The fourth research question aims at uncovering the distribution of hedges across three corpora. From the previous section, the distributional difference of three major patterns (anticipatory it, abstract entity, human subject) are presented. Specifically, the study also explores the linguistic forms that tend to generally associate with certain
97
verb types. This section provides the functional distribution of data across three corpora.
4.5.1 Reference to Other Studies
Reference to other studies or literature is the most frequent function in this study.
In other words, three groups of writer tend to use that-clauses to refer to previous studies. This reflects the attitude that writers are perhaps seeking to use cited research to gain credibility and to reduce responsibility. This type of hedge is equivalent to the
“attribution shield” proposed by Prince et al. (1982:89). This hedging function is commonly observed across three corpora since it has been recognized as quite common in academic writing. In this study, the function is principally achieved through named researcher which is similarly found across the corpora. The clauses which begin with nominalized subjects are the least frequently used to achieve this function which are also similar across the three groups. The verbs could be of various types including cognitive (e.g. believe) and non-factive reporting verbs (e.g. found, suggest, report, indicate) which largely co-occur with animate sources. Examples are shown as follows. Notice that the first three examples are headed by citation and the last three are begun with nominalized subjects such as some theorists, scholars, researchers…etc.
(i) Schachter (1990) indicated that the gap between knowledge of the structure of a sentence studied in isolation and knowledge of what is conveyed in uttering that sentence in a particular…(Chung Cheng 01) –L2 corpus
(ii) Ripich and Griffith (1988) found that when retelling a story, children with LD/LI had difficulty recalling… (Akron 02) -Native corpus
(iii) Saito and Lyster (2012) found that when JL 1s produce and /r/ that is perceived as correct by NSAEs…(CALL 16) –Journal corpus
98
(iv) Some theorists believe that text-based factors determine meaning, while others believe that inside-the-head factors determine meaning (Bernhardt, 1984).
(NTUST 03) –L2 corpus
(v) Many scholars believe that African Americans are ashamed of AAE…Perry
& Delpit , 1998 ) (Miami 31) –Native corpus
(vi) Some researchers have suggested that there may not be a large ‘generic’
academic vocabulary, which is useful for students across the university (Hyland and Tse 2007; Martinez et al. 2009) (AL 05) –Journal corpus
In summary, attributing the responsibility to a third person source creates an impression that the content of a proposition, a concept, or an idea does not belong to the writer, but the originator. With it, not only the responsibility is reduced on the part of a writer, but this can also make the entire writing sound more objective and reliable in tone.
4.5.2 Reference to Result
As common as the previous one, this hedging function is used when reporting a proposition. It might appear in the reporting of results that a writer himself or herself does not want to be fully committed to the claim. Consider the following examples.
(i) …it seems that the only member of the intended treatment group to have expressed a very favorable opinion of extensive reading was Taiwanese female.
(Toledo 42) –Native corpus
(ii) As demonstrated in the quantitative analysis above, L1 texts contained significantly fewer verb errors; however, through qualitative analysis, it appears that certain types of verb errors may be common across Generation 1.5 and L1 students alike. (JSLW 25) –Journal corpus
(iii) Data from the second section suggests that while there are clear cases of count nouns whose referents are individuated…and mass nouns with hardly any
99
individuation of their referent at all, there are… (Ohio 37) –Native corpus
(iv) With regards to the first question, the overall results suggest that the participants view the use of the LRS positively or at least with a certain degree of neutrality… (CALL 18) –Journal corpus
Another means to achieve this function is through the use of cognitive or speculative verbs with either the singular I or plural form we (inclusive or exclusive).
Inclusive we mainly is referred to both readers and writers. Exclusive we only includes authors. Of the three corpora, L2 writers have the fewest instances of the pattern I / We + cognitive verb as indicated earlier in the study. One of the most frequent cognitive verbs they used is know as in we (all) (may) know that… which is only at the frequency of 0.61 per hundred thousand words. They tend to express their viewpoint with the inclusive we to involve readers into the text by assuming shared knowledge with readers. Referring to inclusive we probably sound more general and involve less personal commitment. As suggested by Hyland (2002), he interviewed a couple of L2 students in Hong Kong and discovered that they were not confident enough to use first-person pronouns as they thought the first-person linguistic devices are not allowed to appear in academic writing.
On the other hand, in both native and journal corpus, believe is one of the most frequently used mutual hedges. In native corpus, writers prefer the use of I, whereas in journal corpus we is more the more predominant one, which is not surprising to see that journal writers often use exclusive we instead of I to refer to writers themselves and “to decrease the visibility of the author himself” (Lo, 2010 p.158) since many of the journal articles are multi-authored. This is probably the reason why journal writers have the highest frequency of personal reference.
(i) Some students even spoke against what they saw as material unrelated to an
100
English class. However, I believe that (even in the case of those students who resisted the assignments) the games themselves demonstrate, in meaningful ways, the value of incorporating interactive fiction in to a class. (BGSU 17) –Native corpus
(ii) Although some researchers have appropriately suggested that this lack of observable improvement in ESL writing may be due to flaws in research methods (e.g.
Ferris, 1999, 2004, 2006 ; Truscott, 1996, 2004), we believe that it is at least as important to recognize that weaknesses in instructional methodologies ….. (TESOL 46) –Journal corpus
As from the examples (i) and (ii) shown above, the authors are interpreting the research findings from either the literature or the results based on their opinions or beliefs. Thus, the statements seem to be quite subjective and personal in tone. The readers are given the opportunity to evaluate the proposition.
4.5.3 Reference to Suggestions
Reference to suggestions is the third function identified in the present study.
Typical verbs of providing suggestions for teachers and researchers are suggest and recommend. This function could be achieved through both anticipatory it subject. In L2 corpus, the combination of anticipatory it subject accompanied by the verb suggest is more frequently observed. On the other hand, in native corpus, native writers seem to prefer the use of it is recommended that over it is suggested that. However, the frequency in native corpus is only per hundred thousand words, which still requires further exploration into the use.
(i) In view of the limitations of the present study, the researcher presents the following suggestions for future research. First, it is suggested that students from elementary schools, junior high schools, or college be included. (NCUE 03) -L2 corpus
(ii) There are also many books available that teachers can read to locate new strategies for reading comprehension and vocabulary. It is recommended that teachers chose a book and have an informal book club to discuss ideas and other ways….(BGSU 14 ) -Native corpus
101 Table 4.17 Distribution of Function of Hedges
4.6 Discussion of the Use of Multi-word Hedges across Three Corpora