Activity And Evaluation Reporting Practices In Academic Writing

HYLAND, K. (2002). ACTIVITY AND EVALUATION: REPORTING PRACTICES IN ACADEMIC WRITING.
IN J. FLOWERDEW (ED) ACADEMIC DISCOURSE.LONDON, LONGMAN. PP 115-30.
ACTIVITYAND EVALUATION: REPORTING PRACTICES IN ACADEMIC WRITING
Ken Hyland

2
Reference to prior research is almost a defining feature of the academic research article. Even the most
original paper integrates and represents ideas, concepts, findings and theories from other sources and,
indeed, would be unlikely to reach publication if it did not. Specifically, such reporting represents in a
new situation the way language was used in a previous context, and is defined for my purposes here as
the attribution of propositional content to another source. Its importance in academic discourse lies in
providing an appropriate context of persuasion, demonstrating how the current work builds on and
reworks past utterances to establish intertextual links to the wider discipline. Put simply, academic
writing depends for its success on situating current work in a larger disciplinary narrative
(Berkenkotter & Huckin, 1995; Hyland, 2000; Myers, 1990). Without such links academics could
neither justify their arguments by connecting their research activities to significant work in the field,
nor use this disciplinary knowledge to establish the novelty of their position (Gilbert, 1977;
Berkenkotter & Huckin, 1995).
By acknowledging a debt of precedent, a writer simultaneously accomplishes a number of rhetorical
objectives. Interpersonally, judicious citation enables writers to display an allegiance to a particular
community or theoretical orientation (Latour, 1987). It also helps them to build a credible writer ethos
using citation to construct factual reliability and show they are prepared to stand behind their words.
Scollon (1994), for example, has argued that attribution “constitutes a system for the creation of the
academic self” as it presents to the reader a stance or position of responsibility. Rhetorically, citation
helps to define a specific context of knowledge or problem to which the current work is a contribution.
But while new work reveals its disciplinary credentials by being embedded in a community generated
literature, it must also transcend this literature to create a niche for itself. Writers have to carve out a
rhetorical gap from the weight of consensual knowledge (Swales, 1990), staking a claim to innovation
while taking care to avoid a hostile depiction of previous work. The ability to handle citations to
rhetorically construct a community consensus, and at the same time ensure that criticism stays within
accepted bounds, is a central means of projecting one’s insider status.

3
Despite this importance however, citation represents a feature which students, and particularly non-
native English speakers, find difficult to either use effectively in their writing or understand correctly
in their reading (Campbell, 1990; Matalene, 1985; Pennycook, 1996). NNS writers of academic papers
frequently overuse naïve quotation with no evaluation (Bruce, 1989), and rely on a restricted range of
verbs, such as ‘say’ to introduce these quotes (Pickard, 1993a). This problem is often more than a
deficit of vocabulary, but symptomatic of a larger issue of how to appropriately acknowledge sources
in academic writing. There are considerable cultural differences in how writers from different
backgrounds use prior texts in argument (eg. Bloch & Chi, 1995), and such differences also extend to
notions of authorship and text ownership (Pennycook, 1996; Scollon, 1994). However, because
reporting is such an important convention of academic writing in English, it is clearly worthwhile
equipping students with the means of using it successfully. In this paper I focus on the most salient
form of attribution in academic writing, the use of reporting verbs.
Reporting signals in academic discourse
The significance of citation in academic writing and the obvious difficulties it poses for learners has
led to considerable interest in the ways it is signaled, often as a basis for producing teaching materials.
The literature testifies to the availability of a wide range of reporting structures in English (eg.
Thompson, 1996), although academic writing appears to rely on a sub-set of specialised conventions
from within these, principally direct quote, paraphrase, summary and generalisation (eg. Dubois,
1988). One attribution feature of interest to researchers has been the distinction between integral and
non-integral structures (Swales, 1990: 148), the former referring to cases where the name of the cited
author occurs in the citing sentence, and the latter to where the author appears in parenthesis or is
referred to by superscript numbers. The use of one form rather than the other appears to reflect a
decision to give greater emphasis to either the reported author or the reported message. Research has
also identified the rhetorical effects of various syntactic features, such as thematic position, tense, and
voice, on the reported information (eg. Malcolm, 1987; Oster, 1981; Shaw, 1992) and to the role of
reporting verbs.

4
The use of a reporting verb is one of the most explicit ways of attributing content to another source,
and represents a significant rhetorical choice. The wide range of verbs that can be used to introduce
reports allow writers to convey both the kind of activity reported and whether the claims are to be
taken as accepted or not (Hunston, 1993; Tadros, 1993; Thomas & Hawes, 1994). Most obviously,
verbs such as demonstrate, prove and show reveal the writer’s agreement with a prior statement, while
hedges like suggest, indicate and imply open an “evaluative space” (Thompson & Ye, 1991: 369) in
which the writer can withhold full commitment to present a contrast with a new view.
Thompson and Ye (1991) provide an interesting analysis of the relationships between reporting verbs
and evaluation, and their categorization emphasizes the important distinction between the position of
the reporting writer and the source author. However, their study was confined to the introduction
sections of articles, which ignores a great deal of the reporting which occurs in humanities and social
science papers, and involves a rather complex categorization system which separates evaluation from
reporting and allows considerable overlap between categories. Thomas and Hawes (1994) alternative
taxonomy offers a much clearer description of the network of options available to writers in medical
articles. But while this scheme is easier to apply in identifying the functions of different verbs, it does
not clearly reveal their evaluative potential nor always maintain the distinction between reporting and
reported writer in identifying the source of this evaluation.
So while my study draws on these systems, it also seeks to offer a clearer description of reporting
conventions in academic writing, to identify the preferences for different verbs among various
disciplinary communities and, hopefully, to reveal some of the rhetorical purposes which motivate
those preferences.
Corpus and methodology
The study employs both qualitative and quantitative approaches, comprising frequency counts and text
analyses of a corpus of published articles. The corpus of 80 research articles consists of one paper from
each of ten leading journals in eight disciplines. The disciplines were selected to represent a broad cross-
section of academic practice and to allow access to expert informants. Molecular biology and magnetic

5
physics represent the pure sciences, and mechanical engineering and electronic engineering the applied
sciences. Philosophy and sociology are often categorised as either humanities or social sciences, while
marketing and applied linguistics might be regarded as more applied social sciences. The journals were
nominated by specialist informants as among the most important in their fields, and the articles chosen
at random from current issues, selecting only those based on original data to compare papers with a
similar rhetorical purpose. The scanned articles produced an electronic corpus of just over 500, 000
words after excluding abstracts and text associated with tables and graphics.
To locate reporting verbs, the corpus was computer searched for canonical citational forms such as a
date in brackets, a number in squared brackets, and Latin references to other citations. This search left
many citations unaccounted for however, particularly where writers’ renewed or extended their
discussions of a previously mentioned author without the repetition of a reference. I therefore
concordanced all the names in the bibliographies of these articles and of third person pronouns. A
citation was recorded only where the text attributed words to a specific author, thereby eliminating
general references to schools or beliefs such as “Marxists’ claim...” or “a constructionist model
suggests…”. Examples of self citation were also ignored (see Hyland, 2001). This yielded 2, 287
examples of reporting verbs, about one every 220 words of text.
I then developed a way of categorizing these verbs, first examining samples from the corpus using the
two models discussed above, and then refining the model through a constant comparison of devices in
context. This inductive approach allows a theory to emerge from data and produced the model
described below. Note that in the following discussion I adopt Thompson and Ye’s useful convention
of referring to the person citing as the “writer” and the cited person as the “author”.
Functions of reporting verbs
i. Process functions
Like Thompson and Ye (1991) and Thomas and Hawes (1994), I classified the reporting verbs
according to the type of activity they referred to. This gives three distinguishable processes:

6
1. Research (real-world) Acts. Verbs in this category represent experimental activities or actions
carried out in the real world. They generally occur either in statements of findings (eg. observe,
discover, notice, show) or procedures (eg. analyse, calculate, assay, explore, plot, recover).
2. Cognition Acts. These verbs are concerned with the researcher’s mental processes (eg. believe,
conceptualize, suspect, assume, view).
3. Discourse Acts. These involve linguistic activities and focus on the verbal expression of
cognitive or research activities (eg. ascribe, discuss, hypothesize, report, state).
These categories are not watertight. Agree, for example, is a cognition act which carries a strong
implication of verbal expression, while conclude is ambiguous between a mental and a discourse
process, analyze may refer to physical or intellectual activities, and solve perhaps involves all three
processes. However, it is possible to consistently attribute a particular meaning to all the verbs using
this system and the categories do provide a comprehensive taxonomy for distinguishing the use of
reporting verbs in terms of the primary aspects of the research process.
Analysis of all reporting verbs in the corpus revealed a clear preference for reporting information as
Discourse Acts, with 57% of cases in this category. Most of the remaining instances were reported as
Real-World Acts, and only 8% as Cognition Acts. Where writers chose to represent work as Research
Acts, frequencies divided almost equally between verbs describing the author’s overall findings and
those commenting on research procedures. Table 1 summarises these findings.
Table 1: Distribution of verbs in process categories
Category Totals Average per paper %
Real World Acts 786 9.8 35
Findings 437 5.4 19
Procedures 349 4.4 15
Cognition Acts 193 2.4 8
Discourse Acts 1308 16.4 57
Total 2287 28.6 100

7
ii. Evaluative functions
Within the process categories writers made more delicate decisions, exploiting the evaluative
possibilities of reporting verbs to take either a supportive, tentative, critical, or neutral stance towards
the reported claims. Here I depart from earlier work by representing writers’ choices as a single
overarching scheme of options which includes both the original author’s academic activity and the
reporting writer’s evaluative judgements. The scheme retains Thompson and Ye’s important insight
however that while all reporting is mediated by the reporter, writers can vary their commitment by
employing verbs which imply a personal stance, such as show, demonstrate, fail and ignore, or which
attribute a position to the original author (accuse, believe, dispute, urge).
Each of the process categories therefore has a sub-set of evaluative options as set out in Figure 1 and
elaborated below.
Figure 1: Categories of reporting verbs
Reporting
Research Acts Cognition Acts Discourse Acts
Findings Procedures Positive Critical Tentative Neutral Doubt Assurance Counters
Factive Counter-factive Non-factive Tentative Critical Factive Non-factive
Within the Findings category of Research Acts, writers can acknowledge their acceptance of the
authors’ results or conclusions with factive verbs such as demonstrate, establish, show, solve or
confirm. Alternatively they can adopt a counter-factive stance, portraying the author’s judgements as
false or incorrect (fail, misunderstand, ignore, overlook). While these verbs might seem an effective
way for writers to refute prior research and establish a niche for their own alternative position, such a
direct challenge to the work of a single author was a choice rarely exercised in this corpus.
Academics, it seems, appear to prefer more subtle ways of building on earlier literature. The final
option is to comment on research findings non-factively, with no clear attitudinal signal as to their
reliability (eg. find, identify, observe, obtain). These verbs accounted for the majority (55%) of
references in the Research Findings category. Verbs which referred to procedural aspects of the

8
author’s investigation contained the greatest variety of expression in my corpus with 153 different
verb forms, 36% of the total. As the following examples show, these verbs carry no evaluation in
themselves but simply report research tasks neutrally:
(1) Finally, Eto et al. (1994) reviewed and analyzed the contents of several indicators,.. (Mech)
Green and Wakefield compared the 20 cases they found in the newspaper with… (Soc)
a "layer" coupled-slot finline structure was studied by Mazur[7] and Tech et al [8]… (Phy)
Leow (1995) replicated an earlier study, Leow (1993), which investigated .… (Ling)
Cognition Act verbs, which portray the cited work in terms of a mental process, handle evaluation
rather differently. Instead of explicitly taking a personal stance on the reported information or opinion,
writers here attribute a particular attitude to the cited author. There appear to be four clear options.
Writers can represent the author as having a positive attitude to the material, accepting it as true or
correct with verbs such as agree, concur, hold, know, think, or understand. Alternatively authors’ may
be characterised as having a tentative view towards the reported matter (believe, doubt, speculate,
suppose, suspect), or, more rarely, as taking a critical stance (eg. disagree, dispute, not think). Finally,
the writer can portray the author as holding a neutral attitude towards the proposition (picture,
conceive, anticipate, reflect).
Writers can also employ Discourse Act verbs to convey an evaluation of the cited material, and here
they have the option of either taking responsibility for their interpretation, conveying their uncertainty
or assurance of the correctness of the claims reported, or attributing a qualification to the author. The
examples below show this distinction, indicating how writers can use discourse verbs to express their
own viewpoint (2) or to ascribe it to the author (3):
(2) As demonstrated by Baker et al. [2], the increased rate of flagellar RNA ….. (Bio)
Baddeley proposes a tripartite system of working memory,… (Ling)
Although derived in 1965, recent work [59], verified that (2) still holds true. (Elec)
As Hinde (1979) points out, many unhappy marriages remain intact because of… (Mkt)
Villaruel-Ordaz, et al. (1993) report C. subnuda from the states of … (Bio)
(3) Churchill abandoned speculative naturalism. (Phil)
Eckstein criticizes psychological reductionism, rational choice, and …. (Soc)
Jacoby accuses American intellectuals of a turn to conservatism,... (Soc)

9
…. dismissed by some theorists as being depthless and superficial chatter (eg.
Baudrllard, 1988; Jameson, 1991). (Mkt)
Frenchio urged against changing…. (Phy)
Discourse verbs which express the writer’s view directly can be separated into doubt and assurance
categories. Those which express doubt about reported claims can be further divided into verbs which
are tentative (eg. postulate, hypothesize, indicate, intimate, suggest) and, more rarely, verbs which are
directly critical (evade, exaggerate, not account, not make point). More frequent in the corpus than
expressions of doubt are assurance verbs which introduce cited material in more positive and
conclusive terms, either to neutrally inform readers of the author’s position (non-factive)or to use that
position to support the writer’s own (factive). There is an enormous variety of verbs available for
objectively passing information without interpretation, including high frequency items such as state,
describe, discuss, report, answer, define, and summarize. Alternatively, writers use assurance
discourse verbs to directly bolster their own views. These signal a supportive role for the reported
information in the writer’s argument, often by attributing a high degree of confidence to the
proposition by the original author. Examples include argue, affirm, explain, note, point out and claim.
The final sub-category of Discourse Act verbs, Counters, refer to what are taken to be the cited
author’s own reservations or objections to the correctness of the reported message. That is, instead of
taking responsibility for the evaluation as in Doubt verbs, the writer attributes these reservations to the
original author. Among the examples in the corpus are deny, critique, challenge, attack, question,
warn, refute and rule out. Thompson and Ye call such attributions Author Acts (1991: 371) and
writers generally drew on such author refutations to either support their own opposition to a position,
or to demolish an opposing argument. The overall results are presented in Figure 2 below.

10
Fig 2: Overall distribution of reporting verbs
Factive (41%)
Research Acts Findings (56%) Counter-factive (4%)
786 (35%) Non-factive (55%)
Procedures (44%)
Positive (50%)
Cognition Acts Critical (5%)
193 (8%) Tentative (20%)
Neutral (25%)
Doubt (24%) Tentative (95%)
Critical (5%)
Discourse Acts
1308 (57%) Factive (42%)
Assurance (71%) Non-factive (58%)
Counters (5%)
Disciplinary variations in reporting practices
One issue of central concern to EAP teachers is the extent to which discourse features can be regarded
as universal or as discipline-specific. Increasingly we are becoming aware of major distinctions in the
ways different fields conduct their inquiries, justify their beliefs and structure their arguments (eg
Bartholomae, 1986; Becher, 1989; Bruffee, 1986). We are also more mindful of the ways these
differences are expressed through the interactive and rhetorical features of academic writing (eg.
Bazerman, 1988; Hyland, 1998, 1999, & 2000; MacDonald, 1993). From a pedagogical point of view
then, both postgraduate students and practicing academics need to be made aware of such disciplinary
preferences in order to communicate with their peers using the most conventionally effective
strategies. In terms of reporting structures, my analysis of this corpus suggests that while writers
chose to present cited work in broadly similar ways, there were considerable disciplinary differences
in the overall use of reporting verbs, in the frequencies of individual items, and in preferences for
particular rhetorical categories.
Citing verbs do not occur in a textual vacuum. Language reports are presented in a carefully
constructed reporting context which involves academic writers in conscious decisions about framing
the imported material. Two central options here are how best to represent the original text and what
emphasis to give to the original author. Overall, there were both broad similarities and differences in

11
how writers in different disciplines exercised these options. Considerations of how to present a cited
message mainly involve the extent to which a writer wishes to duplicate the original source material in
his or her argument (Dubois, 1988; Thompson, 1996). The fundamental options are to either employ
short direct quotes (of five or more words), use the original wording extensively in indented blocks,
summarise from a single source, or present a generalised statement from several sources.
An analysis of half the articles in each discipline confirmed the impression that the overwhelming
preference in all eight disciplines was for summary, an abbreviated statement of one work, with
generalisation accounting for most of the remainder. It can be seen in Table 2 that direct quotation,
which still seems to occupy a considerable amount of textbook space, and teacher time, was rarely
employed and did not occur in the science and engineering papers at all. In the hard sciences this may
be partly related to the suppression of explicit researcher intervention, but more generally the decision
of writers to express cited material in their own terms represents the most effective way of supporting
their own positions. Thus while students often overuse direct quotation (Campbell, 1990: 222), expert
writers tend to be sparing with it, and are only likely to employ it when they consider the author’s
original words to be the most persuasive means of presenting their case.
Table 2: Presentation of cited work (%)
Discipline Quote Block Quote Summary Generalization
Biology 0 0 72 38
Physics 0 0 68 32
Electronic Engineering 0 0 66 34
Mechanical Engineering 0 0 67 33
Marketing 3 2 68 27
Applied Linguistics 8 2 67 23
Sociology 8 5 69 18
Philosophy 2 1 89 8
A second decision facing writers in citing material is whether to employ the reporting verb in an
integral or non-integral environment (Swales, 1990). As noted earlier, integral structures (example 4
below) tend to give greater prominence to the cited author than non-integral contexts (as in 5):
(4) …some can be analysed in the manner suggested by Lewiss,… (Phil)

12
Patrick O’Niell (1990) suggests that the growth of black humour is a peculiarly … (Soc)
The study of Tadmor (1986) shows that fly ash and ESP ash are enriched with… (ME)
Barry and Diamond (1984) and Dainty (1985) have argued that water relation… (Bio)
(5) Several experiences have been reported in the literature on the millimeter wave
interaction with snowcover [5-7] … (Phy)
At least one writer (Stark, 1987) suggests that police leniency in ghettos is largely… (Soc)
Reference [6] shows a spatial Euler-Savary analogue based on velocity and ... (ME)
… recently examined modality differences in SLA (eg. Kirster, 1984; Leow, 1985),… (Soc)
In this corpus, at least, there was again little disciplinary variation with an overwhelming preference
for non-integral forms. As Table 3 shows, only the articles in philosophy, which typically consist of
long narratives engaging the arguments of other writers, consistently included the cited author in the
reporting sentence.
Table 3: Representation of author by discipline (%)
Discipline Non-Integral Integral Discipline Non-Integral Integral
Biology 90.2 9.8 Marketing 70.3 29.7
Electronic Eng 84.3 15.7 Applied Ling 65.6 34.4
Physics 83.1 16.9 Sociology 64.6 35.4
Mechanical Eng 71.3 28.7 Philosophy 35.4 64.6
Disciplinary preferences for verb forms
Despite these similarities in message form and author visibility, the distribution of reporting verbs in
the corpus reveals broad disciplinary differences. There appear to be community-based preferences,
both for specific items and the implications carried by particular semantic categories.
While all the papers in my corpus included reporting verbs, they were far more prevalent in the
humanities and social science papers than in the hard sciences. This in fact mirrored the more explicit
role of reference to prior research in these disciplines which occurred overall in the corpus. I have
sought to account for these broad disciplinary differences elsewhere in terms of the distinctions
between hard and soft knowledge domains (Hyland, 2000). Briefly, I believe the higher use of
citations in the soft disciplines can be accounted for by the fact that (i) they typically have less
cohesive and established frameworks of knowledge; (ii) they possess less highly formalized and

13
standardized codes for reporting research; (iii) they are more inclined to explicitly recognize the role
of human agency in constructing knowledge; and (iv) they engage in more recursive patterns of
investigation which involve more diverse and less predictable and abstract subjects than those
typically found in the sciences.
We can see from Table 4 that the density of reporting structures differed substantially, with the
average number of reporting verbs in the engineering and science papers all below the average for the
corpus as a whole. Together there were just 619 reporting verbs in the 40 hard knowledge papers,
amounting to only 27% of all examples, with physics accounting for less than 3% of the overall total.
In terms of density of verbs per 1,000 words, the distribution looks slightly more even, with both
engineering disciplines similar to the marketing figures and biology exceeding all but sociology and
philosophy, which together accounted for 44% of all the reporting verbs in the corpus.
Table 4: Reporting verbs by discipline
Discipline Reporting verbs Most Frequent forms (ranked from left)
per paper per 1,000 words %
Philosophy 57.1 7.3 say, suggest, argue, claim, point out, hold, think 25.4
Sociology 43.6 5.3 argue, suggest, describe, note, analyze, discuss 26.8
Applied Ling. 33.4 4.8 suggest, argue, show, explain, find, point out 28.5
Marketing 32.7 3.5 suggest, argue, find, demonstrate, propose, show 40.4
Biology 26.2 4.9 describe, find, report, show, suggest, observe 55.7
Electronic Eng. 17.4 3.4 propose, use, describe, show, publish, develop 38.0
Mechanical Eng. 11.7 3.1 describe, show, report, discuss, give, develop 30.7
Physics 6.6 2.0 develop, report, study, find, expand 42.4
Totals 28.6 4.6 suggest, argue, find, show, describe, propose, report 27.3
Table 4 also shows the most frequent forms found in each discipline and the proportion of total
reporting verbs which these comprised. Over 400 different verbs were used to introduce citations,
although the seven forms in the totals column constituted over a quarter of all cases and nearly half the
forms in the corpus occurred only once. It appears from the figures that writers in biology, marketing
and physics relied on a more restricted range of items, with the most preferred five or six forms
accounting for over 40% of all cases. Clearly writers have the same potential range of items to draw on

14
and so this preference for a limited number suggests a higher degree of conventionalisation in the
ways that material is reported in these disciplines.
The most common verbs in the corpus were the discourse acts suggest (148 cases) and argue (103),
followed by the research verbs find (92) and show (86), and the discourse verb describe (86).
However, there were clear differences in the forms preferred by broad disciplinary areas. Writers in
the social sciences and humanities, for example, greatly favoured argue (100% of cases), suggest
(82%), and study (70%), while report (82%), describe (70%), and show (55%) occurred mainly in the
science / engineering articles. Of the other forms which occurred more than 25 times in the corpus,
the most marked distributional imbalances were say (95% in philosophy), use (70% electrical),
examine (50% physics), report (53% biology), analyze (40% sociology) and observe (40% biology).
Process categories of reporting verbs
The analysis shows a fairly clear division in the process categories which again corresponds to the
traditional division between hard and soft disciplines (Table 5). Philosophy, sociology, marketing and
applied linguistics largely employed Discourse Act reporting verbs (averaging 61% of the total for the
four disciplines), and the engineering and science papers displayed a preference for Research-type
verbs (50%). Cognition verbs were infrequent in almost all disciplines and were virtually absent in the
engineering and physics papers, which together contained only 11 cases, or 6% of the total.
Table 5: Distribution of verbs in process categories by discipline (%)
Process Biology Physics Elec Eng Mech Eng Mkting App Ling Sociology Phil Totals
Total n 262 66 174 117 327 334 436 571 2287
Research 43.1 56.0 55.2 47.0 31.2 30.5 29.1 23.5 33.5
Cognition 7.2 6.1 2.9 1.7 7.3 10.5 6.9 14.7 8.9
Discourse 49.7 37.9 41.9 51.3 61.5 59.0 64.0 61.8 57.6
These writers therefore clearly drew intertextual links to their disciplines in different ways, and these
broadly reflected the preferred citation and argument forms of their discourse communities. The
predilection for Research verbs and corresponding deficiency of Cognition forms in the science and

15
engineering papers, for example, together help to convey an experimental explanatory schema. This
distribution contributes to an overall impression in the hard discipline papers of research activity as an
inductive, impersonal and empirically-based endeavor. Here the reported facts are shown to emerge
from laboratory activities rather than discursive practices, and the legitimacy of the information rests
securely on the non-contingent, socially invariant standards of laboratory procedures. The following
examples offer some flavour of this:
(6) Grafstrom et al. [4] have studied photothermal currents…. (Phy)
... using special process and design [42], or by adding [101], or removing [83] a mask. (Elec)
The relevant properties from each individual tensile test have been reported in reference (11)
… (Mech)
Marschner & Dell (1994) estimated that external hyphae of AM fungi can deliver up to.. (Bio)
The reasons for this are examined in detail by Yeo et al (1990), Yeo (1994), and … (Bio)
The greater use of Discourse Act forms in the humanities and social sciences, on the other hand, is
more appropriate in an argument schema which more readily regards explicit interpretation,
speculation and complexity as accepted aspects of knowledge. These disciplines are typically more
discursive and examine relationships and features that are more subject to contextual and human
irregularities than those studied in the hard sciences. Writers in the soft disciplines therefore employed
arguments that made greater use of Discourse act forms which expedited the verbal exploration of
such issues, facilitating qualitative arguments which rest on finely delineated interpretations and
conceptualisations, rather than systematic scrutiny and precise measurement:
(7) Lindesmith's (1965) classic work indicated the exclusion of blacks from managerial… (Soc)
Davidson defends this claim on the grounds that without creatures using sentences… (Phil)
Cass Sunstein makes the case that the best thing about republican thought is… (Phil)
Ballard and Clanchy (1991) argue that cultural attitudes toward knowledge range from…(Ling)
Over 25 years ago Hardy (1975) lamented that few managers really knew whether … (Mkt)
Similarly, the fact that 85% of Cognitive verbs occurred in the humanities and social science papers
admits to a greater role for personal interpretation in knowledge negotiation. In these disciplines the
literature is frequently depicted in terms of the cited author’s theorizing and mental activities, thereby
giving prominence to the role, and often the fallibility, of human agency in constructing claims. Thus
through the use of Cognition verbs, the soft domain writers tend to represent research as proceeding

16
from the interpretive operations or verbal accounts of researchers, emphasising the part that reasoning
and argument play in the construction of knowledge:
(8) Chomsky regards scientific work as, by definition, characterized by a high level of … (Ling)
Donnelan believes that for most purposes we should take the demonstatum to be .. (Phil)
Austin seems to assume that a classification of different verbs is eo ipso a … (Ling)
Jerry Fodor thinks that it is irredeemably disjunctive. (Phil)
Julian Hochfeld perceives PS as the only base for a unified ‘science of politics’… (Soc)
Evaluative categories
The evaluative categories show less obvious differences, with writers in all disciplines displaying a
clear preference for neutral stances. While the selection of an appropriate reporting verb allows
writers to signal their assessment of the evidential status of information, and thus demonstrate their
degree of commitment to it, actual refutation is relatively rare. In fact, direct criticism does not occur
in my science and engineering papers at all and is mainly confined to philosophy and sociology, where
it constitutes 70% of all cases in the corpus (46 instances):
(9) Donnelan overlooks intuitions of falsity through keeping… (Phil)
…and the theory of Rawls in particular, unjustifiably ignore an unjust division of labor,...(Phil)
Lillian Faderman has also probably exaggerated the pervasiveness of …. (Phil)
What Travers fails to explain is why, given these differences in local context, .. (Soc)
But they do not explain why elites no longer exhibit ….. (Soc)
It is also worth noting that negation reporting verbs also tend to be largely drawn from the categories
of Discourse and Cognition Acts. Because these focus on authors’ interpretations or texts, rather than
their research, they are potentially a less challenging form of criticism as they avoid any direct attack
on the author’s competence or reputation (cf. Thompson & Ye, 1991:377).
Overall however it appears that if the writer wishes to offer a negative assessment of another’s
viewpoint, this is not typically signaled by the reporting verb. Explicit rebuttal of other researchers is a
serious face-threatening act in academic writing (Myers, 1989), and such violation of interpersonal
conventions is likely to expose the writer to retaliation or the disapproval of publishing gatekeepers
(Hyland, 1998). Refutation is therefore likely to be heavily mitigated and involve considerable
rhetorical work to accomplish (eg. Pickard, 1993b). Indeed, the fact that most rebuttals in the corpus

17
were expressed in terms of omission, an assessment that a particular act was not performed, might
itself be seen as a mitigating strategy.
Also infrequent are Counters, or cases where the writer represents the cited author as holding a
tentative or negative stance to the reported information. Generally, where writers imputed positions to
authors they generally expressed them positively or neutrally as cognition verbs. Again, the vast
majority of counters occurred in the sociology and philosophy papers with almost no examples in the
hard knowledge corpora:
(10) Vernon and Sewell seriously question the materialist premises of Thompson's work,… (Soc)
Tilly has, in fact, complained that RMT ‘identifies the amassing or spending of … (Soc)
…deregulation has been blamed or small grocery failures in Australia (Harty, 1994), … (Mkt)
… which are dismissed by some theorists as superficial chatter (eg Baudrillard, 1988;
Jameson, 1991).. (Mkt)
…the idea of a characterless substructure is rejected by Aristotle in Metaphysics. (Phil)
…cited and criticized by Coates, (1984, p. 111). (Phil)
…..Donald Downs decries restrictions on pornography. (Phil)
However, both Davidson and Wittgenstein explicitly disown the view .. (Phil)
Scientists and engineers were far more likely to report information non-factively, with 66% of all
reporting verbs in these disciplines in this category. These verbs help to rhetorically represent the
norms of impersonality and impartiality which are professed to characterise inquiry the hard
knowledge fields, providing an acknowledgement of prior research without appearing to corrupt it
with personal judgement. This is particularly apparent where writers also chose to remove the cited
author, and thus the implication of human intervention, from the reporting sentence:
(11) Ref. [9] developed finite formulations and corresponding code. (Phy)
Fig. 1 shows the CEMS spectra recorded at room temperature using a He/CH4
gas-flow detector described elsewhere [6], … (Phy)
…properties of a line trajectory in spatial motion are researched by Refs [21-23], … (Mech)
… other solutions have also been published [5-7] (Elec)
Less often, these writers employed reporting verbs to signal their acceptance of prior research (22% of
verbs in the hard knowledge fields). Virtually all the remaining cases were tentative Discourse act
verbs, conveying a implication of uncertainty or less than full commitment:

18
(12) Sacc. is considered a nomen confusum (Hughes, 1958)….. (Bio)
These studies imply that microtubule penetration into… (Bio)
Preliminary data (Gentzis, 1993) indicate U to have predominantly … (Mech)
Burbidge (1992) suggests that a manned cell should not contain more than… (Elec)
A theory based on the demagnetization field has been suggested by Warren et al [13-15]..(Phy)
The fact that reporting verbs were more likely to carry the writer’s evaluative stance in the humanities
and social science papers reflects the more disputational and discursive rhetorical style of these
disciplines. Evaluation allows writers to both engage with the disciplinary literature in a continuing
debate and to negotiate the value of their own contribution by opening a discursive space in that
literature, building on what has gone before. This is particularly evident in the fact that writers in the
soft disciplines were also more likely to evaluate cited information by adding adverbial comment to
the reporting verb, often to strongly align themselves with that position:
(13) He argues there, correctly to my mind, that …. (Phil)
Churchland correctly rejected this move… (Phil)
Benveniste (1969: 116 and 166-168) rightly points out how the act of swearing is … (Ling)
As Dipankar Gupta correctly asserts … (Soc)
Wilson rightly says that many important social scientific theories have…. (Soc)
In sum, the reporting options found in this corpus appear to be more than simply the stylistic
proclivities of individual writers. Instead, they can be seen as reflecting disciplinary preferences for
particular forms of rhetorical expression which, in turn, instantiate the different procedures and
epistemological understandings of particular fields of inquiry.
Some conclusions on reporting and EAP
Although reference to previous work is virtually obligatory in academic writing, I have tried to show
here that the ways writers choose to present information varies according to the discourse communities
they inhabit. The options offered by the network of reporting verbs are an important means by which
writers are able to appropriately link their local contributions into a wider disciplinary framework and
so persuade readers of their ideas. These findings, I believe, have important consequences for novice

19
research writers, and for those who teach them., as once again we have support for the view that
students entering academic communities need to acquire and employ a discipline-specific literacy.
A growing body of research into academic discourse has demonstrated that students have specialized
communicative needs which are defined by the social purposes and rhetorical practices of their
specific target communities. Although initially focused on undergraduates, this research has
progressively responded to the needs of more advanced groups, and the particular demands of research
writing are now recognised as representing a serious challenge to both post-grad students (Paltridge,
1997) and L2 academics seeking to publish in English (Flowerdew, 1999). Although reporting is
often treated in EAP materials as the application of standardised conventions and advice about
avoiding plagiarism (eg. Lester, 1993), students need to know more than the necessity and mechanics
of referring to an existing literature. They also have to present that literature in ways that their readers
are likely to find convincing.
The selection of a particular reporting verb is a delicate choice as it is a crucial means of both situating
one’s work appropriately and communicating with ones peers effectively, a way of engaging with
colleagues and of appealing to the epistemological and interactive understandings of one’s
community. Selecting a particular verb thus not only signals a reported voice, but invokes a precise
context of meaning and judgment which locates the writer in a certain relationship to the reader and
the reported text. This can critically influence a reader’s willingness to go along with a writer and
accede to her claims.
Clearly more work needs to be done. In particular, there has been no space to discuss the typical
rhetorical functions of the reports in which particular categories of verb occur, and I have said nothing
about the effect of negation on the use of specific verbs. Moreover, the study is restricted to research
papers and does not address other academic genres that students are commonly required to read or
produce, such as theses, reports and survey papers. However, the taxonomy suggested here offers a
fairly straightforward and useful way of identifying the meanings of different choices and of mapping
the dimensions of difference between disciplines and genres. The need to distinguish the particular

20
rhetorical features common to specific disciplines is a growing pedagogical imperative in university
writing courses, and reporting is clearly one area where students urgently require more assistance. So,
while various applications and investigations remain, this study may contribute to a fuller
understanding of research reporting and to the development of materials for novice academic writers.
References
Bartholomae, D. 1986. Inventing the university. Journal of Basic Writing, 5 , 4-23.
Bazerman, C. 1988. Shaping written knowledge. Madison: University of Wisconsin Press.
Becher, T. 1989. Academic tribes and territories: intellectual inquiry and the cultures of disciplines.
Milton Keynes, SRHE/Open University Press.
Berkenkotter, C. & Huckin, T. 1995 Genre Knowledge in Disciplinary Communication. Hillsdale,
New Jersey: Lawrence Erlbaum.
Bloch, J. & Chi L. 1995. A comparison of the use of citations in Chinese and English academic
discourse. In D. Belcher & G. Braine (eds.) Academic writing in a second language: essays on
research and pedagogy. Ablex: New Jersey.
Bruce, N. 1989. The roles of analysis and the conceptual matrix in a process approach to teaching
academic study and communication skills. In V. Bickley (ed.) Language teaching and learning
styles within and across cultures. Hong Kong: Institute of Language in Education.
Bruffee, K. 1986. Social construction: Language and the authority of knowledge. A bibliographical
essay. College English, 48: 773-779.
Campbel, C. 1990. Writing with others' words: using background reading text in academic
compositions. In B. Kroll (Ed.), Second language writing. Cambridge: CUP.
Dubois, B. 1988 Citation in biomedical journal articles. English for Specific Purposes, 7: 181-94.
Flowerdew, J. 1999. Problems in writing for scholarly publication in English: the case of Hong Kong.
Journal of Second Language Writing. 8: 243-64.
Gilbert, G. 1976 The transformation of research findings into scientific knowledge. Social Studies of
Science, 6: 281-306.
Hunston, S. 1993 Evaluation and ideology in scientific writing. In M. Ghadessy (ed.), Register
analysis: theory and practice., (pp. 57-73) London: Pinter.
Hyland, K. 1998 Hedging in scientific research articles. Amsterdam: Benjamins.

21
Hyland, K. 1999. Disciplinary discourses: writer stance in research articles. In Candlin, C. & Hyland,
K. (eds) Writing: texts, processes and practices. pp. 99-121. London: Longman.
Hyland, K. 2000. Disciplinary discourses: social interactions in academic writing. London: Longman.
Hyland, K. 2001. Humble servants of the discipline? Self mention in research articles. English for
Specific Purposes.
Latour, B. 1987. Science in action: How to follow scientists and engineers through society. Milton
Keynes: Open University Press.
Lester, J. D. 1993. Writing research papers. 7th
Edition. New York: Harper Collins.
Macdonald, S. P. 1993. Professional academic writing in the humanities and social sciences.
Carbondale: Southern Illinois University Press.
Malcolm, L. 1987 What rules govern tense usage in scientific articles ? English for Specific Purposes,
6: 31-44.
Matalene, C. 1985. Contrastive rhetoric: an American writing teacher in China. College English, 47:
789-808.
Myers, G. 1989. The pragmatics of politeness in scientific articles. Applied Linguistics, 10: 1-35.
Myers, G.1990 Writing biology: Texts in the social construction of scientific knowledge. Madison:
University of Wisconsin Press.
Oster, S. 1981. The use of tenses in reporting past literature. In L. Selinker, E. Tarone and V. Hanzeli
(Eds.), English for Academic and technical purposes (pp. 76-90). Rowley, MA: Newbury
House.
Paltridge, B. 1997. Thesis and dissertation writing: preparing ESL students for research. English for
Specific Purposes, 16 (1), 61-70.
Pennycook, A. 1996. Borrowing others’ words: Text, ownership, memory, and plagiarism. TESOL
Quarterly. 30 (2): 201-230.
Pickard, V. 1993a. Citing previous writers: What can we say instead of "say". Paper presented at
RELC Conference, 1993. Singapore.
Pickard, V. 1993b. Should we be teaching refutation? Concordanced evidence from the field of
applied linguistics. In N. Bird, J. Harris & M. Ingham (Eds.), Language and content. Hong
Kong: Institute of Language in Education.

22
Scollon, R. 1994. As a matter of fact: the changing ideology of authorship and responsibility in
discourse. World Englishes, 13: 34-46.
Shaw, P. 1992 Reasons for the correlation of voice, tense and sentence function in reporting verbs.
Applied Linguistics, 13: 302-19.
Swales, J. 1990 Genre Analysis: English in academic and research settings. Cambridge: CUP.
Tadros, A. 1993 The pragmatics of text averral and attribution in academic texts in M. Hoey (ed)
Data, description, discourse. London, Harper Collins.
Thomas, S. & Hawes, T. 1994. Reporting verbs in medical journal articles. English for Specific
Purposes, 13: 129-148.
Thompson, G. 1996 Voices in the text: discourse perspectives on language reports. Applied
Linguistics, 17: 501-530.
Thompson, G. & Ye, Y. 1991 Evaluation of the reporting verbs used in academic papers. Applied
Linguistics, 12: 365-82.

Activity And Evaluation Reporting Practices In Academic Writing

More Related Content

Similar to Activity And Evaluation Reporting Practices In Academic Writing (20)

More from Michelle Shaw (20)

Recently uploaded (20)

Activity And Evaluation Reporting Practices In Academic Writing