Examining present tense conventions in scientic

writing in the light of reader reactions to three

Dutch-authored discussions
Joy Burrough-Boenisch*
Science Editing and Translation, Boeslaan 3, 6703 EN Wageningen, Netherlands
Abstract
In scientic English it is conventional for present tense to signal general truth (scientic
universality); the past tense is then used to report the authors own research actions and
ndings. Both NS and NNS scientic English may, however, depart from the present tense
convention - for dierent reasons. Data collected froma reception study in which 45 readers from
eight countries evaluated and annotated the same three Discussion sections written by Dutch
biologists (but not yet corrected by an NS) form the basis for a discussion of these reasons.
The NNSs competence in English, mother tongue interference and non-anglophone tense
conventions for reporting past events are dealt with. The readers responses to the preponderance
of present tense in the texts appeared to be inconsistent. Possible reasons for this are suggested
and the implications of the ndings for writers, teachers, editors and reviewers are discussed.
# 2002 Published by Elsevier Science Ltd on behalf of The American University.
1. Introduction: tense conventions in scientic English
Scientic English has certain verb tense conventions that in the more prescriptive
textbooks on scientic writing are often presented as rules, perhaps most suc-
cinctly by Day:
1. Established knowledge (previous results) should be given in the present tense.
2. Description of methods and results in the current paper should be in the past
tense.
3. Presentation (Table 1 shows that. . .) is given in the present tense.
4. Attribution (Jones reported that. . .) is given in the past tense. (Day 1995, p.72)
English for Specific Purposes 22 (2003) 524
www.elsevier.com/locate/esp
0889-4906/02/$22.00 # 2002 Published by Elsevier Science Ltd on behalf of The American University.
PI I : S0889- 4906( 01) 00049- 7
* Corresponding author. Tel.: +31-317-420283; fax: +31-317-416774.
E-mail address: burrough@bos.nl (J. Burrough-Boenisch).
Day stresses that these rules must be adhered to, to prevent readers having dif-
culty in distinguishing between an authors own ndings and what is accepted sci-
entic knowledge or universal truth. Rules 1 and 2 encapsulate a paradigm of
science writing in English, that past tense signals the specic (the method used and
the results obtained in a given experiment or research project) whereas present tense
signals the universal (that a method is standard, that a statement is irrefutable).
It has been said that in English science writing, proper use of tense derives from
scientic ethics (Matthews, Bowen, & Matthews, 1996: p.104). Put simply, the
change of status a piece of information undergoes from being a nding in a parti-
cular experiment to a being a tenet of science is accompanied by a tense change. The
nding is rst reported in the past tense in a research article, because at the time of
writing the results are still research-specic: they have not yet been accepted by the
discourse community and become part of shared scientic knowledge. Used this
way, the past tense is thus a hedging device (Hyland, 1998). But once published in a
reputable journal (after passing the scrutiny of the journals reviewers and editor,
the gatekeepers to the dissemination of scientic knowledge to the discourse com-
munity), the information becomes part of established scientic knowledge. Hence-
forth, other scientists may acknowledge this by referring to it in the present tense.
That NSs of English have to have the tense conventions of scientic English
inculcated by textbooks and by writing courses indicates that these conventions are
a cultural construct that aspiring members of the discourse community need to
master. As Taylor (1989: p.150) notes . . .there are certain conventions about the
use of tensesespecially past and presentin academic writing which are not
intuitively obvious. NNSs of English who aspire to be members of the international
academic discourse community also need to know these conventions, which is why
the writing manuals for such readers take pains to explain the role verb tenses play
in academic (scientic) writing. Whereas scientists and editors prescribe tense use in
scientic English (e.g. Booth, 1984; Day, 1995; Huth, 1999; Matthews et al., 1996;
OConnor, 1991) Weissberg and Buker (1990) and Swales and Feak (1994) give
Table 1
The readers in the reception study
Country Reviewers
a
Non-reviewers Total
US 5 1 6
UK 5 3 8
Netherlands 2 3 5
Germany 2 3 5
Sweden 4 0 4
France 3 1 4
Spain
b
5 4 9
Japan 2 2 4
Total 28 17 45
a
Dened as people who had reviewed at least one English-language article for a scientic journal.
b
Includes six Catalan speakers.
6 J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524
advice based on analyses (their own or by other applied linguists) of papers pub-
lished in a range of scientic disciplines that show that in real life, authors and
journals do not adhere rigidly to the conventions.
1
They attribute rhetorical power
to tense deployment, for example:
The dierences among. . .tenses are subtle. In general, a move from past to
present perfect and then to present indicates that the research reported is
increasingly close to the writer in some way: close to the writers own opinion,
close to the writers own research, or close to the current state of knowledge
(Swales & Feak, 1994: p.184).
The simplistic paradigm of appropriate use of the past and present tenses in sci-
entic writing as advocated by Day, and its renements (as explicated by Swales &
Feak, for example), are part of the culture of science in anglophone countries. They
are inculcated not only by manuals on writing but also by senior members of the
scientic community; for example, Dudley-Evans (1991) reports a supervisor chan-
ging the tense in drafts of an NS biology PhD thesis. NNSs, however, may have
extra diculty in deploying tenses correctly; it has been contended (Hinkel, 1997:
p.289) that the learning of past tenses in English is particularly complex. Klein (in
Dietrich, Klein & Nouyau, 1995: p.7) has noted (in the context of learner utterances)
that language learners in general consciously avoid tense forms to express tempor-
ality, preferring to use other features of discourse, such as temporal adverbs. Given
the connotation of the present tense in scientic English, however, an avoidance
strategy involving present tense verbs could lead to miscommunication.
Another factor aecting NNS deployment of tenses in their academic and scien-
tic writing is the inuence of mother tongue. There is scope for the misguided
transfer of tense meaning, as Hinkel (1997) demonstrated in her study of tense use of
Asian NNSs (Chinese, Korean, Japanese and Indonesian) compared with an NS
control group (American English) in student essays and Cloze tests. She attributed
the deviation from NS practice to the Asian languages having a dierent perspective
on the sense of timelessness conveyed by the present tense and linked her ndings
to the tense conventions in discourse.
Transfer of tense conventions from the NNSs native discourse to English is par-
ticularly likely to aect present tense use because English diers from European
languages in its discourse conventions on present tense. The present historic is
perhaps rather less common in English than in other European languages (Comrie,
1
Note that the conclusions drawn from such research may say more about journal policy on verb tense
than about authors preferences. (Even if an author ignores the journals guidelines to contributors, a
copy editor will emend verb tenses to house style.) For example, McCarthy and Carter (1994) found that
all the verbs in the Abstracts of the British Medical Journal were in the past tense, except for the nal
sentence, which projects forward to future research. They note (p.102) that The shifts in grammatical
choice are a fundamental feature of the genre, adhered to by all contributors (my italics), apparently
overlooking the role of the BMJs copy editors. Yet the BMJs editors are committed to standardising
how information is presented in BMJ articles (Doherty & Smith, 1999) and current BMJ house style sti-
pulates structured (i.e. formulaic) Abstracts.
J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524 7
1976: p.73) and its use in EST discourse is perceived as too informal (Malcolm:
1987, p.38). In French, however, not only is the present historic used widely in his-
tory textbooks and journalese (Monville-Burton & Waugh, 1991), it is institutiona-
lised in scientic discourse. That the present tense is part of the culture of French
science can be seen from the following reaction to the editor of Annales de Derma-
tologie et Ve ne re alogie, who had broken with French convention by instructing
authors to report materials, methods and observations in the past tense (Mate riel
et me thodesou observationa` rediger au passe ).
Alors, pourquoi le passe quand notre souci dans la re daction me dicale est de tre
pre cis et clair, compris de tous et aise ment traduit?
. . .Constamment, le pre sent est plus simple, plus bref. . .et surtout, il est plus juste.
2
(Delauney, 1998: p.568, my emphasis)
Anglophones might view the phrases I have highlighted in bold as demonstrating
quaint francophone linguistic chauvinism, but example 1 shows that this French
tense convention can be transferred to English and may slip through into an Eng-
lish-language journal in spite of a language corrector (acknowledged by the
authors). The result is miscommunication: the citation appeared under the heading
Mapping the rainfall distribution and describes the procedures the authors fol-
lowed in their study; it is not a general description of a standard procedure.
1. The regression models are applied to the 300 reference points that provide for
each of these points a couple of values, Gd(60) and P
0
(60). The variograms of
the rst four PCs show a minimal range of 6 km, so a sampling every 2 km
should be sucient for describing the spatial variability of the relief (in its
general trend). Spline interpolation (Smith and Wessel 1990) of the 300 CP-
derived estimates is then carried out to produce the map of the Gumbel
parameters. . .(Wotling, Bouvier, & Fritz, 2000: p.96)
2. The reception study
Given the cultural dierences in the pragmatics of the present and past tenses in
science mentioned above, it is interesting to examine how NNS scientists cope with
the tense conventions of scientic English and how their peer readers (both NS and
NNS) react. A recent (1999) reception study (i.e. a study to nd out what a parti-
cular category of readers makes of a particular text) has provided an opportunity to
do this. The study was intended to reveal whether there would be any consensus on
obstacles to comprehension in Dutch scientic English and, if so, how it would
manifest; for example, as NSs versus NNSs, or as a dierence between arguably
2
So why use the past tense, given our concern in medical editing to be precise and clear, understood by
all and easily translated? . . .The present tense is constantly simpler, more direct, more concise . . .and,
above all, is more accurate.
8 J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524
more critical readers who review articles for English-language scientic journals and
normal (i.e. non-reviewer) readers of such journals. In brief, its features were:
Three texts: A (1425 words), K (695 words) and R (1150 words), being the
Discussion sections of three research articles on vegetation science. In the
Discussion, an author evaluates and comments on the research results, relat-
ing them to the corpus of knowledge. The Discussion is therefore more
argumentative than the Methods or Results sections, in which the emphasis is
on reporting (i.e. narration), and hence it is more challenging to write
especially for an NNS. [The Introduction section is similarly challenging, but
the choice fell on the Discussion because a previous study (Burrough-Boe-
nisch, 1999) had suggested that Introductions are less important to readers].
The test texts were selected in 1998 from manuscripts by rst-time clients of
an editing and translation service. The criteria for selection were subject
matter and Discussion length (<1500 words), not any particular textual fea-
ture(s). A and R were to be submitted to journals, K had appeared in a
published PhD thesis. None of the texts had been corrected by an English
NS. To provide the readers in the reception study with context, the texts were
presented with their titles and Abstracts.
The authors were Dutch biologists (vegetation scientists) aged 44, 53, and 46.
When I selected the texts I was unaware of their age or their previous pub-
lication record (all had previously published in English). The fact that the
manuscripts were sent for language revision indicates that the authors (or
their superiors) recognised that the English was not optimal. However, these
manuscripts were by no means the worst examples of Dutch scientic English
I have encountered since 1976.
The readers were 45 professional biologists/vegetation scientists from eight
countries, including the UK, the US and (the control group) the Netherlands
(see Table 1), who did the study in their spare time. They had been recruited
by letter or e-mail from the reviewer lists of biology or ecology scientic
journals or from university Web sites. Some of these journal reviewers were
persuaded to recruit non-reviewer colleagues.
On recruitment I collected personal data (e.g. age, mother tongue, languages spo-
ken or read, whether a reviewer). From these data I know that none of the readers
(except the control group from the Netherlands) had any knowledge of Dutch or
Dutch writing conventions. Except for ve of the Spanish group (who had Catalan
as mother tongue) all were NSs of the language of their country of residence.
This paper describes some of the data collected fromthe annotations the readers made
to the texts. The texts were sent stapled together in random order (AKR, ARK, KAR,
KRA, RAK, or RKA), and the readers were asked to make changes to improve the
English and the texts eectiveness and appropriateness. As my aim was to elicit
spontaneous comments and emendations, I took care not to prime readers to look at
any specic linguistic or discourse features of the texts. The main features readers
responded to spontaneously were scientic content, verb tense, redundancy, cohesion,
J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524 9
hedging,
3
and typographical and grammatical errors. Of these, the most striking
response numerically was elicited by the tense use. It was almost invariably triggered
by the simple present tense, which, as Table 2 shows, predominated in the three
texts. In this paper I will therefore focus on this tense. The use of present tense in the
three texts was rarely ungrammatical, as can be seen from Text R (the text with the
most present tense verbs), in the Appendix.
3. Results
3.1. The NS and NNS responses to the present tense in the Discussion texts
The readers response to the verb tense can be expected to reect their awareness
of and attitude to tense conventions in scientic English. I expected the reviewers to
be stronger upholders of such conventions than normal journal readers. This is
because as established (i.e. published) scientists, who have been given guidelines on
reviewing by one or more journal editors, reviewers read to assess scientic and
genre appropriateness rather than primarily for information (Burrough-Boenisch,
1999). Native-speakerhood (i.e. whether a reader was NS or NNS) will also have
been important in this study, in terms of prociency in English and mother tongue
interference.
Nine of the 45 readers reacted to tense use by writing comments in the margins,
sometimes to reinforce emendations. One, a British reader, did so on each text (e.g.
in Text K: ? Was this observed, in which case use past tense, or is it just a statement
of logic, in which case use present tense.) Of the eight other readers who wrote
comments on tense, three (one American, one British, one German) did so on Text
A, one (Spanish) on Text K and four (one American, two British, one French) on
Text R.
Readers emendations of present tense to past tense far outnumbered their emen-
dations from past tense to present tense: 295 versus 24. They were almost invariably
to simple past. Only 6% of the 295 emendations were to the present perfect; they
Table 2
Tense use in the three Discussion texts
Text code Present tense verbs Past tense verbs Future tense verbs
a
N % N % N %
A 69 64 36 33 3 3
K 41 95 2 5 0 0
R 80 90 7 8 2 2
a
All these were modal will with future time reference.
3
Though the use of the simple past tense in research articles is often a hedge (Hyland, 1998) my
hedging data class referred to other hedging devices.
10 J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524
were made at ve points in the texts (four in Text A and one in Text R). Table 3
shows the magnitude of the shift from present tense to past tense that the readers
achieved collectively. Compiled by counting each verb changed from present to past
by at least one reader, this table shows that accepting all these emended verb tenses
would increase the past tense verbs hugely in Texts K and R and greatly in Text A
(the text with the most past tense verbs originally). The table also shows how
implementing all these present to past tense changes would have aected the nal
proportions of past tense verbs in the texts: Texts A and R would have ended up
with well over half their verbs in the past tense.
The mean number of present to past changes per text was 4.29 (S.D.=6.94) for
the NSs and 1.24 (S.D. 3.88) for the NNSs. Reviewers averaged 2.79 changes
(S.D.=6.14) and non-reviewers 1.20 changes (S.D.=2.91). To ascertain the statis-
tical signicance, I performed univariate General Linear Model (GLM) analyses on
the present to past tense changes. This analysis also looked for signicant interac-
tion eects; unless mentioned, none were found. In my rst analysis I looked for
evidence that knowledge of English and of English tense conventions could have
been signicant in motivating these changes. As xed factors I therefore used native-
speakerhood and reviewer status andbecause I was interested in dierences
between textstext. The only signicant eect was that of native-speakerhood: the
NSs (i.e. the American and British readers) were clearly more active in changing
present to past tense [F (1, 123)=8.973; P=0.003]. To explore this eect further, I
repeated the analysis, substituting readers country for native-speakerhood (and
therefore distinguishing between American and British NSs as well as between the
mother tongues of the NNS). The eect was again very signicant [F (7, 90)=2.977,
P=0.007] and at this level of analysis, reviewer status became signicant [F (1,
90)=4.466, P=0.037]. A post hoc analysis (Tukey) revealed no signicant dierence
between the American and British NSs but showed that the British NSs diered
signicantly from the German and Spanish readers, who made very few, if any,
present to past tense changes.
The nding that reviewer status became signicant in the second analysis, which
took account of readers country, suggests that analysing at this level allowed the
performance of the NNS reviewers in the small country groups to emerge as
important. If the present tense was indeed being used to refer to the authors specic
research ndings (as is the case in the rst 10 sentences of Text R), we may infer that
these reviewer readers (and the NS reviewers) were trying to uphold the anglophone
Table 3
The shift from present to past tense that would be achieved by implementing all the verb changes made by
the readers
Text code Increase in
past tense verbs
Proportion of past tense verbs
in text if all changes implemented
A 183% 61% (was 33%)
K 950% 44% (was 5%)
R 757% 60% (was 8%)
J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524 11
convention of using the past tense for ones own research ndings. This could explain
why neither the Dutch readers (expected to share the authors tense perceptions) nor
the French readers (recall the French convention for reporting science in the present
tense) made the fewest tense changes; in both these groups, the reviewers behaved
more internationally than chauvinistically.
3.2. Dierences between texts
I was not surprised that the GLM analysis looking at native-speakerhood found
text to have a signicant eect on the present to past changes [F (2, 123)=3.704,
P=0.027]. Dierence in text length must have been important. But note too that in
Texts K and R, 90% or more of the verbs were in the present tense, compared with
64% in Text A (Table 2). Another dierence was that 20% of Text Rs present tense
verbs were in the passive voice, but Texts A and K had fewer than 5% passive verbs.
The histograms of numbers of changes of present to past tense (Fig. 1) indicate the
dierent response patterns to the present tenses in the texts.
Not only were the frequencies of tense changes per reader in the three texts very
skewed, but it can also be seen that even though the tense use in the texts elicited
vigorous spontaneous comments from some readers, the frequency of tense changes
was generally smallmuch smaller than the potential suggested in Table 3. Fur-
thermore, the largest category was zero changes. Should this zero response be
interpreted as tacit endorsement of the predominance of present tense, or as indi-
cating that verbs were not noticed? (Readers may have been too busy emending
other text features.) The presence of eight overt verb errors (ve in Text A, three in
text K) enabled me to assess readers alertness to verbs. Three of the errors were in the
participle (e.g. Although these species can not be compare directly with perennial
Fig 1. Histograms of present to past tense changes.
12 J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524
species. . .): two of these were emended by over 50% of the readers. One error
(Comparative studies with respect to plant morphology were often done in short
term experiments) could be of tense (are often done) or aspect (have often been
done). There were four errors of aspect (e.g. Still, many national oras are not
fully covering the subgenus.
The raw data showed that the NSs were most alert to these wrong verbs. Uni-
variate GLM (xed factors: native-speakerhood, reviewer status and text) found a
very signicant eect for native-speakerhood [F (1, 123)=22.535, P=0.000].
Reviewer status was not signicant. (That text and the interaction between native-
speakerhood and text were highly signicant is spurious, given the distribution of the
wrong verbs among the texts.) A univariate GLM analysis substituting readers
country for native-speakerhood, followed by a post hoc test, showed that though the
American and British readers corrected signicantly more of the eight wrong verbs
than the Germans, the Americans were not as alert as the British readers, who also
corrected signicantly more wrong verbs than the French, Spanish and Japanese.
Given the NS dominance in emending wrong verbs and arguably inappropriate
verb tenses, could the large zero response to present tenses (Fig. 1) have been an
NNS response, attributable to lack of competence in English? If it were, the readers
who changed tenses would have responded consistently across all three texts. They
did not. Only six of the 45 readers made present to past tense changes in all three
texts. In other words, only six readers seemed to have been applying the rules of
tense use in scientic English consistently. Twelve readers (one American, one
Dutch, two Germans, two Swedes, two French, two Spanish and two Japanese,
together accounting for 27% of the total readers) made no present to past tense
changes in any of the texts. Lack of prociency in English does not necessarily
account for their behaviour: eight of them (including the American) corrected at
least one wrong verbindeed, the American corrected several. I infer that the
reason they did not change the present tenses was because they did not nd them
incongruous or inappropriate. Only four readers (one German, one French, two
Spanish) made no verb-related changes, i.e. did not correct wrong verbs or change
tenses. Arguably, these readers were less procient in English, or did not know the
genre tense conventions, or both.
Another interesting nding is that 27 of the 33 readers who did change present to
past tense made no such changes in at least one of the texts.
3.3. Extreme reader behaviour
The tense use in Texts K and R provoked extreme responses from two readers:
one British reader, B3, changed 14 present tense verbs to past tense in Text K, and in
Text R made 39 such changes. He is clearly visible in Fig. 1 as the extreme outlier in
the histograms for Texts K and R (the other outlier, a Japanese (J1) changed 31
present tense verbs to past tense in Text R). Expressed as proportions of the present
tense verbs in the text concerned, B3s tense changes aected 34% of the present
tense verbs in Text K and 49% of the present tense verbs in Text R. This means that
he changed the proportions of past tense verbs in Texts K and R from 5 and 8 to 37
J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524 13
and 52%, respectively. In other words, he achieved a greater rhetorical shift (changing
the status of the information to mark it explicitly as specic to the study) in Text R
than in Text K. The extreme behaviour of readers B3 and J1 greatly inuenced the
data shown in Table 3; if their changes are ignored, the shift from present to past
shown in that table would be much less dramatic.
Note that though B3 and J1 overlapped on 26 of the verbs they changed from
present to past tense in Text R, there were 19 present tense verbs that only one of
them changed.
Reader J1 was the reader who made the most present to past tense changes (14) in
Text A. In that text, reader B3 was not the NS who changed the most present tenses
to past; ve other British readers surpassed his eight changes.
3.4. Which verbs were changed from present to past tense
Though 95 dierent verbs were changed from present to past, only 27% were
changed by three or more readers. In Texts K and R the highest frequency of a
change from present to past was six, but in Text A there were two present tense
hotspots. The rst was a present tense embedded in a sentence written in the past
tense. Fifteen readers changed it to past tense; eight of these were NSs. Their motive
may have been to correct the tense inconsistency, but from example 2 (the verb is
underlined) it is clear that an alternative solution would be to emend have to to
have; only three (all NNSs) of the 45 readers did so, and one NS deleted the
have.
2. At harvest 4 within the species Holcus and Rumex (from a nutrient rich per-
ennial stage) were found to have the highest root fraction and within the
grasses Holcus have the highest specic root length.
The second hotspot, in the phrase In this research we examine biomass
allocation. . ., occurred at the start of the penultimate paragraph. The 18 readers
who changed the tense presumably reacted to the incongruity of referring to com-
pleted research in the present tense rather than to an inappropriate claim of uni-
versality. Two of the 18 readers opted for a present perfect (have examined rather
than a simple past examined).
4. Discussion of the tense changes
It is clear that the pooled data on present tense changes conceal a wide variety in
the frequency of these changes, in terms of numbers of changes made per reader and
per text and also which verbs were changed. To attribute the tense changes recorded
in this study to readers response to a mismatch between the tense in the text and the
genre convention on tense use is, therefore, to oversimplify. Such a response
would have produced a much clearer signal, one comparable to that of readers
B3 and J1.
14 J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524
Even acknowledging possible reader oversight, the dierences in response suggest
that deciding on appropriate tense use in science reporting is very subjective. The
huge zero response to the present tenses in the texts, plus the fact that any given
present tense verb was rarely changed to past by more than three readers, implies
that most readers accepted the tense use. The exceptions were readers B3 and J1
(both reviewers). When asked to explain his response to Text R (49% of present
tense verbs changed to past tense), Reader B3 wrote:
I think the reasons for my changes were based on correcting unconventional use
rather than errors in grammar. I think this is due to an opinion that results and
conclusions from experimental studies are reporting on past events. In the
context of this paper I believe it is correct to say . . .grasslands are mown at the
end of the growing season, as this reports ongoing events. However, reporting
on your results, for example, both species richness and diversity are strongly
correlated, is actually claiming more than you measured. The correlation
holds only for the period over which the information was gathered. Though it
probably still holds now, it may not do, and should be reported in the past
tense.
It is a long time ago when I was at University (15 years) so I cant really
remember if I was taught it or learned it by imitation. I was never taught
directly about how to write, but it must have occurred through comments on
written work. It may even have been drummed into me at school. However, I
must admit that I still sometimes lapse into the present tense, and have to be on
my guard when I revise texts to ensure I am consistent.
If pressed on my opinion on the aects of the present tense on the impact
of science reporting, I would have to say that it can bias the impact of
what is written. The present tense provides a greater immediacy and certainty
to science writing, but at the expense of humbleness. By that I mean that
science is a collection of hypotheses of various power and consistency, and
our goal as scientists is to refute these hypotheses if we can. Science is not a
eld of certainty and reporting it in such terms is counter productive in the
long-term.
Clearly, B3 conscientiously attempted to uphold the genre conventions of tense
use. Note that in the case of present tense statements in a Discussion being restate-
ments of results, as occurred in all three texts, not only in R, it can be argued
(Malcolm, 1987: p.38) that the past tense is obligatory, as otherwise the situations
referred to would be timeless or at least omnitemporal. But looking at the open-
ing of Text R, one could argue that the obligation is not merely to comply with
genre convention: perusal of the opening lines shows that the author does not
explicitly say she is reiterating her ndings; she puts the onus on the reader to infer
the specicity of her statements. Comrie (1985) has pointed out that when inter-
preting a tense, a reader makes use of other features of the structure of the sen-
tence and knowledge of the real world. In this case, Author R does not supply
other features, so the scientist reader uses specialist knowledge, augmented by
J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524 15
information supplied earlier in the text (in this case, in the Abstract, where the
ndings were reported in the past tense). Instead of changing tense to remove the
ambiguity, readers could have made the specicity of the statements overt, by
inserting textual clues (e.g. by beginning Our results show that in our study
area. . .); none did so.
Previous research has demonstrated the role of textual clues in enabling readers to
infer temporal context. In their reception study using three NNS Introductions of
research articles, Chappell and Rodby (1983) demonstrated that time adverbials
send strong signals about the time domain of a tense. Their NS readers judged that
the most easily comprehensible versions of three texts were those rewritten to
include explicit time signalling by such adverbials. Signicantly for my study, most
of Chappell and Rodbys readers had no problems with the tense switching in the
texts, whose authors were NNS graduate students. The only readers who showed
strong preference for tense consistency (text versions rewritten in a single tense) were
the teachers of English as a foreign language! We may infer that writers have some
leeway in their choice of tenses.
The readers in my study seem to have behaved like the non-teacher readers in
Chappell and Rodbys study. From textual clues about the temporal context, plus
their knowledge of the science involved, they were able to infer that certain present
tenses in the Discussion texts referred to research-specic results rather than to uni-
versal scientic truths. This could explain why for each text the zero tense changes
category contained the most readers. Except for readers B3 and J1, those readers
who did change present tenses to past tense presumably made strategic changesat
points in the text where they felt it was necessary to be reminded of the time-
boundedness of the information being presented. In this way, the occasional past
tense would function as a time framer, in the same way as the adverbials Chappell
and Rodby used.
One sentence from Text A illustrates the options readers chose when changing
successive instances of present tense to past.
3. The fact that they grow faster caused an increased need for nutrients and may
induce a poorer environment in the pots which results in a functional
response in the shoot-to-root ratio.
The nine readers (seven were NSs) who made one or more present to past
changes here did not include B3 or J1. Three (two British, one French) opted for
tense consistency, changing all three present tenses to past: three changed the rst
and second, one changed the second and third, one changed the second only and one
the third only. Those who emended the may did so to the present perfect.
As I did not interview the readers after processing the data, I cannot be certain
why the NSs emended tense and corrected verbs more actively than the NNSs.
However, a study by Vann and Meyer (1984) is relevant here. They reported that US
university lecturers asked to assess the acceptability of English written by NNS stu-
dents of science or social science put tense errors into the least acceptable category of
errors. The latter for the most part are global and/or are relatively rare violations
16 J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524
for native speakers. Unfortunately, Vann and colleagues do not say whether the
lecturers perception of tense error was motivated by wrong grammar, or by a per-
ceived mismatch with genre conventions.
5. Why the Dutch authors preferred the present tense
That Dutch authors tend to overuse the present tense in English can be inferred
from the attention paid to this in style manuals written for Dutch learners of English
(notably, Hannay & Mackenzie, 1996). Before drawing some general conclusions
from the ndings presented above, I will briey discuss the possible reasons for this
tense preference in the three texts.
5.1. Authors lack of prociency in English
This reason is unlikely. Each of the authors (all had previously published in Eng-
lish) had been learning and using English for over 30 years. Furthermore, together
the texts contained only eight verbs that were grammatically or orthographically
wrong. The clinching evidence, however, is that the authors used past tense
conventionally in their Abstracts. (Indeed, in Abstracts A and R that tense pre-
dominated.)
5.2. Inuence of Dutch tense meanings (i.e. semantic transfer)
In Dutch, the onvoltooid tegenwoordig tense [which Sanders et al. (1992, p.107)
contend may well be the most frequently used of all Dutch tenses] is in the form of
a present tense but it also has a continuative use akin to the present perfect tense in
English. They have lived in London since 1972 (which is compatible with their
still living there) is expressed in Dutch by Zij wonen [present tense] al sinds 1972 in
London. Example 4 (from a Dutch-authored research article not used in the recep-
tion study) illustrates this transfer:
4. In the Gelderse Vallei the possibilities for restoration of peat grassland are
investigated in the trial area since 1987.
5.3. Mistranslation/transfer of auxiliaries from Dutch
There is also scope for a Dutch tense form that looks like an English tense form to
be transferred into English. What encourages the transfer is the possibility of form-
ing the perfect (voltooide tijd) in Dutch by using the auxiliary zijn (from the verb to
be) plus the past participle of the verb. When wishing to use a passive construction
(these are common in science: see e.g. Matthews et al., 1996; Tarone & Dwyer, 1981)
a Dutch author might inadvertently transfer the Dutch third person singular form of the
auxiliary (is) into English, where its equivalent is the identical is. Evidence that this
happens is publicly visible in another genre in Amsterdam Schiphol International
J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524 17
Airport. Here, the ocial bilingual notices at the entrances to all public toilets fea-
ture a clock with adjustable hands and the texts:
5a. DIT TOILET IS GECONTROLEERD OM. . .UUR
THIS TOILET IS CHECKED AT. . .OCLOCK
But the correct English translation should be:
5b. THIS TOILET WAS CHECKED AT. . .OCLOCK
Errors of this type occurring in Dutch scientic English would signal habituality
or universal scientic truth. When I examined the passive voice present tense verbs
in the texts (two in A, one in K and 16 in R), however, I found they were embedded
in passages of present tense and were not associated with dates or other time-
bounding devices that would make them incongruous. This makes reason three
unlikely.
5.4. Mistranslation/transfer of auxiliaries from Dutch
This is largely speculative. The present historic is certainly more common in
Dutch than in English, particularly in historical narratives, including those within
academic or scientic texts (for examples, see Burrough-Boenisch, 2000). In line with
the convention in Dutch, Dutch NSs asked to take minutes of meetings in English
will often do so in the present tense, as shown below, in the rst few lines of the
English-language minutes of a professorial selection committee, recorded in the present
tense by a Dutch academic. I have underlined the instances of present tense.
6. Minutes of the meeting held in xxxxxxxxx on ** October 1999
1. Opening and welcome
The meeting is opened by the Chairman at 14.05 hours. Present are in alpha-
betical order: [A, B, C, D, E]. Absent with notication [F]. There follows a
short round of introduction of the members. . .
It is conceivable that this Dutch convention of reporting meeting minutes in the
present tense was adopted from the French, perhaps during French suzerainty
(17951806). An intriguingbut highly speculativepossibility is that French has
similarly inuenced the discourse of Dutch scientists; it was the international mod-
ern language used by Dutch scientists in the early stages of international scientic
communication. German (in which the present tense is also less marked than in
English) overtook French as the international language of Dutch scientists in the
nineteenth century and continued as such until the mid-twentieth century (van Ber-
kel, Helden, & Palm, 1999), when it was displaced by English.
18 J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524
Whatever the reasons for Dutch scientists preference for the present tense, this
usage will be reinforced within the Dutch scientic and academic discourse com-
munity when scientists and academics read draft English-language articles by Dutch
colleagues. Furthermore, as this is not a grammatical error but an error at discourse
level (see James, 1998), it may be overlooked by language correctors, journal
reviewers and editors (see example 1).
When questioned on her tense use, Author R, who had used many present tenses
(Table 2), of which 20% were present passives, said she had thought that in con-
structions such as high loam percentage is correlated with higher soil pH she was
using the perfect tense. Clearly, she was misguidedly transferring Dutch tense
meanings (reason three). She also said that she felt justied in using the present tense
for any of her results that had been shown to be statistically signicant, and would
certainly do so when writing in Dutch. She had been unaware of the pragmatics of
the present tense in scientic English.
6. Conclusions, and implications for writing, teaching and editing scientic English
My ndings on tense use and readers response to it conrm that tense conventions in
scientic English can be (and are) ignored. I have suggested some factors that may cause
NNSs (specically, Dutch authors) to apparently ignore the present tense conventions.
Because this international study was conducted at long range and relied on the goodwill
of busy scientists, many of whom had had to be chivvied into completing the study, I did
not attempt to conduct follow-up interviews to ascertain readers motives.
Though some of the readers in my study (generally the NSs) wished to uphold the
tense conventions, there was disagreement about which verbs needed to be past
tense to signal research specicity. At rst sight, my data suggested that readers
tense changes were motivated by a mismatch with genre convention, but scrutiny of
the response to the three texts showed that this conclusion is an oversimplication.
What motivated a reader to change the tense of a particular verb was often a desire
to achieve tense consistency within a sentence or paragraph (a motive that Chappell
and Rodbys, 1983 teacher readers would presumably have applauded). But the
tense emendation was very subjective: most of those readers who changed tenses
from present to past seem to have applied the tense convention selectively, generally
allowing present tense to remain and be implicitly interpreted as referring to the
ndings specic to the reported research. This acceptance of a tense mix agrees with
the analyses of English research articles carried out by, among others, Swales (1990).
Given that tense use conventions in English scientic and academic writing are not
adhered to slavishly by NSs, it is dicult to be sure about what underlies the way
competent NNSs handle verb tenses in written scientic English. When writing or
reading scientic English, do they rely on their L1 conceptualisation of the tempor-
ality signal conveyed by tense? Or do they behave like NSs who use tense gramma-
tically but unconventionally and yet pass the scrutiny of journal editors and
referees? In scientic English, though writers have leeway to mix tenses and may
out tense conventions, other aspects of the English should be sound. (For example,
J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524 19
there must be sucient explicit clues about the timeboundedness of the verbs). The
response to Text A suggests that there must not be too much tense mixing within sen-
tences and that if a text contains overt verb errors, especially at its beginning, readers
are less likely to accept tense use that deviates fromthe conventions in scientic English.
Perhaps they then also interpret tense mixing as indicating learner ineptitude rather
than as the author attempting to signal stance towards the information being presented.
Clearly, presenting tense use in scientic English as inviolable rules (cf. Day, 1995)
oversimplies actual practice and ignores readers abilities to infer meaning from
context. However, the easiest way for writers to minimise miscommunication about
the generality or specicity of the information being presented, is to keep to the
tense conventions in scientic English. This is certainly the safest option for NNS
writers who may be unskilled in deploying other devices to signal the specicity or
generality of information. To avoid misunderstanding or confusion arising from the
signals conveyed by verb tenses in scientic English, authors, editors and reviewers
should ensure that there are sucient other contextual clues indicating whether
information has a narrow or broad applicability. If this is done, then even uncon-
ventional present tense use, whether attributable to learner English, or to ignorance
of conventions, or to transfer of conventions from another language, or to a desire
to write vividly and forcefully, will not transmit a misleading scientic message.
Acknowledgements
I thank Professors C. de Bot and M. Gerritsen (supervisors), and Dr. F. van der
Slik (statistics) of Applied Linguistics, Nijmegen University, and two anonymous
reviewers. I am especially grateful to all the readers and authors involved in the study.
Appendix. Text R (the text attracting the most tense changes)
Discussion
In our study area variation in soil substrate determines vegetation composition to a
large extend. The loam percentage is practically zero in the sandy soils where spe-
cies-poor heath is the dominant vegetation. Grassland communities prevail on soils
with a higher loam content. Management has been adapted to the soil and vegeta-
tion present: the grasslands are mown at the end of the growing season, whereas the
dwarf shrub communities are left undisturbed. High loam percentage is correlated
with higher soil pH, higher calcium concentration and lower aluminium concentra-
tion in the soil (Fig. 4). Also the nutrient supply ratios are correlated to loam per-
centage. N/P- and N/K ratios are high on the sandy soils and much lower on the silt
and clay soils. Both species richness and diversity are strongly correlated to loam
percentage. Species richness varies but is relatively high in grassland and low in
heath. In this data set, the parent material is the determinant of several measured
parameters including management type. Therefore we rst analysed the data set as a
20 J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524
whole but subsequently analysed the two major vegetation types separately. The soil
pH is the strongest predictor of species composition of a whole set of interrelated
variables, including sand, silt, clay, Ca, AI and biomass (Table 1, Fig. 4). In Fig. 1
and 2 the eect of decreasing pH on species diversity is clear. We conclude that
increased acidity of the soil can be the principal cause for the declining species
richness in nutrient poor grassland and heathland communities, which is in agree-
ment with the results of other studies (Grime, 1979; Houdijk et al., 1993; Roelofs et
al., 1996). However we found that nutrient supply ratios also inuence the species
composition. The eects of the nutrient supply ratio in the vegetation biomass are
rather independent of the pH eect (Fig. 4). For species richness the multiple
regression results indicate also a signicant additional eect of the nutrient supply
ratios measured in the vegetation biomass (Table 2). N/P and N/K ratios in phyt-
ometer plants are stronger correlated with loam percentage in the soil than these
ratios in the vegetation are. Phytometer plants, all of one species, would be expected
to give the most reliable information on soil nutrient availability. Vegetation com-
position is stable during a long time but diers between the plots, which can explain
the lower correlation with soil acidity or loam percentage (Fig. 4).
The amount of phosphorus and potassium available to plants is related to soil
acidity in podzol soils. For plants the availability of soluble phosphates in the soil
decreases below pH 6. Potassium availability decreases with soil pH too, because in
podzol soils soluble potassium is leached more when CEC is low (Bolt & Brugge-
wert, 1976). The availability of nitrogen for plants is primarily determined by the
mineralisation of organic nitrogen in the soil which increases with an increasing
amount of soil organic matter and is less aected by soil acidity. The accumulation
of soil organic nitrogen is largely determined by the atmospheric deposition. Total
atmospheric deposition of nitrogen, corrected for canopy uptake of ammonium, is
3045 kg/ha/year in comparable heathland sites (Bobbink & Heil, 1993). Nitrogen
deposition increases N mineralisation but it may also accelerate soil acidication.
Yet nitrogen deposition may increase N supply, but decreases the supply of P and K
and thus, strongly increases N:P and N:K supply ratios. Our results (Fig. 4) support
the hypothesis that nitrogen deposition has additional negative eects on plant spe-
cies diversity in heathlands by inducing high N:P and N:K ratios.
In this data set most plots are located at sites where the N;P ratio of the vegetation
exceeds 16 (g. 1, 2). This indicates that phosphorus is the most important limiting
nutrient for vegetation growth (Koerselman and Meuleman, 1996), which is usual in
heathland vegetation on podzol soils in the Netherlands (Diemont, 1996). However
the high heterogeneity in soil and vegetation composition of the nature reserve is
illustrated by the broad range of N/P ratios in the data set including N/P ratios
indicating nitrogen limitation. Also the variation in N/K ratio in the data set is high.
Potassium seems to be a limiting nutrient for vegetation growth in at least part of
the plots. The aboveground biomass of the vegetation on these nutrient poor soils is at
its maximum at intermediate nutrient ratios, indicating that not one single nutrient is
strongly limiting (Fig. 3). Balanced nutrient supply ratios favour plant growth. Also
J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524 21
most of the endangered species are found in vegetation with balanced nutrient supply
ratios (Table 4), when more then one nutrient is limiting vegetation growth. Nitrogen
deposition will cause an increase in N:P and N:Ksupply ratios, and is therefore a serious
threat for the species richness in nutrient poor grassland and heathland communities.
The nutrient concentrations and nutrient ratios measured in phytometer plants are
strongly correlated with number of species and diversity (Table 1). The weight of
phytometer plants is strongly correlated with aboveground biomass of the vegeta-
tion. In this data set a phytometer appears a good tool to measure the nutrient
availability for vegetation in the soil.
Six species in this data set are on the red list or are protected in the Netherlands. All
these species strongly declined since 1950. Other species in the data set that strongly
declined after 1950 are Succissapratensis, Euphrasia stricta and Rhynchospora alba
(Mennema et al., 1980, 1985; Van der Meijden et al., 1989). All declining species were
growing in soils with pH exceeding 5 (Table 4). In Dutch heathlands soil pH nowadays
is usually below 5 (Diemont, 1996). Due to the exceptional soil conditions with Ca rich
loamy layers which maintain relatively high soil pH despite the high deposition of
acidifying compounds, our study area has an almost unique vegetation with many spe-
cies that have strongly declined in other areas. Only the common heathland species C.
vulgaris and E. tetralix and S. cespitosus are growing in soils with low pH. Even the
common grass species Molinia caerulea which is growing all over the nature reserve,
prefers soil pH above 5.
The N/P ratio measured in individual species can dier considerable of that of the total
vegetation in which they are growing (Table 5). Competition for nutrients may aect
the uptake ratio and some species certainly are better competitors for one or more
nutrients (Braakhekke, 1980). Genista anglica (leguminosea) can easily acquire nitrogen
by N
2
xation which is shown by an N/P ratio higher than that of the surrounding vege-
tation. The N/P ratios in plants of Parnassia palustris, Gentiana pneumonante, Succisa
pratensis and Rhynchospora alba are lower than those in the whole vegetation. These
species seem to be relatively strong competitors for phosphorus. Specially G. pneu-
monante and R. alba were found in vegetation with high N/P ratio.
Summarising, we can conclude that increasing acidication can be seen as the most
important threat for all declining species in this data set. Increasing N/P and N/K
ratios have additional negative eect on the abundance of endangered species. Fur-
ther research will focus on the inuence of N/P and N/K ratio on plant growth in
some of the declining species.
References
van Berkel, K., Helden, A. V., & Palm, L. (1999). A history of science in the Netherlands: Survey, themes
and references. (Eds.) Brill: Leiden.
Booth, V. (1984). Communicating in science: Writing a scientic paper and speaking at scientic meetings.
Cambridge: Cambridge University Press.
22 J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524
Burrough-Boenisch, J. (1999). International reading strategies for IMRD articles. Written Communica-
tion, 16(3), 296316.
Burrough-Boenisch, J. (2000). When is the present the past? Why this is important for scientists and
translators. Bulletin of the Institute of Translation and Interpreting(December 2000-January 2001), 25.
Chappell, V. A., & Rodby, J. (1983). Verb tense and ESL composition: a discourse level approach. In
M. A. Clarke, & J. Handscombe (Eds.), On TESOL 82: Pacic perspectives on language learning and
teaching (pp. 309320). Washington: TESOL.
Comrie, B. (1976). Aspect: an introduction to the study of verbal aspect and related problems. Cambridge:
Cambridge University Press.
Comrie, B. (1985). Tense. Cambridge: Cambridge University Press.
Day, R. A. (1995). Scientic English: A guide for scientists and other professionals. Phoenix, AZ: Oryx.
Delaunay, M. (1998). E

loge du pre sent. Annales de Dermatologie et Ve ne re alogie, 125, 567568.

Dietrich, R., Klein, W., & Noyau, C. (1995). The acquisition of temporality in a second language.
Amsterdam: John Benjamins.
Doherty, M., & Smith, R. (1999). The case for structuring the discussion of scientic papers. British
Medical Journal, 318(8 May 1999), 12241225.
Dudley-Evans, T. (1991). Socialisation into the academic community: linguistic and stylistic expectations
of a PhD thesis as revealed by supervisor comments. In P. Adams, B. Heaton, & P. Howarth (Eds.),
Socio-cultural issues in English for academic purposes (pp. 4151). London: Modern English Publica-
tions.
Hannay, M., & Mackenzie, J. L. (1996). Eective writing in English: A resource guide. Groningen: Marti-
nus Nijho.
Hinkel, E. (1997). The past tense and temporal verb meanings in a contextual frame. TESOL Quarterly,
31(2), 289313.
Huth, E. (1999). Writing and publishing in medicine. Baltimore: Williams & Wilkins.
Hyland, K. (1998). Hedging in scientic research articles. Amsterdam: John Benjamins.
James, C. (1998). Errors in language learning and use: Exploring error analysis. London: Longman.
Malcolm, L. (1987). What rules govern tense usage in scientic articles. English for Specic Purposes, 6(1),
3143.
Matthews, J. R., Bowen, J. M., & Matthews, R. W. (1996). Successful scientic writing: A step-by-step
guide for the biological and medical sciences. Cambridge: Cambridge University Press.
McCarthy, M., & Carter, R. (1994). Language as discourse: Perspectives for language teaching. London:
Longman.
Monville-Burton, M., & Waugh, L. R. (1991). Multiple meanings in context: the French present tense. In
J. Gvozdanovic, & T. Janssen (Eds.), The function of tense in texts (pp. 183196). Amsterdam/Oxford/
New York/Tokyo: Koninklijke Nederlandse Akademie van Wetenschappen.
OConnor, M. (1991). Writing successfully in science. London: HarperCollins.
Sanders, M., Tingloo, A., & Verhulst, H. (1992). Advanced writing in English: A guide for Dutch authors.
Leuven/Apeldoorn: Garant.
Swales, J. M. (1990). Genre analysis: English in academic and research settings. Cambridge: Cambridge
University Press.
Swales, J. M., & Feak, C. B. (1994). Academic writing for graduate students: Essential tasks and skillsA
course for nonnative speakers of English. Ann Arbor MI: University of Michigan Press.
Tarone, E., Dwyer, S., Gillette, S., & Icke, V. (1981). On the use of the passive in two astrophysical
journal papers. English for Specic Purposes, 1, 123140.
Taylor, G. (1989). The students writing guide for the arts and social sciences. Cambridge: Cambridge
University Press.
Vann, R. J., Meyer, D. E., & Lorenz, F. (1984). Error gravity: a study of faculty opinions of ESL errors.
TESOL Quarterly, 18(3), 427440.
Weissberg, R., & Buker, S. (1990). Writing up research: Experimental research report writing for students of
English. London: Prentice-Hall.
Wotling, G., Bouvier, C., & Fritz, J.-M. (2000). Regionalization of extreme precipitation distribution
using the principal components of the topographical environment. Journal of Hydrology, 233, 86101.
J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524 23
After obtaining degrees in geography and from Oxford and McGill universities,
Joy Burrough-Boenisch began a career in editing in Australia. Since 1976 she has
worked as a freelance science editor and translator in the Netherlands. She is a
founder member and past chair of the Society of English-Native-Speaking Editors.
Her PhD thesis Culture and Conventions: Writing and Reading Dutch Scientic Eng-
lish will be published in the Netherlands in 2002. Important publications in linguis-
tics are Righting English Thats Gone Dutch (Sdu, The Hague, reprinted twice in
1999) and International reading strategies for IMRD articles (1998) Written
Communication 16(3): 296316.
24 J. Burrough-Boenisch / English for Specic Purposes 22 (2003) 524