Table 4.10 shows the frequency and proportion of each non-null arguments and null pronouns. The total number of the pronoun features observed is 399, in which the null arguments(X) were far less than the non-null arguments(Y), indicating the
participants tend to produce obligatory subjects in this task. In addition to using overt subjects to co-index with the previous references, they also repeated previous mentioned common nouns or proper names, as in (44), which is unnatural in English.
Table 4.10 The frequency of the null and non-null arguments in ST task
Pronoun Features LP HP
X Y N X Y N
F % F % % F % F % %
subject 4 1 79 19.8 4.8 10 2.3 113 28.4 8.1
object 13 3.3 4 1 76.5 0 0 8 2 0
it-time 3 0.8 0 0 100 0 0 6 1.5 0
expletive
there 5 1.3 6 1.5 45.5 57.1
0 0 25 6.3 0
that-trace effect 0 0 0 0 0 0 0 0 0 0
V-S order / S-V order 2 0.5 6 1.5 25 0 0 13 3.3 0
Non-Subject Topic 3 0.8 0 0
Proper name 18 4.5 4 1
Common noun 38 9.5 39 9.8
Total 30 7.5 151 37.9 10 2.3 208 52.3 Note: N= X / X+Y (where X=number of instances of null subjects, and Y= numbers of cases where null subjects could have occurred in L1 but did not)
(44) a. The prince find(s) it, so the prince take(s) the shows (shoe), and, go(es) down, go(es) in the town to…to find her. (HP)
b. The bodyguard find the beautiful girl is that Cinderella…so the
bodyguard lead, leads, leads girl to the castle to meet their prince. (LP) c. Mm…a girl his (her) name is…Cinderella. Cinderella…is very very
beautiful. (LP)
As a whole, the LP group performed worse than the HP group in each null property (see N for the percentage of occurrence). The HP learners only dropped subjects (8.1%), while the LP learners produced sentences with not only missing pronouns but S-V inversion as well. All the learners showed no evidence for that-trace effects, which may result from the rare use of this sentence type or the avoidance of
troublesome aspects, as we mention earlier that EFL learners have long-term difficulty in learning this syntactic structure. Among the sentences with null pronouns, we found there appears to be some patterns in their utterances. Firstly, the learners tend to drop the subjects which can be referred to the objects in the preceding adjacent clauses, (neighboring object-subject reference) as the underline parts in (45).
(45) a. But, but at but in the midnight, Cinderella thought [the angel say to her]
[___ must in the twelve o’clock to come back.] (LP)
b. Mmm… unfortunately, [her father married a very wicked woman] and [___ also bringed (brought) two bad sister(s) into her family]. (HP)
Secondly, it is easier for them to drop the subjects in the parallel structures, as shown in (46). This appears to be related to the existence of topic structures in Chinese, which invisible subjects may be bound by a zero topic (Huang, 1984), as in (47).
(46) 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 she [started running, running home.] Accidentally ___ [left… one of her shoes in the palace.] (HP)
d. 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.] (HP)
e. Mmm, in the par, in the party, she..her enjoy, enjoy(ed) the party, and when the time is (was) close to twelve o'clock, ah, Cinderella [is (was) very… hurry], ___[want to go, want(ed) to go back.] (HP)
f. And a… witches (witch) show(ed) up, she… [help(ed)…. her], ___ [gave her a beautiful dress.] (HP)
(47) [TOP woi] ti gang chi quo fan, ti zai kan dianshi.
just eat EXP rice PROG watch TV
‘I just ate and now (I’m) watching TV.’
Thirdly, several LP learners tend to drop the variable objects controlled by verbs with similar meaning. For example, in (48a, b) a party is controlled by the verb join, as the null argument controlled by the verb join and go to. And in (48c) the verb find also controlled the dropped argument.
(48) a. One day, a prince a prince ask, ask every girls to join a party because she he want to ask to search a wife. The step mother and two daughters heard, heard the news, she they didn’t they didn’t want Cinderella to join___.
b. He (She) heard the news that about come, about…have a party, he (she) want, he (she) want to…joining ___ but his…he (she)…cannot go to___, so he (she) very bad… sad.
c. And press (prince) see the take the glasses to find the Cinderella. He find ___very long time, but one day the person tells the prince: Cinderella at that house.
Fourthly, they also dropped objects referring to the previous topic event, as the examples in (49).
(49) a. One day, the prince hold a party. The step mother knows___, she hopes her two daughters to join, but she doesn’t like Cinderella know___.
b. The Cinderella…want to take part in that party, so she…she prays all day long. And the fairies knows___.
Besides, in the data we also found several cases of null expletive it-time and there, as in (50) and (51).
(50) a. Cinderella see, look at the clock, ___already twelve, she is very hurry and told, told the prince, “I must go home and sorry”
b. So when ___ twelve o’clock, Cinderella will to go.
(51) a. Long long time ago, one town have a beautiful.
b. At long long time ago, a town at Europe have a girl brown (born) in the house
c. He (she) heard the news that about come, about…have a party
The learners tend to omit it along with be verb, which is similar to the results in PT task. As for null there, they used the verb have instead of there be, which suggests that they seem to be affected by L1 grammar since in English have and there be represent the same word you (possessive verb or existential verb) in Chinese. This phenomenon is interesting since if we recall the results in the GJ task, we can find that the LP group shows significant difference in detecting the experimental sentence with a null embedded expletive there, as in (52). In contrast with the HP and NS groups, they highly accepted this kind of sentence.
(52) Mindy thinks have hundreds of people in her concert.
In addition, the LP learners also produced sentences without expletive there, as in (53).
This indicates that learners still have difficulty in sentences with expletive there.
(53) a. And …___ appear a old woman to help him to help her and Cinderella can go to the party.
b. Long time ago, in the castle ___ live a girl whose name is Cinderella.
Moreover, the L1 effects can also be observed in the sentences with non-subject topics. In (54), the subject was dropped with a topic at the clause-initial position.
(54) a. One day, the house the white house Ø hold a dancer party.
b. One day, the castle Ø want, want to hold a party in order to prince, prince to find a girl to become his princess.
To summarize, according to the data in the ST task, we still found some evidence of missing subjects and S-V inversion produced by EFL learners. Although the pro-drop sentences were not plenty, especially few in the HP group, they do show some patterns that make pronouns easier to drop. According to the GJ and PT task, we found that some features related to pro-drop parameter are difficult for the EFL
learners to unlearn, such as null objects, non-subject topics, and expletive there. Then how can we account for these phenomena? In the next chapter, we will try to give explanations for the findings above to see whether the pro-drop cases in the data are related to the effects of L1 knowledge or UG, and whether the parameter resetting is possible to Taiwan EFL learners.
The results over the three tasks are summarized as in Table 4.11:
Table 4.11 Summary of the results in the three tasks
GJ(sig.) PT(%) ST(%)
Null Subject vs. Null Expletive (NS) (NE)
SV order vs. VS order With proficiency increases, learners show better performance on this feature.
That-trace effects
Both groups had difficulty dealing with this kind of sentence; they tend to consider that to be optional in English.
No occurrence
p<.05*
(“A>B” means A is more accepted/produced than B)
CHAPTER FIVE
DISCUSSION AND CONCLUSION