The Interpretation of English Reflexives and Pronouns by Adult Speakers of Chinese An analysis of language transfers

The study reports the L2 acquisition of English reflexives and pronouns between native English speakers and adult Chinese speakers of L2 English. It is found that Chinese pronouns are quite similar with English, and Chinese and English pronouns can be either long-distantly binding or referred to others except for the local domain in complex sentences. However, Chinese reflexives are not the same as English reflexives. Chinese reflexives have two types, monomorphemic ziji and polymorphemic ta ziji, and one of the goals on the study is to test whether Chinese speakers of L2 English will domain the local binding in the complex sentences of English reflexives himself/herself. Finally, the results of the interpretation of English pronouns and reflexives show that L1 transfer cannot totally account for the interpretation of sentences. Two theories thus called Full Transfer/Full Access (Schwartz & Sprouse, 1994; 1996) and Partial Access (Tsimpli & Roussou, 1991) probably can account for the results of the present study.

(1) a. Zhang San ai Li Si / ta. In (1b), ta can refer to Zhang San (i) or someone else in the context (k). But ta cannot refer to Li Si (j) because it must be free in the local clause headed by the verb piping (criticize). Additionally, Mandarin Chinese is not an inflectional language, so Mandarin Chinese does not have case markers (Li and Thompson, 1989). The same personal pronoun is used to represent both the subject and the object.

Reflexives in Chinese
Reflexives in Chinese can either be morphologically complex/polymorphemic or morphologically simple/monomorphemic (Kong, 2009). ta ziji is a polymorphemic anaphor and ziji is a monomorphemic anaphor. According to Huang et al (2009), the behavior of the Chinese compound reflexive, ta ziji, is quite similar to the English reflexive.
(2) Zhang Sani zhidao Li Sij piping ta ziji*i/j. ta ziji in (2) can only refer to Li Si(i) but not Zhang San(i) for the reason that Binding Principle A (Chomsky, 1981) states that reflexives must be locally bound. Therefore, Zhang San(i) cannot bind ta ziji because it is in the upper clause. The monomorphemic anaphor, ziji, on the other hand, can be bound locally or long-distantly (Huang et. al., 2009). This is illustrated in (3a): (3) a. Zhang Sani zhidao Li Sij piping zijii/j. ziji can be bound locally by Li Si(j) or long-distantly by Zhang San(i). Besides, the monomorphemic anaphor ziji can play an intensifying role, locating at a non-argumentation position, as shown in (3b). Yet, monomorphemic anaphor ziji is not the main issue in this study

Pronouns in English
Pronouns are a closed class of words (Carter & McCarthy, 2006) that are one of the most commonly seen in the English language (Balogh, 2003). English pronouns, like their Chinese counterparts, must be free in their governing domain. They can be bound, however, by a constituent in a different clause, as illustrated in (4a) and (4b): (4) a. Wendyi hates her*i/j/.
b. Bettyi knows that Wendyj hates heri/*j/k. In (4a) and (4b), her(i) cannot refer to Wendy(j) of the same clause. However, her can refer to others out of the sentence in (4a), or it can refer to Betty of the upper clause within the sentence in (4b) as its antecedent. The syntactic behavior of English pronouns is similar to that of Chinese pronouns.

Reflexives in English
Binding Principle A points out that an anaphor must be bound within its governing domain; that is, it must take a local subject as its antecedent. English is no exception, as examplified in (5a) and 5(b): (5) a. Oliveri hit himselfi.
b. Oliveri thought that Wilsonj hit himself*i/j. c. * Oliver hates self./ * Oliver thought that Wilson hates self.
In (5a), the English reflexive himself must refer to the subject Oliver of the same clause; similarly, in (5b), the English reflexive himself in (5b) must be bound by Wilson of the same clause. The subject Oliver of the upper clause cannot be its antecedent. However, English bare reflexive self syntactically differs from Chinese ziji in that it cannot be an anaphor in (5c), contra to (3a).

Summary
To sum up, English pronouns and Chinese pronouns are similar in the way that they must be free in their governing domain. The English reflexives are very similar to the Chinese polymorphemic/complex reflexives, allowing local anaphors only. What differentiates the two languages is the monomorphemic/simple reflexive, which exists in Chinese but not in English. Chinese complex reflexives allow local and long-distance antecedents. The present study is organised as follows: Section 2 briefly reviews three prior studies concerning the acquisitions of pronouns and reflexives in English, Chinese/Cantonese. Section 3 demonstrates my experiments on second language acquisition on pronouns and reflexives in English and Chinese. Finally, I discuss the study.

Previous studies
According to the previous studies on the Second Language Acquisition (SLA), English pronouns and reflexives have been intensively studied. In the following section, I briefly review three studies which greatly concern about Chinese speakers of L2 English.
2.1 Yip andTang (1998) Yip andTang (1998) report in details on the interpretations of English reflexively him/herself by the Cantonese learners of English based on the transfer hypothesis, specially Positive Transfer Hypothesis. The study, based on the three types of sentence (finite clauses, non-finite clauses and single-clause dative constructions. posited two problematic issues: (I) Cantonese speakers did not consistently treat keoizigei 'him/herself' as a local anaphoric; (ii) many advanced Cantonese speakers of L2 English, acquiring local binding of English reflexives, allowed keoizigei to bind a long-distance antecedent. Additionally, much advanced learners were able to treat the binding properties of L2 as an independent system, without any transfer from their native speakers. Moreover, this study provides Logical Form to account for the findings. Yip and Tang also pointed out that the local binding of English reflexively is readily learnable while others like subject orientation are less so, compared to Yuan (1998). However, Yip and Tang provided specific explanations to the findings, but they did not test English pronouns and Cantonese pronouns in order to prove the Positive Transfer Hypothesis.
2.2 Yuan (1997) Yuan (1997) only investigates the acquisition of Chinese reflexive ziji by English and Japanese learners. Unlike English reflexive, Chinese ziji and Japanese jibun allow a long-distance antecedent as well as a local one. The observations thus showed that it is easier for Japanese learners than for English learners in acquiring the literature no-distance binding of Chinese ziji. Additionally, it is asymmetry in finite and nonfinite clauses in English speakers' grammars of Chinese. Yuan also points out that both English and Japanese learners have trouble in acquiring subject orientation of Chinese ziji and cannot find out the implication relationship between long-distance and subject orientation. There are two examples of the null subject and the null objects of pronouns to be illustrated below: (6) Wo wen ta Li Mingi jintian lai bu lai, ta shuo ei yiding hui lai.
I ask him Li Ming today come not come he say certainly will come ' * I ask him whether Li Ming will come or not, he says e will certainly come.'

(7) Zhang San shuo [e bu renshi e].
Zhang San say not know ' * Zhang San said e didn't know e.'/'Zhang San said (he) didn't know (him).' The findings have shown the more advanced groups show that they successfully acquire the non-null subjects in English. Besides to null subjects, in the structures of null objects, the results of Chinese learners of L2 English compare to the results of the group native speakers with great gaps, meaning that Chinese learners of L2 English do not successfully acquire that English does not allow null objects either. In brief, in acquiring the long-distance properties of Chinese ziji, Japanese speakers are better than English learners since the properties of Japanese jibun are parallel to those of Chinese ziji. Although Yuan investigated both English and Japanese learners on Chinese bare reflexive ziji in various syntactic structures, he did not further investigate whether both English and Japanese learners can well acquire Chinese disyllabic reflexive ta ziji 'him/herself'.
2.3 Kong (2005Kong ( , 2009) Kong (2005Kong ( , 2009 reports that the L2 learners performed native-like in acquiring the locality of complex reflexive ta ziji 'him/herself' whereas they could acquire locality condition of bare reflexive ziji 'self'. Two points designed in Kong's study: one is to test Yuan's hypothesis that Chinese speakers have less trouble unlearning null subjects than null objects based on the assumption that the recognition of S-V Agreement features triggers their unlearning of null subjects while transfer of L1 [+ topic-drop] property causes them problems in unlearning null subjects. The other is to include some syntactic items. For example, non-subject topics in clause initial position which is described and concluded in the section 6, Discussion part. As for native speakers, monosyllabic ziji can be locally and long-distantly bound. According to his results of experiments, Kong concludes that the divergence may be an effect of the inaccessibility of some critical period-associated functional features (Hawkins & Chan, 1997), and what appears to be a parameter resetting scenario is an adjustment of f L2 setting to match with the L1 setting. Therefore, he further claims that the monomorphemic ziji may have been misanalysed as pronouns in English.

Interim summary
To sum up, these three studies put much attention on acquisitions of English reflexives by L1 Chinese speakers, and provide as many sentence types as possible to investigate how Chinese speakers as L2 English acquire English reflexives. Besides three prior studies, there are also many studies to test the acquisitions of binding English reflexives for native speakers of Chinese, and also to study the acquisitions of non-null subjects and objects languages, like English language, of Chinese speaker, which Chinese is the null subjects and objects language. However, there are fewer studies to the acquisitions on binding of English pronouns. Therefore, I am going to focus on the studies of binding on English reflexives and pronouns to test Chinese learners as L2 English whether to acquire the binding principles or not.
There are two research questions that I am going to discuss in the current study: one is that will L1 be a factor in the acquisition of English reflexives and pronouns; the other one is that will advanced learners perform like native speakers in the interpretation of English reflexives and pronouns. The results of my experiments are shown as follows. Based on the results, this study proposes the language transfers (positive/negative transfer) to account for my observations.

Participants
Forty participants took part in the study. Among them thirty were students from Tunghai University in Taiwan. The other ten were native speaker at the same university. All participants were above 18 and the thirty students were divided into three groups based on their scores on a proficiency test administrated by Tunghai University's English Language Center.

Task and scoring
Twenty-four randomized multiple-choice questions made up the task. The twenty four sentences were generated from six sentence-types and each type consisted of four sentences. They were used to test Taiwan speakers of L2 English's interpretations of English reflexives (himself/herself) and English pronouns (him/her). Table 4 illustrates the samples of the six sentences types. After each sentence, three questions followed with either Yes or No answer attached. Learners were asked to choose either one of the two answers to which they thought the reflexive or the pronoun referred, as shown in (8) and (9) In Type 1 (Example 8), the upper clause subject Tom is the only correct binder of the pronoun him. If a "Yes" is chosen for question (i), the learner gets 1. If a "No" is chosen, the learner gets 0. In question (ii), 1 is given if the learner choose "No" and 0 if the learner chooses "Yes". Likewise in question (iii), if "No" is chosen, the learner is scored 1 but 0 if the learner chooses "Yes". In (9), The only correct binder of the reflexive herself is the local subject Claire. A learner will get 1 for choosing "No" for (i), "Yes" for (ii) and "No" for (iii) respectively. "Yes" for (i), "No" for (ii) and "Yes" for (iii) will give the learner 0. Learners' performance was calculated individually and as a subgroup for their interpretation of the pronouns and reflexives concerned. Average scores were then calculated.

Procedure
The participants took the comprehension test separately. The three experimental groups took the test in class during a regular class period. Participants were stimulated not to discuss or check answers with each other. The test was not timed. Participants in control group took the test separately. Table 3 shows that all groups performed native-like in accepting the non-local binder for the pronoun (him/her) than local antecedents. All groups scored above 85%. They all disallow the pronouns (him/her) to be locally bound. It indicates that the property of English pronoun is quite similar to Chinese pronoun, and it shows that learners have no problems in acquiring English pronouns in finite clauses.  Table 4 shows that the experimental groups prefer long-distance antecedents to local antecedents of him/her in English. However, Group 1 shows that there are higher scores than Group 2 and 3, and some participants in Group 2 and 3 accept local binding (both are 5%) and both (both are 8%), long-distance and local antecedents in infinite clauses. By contrast, native speaker of English accept long-distance antecedents only in infinite clauses.

him/her in infinite clauses
Therefore, Group 1 with advanced English proficiency greatly acquire English pronouns in infinite clauses whereas Group 2 and 3 may be influenced by the property of Chinese pronouns.  Table 5 shows that the experimental Group 1 and 3 more accept long-distance antecedents of him/her than local antecedents in dative clauses; the results of Group 2 show 45 % in long-distance antecedents, and 40% in local antecedents, which means participants select either long-distance or local antecedents to refer to the pronouns in dative clauses. By contrast, native speakers of English show that they accept both long-distance and local antecedents, more than select local antecedents only. That is, the results of three experimental groups seem to have more problems in acquiring English pronouns binding in dative clauses than in finite and infinite clauses. This study infers that Chinese pronouns in dative constructions prefer to a long-distance interpretation.

himself/herself in finite clauses
According to Binding Principle, Chomsky (1981) points that Binding Principle A states that an anaphor must be bound in its domain; that is. It must take local subjects as its antecedents. Table 6 shows the scores of three experimental groups increase gradually to prove that they acquire the local binding of English reflexives as long as they expose to more English.
Group1 seems to still have problems in select local antecedents of English reflexives in finite clauses. To compare to native speakers, they only accept local subjects as antecedents in finite clauses. Even if most of participants of Group 3 know to choose local antecedents (25%), there are some to select both (25%), local and long-distance anaphors.

himself/herself in infinite clauses
In infinite clauses, table 7 shows that Group 2 and 3 also acquire to take local anaphors in infinite clauses whereas the score of Group 2 in selecting local antecedents is higher than Group 3's by 5% and is closer to Group 4's (native speakers'). The score of Group 1 in selecting local antecedents is higher than in choosing long-distance anaphors, but the distance to prove that Chinese learners of L2 English successfully acquire English reflexives binding is not oblivious. 3.4.6 himself/herself in dative clauses ① Table 8 shows that three experimental groups do not perform well in dative clauses of English reflexives. By contrast with Group 4, native speakers of English accept long-distance anaphors in dative clauses. Thus, the scores of Group 2 and Group 3 show that they also prefer to select long-distance antecedents. For Group 1, the scores of selecting long-distance and local anaphors are not quite oblivious to show they allow long-distance anaphors or local antecedents in English dative clauses. Similar to the result ① The both results in section 3.4.3 and section 3.4.6 have shown that L2 English learners have much trouble in acquiring dative clauses of English pronouns and reflexives. Thus, the on-going study expands this research to collect more tests of participants from other English departments, and the study will run R and ANNOVA to investigate whether there is any significant to account for the asymmetry results in that the current is a primary step to investigate how L2 English learners in acquiring English pronouns and reflexives. (This study thanks to one of anonymous review to mention the reliable test). of table 5, the L2 English learners on English reflexives in dative clauses cannot syntactically behaves like what native speakers do.

Interim summary
In Chinese reflexives, there are two types which are pointed in section 2, monomorphemic/simple, ziji, and polymorphemic/complex reflexives, ta ziji. On Chinese reflexives, the behavior of the compound reflexives, ta ziji, is quite similar to English; that is, it have to take a local subject as its antecedent. On the contrary, the behavior of the bare reflexives, ziji, can be bound in its local antecedent, or it can also be free to its long-distance antecedent.
Because of Chinese bare reflexive, ziji, it seems there is an influence, an obstacle, for Chinese speakers of L2 English learning English reflexives as illustrated in table 8, 9, and 10. Particularly, Chinese speakers learning English reflexives cannot well perform in infinite clauses and dative clauses. I discuss the analysis in the following section.

Discussion
As far as the interpretation of English reflexives and pronouns is concerned, the results in the current study have shown that Chinese speakers of L2 English have less trouble acquiring reflexives and pronouns in finite and non-finite sentences than in dative sentences. Advanced learners perform much better than their lower proficient counterparts whereas still have problem acquiring reflexives and pronouns in dative sentences. The asymmetry in advanced learners' interpretation of finite/non-finite and dative sentences provides an answer to my second research question which asks if advanced learners can acquire the language form in question. It seems that they can acquire reflexives and pronouns in finite and non-finte sentences but not in dative sentences. The question to ask now is if L1 transfer can account for such an asymmetry. Recall that our first research question is whether or not L1 transfer plays a crucial role in L2 development.
It is evident that L1 transfer influences learners' interpretation of the sentences, given the similarities the two languages share in reflexives and pronouns discussed in section 1. However, L1 transfer alone is inadequate to account for finite/non-finte and dative sentences asymmetry. Dative sentences exist in Chinese as well as in English. Why would transfer occur in finite/nonfinite but not in dative sentences in advanced learners' L2 grammar? Two influential theories may offer an explanation to the observed behavior in the study. But because of the pilot nature of the study, we will review the essence of two theories and discuss their explainabilities to future study.
One of two theories is called Full Transfer and Full Access (FT/FA) of Schwartz andSprouse (1994, 1996). The theory assumes full transfer of L1 grammar in initial state of L2 acquisition. Target language grammar is acquirable provided that there is sufficient in the L2 input to inform the learner that L1 and L2 are different. Based on the input, the learner will be able to acquire the target grammar. The other theory is proposed by Tsimpli and Roussou (1991) called the Partial Access to Universal Grammar hypothesis. Similar to the FT/FA account, it stipulates full transfer of L1 grammar in the early stages of L2 acquisition. However, it differs from the FT/FA hypothesis in the sense that it assumes some functional category-related properties be inaccessible to adult L2 learners. This is because those properties are parametric variant between languages and are inaccessible beyond the critical period. In other words, adult L2 learners will never be able to acquire some L2 properties like native speakers do.
The two theories can provide an answer to the observed behavior in the study. But we will leave this to the future study. What needs to be considered in the future study is the task design. A subtle testing of a wider range of properties in relation to reflexives and pronouns is also required.