The second research question of the present study examines the difficulty hierarchy of the types of English it-clefts, i.e., NP, PP, Zero, FIN CL, N-F CL, ADV P, and ADJ P, as (5) to (11) show respectively3.
(5) It was last week that the superstar came to Taiwan. (NP) (6) It is for my high school teacher that I bought this card. (PP) (7) It may be that the girl will be sent away. (Zero)
(8) It is that she already has a date with others that is her answer. (FIN CL) (9) It is to ask him to sweep that is never easy. (N-F CL)
(10) It is then that they met. (ADV P) (11) It is big that this house is. (ADJ P)
The results are presented and discussed in the following subsections.
4.2.1 Overall Findings
As Table 4-4 shows, both the descriptive statistics and the results of one-way ANOVA for all the subjects on the types of English it-clefts can be seen below.
3 Sentences (5) to (11) are repeated from sentences (48), (54), (59), (62), (70), (75), and (78) in Chapter Two respectively.
Table 4-4 The Overall Results for the Seven Types of English It-clefts
Type M SD F
p-value
1. NP 3.79 0.62 23.402 .000
2. PP 3.65 0.71
3. Zero 3.69 0.71
4. FIN CL 3.10 0.71
5. N-F CL 3.44 0.66
6. ADV P 3.33 0.73
7. ADJ P 2.62 1.05
There was a significant difference among the seven types of English it-clefts (F (6, 553) = 23.402, p < .05). Besides, the LSD post hoc analysis showed that our subjects gave significantly higher scores on NP and Zero than on FIN CL, N-F CL, ADV P, and ADJ P. Still, they significantly scored higher on PP than on FIN CL, ADV P, and ADJ P. Also, they scored significantly higher on N-F CL than on FIN CL and ADJ P, and the scores on FIN CL and ADV P were significantly higher than those on ADJ P.
Accordingly, a difficulty hierarchy for the seven types of English it-clefts was found:
(12) NP (3.79) / Zero (3.69) / PP (3.65) > N-F CL (3.44) / ADV P (3.33) >
FIN CL (3.10) >ADJ P (2.62)
Table 4-5 presents the within-group differences among the seven types of English it-clefts.
Table 4-5 A Comparison among the Seven Types of English It-clefts within Each Group
L I H NS
F .711 1.859 5.503 70.130
p-value
.641 .093 .000 .000As Table 4-5 shows, while there was a significant difference among the seven types of English it-clefts in the H group (F (6, 133) = 5.503, p < .05) and the NS group (F (6, 133) = 70.130, p < .05), no significant difference was found in the L group (F (6, 133)
= .711, p > .05) nor the I group (F (6, 133) = 1.859, p > .05), according to the results of one-way ANOVA. Besides, the LSD post hoc analysis showed that the H group significantly scored higher on NP than on FIN CL, N-F CL, and ADV P, and significantly scored lower on ADJ P than on PP, Zero, N-F CL, and ADV P, and significantly scored higher on Zero than on FIN CL. Thus, a difficulty hierarchy for the H group can be seen as follows:
(13) NP (4.15) / Zero (3.79) / PP (3.72) > ADV P (3.65) / N-F CL (3.63) >
FIN CL (3.28) >ADJ P (2.91)
As for the NS group, they gave significantly higher scores on NP, PP, and Zero than for FIN CL, N-F CL, and ADV P, and they significantly scored lower on FIN CL than on N-F CL and ADV P, and significantly scored lower on ADJ P than on all the other types. Likewise, a difficulty hierarchy for the NS group’s preference is as follows:
(14) Zero (4.11) / PP (3.96) / NP (3.83) > N-F CL (3.25) / ADV P (3.14) >
FIN CL (2.56) > ADJ P (1.33)
Figure 4-2 shows the overall scores on the seven types of English it-clefts for each proficiency group.
Figure 4-2 Overall Performances of Each Group on the Seven Types of English
It-clefts
As shown in Figure 4-2, the L group scored above average on FIN CL type (M: 3.40 >
3.10), N-F CL type (M: 3.45 > 3.44) and ADJ P type (M: 3.16 > 2.62) of English
it-clefts, and their overall performance (i.e., L: NP (3.56) > PP (3.51) > N-F CL (3.45)
> FIN CL (3.40) > Zero (3.34) > ADV P (3.25) > ADJ P (3.16)) further showed that NP was the most acceptable type, while ADJ P was the most unacceptable type for them, although the scores on the seven types were quite close. Besides, the I group scored above average on FIN CL type (M: 3.17 > 3.10) and ADJ P type (M: 3.09 >
2.62) of English it-clefts, and most of its mean scores were higher than those of the L group, except for FIN CL type, N-F type, and ADJ P type. As for the overall
performance of the I group (i.e., I: NP (3.63) > Zero (3.51) > N-F CL (3.43) > PP (3.42) > ADV P (3.26) > FIN CL (3.17) > ADJ P (3.09)), it indicated similar results with the L group, i.e., there was no big difference among the seven types, NP was the most acceptable type, and ADJ P was the least acceptable for them. The performances of the H group on the seven types of English it-clefts were all above average, and almost all of their scores were higher than those of the lower groups, except the types of FIN CL and ADJ P. Besides, the overall performances of the H group as in (13) showed that NP was the most acceptable type and ADJ P was the most unacceptable for them, which was similar with the L and I groups. Last, our native controls only gave scores above average to the types of NP, PP and Zero. In addition, their overall performances as in (14) indicated that the most acceptable type for them was Zero type, rather than NP type. They also considered ADJ P type the most unacceptable type as all the other proficiency groups.
Table 4-6 is a comparison among groups within each type of English it-clefts.
Except for PP, N-F CL, and ADV P, there were significant differences in the other four types of English it-clefts.
Table 4-6 A Comparison among Groups within Each Type
NP PP Zero FIN CL N-F CL ADV P ADJ P
F 3.984 2.426 5.209 6.809 1.074 1.970 27.456
p-value
.011 .072 .003 .000 .365 .126 .000What follows are the detailed results of the LSD post hoc analysis. For NP type, it was found that the H group gave significantly higher scores compared with the L group and the I group. As for Zero type, the scores the H group gave were significantly higher than those the L group gave, while those given by the NS groups were significantly higher than those given by almost all of the groups, except for
those given by the H group. Regarding FIN CL type as well as ADJ P type, the scores given by the NS group were significantly lower than those given by all the other groups.
4.2.2 General Discussion
In this section, the findings reported in Section 4.2.1 are discussed. First of all, a significant difference in statistics was found in the seven types of English it-clefts, including NP, PP, Zero, FIN CL, N-F CL, ADV P, and ADJ P, showing that our subjects’ acceptance of the English it-clefts did vary with types. The difficulty hierarchy shown in the previous section is repeated below:
(15) NP (3.79) / Zero (3.69) / PP (3.65) > N-F CL (3.44) / ADV P (3.33) >
FIN CL (3.10) > ADJ P (2.62)
In Collins’ study on the spoken Lancaster-Oslo/Bergen Corpus and the written London-Lund Corpus (Collins 1991, p.56), the percentage of each type of English
it-clefts is as follows:
(16) NP (50.4%) > PP (21.5%) > Zero (14.5%) > FIN CL (6.8%) >
ADV P (5.9%) > N-F CL (0.8%) > ADJ P (0.1%) The most acceptable English it-clefts in our study were the types of NP, PP, and Zero,
and they were also the types with the highest percentage in Collins’ study (NP: 50.4%, PP: 21.5%, Zero: 14.5%). In addition, N-F CL and ADV P were found the next acceptable, followed by FIN CL. This, however, did not completely agree with Collins’ study, where FIN CL (6.8%) had the most percentage compared with the other two, i.e., N-F CL (0.8%) and ADV P (5.9%). In our study, on the other hand, there was a drop in FIN CL, and a rise in N-F CL. Last, the least acceptable type in our study was the type of ADJ P, and this conforms to the result of Collins as well (ADJ P: 0.1%). In a nutshell, the results obtained in the present study in general had a
similar pattern with Collins’ investigation on the two corpuses, indicating that how much our subjects accepted the type of English it-clefts was somewhat corresponding to how much they existed in the corpuses.
To further discuss the difficulty hierarchy we obtained, we may have a closer look at the features of each type. It was found that the types of NP, PP, and Zero got higher scores than the types of FIN CL and N-F CL, implying that the categorical types, i.e., the former three types, had a better acceptance rate than the clausal types.
This might be because the clausal types of clefts are often lengthy, compared with the categorical types, and thus they are difficult for the L2 learners to process, as compared in the following4.
(17) Categorical Type: It is Christ who lives in me.
(18) Clausal Type: It is in order to improve the quality of life for their children that they work so hard every day.
In (17), the clefted element “Christ” is a noun phrase. In (18), however, the clefted clause “in order to improve the quality of life for their children”, which is much larger in number than the clefted syntactic category in (17). Thus, it is likely that the dissimilarity of acceptance between the categorical type and clausal type of clefts results from the length of cleft sentences in English. Nevertheless, for ADV P and ADJ P, which also carry categorical feature, they did not get higher scores than the clausal types. This might be due to the fact that their frequency rates in everyday conversations are so low that their acceptability decreased. In addition, for the clausal types of FIN CL and N-F CL, there was a higher acceptance rate for N-F CL than for FIN CL. One possible reason might be the dual that in the type of FIN CL most of the
4 The sentences in (17) and (18) are excerpted from questions (1) and (66) in the Acceptability Judgment Task as shown in Appendix A.
time, as shown in the following5.
(19) It is that everyone should promote the protection of the environment that is the main claim of the professor.
Such a cleft sentence causes a garden-path effect, which leads to the ambiguity of the sentence and thus lowers acceptance rate for the subjects (Pritchett 1988).
Moreover, the statistics showed that while the acquisition of English it-clefts varied with the types for the H group and the NS group, it did not for the L group or the I group. For the L group and the LI group, the indifference in the seven types of English it-clefts might be due to the degree of markedness of the structure (Eckman 1977, 1985). The structure of English it-clefts is so marked to the L and I groups that they could not distinguish the differences of the seven types of English it-clefts.
However, both groups still gave high scores to NP and low scores to ADJ P, which accorded with the H group and the NS group. As for the H group and the NS group, their difficulty orders are repeated as follows:
(20) H group: NP (4.15) / Zero (3.79) / PP (3.72) > ADV P(3.65) / N-F CL (3.63)
> FIN CL (3.28) > ADJ P (2.91)
(21) NS group: Zero (4.11) / PP (3.96) / NP (3.83) > N-F CL (3.25) / ADV P (3.14) > FIN CL (2.56) > ADJ P (1.33)
It was found that our H group had already had a similar preference with the NS group, i.e., the scores on NP, Zero, and PP were higher than those on ADV P and N-F CL;
those on ADV P and N-F CL were higher than those on FIN CL; those on FIN CL were higher than those on ADJ P.
With regard to the performances of the groups in each type of English it-clefts, while the learners’ groups had similar perception with the NS group on the types of
5 Sentence (19) is excerpted from question (14) in the Acceptability Judgment Task as shown in Appendix A.
NP, PP, N-F CL, and ADV P, they failed to perform native-like on the types of Zero, FIN CL, and ADJ P. For Zero type, the L2 groups seemed to have difficulty accepting it. This might be due to negative transfer effects, i.e., such type of clefts cannot occur in Chinese. On the other hand, the L2 groups accepted the types of FIN CL and ADJ P better, but the NS group did not accept these types. This implied that the L2 groups had not yet taken lengthy sentences as a disturbance as the natives had, nor did their acceptability conform to the frequency rate in the corpus. Besides, looking at the distribution of the mean scores of each proficiency group, the I group in general gave higher scores on each type of English it-clefts than the L group did, except for the less acceptable types of FIN CL, N-F, and ADJ P. Besides, the H group gave higher scores than the L group and the I group, except for the types of FIN CL and ADJ P, the less acceptable types as well. That is to say, while the Chinese EFL learners’ English proficiency improved, the better they could accept the structure of English it-clefts.