table 4-13 Subjects’ Overall Acceptability of Frequency Adverbs Modifying Events in the Context-free and Context-provided Tasks ......................................... 87 Table 4-14 The p-values for the within-group Differences between the Context-free and the Context-provided Tasks .................................................................. 89 Table 4-15 The p-values for the between-group Differences in the Acceptability of Frequency Adverbs in the Context-free and Context-provided Tasks......... 90 Table 4-16 Overall Acceptability Ratings for Frequency Adverbs Modifying ET2-1 in the Context-free and Context-provided Tasks ......................................... 91 Table 4-17 The Children’s Acquisition of Chinese Frequency Adverbs ........................ 94. ix.

(11) LIST OF FIGURES Figure 2-1 Wu and Ding’s (2005) Analysis of Relation Among Frequency Adverbs and Their Negative Forms ........................................................................... 14 Figure 2-2 The Sub-types of Frequency Adverbs in Zou (2006) ................................... 17 Figure 4-1 Overall Performance of Each Group on the Interpretations of Typical and Atypical Frequency Adverbs ....................................................................... 65 Figure 4-2 A Revised Comparison between Typical and Atypical Frequency Adverbs 68 Figure 4-3 Each Group’s Acceptability of Frequency Adverbs Modifying Telic and Atelic Events ............................................................................................... 74 Figure 4-4 Each Group’s Acceptability Ratings for Frequency Adverbs Modifying the Three Subtypes of Telic Events ............................................................. 81 Figure 4-5 Each Group’s Acceptability Ratings for Frequency Adverbs in the Context-free and Context-provided Tasks ................................................... 88. x.

(12) Chapter One Introduction 1.1 Motivation Frequency adverbs are indispensable for communication in Mandarin (Li 2010). By using frequency adverbs, people can express how often an event happened, happens, or will happen (Biber et al. 2000) so that hearers or readers know the rate of occurrence of that event within a particular period of time. However, the difficulty in interpreting or using frequency adverbs by children has been proved in the literature (e.g., Kuczaj 1975, Huang and Huang 2011). For instance, Kuczaj (1975) found that among his forty subjects who were preschoolers, only nine of them were able to recognize that always is a strong positive and that never is a strong negative 1. Few of them believed that always and never are both positive or both negative. The restriction on frequency adverbs modifying events may also be a problem for children because not all verbs can be modified by frequency adverbs. Zou (2007) argues that frequency adverbs have the semantic feature [+diffuseness], so the events that can be modified by frequency adverbs need to be bounded. The contrast between the events that can be modified by frequency adverbs and those cannot is shown in (1) and (2):. Kuczaj (1975) views the ordered set of frequency adverbs as a positive-negative dimension, with always denoting a strong positive likelihood and never denoting a strong negative likelihood. 1 1.

(13) (1) Ta changchang da ren. he usually hit people ‘He usually hits people.’ (2) ?Ta. changchang renshi. Wang. he usually know Wang ‘?He usually knows Mr. Wang.’. (Huang and Huang 2011:78) xiansheng. Mr.. It is acceptable to modify da ren ‘to hit people’ with changchang ‘usually’ in (1); however, it is unacceptable to modify renshi Wang xiansheng ‘to know Mr. Wang’ with changchang ‘usually2’ in (2). Therefore, the present study aims at exploring children’s interpretations of frequency adverbs and examining if children are aware of the acceptable combinations of frequency adverbs and verb events.. 1.2 Theoretical Background The present empirical study is an attempt to examine children’s understanding of frequency adverbs, which has been classified into typical and atypical according to their semantic features and whether they can be placed in the focus position (Zou 2006). In the classification, typical frequency adverbs include shichan ‘often,’ changchang ‘usually,’ youshi ‘sometimes,’ ouer ’occasionally,’ and so on. In contrast, atypical frequency adverbs are zong(shi) ‘always,’ jielian ‘continuously,’ buduan ‘unceasingly,’ shishi ‘constantly,’ and so on. The present study will investigate whether different types. 2. The different semantic features between the two events will be explained in the next section. 2.

(14) of frequency adverbs are a factor influencing children’s interpretation. Moreover, verb choice is an issue that can be explored. As demonstrated in (1) and (2), some verbs can be modified by frequency adverbs, whereas the others cannot. The distinction lies in the concept of “telicity.” The term “telic,” first introduced by Garey (1957:106), refers to an event with “a natural final endpoint or intrinsic bound” (Smith 1997:19), while “atelic” denotes an event without a natural endpoint. Whether an event is telic or atelic depends on its verb type. Vendler (1957, 1967) classifies English verbs into four types—state, activity, accomplishment and achievement. The features and examples of the four types are shown in Table 1-1.. Table 1-1 Vendler’s Classification of English Verbs Verb types. Features. Selected examples. State. static, durative, atelic. know, believe, love. Activity. dynamic, durative, atelic. run, push a cart. Accomplishment. dynamic, durative, telic. build a house, draw a circle. Achievement. dynamic, instantaneous, telic. die, win, reach (compiled by the author). As shown in Table 1-1, the first two types are atelic, whereas the other two types belong 3.

(15) to telic. Verbs in Mandarin can also be classified into these four types (Lin 2004) 3. In (1), da ren ‘to hit people’ is telic so it can be modified by a frequency adverb; in (2), renshi Wang xiansheng ‘to know Mr. Wang’ is atelic, so it cannot co-occur with a frequency adverb. Whether all telic verbs display the acceptability of combining with frequency adverbs in the same degree remains an unsolved issue. For instance, both the events in (3) and (4) are telic.. (3) Ta. changchang. da. po. he usually hit broken ‘He usually breaks a cup into pieces.’ (4) Ta he jiu then. changchang xiang usually think jiejue. solve. beizi. cup. chu banfu, women out solution we. de wenti DE problem. ‘He usually thinks out a solution, and our problems are solved.’. The difference between (3) and (4) lies in the construction. The construction of (3) is a resultative verb compound but (4) is a verb followed by a particle. Whether different constructions of telic events remain identically acceptable to speakers when a frequency adverb is added in has not yet been explored in literature.. Some researchers (e.g., Tai 1984) contend that there are only three verb types in Mandarin. Because the focus of the present study is the feature of telicity rather than the number of verb types in Mandarin, the debate on the number of verb types in Mandarin is not discussed. 4 3.

(16) It is worth noting that some telic events, which refer to events that people normally do not or cannot do it repeatedly, cannot be modified by frequency adverbs when the subject of a sentence denotes human beings. A typical example is si ‘to die.’ Although it is regarded as an achievement verb (Tai 1984), people normally do not die for more than one time in the real world; therefore, it would sound awkward if it is modified by frequency adverbs, which describe the recurrence of events. In addition to the internal meanings and external combinations of frequency adverbs with verb events, the present study will investigate whether the factor of contextual clues would help the ease of interpretation regarding frequency adverbs. Many acquisition studies examine context effect (e.g., Schatz and Baldwin 1986, Stanovich 1986, Cacciari and Levorato 1989, Qian 1996, Walqui 2000, Friedrich and Friederici 2004, Belke et al. 2005, Hsieh 2008, Hoff 2010, among others). Take the study of Hsieh (2008) on children’s acquisition of metaphorical expressions as an example. Her study compared test items in isolation and in context, and she found that the subjects’ performance was enhanced in virtual of contextual clues. Similarly, the present study will also perform tasks to compare the subjects’ performance on test items with and without contextual clues. Another popular concern in first language acquisition is age effect. Language learning is a process starting from childhood through adolescents to adulthood (Nippold 5.

(17) 2004). When can children successfully acquire the meaning and usage of a word is a fascinating issue. The present study will also address this issue by exploring at what age children understand and use frequency adverbs as adults do. To sum up, the present study will examine children’s acquisition of frequency adverbs and their acceptability of telic/atelic events modified by frequency adverbs. The context effect and age effect will be discussed as well.. 1.3 Research Questions To further understand children’s acquisition order of a set of common frequency adverbs in Mandarin and to closely examine children’s acceptability of frequency adverbs modifying verb events, the present study will address the following research questions: 1) Do typical and atypical frequency adverbs in Mandarin display the same degree of difficulty and contribute to a similar order of acquisition? 2) Can Chinese children distinguish events that can be modified by frequency adverbs from those that cannot be modified by frequency adverbs? 3) Do telic events in Mandarin modified by frequency adverbs display a similar degree of acceptability to speakers? 4) Do contextual clues help Chinese children interpret events with frequency adverbs? 6.

(18) 5) At what age do Chinese children interpret frequency adverbs as adults do?. 1.4 Significance of the Study Children’s acquisition of a set of related words with various degrees has been investigated in literature (e.g., Cohen and Hansel 1956, Townsend 1974, Kuczaj 1975, Hoffner et al. 1990, Bocklisch et al. 2012, among others); however, most of these studies focus on the English language, and there is only one study probing into children’s production of frequency adverbs in Mandarin (Huang and Huang 2011), in which the issue that whether children correctly used the frequency adverbs in sentences has not been discussed. In order to fill the research gap in the language of Mandarin, the present study is an attempt to explore Chinese-speaking children’s interpretations of Chinese frequency adverbs. Furthermore, the issue of the acceptability of Chinese frequency adverbs in sentences has not yet been explored in acquisition studies, so the present study also aims at examining children’s interpretations of semantically acceptable combinations of frequency adverbs and verb events.. 1.5 Organization of the Thesis This thesis is organized as follows: Chapter Two reviews some theoretical studies of Chinese frequency adverbs and empirical studies of the acquisition of related words 7.

(19) with various degrees, such as frequency and probability. Chapter Three introduces the research design of the present study. Chapter Four reports the results and the analysis of the children’s acquisition of Chinese frequency adverbs. Finally, the major findings of the present study and some suggestions for future research are provided in Chapter Five.. 8.

(20) Chapter Two Literature Review In this chapter, theoretical studies of frequency adverbs and empirical studies of vague linguistic terms are explored. Section 2.1 reviews theoretical studies on Chinese frequency adverbs. Except for the frequency adverbs which lie in the two extremes (i.e., congbu ‘never’ and zongshi ‘always’), the meanings of frequency adverbs are vague. That is the reason why most frequency adverbs are categorized as linguistic terms with vague meanings. However, few empirical studies investigated interpretations of vague linguistic terms in Chinese. Therefore, Section 2.2 reviews empirical studies of vague linguistic terms in several languages. Finally, Section 2.3 provides a brief summary of this chapter.. 2.1 Theoretical Studies of Frequency Adverbs This section recapitulates three theoretical studies of Chinese frequency adverbs. Wu and Ding (2005) classify frequency adverbs into five levels according to the frequency of a frequency adverb; Zou (2006) classifies frequency adverbs into typical and atypical by taking their functions into consideration; Piao (2011) identifies five semantic features of frequency adverbs.. 9.

(21) 2.1.1 Wu and Ding (2005) Frequency, according to Contemporary Chinese Dictionary (Xiandai Hanyu Cidian), is the number of times a particular event happens in a period of time. In modern Mandarin Chinese, frequency can be expressed by short terms (e.g., fenfen miaomiao ‘every minute and every second’), adjectives (e.g., qin ‘often’) and adverbs (e.g., zongshi ‘always,’ changchang ‘usually’ and youshi ‘sometimes’). Among the three types, Wu and Ding (2005) find that the number of short terms and adjectives denoting frequency is few and adverbs are the main type of frequency expressions. Frequency adverbs can be classified into several levels according to the frequency that a frequency adverb denotes. Previous literature, such as Hu et al. (1989), Shi (2002) and Bocklisch et al. (2012), group English frequency adverbs into three levels or categories (i.e., high, medium and low) based on semantic meanings of frequency adverbs. In the field of statistics, there are five levels of frequency. The highest frequency indicates that a particular event happens every time in a period of time; the lowest frequency implies that a particular event never happens. The semantic scope of the other three levels is vague. All the above classifications are based on semantic features, but Wu and Ding (2005) argue that these classifications are too subjective. Wu and Ding (2005) not only take semantic features into account but also examine forms of frequency adverbs. They devise a formula to distinguish frequency adverbs. For instance, X is a frequency adverb, youshi ‘sometimes’; S is a subject (e.g., ta ‘he’); 10.

(22) V1 is a verb phrase (e.g., zai jia ‘at home’); V2 is another verb phrase (e.g., zai xuexiao ‘at school’). Then, ta youshi zai jia, youshi zai xuexiao ‘He is sometimes at home and is sometimes at school’ can be expressed as SXV1XV2. Similarly, ta jingchang zai jia, ouer zai xuexiao ‘He is often at home and is occasionally at school’ or ta congbu zai jia, yizhi zai xuexiao ‘He is never at home and is always at school” can be expressed as SYV1ZV2 or SZV1YV2 when Y is jingchang ‘often’ or zongshi ‘always’ and Z is ouer ‘occasionally’ or congbu ‘never.’ Wu and Ding (2005) apply these formulae to differentiate frequency adverbs. Those expressed as SXV1XV2 are mid-level frequency adverbs, and those expressed as SYV1ZV2 or SZV1YV2 are the highest, the lowest, the higher or the lower frequency adverbs. Taking both semantic meanings and syntactic performance into consideration, they classify frequency adverbs into five levels, as shown in Table 2-1.. Table 2-1 The Five Levels and Their Examples of Frequency Adverbs Frequency. Examples. the highest level. zongshi ‘always’. the higher level. changchang ‘usually’. the middle level. youshi ‘sometimes’. the lower level. ouer ‘occasionally’. the lowest level. congwei ‘never’ (Wu and Ding 2005:21) 11.

(23) However, their classification is still too broad. For instance, the meanings of the frequencies denoted by henshao ‘seldom’ and jihubu ‘rarely’ are less frequent than ouer ‘occasionally’ but more frequent than congwei ‘never’ (Zheng 2011). Obviously, there is no room for henshao ‘seldom’ and jihubu ‘rarely’ to stay in Wu and Ding’s five-level classification. Consequently, classifying frequency adverbs into more levels is necessary. After classifying frequency adverbs, Wu and Ding (2005) further analyze frequency adverbs in sentence negation, observing the form of “negation + frequency adverb” in discourse. It is found that the meaning of the frequency adverbs of the highest level in sentence negation is wider because it corresponds to the meanings of the frequency adverbs in the other four levels, as shown in (1).. 12.

(24) (1) Q: Ta zongshi badianzhong dao xuexiao ma? he always eight o’clock arrive school particle ‘Does he always arrive at school at eight o’clock?’ A:. a.. Bu, ta. bu. zongshi badianzhong dao. xuexiao,. no he not always eight o’clock arrive school (ta ouer qidianban lai,)1 zhi neng shuo ta he occasionally seven thirty arrive only can say he shi youshi badianzhong laidao xuexiao. focus marker sometimes eight o’clock arrive school ‘No, he does not always arrive at school at eight o’clock. He usually. b.. c.. d.. arrives at school at seven thirty. It can only be said that he sometimes arrives at school at eight o’clock.’ Bu, ta bu zongshi badianzhong dao xuexiao, no he not always eight o’clock arrive school zhi neng shuo ta jingchang badianzhong dao xuexiao. only can say he often eight o’clock arrive school ‘No, he does not always arrive at school at eight o’clock. It can only be said that he often arrives at school at eight o’clock.’ Bu, [ta bu zongshi badianzhong dao xuexiao,] ta no he not always eight o’clock arrive school he zhishi ouer badianzhong dao xuexiao. only occasionally eight o’clock arrive school ‘No, he does not always arrive at school at eight o’clock. He simply usually arrives at school at eight o’clock.’ Bu, ta congwei badianzhong daoguo xuexiao. no he never eight o’clock arrive school ‘No, he never arrives at school at eight o’clock.’ (Wu and Ding 2005:22). As can be seen in (1), the scope of bu zongshi ‘not always’ includes the affirmative form of the frequency adverbs in the other four levels—youshi ‘sometimes,’ jingchang ‘often,’ ouer ‘occasionally’ and congwei ‘never.’. 1. The words in the parentheses, (), can be omitted. 13.

(25) In addition to being placed before frequency adverbs, negators can also be placed after frequency adverbs. Wu and Ding (2005) also examine the form of “frequency adverb + event” and its negation, “frequency adverb + negation + event,” such as ouer lai ‘to come occasionally’  ouer bu lai ‘not to come occasionally’, concluding the following pattern.. frequency adverb + event. frequency adverb + negation + event. the highest level. the highest level. the higher level. the higher level. the middle level. the middle level. the lower level. the lower level. the lowest level. the lowest level. Figure 2-1 Wu and Ding’s (2005) Analysis of Relation Among Frequency Adverbs and Their Negative Forms. Figure 2-1 shows that the frequency adverb of the highest level followed by a negation is semantically equal to the frequency adverb of the lowest level (e.g., ta zongshi bu lai ‘He always does not come.’= ta congwei lai ‘He never comes.’), the frequency adverb of the higher level followed by a negation is semantically equal to the frequency adverb of the lower level, and so as the other frequency adverbs of the rest of the three levels. Overall, Wu and Ding (2005) classify frequency adverbs into five levels and then 14.

(26) examine frequency adverbs with negation right before and right after frequency adverbs. However, we argue that frequency adverbs should be classified into more than five levels according to the frequencies frequency adverbs denote. In Wu and Ding’s classification, ouer ‘occasionally’ belongs to the lower level and congwei ‘never’ belongs to the lowest level, but there are still some frequency adverbs whose meanings of frequencies are between ouer ‘occasionally’ and congwei ‘never,’ such as henshao ‘seldom’ and jihubu ‘rarely.’ Moreover, the issue of what kind of events can be modified by frequency adverbs is not discussed in Wu and Ding (2005).. 2.1.2 Zou (2006) Zou (2006) examines the scopes of Chinese frequency adverbs defined by previous studies, such as Liu et al. (1983), Xing and Ma (1992), Zhou (1999) and Zhang (2000), and has found that they fail to reach consensus. The main reason behind the different scopes determined in the previous studies is that there is no definite standard which distinguishes frequency adverbs from other adverbs that denote time or momentum, such as yizhi ‘all the times’ and zhanshi ‘for the moment.’ Another problem is the inconsistent classification of frequency adverbs. Since the previous classification is not objective enough, Zou (2006) attempts to offer a precise definition of frequency adverbs and re-classify frequency adverbs into various types by taking their semantic functions into consideration. 15.

(27) He defines frequency adverbs as adverbs which express the number of times an event or an action happens within a particular time. There are three semantic features carried by frequency adverbs: diffuseness, indefiniteness and quantifiableness. First, diffuseness, contrary to continuity, specifies that frequency adverbs indicate the repetition of an event or an action. For instance, changchang ‘usually’ is a frequency adverb while yizhi ‘all the times’ is an adverb of time which possesses the feature of continuity and thus is not a frequency adverb. Second, indefiniteness, also called vagueness, points out that the semantic property of frequency adverbs is indefinite or vague. For instance, changchang ‘usually’ implies that an event or an action happens many times but how many times it actually happens is unknown. Third, quantifiableness means that frequency adverbs reflect quantity. For instance, changchang ‘usually’ implies that its number of times is more than that of ouer ‘occasionally.’ According to Zou (2006), these three features distinguish frequency adverbs from other adverbs and thus the scope of frequency adverbs is clearly defined. After clarifying the scope of frequency adverbs, Zou (2006) provides a new classification of frequency adverbs. Frequency adverbs are divided into two types according to their different functions: that of judgment (Type A), which is to judge the frequency of an event and to identify the focus when accompanied by shi ‘be,’ and that of description (Type B), which is to describe the frequency of an event. Both Types A 16.

(28) and B are further divided into subtypes, as presented in Figure 2-2.. A1: zhongshi. Type A (judgmental). A2. Frequency adverbs. A2-1: tongchang A2-2: changchang. B1: jielian. Type B (descriptive). B2: yizai. Figure 2-2 The Subtypes of Frequency Adverbs in Zou (2006). Those frequency adverbs belonging to Type A can be followed by shi ‘be,’ as shown in (2).. (2) Jingchang often. shi women be we. lia hen suiyi two very free. de jiu DE auxiliary. he guang yi ping beijiu. drink empty one CL wine ‘It is often that we drink one bottle of wine freely.’. (Zou 2006:40). Adopted Chang (2003), Zou (2006) argues that the function of the form “frequency adverbs + shi” is to identify the focus and to emphasize the importance of the following phrase. In (2), therefore, “jingchang shi” serves the function of identifying that it is we, rather than other people, who drink one bottle of wine freely. Most frequency adverbs under Type B can be followed by the adverbial suffix de 17.

(29) to serve the function of describing an event or an action, as shown in (3).. (3) Xiaoniu zaisan. de. qing. xiaoniu again and again DE ask for ‘Xiaoniu asks for a war again and again. …’. zhan …. a war… (Zou 2006:40). Although some frequency adverbs of Type A can be followed by de, such as changchang ‘usually,’ it does not mean that frequency adverbs of Type A possess the function of description. The differences of Types A and B can be perceived from two aspects. First, frequency adverbs of Type A can be negated while those of Type B cannot (e.g., bu changchang ‘not usually’ vs. *bu lu ‘not repeatedly’). Second, when frequency adverbs of Types A and B co-occur in a sentence, Type A must precede Type B. Descriptive frequency adverbs (i.e., Type B) are used to directly describe the frequency of an action so the relation between a descriptive frequency adverb and a verb in a sentence is close. On the other hand, judgmental frequency adverbs (i.e., Type A) are used to judge the frequency of an action and the relation between a judgmental frequency adverb and a verb is not so close compared with the relation between a descriptive frequency adverb and a verb. On account of proximity of a descriptive frequency adverb and a verb in meaning, a judgmental frequency adverb must precede a descriptive one when they co-occur in a sentence. Frequency adverbs of both Types A and B can be subdivided. Frequency adverbs 18.

(30) of Type A are of two subtypes, Type A1 and Type A2, according to whether they can express duration of an event or an action.. (4) … Chenqingyang shubiao li na shuang po xie … chenqingyang schoolbag inside that CL old shoes lao /*changchang bu reng. always/*usually not throw away ‘…. Chenqingyang has not thrown the pair of old shoes in his schoolbag for a while. ….’. (Zou 2006:41). Lao ‘always’ in (4) expresses duration of the action and thus belongs to Type A1 while changchang ‘usually’ cannot serve the same function and falls under Type A2. Frequency adverbs of Type A2 can be further subdivided into Types A2-1 and A2-2. Frequency adverbs of Type A2-1 (e.g., tongchang ‘usually’) can stay in non-focus positions only; those of Type A2-2 (e.g., jingchang ‘often’) can stay in both focus and non-focus positions, as shown in (5).. (5) a.. b.. Wo tongchang I usually. yi ge ren one CL person. dai zai jia stay at house. li. inside. ‘I usually stay at house alone. Wo jingchang yi ge ren I often one CL person ‘I often stay at house alone.. dai zai jia stay at house. li. inside (Zou 2006:42). In (5a), the focus can be yi ge ren ‘one person’ or jia li ‘house’ but not tongchang ‘usually.’ Due to the fact that tongchang ‘usually’ can stay in non-focus positions only, 19.

(31) it belongs to Type A2-1. In (5b), however, the focus can be yi ge ren ‘one person,’ jia li ‘house’ or jingchang ‘often.’ Jingchang ‘often’ can stay in both focus and non-focus positions, so it belongs to Type A2-2. There are also two subtypes of frequency adverbs in Type B. Frequency adverbs of Type B1 can be used to express duration of an event or an action while that of Type B2 cannot, as shown in (6).. (6) Tamen jielian/*yizai chang they continuously/again and again sing ‘They have sung for six years ….’. le PERF. liu nian …. six year … (Zou 2006:43). In (6), jielian ‘continuously’ can be used to modify the duration of the event of their singing and thus belongs to Type B1. On the contrary, yizai ‘again and again’ belongs to Type B2. Among the subtypes of frequency adverbs classified by Zou (2006), only frequency adverbs of Types A2 and B2 are typical frequency adverbs. Frequency adverbs in Type A1 cannot be placed in the focus position, which differs from the general tendency of placing frequency adverbs in the focus position, so frequency adverbs of this type are atypical. While frequency adverbs of Type A2-1 can be placed in a focus position, they have the semantic features of recurrence and duration, which are not typical features of frequency adverbs. Similarly, frequency adverbs of Type B1 20.

(32) have the semantic features of recurrence and duration and thus belong to atypical frequency adverbs. From the perspective of semantic functions and syntactic positions, frequency adverbs can be classified into typical and atypical, but whether the distinction between typical and atypical frequency adverbs exerts influence on children’s acquisition has remained unknown. To conclude, Zou (2006) identifies three semantic features of frequency adverbs: diffuseness, indefiniteness and quantifiableness. He also provides the subtypes of frequency adverbs, as presented in Figure 2-2. Among these subtypes, only Types A2-2 and B2 are identified as typical frequency adverbs; the others are atypical frequency adverbs. Furthermore, Zou (2006) does not discuss the distinction between events that can and cannot be modified by frequency adverbs. The present study will include this issue and explore whether children recognize the distinction.. 2.1.3 Piao (2011) Frequency adverbs can be used to indicate the recurrence of an action or an event or the continuity of a state. There are a series of frequency adverbs, and the difference between neighboring frequency adverbs is blurred. However, when they are classified into different levels, the boundaries become clearer. Piao (2011) classifies Chinese frequency adverbs into three levels: the highest level includes zong(shi) ‘always,’ lao(shi) ‘always’ and so on; the higher level chang ‘often,’ jingchang ‘often,’ 21.

(33) changchang ‘usually,’ and so on; the lower level youshi ‘sometimes,’ ouer ‘occasionally’ and so on. Frequency adverbs are examined by Piao (2011) in terms of their semantic features and morphology, and he identifies five semantic features of frequency adverbs. First of all, quantifiableness means that frequency adverbs can be measured by their quantity. From the perspective of quantity, frequency adverbs of the highest level are permanent; those of the lower level imply few times of recurrence. Second, diffuseness indicates that the action modified by a frequency adverb is diffused and that action happens repeatedly within a period of time. Third, vagueness demonstrates that the actual times of an event modified by a frequency adverb is not precisely specified. Fourth, continuity exhibits that frequency adverbs actually imply a state of continuing for a period of time within intervals. Fifth, irregularity presents that when an action happens is unsure and the interval between the recurrences of an action is unclear. Regarding morphology, some frequency adverbs are formed by two identical phonological syllables, such as lulu ‘repeatedly,’ pinpin ‘incessantly’ and changchang ‘usually.’ According to Piao (2011), frequency adverbs with reduplication display greater intensity than their monosyllabic counterparts. For instance, changchang ‘usually’ implies more frequent than chang ‘often.’ However, we find that the reduplication may not definitely indicate greater intensity of frequency. The difference 22.

(34) of the two may also lie in their distributions—disyllabic frequency adverbs are mainly used in spoken form and their monosyllabic counterparts are used in written form. In other words, their meanings are identical and speakers decide the suitable wording according to the forms or occasions. Piao (2011) further argues that frequency adverbs mainly stay in positions either before subjects or between subjects and predicates. The latter is typical. Sentences with frequency adverbs before subjects are marked and can be restored to the position between subjects and predicates. The function of this type of sentences is to highlight the importance of frequency and to link the previous context.. (7) a.. b.. Wanshang night Ouer occasionally. yi pian qihei. one CL dark renjia chuanghu touchu yi dian family window penetrate one CL. dengguang, …. light. ‘It is very dark at night. Some lights are occasionally penetrated through the windows of the family.’ (Piao 2011:139) Renjia chuanghu ouer touchu yi dian dengguang, …. family window occasionally penetrate one CL light ‘Some lights are occasionally penetrated through the windows of the family.’. As demonstrated in (7a), the frequency adverb ouer ‘occasionally’ emphasizes the importance of frequency and links the two sentences, and it can be restored to the position between the subject and the predicate like (7b). In other words, placing a frequency adverb in either position in Mandarin does not affect the grammaticality or 23.

(35) acceptability of the sentence. Therefore, it is not necessary to examine whether the two positions will influence children’s interpretation of frequency adverbs. Because the latter position is more typical than the former one, all frequency adverbs will be placed between subjects and predicates in the present study. With regard to the selection of frequency adverbs, frequency adverbs require that the events they modify should be able to recur. Thus, events that are unlikely to happen many times cannot be modified by frequency adverbs, as shown in (8).. (8) ?Ta qunian. changchang. daxue. biye.. he last year usually college graduate ‘?He usually graduated from college last year.’. (Piao 2011:139). However, frequency adverbs can modify this kind of events when the subject or object is a collective noun because it indicates a collection or a number of people or things. In this case, the recurrence does not limit to an individual. Rather, the recurrence indicates that the action happens to many individuals belonging to this group. In sum, Piao (2011) classifies frequency adverbs into three levels and examines frequency adverbs from the perspective of semantics and syntax. As discussed above, frequency adverbs should be classified into more levels or categories according to the frequencies frequency adverbs denote. Regarding semantics, in addition to quantifiableness, diffuseness and vagueness, also called indefiniteness, which have 24.

(36) been identified by Zou (2006), Piao (2011) adds two more semantic features— continuity and irregularity. With respect to syntax, the typical position of frequency adverbs is between a subject and a predicate. Besides, events that frequency adverbs can modify are restricted. Whether children detect the restriction of frequency adverbs modifying events deserves to be explored.. 2.1.4 Summary To sum up, the term, frequency adverbs, is generally defined by recent literature as the adverbs that denote approximate number of times an event or action happens within a period of time. Since frequency adverbs are highly related to time, they are believed to be one of the categories of adverbs that denote time in some previous literature, such as Zhao (1979) and Zhu (1982). However, a clear distinction between frequency adverbs and time adverbs has been drawn in Zou (2006). Piao (2011) also identifies five semantic features of frequency adverbs. After clarifying the scope of frequency adverbs, most previous studies classify frequency adverbs into three (Piao 2011) or five (Wu and Ding 2005) levels or types according to the frequencies they denote. However, these previous studies fail to distinguish among frequency adverbs that denote low frequencies (i.e., henshao ‘seldom,’ jihubu ‘rarely’ and congbu 2. The present study uses congbu ‘never’ to replace the more challenging word, congwei ‘never,’ used in Wu and Ding (2005). 25 2.

(37) ‘never’). Both jihubu ‘rarely’ and congbu ‘never’ have the negative element bu. Whether children are able to distinguish between the two frequency adverbs remains an issue. Furthermore, both shao and bu indicate negative meanings in Mandarin. Whether children have the ability to recognize that henshao ‘seldom’ denotes more frequent than jihubu ‘rarely’ and congbu ‘never’ can be explored. Therefore, the present study will extend five types of Chinese frequency adverbs to seven. Moreover, Zou (2006) distinguishes between typical and atypical frequency adverbs. The present study will explore whether the two types of frequency adverbs display different degrees of difficulty and contribute to a different order of acquisition. In the next section, we will review four empirical studies of a set of expressions that denote various degrees of a category.. 2.2 Empirical Studies of Vague Linguistic Terms The aim of the present study is to examine children’s acquisition of Chinese frequency adverbs, whose meanings are vague, so previous empirical studies that discussed people’s interpretations of vague meanings are reviewed. Due to the fact that there are few empirical studies on vague meanings of a set of related words in Mandarin, empirical studies that investigated languages other than Mandarin are included in this section—two studies of people’s interpretations of probability expressions in Hebrew and in English (Beyth-Marom 1982, Hoffner, Cantor and Badzinski 1990) and two 26.

(38) studies of people’s interpretations of frequency expressions in English and in Mandarin (Kuczaj 1975, Huang and Huang 2011). Among these four studies, Kuczaj (1975), Hoffner et al. (1990) and Huang and Huang (2011) explored children’s acquisition of vague linguistic terms while Beyth-Marom (1982) addressed the issue of adults’ interpretations of vague linguistic terms.. 2.2.1 Kuczaj (1975) Children’s acquisition of lexical meanings has been widely studied in literature, and it has been concluded that children acquire partial meaning of a word before knowing its meaning (cf. Clark 1973a). Regarding the acquisition of antonyms, previous studies (e.g., Donaldson and Balfour 1968, Donaldson and Wales 1970, Clark 1971, Palermo 1973) have found that children tend to confuse the meaning of an antonym with that of its opposite. More than children’s acquisition of antonyms, Kuczaj (1975) studied children’s acquisition sequence of a set of related words by investigating five terms denoting various degrees of frequency. Two experiments were conducted. Forty preschoolers, aged from 40 to 56 months, participated in Experiment 1 and thirtytwo preschoolers, aged from 42 to 61 months old, took part in Experiment 2. In Experiment 1, a comprehension task, a judgment and imitation task were designed to examine whether the subjects understood always and never. According to Kuczaj (1975), always and never denote the two extremes of frequency, so the meanings 27.

(39) of the two terms could be regarded as differing along a positive-negative dimension. To be precise, these two terms are strong positive and strong negative respectively, rather than simply positive or negative because The boy eats dinner and The boy always eats dinner are not synonymous. One might consider that “absolute” rather than “positive” better captures the meanings of the two terms in the dimension of frequency, but Kuczaj (1975) argued that the terms are not always used to denote absoluteness. For instance, I always eat breakfast is not necessarily false if the speaker missed his breakfast for one time or another. Thus, Kuczaj’s study distinguished strong positive/negative from positive/negative. For testing the subjects’ comprehension of always and never, the comprehension task consisted of a number of pairs of sentences that included always, never or a VP only. The subjects were asked to answer “Who VPs?” after hearing each pair of sentences. For the judgment and imitation task, the subjects were asked to judge whether the sentences with or without nonsense words were understandable and were asked to repeat the sentences. The purpose of this task was to examine the subjects’ decision on the understandability of the test sentences and whether the subjects deleted nonsense words, always and never when they repeated the test sentences. It was assumed that if a subject treated a term as a nonsense word, he would judge that sentence as not understandable and would delete the term when he repeated the sentence. The results of Experiment 1 demonstrated that never sentences were answered 28.

(40) correctly more frequently than were always sentences though the difference was not statistically significant. Moreover, half of the subjects understood that always is positive and never is negative but not that they are strong positive or negative. Some well knew that always and never are strong positive and negative; some believed always and never are both positive or both negative; still some treated always and never as nonsense words. One of Kuczaj’s friend suggested that children might initially acquire the meaning of sometimes and then contrast always and never to sometimes, which could, at this stage, explain the reason that some subjects believed always and never are both positive or both negative. To test the hypothesis, Kuczaj conducted Experiment 2. Similar to Experiment 1, Experiment 2 consisted of a comprehension task and a judgment and imitation task. What differed from Experiment 1 was that the test sentences included not only always and never but also usually, seldom and sometimes. The results of Experiment 2 suggested that the correct rate of never sentences responded by the subjects was the highest among the correct rates of the five terms. Also, some subjects correctly interpreted always and never but treated sometimes as nonsense. The data, therefore, failed to support the hypothesis proposed by Martin Braine, which predicted that sometimes would be acquired before always and never. In summary, through the two experiments, Kuczaj (1975) found that never was the 29.

(41) easiest among the five frequency adverbs and that never and always were easier than the other three terms lying in the inner positions in the dimension of frequency. Although this study proved that children acquired the meanings of the outermost members before acquiring the meanings of the inner members, the task design was not appropriate. For instance, in a pair of the comprehension task, the subjects heard two sentences, The girl is sitting on the chair and The policeman always sits on the chair, and were asked who sits on the chair. The correct response set by Kuczaj was “the policeman,” but actually “the girl” also sits on the chair. When the subjects answered “the girl,” it might not be the case that they were unable to figure out the meaning of always.. 2.2.2 Beyth-Marom (1982) Previous studies, such as Lichtenstein and Newman (1967), have concluded that the interpretations of probability expressions (e.g., likely, probably, etc.) is extremely vague. However, verbal probability expressions are more common than numerical ones for a number of reasons. First, people are unware of the vagueness. Second, numerical probability expressions are precise values which create a speaker great stress of being completely correct when using them; however, verbal probability expressions are subjective. Third, it is sometimes difficult for people to specify a precise number to express the probability; thus, they prefer verbal expressions to numerical expressions. 30.

(42) Yet, the feature of vagueness causes some problems in communication. Therefore, Beyth-Marom (1982) carried out an experiment to demonstrate the problems caused by verbal probability expressions. Two parts, the isolation part and the context part, were conducted in a professional forecasting organization in Israel. The subjects were all experts in political forecasting. Twenty-seven subjects took part in the isolation part, and thirty-two subjects joined the context part. In the isolation part, thirty Hebrew verbal probability expressions were collected. Most of them were taken from the organization’s own published political forecasts. The remaining seven verbal probability expressions were extreme expressions, such as “not likely” or “near certain,” that were not found in the organization’s publications. The subjects were asked to finish five tasks in this section. Task 1 was a numerical translation task which asked the subjects to give each of the thirty verbal probability expressions a number between 0 and 100 where 0 expressed absolute confidence that the event will not happen; 100 expressed absolute confidence that the event will definitely happen. Task 2 was an event-classification task which asked the subjects to classify 50 events with probability into seven categories according to their probable levels. Task 3 was a probability-expression classification task which asked the subjects to order the thirty probability expressions from the least probable to the most probable. Task 4 was a matching task which asked the subjects to match each probability category 31.

(43) to one of the seven event categories. Task 5 was a second numerical translation task which asked the subjects to give each probability-expression category a number between 0 and 100. In the context part, a questionnaire which included fifteen selected paragraphs was distributed to the subjects. Each paragraph contained a number of words expressing the probability that an event would happen. The subjects were told to write a number, between 0 to 100, to indicate their thoughts on the probability of the happening of the events. The results of the isolation part showed that the subjects were highly consistent in their numerical translation in Task 1 and Task 5. With regard to between-subject consistency, the largest range of the number of an expression was 25, which implied the vague meaning of the expressions. In addition, in the grouping of the seven categories (i.e., I to VII), the expressions in the two extreme categories (i.e., I and VII) were those seldom or never found in the organization’s publications, which indicated that there were more probability words expressing intermediate probabilities than extreme ones. The results of the context part also showed that the subjects were highly consistent. With regard to between-subject consistency, the subjects’ judgments were more variable when a probability expression appeared in context than in isolation. The reason might 32.

(44) be that the given paragraphs might have not been defined clearly, causing huge differences among the subjects’ interpretations. Also, the subjects’ desirability of an event might have influence on their interpretations. Overall, the results revealed that the subjects’ interpretations of most verbal probability expressions were variable as evidenced in the large range of numbers given to each expression. In addition to testifying the vague meanings of verbal probability expressions, the contextual effect was examined in the study and the results revealed that subjects’ judgments were more consistent in the isolation part than in the context part. However, whether the given paragraphs in the context part were clear enough remains a question. Hence, the results of the contextual effect in the study were not convincing.. 2.2.3 Hoffner, Cantor and Badzinski (1990) Children’s acquisition sequence of serially ordered semantic sets is studied in literature. However, children’s comprehension of expressions that denote modality, such as likely and possibly, has not received much attention in the previous literature (Coates 1988). Therefore, Hoffner et al. (1990) attempted to examine children’s understanding of three terms which denote different degrees of likelihood: possibly, probably and definitely. Among these three words, definitely has an absolute meaning. The meaning of the other two words are vague since there is no general agreement on 33.

(45) the percentage of occurring of the events they modify, though adults generally agree that probably indicates a higher degree of likelihood than possibly. To examine children’s understanding of the meanings of the terms, Hoffner et al. (1990) conducted two studies and recruited seventy-eight subjects respectively from various grade levels. Subjects of Study 1 were preschoolers, first, third and fourth graders. Study 2 recruited subjects at preschool, the first, third and fifth grade. In Study 1, a comprehension task was conducted to examine children’s understanding of the three terms. The subjects heard stories about two people who would be likely to engage in an activity, and they were asked to predict which of the two was more likely to do the activity. The three words were taken as pairs, and all the possible combinations and permutations were formed to make a total of six pairs. Six additional pairs were formed from the negative forms of the three terms (i.e., possibly not, probably not, definitely not). Therefore, twelve stories, each of which included two people with one of the above pair of terms modifying the likelihood of engaging in a particular activity, were told to the subjects. In Study 1, the subjects were asked to compare the likelihood denoting by the two of the three terms in a story. However, people are not commonly required to make such comparisons in daily life. To make the experiment closer to real life, the other study was conducted. In Study 2, the subjects heard a story about three children, who watched 34.

(46) a movie where something really scary happened. The subjects then heard a separate statement about each of the three children, describing the child’s perceived likelihood that the scary event would happen. Each statement contained only one of the three terms and the subjects were asked to score the probability from 1, “not scared at all” to 7, “very very scared.” The results of Study 1 revealed that there were significant effects of both grade levels (F (3,74) = 22.43, p < .001) and word combinations (F (2,148) = 30.47, p < .001). As expected, the subjects’ performance improved with grade levels. In addition, their performance was more accurate in distinguishing between definitely and each of the other two terms (i.e., definitely and possibly, definitely and probably) than that of distinguishing between possibly and probably. That is to say, it was easier for them to understand the distinction between certainty (100%) and uncertainty than it was to understand the distinction between two different degrees of uncertainty. Negative counterparts of the statements were expected to be more difficult than the affirmative form (cf. Clark and Clark 1977); however, there were no significant effects. The results of Study 2 demonstrated that there were significant effects of both the form of the statements (F (1,70) = 76.52, p < .001) and grade levels (F (3,70) = 4.59, p < .001). The subjects rated the children as more scared when the statements described the likelihood that a scary event would happen (mean = 5.4) than statements with would 35.

(47) not happen (mean = 2.8). The analysis also displayed an interaction between the form of the statements and the three terms (F (2,140) = 14.75, p < .001). The interaction indicated that the correct rank order of the terms in affirmative forms was the reverse of the correct rank order of the terms in negative forms. That is to say, the greater the likelihood a scary event will happen, the more scared someone felt, while the greater the likelihood a scary event will not happen, the less scared someone felt. An additional analysis revealed that the percentage of the subjects who understood the relative degrees of the likelihood denoting by the three terms (i.e., possibly < probably < definitely) increased with grade level (X2 (3, N = 78) = 7.59, p = .055): the preschoolers 11.1%, the first graders 26.1%, the third graders 35.0% and the fifth graders 52.9%. The percentages indicated that by grade five, more than half of the subjects correctly understood the rank order of the relative degrees of the likelihood the terms denoted. Overall, the subjects’ understanding of the distinction of various degrees of likelihood denoting by the terms increased with grade levels. In addition, distinguishing between definitely and one of the other two terms was easier for children than distinguishing between possibly and probably. This result testified to the general belief that children tend to acquire the meanings of absolutely positives and negatives, such as always and never, before they understand the meanings of inner members in a serially ordered semantic set. However, the subjects’ grade levels were not consistent in the two 36.

(48) studies of Hoffner et al. (1990). The present study will make the subjects’ grade levels identical in all tasks.. 2.2.4 Huang and Huang (2011) In the field of first language acquisition in Mandarin, studies about frequency adverbs have not been received much attention. Huang and Huang (2011) extended Wu’s (2006) finding on the basis of cognitive semantics related to frequency adverbs, aiming at analyzing children’s acquisition of frequency adverbs. The data obtained by Huang and Huang was from children aged one to five years old in a kindergarten in China. They recorded children’s talking, and identified the occurrence of each frequency adverb. Their study classified frequency adverbs into three levels, high-value, medium-value and low-value (cf. Shi 2002), and compiled a table which demonstrated the youngest age of children producing each frequency adverb.. 37.

(49) Table 2-2 The Youngest Age of Children Producing Each Frequency Adverb Ages. 2:6. Levels. 3:6. 4:0. 4:6. 5:0. zong, High-value. zongshi, laoshi, yizhi. lao. changchang, jingchang. Medium-value Low-value. conglai. youshihou (Huang and Huang 2011:77). High-value frequency adverbs are used to indicate that a particular event happens extremely frequently in a period of time. Medium-value frequency adverbs imply a particular event happens many times in a period of time. Only two expressions of this category were found in the data. Low-value frequency adverbs indicate a particular event happens rarely in a period of time. In the data, only youshihou ‘sometimes’ occurred. As can be seen in Table 2-2, it is obvious that the number of frequency adverbs uttered by children under five was small. They uttered only nine frequency adverbs. Moreover, the table also demonstrated that the high-value frequency adverb, lao ‘always,’ was uttered firstly (2:6), followed by the low-value frequency adverb, youshihou ‘sometimes’ (3:6). It was not until the age of four that children uttered the first medium-value frequency adverbs, changchang ‘usually’ or jingchang ‘often.’ In 38.

(50) principle, frequency adverbs uttered earlier were used by speakers more frequently; therefore, it could be inferred that high-value frequency adverbs were used by children more often. Huang and Huang (2011) provided explanations from a psychological perspective to account for the few number of frequency adverbs uttered by children. Children’s language development is based on the experience of daily life. They know concrete objects or the things they can contact earlier than abstract concepts. Therefore, it is reasonable that they acquire expressions that denote concrete objects earlier than those that denote abstract concepts. Frequency adverbs do not have concrete counterparts in the real world, so they are acquired later. Frequency adverbs belong to adverbs so children acquire frequency adverbs later. Moreover, since frequency adverbs are generally used to describe an event that has happened in the past, children need to rely on their memory to think about how to describe the frequency of that event. The development of children’s memory ability must be the basis for children to acquire frequency adverbs. According to Busby and Suddendorf (2005), children start to develop the concept of past and future at the age around four. The subjects of Huang and Huang (2011) were no more than five. That was the reason why the number of frequency adverbs uttered by children in the data was small. Overall, Huang and Huang (2011) found that the number of frequency adverbs 39.

(51) uttered by children under the age of five was small, and that children firstly uttered a high-value frequency adverb rather than a medium- or a low-value frequency adverb. They argued that an event which happens repeatedly leaves people more impression than an event which happens only once. Due to the limitation of children’s cognitive ability, repeated events leave children more impression. However, the study was a corpus-based approach and no experiment was conducted to examine children’s understanding of frequency adverbs. Whether children correctly used those frequency adverbs in context was not explored. Also, standard low-value frequency adverbs, such as henshao ‘seldom,’ and congbu ‘never,’ were not found in the data. Was it because children tended to describe what they did rather than what they did not do in conversation? In other words, is it possible that children acquired low-value frequency adverbs early but seldom uttered them? If this is the case, the memory hypothesis might be rejected.. 2.2.5 Summary Previous empirical studies of people’s understandings of vague meanings have been reviewed in this section and Table 2-3 summarizes the major findings and limitations of these studies.. 40.