The implicit knowledge of constraints on wh movement in english among l2 learners a case study at english faculty of a university in vietnam m a

VIETNAM NATIONAL UNIVERSITY – HO CHI MINH CITY UNIVERSITY OF SOCIAL SCIENCES AND HUMANITIES FACULTY OF ENGLISH LINGUISTICS AND LITERATURE THE IMPLICIT KNOWLEDGE OF CONSTRAINTS ON WH-MO

VIETNAM NATIONAL UNIVERSITY – HO CHI MINH CITY

UNIVERSITY OF SOCIAL SCIENCES AND HUMANITIES

FACULTY OF ENGLISH LINGUISTICS AND LITERATURE

THE IMPLICIT KNOWLEDGE OF CONSTRAINTS ON

WH-MOVEMENT IN ENGLISH AMONG L2

LEARNERS: A CASE STUDY AT AN ENGLISH FACULTY OF A UNIVERSITY IN VIETNAM

A thesis submitted to the Faculty of English Linguistics & Literature in partial

fulfillment of the Master’s degree in TESOL

STATEMENT OF ORIGINALITY This is to certify that the thesis entitled: “THE IMPLICIT KNOWLEDGE

OF CONSTRAINTS ON WH-MOVEMENT IN ENGLISH AMONG L2 LEARNERS: A CASE STUDY AT AN ENGLISH FACULTY OF A UNIVERSITY IN VIETNAM” submitted by Tạ Lê Minh Phước is a record of bona

fide work under the instruction of Dr Nguyễn Thị Như Ngọc and the supervision of the Faculty of English Linguistics and Literature, Ho Chi Minh University of Social Sciences and Humanities

I guarantee that the content of the thesis was not previously published by any others And the literature mentioned is properly acknowledged

Ho Chi Minh City, May, 2022

TẠ LÊ MINH PHƯỚC

ACKNOWLEDGEMENTS

In order to complete my Master thesis, I have been very fortunate to receive a lot of support and guidance from many beloved people, whom I would love to say

“thank you”

First, a huge thank to my instructor Dr Nguyễn Thị Như Ngọc During the time

I conducted my research, she was a whole-hearted mentor, a disciplined instructor and

an inspiring teacher I was highly motivated by her dedication and patience She got

me through the difficulties of writing a thesis and guided me how to orientate my career path

My thesis could have never been possible without the help of Dr Lê Hoàng Dũng and Dr Nguyễn Đăng Nguyên, who taught me the very core theories of Research Methodology during my coursework

I would love to extend my best gratitude to Ms Bùi Huỳnh Thủy Thương Without her, I might not have seen the light of Linguistics, which is my life passion She has always given me valuable pieces of advice and connected me to the world of knowledge

Moreover, the thesis would have been impossible without the support of 103 students participating in the study I also would like to thank them from the bottom of

TABLE OF CONTENTS

Statement of originality i

Retention of use ii

Acknowledgements iii

List of abbreviations ix

List of tables x

List of figures xiii

Abstract xiv

CHAPTER 1: INTRODUCTION 1

1.1 Background to the study 1

1.2 Aims of the study 3

1.3 Research questions 3

1.4 Significance of the study 4

1.5 Scope of the study 4

1.6 Outline of the thesis 5

CHAPTER 2: LITERATURE REVIEW 7

2.1 Implicit knowledge 7

2.2 Implicit learning 8

2.3 Instruction in EFL classrooms 9

2.4 Interlanguage grammar 11

2.5 Universal Grammar (UG) and Language acquisition 13

2.5.1 Universal Grammar in L1 acquisition: an overview 13

2.5.2.1 The initial stage 15

2.5.2.2 The later stage 19

2.6 Wh-movement in English 21

2.6.1 Ross’s constraints 22

2.6.1.1 The Complex Noun Phrase Constraint 22

2.6.1.2 The wh-island constraint 23

2.6.1.3 The Sentential Subject Constraint 24

2.6.1.4 The Adverbial Island Constraint 25

2.6.2 That-trace effect 25

2.6.2.1 Subject extraction 25

2.6.2.2 Object extraction 26

2.6.2.3 Modifier extraction 26

2.7 Studies into the implicit knowledge of constraints on wh-movement in English 27

2.8 Conceptual framework of the study 28

CHAPTER 3: METHODOLOGY 31

3.1 Research design 31

3.2 Context of the study 32

3.3 Sampling 33

3.4 Research instruments 35

3.4.1 Language Proficiency Test 35

3.4.2 Questionnaire 35

3.4.2.1 Construct and design 36

3.4.2.2 Piloting 39

3.4.2.3 Reliability 39

3.4.2.4 Validity 40

3.5 Data collection procedure 40

3.6 Data analysis procedure 41

3.6.1 Intervals of means 41

3.6.2 Mean analysis 42

3.6.3 Inferential tests 45

CHAPTER 4: RESULTS AND DISCUSSION 46

4.1 Normal distribution 46

4.2 Research question 1 48

4.2.1 Complex Noun Phrase Constraint 48

4.2.1.1 Sentential complements 48

4.2.1.2 Relative clauses 49

4.2.2 Wh-island Constraint 50

4.2.3 Sentential Subject Constraint and Adverbial Island Constraint 51

4.2.4 That-trace effect 52

4.2.4.1 Subject extraction 52

4.2.4.2 Object extraction 53

4.2.4.3 Modifier extraction 54

4.2.5 Further discussion 55

4.3 Research question 2 58

4.3.1 The presence of “that” 59

4.3.2 Modifier extraction 60

4.3.3 Subject extraction 62

4.3.4 Further discussion 63

4.4 Research question 3 64

4.4.1 The B1 group 64

4.4.2 The B2 group 67

4.4.3 The C1 group 69

4.4.4 Further discussion 71

4.5 Summary of the major findings 72

CHAPTER 5: CONCLUSION AND IMPLICATIONS 75

5.1 Conclusion 75

5.2 Pedagogical implications 76

5.2.1 Implications for EFL teachers 76

5.2.2 Implications for L2 learners 78

5.3 Limitations and recommendations for future research 79

REFERENCES 81

APPENDICES 90

APPENDIX 1: QUESTIONNAIRE ON THE IMPLICIT KNOWLEDGE OF

CONSTRAINTS ON WH-MOVEMENT IN ENGLISH 93 APPENDIX 2: MINITAB OUTPUT OF PARTICIPANTS’ DEMOGRAPHIC

CHARACTERISTICS 97 APPENDIX 3: MINITAB OUTPUT OF RELIABILITY OF THE QUESTIONNAIRE 98 APPENDIX 4: MINITAB OUTPUT OF JUDGMENTS ON INVESTIGATED

STRUCTURES OF L2 LEARNERS 99 APPENDIX 5: SYNTACTIC THEORIES OF CONSTRAINTS ON WH-

MOVEMENT IN ENGLISH 113 APPENDIX 6: COURSE SYLLABI OF GRAMMAR COURSES AT EF 126

LIST OF ABBREVIATIONS

LIST OF TABLES

Table 2.1 Summary of the constraint on Complex Noun Phrases 23

Table 2.2 Summary of the constraint on wh-islands 24

Table 2.3 Summary of the constraint on that-trace effect 26

Table 3.1 Summary of the participants’ demographic characteristics 34

Table 3.2 Distribution of items in the questionnaire 36

Table 3.3 Cronbach’s Alpha coefficient of the test items 40

Table 3.4 Intervals of means for data analysis 42

Table 3.5 Grammaticality of the items investigated in the study 43

Table 3.6 Criteria in order to confirm the participants’ acquisition of each structure 44

Table 4.1 Shapiro-Wilk test for normal distribution of the structures investigated 46

Table 4.2 Descriptive statistics of wh-movement in sentential complements 48

Table 4.3 Two-sample t-test for judgments on wh-movement in sentential complements (between complement extraction and modifier extraction) 49

Table 4.4 Descriptive statistics of wh-movement in relative clauses 49

Table 4.5 Two-sample t-test for judgments on wh-movement in relative clauses (between complement extraction and modifier extraction) 50

Table 4.6 Descriptive statistics of wh-movement in wh-islands 50

Table 4.7 One-way ANOVA test for subject extraction, object extraction and modifier extraction out of wh-islands 51 Table 4.8 Descriptive statistics of wh-movement in sentential subjects and adverbial

Table 4.9 Descriptive statistics of wh-movement in subject extraction with and without

“that”………52 Table 4.10 Two-sample t-test for judgments on wh-movement in subject extraction with

and without “that”……….53

Table 4.11 Descriptive statistics of wh-movement in object extraction with and without

“that” 53 Table 4.12 Two-sample t-test for judgments on wh-movement in object extraction with and without “that” 54 Table 4.13 Descriptive statistics of wh-movement in modifier extraction with and without “that” 54 Table 4.14 Two-sample t-test for judgments on wh-movement in modifier extraction with and without “that” 55 Table 4.15 Summary of the participants’ implicit knowledge of constraints on wh-movement in English 55

Table 4.16 Tendency of the participants’ judgments on the extractions over “that”

59

Table 4.17 Tendency of the participants’ judgments on modifier extraction 60 Table 4.18 Tendency of the participants’ judgments on subject extraction 62 Table 4.19 Summary of differences between the participants’ interlanguage grammar and native speakers’ grammar 63 Table 4.20 Summary of the B1 level group’s implicit knowledge of constraints on wh-movement in English 64 Table 4.21 Summary of the B2 level group’s implicit knowledge of constraints on wh-movement in English 67

Table 4.22 Summary of the C1 level group’s implicit knowledge of constraints on movement in English 69 Table 4.23 Summary of structures that the 3 groups (B1, B2 and C1) have acquired 71

LIST OF FIGURES

Figure 2.1 Model for L1 acquisition 14

Figure 2.2 Model for Full Transfer Full Access Hypothesis 16

Figure 2.3 The conceptual framework of the study 30

Figure 4.1 Summary of major research findings 74

For L2 learners, besides their explicit knowledge, implicit knowledge is an important faculty contributing to their linguistic performance The study aims to provide an insight into the implicit knowledge of constraints on wh-movement in English among L2 learners at an English faculty of a university in Vietnam with a Universal Grammar (UG) approach Those constraints are chosen because they are not prevalent in L2 learners’ mother tongue and therefore the acquisition should be UG-bound The participants comprised 103 English-majored students at various English levels The research design was a quantitative case study with the employment of an Acceptability Judgment Test and Minitab 19 software for data analysis The findings of the study have revealed that the L2 learners participating in the study clearly showed their implicit knowledge of constraints on wh-movement in English In addition, the higher the language proficiency is, the more likely an L2 learner is to have acquired the implicit knowledge of constraints on wh-movement in English In the study, those above B2 level provided strong evidence that they have acquired the structures investigated Plus, L2 learners’ interlanguage grammar was not identical to native speakers’ grammar but it differed in several aspects This was due to the effect of L1 transfer and intriguingly enough, L2 learners were able to devise their own rules in interlanguage grammar, which was neither L1’s nor L2’s Lastly, pedagogical implications were withdrawn in order to encourage implicit learning among L2 learners and implicit instruction among EFL teachers in tandem with traditional explicit instruction

Key terms: implicit knowledge, Universal Grammar, L2 learners, constraints on

wh-movement, interlanguage grammar

CHAPTER 1 INTRODUCTION

1.1 Background to the study

Han and Ellis (1998) define implicit knowledge as “knowledge of language” Differing from explicit knowledge, which is breakable into analyzed knowledge and metalanguage, implicit second language (L2) knowledge is manifested through naturally occurring language behavior and can hardly be accessed apart from this behavior (Bialystok, 1990) In addition to that, Ellis (2009) proposes that implicit knowledge is intuitive and tacit: a learner may feel intuitive that one sentence is unacceptable but fail to verbalize the problems with it Then, implicit knowledge has been a focus in SLA research besides explicit knowledge so as to provide evidence that L2 learners can both learn explicitly and implicitly

Take a look back in the history, since its early stages, L2 acquisition research has attended to the question whether there is an analogy between the first language (L1) and second language learning processes Initially, researchers compared the utterances produced by L1 and L2 learners regarding structural similarities and differences (e.g., Dulay and Burt, 1974; Hakuta, 1976; Schumann, 1978) Obviously, what actually interests L1 and L2 acquisition is not those explicit structural properties but the cognitive processes that learners go through The analysis of utterances should act as a window into the mental processes of language acquisition In the absence of theoretical frameworks, Chomsky (1980&1981) proposes a theory called Universal Grammar (UG) which is characterized by a system of principles with respect to the universal properties of natural languages UG argues that there is a common mechanism for humans to acquire a language Specifically, by being exposed to the linguistic input of a language, humans are capable of generalizing its rules and learning how to create proper sentences or utterances That is how a child learns L1 through interaction with his/her parents or family members And there must be a definite set of mental structures that enable a child to acquire his/her mother tongue because if the rules are unlimited, one cannot master a language Plus, those principles

are inherent and at an unconscious level (i.e., implicit knowledge), which explains why sometimes one can judge a sentence grammatical or ungrammatical but he/she cannot give thorough explanations Then, the question is whether L2 learners go through the same cognitive processes as L1 learners or whether L2 learners have access to UG The UG-based research trend has long been in vogue in SLA

Then, Selinker (1972) proposes the concept of L2 learners’ interlanguage grammar or linguistic system produced by L2 learners while they are in the process

of learning the target language Interlanguage grammar belongs to neither L2 learners’ native language nor the target language And Selinker claims that L2 learners can never reach the grammar of native speakers For SLA research, the next question is that whether L2 learners’ interlanguage grammar actually differs from the grammar

of native speakers or standard grammar

In the context of Vietnam, in Barnard’s (2009), most Vietnamese EFL teachers agreed that grammar should be taught explicitly and students most benefit from teachers’ clear explanations Moreover, over 2/3 of them disagreed that L2 learners can learn grammar through exposure to authentic materials or natural use of English The results of this study shows that in Vietnam, the mainstream approach to English grammar teaching is explicit instruction, which undermines L2 learners’ ability to learn implicitly Thus, it is of great necessity to research the implicit knowledge of L2 learners

Moreover, in Vietnam, so far, not many studies have been conducted to understand L2 learners’ language acquisition from the UG perspective Hence, the researcher decided to implement this study with a view to contributing another aspect

of SLA among L2 learners in Vietnam, to what extent L2 learners have acquired implicit knowledge of linguistic features in their SLA process

1.2 Aims of the study

First, the study aims to identify to what extent L2 learners at an English faculty, i.e., hereby called EF (where the thesis writer is working), have acquired knowledge

of English beyond the conscious level (i.e., implicit knowledge) To be specific, the research employs constraints on wh-movement in English, which are never taught explicitly to English-majored students at this faculty, as a medium to check whether

they can recognize ungrammatical wh-questions in English without any prior

conscious awareness of such constraints (through written materials or real-life interaction)

Second, the study explores the discrepancies between L2 learners’ interlanguage grammar and standard English grammar The two are not identical even though L2 learners learn the same English like that of native speakers L2 language acquisition can be better understood through the analysis of those differences

Third, the research attempts to identify whether language proficiency contributes to the formation of the implicit knowledge of constraints on wh-movement

in English among L2 learners at the EF By pinpointing this, EFL teachers can maximize students’ ability to learn the L2 both implicitly and explicitly

Based on the aims and objectives, the study shall offer pedagogical implications concerning the development of L2 learners’ implicit knowledge of language especially in the context of Vietnam where L2 grammar is mainly taught explicitly The purpose is not to refute the role of explicit instruction but to posit another possibility of L2 grammar teaching: EFL teachers’ instructional patterns can

be either explicit or implicit

1.3 Research questions

- Research question 1: To what extent have L2 learners at the EF acquired the implicit knowledge of constraints on wh-movement in English?

- Research question 2: What are the differences between L2 learners’ interlanguage grammar and native speakers’ grammar?

- Research question 3: Is there a relationship between the acquisition of the implicit knowledge of constraints on wh-movement in English among L2 learners at the EF and their language proficiency?

1.4 Significance of the study

The findings of this study are expected to introduce the core theories of UG, which assumes that there are innate universal language properties that govern language acquisition and stay hidden in the language learner’s mind Without those properties, one is unable to identify rules or forms of a language and thus he/she cannot fully acquire a language The study shall help to determine the extent to which L2 learners at the EF possess analogous abstract representations in comparison to standard English grammar To be specific, they will be tested on what they believe to

be grammatical in the language being learnt and what they consider ungrammatical Then, the results derived from the study shall consolidate the belief that students are capable of synthesizing grammatical rules unconsciously through exposure to the language per se Therefore, English teachers need not “spoon-feed” every single grammatical point to students, especially English majors like those at the EF Then, the findings shall give more evidence to encourage English teachers to instruct grammar implicitly at appropriate levels rather than highly focus on explicit instruction as found in Barnard (2009)

1.5 Scope of the study

The study focuses on students’ implicit knowledge of constraints on movement in English only, other aspects of grammar are not included To be specific, the study investigated 5 constraints:

wh-1/ Complex Noun Phrase Constraint

2/ Wh-island Constraint

3/ Sentential Subject Constraint

4/ Adverbial Island Constraint

1.6 Outline of the thesis

The thesis comprises 5 main chapters:

The first chapter, Introduction, provides background information on the study

and purposes for conducting the research It also offers the significance and scope of the study

The second chapter, Literature Review, provides the theoretical approach of

the study In this part, various theories with respect to Second Language Acquisition and Universal Grammar In addition, previous studies on constraints on wh-movement

in English are discussed

The third chapter, Methodology, presents the methods utilized to conduct the

study In this part, research questions are elaborated In addition, the research design and procedures for collecting and analyzing data are fully described

The fourth chapter, Results and Discussion, contributes to the analysis and

discussion of the collected data together with the major findings of the study This chapter answers the three research questions, leading to pedagogical implications in chapter five

The last chapter, Conclusion and implications, summarizes the current

research findings and offers implications for EFL teachers in teaching English

grammar to L2 learners and how to motivate L2 learners’ implicit learning besides explicit learning

CHAPTER 2 LITERATURE REVIEW

This chapter offers theoretical background to the study with a focus on Universal Grammar (UG) and Second Language Acquisition (SLA) Also, previous studies on wh-movement in English are prevalent in this chapter as the foundation for the current study In the end, a conceptual framework is presented so as to visualize the flow of the research

2.1 Implicit knowledge

Implicit knowledge is constantly applied in every aspect of life A person who has no professional training in music can recognize an out-of-tune note of a song he

or she is familiar with; a cook can taste a subtle difference in the tastes of two bowls

of soups but he or she cannot explain that As for a fluent speaker of a language, he

or she can detect grammatical irregularities without being able to verbalize the underlying rules Implicit knowledge can be influential even in the absence of awareness and that knowledge is acquired and it is currently affecting processing (Cleeremans, Destrebecqz, & Boyer, 1998, p 406) Implicit knowledge should be distinguished from explicit knowledge, which we know that we know and that we are aware of utilizing (Dienes & Perner, 1999)

In addition, Ellis (2009, p.11-14) proposes certain features of implicit knowledge:

(1) Implicit knowledge is intuitive and tacit: Upon the interaction with an

ungrammatical sentence (e.g., I an apple eat), a learner may feel intuitive that the

sentence is unacceptable in standard English but fail to verbalize the problems with the sentence

(2) Implicit knowledge is automatically processed: Implicit knowledge can

be easily and rapidly accessed through unintentional language use

(3) Implicit knowledge is only evident in learners’ verbal behavior:

Implicit knowledge cannot be verbalized due to its existence in the form of statistically weighted connections between memory nodes This is an important feature to distinguish implicit knowledge from explicit knowledge

(4) There are limits to the acquisition of implicit knowledge: In spite of the

fact that implicit knowledge is learnable, constraints on the ability of learners to fully learn an L2 implicitly are prevalent Therefore, L2 learners may not achieve native speaker proficiency Aging can hinder L2 learners’ ability to learn implicitly (Birdsong, 2006)

It is still a controversy whether implicit and explicit systems are separate Krashen (1981) suggests that the two systems are entirely distinct Additionally, Paradis (2004) postulates that the two types of knowledge reside in different parts of the brain However, there is wide acceptance that implicit knowledge and explicit knowledge interact with each other during the learning process of L2 learners Bialystok (1982) claims that language use typically involves learners relying on both systems to build up messages Moreover, it is probable that L2 learners can develop both implicit and explicit knowledge of the same linguistic feature In conclusion, besides explicit knowledge, L2 learners can consult their implicit knowledge for language processing

To sum up, among L2 learners, implicit knowledge also contributes to language processing together with explicit knowledge This lays the foundation for the research question 1 of the study, which aims to identify the extent to which L2 learners have acquired the implicit knowledge of constraints on wh-movement in English

2.2 Implicit learning

Implicit learning was first used by Arthur Reber (1967, p.317) to describe “a

stimulus environment without intending to and without becoming aware of the knowledge they have acquired”

Awareness is the most crucial factor to identify implicit learning Thus,

‘awareness’ should be defined Schmidt (2001) categorizes two types of awareness: awareness as noticing (involving conscious attention to “surface elements”) and metalinguistic awareness (involving awareness of abstract rules that govern linguistic phenomena) Schmidt argues complete implicit learning is not possible because at least noticing is needed for learning and noticing requires certain extent of awareness Then, implicit learning could be redefined “learning without metalinguistic awareness” On the other hand, Williams (2005) agrees that learning without awareness at the level of noticing is plausible The debate remains ongoing

According to Ellis (2009), implicit learning involves no central attentional resources Ellis (2008) proposed “generalizations arise from conspiracies of memorized utterances collaborating in productive schematic linguistic productions” Therefore, the knowledge acquired is “subsymbolic, reflecting statistical sensitivity

to the structure of the learned materials” In addition, implicit learning happens unconsciously so verbalization is not feasible In EFL classrooms, under the instruction of teachers, L2 learners do not only learn English explicitly but they can also acquire knowledge implicitly

It can be concluded that L2 learners are capable of both explicit learning and implicit learning Therefore, they possess the ability to learn implicitly grammatical points and apply those rules unconsciously

2.3 Instruction in EFL classrooms

Ellis (2005) defines instruction as the process of getting involved in interlanguage development including direct and indirect intervention Direct intervention is featured by ‘a structural syllabus with the preemptive specification of what learners are supposed to learn while indirect instruction derives from a task-

based syllabus in order to provide learners with chances of communicating in the L2’ Nevertheless, explicit and implicit instruction do not align exactly with direct and indirect instruction According to Ellis (2009), unlike explicit instruction which includes the metalinguistic awareness of the target structure, implicit instruction aims

to encourage learners to infer rules without awareness

According to Housen and Pierrard (2006), explicit instruction involves direct attention to the target form; the target form is presented with the frequent use of metalinguistic terminology; communication is interrupted to explain the target form and practice of the target form is highly controlled On the contrary, implicit instruction requires no direct attention to the target form; the instruction is unplanned and spontaneously; the target form is presented in relation to the context with no use

of metalanguage and learners are encouraged to use the target form freely

Ellis (2009) postulates that explicit instruction does not only lead to explicit learning but also implicit learning Likewise, implicit instruction may result in both explicit and implicit learning For instance, when the teacher explicitly explains the rules of English articles to the learners, besides the target feature, they may adopt implicitly and incidentally other lexical items or grammatical features

Regarding the effectiveness of the two instructional approaches, Norris and Ortega (2000) suggest that explicit instruction is more effective in comparison to implicit instruction in their meta-analysis On the other hand, Krashen (1994)

proposes No Interface Hypothesis which supports the fact that simple rules can be

taught with explicit explanations while it is rather hard to use the same strategies for teaching complex rules and thus those should be taught implicitly Tammenga-Helmantel et al (2014) also agree that complex structures should be learned implicitly To sum up, the effectiveness of explicit and implicit instruction in correlation with structural complexity has not been fully verified However, within the limited classroom interaction, all rules cannot be explicitly taught by teachers,

economical and should be applied in the EFL classroom together with explicit instruction in order to maximize the amount of knowledge conveyed

2.4 Interlanguage grammar

The concept of interlanguage grammar (IL) was proposed by Larry Selinker in

1972 According to Selinker (1972), IL is defined as the linguistic system of L2 learners when they attempt to produce meaningful communication in L2 at various levels: phonology, morphology, syntax, semantics, and pragmatics This system is inherently autonomous and patterned In other words, IL is not a collection of random

or unsystematic errors; neither is it the L1 or the native language (NL) nor the L2 or the target language (TL) In other words, IL is a separate transitional linguistic system Another claim made by Selinker (1972) is that IL is typically unconscious thus introspective analysis is not open to L2 learners They are unaware of the linguistic characteristics of the language they are unconsciously using and unable to verbalize the rules involved, which means that IL exists as implicit knowledge in L2 learners’ learning process Interestingly, even though L2 learners can describe L2’s rules that they have consciously learned in the classroom, those rules are not the interlanguage rules they actually use when focused on meaning In addition, Selinker argues that IL always “fossilizes” or it stops evolving at some point before reaching the full L2’s language system Therefore, L2 learners can never reach the proficiency of native speakers

Selinker identifies five cognitive processes that shape the interlanguage linguistic system: (1) native language transfer, (2) overgeneralization of target language rules, (3) transfer of training, (4) strategies of communication, and (5) strategies of learning To be specific:

- Native language transfer involves in the process of shaping interlanguage rules Selinker (1972, 1992) and Weinreich (1968, p.7) suggests that the

cognitive process underlying transfer is called interlingual identifications:

“the perception of certain units as the same in NL, IL, and TL” For instance, L2 learners may perceive NL “cái bàn” as exactly as TL “a table”, and develop the same meaning in IL Plus, errors can be made in TL due to

NL transfer To clarify, Vietnamese learners of English may produce a sentence like “He eat an apple” without the -s at the end of the verb “eat” because in Vietnamese, verbs do not change as the subject changes

- The second cognitive process is overgeneralization of TL rules This process is more common among children and sometimes called developmental process L2 learners seem to have mastered a rule, but they

do not know exceptions To illustrate, L2 learners may overuse the past tense marker -ed for *drinked, *goed, *hitted, etc The overgeneralization error shows that the learners have learned the general rule, they need to proceed to the next step: learning the exceptions

- Transfer of training is the result of rules learned from instructors or textbooks The success of using those rules is not always guaranteed For example, if the teacher explains in the classroom that past perfect tense is considered as “past past”, L2 learners may erroneously use the past perfect for events in the distant past (not the action before another past action) Therefore, EFL teachers should be cautious with their lesson plans and explanations

- Strategies of communication are employed when L2 learners struggle to get meanings across and their language system is not complete as that of native speakers The sole purpose is to communicate and get people to understand even though the linguistic form is not correct Consequently, this may prolong in their IL

- Strategies of learning are L2 learners’ conscious attempts to master the TL but the result is not always successful If L2 learners get confused in their learning process and form incorrect linguistic rules, those may persist in

their IL As the result, L2 learners may reach at interlanguage grammar of neither the L1’s nor the L2’s grammar (Schwartz and Sprouse, 1994) Since its introduction in 1972, the interlanguage hypothesis has considerably impacted the field of SLA and evolved in certain ways (Han & Tarone, 2014) In the process, it is also influenced by UG The question is whether IL is a natural language and constrained by UG In Selinker (1972)’s initial hypothesis, IL is not a natural language Selinker argues that natural languages are produced by UG but as for ILs, unlike native languages, “fossilize and evidence native language transfer” therefore must be produced by other cognitive processes (not UG) In contrast, Adjemian (1976) suggests that ILs are natural languages even though “unlike other natural languages,

IL rule systems are permeable to invasion from NL and TL rules” Thus, ILs are constrained like natural languages and supposed to obey language universals ILs differ from natural languages in the aspect that ILs fossilize where parameters have been set for the L1, and the L2 with different parameters settings must be learned The debate is still ongoing (White, 1990)

UG-To sum up, interlanguage grammar has been intensively researched and believed to be inseparably linked to UG, which leads to the next part with the focus

on UG and Language Acquisition

2.5 Universal Grammar (UG) and Language Acquisition

2.5.1 Universal Grammar in L1 acquisition: an overview

White (2003) claims that a main task required for L1 learners is to reach a linguistic system which accounts for the input, which enables a child to construct linguistic representations and to comprehend and reproduce language In that sense,

UG acts as a “computational system” that decides which structure is possible or impossible Those are invariable principles which remain true across languages Moreover, UG allows a parametric system which explains the diversity from language

to language Basically, the features of language stay in a binary mode like a “switch”

which can vary between two ends For example, the feature “Null Subject”, in Spanish, it is [+null subject] because the subject can be dropped without affecting the grammaticality of the sentence while in English, such rule is not allowed Thus, English is analyzed as [-null subject]

In order to set the parameters, L1 learners need exposure to the target language

At the initial stage (S0), the child is equipped with the language knowledge constituted

by UG Then, the child gets exposed to the primary linguistic data (L1 input) or the environment where that language is mainly spoken Gradually, the child absorbs the input and accumulates a language-specific lexicon The Grammar (G) may be restructured many times before the child reaches a steady state of L1 grammar (Ss)

Figure 2.1 Model for L1 acquisition Source: White (2003, p.3)

2.5.2 Universal Grammar in L2 acquisition

First, as mentioned above, interlanguage grammar is the grammar of native speakers or L2 learners The concept was proposed by Adjémian (1976), Corder (1967), Nemser (1971) and Selinker (1972) Those researchers postulated that mistakes made by L2 learners are not random but rule-governed behavior

non-White (2003) points out that there is a logical problem in L2 acquisition similar

to L1 acquisition: “L2 learners acquire complex and subtle properties of language that could not have been induced from the L2 input” This consolidates Plato’s paradox:

“There is knowledge that we do not know that we know” Therefore, this strongly indicates that interlanguage grammars are constrained by the principles of UG

S0 (Universal Grammar - UG)

L1 input

grammars might be drawn from L1 grammar, not UG Then, White (2003) suggests that in order to prove UG constrained interlanguage grammars, two conditions must

be satisfied:

1/ The phenomenon investigated must be underdetermined by the L2 input, which means the feature cannot be acquired through L2 input

2/ The phenomenon cannot be derived from L1 grammar

This point explains why the researcher chose “Constraints on wh-movement in English” to test among L2 learners whose mother tongue is Vietnamese Firstly, students majoring in English at EF have never been taught explicitly the “Constraints

on wh-movement in English” during their undergraduate program Secondly, Vietnamese is a wh-in-situ language, which means wh-movement is not a distinctive feature of Vietnamese grammar

2.5.2.1 The initial stage

As aforementioned, UG constrains L1 acquisition at the initial stage Then, the question is whether UG transforms itself into a steady-state grammar or it remains as

an independent entity from L1 grammar Bley-Vroman (1990) argues that UG cannot survive after the process of L1 acquisition Therefore, the initial stage of L2 acquisition must be L1 grammar, and gradually L2 grammar is developed from the mother-tongue grammar To argue against this point, Epstein et al (1996), and Flynn and Martohardjono (1994) show that bilingualism poses a problem for Bley-Vronman’s argument because a bilingual child’s acquiring system usually requires different parametric settings Thus, UG must be constant and distinguished from the L1

Next, the problem is whether adult L2 learners get access to UG and whether L1 grammar is involved in L2 acquisition There have been quite a few theories proposed on the problem and linguists have developed three hypotheses so as to

examine the role of UG in L2 acquisition, including: Full Access Hypothesis, Partial

Access Hypothesis, and No Access Hypothesis

(i) Full Access Hypothesis

Schwartz and Sprouse (1996) propose the Full Transfer Full Access

Hypothesis In this theory, L2 learners start out completely with L1 grammar (Full Transfer), however, L1 grammar cannot accommodate the learning process of L2

Consequently, they have to turn to UG as the last resort (Full Access) UG

continuously governs new parametric settings, functional categories, and features through L2 input before interlanguage grammar reaches its stable state (IL Ss)

Figure 2.2 Model for Full Transfer Full Access Hypothesis Source: White (2003, p.61)

The Full Transfer Full Access Hypothesis has received some criticism What

if at the initial stage, L2 learners of different L1s show the same behavior of a particular phenomenon in the L2 Yuan (2001) shows that French-speaking and English-speaking learners of Chinese exhibit the same habit of verb placement in the early stage Therefore, L1 grammar cannot be the starting point in L2 acquisition

Vainikka and Young-Scholten (1996a, b) propose another theory called

Minimal Trees Hypothesis In this hypothesis, the interlanguage initial state is

partially constrained by L1 grammar, which is limited to lexical categories and inaccessible to functional features And by exposure to L2 input, L2 learners shall get access to UG to withdraw certain functional features of the L2 And functional

Universal Grammar (UG)

S 0 = L1

Ss

IL G2

IL

G 1

IL S S

L2 input

hypothesis: given the well-formed L1 grammar and UG inventory of functional categories at hand, why should one lack access to either L1 grammar or UG at the early stage of interlanguage, which is not a logical mechanism

Another theory is Valueless Features Hypothesis of Eubank (1996) Eubank

shows evidence that both lexical and functional features of L1 are prevalent in the early stage of interlanguage development, but feature strength does not transfer, which means that features are valueless (neither strong nor weak) L2 feature strength is acquired upon exposure to L2 This approach was criticized for the unclear point that features are valueless or inert

The three aforementioned theories all agree that L2 learners begin with L1 grammar, UG is accessed later But there are other perspectives claiming that UG

should be the starting point Platzack (1996) proposes the Initial Hypothesis of

Syntax assuming that L1 and L2 acquisition are not different L2 also starts with UG,

all the functional categories are set at default as weak Those settings are separated from L1 settings Then, L2 learners have to figure out which features to set strong in the process of learning the L2 This approach fails to explain some transfer effects of L1 on interlanguage grammar Hence, the fact that L2 learners start off with weak features is questionable

The next one is Full Access Hypothesis by Epstein et al (1998), Flynn and

Martohardjono (1994), and Flynn (1996) In this theory, UG is fully accessed by L2 learners at the initial stage This might undermine the transfer effects of L1 grammar and make it impossible to predict the initial stage of L2 grammar

We have looked at claims about the Full Access Hypothesis of L2 acquisition And it seems that the Full Transfer Full Access Hypothesis and the Full Access are

the most logical explanation for the initial stage of L2 language acquisition: either L1 grammar or UG is the initial stage But in the hypotheses, we can notice that UG has always been prevalent in L2 acquisition

(ii) Partial Access Hypothesis

In this approach, UG is partially accessible to L2 learners Partial Access

Hypothesis was proposed in the 1980s (White, 1986) then renamed in the 1990s as Failed Functional Features Hypothesis (Smith and Tsimpli, 1995; Tsimpli and

Roussou, 1991) Within the scope of this hypothesis, the fossilization of grammars is due to the fact that functional features (e.g., case and agreement) can only be accessed during L1 acquisition Hence, such features are not obtainable among L2 learners

Then, the hypothesis was modified with the novel version named the Interpretability

Hypothesis, which suggests that uninterpretable features that are not instantiated in

L1 are not available for L2 acquisition (Tsimpli et al., 2003; Tsimpli et al., 2007) The explanation is that uninterpretable features are subject to critical period constraints and they are inaccessible to L2 learners To put it simply, language acquisition is more effortless among young learners while adult learners may not achieve full acquisition

On the other hand, interpretable features are accessible to L2 learners even though L2 differs from L1 substantially

The interpretable and uninterpretable features were introduced by Chomsky (1995) in his theory of the Minimalist Program Based on this approach, the interpretable features (e.g., animate – inanimate; human - non-human; singular – plural, etc.) have a semantic role and they have actual meanings while the uninterpretable features (e.g., uninterpretable wh-feature) exist only for syntactic operations They can be strong or weak across various languages These uninterpretable features and their interpretable counterparts must be checked with one another There are two possibilities in the checking process If the uninterpretable feature is strong, it induces its interpretable counterpart to move to a specific position

If the uninterpretable feature is weak, its interpretable counterpart stays in-situ or no movement happens Obviously, English possesses an uninterpretable strong wh-feature while Vietnamese has an uninterpretable weak wh-feature Therefore, The

constraints on wh-movement in English because the uninterpretable strong wh-feature

in English is not accessible to L2 learners as they have been in their adulthood

The Partial Access Hypothesis tends to limit the linguistic competence of L2

learners Even though L2 learners got exposed to L2 later in their life, the likelihood that they have acquired L2’s features (both interpretable and uninterpretable ones) cannot be completely excluded

(iii) No Access Hypothesis

No Access Hypothesis argues that L1 and L2 acquisition undergo different

cognitive processes (e.g., Bley-Vroman, 1989; Clashen, 1988 Clashen; and Muysken, 1989) Contrary to L1 acquisition, L2 acquisition excludes adult L2 learners’ UG accessibility Learning does take place but not as the result of UG’s involvement, it occurs by means of non-linguistic learning strategies including distributional analysis, analogy, hypothesis formation, and testing The explanation for this approach is that the innate capacity for language learning deteriorates with age (Freeman & Long,

1991) However, the No Access Hypothesis seems to oversimplify the cognitive

process of L2 learning and underestimate humans’ innateness for language acquisition

2.5.2.2 The later stage

In the previous part, various hypotheses regarding the nature of grammatical representations adopted by L2 learners in the initial stage of L2 acquisition were discussed In this part, the researcher will examine the later stage of interlanguage grammar development with two primary approaches to this issue The first position argues that there is a breakdown in parametric systems: either global or local, which implies that interlanguage grammar is not fully UG-constrained The second perspective proposes that interlanguage grammar is unimpaired and still parameter-

driven with two subcategories involved: No Parameter Resetting; and Parameter setting and resetting

Global impairment: This implies that there are no parameters at all in

interlanguage grammar Proponents of this approach include Clahsen and Hong (1995) and Neeleman and Weerman (1997), who believe that interlanguage grammar

is construction specific Learning is possible through pattern matching: learners focus

on surface properties, unconsciously taking account of similarities and differences across linguistic forms (Bley-Vroman, 1997)

Local impairment Hypothesis: Beck (1998) argued that breakdown is more

local than global She claimed that parameters are prevalent in interlanguage grammar but in defective forms: feature strength is deemed to be permanently impaired Therefore, it is predicted that no development in this domain is possible even for advanced learners On the other hand, there is a question of the feasibility of this theory: is it safe to assume that impairment is permanent? However, Beck (1997)suggests that this prediction is not the case Beck (1997) suggests that while L2 learners seem to perform well in the L2, the achievement is not the result of grammar itself but the operation of “additional, agrammatical, learned mechanisms” This claim

renders the Local Impairment Hypothesis infallible with “patch-up” mechanisms

No Parameter Resetting Hypothesis: In this theory, the interlanguage

grammar is supposed to get access to parameter settings realized in the L1 only and there is no subsequent resetting in response to the L2 In other words, new parameter values cannot be acquired and L2 learners can never reach native-like criteria of grammar The L1 grammar controls the learner’s representation of the L2 initially and later on Differing from the previous theories, parameters are not assumed to break down, they just cannot be reset And the interlanguage is considered UG-constrained because it has the same properties as the L1 grammar Hawkins and Chan (1997) showed that L2 (=English) learners have access only to those functional features

Chinese-speaking learners of English showed accurate performance on grammaticality judgment task concerning Subjacency violations

Parameter setting and resetting: in this approach, learners can acquire L2

functional categories, features, and feature values that are absent from the L1 grammar In other words, parameters shall be reset to accommodate the L2 grammar Besides, L2 learners may reach at parameter settings of neither the L1’s nor the L2’s grammar (Schwartz and Sprouse, 1994) Deviations from the L1 grammar prove that interlanguage grammars are UG-constrained (e.g., Epstein et al., 1996; Schwartz and Sprouse, 1996)

The researcher believes that L2 learners’ ability to acquire a language (even in their adulthood) is not limited L2 learners can still get access to UG and parameters

can be reset to suit the L2 grammar as proposed in Parameter setting and resetting

However, the link to UG, different from L1 acquisition, is not natural but based As postulated in this study, language proficiency is the key to open the link to UG: if one can reach a certain English level, the link can be reinforced

condition-2.6 Wh-movement in English

Prior to the 1960s, wh-movement was deemed unconstrained For example:

(1) a What did John write _?

b What did you say that John wrote _?

c What did you think that John said that he wrote _?

As can be seen from (1a, b, and c), wh-movement can range from a short to a very long distance as long as the brain could hold the flow of information However, Ross (1967) discovered some restrictions or contexts in which wh-movement could

be hindered Before further analyses are presented, some symbols should be clarified:

*: an ungrammatical sentence

?*: a marginal sentence

2.6.1 Ross’s constraints

All of the following examples are extracted from the questionnaire of the study Even though the examples used are different from the original work of Ross (1967), the structures investigated are ensured

2.6.1.1 The Complex Noun Phrase Constraint

No extraction is allowed out of a complex noun phrase A complex noun phrase

is a noun phrase that takes a clause either as a complement (a sentential complement)

or as a modifier (a relative clause)

As in (2.1.b), when the direct object (complement) “the Coronavirus” is

extracted out of a sentential complement, the wh-question receives a marginal result

+ Modifier extraction

(2.2) a I like the idea that humans can live in space in the near future

b *When do you like the idea that humans can live in space ?

For (2.2.b), extraction of a modifier out of a sentential complement is not

possible

b Relative clauses

+ Complement extraction

(2.3) a John remembered the girl who wore a green T-shirt at the party

b *What did John remember the girl who wore at the

party?

+ Modifier extraction

(2.4) a Minh has a friend who started studying Japanese in 2010

b *When does Minh have a friend who started studying Japanese _?

à (2.3.b) and (2.4.b): Extraction of a complement or a modifier out of a relative clause is not feasible

Table 2.1 Summary of the constraint on Complex Noun Phrases

Extraction of a complement Extraction of a modifier

2.6.1.2 The wh-island constraint

No extraction is allowed out of a wh-island A wh-island is an embedded

clause that contains a wh-word in it

a Subject extraction

(2.5) a I wonder whether Toan will marry Giang next month

b *Who do you wonder whether will marry Giang next month?

b Object extraction

(2.6) a I wonder whether Toan will marry Giang next month

b ?*Who do you wonder whether Toan will marry _ next month?

c Modifier extraction

(2.7) a I wonder whether Toan will marry Giang next month

b *When do you wonder whether Toan will marry Giang _? The extraction of a subject and a modifier out of a wh-island result in an ungrammatical sentence while the extraction of a direct object out of a wh-island leads

to the marginality of the sentence

Table 2.2 Summary of the constraint on wh-islands

Subject extraction Object extraction Modifier extraction

2.6.1.3 The Sentential Subject Constraint

No extraction is possible out of a sentential subject

(2.8) a That John loves English is obvious

b *What is that John loves obvious?

The extraction of a complement out of a sentential subject results in an ungrammatical sentence

2.6.1.4 The Adverbial Island Constraint

No extraction is possible out of an adverbial island An adverbial island is part of a sentence that contains extra information that can be omitted without rendering the sentence ungrammatical Adjunct islands typically include clauses introduced by “because”, “if” and “when”

(2.9) a Ben went home early because he had to do his homework

b *What did he go home early because he had to do _? The extraction of a complement out of an adverbial island makes the sentence ungrammatical

Ross’s constraints are considered as an important contribution to UG Thanks

to Ross’s work, it is possible to explain the ungrammaticality of certain wh-question structures This consolidates the argument that there must be an internal mechanism

of language acquisition with a finite set of rules and constraints, which enables children to learn a language However, Ross’s theory is more descriptive than explanatory and it requires wider theories to further explicate the constraints on wh-

movement, which are presented in Appendix 5 (p.116)

2.6.2 That-trace effect

The focus in this part is the extraction of subjects, objects, and modifiers over the complementizer “that”

2.6.2.1 Subject extraction

(2.10) a I said that John read a book last night

b *Who did you say that _ read a book last night?