GENRE A NA LYS I S : A KEY TO A THEORY OF ESP?

[Md. Mostafa Rashel]

GENRE A NA LYS I S : A KEY TO A THEORY OF ESP?

A much more promising approach to a theory of ESP comes from the analysis of ESP texts . In the same way as the te aching procedu res of ESP are linked to a vi ew of language and learning, all ESP activity has to be linked to a view of text. ESP has in its brief h i s tory adopted various approaches to text analys i s , f rom the early regi s tear n a lysis assoc i a ted with the iden ti f i c a ti on of key gra m m a tical el em ents of s c i en ti f i c com mu n i c a ti on (Ba rber, 1 9 6 2 , Sw a l e s , 1971) to rh etorical analysis assoc i a ted wi t h Trim ble (1985) and Lack s trom , Sel i n ker, and Tri m ble (1972) thro u gh the functional/notional approach associated with the textbooks The Nucleus Series (Bates and Dudley-Evans, 1976) and The Focus Series (Allen and Widdowson, 1974) through to the dominant approach of today, genre analysis (Swales, 1990; Bhatia, 1993). The value of all these approaches to text analysis are that they begin from the idea that the texts used in particular specialist envi ron m en t s , wh et h er that be ac ademic writing, business or other professional activity or wherever the ESP need lies, have particular characters tics that distinguish them from other texts and from the generalized summaries of linguistic features that arise from an approach to text analysis that uses a corpus of differing texts.

The va rious approaches men ti on ed above all have their own va lue and have each moved the activity of finding particular characteristics of ESP texts a stage further. The work of the register analysts have shown that an ability to use certain key grammatical features is vital in ESP work and that other grammatical features of little relevance to ESP work can be ignored. The work of rhetorical analysis first introduced the idea that grammatical features found in specific contexts, such as an academic textbook, might follow rules that in certain su btle ways differ from the gen eral rules as set out in gen eral grammar books. This work also stresses the predom i n a n ce of rhetorical considerations in determining grammatical choice. The notional/functional approach
at its best still strikes me as a very fruitful method of bringing together lexical items that co - occur naturally, i.e. verbs su ch as co n s i s t of and co n t a i n, with passive constructions such as attached to, connected to, mounted on in describing the notion of S t r u c t u re, or items assoc i a ted with the noti on of Q u a n ti ty, su ch as a d e q u a t e , sufficient, enough, too much, excessive etc.

Tony Dudley-Evans, The University of Birmingham


Enduring..........

[Md. Mostafa Rashel]

The findings of gen re analys i s , h owever, bring toget h er the insights of these earl i er
approaches to text analysis, but also a greater sophistication in the examination of the
wri ters' purpo s e . The Moves and Steps that Swales (1990) su ggests for the arti cl e
i n trodu cti on marries the tex tual aw a reness of the regi s ter analysts with a mu ch
broader vi ew of h ow rh etorical con s i dera ti ons govern gra m m a tical ch oi ce . Th e
i n terest in disco u rse com mu n i ty and how the ex pect a ti ons and conven ti ons of
different discourse communities mould the texts that they use has led to this broader
vi ew and placed ESP re s e a rch in a po s i ti on wh ere it can make a meaningf u l
con tri buti on to discussion of h ow ideas are dissem i n a ted and facts cre a ted in
communities.

The early work by Swales foc u s ed on the re s e a rch arti cl e , and in particular the
i n trodu cti on secti on of the re s e a rch arti cle gen re (Sw a l e s , 1 9 9 0 )1. His Cre a ting a
Research Space model (the CARS model) is very well known, but for the sake of clarity
and comprehensiveness I shall outline it here. The model captures the ways in which
academic writers justify and highlight their own contribution to the ongoing research
profile of the field by first establishing a topic for the research and summarising the
key features of the previous research, then establishing a gap or possible extension of
that work that wi ll form the basis of the wri ters' cl a i m s . The model proposes three
main Moves for the introduction and a number of Steps used to express each move:

Move 1: Establishing a Territory

Step 1: Claiming Centrality
and/or
Step 2: Making Topic Generalisations
and/or
Step 3: Reviewing Items of Previous Research

Move 2: Establishing a Niche
Step 1A: Counter-claiming
or
Step 1B: Indicating a Gap
or
Step 1C: Question Raising
or
Step 1D: Continuing a Tradition

Move 3: Occupying the Niche
Step 1A: Outlining Purposes
or
Step 1B: Announcing Present Research
Step 2: Announcing Principal Findings
Step 3: Indicating Research Article Structure

This model (originally presented by Swales in a slightly different form in 1981) has
had a tremendous influence on genre analysis in ESP and on the teaching of academic
wri ti n g, both to intern a ti onal or L1 stu den t s , or to profe s s i onal wri ters wishing to
publish in international journals. A move based approach has also been used for the
a n a lysis of the re s e a rch arti cl e , e.g the abstract (Sa l a ger- Meyer, 1 9 9 0 ) , the met h od s
section (Wood, 1982), the results section (Brett, 1994, Williams, 1999) the discussion
section (Belanger, 1982, Dudley-Evans, 1994) and also for the analysis of dissertations
(Hopkins and Dudley-Evans, 1988).


Tony Dudley-Evans, The University of Birmingham


Enduring..........

[Md. Mostafa Rashel]

A fundamental assumption of all these move-based models is that they are common to
a ll ac ademic disciplines. Swales (1990) notes that different steps may be used in
d i f ferent disciplines, but also su ggests that many of these steps wi ll be wi dely
distributed across the disciplinary areas.

While there is little doubt that the CARS model is frequently found in more or less its
pure form in many disciplines,many researchers who have continued Swales' work on
Moves and Steps have begun to find intere s ting va ri a ti ons in the patterns found in
d i f ferent disciplines. Cl e a rly there wi ll alw ays be excepti ons to the gen eral pattern ;
writers will choose to omit a particular Move or vary the order of Moves or Steps to
suit their particular rh etorical purpo s e . But wh ere it can be establ i s h ed thro u gh
det a i l ed analysis that a particular discipline reg u l a rly and sys tem a ti c a lly uses a
variation on the general model, then this is an interesting finding. An example of this
was Croo kes' finding (Croo ke s , 1986)that the lon ger introdu cti ons found in Soc i a l
S c i en ce arti cles were the re sult of the wri ters' using cycles of Moves so that the
i n trodu cti on might contain more than one Move 2 and Move 3 and that each new
Move 2 would be followed by another Move 3. Another more detailed presentation of
a key va ri a ti on comes in An t h ony's work on introdu cti ons in sof t w a re en gi n eeri n g
(Anthony, 1999). He found that writers seemed to feel a need to justify their research
and that he therefore needed to add one Step Eva l u a tion of Re se a rch to Move 3 to
capture what was happening in the articles introductions he examined. Although this
may seem a relatively small addition, it is clearly more than just an exception to the
rule used by certain writers; it is a distinctive systemic feature of writing in the field of
s of t w a re en gi n eeri n g. The need to use the Step seems to arise from the fact that
s of t w a re en gi n eering is a rel a tively new field and its journals may be re ad by many
en gi n eers from other bra n ches who are not nece s s a ri ly up to date in a ra p i dly
developing field. The same situation seems to explain another major feature of article
i n trodu cti ons in the field that An t h ony (1999: 42) note s : the introdu cti ons are
relatively long and that this is because writers make extensive use of Move 1 - Move 2
c ycles in wh i ch det a i l ed back ground inform a ti on and def i n i ti ons are pre s en ted . As
Anthony (1999: 43) explains, the writers see their task 'as a kind of preaching to the
cannibals'. In this case the cannibals are 'engineers from a wide range of disciplines'
who 'subscribe to the journal in order to acquire results which can be used to solve
their own particular problems (Anthony, 1999: 43).


Tony Dudley-Evans, The University of Birmingham


Enduring..........

[Md. Mostafa Rashel]

It may well be that the need for background information and for justification of the
research carried out is not exclusive to software engineering. It is not too surprising,
but intere s ting non et h el e s s , that in the rel a ted field of com p uter en gi n eeri n g
Posteguillo (1999) has found a similar phenomenon. And, although she did not say so
directly, Cooper's finding that introductions in electronic engineering did not match
Swales' model2 seems to reflect the same phenomenon (Cooper, 1985).

The direction that genre analysis has taken since the early Move and Step analysis is
i n tere s ti n g. Th ere have been two main ten den c i e s ; one under the influ en ce of
sociology of science (e.g. Mulkay, 1985 and 1991, Knorr-Cetina, 1981) has been the
more detailed analysis o f the concept of discourse community (Miller, 1994) and of
actual disco u rse com mu n i ties in practi ce (Sw a l e s , 1 9 9 8 ) , the other has been the
detailed analysis of specific features of language as used in particular genres, such as
h ed ging (Hyl a n d , 1 9 9 8 ) , reporting verbs (Th om p s on and Ye , 1 9 9 1 ; Th omas and
Hawes, 1994) or verbs with inanimate subjects (Master, 1991). Research in the second
of these two ten dencies has incre a s i n gly thrown up intere s ting differen ces bet ween
disciplines and I now wish to report on some of these.

Th ere is intere s ting va ri a ti on in the use of h ed ging bet ween disciplines. Th e
d i f feren ce in this re s pect bet ween ac ademic arti cles publ i s h ed in pre s ti gious and
vers i ons of those papers publ i s h ed in 'popular' journals su ch as S ci en tific Am eri c a n
or New Sci en ti s t is well establ i s h ed . Both Fa h n e s tock (1986) and Myers (1990) have
s h own that wri ters in ac ademic journals wi ll make guarded claims abo ut thei r
f i n d i n gs using many hed ged statem ents (the re sults su ggest that ..., the findings
a ppear to su ppo rt the claim that ...) but wi ll use more con f i dent statem ents in
popular journals tending to pre s ent claims as establ i s h ed fact s . The same
ph en om en on has been ob s erved in arti cles and tex tbooks with wri ters of
tex tbooks pre s en ting theories and experimental findings as established knowl edge and devel oping a con s en sus vi ew of the state - of - p l ay in the discipline in wh i ch
these theories and re sults are seen as uncon troversial (Myers , 1 9 9 2 ) .


Tony Dudley-Evans, The University of Birmingham


Enduring..........

[Md. Mostafa Rashel]

What is interesting is that significant differences in the use of epistemic modality3 are
now being observed between disciplines. Rizomilioti (work in progress) has examined
the downtoners used in journal articles in three disciplines: Biology, Archaeology and
L i tera ry Cri ti c i s m . In each discipline the corpus she co ll ected con s i s ted of
approximately 200,000 words.She found that Archaeology had the highest proportion
with a total number of 2569 downtoners used with an ratio of 12.845 per 1,000 words,
Biology had 1521 occurrences of downtoners making a ratio of 7.605 per 1,000 words
and Literary Criticism had the lowest proportion with 1174 making a ratio of 5.87 per
1,000 words. Rizomilioti has also looked in detail as the use of boosters in the same
articles. Boosters are linguistic devices that w riters use to affirm their confidence in a
claim that they are putting forward; examples include definitely, the modal verb will
and reporting verbs su ch as s h ow and prove. Ri zom i l i o ti reports that the high e s t
occurrence (389 instances) was in the Literary Criticism corpus; this compared with
169 occ u rren ces in the Arch aeo l ogy corpus and on ly 92 occ u rren ces in the Bi o l ogy
corp u s . She argues that the rel a tively high use of boo s ters by wri ters of L i tera ry
Cri ticism arti cles re sults from their wish to 'seclu de certain claims from direct
contradiction' (Rizomilioti,work in progress, p. 24).Literary criticism is a discipline in
wh i ch wri ters make asserti ons abo ut litera ry texts based on their assu m ed spec i a l
insight into their texts rather than on experimental evidence as in scientific disciplines.
Their writing is therefore marked by the use of adverbials seeking affirmation, such as
surely, undoubtedly and definitely and a low use of downtoners.Writers of Archaeology
and Bi o l ogy arti cl e s , by con tra s t , persu ade re aders of the va l i d i ty of t h eir cl a i m s
through the evidence provided in the Results section and therefore have less need to
use these boosters.


Tony Dudley-Evans, The University of Birmingham


Enduring..........

[Md. Mostafa Rashel]

We have noted differences in the use of epistemic modality between genres (academic
a rti cle v. popular scien ce arti cl e , and ac ademic arti cle v. tex tbook) and bet ween
different disciplines. Bloor and Bloor (1993) have taken this kind of analysis a stage
further in that they noted a very interesting difference in the use of hedging devices
within one genre and one discipline. They found that in economic articles published in that discipline's most general academic journal The Economic Journal writers were
either making claims that were related to the real world (field central claims) or claims
that were related to issues within the discipline itself (substantive claims). Bloor and
Bloor noted that the substantive claims tended to be hedged while field central claims
were not.

The variation in hedging between disciplines (and within disciplines) is probably the
most devel oped area of re s e a rch . But other intere s ting work is em er gi n g. Hyl a n d
(1999) has shown how citations practices vary across disciplines and suggests that this
results from differences in epistemological and social conventions of the disciplines.
Hyland analysed citations in 80 articles taken from the fields of Sociology,Marketing,
P h i l o s ophy, App l i ed Linguisti c s , Bi o l ogy, P hys i c s , Mechanical Engi n eering and
E l ectronic Engi n eeri n g. In gen era l , he found that the social scien ce and hu m a n i ti e s
writers used more citation than the science and engineering writers.With the regard to
the use of integral v. non-integral citation structures (an integral structure will name
the cited author in the actual sentence), the overall preference in all the disciplines,
except Philosophy, was for non - i n tegral stru ctu re s . Th ere was, h owever, a mu ch
greater use of integral structures in the four Humanities and Social Science disciplines
com p a red with the Scien ce and Engi n eering disciplines. Th ere were also major
differences in the reporting verbs used by the different disciplines with the science and
engineering writers favouring more neutral verbs such as report, use, develop and the
humanities and social science writers favouring more tentative verbs such as suggest,
claim, argue.


Tony Dudley-Evans, The University of Birmingham


Enduring..........

[Md. Mostafa Rashel]

Discussion and Conclusion
The initial work on Move and Step analysis (Sw a l e s , 1981 and 1990; Bh a ti a , 1 9 9 3 )
suggested that the models proposed were generalised models that applied to academic
a rti cles wri t ten in all ac ademic fiel d s . Cl e a rly it was ack n owl ed ged that there were
s ome differen ces bet ween disciplines, but it was argued that the models propo s ed ,
su ch as the CARS model for the arti cle introdu cti on , a re pro to types and actu a l
examples will vary in the degree to which they conform with this prototype.
What is becoming cl e a r, h owever, is that disciplinary va ri a ti on is mu ch more
significant than allowed for in the original work on genre analysis.We need to devise a
theory that goes beyond the ideas of prototypicality to acknowledge that variation in
the disco u rse stru ctu ring of gen res ref l ecting different ep i s tem o l ogical and soc i a l practices in disciplines is a key factor in genre theory. The danger is that the theory will
become immensely complicated with a proliferation of genres for each discipline. It
could be, however, that the simplicity of Martin's model (Martin, 1989) which places
register between genre and language in the hierarchy, as in the diagram below, will
capture variation without over-complicating the theory:

(PLEASE NOTE THAT WE NEED TO ADD ARROWS

FROM GENRE TO REGISTER TO LANGUAGE)

Genre – Register – Language

This sys tem all ows us to account for differen ces in discipline and bet ween form a l
academic papers and those in popular journals through the use of Field and Tenor, two
of the three components of register in the Hallidayan system (Halliday 1985).

I am thus arguing for a theory of ESP based on text, but one that starts from the point
of view that texts in different disciplines will have different patterns of organisation
rather than variations on one 'common-core' pattern. Common-core patterns are, I
am suggesting, a convenient starting point for pedagogical purposes, but may not have
much basis in actual genre analysis.

The teaching of academic writing will clearly need to reflect this variation. It is clearly
possible to use the generalised CARS model as the starting point for the teaching of
ac ademic wri ti n g, p a rti c u l a rly if one is te aching heterogen eous groups of s tu den t s
f rom different disciplines, but with hom ogen eous groups it may be mu ch more
ef f i c i ent to focus on the specific fe a tu res of the actual gen res that stu dents actu a lly
have to read or write. This is especially the case where students are in an EFL situation
studying their subject course in their first language. Such students will not have the
high proficiency levels in English that most students have in a first or second language
s i tu a ti on and wi ll need , in my op i n i on , a mu ch more stra i gh tforw a rdly linguisti c
a pproach based on the actual texts they use. Th ey do not have the linguisti c
sophistication to deal with issues about the readership and the discourse community
in any depth. They need to see how the Moves and Steps work in the genres they use
and how they are expressed in English.

Tony Dudley-Evans, The University of Birmingham


Enduring..........