For the first and second research questions on forward transfer and backward transfer with respect to L2 proficiency, the key factors in explaining the Chinese learners’ performance on both English and Chinese tasks seem to be the ideas of restructuring and task effect. Therefore, a close examination of the Chinese learners’ L1 and L2 performance data at these two different task types may shed some light on the learners’ SLA process. As indicated above, there is a negative L2 effect on the Level 2 Chinese learners’ use of because-initial information sequencing in the Chinese SCT. It is further suggested that what underlies this negative L2 effect may have been the backward transfer (or L1 attrition) phenomenon. This phenomenon may have been related to the Level 2 Chinese learners’ restructuring
process as well as to the task effect. In what follows, the Chinese learners’ L1 performance data is examined to further support the interpretations suggested by the present author based on SLA theories reviewed in the preceding sections. The results of the DTs in the present study are examined first, since they are consistent with pervious research findings regarding L1 transfer and L2 proficiency, whereas results of the SCTs appear to be different drastically from previous research findings.
5.3.1 L1 and L2 data comparison in the discourse tasks
Table 12 displays the comparative performance data for the Chinese learners in their use of L1 native pattern of because-initial structure in the Chinese DT and English DT.
Table 12. The Chinese ESL group’s use of because-initial structure in the Chinese DT and the English DT across the three proficiency levels
Chinese The Chinese DT The English DT
Speakers Relative
A comparison of the Chinese learners’ L1 Chinese and L2 English performance data for because-initial structure on the DT, as displayed in Table 12, indicates that their L1 and L2 linguistic performances are consistent with previous research findings.
First, in the Chinese DT, the Chinese learners at all proficiency levels show a strong preference for using because-initial structure: 87.1% at Level 1, 99.33% at Level 2, and 82.24% at Level 3. This finding lends strong support to previous research claims that because-initial structure is a fundamental principle for information sequencing in Chinese discourse (Kirkpatrick 1993). Moreover, within the English DT, the frequency with which the Chinese learners use the non-target-like L2 marked pattern of because-initial structure exhibits a downward linear curve:
69.57% at the Level 1, 55% at Level 2, and 27.78% at Level 3. This supports general research claims that there is a negative relationship between L1 transfer and L2 proficiency at the discourse level (e.g. Takahashi and Dufon 1989, Maeshiba, Yoshinaga, Kasper and Ross 1996).
5.3.2 L1 and L2 data comparison in the sentence tasks
Table 13 displays the comparative performance data for the Chinese learners in their use of the L1 pattern of because-initial structure in the Chinese SCT and English SCT.
Table 13. The Chinese ESL group’s use of because-initial structure in the Chinese SCT and the English SCT across three proficiency levels
Chinese Chinese SCT English SCT
Speakers Relative
A comparison of Chinese learners’ L1 Chinese and L2 English performance data for because-initial structure at the sentence level reveals a similarity in the frequency with which they produce this Chinese L1-based discourse pattern. That is, both the L1 and L2 performance data exhibit a U-shaped behavior pattern, with the Level 2 Chinese learners producing the least of this L1 discourse patterns in both the Chinese SCT (29.21%) and English SCT (13.75%). This contradicts previous research claims for a negative relationship between (forward) transfer and L2 proficiency. On the other hand, however, the linguistic performance of both the Level 1 and Level 3 Chinese learners appear to confirm previous research claims.
That is, the Level 1 Chinese learners consistently exhibit a high frequency of use in the L1-based pattern in both the Chinese SCT (68.97%) and the English SCT (69.05%), confirming that the beginning learners adopt a primary L1 discourse strategy for both SCT tests. Additionally, the Level 3 Chinese learners produce a high frequency of use in the L1 discourse pattern in the Chinese SCT (64.84%), but they produce much less of this L1-based pattern in the English SCT (31.58%), confirming that the advanced learners are less vulnerable to L1 transfer in their L2 production. In contrast, the Level 2 Chinese learners not only show less preference for their L1 native discourse pattern in the Chinese SCT, but also show the least transfer effect on the English SCT, suggesting that the intermediate students are still in the unstable, interim stage of SLA.
Some explanations have been suggested for the unusual linguistic performance of the Level 2 Chinese learners in both Chinese and English SCTs. Their low frequency of use in because-initial structure in the Chinese SCT is likely to be
caused by a negative L2 effect, itself being a function of backward transfer from L2 to L1 that comes with these learners’ restructuring in their L1 system. Their high frequency of use in the target-like L2 unmarked pattern of because-medial structure in the English SCT may well have been a manifestation of overuse due to the process of restructuring in their L2 system.
Thus, a close examination of the L1 Chinese and L2 English performance of the Chinese learners at sentence level suggests a high degree of similarity between the two. Transfer of training and task effect are the two most likely determinants of overuse of the target-like L2 pattern of because-medial structure (found in the English SCT) and underuse of the native-like L1 pattern of because-initial structure (backward transfer found in the Chinese SCT). These two conditions may account for the Level 2 Chinese learners’ disparate linguistic behavior on the sentence and discourse tasks in L1 and L2 as displayed in Figure 613.
Transfer of Training
Condition Attention to Form（±） One
＋ － Sentence Task Discourse Task
(grammatical forms) (communication of ideas)
Condition restructuring Two
L1 System L2 System – L1 underuse – L2 overuse
+L1 underuse +L2 overuse
Figure 6. The relationship of L1 underuse and L2 overuse in relation to the task effect and transfer of training in the performance of the Level 2 Chinese learners
13 The author acknowledges his gratitude to one of the two reviewers, whose suggestion has contributed to this
6. Implications
Several implications have emerged from the present linguistic analysis of the written language of the Chinese ESL learners regarding because-initial information sequencing at the sentence and discourse levels and in both Chinese and English.