From the discussion above, we found some evidence of L1 effects on the L2A by EFL learners. Then we would like to consider the issue whether EFL learners reset their L1 value of null subject parameter to the L2 value of null subject parameter. As we mentioned in chapter 2, there are three ways of acquisition: Full Access to UG, Partial Access to UG, and No Access to UG. Full Access to UG seems to be the most questionable in this study since EFL learners, even the advanced learners, still have some problems in acquiring some English features, such as obligatory objects and expletive pronouns. No Access to UG is also excluded here since the developmental progress between the low and high achievers may imply they are not entirely free from the influence of UG. Therefore, we turn to the position of the Partial Access to UG and see whether the parameter resetting is possible.
As explored in Kong’s (2005) study, matrix null subjects and embedded null subjects were treated differently by his informants, by which Kong suggests his learners didn’t reset their L1 parameter since if they did, they should perform no difference to null subjects in matrix clauses or in embedded clauses. In our study, we also found the different judgments of null subjects between the two sentence structures, but diverse from Kong’s study, we also put the factor of animacy into our consideration. For the inanimate part, we found that both learner groups have more difficulty in detecting null subjects in the embedded clause than matrix clause, the same result as in the previous studies; however, the animate part shows the reverse result. When we compare each sentence, we found that the sentence in (69a) shows the parallel structure. If we recall the data in the ST task, we can find that our learners tend to drop subjects in the parallel sentence structure, as in (70). Therefore, it may be the reason why (69a) was more acceptable than (69b), while sentences in (68) with no parallel structure also conform to the predicted result.
(68) Inanimate Subjects:
a. Jane bought a pair of shoes. __ Were very beautiful.
b. Judy likes the white table, but she thinks __ is too small. (accept more) (69) Animate Subjects:
a. Lisa [got up very late.] __ [Didn’t go to school on time.] (accept more) b. Mary likes to go shopping. I believe __ will like this department store.
(70) In ST task:
a. So, the 巫婆 the witch [give her…mm…make his (her) dream come true]…___ [let he (her)…let she (her) go to the party.]
b. And a… witches (witch) show(ed) up, she… [help(ed)…. her], ___ [gave her a beautiful dress.]
Although our results in this point show a bit different from other studies, we still can observe null subjects are treated differently between the matrix and embedded clauses, and apart from null subjects, null expletives also show this kind of pattern.
From the results, parameter resetting seems not to be workable.
Besides, from our data in the ST task, parameter resetting is also not evident since there were still many non-native uses of null subjects in spoken discourse by the EFL learners. Williams (1989) indicated that in spoken discourse native speakers may allow null subjects in the following two descriptive norms (Braidi, 1999 cited form William (1989: 340)):
(71) a. Subject pronouns may be omitted when two clauses represent a
coordinate relationship but without the requisite coordinating conjunction (and, but, or or) present.
e.g. He just walked into the crossfire. Ø Never knew what hit him.
b. Subject pronouns may be omitted when they can be clearly identified by the immediate situational context.
e.g. (while attending a lecture) Ø Sure knows his stuff.
However, in our data the learners produced not only sentences as the norms above but also sentences which are inconsistent with native-speaker use, as shown in (72).
(72) a. Few days later the prince… [want…to find her], so…___ [require his soldier…to find the Cinderella] and…mm…prince find the Cinderella (laugh). (LP)
b. And… she [really… want(ed) to go to the party], so ___ [can't (couldn't) help stop crying.] (HP)
c. And so he, mmm, so [he arranged a party and a, mmm, invite(d) all the girls in the kingdoms (kingdom) to at…attend the party] and [maybe ___
have the chance to dance with the prince]. (HP)
It is not native-like that null subjects appeared in the sentences with coordinate conjunction ‘so’, as in (72a,b) or in the sentences which the first referent was an object in its clause, and the null subject occurred as a subject in an adjacent clause, as in (72c). Although L2 learners broke the native-speaker norms, the reference relationship was clear from the context, which may suggest they are influenced by the discourse-oriented feature in Chinese.
Concluded from the above, the parameter resetting seems impossible; therefore, our research question about whether the clustered properties, missing subjects, that-trace effects and S-V inversion, may reset together is rejected. But we should
notice an interesting phenomenon that Taiwan EFL learners have difficulty dealing with sentenced with that-trace effects, which may toss a question: whether that-trace effects is included in Chinese pro-drop parameter? Findings in the GJ task and the PT task indicate that the EFL learners, even the HP learners, performed considerably worse at this kind of sentence pattern. As we mentioned before, the existence of complementizer in Chinese has been debated. Does it suggest in Chinese grammar there is no complementizer corresponding to that in English? Or if Chinese had this feature, could it mean the resetting of this feature is difficult, or even impossible? This
problem may be explored in future studies.