Applied Linguistics 28/2: 266285

doi:10.1093/applin/amm011

Oxford University Press 2007

Applying a Gloss: Exemplifying and

Reformulating in Academic Discourse
KEN HYLAND
University of London
A great deal of research has now established that written texts embody
interactions between writers and readers, but few studies have examined the
ways that small acts of reformulation and exemplification help contribute to
this. Abstraction, theorisation and interpretation need to be woven into a
text which makes sense to a particular community of readers, and this invariably
involves frequent reworkings and exemplifications as writers assess the
processing needs, knowledge and rhetorical expectations of their readers to
present and then interpret ideas as they write. Known as code glosses in the
metadiscourse literature, these elaborations help to contribute to the creation
of coherent, reader-friendly prose while conveying the writers audiencesensitivity and relationship to the message. Drawing on a large corpus of
research articles, I explore how professional academic writers monitor their texts
for readers in this way to restate information or provide examples as they
construct their arguments. Analysis of the corpus reveals that elaboration is
a complex and important rhetorical function in academic writing, and that
both its use and meanings vary according to discipline.

Constructing explanations for events in the natural and social worlds is often
seen as fundamental to scientific discourse. We look to the sciences for
plausible reasons why things are as they are and writers take care to
represent their arguments by identifying, classifying and interpreting features
of the real world in ways which are likely to make most sense, and be
most persuasive, to their particular communities of readers. This interest
in scientific persuasion has produced a substantial literature devoted
to describing the many linguistic and rhetorical resources which contribute
to such argumentative processes. Surprisingly little, however, has been
written about the many small acts of elaboration, embellishment, and
clarification which occur in the process of creating a plausible argument.
Abstraction, theorisation and interpretation need to be woven into an
overt framework of argument so that an explanation makes sense to its
intended readers, and this often involves frequent expansions, reworkings,
and exemplifications as writers constantly present and then interpret ideas on
the fly. Collectively known as code glosses in the metadiscourse literature
(Hyland 2005), these brief reformulations and exemplifications help to
contribute to the creation of coherent, reader-friendly prose while conveying
the writers audience-sensitivity and relationship to the message. This article

KEN HYLAND

267

is concerned with such acts of interaction, local elaboration, and on-line

clarification. Drawing on a large corpus of research articles, I explore how
professional academic writers constantly monitor their texts to restate
information or provide examples as they construct their arguments. I will
first look briefly at the nature of these acts and then go on to discuss
the study.

ELABORATION AND CODE GLOSSING

Every academic text is written to be both understood and accepted, and
while neither goal is ever completely assured, writers who can successfully
predict something of what their readers will know of their subject
and expect of its presentation are more likely to be convincing. Based
on their assumptions of their readers and their previous experiences
with similar texts, writers constantly monitor their unfolding discourse
to address these expectations, making rhetorical choices which
negotiate appropriate engagement and explicitness. They identify
where readers will need help in interpreting points, where greater
elaboration or specificity is required, where clarification or examples are
needed, and so on. In other words, arguments are underpinned and
supported by small acts of propositional embellishment which serve to
enhance understanding, shape meanings more precisely to the writers goals,
and relate statements to the readers experience, knowledge-base, and
processing needs.
Halliday (1994: 225) discusses this kind of propositional expansion as a
semantic relationship of elaboration, where one clause elaborates on the
meaning of another by further specifying or describing it. In contrast to cases
where a second unit extends the meaning of a previous unit by adding
something new, or qualifies it by reference to time, place, manner, condition,
etc., the secondary clause does not introduce a new element but provides a
further characterisation of one that is already there, restating it, clarifying it,
refining it, or adding a descriptive attribute or comment (1994: 225).
These examples from my research article corpus illustrate this kind of
rhetorical work:
(1) conditional statements merely construct the dependence
of one proposition on the truth of another. Or, in slightly
different terms, the if clause sets up an imaginary world in which
the proposition in the then clause is the case . . .
(applied linguistics article)
These have rested on assumptions that power is zero-sum, that
is, a finite resource which people cannot share. (sociology article)
An annual production capacity of 4000 hours (50 weeks
80
hours/week) was assumed for the system.
(electrical engineering article)

268

EXEMPLIFYING AND REFORMULATING IN ACADEMIC DISCOURSE

A farm that is a cash grain enterprise is about 24% more likely

to adopt no-tillage than, say, a dairy farm, while the
type of ownership (e.g. single owner versus partnership) of
the farm has no effect on the conservation tillage adoption
decision.
(mechanical engineering article)
As can be seen, these meanings can be expressed by the juxtaposition of
noun phrases, in academic discourse often marked off by punctuation,
with the second part enclosed between brackets andoccasionallydashes
or commas. More often, they are signalled explicitly by what Downing
and Locke (1992: 283) call an elaborating conjunction and Blakemore
(1993) an apposition marker, devices such as that is, for example, namely
and in slightly different terms which link such clauses overtly. These
conjuncts provide cohesive, rather than structural, linking and offer
metadiscursive cues which seek to reduce the possibility of pragmatic
ambiguity.
Such connections have traditionally been classified as appositive (Quirk
et al. 1972: 620), where the second unit of text is to be treated as equivalent
to or included in the previous unit (Biber et al. 1999: 876). This classification,
however is both too restrictive, as the connection can also occur across
sentence boundaries (see Burton-Roberts 1993), and too vague, and it is
more productive to see these lexical and parenthetical signals as contributing
to the class of communicative resources known as metadiscourse markers
(Hyland 2004, 2005; Hyland and Tse 2004).
Essentially, metadiscourse is self-reflective matter which makes reference
to the evolving text or to the writer and imagined reader of that text. It
therefore acts to connect, organise and interpret material with regard to
the understandings and values of a particular discourse community
(Hyland 2000). More specifically, these features are examples of code glosses,
or items which supply additional information by rephrasing, explaining or
elaborating what has been said to ensure the reader is able to recover
the writers intended meaning. Such items facilitate argument in
academic discourse by contributing to the logic of the unfolding discourse
rather than to the logic of events as they occur in the real world (Martin
and Rose 2003).

REFORMULATION AND EXEMPLIFICATION

Code glosses are, ostensibly at least, almost always concerned with
clarification of the writers communicative purpose. The term represents a
number of basic communication strategies used in the negotiation of
meaning in many different contexts, occurring in both spoken and written
language, to facilitate the readers understanding. In this paper I will be
concerned with two broad sub-functions of this purpose: reformulation and
exemplification.

KEN HYLAND

269

Reformulation
Reformulation is a discourse function whereby the second unit is a
restatement or elaboration of the first in different words, to present it from a
different point of view and to reinforce the message. In academic writing
such connections are often signalled parenthetically or lexically by what
I shall call reformulation markers as in (2):
(2) Between what Braj Kachru (1988) appropriately calls the
Outer Circle, or the countries where English was brought by
colonization, and the Expanding Circle . . .
(applied linguistics article)
They argued, on the basis of the emerging survey data, that
drug use by young people is becoming so common that it is no
longer regarded as a deviant activity by them. Put another way,
they claim that drug use among young people is becoming
normalised.
(sociology article)
The term natural then functions as a mythic construct in the
context of fashion discourse (Barthes 1983), that is, an amorphous
ideal whose form is continuously reformulated in ways that
sanction present-day standards.
(marketing article)
Reformulations have largely been treated as repairs in unplanned discourse
(e.g. Schegloff et al. 1977), displaying the speakers recognition that the
original formulation was not an appropriate means of achieving communicative success or, in Blakemores (1993: 101) terms, that it did not achieve
optimal relevance. Reformulations in writing, however, must be seen as
part of a plan and therefore purposeful, indicating that the writer is seeking to
convey particular meanings or achieve particular rhetorical effects. Essentially,
reformulation is a discourse function whereby a writer re-elaborates an idea
to facilitate comprehension. This is common in research seminars where
members reformulate other speakers previous utterances in working towards
comprehension (so you are telling us. . ., what you are saying is. . .) (Gonzales 1996
reported in Swales 2004). It tends, therefore, to be regarded as an equivalence
operation so that the two units are different ways of expressing a single idea,
as in this example:
(3) Of a sample of 927 consumers, 844 respondents, or 91%, had
heard of organic foods.
(marketing)
Formulation A (844 people out of 927 in a sample) formulation
B (91% of the sample).
But while adjacent text elements may be about the same thing in that the
subject matter remains the same, meaning is not created solely through

270

EXEMPLIFYING AND REFORMULATING IN ACADEMIC DISCOURSE

propositional material. Rather, meaning is the result of an interactive process

between the producer and receiver of a text in which the writer chooses
formulations which are most likely to secure understanding and agreement.
Reformulations can thus serve a variety of functions beyond summarising or
gisting. While writers may present two versions of the same material,
alternative formulations of a single idea rarely constitute identical meanings
and tend to go beyond strict paraphrase to present what the writer considers
to be the key elements of a prior utterance. I shall return to these meanings
in a later section.

Exemplification
Exemplification is a communication process through which meaning
is clarified or supported by a second unit which illustrates the first by
citing an example. While example can refer to larger discourse units
including case studies, I am restricting its use here to refer to a group, clause,
word, or numerical figure. Once again, these can be signalled by
punctuation, normally parentheses, by linking adverbials and by abbreviations such as, e.g., I collectively refer to these as exemplificatory markers
(Lee 2004: 298):
(4) In recent decades, this pattern has been complicated in most
states by developments such as the lowering of the legal drinking
age and the extension of trading hours for many venues to early
morning or all night.
(sociology)
. . . students find reading in English to be difficult and that
self selected reading did not seem as valuable as other activities
(e.g., required school work, TOEFL study).
(applied linguistics)
Here again, operators like A & P, Dominicks, Jewel, Safeway, Tom
Thumb and Vons have continued with their Hi-Lo strategy and are
successful.
(marketing)
Other units get changed to a more dramatic extent: SI units for
moment of inertia, for example, becoming kg.m rad.
(physics)
Exemplification is a recurrent feature of academic writing and a central
aspect of exposition, a part of the routine ways in which writers in all
fields seek to make their ideas accessible and persuasive. Essentially it is
an appeal to understandings the writer believes are recoverable from
the example: presenting an element of the writers data or experience to
make the abstract more concrete. As a result, it reveals something of the
writers predictions about the readers familiarity with the topic and world
knowledge.

KEN HYLAND

271

It is perhaps clear from the examples above, however, that reformulation

and exemplification are not simple discourse functions but complex rhetorical
categories which can have a range of meanings. I will return to what these
meanings might be in academic writing after describing the text corpora
which forms the basis of the analysis.

CORPUS AND PROCEDURES

The study is based on a research article (RA) corpus of 240 published papers,
three from each of ten leading journals in eight disciplines totalling
1.4 million words. The value of exploring such a large corpus is that it makes
available many instances of the target features in a naturally occurring
discourse, replicating the language using experience of community members.
The disciplines were selected to allow comparisons across a range of
academic knowledge and rhetorical practice, comprising mechanical
engineering (ME), electrical engineering (EE), marketing (Mk), philosophy
(Phil), sociology (Soc), applied linguistics (AL), physics (Phy), and
microbiology (Bio). The journals were nominated by informants as among
the leading publications in their fields, and articles were selected at random
from current issues.
All the texts were electronically scanned and then searched for specific
features which could potentially act as elaborating conjunctions, including
abbreviations and punctuation. A list of 45 potentially productive search
items was compiled based on those listed in grammars (Biber et al. 1999;
Dowling and Locke 1992; Halliday 1994; Greenbaum 1996), previous work
on metadiscourse (e.g. Hyland 2000/2004, 2004), and a close study of the
texts themselves. All cases were examined in context to ensure they
functioned to either reformulate or exemplify statements in the texts and
a small sample was double-checked by a colleague working independently
to verify these interpretations.

ELABORATION IN RESEARCH ARTICLES: CORPUS FINDINGS

Analysis of the article corpus shows that providing examples and
reformulating utterances is an important feature of academic writing, and
indeed, they appear to be more common in academic prose than in other
registers (Biber et al. 1999: 884). There are almost 6,000 occurrences in the
corpus, roughly 25 in each paper, with writers in applied linguistics and
marketing employing the most signals overall (38 per cent of the total) and
67 per cent of all devices occurring in the humanities and social science
papers. Table 1 shows that writers in all the disciplines examined employed
both forms of elaboration and that overall, the so-called hard and soft
knowledge fields contained a similar density of glosses per 10,000 words.
The table also highlights that exemplification was about 16 per cent more
common overall, underlining the importance of clarifying propositions

272

EXEMPLIFYING AND REFORMULATING IN ACADEMIC DISCOURSE

Table 1: Code gloss markers by discipline (per 10,000 words)

Function

Phy

Bio

ME

EE

Phil

Soc

AL

Mkt

Totals

Reformulation
Exemplification
TOTAL

34.9
16.7
51.9

23.0
13.2
36.2

24.2
17.2
41.5

21.7
19.6
41.3

15.3
25.2
40.5

14.7
25.2
39.9

21.0
32.0
53.0

20.7
32.9
53.6

20.6
24.5
45.0

through illustrative material. A closer look, however, shows the imbalance is

largely due to the heavy presence of exemplification in the humanities and
social science articles. In fact, the figures show a rather neat division between
the hard and soft knowledge fields here, with 61 per cent of features in the
science and engineering papers (hard) signalling reformulations and 61 per
cent of those in the humanities and social science papers (soft) indicating
exemplifications.
Interestingly, these preferences point to fundamental differences in the
ways that these broad domains construct knowledge and help to contribute
to our understanding of disciplinary stereotypes.
These differences, at least in part, are a consequence of the fact that the
hard and soft disciplines mediate reality in very different ways. Unlike
scientific knowledge, which tends to be cumulative and tightly structured,
researchers in the humanities and social sciences cannot assume that the
background to a problem, appropriate methods for its investigation, and
criteria for establishing resulting claims are agreed by all readers. Instead, the
context often has to be elaborated anew, its more diverse components
reconstructed for a potentially less cohesive readership (Hyland 2000).
Exemplification plays a role here by helping to index a known and
recoverable reality, keeping the relation between things in the world and
discussion of those things as clear as possible. In other words, examples in
soft knowledge fields represent a heavier rhetorical investment in
contextualisation, perhaps even a need to persuade the reader that the
phenomenon actually exists. Examples are a key means by which writers
engage with their readers in this way, encouraging them to recognise
phenomenon through recoverable experiences and to become involved in
the unfolding text.
I will expand and draw on these general observations about disciplinary
practices below, considering reformulation and exemplification in turn.

REFORMULATION IN ACADEMIC ARTICLES

Reformulation, as I discussed earlier, occurs when a writer rewrites an
utterance by expressing an idea in a different way. Reformulations comprised
about 46 per cent of the total elaborations in the corpus with the
majority in the soft knowledge fields. When we consider the proportion

273

KEN HYLAND

Table 2: Most frequent reformulation markers by discipline (% of total)

Marker

Totals Phy

Parentheses
26.1
i.e. in particular/
25.5
particularly
8.9
that is,
8.6
especially
6.4
in other words
5.3
namely
4.9
specifically,
3.9
or x
3.6
which/this means
3.5
Others
3.3
TOTAL
100

Bio

ME

EE

Phil

Soc

AL

Mkt

35.9 79.1 44.2 39.7

8.4
9.7
6.3
5.2
30.6
7.0 22.3 24.8 28.4 11.9 30.0 40.1
12.4
1.5
2.5
6.4
9.1 16.7
9.5 10.6
1.5
1.2
7.6
6.8
0.9 11.2 22.1 11.5
2.1
4.2
4.3
4.7
4.4 17.3
6.3
6.7
3.3
0.9
3.2
2.6
7.5
7.9
7.7
7.1
4.4
0.6
6.1
3.8 18.1
3.6
3.4
0.7
0.3
0.6
1.4
1.3
3.4
4.0
4.5 11.5
3.2
1.2
3.6
0.9
5.0
8.8
3.6
2.7
3.2
0.6
2.5
3.8 10.3
3.3
2.9
2.0
3.1
3.1
2.3
5.2
4.5
5.6
3.7
1.9
100
100
100
100
100
100
100
100

of cases per 10,000 words, however, almost two-thirds occurred in the

science and engineering papers. Equivalence between two statements is
signalled by reformulation markers and the most common of these in the
corpus were parentheses (26 per cent of all cases), i.e. (25 per cent of cases), in
particular/particularly (9 per cent) and that is (9 per cent), but there were
sizeable disciplinary variations. Table 2 shows the most frequent reformulation markers as a percentage of all such markers by discipline.
There seems to be a marked preference for simple appositional
reformulation in the science and engineering disciplines, where the
reformulation is syntactically separated from the rest of the sentence by
comma intonation (Rouchota 1998), normally brackets. It is interesting to
note that 85 per cent of all cases of this pattern occurred in the hard
knowledge papers. Brackets create a distinctive space within the sentence
which serves to distance the information it contains from the remainder of
the proposition, allowing writers to signal that the enclosure provides
background or illustrative information rather than main ideas. The second
most frequent marker in each science field was the abbreviation i.e. and
together these two signals comprised 72 per cent of reformulation markers
in the four disciplines.
In contrast, there is greater variation in the use of devices in the soft fields,
with the top five markers (that is, i.e., in other words, especially, and brackets)
making up only two-thirds of the total. In all disciplines, however, there is a
clear preference for fixed connectors which exhibit no predictive structure
over more complex forms such as this means that, put another way, to say the
same thing differently, to be more precise, etc. which allow a certain degree
of modification.

274

EXEMPLIFYING AND REFORMULATING IN ACADEMIC DISCOURSE

Explanation
Expansion
Implication
Reformulation
Paraphrase
Reduction
Specification

Figure 1: Discourse functions of reformulations

While there is more to be said about these formal differences, it is more
interesting to look at how they represent underlying semantic preferences.
While the adjacent formulations are notionally equivalent, it is clear that
reformulation can carry a range of meanings. These function to either
expand the original, by explanation or implication, or to reduce it by
paraphrase or specification, as shown in Figure 1.

Expansion
These reformulations restate an idea in such a way as to widen the sense in
which the writer intends it to be understood. What is being expanded, then,
is the readers understanding rather than an idea or a locution. This is
achieved here by offering an explanation or by drawing an implication from
it, thus increasing the accessibility of the original or highlighting the writers
preferred understanding of its meaning.

Explanation
These are situated clarifications which elaborate the meaning of a preceding
unit to make a concept more accessible by providing a gloss or a definition.
Because they function to expand the readers understanding of material,
rather than the material itself, explanations include cases where the
reformulation provides clarification for a technical term (as in the first
example in 6 below) and also where the reformulation provides a technical
term for a concept already expressed in more accessible terms. Once again,
parentheses and that is are common signals, as are defining items such as
known as, called, and referred to as:
(5) Among blacks, increases in nonmarriage have accounted for
the overwhelming share of the post-1960 rise in the nonmarital
fertility ratio, that is, the ratio of nonmarital births to all births.
(Soc)
Due to the lack of success in using several conventional methods,
an unbiased recognition algorithm is proposed based on a novel

KEN HYLAND

275

statistical feature point recognition principle, called the maximum

principle of slope difference.
(EE)
One of the most significant, which has received wide attention, is
the formation of arrow shape voids in the central axis of the
extrudate, known as central bursting or chevroning.
(ME)
The possession of knowledge concerning ethical and
unethical behavior did not, according to Simpson, Banedee,
and Simpson (1994), affect stealing microcomputer software
(softlifting).
(Mkt)

Implication
This function serves to draw a conclusion or sum up the main import of the
prior segment. Despite the signal in other words which is often used here,
the notion of equivalence between statements here extends a rephrasing
to a conclusion of what the writer wants readers to take away from the
statement, as in these examples:
(6) Many reports show growth depressions in AM plants
during the first several weeks of seedling development
which disappear as internal seed reserves become depleted
(e.g. Bethlenfalvay, Brown & Pacovsky, 1982; Koide, 1985).
In other words, short term losses are often compensated by long-term
gains.
(Bio)
She was in direct control of something of which Dans death was a
consequence, and only in this way did she have control over
Dans death. This means that Dans death was not in Shirleys control
except insofar as this something was in her control.
(Phil)
In order to understand the complex decision-making
process involved in the organization of written text, a pedagogical
theory of L2 writing needs to look beyond the background
of ESL writers. In other words, an alternative theory of L2 writing
is needed.
(AL)

Reduction
Reformulations in this category serve to restrict the meaning of what has
been said, narrowing the scope of interpretation by either paraphrase or
specification.

276

EXEMPLIFYING AND REFORMULATING IN ACADEMIC DISCOURSE

Paraphrase
As the example in (2) above illustrates, one function of reformulations is
gisting or restating an idea in different words to provide a summary.
Typically this function is signalled by the forms that is, in other words,
put another way and the use of parentheses:
(7) These people are often active in social change organisations,
but their theories do not provide intellectual support for their
actions, or put differently, do not explain their practices to them.
(Soc)
12 of the 18 (67 per cent) crimes were rape and murder, or a
combination relating to a sexual encounter.
(Soc)
However, the value of past participants as information carriers and
the persuasive influence of supporters in senior managerial
positions soon became abundantly apparent. That is to say, the
value of an internal network of committed true believers in contributing
to this border crossing activity became abundantly clear.
(Mkt)

Specification
Here the reformulation functions not to restate an idea but to further detail
features which are salient to the primary thesis in order to constrain how the
reader might interpret it. By specifying the statement more precisely, the
writer is able to both highlight the specification and simultaneously include it
within the scope of the original formulation. The italicised reformulations in
these examples illustrate this point:
(8) The Britishor more especially the Englishconstitution is
founded on the principle that no Briton is to be obedient to a
foreign prince . . .
(Phil)
. . . they refer to psychoanalysis, to existential phenomenology and
to Marxism (in particular to the earlier works of Marx).
(Soc)
As a result, implementation of a commercial lighting program can
affect the costs and benefits of different stakeholders. Specifically,
it can affect utility rates, the total resource cost to the society, the utility
expenditures, and the total cost to all customers.
(ME)
. . .. is modulated by radiation damping during t (or more accurately
speaking, during the second t/2).
(Phy)

KEN HYLAND

277

Strictly then, reformulation implies X is equal to Y, yet the process of

repackaging information rarely results in a statement which has an exact
equivalent meaning. Rephrasing invariably alters the pragmatic and
rhetorical connotations of the original by moving the reader towards
the writers preferred interpretation. As Cuenca (2003: 1072) observes,
the connective does not express equivalence, but creates the equivalence.
In other words, reformulation in academic writing is a pragmatic or
discoursal equivalence rather than a strictly logical or propositional one. It is
a rhetorical sleight of hand, and readers may need to be cautious in accepting
it at face value.
In fact, when we look closely at the frequencies for these particular
functions, we find that the paraphrase (or gisting) meaning of reformulations is actually relatively rare in the corpus, turning out to be the least
frequent use in all disciplines. Table 3 shows the percentages of each of the
functions listed above based on a close analysis of a randomly generated
sample of 20 per cent of the instances of each feature in each discipline, some
580 items overall. As can be seen, specification is the most common function
of reformulation, accounting for half of all cases in the sample and
emphasising the importance academic argument attaches to precision,
to restricting interpretation and to highlighting the writers understanding
of phenomena.
While specification is the most widely used, and paraphrase the least used,
function overall, it is interesting to note that reformulation is often expressed
differently across disciplines. Physicists seem to employ more explanations,
mechanical engineers more paraphrase, and biologists more specification.
While the figures are too small for a firm assertion here, the distribution
suggests that writers in the soft fields were far more likely to reformulate a
statement as an implication. This might be because they are less able to
confidently presuppose that their readers will possess the necessary
background knowledge to make connections for themselves. Unlike the
hard knowledge fields, research does not always occur within the kinds of
established theoretical frameworks which provide an explanatory schema for
claims (e.g. Kuhn 1970). Linearity and predictability is less assured and

Table 3: Proportions of reformulation functions by discipline (%)

Sub-function

Totals

Phy

Bio

ME

EE

Phil

Soc

AL

Mkt

Explanation
Implication
Paraphrase
Specification
Totals

24.7
17.7
6.7
50.9
100

43.3
13.4
7.5
35.8
100

22.0
5.9
1.5
70.6
100

20.8
17.0
15.1
47.1
100

25.5
12.8
4.3
57.4
100

15.6
26.6
10.9
46.9
100

16.9
20.0
7.7
55.4
100

26.0
24.0
4.4
45.6
100

27.5
22.0
2.2
48.3
100

278

EXEMPLIFYING AND REFORMULATING IN ACADEMIC DISCOURSE

writers draw on a literature which is more dispersed and open to greater

interpretation.
Writers therefore often need to take greater care in situating their research
to demonstrate a plausible basis for their claims, spelling out an inference or
repercussion explicitly rather than running the risk of the reader missing
it altogether.

EXEMPLIFICATION IN ACADEMIC ARTICLES

Exemplification is the second broad type of elaboration in research articles,
representing an appeal to a more familiar and concrete experience which
overrides divergent perceptions. Like reformulation, exemplification plays a
key part in the interactive process between reader and text as the writer
anticipates and responds to the readers possible need for clarification and
provides a more accessible way of perceiving it. It therefore offers an insight
into the writers reading of the audience and what is likely to be known and
persuasive to it. The illustration is assumed to be helpful to the readers
processing of the text and to promote the writers perspective by furnishing
information which may be available but currently not in the readers
consciousness.
Examples in the corpus were signalled in a limited number of ways, with
three-quarters of cases indicated by just three markers: such as, for example,
and e.g. Table 4 shows, however, that the proportions of these markers varied
across disciplines, with mechanical engineering heavily favouring such as,
marketing preferring e.g., philosophy for example, and electronic engineers the
fixed phase an example of. In terms of raw numbers, however, these
proportions mask even bigger disciplinary disparities, with over 80 per cent of
all uses of such as, for example, e.g., and for instance occurring in the soft
knowledge papers.

Table 4: Most frequent exemplification markers by discipline (% of total)

Such as
For example
e.g.
An example of
Like
For instance
say
Others
TOTAL

Totals

Phy

Bio

ME

EE

Phil

Soc

AL

Mkt

29.8
23.8
20.9
9.6
6.3
5.3
1.5
2.8
100

28.2
27.6
15.5
12.1
4.5
4.5
2.1
5.5
100

25.4
23.0
25.9
10.8
7.4
5.3
0.0
2.1
100

48.5
21.0
10.5
10.5
1.0
2.0
2.0
4.5
100

28.6
14.4
16.4
21.9
10.2
2.4
1.4
4.7
100

18.6
29.1
13.3
6.7
14.5
11.5
3.6
2.7
100

27.5
24.3
26.4
10.3
4.3
4.7
0.6
1.9
100

28.7
28.1
26.5
2.5
4.7
5.3
0.5
3.7
100

31.7
23.1
32.8
2.1
2.1
5.7
0.6
1.2
100

KEN HYLAND

279

As noted earlier, exemplification is a particular feature of argument in the

soft knowledge fields, where these markers comprised almost two-thirds of
all elaborations. In these disciplines examples are brief, practical expressions
embedded in the on-going text to illustrate and support a prior locution.
They work in three main ways:
1 by offering an instance of a general category;
2 by providing a parallel or similar case;
3 by giving a precept or a rule.
In the first category, which comprised the overwhelming majority of cases,
the writer elaborates a statement by citing a particular instance to represent
the category of events, people or phenomena mentioned in the immediately
prior text.
(9) In the modern sense, systems thinking has provided
a framework for empirical enquiry in quite diverse disciplines
like Gestalt psychology (Koffka 1935) and biology (von Bertalanffy
1950) . . .
(Mkt)
Some philosophers (e.g., Block) have recently worried that
mental properties or the functionalist conception of them are
epiphenomenal.
(Phil)
The lecturer made several references to himself in the lecture as
a way of establishing personal contact with his audience.
At one point, for example, the lecturer said the reason he was in
Hong Kong was related to the Asian economic miracle. He also
periodically referred to the students, for example saying at one
point, Youre in your second year now so you should know what
I mean.
(AL)
Rowdy acts of misbehaviour, like pushing, arguing, swearing, loudness
and obscenity, are all valued for being part of a continuum of social
rule-breaking which heightens the pleasurable experience of
drinking as time out.
(Soc)
A second, less frequent, type of example helps to provide a more general
sense of the target concept by assembling resources which relate the
unfamiliar to the more recognizable through a parallel or closely similar case,
as in (10):
(10) A tradition may be reflective and designed, like the
deliberations of the Supreme Court, or unreflective and spontaneous,
like sports fans rooting for their teams; it may have a formal
institutional framework, like the Catholic Church, or it may be
unstructured, like mountain-climbing; it may be competitive, like the

280

EXEMPLIFYING AND REFORMULATING IN ACADEMIC DISCOURSE

Olympics; largely passive, like going to the opera; humanitarian, like

the Red Cross; self-centred, like jogging; honorific, like the Nobel Prize;
or punitive, like criminal proceedings.
(Phil)
Euler knew that numbers of the form a b3 form a number
system much like that of the integers.
(Phil)
Overall sociological understanding is extended considerably
by an analysis of its links to cultural resistance and protest
masculinity. But, like accounts which focus too exclusively on the
role of male honour in interaction, they also risk trivializing
victimization.
(Soc)
A third
concept
making
concept

type of example provides a looser kind of association between a

and an instance by illustrating a principle or the application of a rule,
available for readers an image or quality which conjures up the
or practice referred to:

(11) By the same token, putting images in the context of a

story vastly improves recognition, as does depicting objects as
parts of larger scenes (e.g., a tulip in a vase will not be taken for
a fork).
(Phil)
There can be some crossover effects in which postdecision
evaluations of, say, A affect satisfaction with choice B.
(Mkt)
In the experience of any change we may identify a particularly
salient point, such as the moment a long-distance runner crosses the
finishing line.
(Phil)
In the sciences such examples are far less common, partly because these
disciplines deal with more abstract and technical representations of reality
which are less amenable to the creation of instant, in situ concrete examples.
Where they do occur, they almost always fall into the first category
mentioned above, supplying a particular instance to enable the initiated
reader to more fully grasp the concept, process, or phenomenon referred to.
They differ, however, in drawing on a more disciplinary specialised range
of items, rather than the often mundane and everyday ones of the social
sciences, to help readers translate the text into specific actions or things.
These examples help to clarify this:
(12) When values of B become larger, e.g., B 50006900 G, while
the asymmetry is still clearly observable, the general pattern of C3
(B) and C31(-B) becomes similar.
(Phy)

KEN HYLAND

281

We take the (mathematical) point of view that all the familiar

transcendental functions (e.g., sin and all trigonometric, exponential,
hyperbolic, logarithmic, Bessel, etc., functions) must have an argument
which is a pure number and therefore cannot have any
dimensions.
(EE)
Various thermochemical conversion systems such as combustors and
gasifiers have been designed and tested for the recovery of energy
from various biomass materials.
(ME)
When nuclear lysates were prepared in the absence of
phosphatase inhibitors (like sodium vanadate or sodium fluoride),
formation of the LIF-dependent complex was impaired.
(Bio)
These examples refer to things and experiences which are more closely
embedded in the understandings and frameworks of the disciplines
themselves. Scientific and engineering audiences generally share an
acceptance of considerable knowledge of materials, procedures, and the
criteria used to assess research practices and claims because of a more linear
and problem-oriented approach to knowledge construction. Natural scientists
tend to see knowledge as developed through relatively steady cumulative
growth (Becher 1989) and the highly specialised, resource intensive, and
expensive nature of much scientific research means that problems regularly
emerge in an established context. As a result, readers are often familiar with
prior texts and research which not only makes the novelty and significance
of contributions easy to recognise, but also allows arguments to be
formulated in a highly standardised code. Writers can therefore have some
confidence that the examples that they employ will resonate with the
understandings of their readers.
It is important to note that such examples also carry considerable empirical
authority which helps to contribute to the apparently strong claims of the
sciences. Tying examples to the writers data helps reinforce the readers
acceptance of the evidential weight of the interpretation. Scientists and
engineers therefore appeal to specific understandings, rather than more
general areas of knowledge as writers in the humanities and social sciences
often do, to provide an additional persuasive dimension to their discourse.
This rhetorical practice both eliminates idiosyncratic choice, and thus
minimises the presence of the writer, and reinforces the writers argument
with the primacy of evidential truth:
(13) Parasitic mycorrhizal associations can occur at particular
stages in the development of the association. For example, formation
of arbuscular mycorrhizas can depress seedling growth in the first few
weeks following germination.
(Bio)

282

EXEMPLIFYING AND REFORMULATING IN ACADEMIC DISCOURSE

Equation (4) relates explicitly to a duct with an adiabatic

outer surface. It applies to the same extent to a typical channel
of, say, the honeycomb structure of the rotating core of a Ritz-type
regenerator.
(EE)
Nevertheless, even when weak interactions occur between a thin
solid polymer film and its interfaces, dramatic large-scale effects
can occur, e.g. phase separation can be induced in polymer blends simply
by reducing the film thickness.
(Phy)
Such examples are almost never hedged, but present strong, unmitigated
support for the arguments they accompany.
But examples not only function to elaborate statements. In all academic
disciplines they serve to interrupt and break up generalised and conceptual
passages with solid references to real life, or at least to more accessible
phenomena, allowing readers to use their senses as well as their minds.
Killingsworth and Gilbertson (1992: 153), in fact, argue that this is a major
function of examples. They believe examples are a key component of
academic persuasion as they lead readers to a correct version of the sense
experience and save them the trouble of having to do their own
translating of abstract material into language more immediately accessible
to the senses.
The following examples give some flavour of this, as writers intersperse a
concrete instance in a more abstract account:
Clearly, researchers developing models of post-choice valuation in
period on word-of-mouth in period t 1 should control for the
influence of both chosen and forgone alternatives word-of-mouth
in period t 1 on post-choice valuation in period t. For managers,
our results suggest the ability of disappointment- and regretoriented messages to influence the post-choice valuation of
consumers who have chosen a manufacturers brand and of
consumers who have not chosen the brand. For instance, in a recent
advertisement, AT&T depicts a consumer who remained with AT&T long
distance praising their service . . .
(Mkt)
Although the interlanguage continuum provides a perspective
on the regularities in second language learners acquisition
patterns, there is an apparent contradiction between this
systematicity and the variation that is evident in, and characteristic of, learner language. Variation is exhibited when learners
apply a rule in one context but not in another. For example,
a learner who writes You depend on other people and they depend of
you has produced the target language rule in the first clause but not
in the second.
(AL)

KEN HYLAND

283

For all point-propositions W, and value functions V, V(W) is

supposed to remain constant no matter what the agent learns. Call
this constancy; it is a crucial assumption for all Lewis-style antiDAB proofs. Constancy is, in practice, a useful simplification.
When decision theory is used to model problems of rational choice
(for example, whether to go to a party, the pub, or to see Lara) it is
convenient to assume . . .
(Phil)
Examples can therefore make other channels of persuasion available to
support the writers argument, diluting unrelieved technical argument or
flights of abstraction with specific, material description with which the reader
can more immediately identify.

SOME OBSERVATIONS AND CONCLUSIONS

I have sought to make two broad claims in this paper: that elaboration is a
complex and important rhetorical function in academic writing, and that its
use varies according to discipline.
Both frequency information and pragmatic interpretation point to the
routine significance of elaborative code glosses in the argumentation practices
of all disciplines. Seeking to craft a text which is both comprehensible
and persuasive, writers must recognise the processing needs, rhetorical
expectations and argument preferences of their readers, and one way they
achieve this is by frequently reworking utterances to offer a reformulation
or concrete instance of what they have said. Such re-workings are an
important element of the ways ideas and analyses are negotiated in
academic writing and represent writers decisions of the effects they are
having on their readers. But while they help to ground what might
be abstract, ideational material in the understandings and experiences
of readers, they serve the equally important function of bringing a more
interactive element to the discourse, inserting the writer into the text at
key points.
While the notion of proposition is under-theorised, it is nevertheless
analytically helpful to differentiate propositional material, or the ideational
content of discourse, on one hand and material which organises this content
and conveys the writers attitudes to it on the other. To oversimplify this
distinction slightly, we might suggest that writers have something to say and
the ways they choose to say it are influenced by their expectations of how it
will be received by a particular audience. That is, the main purpose of a text
is to be read, and the writers anticipation of this reading has a backwash
effect on the composition of the text, influencing how it is set out and the
position the writer takes towards it. Statements thus, simultaneously, have
an orientation to the world outside the text and an orientation to the
readers understanding of that world through the text itself.

284

EXEMPLIFYING AND REFORMULATING IN ACADEMIC DISCOURSE

Metadiscourse elements such as code gloss signals are therefore a crucial

element of a texts meaning as they work to relate a text to its context by
taking readers needs, understandings, existing knowledge, intertextual
experiences, and relative status into account. Elaborations illuminate how
writers project themselves into their discourses by signalling their attitude
towards both the content and the audience of the text. In other words,
reformulation and exemplification not only support the writers position and
contribute to communicative effectiveness, but at the same time structure the
means by which he or she is able to relate a text to a given social and
interactive context. By making rhetorical choices of this kind, writers also
signal the judgements they are making about their readers. They convey an
understanding of a community and how they wish to position themselves in
this community by conveying audience-sensitivity and projecting a relationship to the message and to readers themselves (Hyland 2000, 2005).
The corpus analysis also shows disciplinary variations in the ways which
reformulation and exemplification are expressed and the functions they
serve, and I have argued that these rhetorical regularities of elaboration
contribute to different knowledge making practices. Essentially, the sciences
and engineering on one side and social sciences and humanities on the other
draw on different linguistic resources in the creation of specialised
knowledge. While this claim is perhaps unsurprising, it is nevertheless
worth making. This is because corpus findings help to explain rather than to
merely confirm our intuitions about disciplinary practices, underlining that
writers rhetorical decisions are informed by the interactions of members of
communities engaged in a common pursuit. In other words, instead of seeing
these glosses as simply regularities of academic style or the result of some
mental processes of representing meaning, we can understand them as
collective responses to a recurring persuasive problem and perhaps appreciate
a little more the extent to which they carry the sanctioned social behaviours
and epistemic beliefs of individual disciplines.
Science and engineering disciplines reconstrue experience in a technical
way, drawing more heavily on reformulation to make observations and
interpretations more specific. The humanities and social science texts, on the
other hand, are more explicitly interpretative, using reformulation to draw
implications for a potentially more diverse readership and exemplification to
shift between abstraction and concreteness, moving from conceptual to realworld reference. These small acts of elaboration thus convey clear
disciplinary meanings where what counts as convincing argument and
appropriate tone is carefully managed for a particular audience.
Such discoursal conventions as preferences for particular kinds of
elaboration are persuasive precisely because they are significant carriers of
the epistemological understandings of community members. It is clear that
writers in different disciplines represent themselves, their work and their
readers in different ways, and the features discussed in this paper support the
disciplinary variations in argument reported elsewhere. Corpus evidence thus

KEN HYLAND

285

reveals how the resources of language mediate disciplinary contexts

in different ways, working to construe the characteristic structures of
knowledge domains and argument forms particular to broad fields of inquiry.
Final version received December 2006

REFERENCES
Becher, T. (1989). Academic Tribes and Territories:
Intellectual Inquiry and the Cultures of Disciplines.
Milton Keynes: SRHE/OUP.
Biber, D., S. Johansson, G. Leech,
S. Conrad, and E. Finegan. (1999).
Longman Grammar of Spoken and Written
English. London: Longman.
Blakemore, D. (1993). The relevance of
reformulations, Language and Literature 2:
10120.
Burton-Roberts, N. (1993). Apposition in The
Encyclopedia of Language and Linguistics. Oxford:
Pergamon Press.
Cuenca, M.-J. (2003). Two ways to reformulate:
a contrastive analysis of reformulation markers,
Journal of Pragmatics 35: 106993.
Downing, A. and P. Locke. (1992). A University
Course in English Grammar. Hemel Hempstead:
Prentice Hall.
Gonzales, P. (1996). The talk and social organisation of problem-solving activities among physicists. Unpublished Ph.D. dissertation. UCLA.
Greenbaum, S. (1996) The Oxford English
Grammar. Oxford: Oxford University Press.
Halliday, M. A. K. (1994). An Introduction to Functional Grammar, 2nd edn. London: Edward
Arnold.
Hyland, K. (2000/2004). Disciplinary Discourses:
Social Interactions in Academic Writing. London:
Longman/Ann Arbor, MI: University of
Michigan Press.
Hyland, K. (2004) Disciplinary interactions: Metadiscourse in L2 postgraduate

writing, Journal of Second Language Writing 13:

13351.
Hyland, K. (2005). Metadiscourse. London:
Continuum.
Hyland. K. and P. Tse. (2004). Metadiscourse
in academic writing: a reappraisal, Applied
Linguistics 25/2: 15677.
Killingsworth, M. and M. Gilbertson. (1992).
Signs, Genres and Communities in Technical
Communication. Amityville, NY: Baywood.
Kuhn, T. (1970). The Structure of Scientific
Revolutions, 2nd edn. Chicago: University of
Chicago Press.
Lee, Y.-A. (2004). The work of examples in
classroom instruction, Linguistics and Education
15: 99120.
Martin, J., and D. Rose. (2003). Working with
Discourse: Meaning beyond the Clause. London:
Continuum.
Quirk, R., S. Greenbaum, G. Leech, and
J. Svartvik. (1972). A Grammar of Contemporary
English. Harlow, Essex: Longman.
Rouchota, V. (1998). Procedural meaning
and parenthetical discourse markers in
A. Jucker and Y. Ziv (eds.): Discourse Markers:
Description and Theory. Amsterdam: John
Benjamins, pp. 97126.
Schegloff, E., G. Jefferson, and H. Sacks.
(1977). The preference for self-correction in
the organization of repair in conversation,
Language 53: 36182.
Swales, J. M. (2004). Research Genres. Cambridge:
Cambridge University Press.