Pedagogical uses of monolingual and parallel

by user








Pedagogical uses of monolingual and parallel
Pedagogical uses of monolingual
and parallel concordances
Ana Frankenberg-Garcia
This paper discusses the use of concordances in the classroom, with particular
reference to the pedagogical implications of the differences between parallel
and monolingual concordances. Examples are given of using the two kinds of
concordances in activities that involve language production, reception,
correction, and testing. It is concluded that monolingual and parallel
concordances have non-conflicting, complementary roles to play.
A concordance is a list of the occurrences of a given word, part of a word,
or combinations of words, together with their contexts, within a corpus of
text. Figure 1, which contains concordances from the Web for the search
string a concordance is, illustrates what a concordance is.
A concordance is a list of all the examples of a specified lexical item which
occur in a corpus.
A concordance is a matchless tool for investigating texts.
Selected concordance
lines from WebCorp1 for
a concordance is
Since words express ideas, themes and motifs, a concordance is highly useful
in detecting patterns of meaning as well.
The student must be able to find how a word is used—and for this
a concordance is invaluable.
Just by looking at the above concordance lines, readers can get the gist of
what concordances are. Their potential for language learning is
enormous, and there is already a significant body of literature
demonstrating how language learners can benefit from their use (for
example, Tribble and Jones 1997 and Aston 2001). Apart from being
employed in the compilation of corpus-based dictionaries, grammars and
syllabuses, concordances can also be utilized directly in the classroom, in
an approach to language learning which, since the pioneering work that
appeared in Johns and King (1991), has come to be known as data-driven
This paper contrasts the use of monolingual and parallel concordances in
data-driven language learning.
ELT Journal Volume 59/3 July 2005; doi:10.1093/elt/cci038
Q The Author 2005. Published by Oxford University Press; all rights reserved.
Monolingual concordances such as the ones in Figure 1 can be obtained
not only from the Web, where provenance is not always clear, but also
from carefully designed linguistic corpora like the British National
Corpus (BNC) and COBUILD’s Bank of English.2
It is imperative that language learners and teachers know exactly where
their concordances come from in order to know what to expect from
them. English concordances from the Web, for example, may not be in
standard English: anyone can publish practically anything on the Web.
Concordances from carefully edited linguistic corpora are safer to use,
but may in turn contain language which is too difficult for beginners or
too general for advanced learners researching specialized technical
topics. Users must learn to choose concordances from corpora that
represent the type of language they are after, and even then it may be
necessary to filter out part of the results returned.
Selected parallel
concordances from
Compara for unp
Parallel concordances are concordances from corpora which contain texts
in one language aligned with their translations into one or more
languages. Parallel concordances are therefore bi- or multilingual.
Figure 2 illustrates what a parallel concordance is by showing a selection
of English-Portuguese concordance lines for the English negative
prefix un taken from the Compara parallel corpus of English and
…trying to get some grievance redressed,
or some unjust law altered …
… a procurar reparar uma ofensa
ou modificar uma lei injusta …
… I am an unsuitable person to be a wife.
… sou pouco adequada para ser
… he’s uncommonly well set up.
… é invulgarmente bem parecido.
… he was unknown to four-fifths of
… quatro quintos de Rouen o
Although parallel concordances are normally associated with translation
studies, translator training, the development of bilingual lexicography and
machine translation, several studies have referred to their potential uses in
second language learning (for example, Roussel 1991 and Barlow 2000).
As discussed in Frankenberg-Garcia (forthcoming), however, when using
parallel concordances in the classroom it is important to bear in mind that:
Parallel concordances are based on translations and encourage
learners to compare languages. They can therefore only be appropriate
in the classroom when learners share the same native language, and
when it is helpful for them to use their first language as a tool for
learning the second language.
Parallel concordances provide access to so many comparable facts of
linguistic performance that it is easy to lose sight of the kind of
information that really matters. There is no point in sieving through
parallel corpora to make second language learners focus on language
Ana Frankenberg-Garcia
differences that do not affect their learning. Swamping learners with
unsolicited language contrasts could even be detrimental to learning.4
Unlike monolingual concordances, which present learners with texts
written in a single language, parallel concordances contain not only
two languages, L1 and L2, but also two types of language: source texts
(ST) and translations (TT). When using parallel concordances for
pedagogical purposes it is necessary to understand the implications of
the differences between L1 –L2, L2 –L1, ST–TT, and TT –ST
In addition to these words of caution, it is also important to note
that parallel concordances come from corpora that are representative
only of (a part of) language that has been translated. Language
learners and teachers cannot expect to find in them words or expressions
that are beyond the scope of the texts these comparatively small corpora
Having drawn attention to the need for special care when using parallel
concordances for pedagogical purposes, in the next section my aim is to
discuss how parallel and monolingual concordances serve different
functions and have non-conflicting, complementary roles to play in the
classroom. While some language learning situations call for the use of
monolingual concordances, others can be dealt with more effectively by
means of parallel concordances.
In either case, it is not necessary to have a computer in the classroom.
Learners can use concordances by themselves, on a self-access basis, and
teachers can simply bring concordance printouts to the classroom or,
with the help of a word processor, turn concordance outputs into
Concordances in
language production
When learners are engaged in language production activities, a question
that often arises is:
How do you say ———— in L2?
The question means learners know how to say something in their L1, and
are looking for a way of expressing it in L2. Monolingual concordances
wouldn’t be very useful here, for learners wouldn’t know where to begin
looking. However, used as a complement to (or instead of) bilingual and
language production dictionaries, parallel concordances can help
learners to find foreign words they don’t know and learn in which
contexts they are appropriate. Provided the parallel corpus used contains
sufficient examples of the term the student is researching, an automatic
search for the term in the student’s L1 renders different ways of
expressing it in the L2. The fact that parallel concordances offer not just
linguistic equivalents, but also the contexts in which different terms are
equivalent can help learners decide which term is appropriate.
Portuguese learners of English who look up festa in a PortugueseEnglish bilingual dictionary, for example, will encounter several possible
translations for the word (party, festival, feast, holiday, celebration, etc.) and
may find it difficult to choose which term to use. If they look up festa in
the L1-L2 direction of a Portuguese-English parallel corpus like
Monolingual and parallel concordances
Compara, they will be able to see not only different ways in which festa
has been rendered in English, but also the different contexts in which
each term was used (Figure 3).
Selected parallel
concordances from
Compara for festa
–Está a dar uma festa?–perguntou.
… faz tempo que não vejo Karl Kroop
nas festas da faculdade.
‘Having a party?’ he said.
… you don’t often see Karl Kroop
at faculty social gatherings.
… semelhante ao som de um antigo
sistema de altifalantes numa festa
de aldeia …
… like the sound of an
old-fashioned tannoy system at
an English village fete …
… os ciclos anuais eram marcados
por festas de família …
… annual cycles were punctuated
by family occasions …
Na festa do Corpo de Deus, dizia
ela, toda a aldeia da família se
associava …
At the feast of Corpus Christi, she
said, her family’s whole town
combined …
… os cultos, as festas, as religiões
que floresciam na sua mocidade.
… the cults, festivals and religions
that had flowered in his youth.
Toda a vida ela sonhara a festa.
She had dreamed of a reception
all her life.
… Tu Bisvat é o nome de uma festa
judaica …
… Tu Bisvat being the name of a
Jewish holiday …
Ela transou com o garoto Ritchie
na festa de Ano Novo.
… não se queria meter em festa
alheia …
She had it off with young Ritchie
at the New Year’s Eve do.
… not wishing to interfere in other
people’s celebrations …
Another way in which parallel concordances can be useful in language
production activities is in helping learners come to terms with the fact
that there are certain words in their mother tongue for which there are no
simple, direct translations available. Parallel concordances that go from
source-texts in L1 to translations in L2 can be especially helpful when
learners have to deal with culturally-bound concepts, difficult to
express in the target language, like the Portuguese word saudade in
English (Figure 4).
Another question that frequently arises when writing or speaking in a
second language is:
Is it okay to say ———— in L2?
The most common way of obtaining answers to questions like these is to
ask a native speaker. But native speakers may not always be available or
may or may not have clear answers. Monolingual concordances from
large, linguistically edited corpora like the BNC or COBUILD’s Bank of
English can be very helpful in these circumstances, for they allow
learners to access the combined intuitions of literally thousands of native
speakers together. In contrast, parallel corpora are often not big or
representative enough to provide conclusive evidence in this kind of
Ana Frankenberg-Garcia
Selected parallel
concordances from
Compara for saudade
… vamos sentir saudades de você …
… we shall all miss you …
… e ele sente espasmos de saudade …
… he has a sudden spasm of
homesickness …
… uma ponta de saudade …
… a stab of longing …
Tenho saudades tuas.
I think about you.
… a lembrança da pátria que trouxe
a saudade ao coração pressago.
… the memory of my native land
that brought a longing to my
anxious soul.
Nenhuma água de Juventa igualaria
ali a simples saudade.
No water from Iuventus could
match simple nostalgia in that.
… que saudade deixava?
… what would he leave behind to
… saudade do antigamente.
A yearning for days gone by …
… vinha-me como uma saudade dos
meus tempos ocupados da repartição.
… I felt almost nostalgic for the
days when I was busy at the office.
A student translating a Portuguese menu into English, for example,
knew you roasted meat but baked bread, but was not sure whether she
should say roast or baked fish. The student’s searches for roast, baked and
fish in an English-Portuguese parallel corpus (C ompara) did not
provide her with an answer. The hits for fish only told her it could be
cooked by frying and boiling, and the hits for roast and baked were not
used in relation to fish. Looking up fish in the 100 million-word BNC
Simple Search was not helpful either, for most of the 50 concordances
retrieved were about live fish. However, her search for roast and baked
enabled her to find information on foods that were typically baked and
foods that were typically roasted, and this enabled her to conclude that
baked fish sounded better than proast fish (Figures 5 and 6).5
Selected concordances
from the BNC for roast
Selected concordances
from the BNC for baked
If you wish to roast your own coffee, you should …
The nuts are picked and dried in the sun, and then placed in a fire to roast.
The 32 guests enjoyed roast turkey …
… to be taken with roast lamb…
… generous portions of roast chicken …
Swathes of roast beef and perfect Yorkshire puddings.
… pulling roast chestnuts out of a grate.
… the baked bread tends to collapse on cooling …
ASADILLO (Baked peppers and tomatoes)
… serve sausage and baked beans by the fire.
A baked potato and some porridge, you know.
… wonderful whole baked fish dishes.
… a freshly baked baguette for breakfast.
… artichoke hearts baked with olive oil and garlic …
… plum tarts and custards, crumbles and baked apples.
Monolingual and parallel concordances
When engaged in language production activities, parallel concordances
can help learners express in L2 what they already know how to say in L1,
and monolingual concordances can provide them (and non-native
teachers) with answers to endless questions about language usage which
require the corroboration of a native-speaker’s intuitions.
Concordances in
language reception
Concordances can also help learners better understand the L2. When
reading in a foreign language, L2 –L1 parallel concordances can help
learners to understand foreign words, meanings and grammar that they
are unfamiliar with. Extracting concentrated examples of parts of the
foreign language that they don’t quite understand, matched to equivalent
forms in their mother tongue can help boost language comprehension.
For example, a Portuguese learner of English familiar with the meaning
of contrast of the word however was having difficulty understanding its
‘no matter how’ meaning in the sentence No programme, however good,
can replace the role of the teacher. He was instructed to look up however in a
monolingual English dictionary, a bilingual English-Portuguese
dictionary, an English-Portuguese parallel corpus (Compara), and
a large, monolingual reference corpus of English (the BNC Simple
Search). He was then asked how well each of these resources enabled
him to understand that particular meaning of however and reported that:
The monolingual dictionary (COBUILD , which happens to be corpusbased) helped him understand the term, but it didn’t help him see
what it meant in his native Portuguese.
The bilingual dictionary (Michaelis) explained the meaning of however
sought ‘in a more direct way’, but the translation given, por mais que,
‘didn’t quite fit in the sentence’.
The monolingual concordances for however retrieved from the BNC
were ‘too many and it took too long to read them all’ (Figure 7).6
The parallel concordances from Compara were very helpful because
they showed him ‘different translations for however in different
contexts. There were too many concordances to read, but a quick
glance at the first few was enough’ (Figure 8).7
However, this is not very easy to achieve.
Claud however found it increasingly hard to earn because of his Communist
past and it was left to Patricia to keep the various wolves from the door.
First four monolingual
concordances from the
BNC Simple Search for
The controlling agents of the status quo may know the power of lies; dissident
sub-cultures, however, are closer to knowing their value’; (cited in Bronski,
Culture-Clash, 41).
However, Newton’s theory of light and colours, first published in 1672 and
extended in his Opticks (1704), provided a new direction for colour theorists.
He concluded that the parallel corpus was the most useful resource, but
the bilingual dictionary could save time if it contained some concrete
examples. The Portuguese translations of the term seemed more
important to the learner than explanations and examples in English only.
For learners with similar learning styles, parallel concordances may
assist language comprehension more than monolingual concordances.
Ana Frankenberg-Garcia
First four parallel
concordances from
Compara for however
Concordances in
language correction
… there’s something special about
every single episode of a sitcom,
however trite and formulaic it
may be …
Actors, however, regard writers
with respect, even a certain awe.
… há qualquer coisa de especial em
cada episódio de uma série televisiva,
mesmo que seja muito banal e siga
sempre a mesma fórmula …
Mas os actores têm respeito, ou até
mesmo um certo temor, em relação
aos autores.
But if it doesn’t occasionally touch
on the deeper, darker side of life,
however glancingly, then the
audience won’t believe in the
characters …
Mas, se não tocar ocasionalmente
no lado mais profundo e mais
obscuro da vida, por muito
superficialmente que seja, o público
deixa de acreditar nas personagens…
Recently, however, it’s been converted
into a glazed and tiled atrium …
Contudo, foi recentemente
transformado num átrio com telhado
de vidro e chão de mosaicos…
Concordances can also be used to help learners correct their errors.
Monolingual concordances can be particularly useful in situations when
learners question their teachers, requesting proof of why something was
marked wrong. A Portuguese learner of English wondered why the word
informatics in the sentence pI don’t like informatics very much, was marked
wrong (to say the same in Portuguese, the learner would have used the
word informática, which means both computers and informatics). A
search for informatics (Figure 9), and one for computers (Figure 10) in the
BNC helped the student establish very quickly that computers was more
appropriate than informatics in the context of her sentence.
Looking up informatics and computers in a parallel corpus like Compara
wouldn’t be as helpful. The corpus is much smaller and contains only
fiction, which is not a genre likely to contain many hits for informatics.
Consequently, it does not provide enough examples for highlighting the
difference between the two terms.
Still, parallel corpora can be very effective for dealing with L2 errors that
can be traced back to the influence of the L1. Portuguese learners of
English, for example, often assume that false cognates like resume and
resumir mean the same. Portuguese to English and English to
Portuguese searches in Compara can show these learners that, in an
unmistakable way, the Portuguese resumir (summarize) is not the same
as the English resume (continue) (Figure 11).
The Faculty of Informatics has four departments: Applied Computing,
Computing Science, Information Systems, and Mathematics.
All the Faculty’s students will, by the time they graduate, have been trained to
bring to their work an awareness of informatics as a key force in social change
and economic advance.
Selected concordances
from the BNC for
The Inter-governmental Bureau of Informatics was set up by the
United Nations.
The systematic study of information technology (informatics) is in its infancy...
Monolingual and parallel concordances
Selected concordances
from the BNC for
Selected parallel
concordances from
Compara for resump (EnPt and Pt-En searches)
He resumes work …
Bob Busby continua o seu
trabalho …
He resumes his interrupted journey
to the toilet …
Conclui a viagem até ao lavabo …
He turned resolutely back towards
the hill and resumed his progress
Resolutamente, deu meia volta,
ficou virado para a colina e
retomou o caminho íngreme.
… life in the plaza was about to
… a vida daquela praça estava
prestes a recomeçar de novo.
A televisão resume-se a linhas.
That’s what TV is–all lines.
… para resumir, Morris Zapp não
conseguia imaginar nada que quisesse
e ainda não tivesse alcançado …
…in short, Morris Zapp could
think of nothing he wanted to
achieve that he hadn’t achieved
already …
Assim lhe chamo, porque resume o
universo, e o universo é o homem.
That’s what I call it, because it
sums up the universe and the
universe is man.
Vou aqui resumindo, como posso,
as esperanças de Natividade.
I summarize here, as best I can,
Natividade’s hopes.
An L2 teacher whose students share the same native language is bound
to find that there are many other problems of cross-linguistic influence
that parallel concordances can help sort out.
Concordances in
language exercises
and testing
It is often the case that when language teachers are preparing exercises
or tests for their students, they need to find or invent sentences which
focus on specific words or grammar structures. While invented
sentences often sound artificial and stilted, finding authentic sentences
without the help of a corpus can be a daunting task. Both parallel and
monolingual concordances make it very easy for the teacher to locate
such structures in the context of authentic sentences, which they can
then use to prepare tests or exercises for their students. Figure 12
contains a mix of monolingual concordances from the BNC containing
adverbs of frequency, which can be easily deleted and used as a wordorder exercise that focuses on the position of those adverbs in the
sentence. Figure 13, in turn, contains parallel concordances from
Compara for the Portuguese preposition com. The English part of
Ana Frankenberg-Garcia
Leading me where I have always wanted to go yet never known how.
But often people stayed at home, finding jobs to do around the house.
Selected concordances
from the BNC for always,
never, and often
The reader is never bored–; but could be disappointed that the book does not
go beyond its 120 pages.
Little Billy always told lies to his mother, but they were never convincing.
… milhares de pessoas morriam
diariamente com SIDA em África …
… thousands died daily of AIDS in
Africa ...
… não podia contar com o apoio dele.
… I couldn’t count on any real
support from him.
estou preocupada com o Joe.
‘I’m worried about Joe.’
Selected parallel
concordances from
Compara for com
–Adeus–respondeu Alistair com uma
voz de alívio.
‘Bye,’ Alistair said in a voice loud
with relief.
the concordance can help Portuguese learners of English to see that the
Portuguese preposition com is not always equivalent to with, its most
common English equivalent.
As many language teaching materials available today are meant for
learners of different first language backgrounds, parallel concordances are
particularly useful to meet a demand for language exercises geared
to monolingual settings, where learners share similar L1-related difficulties.
This paper has attempted to show different uses of parallel and
monolingual concordances in second language learning. While in some
situations parallel concordances are more relevant, in others
monolingual ones are more appropriate. It is believed that the two types
of concordances have non-conflicting, complementary roles to play. With
common sense and practice, it is not hard to establish whether or not
concordances are suitable, and if so, which concordances are best for
which situations.
Revised version received February 2004
1 WebCorp allows use of the World Wide Web as a
corpus. Access is online and free at http://www.
2 The BNC offers free online access to its Simple
Search facility at
where any person can access up to 50
concordance lines from a 100 million word
collection of written and spoken British English
from the early nineties. At a small cost, the
complete service is also available online or on
CD-ROM. The Bank of English is made up of
450 million words of contemporary British and
American English (written and spoken). Its
demo facility can be used free of charge at
CorpusSearch.aspx. A paid subscription service
is also available.
3 Free online access to parallel concordances is
available for the following language pairs:
English –Portuguese (þ1.7 million words of
published fiction)
English –French, English–Spanish, English–
German, and English –Danish (World Health
Organization and European Union documents)
English –Slovene (one million words of
technical documents, fiction, and official texts by
the Slovene government)
Monolingual and parallel concordances
English –Chinese (300,000 English words and
500,000 Chinese characters of mostly legal and
documentary texts)
English-Italian (700,000 English words from
the Brown corpus translated into Italian)
It is also possible for language teachers to build
small, home-made parallel corpora using
software like Multiconcord (http://web.bham.
ac.uk/johnstf/lingua.htm) or ParaConc (http://
The Contrastive Analysis Hypothesis was
criticized precisely for this; not all differences
between languages are relevant to second
language learning (Wardhaugh 1970).
The (corpus-based) Oxford Collocations
Dictionary would also have answered this query
Only the 34th of the 50 concordances supplied
contained the ‘no matter how’ meaning of
however. The learner did not get as far as reading
it. His search would have been more successful
had he tried the more sophisticated part-ofspeech query: however þ ADJ.
The fact that the first four concordances gave the
student the answer he needed was fortuitous.
The ‘no matter how’ meaning of however could
have turned up much further down in the list of
concordances retrieved.
Ana Frankenberg-Garcia
Aston, G. (ed.). 2001. Learning with Corpora.
Houston: Athelstan.
Barlow, M. 2000. ‘Parallel texts in language
teaching’ in S. Botley, T. McEnery, and A. Wilson
(eds.). Multilingual Corpora in Teaching and
Research. Amsterdam: Rodopi, 106–115.
Frankenberg-Garcia, A. 2004. ‘Lost in parallel
concordances’ in G. Aston, S. Bernardini, and D.
Stewart (eds.). Corpora and Language Learners.
Amsterdam: John Benjamins.
Johns, T. and P. King (eds.). 1991. Classroom
Concordancing. Birmingham: The University of
Birmingham Centre for English Language
Roussel, F. 1991. ‘Parallel concordances and tonic
auxiliaries’ in T. Johns and P. King (eds.).
Tribble, C. and G. Jones. 1997. Concordancing in the
Classroom: A Resource Guide for Teachers. Houston:
Wardhaugh, R. 1970. ‘The contrastive analysis
hypothesis’. TESOL Quarterly 4: 123 –30.
The author
Ana Frankenberg-Garcia holds a PhD in Applied
Linguistics from the University of Edinburgh, and
is an auxiliary professor at ISLA, in Lisbon. She is
joint project leader of the COMPARA corpus, and
her current research interests focus on the use of
corpora for language learning and translation
Email: [email protected]
Fly UP