Maket 2013:Layout 1.qxd


Modal Adverbs Used as Hedges in Research Articles

Naira Martirosyan, Tatev Ghazaryan
Yerevan State University

A
cademic writing is nowadays generally seen as an interaction between the writer

and the reader. The main aim of the writer is to build up an interactive and con-

structive relationship with the readers, to persuade them of the correctness of their

claims. Accordingly, the authors of research articles commonly qualify the information

presented by means of items relating to vagueness, uncertainty or tentativeness. These

devices are referred to as hedges.

Thus, despite a widely held view that professional scientific writing is purely objec-

tive and impersonal, hedges are abundant in science: they are a significant means for aca-

demics as they help researchers to “shield’ against readers” possible rejection of their

propositions.

Lakoff used the term hedge to refer to “words whose job is to make things fuzzier or

less fuzzy” (Lakoff 1972:195). This definition became the starting point for several studies

on this phenomenon. Later Brown & Levinson defined hedges as “a particle, word or

phrase that modifies the degree of membership of a predicate or a noun phrase in a set”

(Brown & Levinson 1987:145). They list hedging as one of various “face” preserving

politeness strategies.  For example, in order to avoid disagreement with the hearer the

speaker can make his own opinion safely vague by using hedges such as sort of, kind of,
etc. 

Hyland claims that hedges are “linguistic means used to indicate either a) a lack of

complete commitment to the truth value of an accompanying proposition, or b) a desire

not to express that commitment categorically” (Hyland 1998:1). Thus, they mark uncer-

tainty and help writers to present a proposition as an opinion rather than a fact. 

The overview of the development of the concept clearly demonstrates that the notion

of hedging has expanded by going through roughly three stages. The first stage is main-

ly semantic. The second stage is semantic and pragmatic, modifying the truth value of

the whole proposition and the speaker’s or writer’s commitment or attitude to the propo-

sitional content, with its focus on the speaker-content relationship. And, finally, the third

stage is mainly pragmatic and social, modifying relationships between the reader and the

writer or between interlocuters, or social relationships in a wider context, with its focus

on interpersonal and social relationships.

The writer’s or speaker’s judgments about statements and their possible effects on

interlocutors is the essence of hedging, and this clearly places epistemic modality at the

centre of our interest. Coates (1987:112) describes epistemic modality as concerned with

“the speaker’s assumptions, or assessment of possibilities, and in most cases, it indicates

the speaker’s confidence or lack of confidence in the truth of the proposition expressed”.

The epistemic system is mostly referred to as realizing a continuum of “commitment” on

the part of the writer to the truth value of his or her statements, ranging from uncertain

possibility to confident necessity. Its importance lies in the fact that such tentativeness

Armenian Folia Anglistika Linguistics

74



avoids personal accountability for statements, reducing the author’s “degree of liability”

(Huebler 1983:18), while expressing caution and facilitating open discussions. Epistemic

comment is seen by many writers as being at the heart of language use and a principal

means by which speakers can use language flexibly to adopt positions and express points

of view. Writers cannot avoid encoding their point of view towards whatever they say,

commenting on and committing to or distancing themselves from their propositions.

The concept of hedging is an area of some neglect (Crystal 1995:120) although the

literature contains references to various aspects of hedging over the past 20 years.

Hedges have been referred to as compromisers (James 1983), downtoners (Quirk et al

1972), weakeners (Brown & Levinson 1987), downgraders (House & Kasper 1981), sof-

teners (Crystal & Davy 1975), backgrounding terms (Low 1996) and pragmatic devices

(Stubbe & Holmes 1995). This phenomenon has received most attention in casual con-

versation where it is extremely common and represents a significant interpersonal and

facilitative communicative resource for speakers, who use them when they wish to cre-

ate an informal atmosphere, facilitate turn-taking, sustain discussion, show politeness.

These units have also been treated as a form of “metadiscourse” by some applied lin-

guists who have included them in the elements of a text.

As it has been mentioned above, hedges are abundant in science and play a critical

role in academic writing. They are important to scientists because even the most assured

scientific propositions have an inherently limited period of acceptance. Categorical asser-

tions of truth in these circumstances are decidedly hazardous. What is more, the use of

hedges may have provocative influence, for on the one hand such hedges as possibly,
maybe, admittedly somehow signal the writer’s perhaps limited knowledge, while on the
other hand, hedges like certainly, undoubtedly emphasize the commitment over what
would be indicated by a simple declarative in English. Hence we see that hedging is an

essential element of academic writing. As Crismore & Farnsworth (1990:15) mention:

“hedging is the mark of a professional scientist, one who acknowledges the caution with

which he or she does science and writes on science”.

In everyday conversation hedges are commonly expressed through auxiliary verbs

and by epistemic adjectives, adverbs and lexical verbs. Common conversational forms

include perhaps, I  guess, maybe, quite, sort of , while prosody, tag questions, hesitation
markers, rhetorical questions also function as hedges when they are used to decrease the

strength of an utterance (Coates & Cameron 1988; Holmes 1995). Academic research

writing also includes diverse expressions of hedging which make use of the full range of

epistemic lexical resources.

In academic writing modal adverbs comprise a greater range of items than modal

adjectives or nouns, from which they differ in that the former are “peripheral in clause

structure” (Quirk et al 1972:421). They are not syntactically integrated as an element of

the clause but can appear in a number of positions without affecting the meaning relation

between the clause and the adverb. In the examples below we shall see that the modal

adverb perhaps appears in all three positions without affecting the grammaticality of the
sentence or changing its core meaning. In other words, the presence or absence of the

modal adverb perhaps does not change the information that the speaker conveys. 

Armenian Folia AnglistikaLinguistics

75



e.g.   
a) Perhaps the most widely quoted ages of the universe come from the

age estimates of globular cluster stars.

b) The most widely quoted ages of the universe come from the age esti-
mates of globular cluster stars, perhaps. 

c) The most widely quoted ages of the universe perhaps come from the
age estimates of globular cluster stars.                           (Riess et al 1998) 

It is worth noting that the mobility of adverbs means that they can be selected to operate

on phrasal, sentential and supra-sentential levels. On the supra-sentential level modal

adverbs also perform a text-creating function, expressing various semantic relations,

anaphora and cataphora included. As is known, anaphora implies retrospective connec-

tion, while cataphora indicates prospective connection. In the following examples modal

adverbs illustrate anaphoric reference:

e.g.
a) But modality, as will be seen, does not relate semantically to the verb

alone or primarily, but to the whole sentence. Not surprisingly, therefore,
there are languages in which modality is marked elsewhere than on the
verb or within a verbal complex. (Palmer  1986)

b) With ought to/should the conditions were the carrying out of an obli-
gation; with might they seem to be those of behaving in a sensible or prop-
er way. But this a much stronger notion than mere permission, and it is
clear that might have does not mean ‘You would have acted thus if you had
had permission’. Apparently then, might expresses a stronger kind of
deontic modality than may – a positive suggestion rather than mere per-
mission. (Palmer 1986)

In these examples the modal adverbs not surprisingly and apparently refer to the pre-
ceding sentences, without which the meaning of the sentence introduced by the modal

adverbs would be obscure.

Examples of cataphoric reference are:

a) In a sense, presumably, this is to be interpreted as not being direct obser-
vation; but the point is that Tajik distinguishes just two types. (Palmer 1986)

b) Intuitively, it seems reasonable to assert that an ASC II file contains
7/8 as much information as an arbitrary file of the same size, since we
already know one out of every eight bits before we even look at the file.

(MacKay 2003)

Armenian Folia Anglistika Linguistics

76



Another interesting point to be made is that modal adverbs may be modified by such

words as very, too, most, quite, rather, less and may form degrees of comparison, for
example:

a) Less obviously, interrogative sentences are related, not only to state-
ments, but to relative clauses.                                               (Crystal 1995)

b) In notional terms all three are, in some way, concerned with the
event or situation that is reported by the utterance. Tense, rather obvious-
ly, is concerned with the time of the event, while aspect is concerned with
the nature of the event. (Palmer 1986)

c) More strictly (and significantly), this is so only if the subordinate
clause is part of what is reported.                                         (Palmer 1986)

Research shows that modal adverbs are quite often used in combination with other

modal adverbs or modal verbs:

e.g.

a) It is undoubtedly the case that most perhaps all languages have a
clear way of indicating that the speaker is to be true making a statement
that he believes to be true.                                                   (Palmer 1986) 

Here the writer uses the modal adverbs undoubtedly and perhaps which constitute 2
polarities, as undoubtedly expresses a strong level of modality, while perhaps – a weak
level of modality.

The same can be said about the next example. 

b) Although all languages probably have devices for asking questions,
and although there is certainly a need to recognize the typological catego-
ry of Interrogative, the ways in which questions may be expressed vary
greatly. (Palmer 1986)

On the other hand, writers may use adverbs of the same level of modality to make a

greater impact on the reader.

a) The essential point here is that the absence of marking categories
such as number and tense is precisely the defining characteristic of non-
finite forms, and non-finite forms typically occur in subordinate construc-
tions. More strikingly, perhaps somewhat paradoxically, such forms are
used with coordination, but not with subordination in some Papuan lan-
guages. (Palmer 1986)

Armenian Folia AnglistikaLinguistics

77



b) Almost certainly, no synchronic explanation can be given for the use
of past tense for unreality without being circular, but it may be possible to
speculate, very tentatively, along the following lines, why the situation has
come about historically. (Palmer 1986)

Epistemic expressions in language, then, are likely to originate in such high-level

conceptual operations as well. Hence, an investigation of them can probably not be lim-

ited to a purely linguistic description, but will obviously require a concern with the con-

ceptual systems from which they derive.  

Lyons employs the term “modally harmonic” to refer to contexts where two or more

forms express the same degree of modality, such that “there is a kind of concord running

through the clause, which results in the double realization of a single modality”. Thus

clusters overwhelmingly function to reinforce the epistemic strength of items, what

Banks refers to as “fertilizing hedges”.

As for the classification of hedges, it should be mentioned that there is no unanimous

agreement on it, since there are no unified criteria for the classification of hedges. Prince,

Fraser and Bosk (1982) classified hedging devices into two kinds: 1) approximators that

express fuzziness within the propositional content, i.e. affect the truth-value of proposi-

tions; 2) shields that express fuzziness in the relationship between the propositional con-

tent and the speaker/author. Despite the seeming clearness, the taxonomy was criticized

because of the obvious overlap between the categories. As Hyland has pointed out, hedg-

ing devices “do not fit into a neat scheme of discrete categories which allows one mean-

ing to be clearly distinguished from others” (Hyland 1996:437). Moreover, meanings are

frequently expressed simultaneously. The reason for this difficulty is that hedging

devices are polypragmatic. The two main categories of hedges distinguished in the con-

tent of research articles by Hyland are content-oriented hedges which mitigate the rela-

tionship between propositional content and a non-linguistic mental representation of real-

ity and reader-oriented hedges, which mostly deal with the interpersonal interaction

between the reader and the writer. Now let us observe the following examples:

a) The investigation of Morgan (1998) on mathematical writing is prob-
ably the most relevant research, as it followed the same CDA framework.

(Wagner and Herbel-Eisenmann 2008)

b) Perhaps in this kind of highly standardized production environment,
“thinking outside of the box” to design new and creative production
processes only results in processes that are less efficient and, as a result,
more variable over time. (Hofmann and Jones 2005)

In these examples the writers make clear that they have reservations concerning

whether the stated situations actually obtain. 

In general, content-oriented hedges are a significant communicative resource as they

enable scientific writers to both negotiate the precision of claims and convey an attitude

Armenian Folia Anglistika Linguistics

78



to them, offering the reader a means of assessing the certainty of its truth and degree of

personal commitment to it. 

Summing up, it should be mentioned that the investigation is carried out on a corpus com-

prising research articles from three different disciplines, namely linguistics, economics and

physics. We have selected 10 research articles from each discipline. The research articles we

have selected had to comply with a number of requirements in order to be included in the cor-

pus: all of them came from scientific refereed journals, their publishing span was between

1999 and 2012, the average length of each article, which has already been modified, is about

6,500 to 7,500 words. The modification includes deletion of all the material that is not the

author’s coherent writing such as citations, examples, or the list of references.

Our analysis of the modal adverbs used as hedges in research articles demonstrates

that the frequency of their occurrence is determined by the discipline. The average num-

ber of modal adverbs used per 1.000 words is 3.5, 3 and 0.8 in linguistic, economic and

physics articles respectively.

These results quite clearly indicate that modal adverbs are used slightly more fre-

quently in linguistic articles than in economic articles and that physics articles had the

lowest frequency of occurrence of hedges. Based on these findings one might speculate

that, in comparison to the articles on physics, linguistic and economic articles are some-

what more flexible and less formal in their mode of expression.

The scientific paper constitutes a rhetorically competent product through which sci-

entific knowledge is negotiated. A successful claim depends on linguistic choices which

appropriately convey the writer’s intention, facilitate the smooth exchange of informa-

tion and create conditions for persuading readers of a point of view.

Thus, hedges are not the negative indicators of imprecision and fuzziness as many lin-

guists think, on the contrary, they are an essential means by which speakers and writers

encode a point of view.

References:

1. Brown, P. and Levinson, S. (1987) Politeness. Cambridge: Cambridge-University
Press.

2. Coates, J. (1983) The Semantics of Modal Adverbs. London: Croom Helm Series.
3. Crismore, A. & Farnsworth, R. (1990) Metadiscourse in Popular and Professional

Science Discourse. Newbury Park, CA: The Writing Scholar: Studies in the
Language and Conventions of Academic Discourse. 

4. Hyland, K. (1998) Hedging in Scientific Research Articles. Amsterdam/Philadelphia:
John Benjamins Publishing Company.

5. Hubler, A. (1985) Understatements and Hedges in English.
Amsterdam/Philadelphia: Pragmatics and  Beyond  IV, 6.

6. Lyons, J. (1977) Semantics. Vols 1 & 2. Cambridge: Cambridge University Press.
7. Lakoff, G. (1972) Hedges: A study in meaning criteria and the logic of fuzzy con-

cepts. Chicago Linguistic Society Papers, 8 Chicago: Chicago University Press.

Armenian Folia AnglistikaLinguistics

79



8. Quirk, R.; Leech, G.; Greenbaum, S. and Svartvik, J. (1985) A Comprehensive
Grammar of the English Language. London/New York: Longman.

9. Stubbe, M. & Holmes, J. (1995)  You Know, Eh and other “exasperating expres-
sions”: An  Analysis of  Social and Stylistic Variation in the Use of Pragmatic Devices
in a Sample of New Zealand English. Elsevier Ltd: Language and Communication:
15 (1).

Sources of  Data:

1. Crystal, D. (1995) The Cambridge Encyclopaedia of Language. Cambridge: CUP.
2. David, J. MacKay (2003) Information Theory, Inference and Learning Algorithms.

73. Cambridge: CUP.

3. Hoffman, D. & Jones, L. (2005) Leadership, Collective Personality, and
Performance. // Journal of Applied Psychology. Vol.90, No.3.

4. Palmer, F. (1986) Mood and Modality. Cambridge: CUP.
5. Riess et al., (1987) Observational Evidence from Supernovae for an Accelerating

Universe and a Cosmological Constant. // The Astronomical Journal. Vol.116. USA.
6. Wagner, D. & Herbel-Eisenmann, B. (2008) Just Don’t: The Suppression and

Invitation of Dialogue in the Mathematical Classroom. Educational Studies in
Mathematics. Vol.67, No.2.

ºÕ³Ý³Ï³íáñáÕ μ³é»ñÇ ·áñͳÍáõÃÛáõÝÁ ·Çï³Ï³Ý ·ñ³Ï³ÝáõÃÛ³Ý Ù»ç

¶Çï³Ï³Ý μÝáõÛÃÇ ï»ùëïáõÙ Ñ»ÕÇݳÏÇ ·É˳íáñ Ýå³ï³ÏÝ ¿ Ñݳñ³íáñÇÝë
ѳÙá½Çã Ý»ñϳ۳óÝ»É Çñ ï»ë³Ï»ïÁ` ÷áñÓ»Éáí Ý»ñ³½¹»É ÁÝûñóáÕÇ íñ³: ø³ÝÇ
áñ ·ñ³íáñ ËáëùáõÙ  ³ÝÑݳñ ¿ ³ÝÙÇç³Ï³Ý ß÷áõÙ ³å³Ñáí»É ÁÝûñóáÕÇ ¨ Ñ»ÕÇ-
ݳÏÇ ÙÇç¨, ϳñ¨áñ ¿ ëï»ÕÍ»É »ñÏËáëáõÃÛ³Ý ¹³ßï, áñï»Õ ï»Õ»ÏáõÃÛáõÝÁ áñ³-
ϳíáñ»Éáõ ѳٳñ Ñ³×³Ë û·ï³·áñÍíáõÙ »Ý ѳÙá½í³ÍáõÃÛáõÝ, ³ÝáñáßáõÃÛáõÝ,
³Ýíëï³ÑáõÃÛáõÝ ¨ Ùáï³íáñáõÃÛáõÝ ³ñï³Ñ³ÛïáÕ μ³é»ñ: ²Ûë ï»ë³ÝÏÛáõÝÇó ³-
é³ÝÓݳѳïáõÏ Ñ»ï³ùñùñáõÃÛáõÝ »Ý Ý»ñϳ۳óÝáõÙ »Õ³Ý³Ï³íáñáÕ μ³é»ñÁª Ç
ßÝáñÑÇí Çñ»Ýó ϳñ¨áñ ï»ùëï³Ï³½ÙÇã ¹»ñÇ:

Armenian Folia Anglistika Linguistics

80