2.2 Function of Hedges
2.2.4 Solidarity
In addition, Myers (1989) believed that writing is an act that involves social interactions, for example making claims, sharing assumptions or making criticisms of others’ perspectives. Hedges can be used by writers to maintain interactions with readers (Hyland, 2001, 2005). Hyland (2001) conducted interviews to discover the original writers’ attitude towards their writing. Some of the participants also thought that the use of pronoun we can help bring readers engage in the text and discuss about writers’ arguments. With this strategy, readers are treated as part of the text so that an interaction is developed between writers and readers.
13 2.3 Researchers’ Classification of Hedges
Hedges can be realized in various linguistic forms. Despite there is a large number of hedges, some linguistic forms are considered to be typical hedges. This section will focus on some of the well-known researchers’ classification of hedges.
Similarities and differences of the classifications will be discussed below.
In early 1980s, Prince et al. (1982) believed that hedges can be of two types.
Based on the physicians’ spoken discourse in pediatric intensive care unit, they classified hedges into two types: approximators and shields. When the truth value of propositions is affected, it is an approximator, whereas when the entire proposition is hedged and the truth value of proposition remains unaffected, it is a shield ( Markknen & SchrÖder, 1997 ) .
To put it more specifically, approximators are of two kinds: adaptors and rounders. Adaptors are similar to the hedges proposed by G. Lakoff (1973) such as sort of. Rounders are like approximately or something, which indicates a range of something. This is similar to what other researchers have suggested as indicated earlier (e.g. Holmes, 1984; Dubois, 1987)
According to Prince et al. (1982:85), shields are also further divided into two subtypes: attribution shields and plausibility shields. The former means that the responsibility of the expression is attributed to a third person, not the author himself or herself (e.g. according to her estimates). The latter reflects writer’s uncertain attitude towards the proposition (e.g. I think).
Skelton compared the use of ‘commentative language’ which expresses “the way in which we can express the extent to which we commit ourselves to particular propositions” (1998b:98) in both disciplines of science and humanities. Based on the data, four categories of comment were identified: (1) copulas other than be, (2) modal
14
verbs, (3) clause initials (There is, this is, it is), (4) lexical verbs (propose, believe, report). Skelton believed that comments frequently appear in academic writing as they appear in around “(…) one third and one half of all sentences” (1988b:103).
Salager-Meyer (1994) later suggested five types of hedge: (1) shields: modal verbs, semi-auxiliaries (e.g. appear), probability adverbs (e.g. probably), derivative adjectives, epistemic verbs (e.g. suggest) (2) approximators (e.g. roughly, somewhat, often) (3) expressions of the author’s personal doubt and direct involvement (e.g. I believe) (4) emotionally charged intensifiers (e.g. particularly encouraging, extremely difficult). (5) compound hedges or strings of hedges (e.g. It may suggest that…/It would seem likely that…). In response to the classifications, Varttala (1999) argued that there is an overlap between first and third category. Some of the linguistic devices in the first category may also be used to denote uncertainty and involvement.
Obviously, there are some problems over the categorizations in the taxonomy proposed by Salager-Meyer (1994).
Based on the investigation of hedging use in the past, one of the more widely used models is Hyland’s polypragamic model of scientific hedges (1998a) as illustrated in the figure below.
Hedging
Content-oriented---Reader-oriented Accuracy-oriented---Writer-oriented
Attribute---Reliability
Figure 2.1 A model of scientific hedging (Hyland 1998a, p. 156)
15
Within the model, there are two major types of hedge: content-oriented and reader-oriented. Content-oriented hedges consist of two subtypes with one being accuracy-oriented and the other being writer-oriented. Accuracy-oriented hedges are used to allow writers to make arguments with greater precision. This type of hedge in this case can be further broken down into two subtypes: attribute hedges and reliability hedges. Attribute hedges refer to approximators which are similar to the notion suggested by Prince et al (1982) and Salager-Meyer (1994). Words like quite, more or less, almost, and barely belong to this category. Reliability hedges are to indicate writers’ “subjective uncertainty” of knowledge (Hyland,1998a:166).
Reliability hedges are commonly realized through the use of different grammatical categories: verbs, adverbs, and nouns (e.g. we suspect…, possibility, appears possible, probably) (Hyland, 1998a:167).
Writer-oriented hedges are in common with Prince et al’s (1982) shields as they are used to prevent writers from being personally involved in the making of propositions. To keep distance from the propositions can be realized in the syntactic constructions involving passive constructions it was assumed that or the use of clauses involving abstract entities such as these data indicate that, the model implies that, taxonomic evidence suggests (Hyland, 1998a:172).
As opposed to writer-oriented hedges which is realized through abstract subject entities to show detachment from the audience, reader-oriented hedges, according to Hyland (1998a) are designed to gain readers’ acceptance and allow readers to have space for interpretation or judgment.
From above, researchers have looked into the concept of hedging and divided it into different types based on the available data in different contexts. These classifications have laid the foundation for future studies. Particularly, Hyland’s
16
model was comprehensive in a way that it could help one analyze the data more easily and clearly.
2.4 Single-and Multi-word Hedges
From the previous studies, hedges can be classified in different ways, depending on the data and the author. In this section, the concept of single-and multi-word hedges will be discussed in detail especially the latter, since this is the aim of the present study.
2.4.1 Single-word Hedges
As shown above, researchers over the years have categorized hedges into different types. Some are single-word hedges, while others are multi-word hedges.
Among these linguistic devices, some are more prevalent than others in academic writing. Thus, a deeper analysis over some of the frequently used grammatical categories and syntactic constructions is necessary.
a. Modal Auxiliaries
Modal auxiliaries have been found to be one of the common hedges in various disciplines: applied linguistics (e.g. Lau, 2001; Yeh, 2007; Lin & Liou, 2009), medical (e.g. Varttala, 1999), and science (Hyland 1998a; Li, 1998). Despite they are easy to be found in a text, they are difficult to be judged whether it is a hedge as some modal auxiliaries may carry multiple meanings. To better understand their meaning, it often depends on the analysis of context, which is believed to be a strenuous task (Holmes, 1998b).
In general, according to previous researchers (Coates, 1983; Palmer, 1990) modal auxiliaries are divided into non-epistemic or root meaning (obligation, ability,
17
permission, and volition) and epistemic meaning (one’s attitudes and beliefs). This section will limit the discussion only to epistemic meaning since hedges are largely realized by epistemic modal auxiliaries. The modal auxiliaries in English which are associated with epistemic meaning are can, could, may, might, will, would, should, Firstly, consider the first pair of auxiliaries can and could. Can is associated with epistemic meaning only when it appears in interrogative or negative conditions, whereas could is often used to suggest possibility in hypothetical situations when it refers to epistemic meaning (Hyland, 1998a). Secondly, consider may and might. May and might are almost the same in meaning and might is a form which indicates lesser degree of certainty. To check if the meaning of may indicates the voice of epistemic possibility, may can be easily replaced by the phrase possible that. In addition, may can be used to denote habitual or future events (Palmer, 1990). Thirdly, consider the last pair. Will is used to denote epistemic meaning when it is referring to prediction.
Would is a tentative form of will, which can be almost interchanged with words like perhaps or It would be reasonable to concluded that… (Palmer, 1990:48). Also, when would is used to express epistemic meaning, it is used as marker that signals prediction. Lastly, should is associated with epistemic meaning when it indicates writers’ attitudes and assumptions based on the propositions (Hyland, 1998a).
b. Epistemic Lexical Verbs
Epistemic lexical verbs comprise a large class of hedging strategies. According to Hyland (1998a, 1998b), epistemic lexical verbs can be divided into two major categories: judgmental and evidential verbs.
Judgmental verbs are the type which is associated with a writer’s uncertainty of evaluation. This category can be further broken down into speculative (e.g. believe, speculate, suspect, assume) and deductive (e.g. calculate, estimate, conclude, deduce,
18
infer, imply). The former has the sense of speculation or prediction, whereas the latter is associated with inferential reasoning, or calculation. Judgmental verbs often co-occur with the inclusive pronoun we, anticipatory it, and abstract noun (e.g. the study). On the other hand, evidential verbs carry two functions. One is used to report truths from previous studies which can be achieved through verbs of ‘hearsay’
evidence (e.g. show, predict). The other function is to indicate one’s feelings, attitudes, or opinions through verbs of senses (e.g. appear, seem). Verbs of senses are frequently used with anticipatory it and inanimate subject.
Nevertheless, Hyland’s categorization is not entirely clear-cut, for there might be some fuzziness between judgmental and quotative category. Certain verbs (e.g.
believe) which are believed to be judgmental could also be counted as members of quotative category. This problem of ambiguous categorization is also recognized in Varttala (2001) who also investigated hedges in science discipline. To make the classification of data easier and less ambiguous, Varttala (2001) came up with a revised classification of three major categories. The three categories are tentative cognition verbs (verbs of thinking or mental activities), tentative linking verbs (seem, appear), and non-factive reporting verbs (verbs used to report one’s own or other’s findings).
19 2.4.2 Multi-word Hedges
Over the past years, single-word hedges have drawn widespread attention in research writing. Their possible functions and linguistic environments have been identified and analyzed in various disciplines. Multi-word hedges, on the other hand, have not been as much the main focus as single-word hedges as they are less often discussed by researchers in the community. Multi-word hedges take a variety of linguistic forms. Nevertheless, the discussion here will only focus on that-clauses.
a. Anticipatory it + that
One of the common linguistic devices is the use of anticipatory it (Hyland, 1998a;
Hewings & Hewings, 2002). In this linguistic structure, a long piece of information is placed at the back of a sentence, and a dummy pronoun it comes at the initial position in a sentence (Quirk et al.,1985). This structure enables writers to conceal their attitude and thus reduce the responsibility of making personal evaluations, which can be regarded as a protective device of fending off oppositions or criticisms from others (Hyland, 1998a, 1998b). This strategy can be realized in the forms of verb phrases and adjective phrases. Example sentences from the data in Hyland (1998b:16-17) are illustrated below.
(i) It is evident that the enzyme and its rnRNA are present in high concentrations throughout the degradation of the old clove and the formation of the new bulbs. (Bio) (ii) It was shown that the algorithm can function well during actual weldingwhen various noises are present. (EE)
b. Abstract Entity + that
The clauses which begin with abstract entities do not involve human volition.
Previous studies have consistently recognized this type of clause could function as hedges to prevent the author from being personally committed to the proposition
20
(Charles 2006, Hyland 1998a). In Charles’ study (2006), abstract entities could be further subdivided into two types: research nouns and text nouns. The former may refer to the process and result of a study, for example data, observation, and result.
The latter is associated with the elements in a text such as chapter, tables, and figures.
For this strategy, lexical verbs play an important role. Judgmental verbs, sensory verbs as well as deductive verbs commonly occur with abstract entities. Below are some of the example sentences from (Hyland 1998b:17).
(i) X-ray diffraction analysis shows that…,
(ii) The result presented here clearly indicated that…
c. Human Subject + that
In addition to using anticipatory it and abstract entity, one can also attribute responsibility using human subjects. To begin sentences with human subjects there are a few options. One can choose that clauses headed by different forms of human subjects including first person pronoun, nominalizations, and named researcher.
First Person Pronoun
First person pronoun is employed to convey writers’ own attitude or opinion so that the proposition is subjective. They are usually in the form of first-person pronouns such as I or we (Hyland, 1998b). The singular form I and the plural form we are different in their pragmatic function. The pronoun I shows that the writer is being personally committed to the proposition. The pronoun we are a bit complex, which include two kinds: inclusive and exclusive. When the pronoun we is referred to people in general including both reader and writer, it is an inclusive we. (Kuo, 1999;
Scheibman, 2004; Charles, 2006). On the other hand, an exclusive we refers to the
21
author and other associates. (Harwood, 2005; Charles, 2006). An inclusive we helps
“invite readers to orient themselves to the discourse and engage in a dialogue”
(Hyland, 1998b:178). People in general can contribute their alternative views, beliefs, and objections to the text (Salager-Meyer, 1994; Hyland, 1998a, 1998b; Kuo, 1999;
Koutsantoni, 2006). In addition to building relations with readers, the use of inclusive pronoun we can also be used to indicate that “there could be others who share propositional responsibility” (Charles, 2006:507). In other words, the responsibility is distributed with the use of inclusive pronoun.
Below is a synthesized table of multi-word hedges compiled from some of the previous studies by Hyland and Milton. The present study will use Table 2.1 as a reference to identify the potential hedging devices.
22 Table 2.1Patterns and Examples
Pattern 1 Examples
It + (modal aux) + (be) + verb + that
1. It was suggested that rays be returned from the eye’s lens, at angles close to the Brewster angle, which would in fact constitute a retrodirector.
2. Although the error increases when 1/1 is less than 0.01 or larger than 1.0, it seems that the ratio d/1 in the range from 0.01 to 1.0 gives accurate, stable results.
(Hyland, 1998b:17)
Pattern 2 Examples
It is + adjective + (to + verb) that
1. It is therefore possible that cv. Firstmore might be different not only in its chromosomal arrangement but also in displaying a particular mechanism..
(Hyland, 1998a:166) 2. It is possible to argue, therefore, for a comprehensive Bill of Rights, with an inclusion of privacy, tied to a US-style Freedom of Information Act.
(Hyland & Milton, 1997:194) Pattern 3 Human Subject Examples
Pronoun + (modal verb) verb + that
1. I believe that the major organizational principle of thylakoids is that of continuous unstacking and restacking of sections of membranes…
2. We speculate that the inhibitor interacts specifically with a protease…
3. We infer that the rate becomes limited by the rate of regeneration of RuBP because both the calculated rate of electron transport and…
(Hyland, 1996:15-16) Pattern 4 Human Subject Examples
Nominalized Subject + that 1. Many students claim that the school text books are rather boring and hard to understand.
2. Many critics argue that reading such books is disadvantages.
(Hyland & Milton,1997:198) Pattern 5 Human subject Examples
Named Researcher + that 1. Quirk et al (27) also noted that there was no
23
consistent effect of reduced Rubisco content on the stomatal conductance of their antisense plants but observed that conductance was insensitive to CO2 (30).
(Hyland, 1998a:180-181) 2. Sorokin’s study predicted that the creation of
quenching centres...
3. ..Henninger et al [3] speculated that tuber size is correlated with extent of necrosis in field-grown potato.
(Hyland,1998a:124)
Pattern 6 Examples
Abstract entity + verb + that 1. Third, the present work indicates that the aromatic ring to which the…
2. Taxonomy evidence suggests, therefore, that a species…
3. The model implies that the function of grana is to…
(Hyland,1998a:171-172)
24
2.5 Previous Studies on Hedges by Natives and L2 Learners
The above discussion has introduced some single-word and multi-word hedges.
The following sections will specifically discuss how L2 learners differ in the use of hedges when compared to the writing by natives or experts.
2.5.1 Overall Distribution of Hedges in L2 and Native Writing
Previous studies which had focused on the investigation of hedges and boosters in native and L2 writing in different genres have revealed L2 learners incline to hedge less. As a result, their writing might sound more assertive. This phenomenon has been found to be consistent in many studies.
Allison (1995) examined essays written by ESL undergraduates in Hong Kong and suggested that although they had received many years of writing experience, hedges were less used. As a result, their writing style remained overstated. For example, learners might say “something is ‘always’ the case” (Allison, 1995:10). In the similar vein, Milton and Hyland (1996, 1997) discovered L2 writers used more certainty markers than native speakers, which make their statements sound more categorical. Similar linguistic phenomenon is also evident in more recent studies.
Chen (2010) noted that making appropriate degree of propositions is difficult for L2 learners to achieve. While native speakers may express the propositions with moderate degree, L2 writers tend to overuse boosters which make the arguments sound more assertive. The poor performance on the rhetorical strategy could bring serious consequences to academic communication, which would influence learners’
academic ability in reading texts and making arguments .
In Taiwan, Chen (2005) investigated how Taiwanese applied linguists differ from writers of TESOL Quarterly in expressing doubt and uncertainty. Two corpora were
25
built with one compiled with ETA conference papers and the other consisted of TESOL Quarterly journal articles. The results were categorized on a continuum from certainty, probability to possibility, and they suggested that Taiwanese applied linguists were found to hedge less compared to their professional counterparts. For example, may is the most frequently used device in both groups, but its frequency in TESOL Quarterly is almost twice as much as in ETA corpus.
The same problem was also found in learners of other languages. Hinkel (2005) made a comparison between L2 learners’ (e.g. Chinese, Japanese, Korean, Indonesian, Vietnamese, Arabic) academic essays to native speakers’ writing. Hedges and intensifiers in NS and NNS corpora were identified based on a compiled list of epistemic hedges (clearly, mostly, relatively), lexical hedges (kind of, may be), possibility hedges (perhaps, possibly), downtoners (a bit, nearly, partly), assertive pronouns (words start with any- or some-), adverbs of frequency (often, frequently, usually, occasionally), emphatics…etc. The results indicated that L2 learners were found to employ more linguistic devices that serve to create hyperbolic impression in writing.
2.5.2 Verbal and Nonverbal Types in L2 and Native Writing
This section provides an overview over the use of both verbal and non-verbal hedges in L2 and native writing so as to understand the extent to which L2 writers are similar or different from native writers.
a. Verbal Types
Verbs have been considered to be one of the most frequently used in L2 writing.
Studying the distribution of grammatical categories in applied linguistics, Li (1998) reported that common verbs that appear frequently in L2 corpus include propose,
26
suggest, assume, argue, report, indicate, show. Among them, suggest was the most frequently used verb. Yeh (2007) also discovered that lexical verbs were one of the most heavily used grammatical category. Among the lexical verbs, the ones which are frequently used are indicate, suggest, seem. Different studies on other disciplines also confirmed similar finding. Lau (2001a) examined 100 published research articles across various disciplines written by Taiwanese Ph.D. students in National Tsing Hua University. The hedges were categorized into five grammatical categories. Among the hedges, lexical verbs ranked the highest (42.6%). For lexical verbs, the most frequently used was show. The high number of occurrence of this word, according to the author, might be related to their general meaning. For example, the term show can be defined as to display, to report, or to demonstrate. More studies on other learners of
suggest, assume, argue, report, indicate, show. Among them, suggest was the most frequently used verb. Yeh (2007) also discovered that lexical verbs were one of the most heavily used grammatical category. Among the lexical verbs, the ones which are frequently used are indicate, suggest, seem. Different studies on other disciplines also confirmed similar finding. Lau (2001a) examined 100 published research articles across various disciplines written by Taiwanese Ph.D. students in National Tsing Hua University. The hedges were categorized into five grammatical categories. Among the hedges, lexical verbs ranked the highest (42.6%). For lexical verbs, the most frequently used was show. The high number of occurrence of this word, according to the author, might be related to their general meaning. For example, the term show can be defined as to display, to report, or to demonstrate. More studies on other learners of