Content uploaded by Sylviane Granger
Author content
All content in this area was uploaded by Sylviane Granger on Dec 22, 2017
Content may be subject to copyright.
World
Engli.dirr.
VoI.
IS.
No.
I,
pp.
17-21.
1996.
Connector usage in the English essay writing
of
native and
non-native EFL speakers
of
English
SYLVIANE
GRANGER*
and STEPHANIE TYSON*
ABSTRACT:
In
this study we focus
on
cohesion in discourse, and more specifically
on
connector usage.
Acknowledging the importance
of
combining a top-down and a bottom-up approach in the study of dis-
course, we adopt a bottom-up approach which is favoured by
our
methodology.
In
the first section we evalu-
ate previous studies
of
learner connector usage and the literature
on
contrastive French-English connector
usage. We hypothesize that we will discover a general overuse of connectors
by
learners and use the ICLE
corpus of learner English
to
test the hypothesis.
Our
study reveals
no
overall overuse
of
connectors by
learners and thus contradicts the initial hypothesis.
A
more qualitative
look
shows strong evidence of over-
use and underuse
of
individual connectors, as well as semantic, stylistic and syntactic misuse. We conclude
that learners should not be presented with lists
of
‘interchangeable’ connectors but instead taught the
semantic, stylistic and syntactic behaviour of individual connectors, using authentic texts.
COHESION AND DISCOURSE STUDIES
Over the last twenty years, interest in written English discourse
-
both the native speaker
and learner varieties
-
has grown dramatically. In particular, widespread studies of
cohesion and coherence have been carried out, sparked
off
by the publication in 1976 of
Halliday and Hasan’s
Cohesion in English.
The aim of many of these studies, particularly
those conducted in the USA in the early 1980s, was to discover whether there was a sig-
nificant correlation between use of cohesive ties and coherence and whether these could then
be linked to writing
proficiency/development.
On the whole, no such correlation has been
found.
For
Tierney and Mosenthal(l983:
225)
for example, cohesion was pervasive in all
the texts they studied but
it
was ‘causally unrelated to coherence.’ Neuner (1 987) finds no
statistical difference in the number of individual cohesive ties used in good and poor native
American English essays, although he identified a difference in the use
of
cohesive chains.
Witte and Faigley (1 98
1:
200)
state that ‘cohesion and coherence interact to a great degree,
but a cohesive text may be only minimally coherent.’
Although quantitative studies of cohesive ties can be instructive (Pritchard (1 98
1)
for
example, found a higher incidence of cohesive ties in ‘problem passages’ of student essays),
this is a completely different exercise to evaluating how these ties are used. As Hartnett
(1986: 143) concludes ‘Using cohesive ties successfully is apparently not easy. Both good
and poor writers may use the same kinds
of
cohesive ties, but they use them differently.’
When studying learner language in particular, it is necessary to combine a quantitative and a
qualitative approach, comparing frequency and semantic/syntactic use.
Another problem with many of these studies is that they are very small-scale. Connor’s
1984 study for example, in which cohesion density is found not to be
a
discriminating factor
between native and ESL students, is based on
six
essays.
Of
course, when one considers the
amount of time needed to identify cohesive tiedcoherence relations manually, the limita-
tion is understandable. However, size of sample
is
not the only problem: in fact, as Khalil
*Centre for English Corpus Linguistics, College Erasme, Place Blaise Pascal
1,
UniversitC Catholique de
Louvain,
1348
Louvain-la-Neuve, Belgium.
0
Blackwell Puhlishers
Lrd.
IYY6.
108
Cowley
Road, Oxford
OX4
IJF.
UK
and
23H
Main
Street,
Suire
S(11.
MA
02
142.
USA.
18
Sylviane Granger and Stephanie Tyson
(1
989)
points out, very few empirically-based studies exist at all. There is a pressing need
for large-scale studies in order to obtain a more accurate description
of
cohesion/coher-
ence problems in EFL/ESL student writing.
THE INTERNATIONAL CORPUS
OF
LEARNER ENGLISH
This need for empirical data formed the rationale behind the International Corpus
of
Learner English (ICLE) project (see Granger,
1993).
The aim
of
the project, launched in
1990,
was togather and computerize a large corpus
of
advanced EFL learner writing, with a
view to investigating the interlanguage
of
advanced learners from various mother-tongue
backgrounds in the light
of
the major advances which have been made in applied linguistics
and computer technology. Nearly five years later, ten language backgrounds are represented
in the corpus (French, Dutch, German, Spanish, Swedish, Finnish, Czech, Polish, Japanese
and Chinese) and when fully complete, the corpus will contain about two million words.
We have called the area
of
research we are involved in ‘contrastive interlanguage analysis’
(or CIA) which involves comparing and contrasting what non-native and native speakers
of
a
language do in a comparable situation. In concrete terms, the different non-native English
varieties are compared with native speaker English and with each other. The results
of
this
interlanguage analysis are then examined in the light
of
classic contrastive analysis
of
the
native languages (for a full discussion
of
the CA-CIA methodology, see Granger, forthcom-
ing). The aim
of
this methodology is to identify and distinguish between L1-related and uni-
versal features
of
learner language and thus
to
be able to draw a clearer picture both
of
advanced interlanguage and
of
the role of transfer
for
the different mother-tongue back-
grounds.
If this goal is
to
be achieved,
it
is essential that the data under investigation be compar-
able: one of the most effective arguments against the results
of
much of the contrastive/
error analysis conducted in the past has been that the data are rendered meaningless by
fundamental differences in both data and research methods (for example, studies have used
subjects with different levels of language ability, performing different tasks, etc.). In order to
achieve this comparability, four variables have been controlled in the ICLE corpus:
Type
of
learner
-
EFL
tzot
ESL.
As
Nickel
(1
989:
298)
observes
in
an
interesting article, the
contradictory
results
obtained
in
applied linguistics, particularly as regards
the
question
of
transfer, are
in
part attributable
to
the lack
of
a clear
distinction
between EFL and ESL.
(ii)
Stage
of
learner
-
advanced.
The extent
of
L1
influence
is often
posited
as
being
in
direct
proportion
to
the learner’s
stage
of
progress,
and
we must therefore
be
sure
that
we are
com-
paring learners at a
similar
stage
of
progress
if
we
are
to
be
able
to
draw
any
conclusions
about
‘advanced
or
‘intermediate’ learners. For the
purposes
of
our
project, ‘advanced’ refers
to
university
students
of
English, usually
in
their
third
or
fourth
year
of
study,
who
therefore
make
relatively
few
morphosyntactic
errors
but
for whom a
significant
number
of
discourse-
level problems remain.
Text type
-
essay writing.
Many
features
of
language
are
extremely genre-sensitive,
so
the
type
of
task
set
will
significantly alter the results obtained. Therefore,
if
meaningful statements
are
to
be
made
about
differences
in
usage,
the
types
of
discourse
under
study
must
be
compar-
able. For example,
connectors,
the topic
of
this paper, are likely
to
feature
much
more
prom-
inently
in
argumentative
essay
writing
than in
a description
of
a
holiday
outing
or
of
how
an
engine
works. The other
obvious
attraction
of
studying
essay
writing
is
that
it
provides
complete texts
which
are particularly
well
suited
to
the
study
of
cohesion,
coherence,
and
other
textual
problems
which
remain
prevalent
at
an
advanced level.
(i)
(iii)
0
Blackwell
Pubhherr
Ltd.
IYYh
Connector usage in EFL essay writing
19
(iv)
A
native speaker
corpus
of
similar writing.
Just as
it
is essential
for
the learner writing to be
comparable in terms
of
genre,
so
it is essential for there to be a control native speaker corpus
composed
of
exactly the same type
of
writing. Comparing learner argumentative essay writ-
ing with native speaker letter writing would tell
us
nothing conclusive about differences in
language usage.
For
this reason, a native speaker corpus
of
expository/argumentative
writing
is being assembled in Louvain (the Louvain Corpus
of
Native Essay Writing [LOCNESS]).
AN
OVERUSE
HYPOTHESIS
The main body of the current study into connector usage compares a sample (89,918
words) of the French mother-tongue sub-component
of
the ICLE corpus with a sample
(77,723 words)
of
writing from the control corpus
of
English essay writing. Our hypothesis
was that the French learners would overuse connectors in their essay writing. There were
several reasons for formulating this hypothesis, not least the fact that we had been struck
when correcting essays by seemingly constant overuse and misuse of connectors. The few
empirically-based studies that have been conducted in this area also suggested overuse.
Milton and Tsang (1 993: 239), for example, in their corpus-based study
of
Hong Kong
students’ use
of
connectors, conclude ‘there is a high ratio
of
overuse of the entire range
of
logical connectors in our students’ writing, in comparison to published English.’ Field and
Yip
(1
992) also find that Cantonese mother-tongue learners use far more linking devices
than their English-speaking counterparts. Evensen and Rygh
(1
988) compare the writing
of
a
class
of
students in their mother tongue (Norwegian) with that in English. They also
find that the students use connectors slightly more in EFL discourse. The overuse hypo-
thesis is also strongly supported by the contrastive literature on French/English con-
nector usage.
In
Vinay and Darbelnet
(1
977: 222), for example, we read that ‘Le franqais
. . .
affectionne les articulations
et
se passe difficilement des precisions qu’elles peuvent
apporter dans
le
deroulement de la pensee. L‘anglais, au contraire, fait beaucoup moins
appel aux articulations explicites, donc laisse au lecteur le soin de supplier lui-meme
les articulations.’ In translation therefore, they conclude, ‘traduire du franqais articule en
anglais, c’est se resigner
a
laisser les charnieres implicites dans une large mesure’ (Vinay and
Darbelnet, 1977: 223). Similarly, Newmark (1 988: 59) states that in translation into
English, connectives are ‘often rightly and deliberately omitted
. .
.
their purpose is partly
phatic, i.e., they are used partly to maintain the reader’s interest.’ Hervey and Higgins
(1992:
49)
in their book
Thinking Translation
also state clearly it
is
more common in
French than in English for texts to be explicitly structured by the use
of
connectives (‘or,’
‘donc,’ ‘ainsi,’ ‘en effet’ and
so
on) and that a text translated from French into English where
all the French connectives are rendered
by
an English ‘equivalent’ will appear ‘stilted,
pedantic
or
patronizing,’ words which EFL teachers are frequently heard using about their
students’ writing.
If transfer plays a role in foreign language production, and it is nowadays almost uni-
versally recognized that it does, then this leads us to suggest that we will find a general
overuse of connectors in the French students’ writing.
METHODOLOGY
The first task was
to
select the connectors for study. Although all the studies above point
to connector overuse, few
of
them deal with the same list. It is, however, important to define
0
Blackwell
Publishers
Ltd.
I996
20
Sylviane Granger and Stephanie Tyson
clearly the object of study
if
results are to be meaningful and open
to
comparison with other
research. The choice of connectors for this study was based on the list
of
connectors in
Quirk
et
al.’s
(1985)
Comprehensive Grammar
of
the
English
Language.
However, we
excluded temporal connectors, which were felt
to
be external to the argumentation, and
included those connectors which are referred to in the grammar as ‘certain attitudinal dis-
juncts which have clear cohesive links’ and some ‘emphasizers’ which seem to add a new
point that strengthens the argument
or,
in
the case of ‘in fact’ for example, give a new turn to
the argument:
It
is a widely held belief in the
UK
that the signing
of
the Maastricht Treaty marked the end
of
an
era. In fact, nothing has changed.
These connectors are here termed corroborative connectors, following Ball’s term (Ball,
1986). The final list contained 108 connectors.
The next stage was to extract from the computerized ICLE corpus every instance of each
of these connectors by applying TACT concordancing software. This software provides
raw frequencies
of
particular words and strings
of
words and displays these words and
phrases within five lines
of
context. The contextualization was essential as the
first
major
manual task of the study was that
of
disambiguation, as many connectors have two
or
more
different uses (for example
‘so’
can function both as adverb and as connector). Once the
corpora are tagged, much
of
this laborious manual stage will be done automatically (for a
comparison of analysing raw and tagged corpora, see Tyson, 1995).
When this work was completed, overall frequencies were calculated (raw frequencies
and per
100,000
words). Frequencies for individual connectors were then examined and
finally, cases
of
stylistic, semantic and syntactic misuse were analyzed.
RESULTS AND DISCUSSION
In
the discussion which follows, the terms
NNS
and
NS
are used to refer
to
the non-native
speaker and native speaker groups respectively. The overall
figures
demonstrated
incontestably that the overuse hypothesis was invalid (see Tables 1A and 1B). However, a
more interesting pattern of overuse and underuse began
to
emerge when we looked at the
use of individual connectors (Tables 2A and 2B).
The figures in Table 2A reveal that the learners seem to overuse connectors which
perform particular functions: corroborating the argument (‘indeed,’ ‘of course,’ ‘in fact’),
giving examples (‘e.g.,’ ‘for instance,’ ‘namely’), and adding points to the argument (‘more-
over’). Overuse of additive and appositive connectors has also been noted in the writing of
Hong Kong students by Field and Yip (1 992).
Table 2B shows that there is underuse
of
connectors which contrast (‘however,’ ‘though,’
‘yet’) and develop the argument (‘therefore,’ ‘thus,’ ‘then’). In other words, the learners use
most frequently those connectors which add to, exemplify,
or
emphasize a point, rather
than those which change the direction
of
the argument
or
take the argument logically
forward. This would seem to indicate that one might find
a
different type of argumentation
in the two sets of essays,
a
hypothesis that
is
beyond the scope of this paper but which could
give rise to a contrastive study
of
FrencNEnglish argumentative discourse. The cases of
underuse were unexpected. In fact, very little has been written about connector underuse,
undoubtedly because it is much more difficult to spot. This is the advantage of large-scale
empirical studies, which draw attention
to
differences that otherwise go unnoticed.
0
Blackwell
Publishers
Lld.
IYYO
Connector usage
in EFL
essay writing
21
Table
1A.
Overall figures
for
connectors
NNS NS
Total number
of
connectors
976 916
Overall typehoken ratio
15.25 14.77
Total number
of
words
in
corpus
89,918 77,713
Nore:
In
Tables
1-6,
NNS
=
non-native speakers;
NS
-
native speakers
Table
1B.
Figures for connectors based
on
100,000
words per variety
NNS NS
Total number
of
connectors
1085 1178
Table
2A.
Non-native speaker overuse of
connectors: raw frequencies
Connector NNS
NS
Actually
Indeed
Of
course
Moreover
e.g.
For
instance
Namely
On the contrary
16
65
53
52
19
54
20
34
2
16
20
1
3
1
3
3
Table
2B.
Non-native speaker underuse of
connectors: raw frequencies
Connector NNS NS
However
Instead
Though
Yet
Hence
Then
Therefore
Thus
47
0
2
17
3
27
33
36
197
13
16
46
12
65
123
56
Q
Blackwell
Publishers
Ltd.
1996
22
Sylviane Granger and Stephanie Tyson
The cases of overuse seemed
to
provide evidence of mother-tongue influence. Perhaps
the most striking example is overuse and misuse of connectors in the corroborative
category. The learners seem not to recognize that connectors such as ‘in fact’ and ‘indeed’
lead the reader
to
expect some new information, but instead they equate them with French
connectors such as ‘en fait,’ which say to the reader ‘Here I am in the text’ and which are
often used as stylistic enhancers. However, connectors in English do not serve this function.
The examples below show this type of misuse, which has the effect
of
mystifying the reader,
who after reading ‘as a matter of fact’ is led to expect some corroboration of the previous
statement. However,
all
that follows in the first example is reformulation. In the second
example, there
is
no obvious logical connection between the two sentences:
[
11
Military service forces the young man to postpone the beginning
of
his
professional career but
also
of
his starting a family.
As
a
matter offact,
it seems impossible for him to get married and
to
have children if he has to stay away from home for one year.
121
the gulf war has established a special relationship between the two countries to the detriment
of
Washington’s relations with the Continent.
Indeed,
the gulf war has shown a flagrant lack
of
preparation and a profound disagreement among the allied armies.
In order to test whether transfer lay at the root
of
this type of overuse,
we
compared the
French learner figures with figures from a comparable corpus of German learner writing.
The results are shown in Table
3.
It is reasonable to conclude from these results that
German learners also overuse corroborative connectors to some extent. However, it is
likely that the significant overuse
of
‘indeed’ is transfer-related, especially since it is often
viewed by French speakers as the translational equivalent
of
‘en effet,’ a very common con-
nector in written French.
Table
3.
Comparison of French and German
NNS
use
of
corroborative connectors (per
100,000
words)
Connector
F”S
GNNS
Actually
18
20
Indeed
70
15.5
Of course
52
51
Another much overused connector is ‘moreover,’ which was misused almost
50
percent
of
the time. Typically, the French learners used it to reformulate, as
in
the example below, or
add a point, rather than to add
a
final powerful argument to convince the reader of a
particular point:
131
When he (the prisoner) will be released, his situation will
be
very painful because he will have,
alone, to re-adapt and re-integrate himself
in
a society he has been excluded from.
Moreover
he will have the greatest difficulties
of
integration because
of
his past as a prisoner.
Overuse of ‘on the contrary’ was also primarily due to misuse, probably due to a confu-
sion with the French ‘au contraire,’ which can be used to express
both
a
concessive and anti-
0
Blackwell
Publishers
Lld.
IVY6
Connector usage in EFL essay writing
23
thetic link. The learners seem not to recognize the extra specificity of the antithetic link.
Thus:
141
This kind
of
union will be economic. Therefore,
I
think nobody will have to fear for his cultural
identity.
On
the
contrary
if Europe achieves
a
political union one day, the European citizen
will have to destroy what made him belong to his previous nation.
(Possible alternatives here would be ‘on the other hand’ or ‘however.’)
The transfer explanation is upheld if we look at figures from the German corpus: 6 per
100,000
words, as opposed to the French
NNS
28 per
100,000
words.
Apart from a lack of understanding of the semantic properties of particular connectors,
learners also seem unaware of their stylistic restrictions. The style-sensitivity of connectors
is discussed by Biber (1988), who claims that the frequency of connectors ranges widely
according to the task, and by Altenberg (1986) in a study of the style-sensitivity of con-
nectors in speech and writing. In a study of Hong Kong connector usage, Field and Yip
(1
992: 26) find that learner writers ‘give confusing signals of register,’ using very formal
devices such as ‘moreover’ alongside devices preferred in speech such as ‘anyway’ and
‘actually.’ Colloquially marked uses of connectors such as ‘anyway’ and
‘so’
were found to
be common in the French learner writing, as in the examples below:
151
Moreover, the United States
of
America and England
do
share the same national language
but
do not form a nation.
Anyway,
the point
is
that language creates cultural clusters
161
I.
.
.I
tunnel under the English channel.
So,
now the cutdown has been
171
I.
.
.I
to a national feeling.
So,
we
see
that the evolution
Such inappropriate usage
is
most likely attributable to the fact that throughout school
and university, relatively little emphasis is put on matters of style. Distinctions between
spoken and written style are likely to be touched on but not focused on until the end of a
university course. Another significant reason is the lack
of
detailed description of con-
nector usage in the dictionaries. Martel
(1
991:
1
59)
for example speaks of ‘le peu de succes
avec lequel les dictionnaires bilingues et unilingues reussissent
a
decrire individuellement
des connecteurs d’usage pourtant frequent.’
There was further evidence
of
the learners’ lack
of
style-sensitivity when the figures were
compared with those for connector usage in the LOB subcorpus of informative writing
(Table
4).
The frequencies for the learner writers were strikingly close to the LOB figures,
the reason being that the LOB figures comprise different types of general writing and are
not specific to one genre.
Table
4.
Frequency of
therefore, thus
and
however
per
100,000 words in the Louvain
NS
and
NNS
corpora and the
LOB corpus
Louvain
NS
LOCNESS Louvain
NNS
Connector Corpus
Corpus
LOB
corpus
Therefore
136.7 36.7
Thus
62.2
40
However
253.4 52.2
47
35
72
0
Blackwell
Publisher\
Ltd.
1YY6
24
Sylviane Granger and Stephanie Tyson
The differences
in
the results of the
two
native speaker corpora underline the importance
of having corpora that are comparable in text type. The current study compares
two
corpora written by university level arts students on similar topics. If we compared the
learners’ results only with more general corpora, such as the
LOB
or the Brown corpora,
the results would undoubtedly have been quite different, as we have demonstrated in the
example above.
As
well
as cases of semantic and stylistic misuse, there is also clear evidence of different
syntactic positioning of connectors, with consistently sigmficant overuse of sentence initial
connectors. Table
5
gives the overall figures and the percentage of total connectors which
are sentence initial for each group.
Table
5.
Sentence-initial (SI) use of connectors
NNS
NS
NNS
(percentage)
NS
(percentage)
Total
SI
472 48.4 262
28.6
SI
typehoken ratio 8.9 6.1
Number
of
SI
types
53
43
It
is likely that this tendency for learners to place connectors in initial position is not
language-specific. For example, Field and Yip
(1992)
find that sentence-initial position, or
ISP as they call it, is the most common position for all
L2
writers and that
‘L1
writers used
the
NIP
(non-initial position) significantly more than
L2
writers’
(1992: 22).
They also find
that the impression of overall overuse of connectors is increased by
L2
preference for
placing connectors in sentence and paragraph initial position: ‘The impression of too many
devices in the
L2
scripts may be compounded by a strong use of the initial sentence and
paragraph position’
(1 992: 25).
They continue ‘it is possible that the NIP position may have
a positive effect beyond variation in that it points the reader more firmly to content than the
ISP position’
(1 992: 26).
Use of the ISP therefore points the reader towards the organiza-
tion of the text, towards the role of the writer, just as the overuse
of
phatic connectors such
as ‘actually’ and ‘indeed’ was found to do.
A
small study we conducted into sentence-initial use of ‘however,’ ‘indeed’ and
‘so’
seemed to corroborate this to some extent. Taking data from
two
other sub-corpora, the
Dutch and Chinese, we compared their figures with the figures for the French learner and
English native speaker corpora (Table
6).
The results show that
in
most cases, there is a
higher use of sentence-initial position by learners, and this is especially striking
in
the case
of
‘so.’
Dutch learner use of ‘however’ goes against this tendency, though.
In
any case, more
work on large samples would need to be conducted before reaching any firm conclusions.
PEDAGOGICAL IMF’LICATIONS
Even at a reasonably advanced level, connectors are difficult to master. We have seen
that French learner connector usage differs widely from that of their
NS
counterparts:
this
is
due to an inability to differentiate stylistically, insufficient knowledge of semantic
Q
BlackweU
Publishers
Lld.
1996
Connector
usage
in
EFL
essay
writing
25
Table 6. Sentence initial
(SI)
however,
indeed
and
so
by native speakers
of
English, French, Dutch and Chinese, based on
100,000
words per variety,
expressed in raw figures and as percentages of total use
of
these connectors
Connector
SUNS
SI/F
NNS
SI/D
NNS
SIKH
NNS
Indeed
9 39
0
6.5
so
14.1 53.4 17.7 73.3
However
126.1 35.6 39 128.9
(43.6%) (53.9%)
-
(5 8.6%)
(1
4.2%) (6 8.6%) (4
1
.6%) (59.8%)
(4 9.7%)
(68%)
(25.6%)
(8
7.4%)
restrictions placed on individual connectors, and inexperience in manipulating connectors
within the sentence structure.
There are various teaching methods which
it
would seem sensible
to
adopt
if
these prob-
lems are
to
be addressed. First,
it
is essential
to
teach students that connectors in English
should not be used as ‘stylistic enhancers’but should be thought of as higher-level discourse
units.
It
is necessary
to
place more emphasis on
how
to
use connectors, laying stress on
examining their use in authentic texts. The distinctions between connectors, whether
syntactic or semantic, are often slight and difficult to grasp and therefore, as Zamel(l984)
says, students must learn to differentiate individual linking devices semantically. As for
syntax, students need to learn the flexibility
of
connector-positioning, again by studying
authentic texts.
We agree with Crewe (1
990)
that misleading lists of so-called interchangeable con-
nectors often found in textbooks should be avoided at all costs. We also agree with Zamel
(1 984: 1 16) that ‘learning when not
to
use them (connectors) is as important as learning
when to do
so.’
French students’ attention needs to be drawn in particular to the danger of
overusing corroborative and additive connectors.
Finally,
it
would also be useful, as Khalil
(1
989) points out, to incorporate contrastive
rhetoric statements into the teaching
of
connectors.
CONCLUSIONS
This study demonstrates some
of
the problems students face in using connectors in dis-
course and proposes ways
of
tackling these problems. However, much more research needs
to be done before ideal teaching methods and materials can be developed. Granger (forth-
coming) speaks
of
the need for more French/English contrastive studies in order
to
get a
better description
of
the problems learners are likely to face. This is certainly necessary in
the area
of
connectors. Contrastive studies into connectors could in turn give rise to better
dictionaries which, as we have seen, come in for criticism. And problems with connectors
are not limited
to
French learners
of
English. Tricas (1990: 529) for example, in an article
on French/Spanish concessive connectors, calls for ‘un bon dictionnaire de connecteurs’
which would be ‘un bon instrument de travail dans l’activite traductrice.’
It
is also necessary
to
conduct more contrastive interlanguage investigations.
It
seems
0
Blackwell
Puhlibhen
Lfd.
l9Y6
26
Sylviane
Granger
and Stephanie
Tyson
that transfer is one
of
a number
of
explanations for problems at discourse level, and that
universals also play
a
part. Computerized learner corpora can make a major contribution in
analysing the relative effects of transfer and universals, one
of
the main questions in Second
Language Acquisition research.
We raised earlier the possibility that overuse
of
some semantic categories of connector
and underuse of others may be due to a difference in French/English argumentation.
If
statements about contrastive rhetoric are
to
be made, then much more research needs to be
carried out in this area.
But we should not forget that connectors represent one small aspect
of
cohesion.
Although we need to teach their correct use, we should not do
so
to the detriment
of
other
forms
of
cohesion. Furthermore, coherence should be the primary discourse consideration:
no matter how much students study connectors or any other aspect
of
cohesion, an incoher-
ent message will always remain
so.
However, increased mastery of cohesive devices will
certainly help students to express relations more clearly. We hope that heightened aware-
ness
of
the semantic, stylistic and syntactic properties
of
connectors will lead students
to
think more carefully about the ideas these connectors are linking.
REFERENCES
Altenberg, Bengt (1986) Contrastive linking in spoken and written English.
In
English in Speech
and
Writing.
Edited by
G.
Tottie and
1.
Backlund. Uppsala: Semqvist
&
Wiksell International. pp. 13-40,
Ball, William J.
(1986)
Dictionary
of
Link Words in English Discourse.
London: Macmillan.
Biber, Douglas
(1
988)
Variation across speech and writing.
London: Cambridge University Press.
Connor, Ulla
(1
984) A study of cohesion and coherence in English as a second language students’ writing.
Papers
Crewe, William
(1990)
The Illogic
of
logical connectives.
ELTJournal,
44(4), 3 16-325.
Evenson, Lars
S.
and Rygh, Irmgard L.
(1
988)
Connecting L
I
and FL in discourse-level performance analysis.
Papers andstudies in Contrastive Linguistics,
22, 133-1 78.
Field, Yvette and Yip, Lee M.
0.
(1992) A comparison of internal cohesive conjunction in the English essay
writing
of
Cantonese speakers and native speakers
of
English.
RELC
Journal,
23( 1). 15-28.
Granger, Sylviane
(1
993) The International Corpus of Learner English.
In
English Language
Corpora:
Design,
Analysis and Exploitation.
Edited by
J.
Aarts,
P.
De Haan and
N.
Oostdijk. Amsterdam
&
Atlanta: Rodopi,
Granger, Sylviane (Forthcoming) From CA
to
CIA and back: an integrated contrastive approach
to
computerized
bilingual and learner corpora.
In
Languages in Contrast. Text-based Cross-linguistic studies.
Edited
by K. Aijmer, B. Altenberg and M. Johansson. Lund Studies in English, Lund: Lund University Press.
1996.
in Linguisrics
17(3).
pp. 57-71.
Halliday, Michael A. K. and Hasan, Ruqaiya (1976)
Cohesion in English.
London: Longman.
Hartnett, Carolyn
G.
(1
986) Static and dynamic cohesion: signals
of
thinking in writing.
In
Functional Approaches
Hervey, Sandor and Higins, Ian
(1
992)
Thinking Translation.
London: Routledge.
Khalil, Aziz (1989) A study of cohesion and coherence in Arab EFL College students’ writing.
System,
17(3),
Martel, Ghislaine
(1
99
1)
Les connecteurs: problemes particuliers
i
I’analyse differentielle du discours.
Langues el
Linguistique,
17,
159-
167.
Milton, John and Tsang, Elza
S.
C.
(1
993)
A Corpus-based study
of
logical connectors in EFL students’ writing:
directions for future research.
In
Studies in Lexis.
Edited by
R.
Pemberton and E.S.C. Tsang. Hong Kong: The
Hong Kong University
of
Science and Technology. pp. 2 15-246.
to Writing.
Edited by
B.
Couture.
London:
Pinter. pp.
1q2-15
1.
359-37
1.
Nickel, Gerhard
(1989)
Some controversies in present day error analysis.
IRAL,
XXVII,
293-305.
Neuner, Jerome L.
(1
987) Cohesive ties and chains
in
good and poor freshman essays.
Research in the Teaching
of
Newmark, Peter
(1988)
A
Textbook
of
Translation.
Englewood Cliffs: Prentice-Hall International ELT.
Pritchard, R. J.
(1981)
A study
of
the cohesion devices in the good and poor compositions of eleventh graders
(doctoral dissertation, University of Missouri-Columbia,
1980).
Dissertation Abstracts International,
42,
688a.
English,
21.92-103.
0
Blackwrll
Publishers
Ltd.
1996
Connector usage
in
EFL
essay writing
27
Quirk. Randolph, Greenhaum, Sidney, Leech, Geoffrey and Svartvik, Jan
(1
985)
A
Comprehensive Grammur
of
Tirrney,
R.
and Mosenthal,
J.
(
1983)
Cohesion and textual coherence.
Research in
the
Teaching
of
English,
17,
Tricas. Mercedes
(
I
990)
Largumentation concessive francaise et espagnole.
Meta,
35(3),
529-537.
Tyson. Stephanie
(I
995)
Manual versus automatic analysis
of
connectors: chalk and cheese’? Paper presented at
the Symposium on Computer Learner Corpora: Grammar. Lexis and Discourse, Louvain-la-Neuve,
January
28,
1995.
the
English Langrtrrge.
London: Longman.
2
15-229.
Vinay,
J.
P. and Darhelnet.
J.
(
1977)
.Sfyli.s/iyrcc~
C’otnparPe drc franqais
et
de
I’unglais.
Didier, Paris.
Witte. Stephen P. and Faigley. Lester
(1
98
1
)
Coherence, cohesion and writing quality.
College Composition and
Zamel. Vivian
(1
9x4).
Teaching those missing links in writing. In
Composing
in
a
Second Language.
Edited by
(Received
17
March
1995.)
Coniniiozit.rclion,
32.
189-203.
S.
McKay. San Francisco State University, Cambridge: Newbury House.
pp.
I
10-1
22.