Chapter 1 Introduction
In the domain of second language (henceforth L2) acquisition, it has been central issue whether native-like grammatical proficiency is possible beyond critical period, namely post-puberty.
Believing that humans' brain are "genetically pre-programmed" with knowledge of language, Chomsky (1965, 1981) postulated a ground-breaking theory of Universal Grammar (UG) to explain how children acquire language. As a part of humans' innate biologically endowed language faculty, UG is "rich and narrowly constrained in structure and rigid in its essential operations" (Chomsky 1972 quoted in Aitchison 2008: 107).
Using "Principles and Parameters" approach, Chomsky hypothesises that UG consists of invariant principles as well as parameters that account for cross-language variation. Principles are universal to all human languages, whereas parameters allow for cross-language variation, which can be set like a light switch that can be turned on or off. Learning a language, in this view, means re-setting parameter values in a particular way. Though this prevailing approach was widely adopted by linguist for investigating the underlying mental representation of L2 learners, the argument has met with much resistance.
Empirical studies show that L2 learners have difficulty in acquiring Chinese wh-questions (e.g. Chen & Hong 1998; Gao 2009; Yuan 2006; Yuan 2007a; Yuan 2007b). Yuan's (2007a, 2007b) study, along the line with Hawkins' (1997, 2006) hypothesis, suggests that the lexical morphological feature [+wh] encoded in particle ne used by English and Japanese speakers to specify Chinese wh-questions is permanently deficient. This has led to persistent variability in their acquisition of L2 Chinese wh-questions.
The core of this study is to understand and to explain the cause of persistent variability in L2 acquisition. Specifically, the role of first language (henceforth L1) in the acquisition of yes/no questions and wh-questions by Chinese L2 learners of native English speakers will be examined. By revisiting the role of "parameter-resetting" in L2 acquisition, this study adopts a revolutionary approach proposed by Lardiere (2008), namely Feature Reassembly Hypothesis, to investigate the development problem of L2 learners.
Research Topic and Goal
This study aims to substantiate the claim the parameter setting and resetting approach is inadequate to address issues in L2 acquisition such as developmental stages and persistent variability. The role of L1 and feature re-assembly on the acquisition process of L2 is therefore to be located and investigated subsequently in this study.
A growingnumber of studies reportedin the literature of L2 research literature on the failure of L2 learners to produce morphological inflection associated with the acquisition of functional categories and their features (such as Lardiere 2008; Yuan 2007a, 2007b). Question formation, in particular has been a very challenging task for L2 learners whose L1 questions are structured very differently. It is argued that L2 acquisition involves figuring out how features should be reassembled into new configurations in the target language (Lardiere 2005, 2008). Along the line of Lardiere, it is postulated that L2 acquisition is a complex learning process that goes far beyond the simple parameter "switch-setting".
Based on above notion, this study intends to contribute to understanding of feature-assembly. The main goal of this study is to testify hypothesis that, rather than parameter setting, it is the features re-assembly that poses difficulties to L2 learners. An empirical study will be conducted to investigate whether native English speakers are able to re-assembly or reconfigure features in their L2 acquisition of Chinese yes/no questions and wh-questions.
By investigating and analysing the interlanguage of native English speakers, this study endeavours to capture the nature of L2 Chinese learning problem in terms of feature re-assembly. On the other hand, approaches that have been adopted along the line of L1 influence, namely the Full Transfer/Full Access Hypothesis (Schwartz and Sprouse 1994, 1996) and Feature Re-assembly Hypothesis (Lardiere 2008), as well as competing hypothesis the Representational Deficit Hypothesis (Hawkins & Chan 1997) and Interpretability Hypothesis' (Hawkins & Hattori, 2006; Tsimpli & Dimitrakopoulou, 2007) will be compared and examined in this study.
In this paper, I will firstly discuss typology differences between Chinese questions and English questions, including yes/no questions and wh-questions. Next, I will briefly review I will review the L2 acquisition theories related to this study as well as important studies that have been carried out to investigate acquisition of yes-no questions and wh-questions. Finally, I will propose predictions based on hypotheses adopted in this study.
A feature is a set of values and the available options for their realisation on linguistic elements (Kibort 2008). Within Minimalist framework, Chomsky claims that an universal set of linguistic features is available as a part of the humans biology endowment, along with a computational mechanism that constrains the combination and interpretation of the features.
Chomsky's (1995) claim that features are the locus of parametric variation among languages. Lexical items are defined as the associations of three sets of features: phonological (such as Inflectional Class), syntactic (such as Gender, Number and Person), and semantic (such as Tense, Aspect and Polarity) features. Lardiere (2009), consistent with Chomsky's interpretation, describes that features as the fundamental units that make up the lexical items of languages.
Lexical and functional elements are hypothesised of as bundles of features. It is also postulated cross-linguistic variation springs from different combination of features in different languages. Syntactic differences among languages are determined by items that make up the functional categories, such as C (Complementizer), I (Inflection), and D (Determiner). Each of these items comprises sets of one or more features such as [±wh] and [±past]. In other words, using biologically metaphor, features are the DNA of human languages, that is, the bundles that consisted of "genes" (the functional categories) which determine the structure of languages (Goodluck, Liceras & Zobl, 2008).
Interpretable and uninterpretable features
Formal features are further classified into two categories: interpretable features and uninterpretable features (Chomsky, 1995).
Features for Person, Number and Gender of a noun or pronoun are interpretable and required for semantic interpretation. They have a role to play in the semantics of the noun or pronoun, which restrict noun and pronoun denotation. The same features appear on a verb, auxiliary or adjective are uninterpretable and required for formal reasons, such as triggering movement. They have no semantic value, as they do not restrict the denotation of these categories.
Interpretable features survive until logical form (LF), whereas uninterpretable features must be checked and deleted during the derivation to prevent crash at interface level (i.e. LF). In the case of wh-constructions, an interrogative Category (C) has an uninterpretable [wh] which needs to be checked and eliminated by LF. The uninterpretable [wh] feature, acts as probe, looks down the 'tree' for the closest element containing the same feature. If C find its goal, C agrees with it. If the [wh] feature on interrogative also has the Empty Category Principle (EPP) property, the wh-constituent that contains that wh-goal will move the SpecCP to check the strong [wh] feature of C.
Chapter 2 Theoretical Background
During the past decades, a significant number ofstudies of L2 acquisition literature has reported that L2 speakers' proficiency divergesignificantly from native speakers in spite of positive evidence providedby the L2 input (e.g. Hawkins andChan, 1997; Hawkins & Hattori, 2006; Lardiere 2008; Prvost and White 1999, 2000; Tsimpli & Dimitrakopoulou, 2007). Despite the undisputable fact of variability outcomes in L2 grammars, the driving force behind the scene remains controversial.
It is argued that L2 learners' initial state influences developmental stages and ultimate attainment of L2 linguistic competence. In this study, therefore, I will focus on relevant L2 acquisition approaches to address and examine L2 learning problems at initial state as well as developmental stage, in particular, concerning the acquisition of functional categories and remapping of features associated with the functional categories.
Beyond Parameter Setting and Parameter Re-setting
Over the past two decades, "Parameter (re)setting" has been used as a metaphor to describe L2 acquisition. This conception seems to offer a promising approach to investigate L2 acquisition.
In early works within generative framework, the nature of L2 initial state has been linked to L1 transfer. Full Transfer/Full Access hypothesis (Schwartz and Sprouse 1994, 1996) proposed that L1 parameter values are the starting point of L2 acquisition.
To go from the L1 to the L2, learners will often have to reset existing parameters or reassign values to them. Failure to do so will mean that the learner does not attain the L2. (Haegeman 1988 quoted in Lardiere 2008:108)
This suggests that failure to reset parameter from L1 value to L2 value could lead to non-native outcome. As certain parameter values in L1 may not compatible with those in L2, this poses difficulty to L2 learners going through development stages.
Parameter (re)setting approach, however, often failed to cope well with some prominent issues in L2 acquisition. To investigate the deficiency of parameter setting approach, it is worthwhile to take a look at three aspects of L2 acquisition: developmental stages, cross-linguistic variations and persistent variability.
Language Acquisition Device (LAD) is an innate system in human brain that facilitates language development (Chomsky 1965). This system provides human brain the capacity to make referent-symbol associations (that is, to construct grammar) in the correspondence to primary linguistic data. UG, a component within LAD, consists of 2-layer system: hard-wired principles, which are universal and applicable to all languages; and partial wired parameters, a finite set of language properties whose values are "switchable" or "tuneable" and vary among languages.
This model provides a solution to "the logical problem" of language acquisition: input underdetermines output. This phenomenon states that L2 learners' output (performance) seems to be more complex and sophisticated than is evident in their linguistic input. UG and Principles and Parameters, thus, could be account for initial state of L2 acquisition.
As shown in above statements, however, we could notice that UG is never a developmental theory; it can only interact with developmental theories. UG does not aim to explain how and why grammar develops in a particular way, and how learners acquire language knowledge. In fact, Chomsky tried to use this theory to explain "representation problem" of learnerswhat learners come to now, what learners grammars are like (White 2003:37). This suggests that we cannot solely rely on Principles and Parameters approach to investigate variability problem in developing L2 system.
Variation among languages is always equated to different settings of parameter. In Chomsky's generative framework, it is claimed that "a few changes in parameters yield typologically different language" (Chomsky 1981 quoted in Aitchison 2008:111). Accordingly Kayne (2000:8) calculates that "the number of independent binary-valued syntactic parameters needed to allow for 5 billion syntactically distinct grammar is only 33".
Baker (1996:7), in addressing this issue, states: One might expect that more and more parameters comparable to the Pro-Drop Parameter would be discovered, and that researchers would gradually notice that these parameters... themselves clustered in nonarbitrary ways...It is obvious to anyone familiar with the field that this is not what has happened.
Newmeyer (2004:196) later expressed his concern on the overwhelming number of parameters: "If the number of parameters.... is in the thousands (or, worse, millions), then ascribing them to an innate UG to my mind loses all semblance of plausibility."
Following the introduction of Minimalist Program (Chomsky 1995, 2001), however, work is beginning to focus on "feature" rather than "Principles and Parameters" approach (such as Ionin, Ko & Wexler 2008; Lardiere 2008; Sorace 2000). In this more current model, features are the locus of parametric variation among languages. Cross-linguistic variation, expressed by parameters in the past, is restricted to divergence in the strength value of uninterpretable features. These uninterpretable features are associated with functional categories such as C (Complementizer), I (Inflection), and D (Determiner).
This model, hence, suggests that language variation is in the lexicon (Baker 2008; Chomsky 2001). Based on this notion, Lardiere (2009) claims that lexical items that make up functional categories are responsible for the variations of syntactic among languages.
Persistent Variability in L2 data
Typically, parameter setting is an "all or nothing" phenomenon, thus changes in parameter setting suggests an abrupt change in the interlanguage of L2 learners (Kemenade & Vincent 1997).
It is worthwhile to take a look at "No Parameter Resetting Hypothesis" suggested by the studies of Hawkins & Chan (1997), whose study focused on the acquisition of wh-movement of L2 English by L1 Chinese/French speakers. This study attributed non-native-like outcomes to the failure of learners to acquire the [wh] feature which is absent in their L1. "No Parameter Resetting Hypothesis" claim that parameter resetting is impossible if the functional features are absent in the L1. In other words, ultimate attainment is inaccessible if L2 learners fail to parametrically distinguish wh-movement from wh-in-situ languages, that is, to switch from L1 weak [wh] feature strength to L2 strong [wh] feature strength .
As we know, however, L2 learners go through a series of gradual stages in the acquisition process of L2, instead of immediate change. Accordingly, Lardiere (2008, 2009) argues that parameter resetting accounts are insufficient to capture the nature of the learning problems faced by L2 learners. In this case, Feature Reassembly Hypothesis proposed by Lardiere (2008, 2009) offers a refreshed and critical look at the parameter-resetting paradigm.
Feature Re-assembly Hypothesis explains the phenomenon of persistent variability in L2 data relevant to morphological inflection. Lardiere challenges and proclaims that parametric selection has underestimated complex learning problem of L2 acquisition. L2 acquisition is a process of assembling the targetedlexical items of a second language (L2) and it goes beyond the simple "switch setting". It requires L2 learners to"reconfigure features from the way these are represented in thefirst language (L1) into new formal configurations on possiblyquite different types of lexical items in the L2" (Lardiere, 2009).
Theory of Initial State and Developmental Stages: FTFA
The significance of L1 transfer has provoked controversy on the battlefield of L2 acquisition research since early 1990s. Despite the numerous studies that have been carried out over the past decade, there still remains a level of uncertainty in the field concerning where, how and to what extent L1 influences in the L2 learners' grammar.
Full Transfer/Full Access (FTFA) Hypothesis implication on initial state and developmental stages of Chinese questions acquisition by native English speakers will be examined. Also, to investigate learning problem beyond initial state in particular, despite of FTFA, predictions of two current competing models, namely, Interpretability Hypothesis (Hawkins & Hattori 2006; Tsimpli & Dimitrakopoulou 2007) and Feature Re-assembly Hypothesis (Lardiere 2005, 2008) will be analyzed on the basis of their different predictions on L2 acquisition.
Full Transfer Full Access Hypothesis(Schwartz & Sprouse, 1994, 1996)
FTFA (Schwartz & Sprouse 1994, 1996) could be interpreted as a model of "parameter resetting" and "non impairment". Schwartz & Sprouse (1994, 1996) posits that L2 learners start their acquisition with L1 knowledge. This is claimed by "full transfer" part of FTFA, meanwhile "Full Access" part claims that L2 learners have access to UG when L1 grammar is insufficient to accommodate properties of L2 input. These UG resources include new parameter settings, functional categories and feature values (White, 2003).
"Full Transfer" part of this model describes different starting point of L1 and L2 learners: L1 learners begin with a set of unfixed parameters whereas L2 learners begin with fixed parameters. This predicts that L2 learners with different L1s will behave differently with native speakers in developmental stages. Hence ultimate attainment or full competence of L2 is not predicted in this model (Gass & Selinker 2001).
Despite of supported by studies carried out within Parameter and Principles framework (Yuan 1998; Slabakova 2000 quoted in White 2003), FTFA model does not always cope well with issue concerning L1 influence strength and degree. For example, Yuan's (2007b) study reveals that there is no L1 transfer of the L1 English [+wh] feature strength in the L2 data. This is compatible with Platzack (1996)'s "Initial Hypothesis of Syntax" which predicts that properties of the L1 grammar is not available in the L2 initial state. To address this issue beyond initial state, further detail will be discussed in next session of Impairment Hypothesis.
Theories Beyond Initial State
There has been considerable debate, whether uninterpretable features are available to adult L2 learners. If speakers' L1 has not selected a feature for the assembly of lexical items, there are two possibilities predicted for L2 acquisition beyond critical period:
- Failed Functional Features Hypothesis and Interpretability Hypothesis: The inactivated features of L1 are not available to L2 adult learners.
- Missing Surface Inflection Hypothesis and Feature Re-Assembly Hypothesis: The feature is still available for selection.
In order to measure the possibility of applying these hypotheses in this study, further detail will be discussed in next session of this chapter.
Representational Deficit Hypothesis (RDH)
The Failed Functional Features Hypothesis (FFFH) (Hawkins & Chan 1997; Smith & Tsimpli 1995), in contra to FTFA, is a "no parameter resetting", and "impairment" model. This hypothesis postulates that features which determine parametric differences are subject to critical period. FFFH assumes full transfer of L1 in the L2 initial state and predicts that properties associated with uninterpretable features which have not been activated in the L1 grammar are inaccessible beyond critical period. This will consequently pose a learning problem for adult L2 learners.
Hawkins (2005) later reformulated FFFH and claim that uninterpretable features not selected from the UG during the critical period will disappear after puberty. Interpretability Hypothesis (Hawkins & Hattori 2006; Tsimpli & Dimitrakopoulou 2007), in line with FFFH, claims that the full resources of UG are inaccessible to adult L2 learners. Unavailability of uninterpretable formal features thence induces learning problems even at advanced level of L2 proficiency (Tsimpli & Dimitrakopoulou 2007).
Such claims for impaired L2 mental representation contrast with Missing Surface Inflection Hypothesis. This study contributes to the debate by investigating both children and adult L2 data at different level of proficiency, arguing against the standpoint that L2 learners' variable use of inflection indicates in syntactic representations impairment.
Both Yuan's studies on acquisition of Chinese questions by L1 Japanese speakers (Yuan, 2007a) and English speakers (Yuan, 2007b) analyze L2 data on the basis of "lexical morphological feature defect account" and claim that features of lexical item cannot be acquired in L2 acquisition. In other words, in line with Hawkins, Impairment Hypothesis has been adopted in his studies.
Yuan's (2007b) study reveals that there is no L1 (English) transfer in the L2 acquisition of Chinese wh-questions. However differentwh-words behave differently at different L2 Chinese proficiency levels. Yuan claims that this is attributed to the breakdowns at the lexical morphology-syntax interface and morphosyntax-semantics interface levels.
Yuan's another study (2007a) suggests that lexical morphological feature [+wh] encoded in particle ne used by Japanese speakers to specify Chinese wh-questions is permanently deficient. This has led to persistent variability in their acquisition of L2 Chinese wh-questions.
However, Yuan's method in his studies might be insufficient to depict the full picture of L2 acquisition nature. As grammatical judgement alone was used in his both studies, L2 learners' genuine ability might not been examined objectively. Therefore it is worthwhile to re-examine the process of L2 acquisition, in particular with respect to features remapping, by carrying out a study with the combination of grammatical judgement task and elicited production task, which would give a closer account of learning process.
Missing Surface Inflection Hypothesis
A number of L2 linguist (Lardiere 2008; Prvost & White 2000; Rule & Marsden 2006; Schwartz 2002) claim that there is a disassociation between morphology and syntax, that is, morphology and syntax have to be acquired independently. Knowledge of the abstract properties of functional categories is claimed to be available in mental representation of L2 learners. Missing inflection morphology in learner's output cannot account for the claim that learners syntactic representational is impaired. The only problem lies in the process of manifestation, i.e. from features re-mapping to overt morphological forms. This bearing in L2 research is called Missing Surface Inflection Hypothesis (henceforth MSIH).
Feature Re-Assembly Hypothesis
Responding to Hawkins & Chan (1997), Lardiere argues that persistent variability in L2 acquisition is not caused by lack of parameterized formal feature, but rather a set of features bundles mapped into one form, which is much more complicated to acquire than simple parameter switch-setting.
Some recent studies (McCarthy 2008; White 2008) have suggested that analyzing L2 data using Feature Re-Assembly approach may provide a more practical solution to the issue to of persistent variability in developmental stage.
Against RDH Hypothesis, McCarthy (2008) and White (2008) tested MSIH and argue that none of these hypotheses can fully account for the L2 data presented in their studies. McCarthy suggests "the locus of the 'deficit' is not that the learners do not have a syntactic representation for gender but rather that it appears to be located in the feature representations in the morphological domain (McCarthy 2008: 484).
Chapter 3 Syntactic Background and Relevant Studies
This study to investigate L2 acquisition of Chinese questions by native English speakers is driven by the fact that these two languages are typologically different in terms of questions formation. The deviation in the L2 grammars of English native speakers at different levels of proficiency in Chinese is a theoretically appealing issue. It will provide a better understanding of nature and process of L2 acquisition, as well as an opportunity to grasp the concept about acquisition difficulties confronting L2 learners, which may be exacerbated by features re-assembly in the lexical items of L2.
From the typological point of view, English and Mandarin Chinese differ in the way in which a question is formed.
Formation of English yes/no questions involves placing a dummy do or an auxiliary at the beginning of the question so that these grammatical items precede the subject. In wh- interrogatives (or the so-called wh-questions), the wh-words undergoes movement to the sentence-initial position, as well as the operation of subject-auxiliary inversion.
Chinese does not have subject-auxiliary inversion. Wh-Interrogative words stay in the position where they are interpreted, hence the so-called Wh-in-situ. Despite this, Chinese employs clause-final question in questions. In Mandarin Chinese, yes-no questions are prototypically "typed" by the question particle ma whereas wh-questions are "typed" by question particle ne (Cheng 1991). Likewise, Cheng and Rooryck (2000: 2) assume that insertion of wh-particle checks the Q-feature in C allowing the wh-words to stay in situ.
Yes-no questions are interrogatives that can be answered by "yes," such as dui, shi, or "no," such as bu, bu dui, bu shi.
Chinese yes-no questions could be expressed in various ways. Some earlier syntactic approaches emphasised that there are two ways to form yes-no questions in Chinese: A-not-A question forms and the question particle ma. (Li & Thompson 1979; Huang 1991; Li 1992; McCawley 1994 quoted in Schaffar and Chen 2001).
In other words, Chinese yes-no questions can be formed by merging of a question marker ma at C, and furthermore, by no overt merge or move to C at all, as in A-not-A questions. (Zhang 1997)
In Chinese, A-not-A questions can be formed by various way: the Verb-not-Verb type, the Adjective-Not-Adjective type, the Auxiliary-Not-Auxiliary type and the Adverb-not-Adverb type.
- Ta xihuan bu xihuan zhe ben shu? he like not like this book Does like this book? Adjective-Not-Adjective type
- Xingjiapo ganjing bu ganjing? Singapore clean not clean Is Singapore clean? Auxiliary-Not-Auxiliary type
- Zhangsan hui bu hui lai? Zhangsan will not will come Will Zhangsan come? Adverb-not-Adverb type
- Ni zuo shi zixi bu zixi? You do thing careful not careful Do you do things carefully?
- Xinjiapo ganjing ma? Singapore clean Q Is Singapore clean?
- Zhangsan hui lai ma? Zhangsan will come Q Will Zhangsan come?
- Ni zuo shi zixi ma? You do thing careful Q Do you do things carefully?
Yes-No Questions with Particle ma
As assumed in current literature (Cheng et al. 1996: 63; Zhang 1997) ma in yes-no questions checks the strong feature of interrogative C. Zhang (1997) claims that the uninterpretable [Q] of C can be checked by the interpretable [Q] feature of ma.
Ta xihuan zhe ben shu ma? he like this book Q Does like this book?
Previous Studies: Yes-No Questions
Chen & Hong (1998) carried out a study to investigate if L2 acquisition is dominated by UG and L1 within the framework of parameter setting. English and Japanese-speaking learners have been examined based on their performance in Preference Task and Restructuring Task on the interpretation of Chinese A-not-A questions. The results suggest that UG and parameter setting is operative in L2 acquisition, but L1 influence is not significant.
Performance of L2 learners relied not only on Head Parameter, but also on learners' level of Chinese proficiency as well as A-not-A questions categories. Learners have found performed better in V-not-V and Aux-Not-Aux compared to Adj-Not-Adj and Adv-not-Adv type of questions. Chen & Hong state that despite learners are aware of syntactic features of each of this categories, there were fewer responses to Adj-Not-Adj and Adv-not-Adv type In my opinion, this suggests that feature associated with wh-lexical items are responsible for the delay of the acquisition. Hence in this study, an investigation on this type of questions will provide more in-depth perspective within the framework of Feature Re-Assembly Hypotheses, in spite of Parameter (Re)Setting Hypothesis.
There are three major linguistic differences between Chinese wh-questions and English wh-questions.
Firstly, while wh-words in English wh-questions have to move to sentence initial position, wh-words in Chinese wh-questions remain in-situ, that is wh words do not undergo overt wh-movement, as in (9).
Secondly, there is a optional question particle ne in Chinese wh-questions, as in (10), Cheng (2003) further explains that in languages with final question particles, the SpecCP is always occupied, that is, the valuation of force is carried out. This provides the explanation why wh-words in Chinese do not undergo overt movement. Whereas there is no question particle in English wh-questions, wh-phrase movement from its base-generated position to SpecCP is necessary so that the Co could be valued with the specification of force.
Thirdly, wh-words in Chinese wh-questions could be located inside islands, such as Sentential Subjects (as in 11), Complex NP (CNP) (as in 12), and Adjunct Clause (as in 13) and while English wh-questions does not allow wh-words to be extracted from these islands as this would violate Subjacency Principle.
Zhangsan mai-le shenme? Zhangsan buy-PERF what What did Zhangsan buy?'
Zhangsan mai-le shenme ne? Zhangsan buy-PERF what Q What did Zhangsan buy?'
[Shui qu bisai] bijiao heshi? Who go competition relatively appropriate *Whoi is [ti to go to competition] more appropriate?
Zhangsan ma-le [shui xie de shu]? Zhangsan buy-PERF who write DE book *Whoi did Zhangsan buy the book [that ti wrote]?
Xiaowang [yinwei ni zuo le shenme] er shengqi? Xiaowang because you do-PERF what and angry * Whati is Xiaowang angry about [because you did ti]?
Previous Studies: Chinese Simple wh-questions
Yuan (2006) reveals that all English learners at different proficiency of Chinese accepted simple Chinese wh-questions with wh-words such as when, where, how and why remain in-situ. Yuan concludes that there is no L1 transfer at initial state. This means, contradicted to the assumptions of FTFA, L2 Chinese learners do not demonstrate their L1 English parameter setting of [+wh] strong feature strength in their L2 grammar. "The Initial Hypothesis of Syntax" (Platzack 1996) postulating that the initial state includes functional categories with all features at default strength, that is weak strength (quoted in White 2003), seems to be adequate to explain this phenomenon.
However, the same study also shows that the Beginner group failed to reject ungrammatical Chinese wh-questions, i.e. those with wh-words moved to the sentence initial position. This leaves the picture of the initial state remains puzzled, as both FTFA and "The Initial Hypothesis of Syntax" are unable to make full predictions on this mixed results.
WH-Questions with Particle Ne
Yuan's (2006) study shows that all English learners at different proficiency levels of L2 Chinese accept Chinese wh-questions without question particle ne. Similarly, Intermediate, Post Intermediate, and Advanced groups of L2 Chinese learners (native English speakers) accepted Chinese wh-questions with question particle ne.
However, the Beginner and Post-beginner groups rejected this kind of wh-questions, indicating that valuation of Co of Chinese wh-questions by phonetically unrealised ne is preferred by L2 Chinese at beginner levels. This highlights an interesting phenomenon to be investigated: is there any L1 influence at the initial state, and feature re-assembly at post-initial state? Unfortunately, Yuan did not provide further explanation to address this issue thus this leaves a gap to fill.
Previous Studies: Chinese Complex wh-questions
Yuan's (2006) study indicates that wh-words with different grammatical functions in Chinese complex wh-questions do not develop in a uniform fashion in L2 Chinese grammars.
Post Beginner and Intermediate group accepted Chinese wh-questions with a wh-argument words (such as who, what) inside a Complex Noun Phrase (CNP) but rejected wh-questions with a wh-adjunct words (such as when, where) in a CNP. Moreover, none of the learner groups, including advanced learners, accepts wh-questions with wh-adverbial words (such as how, why) inside a CNP.
For wh-argument words, Yuan (2006) argues that this is irrelevant to Subjacency Principle and this phenomenon happens merely due to L2 Chinese learners have mastered the basic structure of wh-questions with an embedded CNP. "Deactivation" of Subjacency Principle at post-beginner level however, in my opinion, suggests that L2 learners have successfully re-assembled features associated to wh-words into their new L2 grammar.
Parodi & Tsimpli (2005)'s Optionality Hypothesis suggested that 'real optionality' (unconstrained) is found in learners with lower proficiency whereas advanced learners perform differently depending on structure-based properties and the morphological richness of L1 compared to L2. By adopting this hypothesis, Yuan (2006) concludes that the variability in different L2 development stages is due to "apparent optionality" (constrained) of Chinese wh-word how. However, in another paper, using same L2 data and results, Yuan (2007) attributed the persistent variability at advanced level to "lexical morphological feature defect" account, instead of "apparent optionality". This also means he changed his standpoint from supporting feature re-assembly (Yuan 2006:160) by stating that "we cannot take it for granted that features of lexical items can be acquired by L2 learners"(Yuan 2007).
As previous studies (such as Hawkins & Chan 1997; Lardiere 1998; Prevost and White 2000; Smith and Tsimpli 1995) have focused on the relationship between functional categories and morphological, Yuan has made contribution by linking the relationship between lexical morphological and syntax in his studies (2007a, 2007b). This undoubtedly is inspirational and motivated for more investigation into the acquisition of Chinese questions from different angle and perspective.
Chapter 4 Hypotheses and Predictions
The three theories outlined in Chapter 3, namely Missing Surface Inflection Hypothesis, Representational Deficit Hypothesis and Feature Reassembly Hypothesis, all make different predictions for English learners of L2 Chinese.
Full Transfer/Full Access predicts that native English speakers of L2 Chinese learners would initially transfer their complete knowledge of the L1 into the L2 (full transfer) but they have access to UG (full access) and that re-structuring 'parameter re-setting' is possible.
As discussed in Chapter 2, English is a language that requires movement of wh-phrases in syntax whereas wh-phrases stay in-situ in Chinese. In such case, FTFA predicts that native English speakers would initially transfer L1 properties on wh-movement (i.e. properties of C and wh-lexical items) to L2 grammar but would eventually acquire non-movement of wh-words by accessing properties in UG. "Full Access" tells us that it would not impose any difficulty to L2 learners to acquire variable expressions related to wh-in-situ and questions particle ma (obligatory for yes-no questions) and ne (optional for wh-questions).
If English L2 learners fail to attain native-likeness in the acquisition of Chinese yes-no questions and wh-questions, evaluation of the issue under an alternative model will be necessary to better understand the issue.
Feature Reassembly Hypothesis
Against RDH, it is expected that the findings of this study will suggest that the most challenging problem for learners does not reside in parametric selection. Instead, the L2 learners whose L1 wh-words are lexicalized differently with L2 will face the difficulty to remap these features into a different L2 configuration. This study adopts Feature-Reassembly approach (Lardiere 2008) is best account to interpret L2 acquisition data.
If the L1 (English) and L2 (Chinese) share a same feature although it is used in different ways (e.g. overt movement and covert movement), it is predicted that L2 learners will still be able to reassemble the L1 features to use in the L2. However, Feature Reassembly does not make any assumptions regarding features of the L2 that are not present in the L1. For example, question particle ma as outlined in chapter 2. Hence this will be investigated by the hypothesis of FTFA.
Research Questions and Method
An empirical study will be conducted to determine whether native English speakers are able to figure out how to remap relevant features of wh-question constructions into different formal configurations in the L2 Chinese.
The major concern of this study is whether L1 and feature re-assembly play a role in L2 acquisition, and if so, how these factors might contribute to different learning outcomes. Research questions pertaining to the research topic are as follows.
- Will English L2 learners of treat L2 (Chinese) like their L1 (English) at initial state when they acquire yes/no questions and wh-questions?
- If FTFA's assumption is correct, this study would tend to prove if there is any L1 transfer at initial state. Will English-speaking L2 learners of Chinese value the ambiguous Co by (1) merging a question particle (ma and ne) into Co, which is adopted by Chinese questions or (2) moving a wh-phrase to SpecCP, which is adopted by English questions.
- Will English speaking learners apply their L1 knowledge of Subjacency in the acquisition of Chinese wh-questions?
- If the native English speakers have problem to interpret Chinese interrogative expressions correctly, is it persistent (as predicted by RDH), or it will be ultimately acquirable( as predicted by Feature Re-assembly Hypothesis) by advanced Chinese learners ?
In order to capture the nature of Chinese yes-no questions and wh-questions by native English speakers, three tasks will be carried out to test their interpretation, that is,
- Grammatical Judgement Task: include acceptance of grammatical and ungrammatical Chinese yes-no questions and wh-questions to test L2 learners' underlying linguistic knowledge
- Oral Production task: to test their spontaneous production of questions based on pictures
- Written Production Task: requires L2 learners to rearrange scrambled words to form questions in correct orders.
- Baker M (1996) The Polysynthesis Parameter. Oxford: OUP.
- Baker M (2008) "The Macroparameter in a Microparametric World" in T. Biberauer eds The Limits of Syntactic Variation . Amsterdam: Benjamins, 351-74.
- Chen CYDoris& LC Portia Hong (1998) "UGand A-not-AQuestionsin L2 Chinese" in Journal of National Taiwan Normal University 43(2): 57-80.
- Cheng Lisa LS ( 2003) "Wh-in-situ" in Glot International7.4 :103-109 & 7.5:129-137.
- Cheng Lisa LS (2009) "Wh-in-situ from the 1980s to now" in Language and Linguistics Compass3.3: 767-791.
- Chomsky N (1981) "Knowledge of language: Its elements and origins" in Philosophical Transactions of the Royal Society of London , 295:223-234.
- Chomsky N (1986)Knowledge of Language: Its Nature Origin, and Use. New York: Praeger.
- Chomsky N (2002) On Nature and Language. Cambridge University Press.
- Ionin T, Ko H & Wexler K "The role of semantic features in the acquisition of English articles by Russian and Korean speakers" in Liceras J M, Zobl, H & Goodluck H eds The role of formal features in second language acquisition. New York: Lawrence Erlbaum, 226-68.
- Kayne RS (2000) Parametersand Universals. New York: Oxford University Press.
- Kibort A (2008) "A typology of grammatical features" Grammatical Features. 11 January 2008. Amended 17 January 2010. http://www.features.surrey.ac.uk/inventory.html.
- Lardiere D (2008) "Feature assembly in second language acquisition". in Liceras J M, Zobl, H & Goodluck H eds The role of formal features in second language acquisition. New York: Lawrence Erlbaum, 106-140.
- Lardiere D (2009) "Some thoughts on the contrastive analysis of features in second language acquisition" in Second Language Research, 25(2): 173-227.
- Liceras J M, Zobl, H & Goodluck H (2008) "Introduction: Formal Features in Linguistic Theory and Learnability: The View From Second Language Acquisition" in Liceras J M, Zobl, H & Goodluck H eds The role of formal features in second language acquisition. New York: Lawrence Erlbaum, 1-11.
- McCarthy C (2008) "Morphological variability in the comprehension of agreement: an argument for representation over computation" in Second Language Research 2008 (24): 459-486.
- Newmeyer F J (2004) "Against a parameter-setting approach to typological variation" in Pica P ed, Linguistic Variation Yearbook 4,181-234.
- Rule S & Marsden E(2006)"The acquisition of functional categories in early French L2 grammars: the use of finite and non-finite verbs in negative contexts" in Second Language Research 22 (2):188-218.
- Schaffar W & Chen L S (2001) "Yes/no questions in Mandarin and the theory of focus" in Linguistics 39(5):837-870
- White L (2003) Second Language Acquisition and Universal Grammar. Cambridge University Press.
- White L (2008) "Some puzzling features of L2 features" in Liceras J M, Zobl, H & Goodluck H eds The role of formal features in second language acquisition. New York: Lawrence Erlbaum, 300-326.
- Yuan B P (2006) "Variability at Different Interfaces in the Development of English Speakers' L2 Chinese Wh-questions" in O'Brien M G, Shea C & Archibald JedsProceedings of the 8th Generative Approach to Second Language Acquisition Conference (GASLA 6), 153-161, Somerville, MA: Cascadilla.
- Yuan B P (2007a) "Japanese speakers' second language Chinese wh-questions: a lexical morphological feature deficit account." in Second Language Research 23(3): 329-357
- Yuan B P (2007b) "Behaviours ofwh-words in English speakers' L2 Chinesewh-questions: Evidence of no variability, temporary variability and persistent variability in L2 grammars" in Bilingualism: Language and Cognition 10(3):277-298. Cambridge University Press.
- Zhang N (1997) Syntactic Dependences in Mandarin Chinese. PhD. Thesis. University of Toronto.