For the overall frequency distribution of three major hedging patterns (anticipatory it, abstract entity, human subject), the total number of each category is calculated. Table 4.1 summarizes the results in three categories across three different groups. L2 writers have the highest number of occurrence 1,699 times (211.32 per hundred thousand words). Native writers come in second with 1,111 times (136.67 per hundred thousand words). Journal writers have the least number. The total number of occurrences is 1,121 (133.72 per hundred thousand words).
The overall findings suggest that-clause complementation is a prevalent grammatical feature in academic writing as it is used widely in the three corpora. Both native and journal writers share similar results in terms of frequency. For anticipatory it clauses, native writers have 363 times out of the corpus size of 816,182 words (44.47 per hundred thousand words). Journal writers also have a similar finding.
There are 356 instances out of the corpus size of 836,168 words (41.71 per hundred thousand words). With human clauses, native writers have 466 times, and journal writers have 352 times (57.62 and 40.76 per hundred thousand words).
L2 writers, however, show twice as much the result as the other two groups. With the frequency of three patterns, there are more instances of hedges in L2 corpus than
51
in the other two corpora. Based on the results, it is possible that there are some particular overused and underused hedges which would need to be tested by a log-likelihood calculator, which would be explained further in the chapter.
Table 4.1 Total Frequency and Overall Distribution of Three Categories
Clause
Note: RF=Raw Frequency NF=Normalized Frequency per 100,000 words
52
4.2 Distribution of Verbal and Nonverbal Types
From the previous section, it is noted that the overall distribution of clauses are slightly different in L2 when compared to native and journal corpus. Regarding the distribution of different verbal and nonverbal types in different clauses, in this section the present study aims to answer it by classifying data into verbal and nonverbal to find out the proportion of verbal and nonverbal types in the three clauses. In this case, verbal part includes three types of verb. Nonverbal part includes adjectives and nouns.
4.2.1 Distribution of Verbal Types across Three Corpora
The present study categorizes verbs into three different categories: cognitive, sensory, and non-factive reporting verbs. Table 4.2, 4.3, 4.4 present the distribution of verb categories in different clause types.
a. Cognitive Verbs
Cognitive verbs refer to one’s mental processes or beliefs. From the three groups, L2 writers are found to have the highest frequency of cognitive verbs in anticipatory it clauses with 14.46 per hundred thousand words. Native and journal writers, however, do not seem to use as frequently as L2 writers do. Both of the two groups only attained around the frequency of 5.00 or 6.00 per hundred thousand words. That is to say, native and journal writers do not use as many cognitive verbs in anticipatory it clauses as L2 writers. Similar findings are also shown in abstract entity clauses. In comparison to native and journal corpus, L2 writers use twice as much as the other two groups. As the data presented below, the frequency is 5.35 per hundred thousand words for L2 writers, whereas only 1.47 and 1.80 per hundred thousand words are indicated in native and journal groups, respectively. With the third category—human clause, the data among the three groups do not seem to show as much contrast as the previous two clause types.
53
Table 4.2 Frequency of Cognitive Verbs in Clause Types
Clause Type L2 Native Journal
Anticipatory it 117 (14.46) 54 (6.63) 49 (5.87)
Abstract entity 44 (5.35) 12 (1.47) 29 (1.80)
Human clause 73 (8.85) 99 (12.37) 71 (7.68)
Total 234 (28.66) 165 (20.47) 149 (15.35)
b. Sensory Verbs
Sensory verbs do not occupy as much proportion as cognitive verbs which are shown similar in three groups. Since sensory verbs are generally associated the perceptions and feelings, this category is not used as frequently as the previous category in academic writing. Sensory verbs are found only in two clauses:
anticipatory it and human clauses. Nevertheless, L2 and journal writers seem to have employed slightly higher number of instances with 6.56 per hundred thousand words in L2 and 7.53 per hundred thousand words in native corpus. Journal writers achieved the highest number of sensory verbs which is probably because the use of sensory verbs in first-person clauses is used more frequently. Native writers lie somewhere in between the results from the other two groups. Further discussion regarding which sensory verb occupies the highest or the smallest proportion in the clauses will be further explained in the next section.
Table 4.3 Frequency of Sensory Verbs in Clause Types
Clause Type L2 Native Journal
Anticipatory it 54 (6.56) 44 (5.38) 63 (7.53)
Human clause 39 (3.99) 48 (6.11) 56 (6.60)
Total 93 (10.55) 92 (11.49) 119 (14.13)
54 c. Non-factive Reporting Verbs
Non-factive reporting verbs occupy the largest proportion in three clause types.
The difference is quite obvious among the three groups. The data suggests L2 writers have the highest number of instances in three clause types. The reason for being the highest out of the three groups could be attributable to the fact that L2 writers tend to rely on certain verbs which would lead to high frequency in the total number of results. Further details will be discussed in the following sections.
Table 4.4 Frequency of Non-factive Reporting Verbs in Clause Types
Type L2 Native Journal
Anticipatory it 267(32.43) 112(13.73) 119(14.24)
Abstract entity 534(64.86) 270(33.11) 413(49.45)
Human clause 536(69.56) 231(28.34) 202(23.61)
Total 1337 (166.85) 613 (75.18) 734 (87.30)
4.2.2 Distribution of Nonverbal Types across Three Corpora
In this current study, non-verbal types include adjectives and nouns. Following verbs, adjectives constitute the second largest proportion of multi-word hedges. L2 writers do not seem to have used as many adjectives as native and journal writers do.
The data reveals L2 group uses less than half of the number of adjectives than native and journal group. Similar tendency is also observed in noun category. Although native and journal writers do not seem to have used as many nouns as adjectives, both groups show they have the tendency of employing nouns in hedging constructions.
More information on the specific use of adjectives and nouns will also be delineated in the next section.
55
Table 4.5 Frequency of Nonverbal Types across Three Corpora
4.3 Distribution of Multi-word Hedges across Three Corpora
To address the second research question, this study first provides the top 25 multi-word hedge list to give an understanding of which ones are particularly preferred by one group or another. Following the presentation of differences in ranking, an analysis of different verb types which tend to collocate with certain subject types are explained.
4.3.1 Top 25 Multi-word Hedges across Three Corpora
Table 4.6 presents top 25 multi-word hedges across three corpora. The multi-word hedges are ranked in terms of frequency. Mostly, the hedging constructions employed in the three groups are in the clauses headed by abstract entity, anticipatory it clauses, or human subject clauses. With abstract entity clauses, three groups consistently employ verbs suggest, show, indicate as in the results suggest that, the studies show that, the results indicate that. This structure is mostly used to report findings or results. However, some slight differences are shown in the use of other clauses. Anticipatory it-clauses controlled by verbs indicate difference among L2, native, and journal groups. It is (was) found was more frequently used to suggest results in L2 corpus than in native and journal corpus. L2 writers tend to use constructions such as it is (strongly) suggested that and it is hoped that to express their suggestions or expectations before or after the study. In comparison, native and journal writers do not seem to frequently use it is (strongly) suggested that to express
Type L2 Native Journal
Adjectives 43 (6.22) 134 (16.41) 118 (14.11)
Nouns 0 19 (2.33) 8 (0.96)
Total 43 6.22 153 17.39 126 14.82
56
one’s suggestions. It is hoped that was also found absent from the top 25 list in native and journal groups. Rather, it seems that they would use active constructions, for example, (the studies, the results) suggested that, X (2001) suggested that, X (1988) found that to convey tentativeness. X (2001) suggested that ranks 7th in native corpus and 13th in journal corpus. The multi-word hedge X (2001) found that ranks 11th in L2 corpus, whereas this same construction ranks first and third in native and journal corpus, respectively. Another area that is worth noting is the use of multi-word hedge it is (should) (be) noted that. The use of this hedge is absent from the L2 corpus, but it is a mutual hedge in native and journal corpus. Last, one more difference is the sensory verbs seem and appear. For L2 writers, it seems that is the only multi-word hedge that appears on the list. In addition to it seems that, native and journal writers also use it appears that which ranks 18th and 9th in native and journal corpus, respectively.
The results also suggest that L2 writers do not use as many adjective controlled clauses as native and journal writers do since adjective controlled clauses were absent from the list in L2 corpus. On the other hand, native and journal groups use more of the hedges controlled by adjectives as shown in underlined (e.g. it is possible that, it is clear that, it is evident that, it is likely (not unlikely) that).
57
Table 4.6 Top 25 Multi-word Hedges across Three Corpora
No. L2
58
59 4.3.2 Anticipatory it Clauses
Figure 4.1 Frequency of Structural Patterns in Anticipatory It-Clauses
Based on the data, anticipatory it-clauses are divided into five sub-patterns: it + (modal) (be) verb, it + be + adjective, it + noun. From the figure above, it is evident that L2 writers differ from native writers in some aspects. The predominant structural pattern for the L2 group is it + (modal)+(be) + verb (e.g. it is (was) concluded that) and the least frequent pattern is it + noun. Native and journal writers, on the other hand, are slightly the opposite. They showed higher frequency in the two aspects, nouns and adjectives.
4.3.2.1 It + (modal) + (be) + verb
Epistemic lexical verbs constitute the greatest number in anticipatory it clauses in the three groups. The verbs can be presented in either passive or active forms.
Based on the data from the corpora, passive forms occur more frequently in anticipatory-it clause than in any other clause type. The results are not surprising since passive forms are a typical linguistic feature in academic writing, which has been acknowledged by many studies (Salager-Meyer, 1994; Hyland, 1998; Varttala, 2001;
Vold, 2006; Hewings & Hewings, 2002). Despite the consistencies, there are still some slight variations in the three corpora. Table 4.7 below summarizes the ranking of some frequent verbs along with the corresponding verb types, raw frequency, and
60
Comparing the total frequency of three verb types in the three corpora, L2 writers are found to be the highest which is 53.45 per hundred thousand words, followed by native and journal groups. From the total number of frequency, it is evident that both native and journal writers are roughly equal. As for the overall distribution of each verb type in anticipatory it-clauses, non-factive reporting verbs appear to be the most popular. In L2 corpus, non-factive reporting verbs accounted for more than half of the total occurrences at the frequency of 32.43 which occupies approximately 60% (32.43 / 53.45) of the three categories. Similar results are also apparent in native corpus. Native writers have the frequency of 13.73 in this category which is around 50% (13.73 / 25.74) of the three categories. Nevertheless, journal writers do not show as much contrast among the three categories as L2 and native writers. This finding might be due to the fact that the length of journal articles is shorter than the articles found in native and journal corpora.
Evidently seen from the findings, it could be concluded that non-factive reporting verbs are used more frequently by three groups of writers, which is consistently the same across three corpora.
Table 4.7 Frequency of Verbs in Anticipatory it-Clauses
L2 Native Journal
Cognitive RF NF Cognitive RF NF Cognitive RF NF
hope 41 4.98 assume 13 1.59 assume 13 1.55
infer 18 2.19 hypothesize 14 1.72 predict 8 0.96
believe 16 1.94 expect 9 1.10 expect 7 0.84
assume 16 2.19 infer 7 0.86 hypothesize 7 0.84
expect 15 1.82 believe 6 0.74 believe 5 0.60
imply 7 0.85 imply 3 0.37 hope 5 0.60
speculate 4 0.49 hope 2 0.25 imply 3 0.36
infer 1 0.12
Subtotal 117 14.46 Subtotal 54 6.63 Subtotal 49 5.87
61
As can be seen from Table 4.7, the verbs are categorized into three major types:
cognitive (related to one’s mental processes, beliefs, speculations, expectations), sensory (refer to one’s feelings, opinions), and non-factive reporting verbs (refer to the report of research findings). Nevertheless, variations are shown in some of the verb usage. The present study points out some of the syntactic constructions along with example sentences that appear to have noticeable differences.
Sensory RF NF Sensory RF NF Sensory RF NF
seem 43 5.22 seem 25 3.06 seem 32 3.82
appear 11 1.34 appear 19 2.32 appear 31 3.71
Subtotal 54 6.56 Subtotal 44 5.38 Subtotal 63 7.53 Non-factive RF NF Non-factive RF NF Non-factive RF NF
suggest 78 9.47 note 39 4.78 note 42 5.02
found 66 8.02 conclude 20 2.45 found 29 3.46
show 31 3.77 found 16 1.96 suggest 10 1.20
conclude 21 2.55 recommend 13 1.59 show 8 0.96
note 13 1.58 suggest 11 1.35 demonstrate 5 0.60
observe 9 1.09 indicate 5 0.61 conclude 5 0.60
indicate 9 1.09 propose 3 0.37 state 5 0.60
recommend 9 1.09 claim 2 0.25 claim 4 0.48
claim 7 0.85 reveal 2 0.25 indicate 3 0.36
state 5 0.61 demonstrate 1 0.12 propose 3 0.36
report 5 0.61 observe 2 0.24
reveal 5 0.61 recommend 1 0.12
See 5 0.61 reveal 1 0.12
Say 3 0.36
demonstrate 1 0.12
Subtotal 267 32.43 Subtotal 112 13.73 Subtotal 118 13.24 Total 438 53.45 Total 210 25.74 Total 230 26.64
62 a. Cognitive Verbs
In cognitive category, one of the interesting differences is hope. In L2 corpus, the verb hope has the highest frequency with 4.98 per hundred thousand words which occupies around 30% of the entire cognitive category. That is to say, L2 writers tend to heavily rely on this verb. The verb hope is used to express some of the possible benefits or contributions of a research study would bring to language teaching and learning. For example, in examples (i) (iii) authors have expressed their expectations by stating that it is hoped that this study will provide…it is hoped that this study will shed light on…In comparison with native and journal corpus, the tendency of expressing one’s own expectations in the pattern it is / was hoped that is relatively low in the other two corpora. Both native writers and journal experts exhibited quite low frequency with only around 0.25 for natives and 0.6 per hundred thousand words for journal writers. The infrequent use of this verb might be attributable to the fact that it is a “private verb”. According to Biber et al. (1988:105, cited in Granger &
Paquot, 2009), “private verbs” (love, want, hope, feel) are associated with one’s own feelings, emotions, and attitudes, and this verb type seldom appears in written academic texts.
(i) It is also hoped that this study will provide teachers with an effective e-learning and e-testing instruction instrument to carry out their teaching and evaluation. (NKNU 03) – L2 corpus
(ii) By improving and upgrading the quality of team-teaching, it is hoped that the results of the study can help EFL learners to learn how to incorporate appropriate narrative structure and evaluative devices to produce an engaging and coherent narrative. (NKUST 03) – L2 corpus
(iii) There have been scare studies discussing Taiwanese students' English learning characteristics and their language learner autonomy. It is hoped that this study will shed light on language learner autonomy itself and its relationship with
63
Taiwan students' English learning characteristics. (NCUE 03) – L2 corpus
(iv) There have not been many studies to have implemented a Direct Instruction writing program; therefore, it was hoped that this study would add to the existing literature base. (Ohio State 39) – Native corpus
(v) It was hoped that by using multimedia program with its guided tasks might have contributed to the development of learners’…(ESP 62) – Journal corpus
b. Sensory Verbs
As pointed out earlier, sensory verbs are the smallest proportion of the entire corpora, which is consistently true in the three corpora. The two main verbs which fall under this category are seem and appear. They are widely acknowledged as the typical markers of tentativeness in academic writing. This finding is also evident in a number of studies (e.g. Varttala, 1999 ; Hewings & Hewings, 2002) in which they found that these two are typical hedging strategies in academic writing. The present study found that seem and appear are either used to report the observed findings in their own present study or provide their own interpretation to the existing literature so as to attenuate the force of evaluation and to reflect tentativeness and uncertainty in researchers’ interpretation.
(i) It seems that NESTs and NNESTs teach differently, which is why many schools want to recruit NESTs for their students. (NTUST 04) – L2 corpus
(ii) It seems that Taiwanese students are used to adapting themselves to different learning contexts such as tests, assignments, in-class English learning activities.
(NCUE 03) – L2 corpus
(iii) In consideration for the affective benefit of extensive reading, it seems that the only member of the intended treatment group to have expressed a very favorable of extensive reading was a Taiwanese female. As for Saudi Arabian male 2… (Toledo 42) – Native corpus
64
(iv) …students were very engaged in the reading rehearsing of this text. It seems that students were unable to identify and articulate the use of high-interest, appropriately-leveled books as motivational to their reading in the classroom…(BGSU 05) – Native corpus
(v) From student evaluation it seemed that the virtual department did offer them the communication and working opportunities they needed. (CALL 21) – Journal corpus
(vi) Lyster (2007,p.151), who advocate practice that is “inherently repetitive and psychologically authentic.” It seems that, in order for writing tasks to be truly authentic, students would need to focus on the accurate production… (TESOL 46) – Journal corpus
In addition, the present finding also indicates L2 writers seldom use it would seem / appear that. On the other hand, native and journal writers would use this construction to attenuate the force of a proposition either in literature or results section. It could be either used in evaluation to previous researchers’ belief or to one’s own research findings. In native corpus, this string of hedges occurs 12 times in 7 texts which is around 14% (7/51) for the dispersion rate. In journal corpus, it appears 8 times in 7 texts. The dispersion rate is about 6.4%. Although the frequency is not too high, the data could still show some tendency of native writers.
(i) Hughes argue that… It would seem that the best language placement tests are those created in-house, with a specific population in mind. (Youngstown 50)–Native corpus
(ii) Whether or not the AR [Accelerated Reader] accomplishes each of these goals is much of debate. For example, it would appear that the goals of Accelerated Reader do not cover the social aspects…(BGSU 15) –Native corpus
(iii) In politics, although much less frequent overall, almost equal numbers of SHOW and ARGUE verbs occur…Thus it would seem that the extensive use of a hidden averral with a SHOW verb is particularly characteristic of writing in materials.
65 –Journal corpus (AL 08)
(iv) The study focused on all four skills. Therefore, it would appear that Base strategies are context specific. (FLA 89) –Journal corpus
c. Non-Factive Reporting Verbs
Non-factive reporting verbs refer to verbs which are used to report writers’
research findings from previous studies (Varttala, 2001). Many verbs were found to be mutually shared across three groups, but there are differences in terms of frequency.
L2 writers tend to have used certain verbs more frequently than others.
In L2 corpus, the top three markers are suggest, found, and show. Taking these three verbs to compare with the ones in native and journal corpus, it was found that L2 writers have a much higher frequency in these verbs. The three verbs account for around 60%, which is over half of the entire verb category. That is to say, L2 writers seem to heavily rely on these verbs more than other verbs. In native corpus, only suggest and found were in anticipatory it clauses. These two verbs only make up around 25% of the verb category. In other words, native writers do not seem to have
In L2 corpus, the top three markers are suggest, found, and show. Taking these three verbs to compare with the ones in native and journal corpus, it was found that L2 writers have a much higher frequency in these verbs. The three verbs account for around 60%, which is over half of the entire verb category. That is to say, L2 writers seem to heavily rely on these verbs more than other verbs. In native corpus, only suggest and found were in anticipatory it clauses. These two verbs only make up around 25% of the verb category. In other words, native writers do not seem to have