Iberica 13
Ibérica 39 (2020): 215-242
ISSN: 1139-7241 / e-ISSN: 2340-2784
Abstract
Academic writers represent themselves in their texts in different ways, notably
through use of first person pronouns to construct an authorial voice and
enhance arguments. This study examines how expert writers in the disciplines of
Literature and Computer Science use first person pronouns. The hypothesis is
that in the absence of objective fact, Literature writers resort to frequent use of
first person pronouns backed by stronger authorial roles to build credibility and
convince readers, while Computer Science writers avoid first person pronouns in
line with conventional wisdom in the hard sciences. The findings suggest that the
general dichotomy between hard and soft sciences regarding first person
pronouns usage may not apply in all cases. Our study discusses the similarities
and differences in the disciplinary conventions in Literature and Computer
Science, thus making contributions towards pedagogy and scholarship of the
role of first person pronouns in voice construction in academic texts.
Keywords: authorial voice, academic writing, English Literature, Computer
Science, journal articles.
Resumen
La voz del autor en la escritura académica: un estudio comparativo de artículos
de investigación en revistas de literatura inglesa y de ciencias de la computación
Los autores de artículos académicos se representan a sí mismos en sus textos de
diferentes maneras, en especial a través del uso de pronombres de primera
persona para construir su voz como autores y realzar sus argumentos. Este
Authorial voice in academic writing: A
comparative study of journal articles in
English Literature and Computer Science
Yin Ling Cheung & Louwena Lau
National Institute of Education, Nanyang Technological University
(Singapore)
yinling.cheung@nie.edu.sg, llsw1301@gmail.com
215
Ibérica 39 (2020): 215-242
YIn LIng CHEung & LOuWEnA LAu
estudio examina cómo los escritores experimentados de dos disciplinas
(literatura y ciencias de la computación) emplean pronombres de primera
persona. La hipótesis es que los investigadores de Literatura, en ausencia de
datos objetivos, recurren a un uso frecuente de los pronombres de primera
persona y a la adopción de una voz mucho más personal para otorgar
credibilidad a su discurso y convencer a los lectores, mientras que los
investigadores de Informática evitan los pronombres de primera persona, como
es habitual en las ciencias puras y aplicadas. Los resultados sugieren que la
dicotomía general que suele establecerse entre las ciencias “duras” (puras y
aplicadas) y las “blandas” (humanidades y ciencias sociales) en relación con el
empleo de pronombres de primera persona no siempre se manifiesta. Asimismo,
este trabajo discute acerca de las similitudes y diferencias en las convenciones
disciplinares en literatura y en ciencias de la computación, de tal forma que
contribuye a la pedagogía y al estudio del papel de los pronombres de primera
persona en la construcción de la voz del autor de textos académicos.
Palabras clave: voz del autor, escritura académica, Literatura inglesa,
Informática, artículos de investigación.
1. Background of the study
In the study of academic writing, the projection of a writer’s identity on
the page has been a continuing focus of research interest. The
interweaving of language and identity suggests the presence of the writer
is inevitable inwriting and has been explored through the concepts of
voice and stance. While they are acknowledged as central to the social
interaction between reader and writer in a text, definitions of each vary.
According to Hyland (2012), stance largely involves the writer’s expression
of attitudes and assessment of knowledge in a text to say something new,
while voice, which is more reader-oriented, concerns the framing of
knowledge according to disciplinary conventions. Hewings (2012) sees
stance as the textual characteristics involved in persuasion, evaluation, and
judgements, while voice takes on a wider perspective, involving the
construction of writer identity. Research into voice thus helps to locate
the “person behind the written word” (Hirvela and Belcher, 2001: 85),
perceived by the reader. Whether this is an actual person, a persona or
identity created to suit audience, context and purpose, the creation of a
strong authorial voice is important in building a credible “discoursal self ”
(Ivanič, 1998: 24) to persuade and engage readers and is considered
integral to good writing.
216
At higher level of academic writing, constructing an appropriate authorial
voice that aligns with disciplinary conventions is considered essential if
writers are to be deemed competent by the discourse community (Harwood,
2005c; Hyland, 2010). The ability to frame knowledge in ways the reader
values, by reshaping the language used by others in the same institutional
context, adds to writer-reader interaction.
A strategy used by writers to create authorial voice is the use of first person
pronouns, probably the strongest indicator of a writer’s presence in the text.
Some studies focusing on first person pronouns across disciplines have
revealed differing usage conventions for first person pronouns in the so-
called hard and soft sciences. In the hard sciences, the writer’s presence in
the text is frequently downplayed, heightening perception of the objectivity
and reliability of research activities and methods. Conversely, in the
humanities and social sciences, personal reference appears to be more
important in establishing credibility and making clear one’s contribution to
the field in the absence of empirical proof (Hyland, 2005). While the
dichotomy between hard and soft sciences is used as a convenient way of
categorising academic disciplines (Harwood, 2005a; Hyland, 2001, 2002), the
reality is disciplinary conventions are dynamic and these assumptions may
not always hold. for example, Sword (2012) found higher percentages of
first person pronouns in the hard disciplines compared to the soft ones in
her study comprising 500 articles across 10 disciplines. Hyland and jiang
(2016) found that first person pronouns in the hard disciplines had increased
while certain soft disciplines featured fewer such pronouns in 360 journal
articles from four disciplines over 50 years.
In many studies so far, it has been found that the first person plural, we, is
the most prominent in academic writing. According to mcCarthy (2015),
some researchers still insist on the collective we even though articles may be
individually authored. This is especially true of the hard sciences, where
there is need to balance significance of one’s research but portray a modest
persona in doing so. It is not uncommon to find the first person plural in
single-authored articles performing an assortment of roles, while the first
person singular I is barely visible or may be completely absent (Harwood,
2005c; Hyland, 2001; Kuo, 1999; Yakhontova, 2006). due to its prevalence,
some studies have focused on the inclusive and exclusive use of we which
can be used to refer to the writer, the writer and the reader, the writer and
disciplinary colleagues or humankind in general (mcCarthy, 2015; Wu &
Zhu, 2014). Writers have taken advantage of this ambiguity to achieve aims
A COmpARATIvE STudY Of jOuRnAL ARTICLES In EngLISH LITERATuRE And COmpuTER SCIEnCE
Ibérica 39 (2020): 215-242 217
such as building rapport with readers, thereby expediting acceptance and
concurrence of claims (Harwood, 2005b); diminishing responsibility by
attributing a point of view to a whole community; or highlighting current
issues.
Though fewer in number, other studies have investigated the full range of
first person pronouns forms, providing insights into how writers differ in
their use of these pronouns in different disciplines. for example, Carciu
(2009) found in her study of biomedical writing that our, considered a
weaker form compared to we, was used to convey a more tentative position,
whereas Lafuente millán (2010) saw it as a strong marker of ownership in his
study across four different disciplines.
At present, a complete mapping of all disciplinary conventions remains to
be achieved; many sub-disciplines and smaller fields such as Statistics,
performing Arts, and Literature, have yet to be examined. factors such as the
pressure to publish and perception of their readers as a specialised audience
requiring less explicit engagement, could change the way writers present
themselves and interact with their audience (Hyland & jiang, 2016). There
are also limitations in the taxonomies developed by researchers including
Beerits (2016), Harwood (2005a), Lafuente millán (2010), macgrath (2016)
and Sheldon (2009) to facilitate the qualitative analysis of first person
pronouns because they were developed specifically for the corpora in the
respective studies and hence may not be generalisable. Therefore, much
scope remains for research into how writers use first person pronouns to
achieve the goals of conveying their personal intentions in alignment with
disciplinary conventions, as they interact with audiences in different
disciplines.
By comparing the use of first person pronouns in Literature and Computer
Science, we aim to investigate the general patterns of first person pronoun
use in hard and soft disciplines. According to Biglan (1973), hard science
fields are paradigmatic; that is, all members of the field subscribe to a shared
set of underlying theories. In Computer Science, well-designed experimental
work to test theories or provide proof is valued (Zobel, 2014). The authorial
voice thus takes on the role of the researcher in an empirical process,
whereas in Literature, one seeks to coax readers to accept the “informal logic
and audience appeals” (Wilder, 2005: 83) through a strong authorial voice, in
the absence of empirical evidence. The views proffered in literary writing are
personally motivated and multiple interpretations of works increase an
YIn LIng CHEung & LOuWEnA LAu
Ibérica 39 (2020): 215-242218
understanding of an issue (Stevens & Stewart, 1996). This view of the soft
and hard fields was affirmed in a study of RAs in Second Language Writing
(SLW), a sub-discipline of Applied Linguistics and Computer networks and
Communications (CnC), a sub-discipline of Computer Science (Zhang &
Cheung, 2017). It was found that in CnC papers, readers were assumed to
possess a certain level of background knowledge and thus writing was more
tightly-structured and standardised, following predictable paths. In SLW,
knowledge is more subjective, varying due to the influence of social, cultural
and historical factors. Consequently, writers had to argue for the validity of
their own study against alternative interpretations.
given the differences in structuring knowledge, epistemic beliefs and
research processes, the two contrasting fields of Literature and Computer
Science can serve as foils for one another by throwing peculiarities of the
other into sharp relief, thereby uncovering patterns of use to add to our
knowledge of academic genres and contribute to pedagogy and student
understanding of voice construction. The research questions for this study
are as follow:
1. What disciplinary variations emerge in the frequency of first
person pronoun forms and roles in Literature and Computer
Science RAs?
2. How do the functional roles behind the first person pronouns
differ in Literature and Computer Science RAs?
2. Analytical framework
Tang and john’s (1999) taxonomy, based on Systemic functional Linguistics
(SfL), links linguistic forms of first person pronouns to their various
functions in the text. The taxonomy was adopted for this study to analyse
first person pronoun functions in hard and soft disciplines. The taxonomy
comprises six possible identities for first person pronouns in academic
writing imbued with degrees of authorial power.
The taxonomy has provided the basis for other researchers to scrutinise a
variety of corpora including RAs written by L1 and L2 researchers within
one discipline (e.g. dontcheva-navrátilová, 2013); across disciplines (e.g.
mcgrath, 2016); across different languages and disciplines (e.g. muñoz,
2013; Sheldon, 2009); and in phd theses (e.g. Starfield & Ravelli, 2006). The
A COmpARATIvE STudY Of jOuRnAL ARTICLES In EngLISH LITERATuRE And COmpuTER SCIEnCE
Ibérica 39 (2020): 215-242 219
adaptations involved conflating identities or adding new identities to Tang
and john’s (1999) framework. However, these new roles were specific to the
corpora in the respective studies.
for this study, two new roles, Researcher and Explainer, were added to better
accommodate the dataset and facilitate qualitative analysis (see Table 1).
• ‘I’ as Researcher.
The Researcher describes, explains and provides the rationale for
conducting research by providing an overview of the landscape,
justifying the approach and defining terms. In this role, the writer
demonstrates his expertise of the topic, drawing on his knowledge of
the field, the research done so far and pulling in references from
external sources when necessary and appropriate to provide a
rationale for the current study. This role is more prominent in the
literature papers where the Researcher weaves the perspectives
proposed into an argument in the absence of a methodological
section. In other words, the Researcher is a guide through the
argument and not the structure of the text, thus while the Architect
has an overview of the flow of the article, the Researcher has an
overview of the flow of the proposition or argument.
• ‘I’ as Explainer of the Research process.
The Explainer is akin to a class teacher demonstrating application of
formulas and explaining assumptions. The Explainer interprets results
and information presented in diagrams and tables for the research
process and is pertinent only to Computer Science RAs. unlike the
guide role which is involved with textual organisation throughout the
article, the Explainer specifically takes the reader through the steps in
the research process detailed in the article. Hence, it is mainly
associated with the methodology section of the Computer Science
papers.
In addition to the new roles detailed above, some adjustments have
been made to the guide and Architect roles: the guide draws the
attention of the reader to a specific point in the article, while the
Architect controls the overall flow of the article and is commonly
found near the beginning of the article or section where the author
informs the reader of the structure of the whole article or the
following section. The revised coding scheme is shown as follows:
YIn LIng CHEung & LOuWEnA LAu
Ibérica 39 (2020): 215-242220
3. Methodology
To address the research questions, a corpus of 160 RAs (80 Computer
Science and 80 Literature) totalling about 1.3 million words was collated
from journals listed in the Arts and Humanities, and Science citation indexes
by Clarivate Analytics. Eleven journals selected from each discipline capture
intra-genre variation (see Appendix A). Single-authored papers were
selected, to avoid bias or inadvertently increase the frequency of first person
plural pronouns in the study due to multiple-authored papers. The articles
span 20 years from 1991 to 2011, the time frame determined by accessibility
A COmpARATIvE STudY Of jOuRnAL ARTICLES In EngLISH LITERATuRE And COmpuTER SCIEnCE
Ibérica 39 (2020): 215-242 221
Meaning Typical expressions
‘I’ as Representative “…generic first person pronoun, used as a proxy
for a larger group of people …reduces the writer
to non-entity.” (p. S27)
our languages, our
nation’s bicentennial,
our society, enable us, taught us
‘I’ as Guide “…writer as guide during the reader’s
journey…locates reader and writer together in the
time and place of the essay, draws reader’s
attention to points plainly visible within the essay.”
(p. S27)
as we saw, as we have seen, we
now turn, let us denote, let us
assume,
we have the following
‘I’ as Architect “…foregrounds person who writes, organises,
structures and outlines the material in the essay.”
(p. S28)
In this paper we will, In this
section we discuss, we prove
later, we conclude in section 5, In
the rest of the paper
‘I’ as Recounter
of the Research
Process
“…recounts various steps in the research
process…which might include reading source
texts, interviewing subjects, collecting data and
so on.” (p. S28)
we ran the simulation, our training
set consisted of, we have also
performed tes
‘I’ as Explainer
(new role found
only in Computer
Science papers)
….demonstrates step-by-step how various
formulas are applied, explains assumptions
made… interprets results which relate to findings
resulting from research.
we get, we have, we assume, we
conclude that, we find, we know,
we can say
‘I’ as Researcher
(new role)
… explains and provides rationale for carrying
out research, provides overview of landscape,
justifies approach and defines terms…references
material external to the topic or text, draws from
the writer’s own experience.
we define, we discuss we refer
the reader to, the models we
employ, we are never told, turns
our attention towards, reminds us,
we are told
‘I’ as
Opinion Holder
“…shares an opinion, view or attitude with regard
to known information or established facts.” (p.
S28)
we believe, we feel, we
leave as a future challenge, I
argue, we wish, we suggest
‘I’ as Originator “…involves the writer’s conception of ideas or
knowledge claims advanced in essay…calls for
the writer to present or signal these as new in the
essay.” (p. 29)
we find, we introduce, I argue, I
affirm, I suggest, my contention,
my claim, my point
Table 1. Coding scheme for first person pronoun roles (adapted from Tang & John, 1999).
to soft copies of the articles due to journals’ moving walls. Each of the 20
years is represented in the dataset for both disciplines, though they were not
pulled with equal distribution from each year. The intention was to capture
changes in academic writing in the disciplines, but no clear diachronic trends
emerged from the data.
The AntConc version 3.4.4 concordancer was used to identify all
occurrences of I, we, us, our, ours, me, my and mine, in the articles, excluding
abstracts, footnotes, acknowledgements, references, tables, diagrams and
captions. The corpus was compiled manually to ensure instances of first
person pronouns highlighted by the concordancer were valid. descriptive
statistics of the corpus are shown in Table 2.
The plural form we, which was most frequent in both disciplines, was
classified into inclusive and exclusive uses to determine which was more
dominant in the respective sub corpora. possible writer intentions are
discussed.
3.1. Coding the data
3.1.1. Pronouns and co-text
To control disparities in the data due to varied article lengths, frequencies of
first person pronouns were normalised per 10,000 words. All first person
pronouns were manually coded. Each pronoun was examined in context to
determine its functional role. during coding, the linguistic environment
surrounding the pronoun was important to determine its role. factors taken
into consideration included (a) immediate co-text, i.e., text within the same
sentence; (b) wider linguistic environment, i.e., sentences which occur before
or after the sentence featuring the pronoun; (c) verb which co-occurs with
the pronoun; and (d) position of pronoun in the article. for example, in (1),
the immediate co-text suggests this could be classified as Explainer since it
involves the discussion of an equation:
YIn LIng CHEung & LOuWEnA LAu
Ibérica 39 (2020): 215-242222
Each of the 20 years is represented in the dataset for
b
T
Literature RAs Typical expressions
No. of RAs 80 80
Total no. of words 662165 638632
Mean no. of words 8277 7983
Table 2. Descriptive statistics of the corpus.
T
3
(1) Furthermore, we will consider problems where f is restricted to a bounded
domain Ω e Rn. This is particularly important when the variables x are
related to physical quantities which are bounded in the real world. (CS26)
However, the sentences are in the Introduction of the article. This is unusual
as the Explainer role normally occurs after information such as the purpose
of the paper, the approach and definition of terms. Considering preceding
and following sentences, as well as the auxiliary verb will, which suggests
future intention, we coded it as the Researcher, discussing what he intends
to show in the paper.
3.1.2. Overlapping of roles
Some researchers referencing Tang and john’s (1999) taxonomy either
conflated the roles of guide and Architect (dontcheva-navrátilova, 2013)
or excluded one (Lafuente millán, 2010; martín, 2003). In this study, the
guide draws the attention of the reader to a specific point in the article,
while the Architect controls the overall flow of the article. The guide can
reference what is ahead or has gone before and can be found throughout the
article, but the Architect role is commonly found near the beginning of the
article or section to inform the reader of the structure of the whole article
or following section.
Overlaps in other categories have been found in previous studies (Leedham
& fernandez parra, 2017; mcgrath, 2016) and are also encountered in the
current corpus. In (2), the phrase In this paper, suggests we could be classified
as Architect, but the verbs present and are derived in the following sentence
suggested that an Originator role was more appropriate and this pronoun
was coded as such.
(2) I n th i s pa pe r, w e pr e se nt the Sugeno integral semantics of linguistic
quantifiers in which a quantifier is represented by a family of fuzzy measures
[35] and the truth value of a quantified proposition is computed by using
Sugeno’s integral [35]. Several elegant logical properties of linguistic
quantifiers are derived including a prenex normal form theorem. (CS59)
Adopting mcgrath’s (2016) approach in general, where two roles seem to be
present in one pronoun, it would be assigned the role for which the case is
stronger.
A COmpARATIvE STudY Of jOuRnAL ARTICLES In EngLISH LITERATuRE And COmpuTER SCIEnCE
Ibérica 39 (2020): 215-242 223
3.1.3. Dual roles in one sentence
In literary texts that feature complex sentences, different pronouns can be
perceived to perform different roles in the same sentence. for example, the
presence of the adverbs earlier and now in (3) may suggest that the two
pronouns should be classified as guide but the verb suggest and adverb also
seem to indicate that the author is putting forth a claim that the novel is
taking on a different function, thus giving the second pronoun an Originator
role.
(3) I h a ve n o t ed e ar l i e r that the novel may be read as hysterical symptom;
what I w o ul d n o w l ik e to su g g es t is that the novel also performs the
function of masquerade. (LIT45)
3.1.4. Subjectivity in coding
during data coding, a degree of subjectivity can occur as interpretation of
language can vary. furthermore, unfamiliarity with Literature and Computer
Science could be a factor influencing interpretation.
3.2. Reliability measure
An mA in Applied Linguistics student enrolled in a university in Singapore
served as co-rater. Two rounds of inter-rater agreement were calculated for
separate sub-sets of data using Cohen’s Kappa and final inter-rater reliability
was 0.82. The second author of this paper looked through the whole corpus
to ensure consistency in coding. means and standard deviations of pronouns
in each role were calculated before independent t-tests were carried out with
IBm SpSS 23 to determine the differences between the two subcorpora. The
alpha was set at .05 for all the inferential analyses in this study.
4. Results and discussion
The findings reveal the number of first person pronouns in Computer
Science RAs was more than 2.5 times the number in Literature articles (see
Table 3), unlike the findings in previous studies (Harwood, 2005b; Lafuente
millán, 2010). Authorial voice was conveyed without use of these pronouns
in only eight Computer Science articles, whereas all the Literature articles
featured at least one instance of a first person pronoun (see Table 4).
YIn LIng CHEung & LOuWEnA LAu
Ibérica 39 (2020): 215-242224
Therefore, the difference in first person pronoun use between Computer
Science and Literature RAs is not whether these are absent or present, but
rather, how and the degree to which writers in these disciplines utilise these
self-mention markers to persuade readers and present results.
While associating certain characteristics with hard and soft disciplines is
convenient for classification, the findings here suggest these should not be
always assumed to be the case (cf. Sword, 2012). As competition is more
intense in a large field like Computer Science, the use of first person
pronouns can act as a strong persuasive force, facilitating solidarity with the
reader (Hyland, 2010) and engendering a promotional tenor (Harwood,
2005b) in a crowded research space. The increasing recognition of human
agency in the process of scientific discovery could also account for a higher
frequency of first person pronouns in the Computer Science RAs. In the
case of Literature, while first person pronouns are employed to create a
credible persona and strengthen arguments, writers also use other strategies
to construct a strong authorial voice, such as framing opinion as assumed
truth, using agentless passive constructs and extensive use of metadiscourse
markers.
Though all RAs were single-authored, plural pronouns (we, us, our, ours)
dominated both fields. Conversely, the singular pronouns (I, me, my, mine)
occurred three times more in Literature. Table 5 summarises the descriptive
statistics of the first person pronouns in the corpus.
A COmpARATIvE STudY Of jOuRnAL ARTICLES In EngLISH LITERATuRE And COmpuTER SCIEnCE
Ibérica 39 (2020): 215-242 225
AUTHORIAL VOICE IN ACADEMIC WRITING: A COMPARATIVE STUDY OF JOURNAL ARTICLES IN
ENGLISH LITERATURE AND COMPUTER SCIENCE
Ibérica 39 (2020): xx-xx
a strong persuasive force, facilitating solidarity with the reader (Hyland, 2010)
and engendering a promotional tenor (Harwood, 2005b) in a crowded research
space. The increasing recognition of human agency in the process of scientific
discovery could also account for a higher frequency of first person pronouns in
the Computer Science RAs. In the case of Literature, while first person pronouns
are employed to create a credible persona and strengthen arguments, writers also
use other strategies to construct a strong authorial voice, such as framing opinion
as assumed truth, using agentless passive constructs and extensive use of
metadiscourse markers.
Total Per 10,000 Words
Liiterature 662165 30.00
Computer Science 638632 76.24
Table 3. Raw and normalised occurrence of first person pronouns in RAs
Literature Computer Science
No. of articles % No. of Articles %
I 8 10% 71 89%
We 69 86% 73 91%
Us 49 61% 62 78%
Our 55 69% 50 63%
Ours 2 3% 2 3%
Me 2 3% 6 8%
My 5 6% 48 60%
Mine 0 0% 5 6%
Table 4. Articles featuring occurrences of first person pronouns
Though all RAs were single-authored, plural pronouns (we, us, our, ours)
dominated both fields. Conversely, the singular pronouns (I, me, my, mine)
occurred three times more in Literature. Table 5 summarises the descriptive
statistics of the first person pronouns in the corpus.
Literary scholars’ use of first person pronouns spreads out among the eight
pronominal forms, probably due to the language proficiency which literary
writers are expected to display, resulting in a wider range of expressions and
textual structures in the construction of a strong authorial voice.
4.1. Analysis of most commonly occurring pronominal forms
4.1.1. We the most preferred pronoun
The we pronoun occurred four times as much in the Computer Science
subcorpus as in the Literature subcorpus, in line with the collaborative nature
of knowledge creation in hard sciences, where computer scientists mark
their individual contributions and also display appropriate collegiality. In
Literature, the high frequency of we demonstrates that, while the
individualistic, unique point of view conveyed via the I is expected, this has
to be tempered by a more communal tone at appropriate junctures.
However, writers exploit the ambiguity of the first person plural, switching
seamlessly between its inclusive and exclusive forms to achieve desired
effects within the same article. In (4), where the writer regards readers to be
literary scholars, it is inclusive; but it excludes those not in this category.
(4) As literary critics working within interdisciplinary studies, w e are interested
in authors who drew the medical laboratory into the imaginative landscape of
the novel…(LIT69)
Inclusive we dominated (90%) in the Literature RAs, where it strengthened
arguments by creating a more familiar tone to enhance interaction with
readers. Though we occurred four times more in the Computer Science
articles, it was distributed between the exclusive (56%) and inclusive (44%)
YIn LIng CHEung & LOuWEnA LAu
Ibérica 39 (2020): 215-242226
YIN LING CHEUNG & LOUWENA LAU
Ibérica 39 (2020): xx-xx
Literature Computer Science
Raw Mean SD Raw Mean SD
We 837 10.48 15.99 3871 48.44 53.57
Our 220 2.75 3.94 537 6.71 10.33
Us 287 3.58 3.57 271 3.34 5.42
Ours 2 0.03 0.16 5 0.06 0.46
I 504 6.30 5.01 161 2.00 9.47
My 113 1.41 1.78 21 0.26 1.43
Me 17 0.21 0.79 3 0.04 0.25
Mine 5 0.06 0.26 0 0.00 0.00
TOTAL 1985 3.10 7.11 4869 7.61 25.03
Table 5. Mean and SD for first person pronouns in RAs
Literary scholars’ use of first person pronouns spreads out among the eight
pronominal forms, probably due to the language proficiency which literary
writers are expected to display, resulting in a wider range of expressions and
textual structures in the construction of a strong authorial voice.
4.1. Analysis of most commonly occurring pronominal forms
4.1.1. We the most preferred pronoun
The we pronoun occurred four times as much in the Computer Science subcorpus
as in the Literature subcorpus, in line with the collaborative nature of knowledge
creation in hard sciences, where computer scientists mark their individual
contributions and also display appropriate collegiality. In Literature, the high
frequency of we demonstrates that, while the individualistic, unique point of
view conveyed via the I is expected, this has to be tempered by a more
communal tone at appropriate junctures.
However, writers exploit the ambiguity of the first person plural, switching
seamlessly between its inclusive and exclusive forms to achieve desired effects
within the same article. In (4), where the writer regards readers to be literary
scholars, it is inclusive; but it excludes those not in this category.
(4) As literary critics working within interdisciplinary studies, we are
interested in authors who drew the medical laboratory into the
imaginative landscape of the novel…(LIT69)
Inclusive we dominated (90%) in the Literature RAs, where it strengthened
arguments by creating a more familiar tone to enhance interaction with readers.
Though we occurred four times more in the Computer Science articles, it was
distributed between the exclusive (56%) and inclusive (44%) forms. The slightly
higher number for the exclusive we in Computer Science could be attributed to
the discipline’s preference for using we rather than I to reflect the communal
nature of knowledge conventions (Yakhontova, 2006; Harwood, 2005a) as the
former is used as a strategic tool to maintain a balance of authority and humility.
forms. The slightly higher number for the exclusive we in Computer Science
could be attributed to the discipline’s preference for using we rather than I to
reflect the communal nature of knowledge conventions (Yakhontova, 2006;
Harwood, 2005a) as the former is used as a strategic tool to maintain a
balance of authority and humility. Since scientists are expected to display a
modest collegial persona, the use of we over I may reduce the force of
imposition (myers, 1989).
We was used extensively in the Explainer role to take the reader though the
methodology in Computer Science RAs. By alternating between the inclusive
and exclusive we, writers gave the impression readers were participants in a joint
process through the methodology and drew on the strength of the discourse
community (Harwood, 2005b) to establish themselves as confident professionals
with strong technical expertise. In (5), through the use of the inclusive we, the
writer assumes readers possess shared knowledge about the field.
(5) Note that when w e test the past temporal operators in the if clause against
the past history, w e do not have to traverse the entire past history of temporal
predicates. (CS20)
Where we is used for structuring the text, making decisions for methods or
processes involved and making claims, it is exclusive. In (6), although we is
used, in this context, it is the writer who is putting forth the proposal for the
control schemes, thus showing his expertise in the field.
(6) In the following w e will propose two control schemes that will accomplish the
above objective. (CS2)
for the Literature RAs, the role behind the highest number of we pronouns
is that of Researcher, while it was the second highest for Computer Science.
plural pronouns were used to involve readers in the elaboration of
arguments and persuade them about the soundness of the interpretations
proposed by the writers.
4.1.2. Our more frequent in computer science than literature
Similar to findings in Lafuente millán’s study (2010) where the possessive
adjective our was more frequent in hard disciplines, it is the second most
frequently used pronominal form in Computer Science occurring more than
twice as frequently compared to Literature.
A COmpARATIvE STudY Of jOuRnAL ARTICLES In EngLISH LITERATuRE And COmpuTER SCIEnCE
Ibérica 39 (2020): 215-242 227
In the Computer Science RAs, our was most often (64%) used to express
ownership. At other times (26%), it complemented the inclusive we and built
communality by suggesting the reader’s joint involvement with the writer in
the empirical process, as seen in (7) or when discussing the field with the
reader. To a smaller extent, our was also used by writers to hedge their claims
(10%) such as in (8) below, where the writer takes on the role of Opinion
Holder to offer his perspective, using the plural pronoun “our feeling is
that…”, opening the space for others who may feel differently.
(7) We may thus restrict o ur attention to a formula of the form (u1,…uk) *
(u1,…uk, x), where the um+1,…uk are actually xi quantified from further
outside. (CS29)
(8) Ou r feeling is that it is not easy to find a simple and significant extension to
evaluable formulas because there are relatively few formulas that are
equivalent…(CS12)
Literary scholars used the pronoun inclusively to suggest writer and reader
agree on something or that what they say can be generalised to humankind
in general. The pronoun in (9) seems to implicate a larger community
beyond the writer and reader of the article as morals and morality are usually
upheld at societal level.
(9) These terms are unmistakably moral but are no longer a part of o u r moral
vocabulary. (LIT2)
4.1.3. Similar frequencies of Us in both subcorpora
The pronoun us was third most frequent in the corpus but used to different
effects in the two disciplines. In Computer Science, almost half (46%) were Let
us imperatives, commonly followed by verbs such as consider, assume, suppose, take
and look. The reader is assumed to be an intellectual equal and invited to
participate in examining the methodology and evaluation in the research
process. The writer in (10) presupposes the reader has ability to analyse the
phenomena under discussion. Such invitations exert a strong rhetorical effect
on attracting readers’ interest, endorsing the recommended approach and
acknowledging validity of the writer’s work (fløttum, Kinn & dahl, 2006).
(10) Let us here analyze the simplifications in the discretization of the continuous
mixture that its usage enables. (CS61)
YIn LIng CHEung & LOuWEnA LAu
Ibérica 39 (2020): 215-242228
At times Let us did not include the reader (4%) but rather assumed their
indulgence or ratification, as conveyed by the expressions Let us put it bluntly
or Let us make the following comments.
In contrast, Let us was almost absent from Literature RAs, constituting only
4% of the us pronouns in these papers as the writer does not involve his
reader in an empirical fact-finding process. In the Literature papers, us was
used mainly to enhance reader involvement and textual coherence as
illustrated by (11) where the pronoun is used to remind readers of how an
insight into socio-economic conditions can add to deeper understanding and
appreciation of the text.
(11) Social and economic conditions also help us to resituate texts in their cultural
moment. (LIT18)
4.1.4. I the second most common pronoun in literature RAs
Based on the greater subjectivity of Literature, I was the second most
frequent pronoun in the Literature RAs. It was present in 89% of Literature
RAs compared to 10% of Computer Science RAs. The unambiguous
reference to the writer could be reflective of the discipline that values
originality in interpretation and relies on an assertive authorial voice to carry
the arguments in the absence of empirical evidence. In (12), the singular first
person pronoun takes on the role of Originator as the writer openly backs
her claim with I would argue and differentiates her view from that of another
member in the field, by beginning the sentence with On the contrary.
(12) On the contrary, I would argue that the women are seen to be using the
limited means available to them to settle their disputes, and that the level of
resourcefulness which they are obliged to display in doing so is evidence of the
gender-specific nature of their efforts. (LIT39)
The frequency of the first person pronouns in this study has also uncovered
intradisciplinary variations. In the Computer Science subcorpus, I was absent
in a majority of the articles but was heavily used in a handful. This finding
is consistent with previous studies of expert writing in Computer Science
where the occurrence of I was low (Harwood, 2005a) or completely absent
(Kuo, 1999), thus making it hard to generalise about the use of the pronoun.
Similarly, in the Literature RAs, though the I pronoun was used by most
writers, it was absent in some articles, possibly due to factors such as
A COmpARATIvE STudY Of jOuRnAL ARTICLES In EngLISH LITERATuRE And COmpuTER SCIEnCE
Ibérica 39 (2020): 215-242 229
sociocultural or individual author preferences which may contribute to the
construction of authorial voice of writers in academic fields.
4.1.5. My is used to claim ownership
The possessive adjective my was featured in more than half the number of
articles in the Literature RAs but appeared in a mere five articles in Computer
Science. The majority of the uses of my in Literature (85%) and Computer
Science (77%) conveyed ownership while the remainder were perceived to be
hedges or referential. previous studies have found my to be non-existent in the
hard sciences. Harwood’s (2005a) study of writing in Computer Science does
not feature the pronoun while in Lafuente millán’s (2010: 42) study, my was
found to be non-existent in the hard science disciplines but “slightly more
frequent” in Applied Linguistics and Business Studies. However, in doncheva-
navrátilová’s (2013: 16) paper, they were rare. The presence of my in
Computer Science in this study suggests that conventions shape rather than
dictate writing in the disciplines and there is leeway for individual choices in
the use of these pronouns. In Literature, the use of my is important to convey
the assertiveness of the writer when making a claim as in (13), where the
Originator role is fronted by the possessive adjective. The writer makes clear
her difference of opinion from an exhaustive study in unequivocal terms, taking
ownership of her point of view.
(13) Celeste Turner Wright remarks in her exhaustive study of this figure in
Elizabethan literature... In m y view, however, the antiusury tradition did
not simply become extinct as England embraced the ethos of capitalism.
(LIT20)
Apart from expressing ownership, the pronoun was used as a hedge in small
number of instances (6%) where the writer acknowledges the subjectivity of
the argument or limits of his/her expertise in the role of a Researcher as
seen in (14).
(14) To the best of m y knowledge there is no hard evidence to support the links
between Cantos and Kestoi, no document, letter, no explicit reference.
(LIT34)
Similarly, in Computer Science, the pronoun expressed ownership (77%),
though there were a higher number of hedges (19%), perhaps reflecting the
more communal culture of the discipline. In (15) below, the writer cites his
YIn LIng CHEung & LOuWEnA LAu
Ibérica 39 (2020): 215-242230
experience, allowing for others who may have had different experiences and
uses It is suggested in the next sentence to lessen assertiveness in putting across
his proposition.
(15) In particular, m y experience with PDMOSA says that PDMOSA
algorithm performs reasonably well for continuous function problems with not
too many variables. It is suggested that all algorithms should be used to
generate a larger set of optimal solutions... (CS50)
The examination of first person pronoun frequencies in the corpus reveal
that they are important in creating a credible persona to persuade the reader,
acknowledge works of others and promote one’s own contributions
(Hyland, 2009). These objectives would have influenced the frequency and
use of the pronouns in the corpus.
4.2. Functional roles of first person pronouns
4.2.1. Continued learning
The most powerful roles in the Tang and john (1999) taxonomy adopted for
this study are those of Researcher, Opinion Holder, and Originator. These
account for about 85% of first person pronouns in the Literature subcorpus.
They account for less than 30% of roles fronted by first person pronouns in
the Computer Science RAs. This suggests that literary scholars used first
person pronouns to project themselves assertively in their texts.
differences in the functional roles of the first person pronouns in Literature
and Computer Science were found to be statistically significant for
Representative, guide, Architect and Opinion Holder, though Cohen’s d
ranged between modest to moderate (see Table 6).
A COmpARATIvE STudY Of jOuRnAL ARTICLES In EngLISH LITERATuRE And COmpuTER SCIEnCE
Ibérica 39 (2020): 215-242 231
AUTHORIAL VOICE IN ACADEMIC WRITING: A COMPARATIVE STUDY OF JOURNAL ARTICLES IN
ENGLISH LITERATURE AND COMPUTER SCIENCE
Ibérica 39 (2020): xx-xx
Literature Computer Science
Mean SD Mean SD t df p
Representative 0.01 0.02 0.00 0.02 1.95 158 0.05
Guide 0.03 0.04 0.11 0.14 -5.14 158 0.00
Architect 0.02 0.02 0.06 0.08 -5.09 158 0.00
Researcher 0.20 0.23 0.17 0.27 0.88 158 0.38
Opinion Holder 0.07 0.09 0.03 0.07 3.00 158 0.03
Originator 0.64 0.05 0.06 0.08 0.26 158 0.80
Tabe 6. Independent t-test findings for Functional Roles (!=.05)
4.2.1. Representative – one among others
The slightly higher number of pronouns in Literature could be due to the
discipline’s association with the human condition. In the Computer Science
subcorpus, one article in the sub-field of Artificial Intelligence accounted for
almost half the occurrences of the role where the writer was discussing the
developments in the field to date. Since reference is made to many rather than
the few, this role is represented by the plural pronouns we, us and our as the
writer identifies as one in a larger group in (16) below.
(16) … Shaw stresses in Pygmalion that such clues, in an era unduly
sensitized to the social import of language, may indicate not only our
past, and our present, but may also determine our future:…(LIT11)
4.2.2. More overt textual organisation in computer science through guide
and architect
Personal intrusions into the text for the purpose of organising discourse are less
frequent in the Literature RAs, similar to Dontcheva-Navrátilová’s Applied
Linguistics corpus (2013). In a discipline known for its rhetorical dexterity
reflected in the journal articles presented as a single, continuous essay, such
overt organisation of the text in the Architect role might be less preferred, unlike
Computer Science RAs which feature marked sections. When both subcorpora
are compared, these roles occur more frequently in the Computer Science
subcorpus, in contrast to Lafuente Millán’s (2010) observation that more first
person pronouns were used in the soft science fields to organise information
because of less linear arguments and more flexible structures. This attests to the
uniqueness of the disciplinary conventions in Literature.
Literature scholars weave textual signposting into their essays and metadiscourse
markers such as frame markers, transitions and code glosses were used to
structure flow of arguments in Literature articles, accounting for lower frequency
4.2.1. Representative – one among others
The slightly higher number of pronouns in Literature could be due to the
discipline’s association with the human condition. In the Computer Science
subcorpus, one article in the sub-field of Artificial Intelligence accounted for
almost half the occurrences of the role where the writer was discussing the
developments in the field to date. Since reference is made to many rather
than the few, this role is represented by the plural pronouns we, us and our
as the writer identifies as one in a larger group in (16) below.
(16) … Shaw stresses in Pygmalion that such clues, in an era unduly sensitized
to the social import of language, may indicate not only o u r past, and o u r
present, but may also determine o u r future:…(LIT11)
4.2.2. More overt textual organisation in computer science through
guide and architect
personal intrusions into the text for the purpose of organising discourse are
less frequent in the Literature RAs, similar to dontcheva-navrátilová’s
Applied Linguistics corpus (2013). In a discipline known for its rhetorical
dexterity reflected in the journal articles presented as a single, continuous
essay, such overt organisation of the text in the Architect role might be less
preferred, unlike Computer Science RAs which feature marked sections.
When both subcorpora are compared, these roles occur more frequently in
the Computer Science subcorpus, in contrast to Lafuente millán’s (2010)
observation that more first person pronouns were used in the soft science
fields to organise information because of less linear arguments and more
flexible structures. This attests to the uniqueness of the disciplinary
conventions in Literature.
Literature scholars weave textual signposting into their essays and
metadiscourse markers such as frame markers, transitions and code glosses
were used to structure flow of arguments in Literature articles, accounting
for lower frequency of the Architect. The guide role occurs more frequently
than Architect as they often make reference to specific points in the text
rather than its overall structure.
The Architect function emerged more frequently in Computer Science with
more than 300 instances compared to 72 in Literature. The higher
occurrence of the Architect role in making explicit the organisation of the
YIn LIng CHEung & LOuWEnA LAu
Ibérica 39 (2020): 215-242232
paper could be reflective of disciplinary discourse where knowledge-making
practices display a linear progression and steps in the process have to be
evident for other researchers to replicate methods.
Although Architect seems to be a textual function, this role was used by
writers in both disciplines to establish expertise by differentiating their work
from others, such as the first I in (17) and we in (18).
(17) It is these plays that I want to consider here; and since I believe the issue of
truth and promise to be a source of major misunderstanding in the
interpretation of Henry IV, m y primary emphasis will be on this, the
greatest of the histories. (LIT38)
(18) Despite the fact that the original two-estimate system considered a and b as
absolute endpoints, in this paper w e define these estimates as determined
fractiles. (CS48)
In Computer Science, the guide role is invoked when the first person
pronoun is used with adverbs such as “now”, “next”, “first” and “here”. In
Literature RAs, expressions such as “As we shall see”, “As I have already
shown” signal the guide role. The guide reminds readers of what has been
discussed or what is to follow. In (19), the reader is reminded of a topic
which was discussed previously with the phrase We will return. The use of I
with the verbs want make for assertive claims of ownership over the content
and flow of argument, even though in a weaker functional role.
(19) We will return to the discrepancy between story and discourse regarding the
mode of historical representation, but for the moment, I want to consider the
implications of the “monumentalism” through which the characters approach
history. (LIT19)
Let us and Let’s are also used in this category in Computer Science articles to
invite readers into the process as willing participants in the textual dialogue
as seen in (20).
(20) Le t’ s consider the water jug problem in AI [53,73]. (CS10)
4.2.3. Recounter of the research process
no instances of this role were found in the Literature subcorpus. References
about the writer’s previous experiences featured in some articles were related
A COmpARATIvE STudY Of jOuRnAL ARTICLES In EngLISH LITERATuRE And COmpuTER SCIEnCE
Ibérica 39 (2020): 215-242 233
as personal anecdotes such as (21) rather than steps in an experimental
process.
(21) A guard, happily waiving the rules, took m e down to the stacks, and there
it was, its soft brown and buff cover as I had last seen it well over half a
century before in the guest lounge of a holiday camp patronized by skilled
low-paid workers. (LIT49)
In the Computer Science subcorpus, this role was second least frequently
fronted by first person pronouns, comprising 2% of the pronouns used,
perhaps because the focus is on methodology detailed in the RAs. The
exclusive we and I front these roles which reinforce the researcher’s expertise
by summarising the procedures which may have been carried out prior to the
writing of the paper as seen in (22).
(22) In order to assess the contribution of the board’s operationalized code
provisions and cases, I conducted a series of experiments, including ablation
experiment. (CS46)
4.2.4. Explainer of the research process
The Explainer is a new role added to the Tang and john (1999) taxonomy
based on analysis of first person pronouns in the Computer Science
subcorpus. In this role, the we pronoun collocates with verbs such as “have”,
“obtain”, “get”, “assume”, “find” and “note” to guide readers through
complex calculations involving models and formulas. By taking advantage of
the fuzziness between the inclusive and exclusive forms of the first person
plural pronoun, writers can create the impression that readers are
participants in the process, even if we may refer to the multiple researchers
who could be behind a single-authored paper. This role accounted for 54%
of the total number of first person pronoun references in the Computer
Science subcorpus as proof of formulas and calculations involved in the
steps or stages in the articles. In (23), the writer reinforces this joint process
with the reader not only by using the first person plural, but also issuing an
invitation to do so with the phrase Let us consider.
(23) L et u s stress that isomorphism is necessary but not sufficient for
ag-equivalence, or, in other terms, ag-equivalence is a stronger condition of
topological similarity. For showing this, w e present the following example.
L e t u s consider the two isomor phic polyhedral S and S’ in
Fig. 14. (CS41)
YIn LIng CHEung & LOuWEnA LAu
Ibérica 39 (2020): 215-242234
use of the exclusive we and possessive pronouns mark the writer’s ownership
of the methodology as illustrated in (24), while in (25), the pronouns are
used to show they are up-to-date with developments by commenting on
results and comparing these to findings by other researchers.
(24) O ur development emphasizes 2-dimensional arrays. Since w e do not build
fractional rows or columns, the values must be integers. (CS39)
(25) Overall, o u r computed results agree well with the experimental data, the
largest discrepancy being of the order of 5%. Ou r results are as accurate
as Veeramani et al (2007) but not as satisfactory as those of Feng and
Michaelidis (2009b). (CS78)
4.2.5. The researcher – Providing the background and developing
arguments
The Researcher sets the stage for the research process by providing
background information such as research carried out thus far, the rationale
for the chosen approach and the goals of the study undertaken. The role is
more prevalent in Literature RAs, referenced by some 60 percent of first
person pronouns. more amenable writer-reader interaction through the use of
the plural pronouns is important in Literature where the argument must be
sustained throughout the article. In fact, occurrences of the plural pronouns
we, us and our were the most commonly used pronouns in this role in the
Literature subcorpus. In (26), the Researcher leads the discussion with the
reader, elaborates and provides evidence to support propositions put forth.
(26) Before w e can make any conjecture, w e should step back for a moment and
consider the conditions under which w e might rightly claim to have found a
solution. (LIT50)
Writers in the corpus also discussed the limitations of their studies in this
role. In Computer Science articles, computer scientists display humility and
communality by highlighting limitations of their own methods or research as
seen in (27). In Literature, the limitations discussed may be attributed to lack
of details in the text as shown in (28).
(27) However, I have not compared GAC-On-X against bounds or range
consistency, so can offer no conclusions on the relative merits of different levels
of consistency. (CS76)
A COmpARATIvE STudY Of jOuRnAL ARTICLES In EngLISH LITERATuRE And COmpuTER SCIEnCE
Ibérica 39 (2020): 215-242 235
(28) For example, w e cannot, and therefore, do not gain access to the “tone of
fervid veneration [and] religious regard” with which Garth speaks of his
vocation (p.250). (LIT58)
The Researcher in both disciplines employed rhetorical questions and
examples to establish expertise or make claims. These provided
opportunities for writers to demonstrate knowledge of a topic or justify a
differentiated approach to an issue. However, where the first person
pronouns are used in declaring new claims as a result of their arguments and
justifications in the Researcher role, the role will be considered that of
Originator. for example, in (29) below, the first instance of the personal
pronoun I would be deemed an Originator putting forth a different (and
therefore original) perspective, while the Researcher (using we) takes over
with the explanation and justification of that perspective.
(29) …what makes it both natural and necessary to ask the question in the first
place – is, I would suggest, the quaint obscurity of the term ‘irreverence’ and
of its contrary ‘reverence’…We feel too uncomfortable using ‘reverence’ in
some of its traditional applications because of a certain undemocratic,
patriarchal, and superstitious cast that the term seems to have acquired.
(LIT2)
In the extract from a Computer Science paper (30) below, the role behind the
pronoun has been coded as the Researcher when making a reference to an
external source in explaining the theory under discussion. Whereas in (31),
drawn from the same paper, the pronoun fronts a new perspective about
polynomials and thus would be considered Originator.
(30) As in Lankford’s approach, w e get a constraint-solving problem, but which
is now in the first order theory of the reals. (CS57)
(31) As in Lankford’s approach, w e get a constraint-solving problem, but which
is now in the first order theory of the reals. (CS57)
4.2.6. Greater prominence of opinion holder in literature papers
The Opinion Holder is more pertinent in Literature than Computer Science
(see Table 6) since interpretations carry the ideational content in Literature
RAs. This role occurred more than twice the number of times in the
Literature than the Computer Science subcorpus. Computer scientists use
YIn LIng CHEung & LOuWEnA LAu
Ibérica 39 (2020): 215-242236
less of this role as in scientific fields, knowledge previously confirmed by the
discourse community takes on the mantle of universal truth and opinion
matters less than objective fact (Hyland, 1998).
The use of first person pronouns with verbs such as “believe”, “wish”,
“propose”, “prefer” and modals such as “can”, “could” and “might” allow
for alternative views thus inviting readers to add their voices to the debate
(doncheva-navrátilová, 2013; Carciu, 2009).
In the Opinion Holder role, apart from offering opinions about previous
research or issues in question, writers also evaluated their own work. Though
the plural pronouns were most common in portraying a humble demeanour
seen in the use of modals and the verb seem in (32), the singular possessive
pronoun mine in the phrase mine is admittedly a minority view was used to
downplay one’s interpretation in (33).
(32) Still, one may ask whether some of the n factors in Agarwal and Matoušek’s
higher dimensional bounds could be replaced by polylogarithmic factors. Ou r
method does not seem to yield new results here. (CS74)
(33) As I see it, Ennui is an especially good site for just this kind of analysis,
although m i ne is admittedly a minority view: even so historically and
theoretically astute a reader of Edgeworth as Seamus Deane holds to the
position that Edgeworth’s fiction is “not an analysis but a symptom of the
colonial problem the country represented.” (LIT43)
While writers may avoid threatening the positive face of disciplinary
colleagues, (34) and (35) seem to go against this grain. In (34), the singular
first person pronoun and verb argue could make for more disagreement.
despite the use of our to minimise the visibility and agency of the writer in
(35), the noun inability and adverb + adjective very alarming seem to be critical
of the writer.
(34) Though some readers and critics consider his aesthetic tenets to be ironic - a
hasty conclusion, I would argue - one of Wilde’s maxims from the preface to
The Picture of Dorian Gray (1890) explains the Dwarf ’s tragic mistake.27
(LIT79)
(35) The inability of o ur most ingenious designers to make even the simplest
systems completely foolproof is very alarming. (CS17)
A COmpARATIvE STudY Of jOuRnAL ARTICLES In EngLISH LITERATuRE And COmpuTER SCIEnCE
Ibérica 39 (2020): 215-242 237
4.2.7. Originator – linking authors and their claims
In the corpus, the Originator role came to the fore when both Computer
Science and Literature writers used self-markers to link themselves with their
claims, suggesting to the reader that both the claim and claim-maker were
worth noting (Harwood, 2005c).
The most unambiguous form of self-reference, the singular pronoun I, was
preferred by literary scholars when making claims and declaring their
interpretations. Almost half the total number of occurrences of I in the
Literature subcorpus were concentrated in this role, often accompanied by
the verb “argue”. This suggests that writers are expected to emerge in their
writing to put forth their claims as the credibility of the person behind the
claim is of utmost importance in absence of objective fact (Hyland, 2007).
In (36), the writer takes ownership of her readings through use of the
singular pronouns, but also limit her claims to one possible interpretation.
(36) … in m y readings of Moll Flanders and Roxana, I interpret the figure of
the woman as the embodiment of a “purer” or more purely imaginative,
version of capitalism... (LIT20)
The lower occurrence of Originator in Computer Science articles compared
to other roles could be due to the focus on methodology where the
Explainer role is more pertinent. In Computer Science RAs the role was
marked by the exclusive we rather than I as seen in (37). The writer hedges
the use of the exclusive we with the adverb “quite” and comparative adjective
“simpler”, rather than declaring his procedure is “better than” Collani and
Sheil’s (1989).
(37) Therefore, w e conclude that the above maximization procedure is not only
quite efficient, but also simpler to solve than Collani and Sheil’s procedure
(1989). (CS16)
Writers in both disciplines also used the less visible us and our in (38) and (39)
to downplay their claims about the wider implications or limit the
generalisability of their conclusions. However, these are balanced by “but
must always” in (38), and “important implication” in (329) to stress their
commitment to their claims.
(38) Disgrace urges u s to see that human rights can never be certain or absolute
outside of a purely fictional state, but must always proceed from a tragic
YIn LIng CHEung & LOuWEnA LAu
Ibérica 39 (2020): 215-242238
recognition both of the perplexities of the human condition and of the
insurmountably difficult, ethically precarious real-world choices that
inescapably must be made. (LIT59)
(39) An important implication of o ur results is that there’s a randomized
polynomial time approximation scheme for the permanent that works for
almost every bipartite graph. (CS19)
6. Conclusion
This study contributes to the scholarship of first person pronouns in
academic writing in two ways. firstly, our findings appear to challenge the
generally established view that a hard science features fewer instances of first
person pronouns compared to a soft discipline. The frequency of first
person pronouns in Computer Science RAs exceeds the number found in
the Literature subcorpus by 2.5 times. This finding seems to be a departure
from previous studies (e.g. Lafuente millán, 2010) and shows greater use of
first person pronouns in academic writing is not peculiar to the soft
disciplines. This reversal of first person pronoun frequencies associated with
the hard and soft paradigm calls into question the approach of studies that
cluster several disciplines together into ‘hard’ and ‘soft’ categories for
comparison and analyses of linguistic features, including first person
pronouns. Instead, it could be more fruitful to focus on comparisons of
single disciplines, one from each of the hard and soft sciences to identify
common and dissimilar features across the hard/soft divide.
Another contribution of this study is the addition of the new roles of
Researcher and Explainer to the Tang and john (1999) taxonomy, which
rendered it an efficient analysis framework for the requirements of the data.
The role of Researcher is germane across both hard and soft disciplines in
this study and may be used in other disciplinary corpus studies of first
person pronouns. Whether RAs contain a specific methodological section,
researchers are required to discuss and justify their approach, elaborate
arguments, cite references and draw from his or her own knowledge.
Similarly, the role of Explainer could also be pertinent in the analysis of
writing in other disciplines where experimental or methodological processes
are sequentially detailed in RAs.
Apart from contributing to theory and research, at the practice level, our
study may guide novices in academic writing for whom the lack of familiarity
A COmpARATIvE STudY Of jOuRnAL ARTICLES In EngLISH LITERATuRE And COmpuTER SCIEnCE
Ibérica 39 (2020): 215-242 239
with disciplinary conventions is a stumbling block. It is assumed that
newcomers will automatically develop target discourse practices. By
highlighting linguistic resources available in different disciplines and their
rhetorical effects on disciplinary knowledge, our study may help novice
researchers and those seeking publication in postgraduate programmes with
decisions about when and how to intrude into their texts in alignment with
disciplinary conventions.
EAp teachers and instructors could utilise the findings to raise awareness of
the continuum of authorial presence and authority to the attention of new
writers in the disciplines. This can assist them in thinking about how to
modulate the degree of intrusion into their text and its effect on readers
(Beerits, 2016). for students who are required to be familiar with academic
writing conventions across multiple disciplines could build an inventory of
first person pronoun practices and conventions in an informed manner. This
paper provides valuable reference material for authors penning guidebooks
for academic writing who can cite the findings to illustrate relevance of
personalised pronouns to authorial voice and point out that high levels of
writer-reader interactivity are not always the domain of softer sciences.
Because each individual discipline has its own particular purposes, practices
and norms, our findings may not be generalisable beyond the specific
disciplines of Literature and Computer Science. The taxonomy applied is
specific to the data examined here and may not be relevant with different sets
of data. Though individual preference could have accounted for
intradisciplinary variations found in the subcorpora, interviews with the RA
writers were beyond the scope of this study. Similarly, preferences for
pronoun forms in specific functional roles were not examined in detail here
and could be an area for future research.
Article history:
Received 15 November 2018
Received in revised form 11 June 2019
Accepted 26 April 2020
References
YIn LIng CHEung & LOuWEnA LAu
Ibérica 39 (2020): 215-242240
Beerits, L. (2016). “Understanding I: The rhetorical
variety of self-references in college literature
papers”. College Composition and Communication
67(4): 550-575.
Biglan, A. (1973). “The characteristics of subject
matter in different academic areas”. Journal of
Applied Psychology 57(3): 195-203.
Carciu, O. M. (2009). “An intercultural study of
first-person plural references in biomedical
writing”. Ibérica 18: 71-92.
A COmpARATIvE STudY Of jOuRnAL ARTICLES In EngLISH LITERATuRE And COmpuTER SCIEnCE
Ibérica 39 (2020): 215-242 241
Dontcheva-Navrátilová, O. (2013). “Authorial
presence in academic discourse: Functions of
author-reference pronouns”. Linguistica Pragensia
1: 9-30.
Fløttum, K., T. Kinn & T. Dahl (2006). “‘We now
report on...’ versus ‘Let us now see how...’: Author
roles and interaction with readers in research
articles” in K. Hyland & M. Bondi (eds.), Academic
Discourse across Disciplines, 203-224. Bern:
Peter Lang.
Harwood, N. (2005a). “‘Nowhere has anyone
attempted…in this article I aim to do just that’: A
corpus-based study of self-promotional I and we in
academic writing across four disciplines”. Journal
of Pragmatics 37: 1207-1231.
Harwood, N. (2005b). “‘We do not seem to have a
theory . . . The theory I present here attempts to fill
this gap’: Inclusive and exclusive pronouns in
academic writing”. Applied Linguistics 26(3): 343-
375.
Harwood, N. (2005c). “‘I hoped to counteract the
memory problem, but I made no impact
whatsoever’: Discussing methods in computing
science using I”. English for Specific Purposes 24:
243-267.
Hewings, A. (2012). “Stance and voice in
academic discourse across channels” in K.
Hyland, & C. Sancho Guinda (eds.), Stance and
Voice in Written Academic Genres, 187-201.
Basingstoke, UK: Palgrave Macmillan.
Hirvela, A. & D. Belcher (2001). “Coming back to
voice: The multiple voices and identities of mature
multilingual writers”. Journal of Second Language
Writing 10: 83-106.
Hyland, K. (1998). “Persuasion and context: The
pragmatics of academic metadiscourse”. Journal
of Pragmatics 30: 437-455.
Hyland, K. (2001). “Humble servants of the
discipline? Self-mention in research articles”.
English for Specific Purposes 20: 207-226.
Hyland, K. (2002). “Options of identity in academic
writing”. ELT Journal 56: 351-358.
Hyland, K. (2005). “Stance and engagement: A
model of interaction in academic discourse”.
Discourse Studies 7(2): 173-192.
Hyland, K. (2007). “Different strokes for different
folks: Disciplinary variation in academic writing” in
K. Fløttum (ed.), Language and Discipline
Perspectives on Academic Discourse, 89-108.
Newcastle upon Tyne, UK: Cambridge Scholars.
Hyland, K. (2012). “Undergraduate understandings:
Stance and voice in final year reports” in K. Hyland
& C. Sancho Guinda (eds.), 134-150.
Hyland, K. & F. Jiang (2016). “Change of attitude?
A diachronic study of stance”. Written
Communication 33(3): 215-274.
Ivanič, R. (1998). Writing and Identity: The
Discoursal Construction of Identity in Academic
Writing. Amsterdam: John Benjamins.
Kuo, C. H. (1999). “The use of personal pronouns:
Role relationships in scientific journal articles”.
English for Specific Purposes 18: 121-138.
Lafuente Millán, E. (2010). “‘Extending this claim,
we propose…’ The writer’s presence in research
articles from different disciplines”. Ibérica 20: 35-56.
Leedam, M. & M. Fernández-Parra (2017).
“Recounting and reflecting: The use of first person
pronouns in Chinese, Greek and British students’
assignments in engineering”. Journal of English for
Academic Purposes, 11: 345-356.
Martín, P.M. (2003). “Personal attribution in
English and Spanish scientific texts”. BELLS
(Barcelona English Language and Literature
Studies) 12. URL: http://www.raco.cat/index.
php/Bells/article/view/82918/108653 [12/06/2017]
McCarthy, W. (2015). “A matter of style”.
Interdisciplinary Science Reviews 40(2): 95-100.
McGrath, L. (2016). “Self-mentions in
anthropology and history research articles:
Variation between and within disciplines”. Journal
of English for Academic Purposes 21: 86-98.
Muñoz, M.C. (2013). “The ‘I’ in interaction:
Authorial presence in academic writing”. Revista
de Lingüística y Lenguas Aplicadas 8: 49-58.
Myers, G. (1989). “The pragmatics of politeness in
scientific articles”. Applied Linguistics 10(1): 1-34.
Pennycook, A. (1994). “The politics of pronouns”.
ELT Journal 48(2): 173-178.
Sheldon, E. (2009). “From one I to another:
Discursive construction of self- representation in
English and Castilian Spanish research articles”.
English for Specific Purposes 28: 251-265.
Starfield, S. & L.J. Ravelli (2006). “‘The writing of
this thesis was a process that I could not explore
with the positivistic detachment of the classical
sociologist’: Self and structure in New Humanities
research theses”. Journal of English for Academic
Purposes 5: 222-243.
Stevens, B. K. & L. L. Stewart (1996). A Guide to
Literary Criticism and Research, 3rd ed. Fort
Worth, TX: Harcourt Brace College Publishers.
Sword, H. (2012). Stylish Academic Writing.
Cambridge, MA: Harvard University Press.
Cheung Yin Ling earned her ph.d. in Linguistics from purdue university,
uSA. She is Associate professor at the national Institute of Education,
nanyang Technological university, Singapore. She specializes in academic
writing and writing instruction. She has published in journals such
as System, The Asia-Pacific Education Researcher, and RELC Journal.
Louwena Lau earned her m.A in Applied Linguistics from national
Institute of Education, nanyang Technological university, Singapore. She
specializes in English as a second language writing.
Appendices
YIn LIng CHEung & LOuWEnA LAu
Ibérica 39 (2020): 215-242242
Tang, R. & S. John (1999). “The ‘I’ in identity:
Exploring writer identity in student academic
writing through the first person pronoun”. English
for Specific Purposes 18: 23-39.
Wilder, L. (2005). “The rhetoric of literary criticism
revisited: Mistaken critics, complex contexts and
social justice”. Written Communication 22(1): 76-
119.
Wu, G. & Y. Zhu (2014). “Self-mention and
authorial identity construction in English and
Chinese research articles: A contrastive study”.
Linguistics and the Human Sciences 10(2): 133-
158.
Yakhontova, T. (2006). “Cultural and disciplinary
variation in academic discourse: The issue of
influencing factors”. Journal of English for
Academic Purposes 5: 153-167.
Zhang, W.Y., & Y.L. Cheung (2017).
“Understanding ENGAGEMENT resources in
constructing voice in research articles in the fields
of computer networks and communications and
second language writing”. The Asian ESP Journal
13(3): 72-99.
Zobel, J. (2014). Writing for Computer Science.
London: Springer.
AUTHORIAL VOICE IN ACADEMIC WRITING: A COMPARATIVE STUDY OF JOURNAL ARTICLES IN
ENGLISH LITERATURE AND COMPUTER SCIENCE
Ibérica 39 (2020): xx-xx
Cheung Yin Ling earned her Ph.D. in Linguistics from Purdue University, USA.
She is Associate Professor at the National Institute of Education, Nanyang
Technological University, Singapore. She specializes in academic writing and
writing instruction. She has published in journals such as System, The Asia-
Pacific Education Researcher, and RELC Journal.
Louwena Lau earned her M.A in Applied Linguistics from National Institute of
Education, Nanyang Technological University, Singapore. She specializes in
English as a second language writing.
Appendices
APPENDIX A: LIST OF LITERATURE AND COMPUTER SCIENCE JOURNALS
Literature Journals Computer Science Journals
1 College Literature ACM Transactions on Modeling and Computer Simulation
2 Contemporary Literature Applicable Algebra in Engineering, Communication and Computing
3 Modern Fiction Studies Artificial Intelligence
4 Modern Philology Computer Standards & Interfaces
5 New Literary History Computers & Chemical Engineering
6 PMLA Computers & Electrical Engineering
7 Nineteenth Century Literature Computers & Industrial Engineering
8 Studies in English Literature IEEE Transactions on Computers
9 Studies in Philology IEEE Transactions on Reliability
10 The Review of English Studies Journal of the ACM
11 Twentieth Century Literature The International Journal for Computation and Mathematics in
Electrical and Electronic Engineering