英語冠詞之第二語言習得

全文

(1)國立立臺灣師範大學英語學系碩士論論文 Master Thesis Graduate Institute of English National Taiwan Normal University. 英語冠詞之第二語言習得 Second Language Acquisition of Articles in English by Chinese Learners. 指導教授：陳純音博士 Advisor: Dr. Chun-yin Doris Chen 研究生：張庭芝 Graduate: Ting-jhih Tiffany Chang. 中華民國一百零零三年年六六月 June, 2014.

(2) 摘要. 本研究旨在探討以中文為母語的學生對英語冠詞之習得狀狀況，研究議題包括不不同冠詞類類型和其所屬的語意特徵之難度度順序、題型效應、語境強化之影響及語言程度度效應。本研究採兩兩個實驗題型:選擇題與填充題，試題皆以對話呈現。受試者為八十位以中文為母語的大一非英語系學生及二十位英語母語人士，並依據其英語程度度分為初、中初、中高及高級四組。實驗結果發現，三種類類型的冠詞和其語意屬性之難易易度度會受到第一語言轉移及英文冠詞系統本身的複雜度度影響。在三種冠詞中，以零零冠詞表現最佳，也最容易易習得；定冠詞 the 表現最不不穩定，習得時間最晚。在冠詞的三種語意特徵的習得上，以[+SR, +HK]最易易習得而[-SR, -HK]最為困難。此外，定冠詞 the 較與[+HK] 之語意屬性有較高相關性。在題型效應方面，當題型較受控制，受試者面對選擇題與填充題的表現大致相同。最後，英語程度度也證實會影響冠詞的習得。受試者英語程度度越高，冠詞的表現越好，對冠詞功能的理理解和使用皆會隨著英語能力力的提升而有所增進。. 關鍵字: 英文冠詞難度度次序、英文冠詞語意特徵難度度順序、第一語言轉移、題型效應、語言程度度效應、第二語言習得. i.

(3) ABSTRACT. The present study aims to investigate the second language acquisition of the English article system by Chinese EFL learners. The research questions addressed in the present study include the difficulty order of the three English article types and their semantic features, the task effect, the input enhancement effect and the proficiency effect. The study employed a comprehension task (i.e., multiple-choices) and a production task (i.e., fill-in-the-blankets) in the format of conversations. The experimental group comprised eighty non-English majoring freshmen in Taiwan who were divided into four proficiency groups (i.e., low, low intermediate, high intermediate and advanced) according to their English proficiency level. The control group composed of twenty native English speakers. The overall results showed that the three article types and their underlying semantic features exhibited different degree of difficulty, and the sequence was subject to the L1 transfer effect and the complexity of the L1 English article system. The definite article the and the feature [-SR, -HK] were the most difficult while the null article and the feature [+SR, +HK] were the easiest. Also, the was found more associated with the feature [+HK] than [+SR]. In response to the task effect, the results did not yield a significant difference, implying the subjects’ competence and performance were similar in acquiring English articles when the production task was controlled. Finally, the proficiency effect did play an influential role in the acquisition of English articles. In sum, the higher the proficiency, the better the performance.. Keywords: difficulty order of English articles, difficulty order of English article features, L1 transfer, task effect, proficiency effect, L2 acquisition. ii.

(4) ACKNOWLEDGEMENTS. This thesis is dedicated to every person who has ever offered me help and concern, for the thesis would not have been completed without the assistance of many people. First of all, I would like to express my utmost appreciation and gratitude to my advisor Professor Dr. Chun-yin Doris Chen, who has been a tremendous mentor for me and has inspired me in so many ways from teaching to academic study. I would like to thank you for encouraging my research and for allowing me to grow as a researcher. Your advice on both research as well as on my life have been priceless. I would also like to thank my committee members, professor Rueih-Lirng Sharon Fahn and professor Jen-I Li for serving as my committee members even at hardship. I also want to thank you for letting my defense be an enjoyable moment, and for your brilliant comments and suggestions. Also, I am indebted to the teachers of the Freshman English classes, and my school sister and brother, Carol Liao and Matt Ku for helping me in recruiting the subjects. I am also grateful to all the participants and to all of my friends who supported me in writing, brainstorming and incented me to strive towards my goal. A special thanks is to my beloved family. Words cannot express how grateful I am to my parents. Your support for me was what sustained me thus far.. iii.

(5) TABLE OF CONTENTS. CHINESE ABSTRACT ................................................................................................. I ABSTRACT .................................................................................................................. II ACKNOWLEDGEMENTS ......................................................................................... III TABLE OF CONTENTS .............................................................................................IV LIST OF TABLES .......................................................................................................VI LIST OF FIGURES ................................................................................................... VII LIST OF ABBREVIATIONS ................................................................................... VIII CHAPTER ONE INTRODUCTION .......................................................................... 1 1.1 Theoretical Background and Motivation ............................................................. 1 1.2 Research Questions .............................................................................................. 5 1.3 Significance of the Study ..................................................................................... 5 1.4 Organization of the Thesis ................................................................................... 6 CHAPTER TWO LITERATURE REVIEW ............................................................... 7 2.1 Theoretical Studies of Articles in English............................................................ 7 2.1.1 Hawkins (1978) ............................................................................................. 7 2.1.2 Huebner (1983) ............................................................................................. 9 2.1.3 Lyons (1999) ............................................................................................... 12 2.1.4 Ionin et al. (2004) ........................................................................................ 15 2.1.5 Summary ..................................................................................................... 18 2.2 Previous Empirical Studies of English Articles ................................................. 19 2.2.1 Thomas (1989) ............................................................................................ 19 2.2.2 Robertson (2000) ........................................................................................ 21 2.2.3 Lu (2001) .................................................................................................... 23 2.2.4 Ionin et al. (2008) ........................................................................................ 24 2.2.5 Summary ..................................................................................................... 25 2.3 Summary of Chapter Two .................................................................................. 28 CHAPTER THREE A REVISED CLASSIFICATION OF ENGLISH ARTICLES 29 3.1 Rational for the Revised Classification.............................................................. 29 3.2 Type One: The Definite Article The ................................................................... 31 3.3 Type Two: The Indefinite Article A .................................................................... 33 3.4 Type Three: The Null Article Ø ......................................................................... 34 iv.

(6) 3.5 A Comparison of the Article System between English and Chinese ................. 35 3.6 Summary of Chapter Three ................................................................................ 37 CHAPTER FOUR RESEARCH DESIGN ............................................................... 38 4.1 Subjects .............................................................................................................. 38 4.2 Methods and Materials ....................................................................................... 39 4.2.1 The Comprehension Task ............................................................................ 41 4.2.2 The Production Task ................................................................................... 42 4.3 Procedures .......................................................................................................... 43 4.3.1 Pilot Study................................................................................................... 43 4.3.2 Formal Study ............................................................................................... 45 4.3.3 Scoring And Statistical Analysis ................................................................. 46 4.4 Summary of Chapter Four ................................................................................. 46 CHAPTER FIVE RESULTS AND DISCUSSION ................................................... 47 5.1 Difficulty Order of English Articles .................................................................. 47 5.1.1 Overall Findings.......................................................................................... 47 5.1.2 General Discussion ..................................................................................... 51 5.2 Difficulty Order of Article Features ................................................................... 53 5.2.1 General Findings ......................................................................................... 53 5.2.2 Article Features for Each Article Type ........................................................ 56 5.2.3 General Discussion ..................................................................................... 63 5.3 Task Effect ......................................................................................................... 66 5.3.1 Overall Findings.......................................................................................... 66 5.3.2 General Discussion ..................................................................................... 68 5.4 Proficiency Effect .............................................................................................. 70 5.5 Summary of Chapter Five .................................................................................. 72 CHAPTER SIX CONCLUSION .............................................................................. 73 6.1 Summary of the Major Findings ........................................................................ 73 6.2 Pedagogical Implications ................................................................................... 74 6.3 Limitations of the Present Study and Suggestions for Future Research ............ 76 BIBLIOGRAPHY ........................................................................................................ 77 APPENDIX A THE COMPREHENSION TASK........................................................ 81 APPENDIX B THE PRODUCTION TASK................................................................ 85 APPENDIX C RESULTS OF THE PILOT STUDY ................................................... 89 APPENDIX D CONSENT FORM .............................................................................. 93. v.

(7) LIST OF TABLES. Table 2-1 A Compared Article Classification from the Previous Empirical Studies ... 18 Table 2-2 Major Findings and Limitations of the Previous Empirical Studies ............ 26 Table 3-1 A Comparison of English and Chinese Articles ........................................... 36 Table 4-1 The Distribution of the Test Items Used in Both Tasks ............................... 41 Table 4-2 A Test Sample Used in the CT ..................................................................... 42 Table 4-3 A Test Sample Used in the PT ..................................................................... 43 Table 5-1 Subjects’ Overall Performance on the Three Article Types of English ....... 47 Table 5-2 The p-value for the Within-Group Difference among Article Types ........... 48 Table 5-3 The Overall Answer Rates for Each Article Types ...................................... 48 Table 5-4 The p-value for the Three Article Types of English..................................... 50 Table 5-5 Subjects’ Overall Performance on the English Article Features .................. 53 Table 5-6 The p-value for the Within-Group Difference among Article Features ....... 55 Table 5-7 The p-value for the Between-Group Difference of Article Features ........... 55 Table 5-8 Subjects’ Overall Performance on Each Article Type.................................. 56 Table 5-9 The p-value for the Within-Group Difference by Each Article Type .......... 61 Table 5-10 The p-value for the Between-Group Difference by Each Article Type ..... 62 Table 5-11 Subjects’ Overall Performance on the Task Effect .................................... 67 Table 5-12 The p-value for the Within-Group Difference between the Tasks ............. 67 Table 5-13 The p-value for the Between-Group Difference of the Task Effect ........... 68. vi.

(8) LIST OF FIGURES. Figure 3-1 A Revised Frame for English Articles ........................................................ 30 Figure 5-1 Overall Performance of Each Proficiency Group on Article Types ........... 49 Figure 5-2 Overall Performance of Each Proficiency Group on the Article Features . 54 Figure 5-3 Overall Performance of Each Proficiency Group on the Features of The.. 57 Figure 5-4 Overall Performance of Each Proficiency Group on the Features of A ..... 59 Figure 5-5 Overall Performance of Each Proficiency Group on the Features of Ø ..... 60 Figure 5-6 Overall Performance of Each Proficiency Group on Task Effect .............. 67 Figure 5-7 The Developmental Stages of English Articles .......................................... 70. vii.

(9) LIST OF ABBREVIATIONS. BA. causative co-verb (ba). CL. classifier. DE. nominalizar (-de). DUR. duration (-zai). EXP. experiential aspect (-guo). PL. plural (-men). PFV. perfective aspect marker (-le). POSS. possessive (-de). Q. question particle (-ma). Rex. response to expectation (ne). viii.

(10) CHAPTER ONE INTRODUCTION. 1.1 Theoretical Background and Motivation According to the Oxford English Corpus over a billion words, English articles the and a are among the top ten frequent words. Nonetheless, researchers have pointed out that English articles remain a constant challenge particularly to L2 learners whose L1 with or without an article system. To uncover the cause of the learning difficulty, numerous studies have attempted to explain English articles via different linguistic domains such as syntax (Schachter 1988), semantics (Ionin & Wexler 2003, Ionin et al. 2004, 2008), pragmatics (Chen 2004) and phonology (Goad & White 2008). One of the recognized causes for the learning difficulty is the mismatch of the morphological mapping from the underlying features to the surface representations as Lardiere (1998, 2000) suggested. Besides the form-and-meaning mismatch, L1 transfers is also known to be an influencing factor. According to the empirical findings of Huebner (1985), Thomas (1989), Lardier (2000) and many others, the most two common errors found in the L2 learners’ acquisition of English articles are omission or overuse of the articles in obligatory contexts or inappropriate contexts. Examples commonly made by articleless learners such as Chinese are shown in (1):. (1) a. She made *(a) phone call to someone. b. She takes a /*the bath. (White 2008:253). While in English NPs may be preceded by articles the, a and Ø based on the singularity/plurality and count/mass properties of the NPs, NPs in Chinese, as Li and Thompson (1981) suggest, do not have an article system as English does. Chinese 1.

(11) resorts to word order, demonstratives or measure words to express the meanings carried by English articles (cf. Cheng & Sybesma 1999). Furthermore, according to Cheng and Sybesma (1999), a bare NP in Chinese can denote a definite, indefinite or generic interpretations, as sentences from (2a) to (2c) demonstrate, respectively.. (2) a. Hufei he-wan-le Hufei drink-finish-PFV ‘Hufei finished the soup’. tang. soup. qu le. b. Hufei mai shu Hufei buy book go PFV ‘Hufei went to buy a book/ books.’ c. Wo xihuan gou. I like dog ‘I like dogs.’ (Cheng & Sybesma 1999:510). Consequently, sentences in (2) seem to give a clue to why Chinese L2 learners lacking an English article system may produce the reported errors in (1). They may not think that English articles are essential to the count and mass properties of NPs and nor do they think that the choice of English articles is highly discourse-related, involving the perspectives between the interlocutors (cf. Givon 1978). Nonetheless, with the aforementioned accounts, is it appropriate to hypothesize that the observed errors made by Chinese speakers are solely due to the lack of the article system in Chinese, or are there any other factors that affect the acquisition of English articles? Is there a difference between the unconscious knowledge attained by Chinese learners who lack an English article system and that possessed by the English native speaker? These are the main questions the present study aims to investigate. The articles a and the cross-linguistically can encode different semantic features like referentiality, specificity and definiteness. These three features are highly 2.

(12) discourse-related, involving whether a speaker intends to refer to an entity existed in the universe of discourse. It may refer to the identity of a reference which carries some noteworthy features located by the interlocutors (i.e., identifiability). The first term ‘referentiality’ in the present study has been applied to a situation where “the speaker’s intent to ‘refer to’ or ‘mean’ a nominal expression to have non-empty references- i.e., to ‘exist’ – within a particular universe of discourse” (Givon 1978:293). By definition, John in (3a) is considered referential, while those in (3b) is non-referential.. (3) a. John, who has a Ph.D. in astrophysics, cheated on the exam. (Fodor & Sag 1982:361) b. Those who can, do. Those who can’t teach. (Chen 2009:1658). In (3a), the existence of John has been presupposed with a non-empty reference. That is why the non-restrictive relative clause who has a Ph.D. can attach felicitously to the proper noun John, functioning to add further elaboration about the property possessed by John. On the contrary, those in (3b) has an empty reference because the speaker does not commit himself to the existence of any specific individuals fitting the description. Hence, those in (3b) is taken as non-referential. The next term comes to ‘specificity.’ While a variety of definitions of specificity have been suggested, the current study adopts the definition suggested by Ionin et al. (2004:5), who consider it as “the speaker intends to refer to a unique individual in the set denoted by the NP and consider this individual to possess some noteworthy property.” To illustrate, sentences in (4) are cases in point.. 3.

(13) (4) a. Every day the chef comes to cook the dinner for us. b. Every day a chef comes to cook the dinner for us. (Chen 2004:1134). A specific NP, like the chef in (4a), denotes that its identity is only known to the speaker and is considered bearing some unique property that can differentiate itself from others in a set; otherwise, it is non-specific as in (4b). However, from the ‘definiteness’ point of view, whenever a speaker assumes a hearer is in a position to access and identify a conceptual representation of the intended referent, it is definiteness (cf. Lyons 1999). That is to say, the criteria of definiteness of an NP in English is met when the identity of the NP is mutually held by the speaker and the hearer. In English, definiteness can be manifested in various ways, one of which is via articles. The article the encodes the property of definiteness, while the indefinite article a encodes the property of indefiniteness. Accordingly, the specific NP the chef in (4a) is also a definite NP, suggesting the identity of the chef is both known to the speaker and the hearer, while the non-specific NP a chef in (4b) is an indefinite NP, indicating the hearer has no prior knowledge to which chef the speaker refers. Contrary to the argument from Lardiere (1998, 2000) that the acquisition of morphological markers may subject to L1-transfer, Larsen-Freeman & Long (1991) and Ellis (1996) point out that transfer of parameter setting exists for some but not all domains of L2 acquisition. For instance, Flynn et a1. (1991) indicate that L2-learners do not transfer lexical specifications of control verbs like promise and tell from their L1 onto their L2, even in cases where L1 and L2 match. As suggested by White (2008), speakers of inflectionally rich languages also initially omit verbal inflection in their L2.. 4.

(14) Furthermore, it is generally claimed that linguistically unmarked1 features of L1 tend to transfer, while marked ones will not (cf. Eckman 1977, Rutherford 1984 among others). Therefore, the learning difficulty generally results from L1-L2 differences involving greater L2 markedness, not from differences involving less L2 markedness. In conclusion, whether L1-transfer can affect the article acquisition by L2 learners still requires further theoretical and empirical investigation.. 1.2 Research Questions The present study aims to address the following research questions: 1) What is the difficulty order of the types of English articles and their possible causes? 2) What is the difficulty order of the features underlying English articles for Chinese EFL learners? 3) What influence do task types have on the acquisition of English articles by L1-Chinese learners of English? 4) Is L2 proficiency a factor influencing Chinese EFL learners’ acquisition of English articles?. 1.3 Significance of the Study English articles have sparked great interests in different domains of linguistics such as syntax (Leung 2001), semantics (Ionin 2003) and phonology (Goad & White 2008) in L2 acquisition studies. However, literature (Huebner 1985, Thomas 1989, Lu 2001, Ionin et al. 2004, 2008, Hawkins et al. 2006) has offered inconsistent findings of the performance of L2 learners from various L1 background, and showed limitations in. 1. The notion of markedness is defined in terms of complexity, relative infrequency of use, or those deviate from the canonical forms of the language (Ellis 1996). 5.

(15) their research methods. In order to supplement previous findings and to increase the legitimacy of the results, the present study included a consistent design and a larger subject pool. Both a comprehension task and a production task were conducted as well. Furthermore, all the marked functions of English articles that were either ignored in the test materials or misplaced in the categorization of the previous works were re-categorized and examined.. 1.4 Organization of the Thesis The organization of the study is as follows. Chapter Two reviews and discusses the theoretical studies that proposed different frameworks for the English articles and four empirical studies conducted based on the introduced theoretical frameworks. Chapter Three presents a revised classification of English articles incorporating the previous frameworks. Chapter Four introduces the research design of the study, including the information of the subjects, materials, experimental procedures and scoring. Chapter Five reports and discusses the results of the present study. Finally, a conclusion and suggestions for future research are given in Chapter Six.. 6.

(16) CHAPTER TWO LITERATURE REVIEW. In this chapter, theoretical and empirical reviews along with a new classification of the English article system are discussed. Section 2.1 discusses four theoretical studies on the features and the classifications of English articles, and Section 2.2 reviews the empirical studies conducted under different frameworks. Section 2.3 presents a modified new classification of English articles examined in the present study and explores the possible L1-transfer effect. Finally, a summary of this chapter is presented in Section 2.4.. 2.1 Theoretical Studies of Articles in English This section summarizes four theoretical studies on the frameworks of English articles. The analytic perspectives in the following studies are both semantic and pragmatic oriented although each framework is defined according to different criteria. Hawkins (1978) and Lyons (1999) generalize the underlying features of English articles by identifying which NP contexts an article can be placed in. Huebner (1983) classifies NP contexts based on the referentiality and identifiability of references, and analyzes the articles in the classified category. Ionin et al.’s (2004) study is the only research that incorporates discourse contexts within the schema.. 2.1.1 Hawkins (1978) Hawkins (1978) classifies English articles the and a based on three semantic/pragmatic concepts: shared sets, inclusiveness and exclusiveness. According to Hawkins, the referent of an NP marked by the definite article the can be referred and located in sets shared by the interlocutors, and the shared sets can be constructed via 7.

(17) several sentence types like anaphoric use as in (1), immediate situation use as in (2), larger situation use as in (3), general situation use as in (4) and associative anaphoric use as in (5).. (1) (2) (3) (4) (5). Fred brought me a bucket, but the bucket had a hole in it. (Hawkins 1978:110) Pass me the bucket, please. (Hawkins 1978:111) The Prime Minister has just resigned. (Hawkins 1978:116) The moon was very bright last night. (Lyons 1999:3) I had to get a taxi from the station. On the way the driver told me there was a bus strike. (Lyons 1999:3). Nonetheless, the shared set seem not exclusive to the definite article the. For a sentence like A member of parliament has just died (Hawkins 1978:173), the referent of a member who has just died is marked by the indefinite article a and can be referred to and located in the background knowledge (i.e., the parliament of the speaker and the hearer’s country) shared by the interlocutors. Hence, Hawkins further suggests that the definite article the must be different form the indefinite one a by carrying some other distinct properties. Hawkins thus draws attention to the fundamental difference between the and a via two features: ‘inclusiveness’ and ‘exclusiveness.’ He considers the function of the is to introduce a referent to the hearer and to instruct the hearer to locate the referent in some shared sets. More importantly, the refers to the totality of objects or mass (i.e., inclusiveness) within the set satisfying the referring expression. As for the indefinite article a, Hawkins argues that a-marked NPs are ambiguous, rendering a specific and a non-specific reading as in (6a) and (6b), respectively.. (6) a. Bill lost a finger in the war. b. Pass me a bucket.. (Hawkins 1978:180) (Lyons 1999:261) 8.

(18) The indefinite article can only be used to refer to objects existing in some shared set and render a specific reading only if the indefinite description can be understood as referring to a proper subset of the total number of objects in a shared set (Hawkins 1978:204). If a hearer cannot construe the reference for not-all objects in the shared set, he is then forced not to locate the referents there at all, yielding the indefinite article to serve as an existential quantifier generating the NPs with non-specific readings. To summarize, Hawkins (1978) claims that it is ‘inclusiveness’ (i.e., reference to all) versus ‘exclusiveness’ (i.e., reference to not-all) that distinguishes the definite article from the indefinite one, and that the hearer’s cooperation plays an important role in locating the referent in the discourse. However, Lyons (1999) has pointed out that ‘inclusiveness’ is insufficient to explain the sentences in (7).. (7) a. In a room with three doors, one of which is closed Close the door, please. b. Meet me at the horse-trough tonight.. (Lyons 1999:14) (Lyons 1999:263). ‘Inclusiveness’ cannot account for the NP the door in (7a), for what the speaker refers to is the door that closed rather than all the doors in the immediate situation. In (7b), although the hearer appears not knowing about where the horse-trough is, the definite article the is still applied.. 2.1.2 Huebner (1983) Huebner (1983) characterizes the English article system based on the dichotomy of two plus and minus semantic features of the NP context: [+/- Specific Referent (SR)] and [+/- Assumed known to the Hearer (HK)], originated from Bickerton (1975).1. 1. Bickerton (1975) proposed a semantic wheel for NP references dichotomized by the semantic features: [+/-SR] and [+/-HK]. The feature [+/-SR] refers to ‘specific referent’, concerning the 9.

(19) Huebner suggests that these two semantic features yield four cross-classified environments based on the lexical properties (i.e., singular or plural, count or mass) of the pre-noun contexts as (i) [+SR, -HK] Specific indefinite: a, Ø; (ii) [+SR, +HK] Specific definite: the; (iii) [-SR, +HK] Generic: a, the and Ø; (iv) [-SR, -HK] Indefinite generic: a, Ø. For the NPs first introduced into the discourse in English that are known to the speaker but not to the hearer, they are specific indefinites and are included in the category of [+SR, -HK] as in (8).. (8) Bill sold a car.. Once an NP has entered the discourse and reappeared in the previously mentioned environment, it commonly becomes a referential definite like the house in (9a). An NP may also be qualified as a referential definite and classified into the [+SR, +HK] if it meets some other criteria as the rest of the examples in (9) demonstrate.. existence of an NP in the world (i.e., whether the noun an article depends on is used referentially), and [+/-HK] ‘assumed known to the hearer’ (i.e., whether the noun which an article is depended on is identifiable by the listener). 10.

(20) (9) a. He bought a house, but this is not the house he bought.2 b. Ashley saw that the angry old lion had a gunshot wound in the neck. (James & Jing 2008:486) c. Tennis is the best sport for people of all ages. (Folse, Solomon & Smith-Palinkas 2004:17) d. Is the moon bright tonight? (James & Jing 2008:484) e. The captain of the women’s basketball team came to lunch with us today. (James & Jing 2008:467). In (9b), the referent of the neck can be specified by entailment from the lion since the neck is a body part of lion. In (9c), the referent of the best sport is by definition specific based on their superlative structures. The moon in (9d) and the captain in (9e) are also referential definites because their referents can either be drawn from the general knowledge or the specific knowledge shared by the speaker and the hearer. NPs classified as [-SR, +HK] are generics and can be marked with either a, the and Ø as exemplified in (10).. (10) a. A corkscrew is a gadget for getting corks out of bottles. (Hewings 2005:90) b. The American love to drink pop. (James & Jing 2008:476) c. Tigers are native to India. (Folse, Solomon & Smith-Palinkas 2004:19). The referents of generics are random and arbitrary and non-specific. They are simply a particular member belonging to the total class that neither the speaker nor the hearer has in mind. However, under the context of general knowledge, the referents in (10) are all known to the hearer and thus can be classified as [+HK]. The last type of NPs in (11) are non-referential, belonging to the category [-SR, -HK]. The NPs of this type are used to name a class that another NP is asserted to as in. 2. Retrieved from http://www.learnersdictionary.com/definition/the 11.

(21) (11a), or to refer to an unspecified member of a class as in (11b). This type differs from the generics in that the hearer in the discourse is unable to identify the referent of the NP from either the shared back group knowledge or the universal knowledge.. (11) a. I think a good receipt will help you keep in shape. b. I guess I should buy a new car.. To recapitulate, Huebner (1983) classifies English articles according to NP environments. The criteria used to divide the NP environments are based on the existence of the referent in the world. Nevertheless, Huebner does not clearly specify how the term a specific referent should be defined (cf. Ionin 2003), nor does he explains to what degree an NP should be classified as known to the hearer. To illustrate, according to Huebner’s definition, the NP the dog in (12) should be categorized as [+SR, -HK], yielding two article choices a (an) or Ø since only the speaker rather than the hearer knows which dog he refers to.. (12) Beware of the dog.. (Hawkins 1978:112). However, the felicitous article use in (12) is the. Last, since the features proposed in Huebner (1983) can only categorize the environment of the noun phrases rather than the articles themselves in English, the underlying properties of the definite and indefinite articles remains unclear.. 2.1.3 Lyons (1999) According to Lyons (1999), the differences between the definite and indefinite articles are made by their underlying properties and functions.. 12.

(22) Concerning the use of the definite article, Lyons proposes four semantic features that may each capture at least some underlying properties of the English article the. First, the definite article the under a familiarity account signals that the entity denoted by the noun phrase is familiar to both the speaker and the hearer as illustrated in (13).. (13) a. Just give the shelf a quick wipe, will you, before I put this vase on it. b. I saw a woman wearing sunglasses walking into the church. The woman seemed very upset. c. The moon was very bright last night. (Lyons 1999:3). The mutual familiarity of a certain entity can be achieved via the situational usage (i.e., physical situation) as in (13a), anaphoric usages (i.e., linguistic situation) as in (13b) or association from larger knowledge as in (13c). Nonetheless, familiarity cannot account for the bold NPs in (14), for their referents are apparently unfamiliar to the hearer.. (14) a. They’ve just got in from New York. The plane was five hours late. b. The president of Ghana is visiting tomorrow. c. The bloke Ann went out with last night phoned a minute ago. (Lyons 1999:3). Hence, Lyons proposes that the possessed the second feature ‘identifiability,’ which states that the use of the definite article is to direct the hearer to the referent of an NP by signaling that he is in a position to identify it. Thus, it is reasonable to say that under “the familiarity account, the tells the hearer that he knows while under the identifiability account, the tells him which he can work out (Lyons 1999: 6).”. 13.

(23) Although most of the NPs marked by the can be explained by the feature ‘familiarity’ and ‘identifiability,’ the third semantic feature ‘uniqueness’ is needed for sentences in (15).. (15) a. A Nurse is entering an operation theatre. I wonder who the anesthetist is today. (Lyons 1999:7) b. Beware of the dog. (Lyons 1999:9) c. The winner of this competition will get a week in the Bahamas for two. (Lyons 1999:9). It appears that the speaker cannot identify the referents of the definite NPs in (15) nor does he necessarily expect the hearer to be able to do so. Both interlocutors know there is or will be such an individual, but could not identify it. The uniqueness of the definite article is usually related to a particular context, but it can be absolute. Its criteria are particularly attractive in cases where the referent is hypothetical, potential, or in the future as the NPs in (15) illustrate. The uniqueness account also provides a convincing explanation for why the definite article the is compatible with superlatives first, only, etc. The only weakness of ‘uniqueness’ is its incompatibility with the NPs that are plural as in a sentence like The teacher gave out the prizes, or mass as in The tea is hot, because the referents of which are not the only one in the shared set and logically contradict the concept of uniqueness. Hence, the fourth semantic feature ‘inclusiveness,’ which can be defined in a way that the referent is to the totality of objects or mass in the context, satisfying the description (Hawkins 1978). As for the indefinite article a, Lyons (1999) argues that an indefinite NP may be used to denote a particular entity as in (16a), or an arbitrary member of the class described by the NP as in (16b).. 14.

(24) (16) a. I haven’t started the class yet; I’m missing a student – Mary’s always late. b. I haven’t started the class yet; I’m missing a student – there should be fifteen, and I only count fourteen. (Lyons 1999:170). A non-specific NP a student in (16b) describes but not refers to, while a specific NP a student in (16a) is a referring expression. Lyons (1999) indicates that for non-specific indefinite NPs, a serves as a logical operator and a narrow scope is applied, while those referring specific NPs marked by an indefinite a have wide scopes. To summarize, Lyons (1999) proposes four semantic features: ‘familiarity,’ ‘identifiability,’ ‘uniqueness’ and ‘inclusiveness,’ which represent part of the underlying properties of the definite article the. Nonetheless, none of the aforementioned features can solely encompass all the usages of the definite article, and each feature has its drawbacks. What’s more, Lyons dwells very little on the use of the indefinite article a, only suggesting that the reading of a can be specific or non-specific. As a result, there seem to be two distinguished systems used to capture English articles a and the, and both of which do not render a comprehensive account.. 2.1.4 Ionin et al. (2004) Ionin, Ko and Wexler (2004) propose that articles cross-linguistically can encode different semantic features and that they should be governed under two semantic parameter settings: the definiteness setting and the specificity setting. The [+Definite] setting and [+Specific] setting are both discourse oriented; that is, they are related to the mind of the speaker and the hearer in the discourse. An informal definition of the two article settings is that “If a Determiner Phrase (DP) of the form [D NP] is [+Definite], then the speaker and hearer presuppose the existence of a unique individual in the set denoted by the NP; if being [+Specific], then the speaker intends to refer to a unique 15.

(25) individual in the set denoted by the NP and considers this individual to possess some noteworthy property.” (Ionin et al. 2004:5). The examples of each article setting are demonstrated below. Sentence (17a) represents [+Definite, +Specific], (17b) [+Definite, -Specific], (17c) [-Definite, +Specific] and (17d) [-Definite, -Specific].. (17) a. We are trying to find the murderer of Miss Andrews – his name is Roger Williams, and he is a well-known criminal. b. We are trying to find the murderer of Mr. Peterson – but we still don’t know who he is. c. I would like to sell you a beautiful silver necklace. It is very valuable – it has been in my family for 100 years! d. I’d like to get to know my classmates. I am planning to find a new good friend! I don’t like being all alone. (Ionin et al. 2004:27-29). According to Ionin et al. (2004), the use of the indefinite article a in the discourse informs the hearer that a new entity is introduced. In Standard English, article choices depend on the notion of definiteness, not specificity. Hence, in [+Definite] contexts like (17a, b), where a shared knowledge is possessed by both the speaker and the hearer, the definite article the is used, while in [-Definite] contexts like (17c, d), in which no such shared knowledge exists between the speaker and the hearer, indefinite articles a (an) or some are used. Furthermore, there is no marker for the [+Specific] feature in the English article system, but the feature does exist in the demonstrative this, applicable to attest the specificity as in (18).. 16.

(26) (18) a. Peter intends to marry a/this merchant banker – even though he doesn’t get on at all with her. b. Peter intends to marry a/?? this merchant banker –even though he hasn’t met one yet. (Lyons 1999:176). The infelicitous use of the demonstrative this in (18b) points out the fact that the speaker has no presupposed knowledge about who the merchant banker is, and thus the merchant banker is unspecific. Accordingly, since the speaker in (18a) believes that there is no noteworthy property possessed by the NP a merchant banker in (18a), the reference thus satisfies the feature [-Definite, + Specific]. However, the reference of the counterpart NP in (18b) satisfies [-Definite, -Specific], receiving a quantificational interpretation that the speaker merely asserts one merchant banker out of a group of merchant bankers that is available (cf. Fodor & Sag, 1982). Again, the felicitous use of the indefinite article a in both sentences (18a, b) denotes that the feature of specificity does not affect English article choice. No matter an NP is [+Specific] or [-Specific], as long as it is [-Definite] the indefinite article a is used. To summarize, according to Ionin et al. (2004), the two semantic features: [+Definite] and [+Specific] play crucial roles for L2 learners to choose a particular article under a particular NP setting. Unfortunately, the proposed features are subjected to the following issues. First, the two features are insufficient to explain the sentence in (15a) I wonder who the anesthetist is today. Although the speaker and the hearer do not share mutual knowledge of the reference the anesthetist, the definite the is used. Second, most of the [+Definite] test samples in Ionin et al. (2004) are presented in a fixed pattern article + NP + of NP, disregarding other possible usages of the definite article the. It seems that only by following such sentence pattern can the two features apply. As. 17.

(27) a result, a better account is needed about whether the feature [+Definite] alone can capture all the definite usages in English.. 2.1.5 Summary A compared table of the theoretical reviews is presented in Table 2-1.. Table 2-1 A Compared Article Classification from the Previous Empirical Studies Work. Main proposals. Limitations. Hawkins (1978). 1. The classified NP contexts: anaphoric use, immediate situation use, large situation use & associative use. 2. Semantic feature of articles: The: inclusiveness A: exclusiveness. 1. Fail to explain some sentences 2. Fail to explain why the cooperative effort from the hearer does not involve in some sentences.. Huebner (1983). 1. The classified NP contexts: [+/-SR] & [+/-HK] 2. The feature of the articles The: NA A: NA. 1. Lack crucial feature to distinguish the & a 2. Fail to explain some sentences. Lyons (1999). 1. Overlapping features 1. The classified NP contexts: 2. No consistent system for the NA function of the & a 2. The feature of the articles: The: familiarity, identifiability, uniqueness & inclusiveness. A: specificity or referential. 1. The classified NP contexts: [+definite] & [+specific] 2. The feature of the articles: Ionin et al. The [+definite, +specific] & (2004) [+definite, +specific] A [-definite, +specific] & [-definite, -specific]. 18. Fail to explain some sentences.

(28) In order to investigate the function and the feature of English articles, two main approaches are proposed in the previous studies. One is to generalize the underlying features of English articles based on the environment they are located (Hawkins 1978, Lyons 1999), and the other is to categorize the NP contexts and analyze the articles that can be placed under them directly (Huebner 1983, Ionin et al. 2004). Notice that only the classifying features of the NP settings in Ionin et al. (2004) take the context of the discourse into consideration. What’s more, even though each of the previous works defines the feature of NP environments with different terms, it appears that the function of English articles relate closely to the discourse linguistically or non-linguistically and involves the participation of the speaker and hearer.. 2.2 Previous Empirical Studies of English Articles In this section, four empirical studies concerning English articles are reviewed. Thomas (1989), Robertson (2000) and Lu (2001) focused on Chinese learners’ performances on the acquisition process of English articles, while Ionin et al. (2008) aimed to investigate the trigger of the observed errors made by L2 learners.. 2.2.1 Thomas (1989) Thomas (1989) attempted to investigate four research questions in her study. First, would L2 learners undergo the same four-stage article acquisition sequences as L1 learners proposed by Czike (1986).3 Second, would the-flooding found in the previous studies (cf. Hakuta 1976, Huebner 1985, Parrish 1987 and Master 1987) be verified?. 3. Czike (1986) proposes a four-stage sequence of the article acquisition for L1 learners. Stage 1: Children mark all referential nouns as [+SR] regardless of the feature [+/- HK] with either a or the. And no article use is found in either the environments of non-referential nouns or generics. Stage 2: Children employ the in [+SR] contexts and a in [-SR] contexts. Here a high rate of errors of the first mentioned nouns jointing with the definite article the might be found. Stage 3: Children reintroduce a into [+SR] environment. Stage 4: Children have fully acquired adults’ system of noun classification. 19.

(29) Third, which NP environment may induce most error of English articles by L2 learners? Fourth, what is the crucial feature associating with English articles a and the. Thomas (1989) adopted Huebner’s (1983) framework by dichotomizing English articles on the basis of two features [+/-SR] (i.e., Specific Referent) and [+/-HK] (i.e., Hearer Known), classifying NPs into four contexts: [+SR, +HK], [+SR, -HK], [-SR, +HK] and [-SR, -HK]. Thirty L2 adult learners representing nine languages (a large proportion of which was consist of Japanese and Chinese) were recruited with their age ranged from 24 to 46. Their average length of residence in English-speaking countries was 10 months. These thirty participants were later cross-sorted into three different proficiency levels high, mid and low, and two groups with the article system [+Art] or without the article system [-Art] in their L1. A picture-description task was employed in this study. The materials used in the picture-description task were eight pairs of picture cards, one as the target picture and one as a distracter. These cards represented varying scenes from simple to complex and from static to dynamic. The subjects were tested in pairs, seated back to back, with one member playing the speaker and the other the listener. The speaker needed to describe the picture he/she saw on the card to the listener, and the listener was later required to distinguish the correct picture card from the distracter. The results of the study showed that both the [+Art] and [-Art] groups in general were more accurate at using the English definite article the than using the indefinite one a. Compared with the [+Art] group, the [-Art] group produced more null article in both a and the contexts. Furthermore, comparing to the frequency count of the three articles used in [+SR, -HK] contexts, the use of the was relatively higher than the other articles although the frequency did not reach a significant difference (p < .025, one-tailed). The effect of proficiency levels appeared to be less marked, yet at each level the participants 20.

(30) performed much better with the than with a. Contrary to the previous studies, Thomas indicated that the task could not elicit more article usages under the environment of generics [-SR, +HK] as the previous studies (Hueber 1985, Master 1987) did, nor did she observe the flooding effect of the (the frequency rate of the marking was below 50%). Hence this study likewise was not able to draw a solid conclusion by associating the to either the feature of [+SR] or [+HK].. 2.2.2 Robertson (2000) Robertson (2000) investigated the article acquisition of English articles by adult Chinese learners following Hawkins’s (1978) framework. The study aimed not at what types of errors were be made by L2 speakers, but at why the mistakes produced by Chinese L2 speakers were unsystematic. The inconsistent errors in the choice of English articles according to Robertson (2000) appeared to be typical evidence for the Optionality Hypothesis4 and they were triggered by Chinese learners’ difficulty in correctly mapping the semantic and pragmatic features of definiteness and referentiality to the syntactic and lexical forms of the target language. Eighteen Chinese origin subjects (China, Taiwan, Malaysia and Singapore) who were postgraduate students at Leicester University were recruited and were divided into nine pairs. Within each pair, there was one subject serving as a speaker and the other a hearer. Speakers would receive a sheet with a diagram drawn on it in two colors red and blue. The speaker needed to describe the diagram to the hearer seated on the other side of the table. The hearer who was not able to see the sheet was asked to recreate the diagram based on the description made by the speaker. Four diagrams of varying complexity were used in the experiment to help the experimenter to analyze the effect 4. The term ‘optionality’ here, based on Robertson (2000), refers to the use of two or more competing forms with the same meaning (i.e., the variable use and non-use of a particular feature with no difference in meaning between the two forms). 21.

(31) of task difficulty on the accuracy. The coding criteria in the study did not exclude reformulations and repetitions from the frequency count out of the echo contexts for the reason that the author believed it may render a much reliable result to the inter-language of L2 learners. Salient findings were found when two variables: linguistic distinction (i.e., definite vs. indefinite) and pragmatic distinction (i.e., echo vs. non-echo contexts) were cross-examined. First, the finding conformed to the previous studies (Huebner 1985, Parrish 1987), showing that accuracy rates were higher in the definite contexts than in the indefinite contexts. Second, accuracy in echo contexts was markedly lower than it was in non-echo contexts. Third, the omission of English articles was more frequent in echo contexts. Finally, the lowest accuracy was found in the immediate situation use followed by the associative use. To account for the omission found in the data, Robertson claimed that the omission tendency was known as one of the characteristics of discourse-oriented languages like Chinese. According to Robertson (2000), as long as the definite article of Chinese was pragmatically redundant and can be recovered from the context, it was more likely to be omitted. In addition, the article might be dropped if it was within the scope of the determiner of an immediately preceding and co-referential NP. The familiar socio-pragmatic principle of accommodation in interaction might also play a part for the tendency of omission since the second speaker inclined to follow the article use of the first interlocutor by omitting the article when the first interlocutor did not use one.. 22.

(32) 2.2.3 Lu (2001) Lu (2001) conducted an empirical study of the acquisition order and process of English articles by Chinese learners. Following Bickerton (1975), Lu intended to examine in which semantic context the overgeneralization of English articles a, the, and Ø might appear and cause the greatest learning difficulty. The participants of the study comprised a total of fifty-five (32 graduates and 6 undergraduates from two different schools) Mandarin Chinese speakers aged from 17 to 37. They were divided into three groups: advanced, upper-intermediate, and lower-intermediate, based on their TOEFL scores. The material in use was a multiple-choice cloze task composed of 58 items in one discrete sentence part and one descriptive paragraph part. The participants were asked to circle the best answer among a, the, and Ø. The measurement employed for the data were Supplied in Obligatory Contexts (SOC), Target-Like Use (TLU), and Used in Obligatory Contexts (UOC). The means of SOC and TLU for each group were compared to identify the acquisition order in terms of article accuracy, while the means of the compared UOC were to investigate over- or underuse of the articles. The findings in Lu (2001) were as follows. First, the acquisition order across the groups were identified as the = a > Ø in SOC, and the > a > Ø in TLU. However, the sequence in TLU was argued to be more reliable because UOC for a went up (109.52% 122.63%. 136%) with a decrease of proficiency, indicating that there was an. inflation for a in SOC. What’s more, the one-way ANOVA for the group effect on the overall TLU means of the three articles combined and the article effects on the overall TLU means of the three groups combined were both found to be significant (p < .0125 and p < .0125, respectively). Second, according to UOC 79.05%. 93.61%. 103.06% from inter-mediate to advanced level, the use of Ø increased with proficiency, while UOC for the remained very stable (94.25% - 97.22%). The results supported 23.

(33) Chaudron and Parker’s (1990) hypothesis of U-shaped behavior for a and Ø, and indicated the completed the U-shaped behavior prior to a. Accordingly, Lu found that lower-intermediate learners had more difficulty in a or Ø than the, and they tended to misuse a for Ø. Third, concerning the semantic context for English articles, [+SR, -HK] was recognized to be the key feature that caused the-flooding, and it should be [+SR] rather than [+HK] that was associated with the. Furthermore, Lu indicated that Chinese learners had difficulties distinguishing [+/-HK] that caused the misuse of the for a or Ø and [+/-Countability] that led to the misuse of a for Ø, or Ø for a.. 2.2.4 Ionin et al. (2008) Ionin et al. (2008) aimed to examine three feasible sources of linguistic roles: L2 input, L1 transfer and innate linguistic knowledge that might trigger the fluctuation effect in the process of English article acquisition. Three groups were included in the study. A controlled English adult group (6 people) and two experimental groups that were divided into a [+Art] L1 Russian group (23 people) and a [-Art] L1 Spanish group (24 people). The participants were required to take a proficiency task and an article elicitation test. In the elicitation task, there were 60 short dialogues involving 36 target items and 24 fillers to elicit a particular target word under the time limit of 45 minutes. Among the 36 target items, there were three categories of six items each aiming at eliciting the, and three categories of six items each targeting at eliciting a. Of each category, half of the targets included scope interactions with an intentional modal operator and the other half did not. The findings of Ionin et al. (2008) echoed the results of Ionin et al. (2003, 2004). Overuse of the was found in the specific indefinite condition and a in the non-specific definite condition. Performance of the participants was improved when the feature of definiteness was in accordance with the feature specificity or with the elevation of their 24.

(34) proficiency. When comparing the Russian group with the Spanish group, it revealed that the Spanish learners were more native-like because they achieved a higher accuracy of the in [+Definite, -Specific] setting and a in [-Definite, +Specific] setting; however, their performance of the in [+Definite, +Specific] was unexpectedly lower than the Russian learners. According to Ionin et al., such lower performances of the Spanish group was not caused by the infelicitous choice of a but by omission due to L1 transfer because omission is commonly adopted in the same pattern in Spanish. Overall, the Spanish learners were highly accurate in the article use in most of the cases and they did not manifest the fluctuation effect. Nonetheless, when cross-examined the statistical test of the variables definiteness, specificity and the proficiency level, the two groups demonstrate different results. On the one hand, definiteness and specificity appeared to have highly significant effects on the article uses while no effect of proficiency level could be found in the Russian group. Furthermore, definiteness and specificity interacted significantly when use of the was measured and marginally when use of a was measured, indicating that specificity effect was stronger with indefinites than with definites. On the other hand, there was no significant effect of specificity but only definiteness and proficiency level in the Spanish group.. 2.2.5 Summary Table 2-2 summarizes the major finding and limitations of the empirical studies reviewed in this chapter.. 25.

(35) Table 2-2 Major Findings and Limitations of the Previous Empirical Studies Work. Major findings. Limitations. Thomas (1989). 1. Overgeneralizing the in first-mention [+SR, -HK] contexts. 2. Overgeneralizing Ø in L1 [-Art] group. 3. Accurate use of a was significantly delayed. 4. No flooding of the was found.. 1. Subjects: a. Unbalanced grouping b. No control group 2. Design: a. Unable to elicit the crucial evidence for the use of articles in generic contexts b. Only one task. Robertson (2000). 1. Article omission was not systematic. 2. Accuracy: a. Definite contexts > Indefinite contexts b. Non-echo contexts > Echo contexts. c. Ø appeared more often in echo contexts.. 1. Subjects: a. Small subject pool b. No control group c. No classification of proficiency 2. Design: Only one task. 1. Tendency: a. Ø goes through a flooding-then-trickling process b. A decreases with. 1. Subjects: No control group 2. Design: a. Lack task comparison b. Task materials were not comprehensive enough.. Lu (2001). Ionin et al. (2008). proficiency c. Misuse a & Ø 2. Learning sequence: the > a > Ø 3. Difficulty in distinguishing [+/-HK] & [+/-Countability] 1. L2 learners’ performance 1. cannot be a result of domain-general statistical learning. 2. L1 Russian speakers are under 2. the influence of specificity and definiteness. 3. L1 Spanish speakers are under 26. Subjects: a. Small subject pool b. Unbalanced grouping and background Design: a. Only one task b. Ambiguous test items.

(36) the influence of definiteness and proficiency.. General observation made by all the previous studies is summarized as follows. First, L2 learners regardless of their L1 having or lacking article systems tended to master the definite article the faster and more accurately than the indefinite article a in general contexts. Second, overgeneralization of the definite article the seemed to appear in most of the cases although the degree of the-overuse was inconsistent from research to research. Third, many studies showed that L2 learners whose L1 with an article system tended to outperform those L2 learners without an article system in their L1. Some of the common errors made by L1 [-Art] learners were like article dropping, fluctuation of article use, overuse of the and so forth. However, whether the errors were made systematically or not was still under a heated debate. Last, it was suggested by some researchers (Paradis & Zdorenko 2008, Ionin 2003, Ionin et al. 2004, 2008) that L1 transfer, L2 input and Universal Grammar might be the factors that caused the misuse and the fluctuation of English articles. In terms of the limitations of the previous studies, the subject pool and the way of subject grouping in all of the previous researches appear to be quite small, inconsistent and unbalanced. There was a wide discrepancy in the subjects’ background such as their age, the spam of their language exposure and their social statuses in most of the studies. Three out of the four studies did not have a control group for the results to compare to. As for the design and the material, three out of the four studies (Thomas 1989, Robertson 2000, Lu 2001) did not include both a comprehension and production task. Although the elicited article use from a picture-description task (Thomas 1989, Robertson 2000) may be more natural than a force-choice task (Lu 2001, Ionin et al. 2008), it is relatively hard to elicit all types of articles and to construct appropriate 27.

(37) contexts from pictures. A force-choice task may also encounter similar dilemma by constructing ambiguous contexts that are difficult to provide a clear-cut between indefiniteness and unspecificity. Without a systematic grouping and task materials, the acquisition sequence of English articles may not be reliable.. 2.3 Summary of Chapter Two In this chapter, theoretical and empirical studies concerning the English articles have been reviewed. Section 2.1 presents four theoretical reviews of the English article system based on different classification frameworks. Section 2.2 demonstrates the empirical research based on various theoretical approaches and discusses the general tendency in article usage made by L2 learners at different proficiency levels whose L1 lack an article system. The revised framework of English article system is introduced in the next chapter.. 28.

(38) CHAPTER THREE A REVISED CLASSIFICATION OF ENGLISH ARTICLES. In this chapter, a revised classification of English articles that integrates the analyses of the previous studies (Hawkins 1978, Huebner 1983, Lyons 1999, Ionin et al. 2004) is presented.. 3.1 Rational for the Revised Classification The revised classification was based on Huebner’s (1983) framework with modified criteria in order to incorporate some uncovered usages proposed by Hawkins and Lyons. According to Huebner’s classification, the definite article the can only be used in the [+SR, +HK] (referential definite) and [-SR, +HK] (generic) contexts, and the indefinite article a and the null article are the only two possible articles that can appear in the [+SR, -HK] context. However, such classification fails to explain why the indefinite-article-marked NP an angle (i.e., a noun with an indefinite article) in There’s an angle standing before me when used sarcastically can appear in the [+SR, +HK] context and delivers a definite reading, why the definite-article-marked NP the horse-trough (i.e., a noun with a definite article) in Meet me at the horse-trough tonight that fits in the [+SR, -HK] context is excluded in Huebner’s framework, and why a nonspecific NP the doctor in I wonder who the doctor is today that should be classified into the [-SR, -HK] context is marked by the definite article the. To complement the previous taxonomy of the English article system, the author re-classified English articles into three types: Type One the, Type Two a and Type Three Ø based on their forms with each being sub-characterized by their reference. 29.

(39) features: Specific Referent [+/-SR] and Assumed Known to the Hearer [+/-HK] and their referential interpretation into three settings1 as Figure 3-1 shows.. __+ N. [+SR] [+HK] definite readings the, a. [-SR] [-HK]. [-HK]. indefinite reading the, a, Ø. indefinite readings the, a, Ø. Figure 3-1 A Revised Frame for English Articles. The features in use are defined as below (cf. Ionin & Wexler 2003). (19) Specific Referent [+/-SR]: An NP is defined as [+SR] if the speaker intends to refer to a unique individual in the set denoted by the NP and considers this individual to possess some noteworthy property; otherwise it is [-SR] (20) Assumed Known to the Hearer [+/-HK]: If the referent of an NP is known or identifiable from the contextually relevant domain to the hearer, it is classified as [+HK]; otherwise, it is [-HK].. In the present modified frame, the referential functions (i.e., [+/- SR] and [+/- HK]) and the referential interpretation of English articles were incorporated to reanalyze English article types (cf. Fodor & Sag 1982). An NP which may be manifested as a + NP, the + NP, or Ø + NP on the first layer first went through the defined NP criterion [+/-SR] in (19) and then was separated into either [+SR] or [-SR] groups on the second 1. The two features [+/- SR] and [+/- HK] originally can yield four settings. However, since the present study aimed to investigate the core functions of English articles that are investigated under a consistent semantic frame, the generic use (generates from [-SR, +HK] settings) that involves other formal mechanisms (cf. Prasada 2000) was excluded in the present study. 30.

(40) layer. A further division was made based on the criterion [+/-HK] in (20), which brought about four different sub-categories on the third layer. The outcome of the third layer was re-analyzed according to their referential interpretations (i.e., the definite reading and the indefinite reading) according to the perspective of the hearer. A detailed explanation of each English article type is described as follows.. 3.2 Type One: The Definite Article The Type One comprises the NPs marked by the definite article the and is classified into three categories: (i) [+SR, +HK] settings with definite readings; (ii) [+SR, -HK] settings with indefinite reading; (iii) [-SR, -HK] settings with indefinite readings. The bold NPs in (21, 22) belong to the [+SK, +HK] setting with definite readings.. (21) A: Fred brought me a bucket, but the bucket had a hole in it.2 B: Didn’t he check it before handing it to you? (22) Caicai ganggang fasheng guess just.now happen zhe ge this CL. nuhai girl. sheme shi? Wo yudao what incidence I met. you-zhe have-ZHE. yi-to one-CL. hong red. yi-ge one-CL. nuhai, girl. toufa. hair. ‘Guess what? I met a girl and the girl has red hair.’. The target English sentence (21) and its Chinese counterpart (22) are both categorized by Hawkins (1978) and Lyons (1999) as the anaphoric use. The use of the definite article the in (21) and the demonstrative zhe ‘this’ in (22) is obligatory. The omission of either one of them would cause the sentences infelicitous. According to Lyons (1999), although Chinese does not have an article system, the demonstratives in Chinese are gradually functioned as the English definite article in 2. To remain consistency of the samples and the purpose of this research, the target sentences are expanded into conversations. The sentence Fred brought me a bucket, but the bucket had a hole in it is originated from Hawkins (1978:110). 31.

(41) many ways. In Chinese, the demonstrative na ‘that’ or zhe ‘this’ may be adopted according to the proximal of the physical or linguistic context. Zhe ‘this’ is psychologically, linguistically or physically more approximate than na ‘that’ (cf. P. Chen 2004). Therefore, when a referent has just been introduced into the discourse like zhe ge nuhai ‘this girl’ in (22), the demonstrate zhe ‘this’ is preferred. The NPs in (23, 24) are classified under the setting of [+SR, -HK] because the speaker has a specific referent the dangerous dog in mind and intends to refer to it even if the hearer does not know which dog it is.. (23) A: B:. Beware of the dog.3 A boy was bitten last night. Thank you for the warning.. (24) Ni yao xiaoxin (*na-zhe) go. Zhe fujin hen you need beware aathat-CL dog this nearby very ‘Beware of the dog. This area is very dangerous.’. weixian. dangerous. The use of the demonstrative is infelicitous for it changes the original meaning in (23). It creates a scenario that the dog the speaker refers to is right in front of the hearer eye because the use of the demonstrative na ‘that’ in Chinese indicates that the hearer can locate and identify the referent of the reference in the physical or linguistic context. Hence, the demonstratives in Chinese are more similar to that rather than the in English.. (25) A: I wonder who the anesthetist is today. B: I have no idea. Let’s wait and see. (26) Bu zhidao juntan-de (*zhe/ na-wei) yisheng shi not know today-DE aathis/that-CL doctor is ‘I wonder who the doctor is today.’. 3. The sentence Beware of the dog is taken from Hawkins (1978:112). 32. (Lyons 1999:7) na which. wei. person.

(42) The anesthetist in (25) and its Chinese counter part yisheng ‘doctor’ in (26) belong to the setting of [-SR, -HK] with an indefinite reading. Notice that the use of the demonstratives zhe ‘this’ and na ‘that’ here makes the sentence infelicitous.. 3.3 Type Two: The Indefinite Article A Type Two comprises the NPs marked by the indefinite article a and is classified into three categories: (i) [+SR, +HK] settings with definite readings; (ii) [+SR, -HK] settings with indefinite readings; and (iii) [-SR, -HK] settings with indefinite readings. The NPs in sentences (27, 28) belong to the [+SR, +HK] setting, rendering the definite reading.. (27) A: Can you believe there is a princess standing right in front of me? B: Mommy! You’re embarrassing me! (28) (You ge) meinu lai-le. have CL beaty come-PFV ‘Here comes a beauty.’. It is typically suggested in the literatures that when an indefinite article is used in English, it serves as an indefinite NP and delivers an indefinite reading. However, the response in (27) implies not only the speaker does have a specific referent in mind, but also both the speaker and the hearer understand to whom the princess refers. Such uses often appear in discourses that show irony, joke, flattery and so forth (cf. Von Heusinger 2002). Similar use is found in the Chinese counterpart ge meinu ‘a beauty’ in (28) although one should note that identical meaning can be made by the bare NP meinu ‘beauty’ in Chinese as well. The NPs in sentences (29, 30) belong to the [+SR, -HK] setting.. 33.

(43) (29) A: Guess what? I bought a car. B: Really? How much is it?. (Lyons 1999:165). (30) Caicai ganggang fasheng sheme shi? Wo yudao guess just.now happen what incidence I met. yi-ge nuhai. one-CL girl. ‘Guess what? I met a girl.’. Foder and Sag (1982) argue that English indefinite article a may be either quantificational or referential. Here the reference of a car in (29) is referential functioning as a specific reference, for a functions to guide the hearer to recognize who is being referred to. To parallel the English NP a car, a cardinal number yi ‘one’ and a classifier ge are needed, as the bold NP in (30) shows. However, notice that the omission of the classifier yi-ge ‘one’ in Chinese in (30) turns the indefinite reading into a generic one. The NPs in (31, 32) belong to the [-SR, -HK] setting. Heim (1982) suggests that the English indefinite article should embody a quantificational force like a student in (31). The Chinese counterpart (yi) ge tongzi ‘a bucket’ in this NP type may either be a bare NP tongzi ‘bucket’ like in (32) or with a cardinal number one.. (31) A: I’m missing a student – there should be fifteen, and I only count fourteen. B: Who is not here? (32) Qing di (yi) ge tongzi gei please pass (one) CL bucket to ‘Pass me a bucket, please.’. wo. me. 3.4 Type Three: The Null Article Ø Type Three comprises the NPs marked by the null article Ø and is classified into two subtypes: (i) [+SR, -HK] settings with indefinite readings; and (ii) [-SR, -HK] settings with indefinite readings. The NPs in sentences (33, 34) fall into the [+SR, -HK] setting. 34.

(44) (33) A: There are squirrels on the tree. B: Where? I can’t see them. (34) Shu-shang you tree-on have. songshu. squirrel. ‘There are squirrels on the tree.’. The speaker addressing the plural NP squirrels in (33) intends to refer to a unique group (i.e., a group of squirrels) and considers them to possess some noteworthy property known to the speaker. However, from the hearer’s point of view, this referent is unknown and thus should give an indefinite reading. Accordingly, the existential construction there is is used in (33). To introduce the existence of an unknown reference in Chinese, the verb you ‘have’ is needed as the sentence (34) manifests. An unknown number of the NPs following the verb you ‘have’ is presented as a bare form like songshu ‘squirrel’ in (34). The last subtype in the null article use is non-specific references like the bare English NP new friends and the bare Chinese NP xin pengyou ‘new friends’ in (35, 36), respectively. They are under the setting of [-SR, -HK] with indefinite readings, for the speakers of the both sentences do not intent to refer to any specific referents and do the hearers are assumed not knowing the referring NPs.. (35) A: Lily said it’s hard to make new friends in her new school. B: So bad to hear that. I miss her a lot. (36) Wo xiwang neng kuaidian jiaodao I hope can hurry make ‘I hope I can make new friends soon.’. xin new. pengyou. friend. 3.5 A Comparison of the Article System between English and Chinese This section discussed the typological difference between the article system of English and Chinese. The similarity and difference of English articles and their 35.